“This book is both timely and important. It is also highly educational. It can help us better understand

the opportunities and challenges renewable energy creates and learn from the experience of others. In

this way, we will all, whether in politics, business, education, or at home, be better equipped to make,

with confidence, the vital decisions that will lead to the energy system we seek for the future—a system

that is at once secure, sustainable, and economically robust.”

—From the Foreword by GÜNTHER OETTINGER, European Union Energy Commissioner

“A comprehensive and timely review of the international experience in fostering the use of renewable

energy sources in the electricity industry. Understanding the political, business, and technological

drivers behind the clean energy revolution is important as many nations aim to enable low-carbon

development paths that will bring economic prosperity to their citizens, while at the same time

reducing human impacts on the environment.”

—BLAS PÉREZ HENRÍQUEZ, executive director, Center for Environmental Public Policy at the Goldman School of Public Policy, University of California, Berkeley

“This book should be devoured by thoughtful policymakers considering how renewables will fit into

decarbonized electric systems. Bringing varied perspectives on this fundamental question, these

scholars make a provocative and informative contribution to a topic of critical importance to

governments around the world.”

—SUE TIERNEY, managing principal, Analysis Group, Inc., USA

Reflecting its reliance on fossil fuels, the electric power industry produces the majority of the world's greenhouse gas emissions. The need for a revolution in the industry becomes further apparent given that "decarbonization" means an increasing electrification of other sectors of the economy—in particular, through a switch from gasoline to electric vehicles. Of the options for producing electric power without significant greenhouse gas emissions, renewable energy is most attractive to policymakers, as it promises increased national self-reliance on energy supplies and the creation of new industries and jobs, without the safety and political concerns of nuclear power or the unproven technology of carbon capture and storage.

Drawing on both economic theory and the experiences of the United States and EU member states, Harnessing Renewable Energy in Electric Power Systems addresses the key questions surrounding renewable energy policies. How appropriate is the focus on renewable power as a primary tool for reducing greenhouse gas emissions? If renewable energy is given specific support, what form should that support take? What are the implications for power markets if renewable generation is widely adopted? Thorough and well evidenced, this book will be of interest to a broad range of policymakers, the electric power industry, and economists who study energy and environmental issues.

BOAZ MOSELLE is a director and JORGE PADILLA a managing director at LECG, an international consulting group focusing on “economic and financial analyses to provide objective opinions and advice regarding complex disputes and inform legislative, judicial, regulatory, and business decision makers.” RICHARD SCHMALENSEE is dean emeritus and professor of economics and management at MIT's Sloan School of Management, and director of the MIT Center for Energy and Environmental Policy Research.

RFF Press strives to minimize its impact on the environment

Energy / Economics / Climate

www.rffpress.org

HARNESSING RENEWABLEENERGY in Electric Power Systems

HA

RN

ES

SIN

G R

EN

EW

AB

LE

EN

ER

GY

in E

lectric

Po

we

r Syste

ms T

heory, Practice, Policy

E D I T E D B Y

Boaz MoselleJorge PadillaRichard Schmalensee

Boaz M

oselleJorg

e PadillaR

ichard Schm

alensee

Co

ver im

ag

e: “S

ola

r Pan

el” ©

Lu

is Ped

rosa

/istockp

ho

to.c

om

HARNESSINGRENEWABLE

ENERGYin Electric Power Systems

HARNESSINGRENEWABLE

ENERGYin Electric Power Systems

Theory, Practice, Policy

Edited by

Boaz Moselle, Jorge Padillaand Richard Schmalensee

Washington, DC • London

First published in 2010 by RFF Press, an imprint of Earthscan

Copyright © Earthscan 2010

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in anyform or by any means, electronic, mechanical, photocopying, recording, or otherwise, except as expressly permittedby law, without the prior, written permission of the publisher.

Earthscan LLC,1616 P Street, NW, Washington, DC 20036, USAEarthscan Ltd, Dunstan House, 14a St Cross Street, London EC1N 8XA, UK

Earthscan publishes in association with the International Institute for Environment and Development

For more information on RFF Press and Earthscan publications, see www.rffpress.org and www.earthscan.co.uk orwrite to earthinfo@earthscan.co.uk

ISBN: 978-1-933115-90-0

Copyedited by Joyce BondTypeset by KerryPress LtdCover design by Circle Graphics

Library of Congress Cataloging-in-Publication Data

Harnessing renewable energy in electric power systems : theory, practice, policy/edited by Boaz Moselle, JorgePadilla, and Richard Schmalensee.

p. cm.Includes bibliographical references and index.ISBN 978-1-933115-90-0 (hardback : alk. paper)

1. Electric power production. 2. Electric power production—Environmental aspects. 3. Electric powerproduction—Economic aspects. 4. Renewable energy sources. I. Moselle, Boaz. II. Padilla, Jorge 1983– III.Schmalensee, Richard.TK1001.H37 2010333.793’2--dc22

2010017827

A catalogue record for this book is available from the British Library

At Earthscan we strive to minimize our environmental impacts and carbon footprint through reducing waste,recycling and offsetting our CO2 emissions, including those created through publication of this book. For moredetails of our environmental policy, see www.earthscan.co.uk.

Printed and bound in the UK by MPG Books

The paper used is FSC certified

The findings, interpretations, and conclusions offered in this publication are those of the authors.They do notnecessarily represent the views of Resources for the Future, its directors, or its officers.

About Resources for the Futureand RFF Press

Resources for the Future (RFF) improves environmental and natural resource policymaking world-wide through independent social science research of the highest caliber. Founded in 1952, RFF pio-neered the application of economics as a tool for developing more effective policy about the use andconservation of natural resources. Its scholars continue to employ social science methods to analyzecritical issues concerning pollution control, energy policy, land and water use, hazardous waste, climatechange, biodiversity, and the environmental challenges of developing countries.

RFF Press supports the mission of RFF by publishing book-length works that present a broad range ofapproaches to the study of natural resources and the environment. Its authors and editors include RFFstaff, researchers from the larger academic and policy communities, and journalists. Audiences for publi-cations by RFF Press include all of the participants in the policymaking process—scholars, the media,advocacy groups, NGOs, professionals in business and government, and the public. RFF Press is animprint of Earthscan, a global publisher of books and journals about the environment and sustainabledevelopment.

Resources for the FutureDirectors

Vicky A. BaileyTrudy Ann Cameron

Preston ChiaroW. Bowman Cutter

John M. DeutchMohamed T. El-Ashry

Daniel C. Esty

Linda J. FisherKathryn S. FullerDavid G. Hawkins

Deborah HechingerPeter R. KaganRubén KraiemFrank E. Loy

Michael A. MantellPeter J. Robertson

Richard SchmalenseeRobert N. Stavins

Joseph StiglitzMark R. Tercek

OfficersLawrence H. Linden, ChairSteven W. Percy, Vice ChairPhilip R. Sharp, President

Edward F. Hand, Vice President–Finance and AdministrationLaurel L. Harvey, Vice President–Development and Corporate Secretary

Mark A. Cohen, Vice President–Research

Editorial Advisors for RFF PressWalter A. Rosenbaum, University of Florida

Jeffrey K. Stine, Smithsonian Institution

Contents

About the Contributors ixList of Figures andTables xiiiAcknowledgments xixAcronyms and Abbreviations xxi

Foreword xxviiGünther Oettinger

Chapter 1. Toward a Low-Carbon Future in Electricity? 1Boaz Moselle, Jorge Padilla, and Richard Schmalensee

Part I. Technology 5

Chapter 2. Renewable Energy Technologies for Electricity Generation 7Godfrey Boyle

Part II. Renewables, Climate Change, and Energy Policy 31

Chapter 3. Renewables Forecasts in a Low-Carbon World: A Brief Overview 33ErinT. Mansur

Chapter 4. Renewable Generation and Security of Supply 51Boaz Moselle

Chapter 5. Market Failure and the Structure of Externalities 69Kenneth Gillingham and James Sweeney

Chapter 6. Renewable Energy, Energy Efficiency, and Emissions Trading 93José Goldemberg

Part III. Renewable Generation and Electric Power Markets 111

Chapter 7. Electricity Wholesale Market Design in a Low-Carbon Future 113WilliamW. Hogan

Chapter 8. Energy Regulation in a Low-Carbon World 137Richard Green

Chapter 9. Building Blocks: Investment in Renewable and NonrenewableTechnologies 159James Bushnell

Chapter 10. Developing a Supergrid 181Christian von Hirschhausen

Part IV. National Experiences 207

Chapter 11. Renewable Electricity Generation in the United States 209Richard Schmalensee

Chapter 12. The European Union’s Policy on the Development of RenewableEnergy 233Christopher Jones

Chapter 13. UK Renewable Energy Policy since Privatization 253Michael G. Pollitt

Chapter 14. Experience with Renewable Energy Policy in Germany 283HannesWeigt and Florian Leuthold

Chapter 15. Renewable Electricity Support: The Spanish Experience 309Luis Agosti and Jorge Padilla

Conclusions: Whither Renewable Generation? 327Boaz Moselle, Jorge Padilla, and Richard Schmalenesee

Index 335

viii Harnessing Renewable Energy in Electric Power Systems

About the Contributors

Luis Agosti is a senior consultant at LECG, an economic and financial consulting firm, where heprovides research and analysis in the field of applied microeconomics, including the economics ofcompetition, regulation, and finance. He also belongs to the Energy Practice. During his years inconsultancy, he has specialized in the electricity and gas sectors. His work has been focused in the analysisof regulation in the electricity and natural gas sectors, with special emphasis in the design of the wholesalepower market in Spain.

Godfrey Boyle is a professor of renewable energy and director of the Energy and Environment ResearchUnit in the Faculty of Mathematics, Computing, and Technology at the UK Open University. He is alsoa visiting professor at The Energy and Resources Institute University in New Delhi. He is coauthor andeditor of Renewable Energy: Power for a Sustainable Future (2004) and Renewable Electricity and the Grid(Earthscan 2009).

James Bushnell is an associate professor and the Cargill Chair in Energy Economics in the Departmentof Economics at Iowa State University, the director of the university’s Biobased Industry Center, and aresearch associate of the National Bureau of Economic Research. He has served as a member of theMarket Monitoring Committee of the California Power Exchange and is currently a member of theMarket Surveillance Committee of the California Independent System Operator. He has testified onregulatory and competition policy issues before numerous state and federal regulatory and legislativeinstitutions and consulted on energy issues throughout the United States and internationally. He haswritten extensively on the regulation, organization, and competitiveness of energy markets.

Kenneth Gillingham is an environmental and energy economist who worked for several years inWashington, DC, at the White House Council of Economic Advisers, Resources for the Future, and theJoint Global Change Research Institute. Following this, he was a Fulbright Fellow at the University ofAuckland in New Zealand and is currently at Stanford University. He has published journal articles ontopics such as solar energy policy, biofuels policy, energy efficiency, and transportation policy.

José Goldemberg, a physicist, is a professor emeritus of the University of São Paulo, Brazil, of which hewas the rector between 1986 and 1990. He has served as president of the Energy Company of the State ofSão Paulo and Brazil’s secretary of state for science and technology. He chaired the World Energy

Assessment of the United Nations Development Programme. Time magazine honored him as one of its“Heroes of the Environment,” and the Asahi Glass Foundation of Japan awarded him its 2008 Blue PlanetPrize.

Richard Green is the director of the Institute for Energy Research and Policy and a professor of energyeconomics in the Department of Economics at the University of Birmingham. He has held visitingpositions at the Office of Electricity Regulation, University of California Energy Institute, and Massa-chusetts Institute of Technology. He is also an author or coauthor of numerous articles published in suchjournals as the Journal of Political Economy, Journal of Industrial Economics, Oxford Economic Papers, and FiscalStudies.

Christian von Hirschhausen is a full professor of economic and infrastructure policy at Berlin Univer-sity of Technology and a research professor at DIW Berlin. Before that, he held the chair of energyeconomics and public sector management at Dresden University of Technology. He is the scientificdirector of several programs that cover a wide range of research in energy and related fields. His recentresearch on contracts and institutions in the utilities was published in The Energy Journal. Additionalresearch focuses on the European economy, energy policy, and economic policy in transitional countries.

William W. Hogan is the Raymond Plank Professor of Global Energy Policy at the John F. KennedySchool of Government, Harvard University, and the research director of the Harvard Electricity PolicyGroup, which is examining alternative strategies for a more competitive electricity market. He has been amember of the faculty of Stanford University, where he founded the Energy Modeling Forum, and is apast president of the International Association for Energy Economics. His current research focuses onmajor energy industry restructuring, network pricing and access issues, market design, and energy policyin nations worldwide.

Christopher Jones was the director of New and Renewable Sources of Energy, Energy Efficiency andInnovation at the Directorate-General for Energy and Transport, European Commission, until January15, 2010, when he became the head of the cabinet of Commissioner Andris Piebalgs, responsible for EUdevelopment policy. He is also a visiting professor of law at the College of Europe in Bruges, Belgium. Hehas authored or coauthored and edited several books and numerous articles on EU energy law andcompetition policy. He is also a member of the advisory or editorial boards of a number of academicjournals.

Florian Leuthold is a research associate at the chair of energy economics at Dresden University ofTechnology. His current research topics include renewable energy, environmental policies, and electricitymarket modeling. He coauthored several recent journal articles published in Energy Economics, EnergyPolicy, and Utilities Policy.

Erin T. Mansur is an associate professor in the Department of Economics at Dartmouth College and afaculty research fellow at the National Bureau of Economic Research. He has coauthored numerousarticles published in economics journals, including the American Economic Review, Review of Economics andStatistics, Journal of Law and Economics, Journal of Environmental Economics and Management, Journal ofIndustrial Economics, and Journal of Economics and Management Strategy.

Boaz Moselle is an economist specializing in energy markets and a director of the economic andfinancial consulting firm LECG. He was previously a managing director at the UK energy regulatorOfgem. He has served as a consultant to governments and corporations in many countries on a widerange of energy policy, competition, and regulatory matters; testified before UK parliamentary commit-

x Harnessing Renewable Energy in Electric Power Systems

tees on renewable generation, security of supply, and the structure of the British power market; andparticipated in a group advising the European Commission on energy, competitiveness, and environmen-tal issues. He is a member of the Advisory Board of the Centre for Economic Learning and SocialEvolution (ELSE) at University College London. He has written books and articles for scholarly journals,including the Electricity Journal; Journal of Risk and Uncertainty; Journal of Law, Economics and Organization;Journal of Economic Theory; Northwestern Journal of International Law & Business; Economic History Review;Peace Economics, Peace Science and Public Policy; and Proceedings of the Edinburgh Mathematical Society.

Jorge Padilla is the European chief executive officer of LECG and advises companies on a variety ofcompetition policy and regulatory issues, covering a wide range of industries, including the electricityand natural gas sectors. He is a research fellow of the Centre for Economic Policy Research and theCentro de Estudios Monetarios y Financieros and is or has been a member of the editorial boards ofCompetition Policy International, Review of Economic Studies, Spanish Economic Review, and InvestigacionesEconómicas. He coauthored The Law and Economics of Article 82 EC (2006) and has written numerousarticles on competition policy and industrial organization, published in such journals as the AntitrustBulletin, Economic Journal, European Economic Review, International Journal of Industrial Organization, Journal ofEconomics and Management Strategy, Journal of Economic Theory, RAND Journal of Economics, Review ofFinancial Studies, and World Competition.

Michael G. Pollitt is an assistant director of the Economic and Social Research Council (ESRC)Electricity Policy Research Group, as well as a reader in business economics at Judge Business School,University of Cambridge, the director of studies in management and economics, and a fellow of SidneySussex College. He is currently an external economic adviser to Ofgem, the UK energy regulator, andhas also advised the ESRC, Norwegian Research Council, DTI, World Bank, and European Commis-sion. Additionally, he has consulted for National Grid, AWG, Eneco, Nuon, Roche, and TenneT.

Richard Schmalensee is the John C. Head III Dean Emeritus and the Howard W. Johnson Professor ofEconomics and Management of the Sloan School of Management at the Massachusetts Institute ofTechnology (MIT) and director of the MIT Center for Energy and Environmental Policy Research. Hehas served as a member of the President’s Council of Economic Advisers and serves on the NationalCommission on Energy Policy and the National Academy of Sciences Committee on America’s ClimateChoices. He is an expert on regulation, antitrust, and energy and environmental policy.

James Sweeney is a director of the Precourt Energy Efficiency Center and a professor of managementscience and engineering at Stanford University. His professional activities focus on economic policy andanalysis, particularly in energy, natural resources, and the environment. He has also advised in a number ofenergy litigations in natural gas, oil, and electricity industries in the United States and New Zealand. Hewas one of the editors of the three-volume Handbook of Natural Resource and Energy Economics (1985–1993).

Hannes Weigt is a research associate at the chair of energy economics at Dresden University ofTechnology. His current research topics include renewable energy and environmental policies. He is alsoa coauthor of numerous journal articles published in Energy Economics, Energy Policy, Applied Energy, andthe Electricity Journal.

About the Contributors xi

List of Figures and Tables

Figures

2.1 Evolution of price of PV modules and systems in selected IEA reporting countries,1997–2007, allowing for the effects of inflation 15

2.2 Projected convergence of utility prices and PV generation costs 152.3 Incremental investment costs of a 20% wind contribution to U.S. electricity

demand in 2030 182.4 Annual offshore wind capacity buildup in the United Kingdom, Sweden,

Denmark, the Netherlands, and Germany, plus total cumulative installed capacity,1989–2009 19

2.5 Marginal resource costs for different levels of renewable generation by sector in theUnited Kingdom in 2020 22

2.6 Day-ahead forecast of wind power output compared with monitored output,for one month in Germany 26

3.1 Wind power worldwide capacity and annual growth rate, 1997–2008 353.2 Share of renewable electricity by EU country in 2007 and targets for 2010 373.3 Expected electricity generation and capacity from nonhydropower renewable

sources with and without ARRA, 2006–2030 434.1 World coal reserves 544.2 World uranium resources 544.3 World gas reserves: top 10 countries 554.4 Capacity credit values 615.1 Economies of scale 765.2 Illustrative incremental benefits from additional cumulative installations: LBD 876.1 World primary energy supply shares of 516 exajoules 936.2 World energy consumption, 1850–2000 946.3 World fossil-fuel CO2 emissions and growth rate, 1990–2005 95

6.4 Contributions of greenhouse gases to global warming in 2006 966.5 Annual growth rates of renewable energy capacity, 1998–2008 976.6 Contribution of renewables to the total energy supply 986.7 Projection of renewable energy production up to 2030 986.8 Energy produced from distributed generation in European countries 996.9 Evolution of energy intensity in a number of countries and regions, 1990–2005 1006.10 Energy savings in the OECD countries, 1973–1998 1016.11 Carbon intensity: CO2 emissions per GDP (PPP), 1970–2005 1016.12 The evolution of industrial subsector energy intensities, 1974–1998 1026.13 Carbon price, January 2006–November 2007 1046.14 Cap and ETS emissions in the European Union 1056.15 Distribution of CDM projects by type 1076.16 Carbon market estimates, 2002–2008 1076.17 Investments in green stimuli 1086.18 Sector breakdown of global green stimuli 1088.1 Models of regulation 1409.1 The four WECC subregions 1649.2 CEMS load and wind production for August 1669.3 CEMS load and wind production for December 1679.4 Annual distribution of CEMS load net of wind 1679.5 Energy market prices for August 1699.6 Energy market prices for December 1699.7 Equilibrium capacity for energy-only market 1709.8 Highest 150 prices CA 1729.9 Highest 150 prices RMPP 1729.10 Equilibrium capacity investment with $25/ton CO2 17510.1 Solar radiation in the United States 18410.2 JCSP’08 wind energy scenario conceptual transmission overlay 18610.3 Integrating 200,000 MW of renewable energy in the U.S. grid

(exemplified for wind power) 18810.4 Airtricity’s European offshore supergrid proposal 18810.5 The North Sea wind power transmission ring 18910.6 The Desertec HVDC overlay network (physical map) 19010.7 Installed capacities for the BAU and TD scenarios, 2010–2050 19210.8 HVDC grid in 2050 (TD scenario) 19510.9 Electricity price changes with reinforced interconnectors to Scandinavia in 2050 19610.10 Required generation and transmission investment for the EEA-MENA supergrid 19611.1 Nonhydro renewable energy consumption as a percentage of total U.S. energy

consumption, 1978–2008 21011.2 Share of nonhydro renewable electricity in total generation, 1989–2006 21111.3 U.S. electricity generation from nonhydro renewable energy, 1990–2007 21111.4 Federal expenditure on R&D in renewable technologies, 1978–2007 21311.5 Wind electricity capacity addition, 1994–2007 21611.6 Nonhydro renewables in California, 1990–2007 221

xiv Harnessing Renewable Energy in Electric Power Systems

11.7 Electricity generation by nonhydro renewable sources in California, 1990–2007 22211.8 Electricity generation by nonhydro renewable sources in Texas, 1990–2007 22314.1 Primary energy consumption in Germany, 2007 28414.2 Energy R&D funding in Germany, 1974–2007 28514.3 Wind turbine installations, 1990–2008 28514.4 Electricity generation from RES, 1990–2007 28714.5 Share of RES from total power generation in 2030 for different future scenarios 28814.6 Annual average wind output, 2000–2004, and expected revenues

for onshore wind 29014.7 Estimated differential costs and avoided external costs as a result of German

RES production 29214.8 Installed RES generation capacities, 1990–2007 30314.9 Cost reduction for RES generation 30315.1 Effectiveness and efficiency of support schemes for onshore wind generation

in Europe in 2006 31015.2 Renewables mix in EU countries, 2007 31115.3 Evolution of demand and share of the SR 31215.4 Distribution of SR production between fixed price and market options 31315.5 Renewable energy growth till 2020 and contribution to final energy consumption 31515.6 Effect of renewable energy promotion in the wholesale generation market 31815.7 Actual versus forecast generation from wind, November 2, 2008 31915.8 Actual versus forecast generation from wind, January 23–24, 2008 320

Tables

2.1 Annual PV capacity installed/predicted: market results versus Solar Generation (SG)scenario predictions since 2001 16

2.2 Projected solar PV electricity output and detailed associated projections for 2030in the SG V scenario 16

2.3 Projected technically available, environmentally constrained, and economicallycompetitive potential for wind energy in 27 EU countries in 2020 and 2030 18

2.4 Prices paid under proposed UK feed-in tariffs for electricity from 2010 and annualdigression rates from 2010 onward 23

2.5 Summary of economic characteristics of renewable electricity generationtechnologies 24

3.1 Total installed capacity of wind energy in 2007 and 2008 353.2 Capacity of renewable resources by type and country in 2008 363.3 IEA world energy outlook predictions, showing generation by energy source 403.4 Electricity generation by fuel type for the three WETO scenarios 413.5 EPA analysis of the Waxman–Markey Bill, showing generation by energy source,

2020 and 2025 443.6 A summary of model predictions for 2030 464.1 EU import dependence, 2005 524.2 Energy consumption and import dependence by 2020 585.1 Some potential policy instruments 80

List of Figures andTables xv

5.2 Sources of market failure and some illustrative potential policy instruments 816.1 Electricity, heat, and transport fuel from renewable sources 966.2 Primary energy production from new renewables 976.3 Additional rate of investment in power generation technologies needed in the

electricity sector every year until 2050 996.4 The carbon market: Volumes and values for emissions trading, 2006–2007 1006.5 Distribution of JI projects by type 1066.6 CDM projects grouped by type 1066.7 Volumes and values for CDM and JI, 2006–2007 1068.1 EU electricity generation from renewable sources 1439.1 Summary statistics of demand 1659.2 Aggregate generation levels 1669.3 Thermal generation costs from EIA 1689.4 Thermal generation costs used in simulations 1689.5 Equilibrium results for energy-only market 1709.6 Capacity payments resulting from $1,000 price cap 1719.7 Investment levels with a capacity market 1739.8 Summary of the revenues of the hypothetical average wind turbine 1749.9 Wind revenues with carbon price at $25/ton 1759.10 Impact of intermittency on average thermal costs 17610.1 Characteristics of high-voltage transmission technologies 18310.2 Studies for large-scale integration of renewable energy sources 18410.3 JCSP’08 main scenario assumptions and results 18610.4 Exogenous and endogenous parameters of the model 19210.5 Number of HVDC transmission cables and CSP generation capacity in

MENA (TD scenario) 19411.1 Actual and potential NHR generation in the United States 21211.2 Leading NHR generation states 21211.3 Subsidies and support to electricity production by technology, 2007 21711.4 State renewable portfolio standards (RPSs) 21811.5 State voluntary renewables goals 21911.6 Other state policies to promote NHR generation 22011.7 Leading wind generation states 22512.1 Member state targets 24013.1 Estimates of the likely economic potentials for different renewable technologies

in the United Kingdom 25513.2 Examples of estimated costs of technologies for the United Kingdom in 2005 25613.3 Costs of electricity sector decarbonization to 2020 25713.4 RO targets and delivery against targets 26013.5 Banding of ROCs from April 1, 2009 26113.6 Financial support for renewables in the United Kingdom (nominal) 26213.7 Support for renewable energy in 2007–2009 26313.8 Renewable electricity generation in the United Kingdom, 1990–2008 26413.9 Differences among leading wind countries in Europe 270

xvi Harnessing Renewable Energy in Electric Power Systems

14.1 Gross employment figures of the RES sector in Germany 29514.2 Impact of emissions trading and RES support on the German economy 29614.3 Feed-in tariffs in Germany to EEG 2008 and EEG 2004 29815.1 Generation mix in Spain, December 2008 31115.2 Evolution of installed renewable and other SR capacity 31215.3 Evolution of incentives for renewable generation 31415.4 Capacity targets and current deployment status, September 2009 31615.5 Implicit subsidies for the SR, 2008 316

List of Figures andTables xvii

Acknowledgments

This book would not have been possible without the support of Luis Agosti, David Black, and DhirenPatki. Special thanks go to Anne Layne-Farrar.They are owed a substantial debt of gratitude.Thanks alsogo to Don Reisman at Resources for the Future, who published the English version of the book, and JuanJose Pons from Marcial Pons, in charge of the Spanish edition of the book, for their support. The viewsand opinions expressed in the chapters of this book are the sole responsibility of their authors.

List of Acronyms and Abbreviations

AC alternating currentACER Agency for the Cooperation of Energy RegulatorsAP average participationARRA American Recovery and Reinvestment Act of 2009a-Si amorphous siliconAWEA American Wind Energy AssociationAZNM southwest (Arizona–New Mexico) region (of WECC)BAU business-as-usualBEITC Business Energy Investment Tax CreditBMU Federal Ministry for the Environment, Nature Conservation and

Nuclear Safety of GermanyBoS balance-of-systemBPA Bonneville Power AdministrationBtu British thermal unitBWEA British Wind Energy AssociationCA California region (of WECC)CAES compressed-air energy storageCAISO California ISOCBA cost–benefit analysis; cost–benefit analysesCCC Committee on Climate ChangeCCGT combined cycle gas turbineCCS carbon capture and storageCDM Clean Development MechanismCdTe cadmium tellurideCECRE Centre for the Control of the Special RegimeCEMS continuous emissions monitoring systemCER certified emission reductionCERT Carbon Emissions Reduction Target

CH4 methaneCHP combined heat and powerCIGS copper indium gallium diselenideCIS copper indium diselenideCO2e carbon dioxide equivalentCPUC California Public Utilities CommissionCRW combustible renewables and wasteCSC current source converterCSP concentrated solar powerDC direct currentDLR German Aerospace CenterDNA designated national authorityDNI direct normal irradianceDOE (U.S.) Department of EnergyEEA European Environment AgencyEEA-MENA European Economic Association–Middle East North AfricaEEG Erneuerbare Energien Gesetz (Renewable Energy Source Act)EEX European Energy ExchangeEIA (U.S.) Energy Information AdministrationEJ exajouleENTSO-E European Network of Transmission System Operators for ElectricityEPA (U.S.) Environmental Protection AgencyEPIA European Photovoltaic Industries AssociationERU emission reduction unitETS (EU) Emission Trading SystemEUMENA Europe, the Middle East, and North AfricaEV electric vehicleEVA Energy Ventures AnalysisFACTS Flexible AC Transmission SystemsFERC Federal Energy Regulatory CommissionFIT feed-in tariffFTR financial transmission rightGaAs gallium arsenideGDP gross domestic productGHG greenhouse gasGt gigatonsGW gigawattsGWEC Global Wind Energy CouncilGWh gigawatt-hoursGWP global warming potentialGWth gigawatts-thermalHFCs hydrofluorocarbonsHR heat rateHVAC high-voltage alternating current

xxii Harnessing Renewable Energy in Electric Power Systems

HVDC high-voltage direct currentIEA International Energy AgencyIER Institute of Energy Economics and the Rational Use of EnergyIET international emissions tradingIGBT insulated gate bipolar transistorsIGCC integrated gasification combined cycleIHSGI IHS Global InsightIPCC Intergovernmental Panel on Climate ChangeIPP independent power producerISO independent system operatorITC investment tax creditJ jouleJCSP’08 Joint Coordinated System Plan 2008JI joint implementationkgCO2/kWh kilograms of CO2 per kilowatt-hourkWh kilowatt-hourkWy kilowatt-yearLBD learning by doingLMP locational marginal pricingLNG liquefied natural gasLPG liquefied petroleum gasL-RES large-scale renewable energy sourcesm/s meters per secondMAP Market Incentive ProgramMAPP Mid-Continent Area Power Poolmcm/day million cubic meters per dayMENA Middle East and North African regionMISO Midwest ISOMSP Mediterranean Solar PlanMtoe million tons of oil equivalentMW megawattsMWh megawatt-hourMWhRES megawatt-hour of renewable energyN2O nitrous oxideNEMS National Energy Modeling SystemNERC North American Electric Reliability CouncilNETA New Electricity Trading ArrangementsNFFO Non-Fossil Fuel ObligationNFPA Non-Fossil Purchasing AgencyNHR nonhydro renewableNIMBY not-in-my-backyardNOx nitrogen oxidesNPV net present valueNREL National Renewable Energy Laboratory

List of Acronyms and Abbreviations xxiii

NWPP Northwest Power Pool region (of WECC)NYISO NewYork Independent System OperatorO&M operation and maintenanceOECD Organisation for Economic Co-operation and DevelopmentOGT open-cycle gas turbinesOMA Office for Metropolitan ArchitectureOR Régimen Ordinario (Ordinary Regime)ORC organic Rankine cycleOTC over-the-counterOTEC ocean thermal energy conversionOWC oscillating water columnPER Plan de Energías Renovables (Renewable Energy Plan)PFCs perfluorocarbonsPFER Plan de Fomento de las Energías Renovables (Plan to Promote

Renewable Energy)PJM Pennsylvania–New Jersey–Maryland InterconnectionPOLES Prospective Outlook on Long-term Energy Systemsppm parts per millionPTC production tax credit; parabolic trough concentratorPURPA Public Utilities Regulatory Policies Act (of 1978)PV photovoltaicPVPS Photovoltaic Power SystemsR&D research and developmentRD Royal DecreeRD&D research, development, and demonstrationRDL Royal Decree-LawREC renewable energy certificate; renewable energy creditREE Red Eléctrica de EspañaREPI Renewable Energy Production IncentiveREPTC Renewable Electricity Production Tax CreditRES renewable electricity standardRGGI Regional Greenhouse Gas InitiativeRHI Renewable Heat IncentiveRMPP Rocky Mountain Power Pool region (of WECC)RO Renewables ObligationROC Renewables Obligation CertificateRPS renewable portfolio standardRRETC Residential Renewable Energy Tax CreditRTO regional transmission organizationRTP real-time pricingSEGS Solar Energy Generating SystemsSERC Southeastern Electric Reliability CouncilSF6 sulfur hexafluorideSG Solar Generation

xxiv Harnessing Renewable Energy in Electric Power Systems

SGIP Self Generation Incentive ProgramSHPPs small hydroelectric power plantsSO system operatorSO2 sulfur dioxideSOx sulfur oxidesSPP Southwest Power PoolSR Régimen Especial (Special Regime)StrEG Stromeinspeisegesetz (Electricity Feed-In Act)tcf trillion cubic feettcm trillion cubic meterstCO2 metric ton of CO2

TD technological developmentTEAM transmission expansion assessment methodologyTGC tradable green certificateTOU time-of-useTREC Trans-Mediterranean Renewable Energy CooperationTSO transmission system operatorTVA Tennessee Valley AuthorityTWh terawatt-hoursUCTE Union for the Co-ordination of Transmission of ElectricityUNFCCC United Nations Framework Convention on Climate ChangeVSC voltage source converterWECC Western Electricity Coordinating CouncilWECS wind energy conversion systemWEM World Energy ModelWETO World EnergyTechnology Outlook—2050WTO World Trade Organization

List of Acronyms and Abbreviations xxv

Foreword

Energy is the foundation stone of prosperity, security, and peace. Yet the European Union’s economicdependence on fossil fuels, of which the lion’s share is imported, is exposing the EU to threats fromclimate change, global geopolitics, and resource competition from rising economies. These risks couldundermine our industrial and social model and ultimately European stability itself.

In response to these challenges, the 27 member states of the European Union unanimously agreed in2007 to binding targets to change the basis of Europe’s energy supply. By 2020, the EU must reduce itsgreenhouse gas emissions by 20%—and by 30% when the conditions are right—and increase the share ofrenewable energy in energy demand to 20%. Energy efficiency must improve by one-fifth. These targetsare at the heart of the European Commission’s broad economic strategy for the year 2020, known asEurope 2020.

Meeting these targets not only is essential for the climate change challenge, but also will dramaticallyimprove security of supply by making the EU economy more efficient, decoupling growth from resourceuse, and creating more than 1.5 million extra jobs.The EU economy will be strengthened, with savings ofat least €60 billion ($81.6 billion) in decreased oil and gas imports, which can be invested in the domesticeconomy.

To do this, we must—within just a few years—generate more than one-third of our power fromrenewable sources of electricity. A significant part of our heating and transport must be based on renew-able fuels. It is not enough to tinker around the edges. We need huge, practical, and concrete initiativesinvolving society as a whole. And we need to persuade our international partners to follow the same track.

The benefits are easy to see: lower imports, new jobs, cleaner air, more stable international energymarkets and prices, lower energy bills, greater consumer empowerment. Renewable energy offers solu-tions to many of our problems—security, economic, and environmental.The switch to an economy witha large share of renewable and low-carbon sources is the only way to ensure sustainable economic growththat brings benefits to all parts of society.

Never before have businesses and consumers had to face such a daunting task. New energy sourcescall for new networks to bring them to customers. Intermittent generation requires new approaches tobalancing demand and supply. New and renewable technologies must become much more efficient andcommercially attractive. At the same time, we have to deal with economic uncertainty and growing globalcompetitiveness.

It is unprecedented, even in a buoyant economy, for markets alone to make the types of investmentsand projects our system needs. The cost–benefit analysis for renewable energy is not simple for manyinvestors. In addition, today’s credit crunch risks undermining many advances industry has made in recentyears. We also have to face up to the adverse impact on investment and investment planning that couldensue from the weak Copenhagen climate agreement.

Moving to a low-carbon economy calls for investments in the order of billions of euros. But climateand security risks also have a cost, which could be several times higher.There is a growing consensus thatthe investments we make today will pay for themselves many times over in the future, in terms of cheaperenergy, greater energy security, new businesses and markets, a cleaner environment, and climate changemitigation. We cannot afford to miss out on the huge technology and employment opportunities andgeopolitical security of renewable power.

Our 2020 targets are a stepping-stone toward the EU vision for a decarbonized power supply andtransport sector by 2050. Let us not forget that the system we have in place today is largely a result oftechnologies and decisions from 40 years ago, or even more, so our 2050 timescale is not as ambitious asit sounds. It simply reflects the reality of how energy systems develop.

To get there, the commission is pursuing a number of interrelated priorities: to implement fully andeffectively an internal market in Europe; to enforce renewables and greenhouse gas emission targets andstrengthen the legislative framework to give greater certainty to investors; to create a solid framework fornew network infrastructure investments; to promote greater collaboration and cohesion in research,including industry-led initiatives; to boost energy efficiency, notably in buildings, appliances, and trans-port; and develop partnerships with other countries, both producers and other consuming nations.

It is vital to inform, involve, educate, and motivate the widest possible public. Every citizen can play apart in the low-carbon revolution.

This book is therefore both timely and important. It is also highly educational. It can help us betterunderstand the opportunities and challenges renewable energy creates and learn from the experience ofothers. In this way, we will all, whether in politics, business, education, or at home, be better equipped tomake, with confidence, the vital decisions that will lead to the energy system we seek for the future—asystem that is at once secure, sustainable, and economically robust.

Günther OettingerEuropean Union Energy Commissioner

xxviii Foreword

1

Toward a Low-Carbon Futurein Electricity?Boaz Moselle, Jorge Padilla, and Richard Schmalensee

Wide consensus exists that the world’s energysystem requires a significant transformation

in order to address the issue of climate change. Inparticular, the electric power system needs toswitch from its carbon-intensive present, wheremost electricity is produced using gas and coal, toa low-carbon future where electricity is producedusing some combination of renewable sources,such as hydro, wind, biomass, tidal, and solarenergy, along with nuclear power and carbon cap-ture and storage (CCS) technology, which wouldallow the continued use of fossil fuels, with thecarbon dioxide (CO2) emitted from combustionbeing stored underground rather than escapinginto the atmosphere. A radical change is neededbecause at present, power generation produces amajor part of the world’s anthropogenic (orhuman-caused) carbon dioxide emissions, reflect-ing its reliance on fossil fuels, and moreoverbecause decarbonization is likely to requireincreasing electrification of other sectors of theeconomy—in particular, a switch to electric orplug-in hybrid vehicles.

Of the options for producing electric powerwithout significant greenhouse gas emissions,renewable generation technologies such as hydro-electricity, wind power, solar power, and biomassare generally more attractive to policymakers thaneither nuclear power, which brings with it con-

cerns about safety and waste storage as well asdifficult political challenges, or CCS, a new and asyet unproven technology. Renewables are alsoperceived to have other important attractions:increased national self-reliance contributing tosecurity of energy supply, industrial policy goalsfor countries that see themselves as leaders indeveloping new green technologies, and domesticjob creation.

This book therefore addresses a set of keyquestions concerning the role of renewable elec-tricity generation in addressing climate change—questions that are central to the current concernsof policymakers, the electric power industry, andeconomists who study energy and environmentalissues:

• At the highest level, is it right to focus onrenewable power as a primary tool for reduc-ing greenhouse gas emissions? Althoughenvironmentalists and many politicians arestrong supporters of renewable energy, manyeconomists question whether it should besingled out for support rather than beingencouraged through a broad approach toreducing carbon emissions (e.g., via a tax onCO2 emissions).

• If renewable energy is to be given specificsupport, what form should that take? What is

the right balance between support forresearch and development and for deploy-ment of renewable generation? Should allrenewable technologies receive equal levels ofsupport? If not, how should the systemchoose among technologies? What is theright mechanism for supporting renewablegeneration: straight financial payments, taxbreaks, guarantees of prices at levels highenough to support investment, or a market-based method involving the trading of greencertificates?

• At a more technical level—but of greatimportance—what are the implications forpower markets of the widespread adoption ofrenewable generation? Is this trend compat-ible with the shift (or aspiration) toward lib-eralized, competitive markets that is seen inthe European Union and much of the UnitedStates? What are the implications for systemreliability of the use of intermittent resources,such as wind and sunshine, for a large part ofgeneration capacity? Will markets provideneeded backup generation, or is regulatoryintervention required?

The various chapters in this book provide objec-tive analysis of these questions, drawing on eco-nomic theory and evidence, including the experi-ences of U.S. states and EU member states thathave devoted considerable resources to the pro-motion of renewable generation over the pastdecade or more.The focus is on the application ofrigorous economic analysis to derive insights thathave practical implications for policy in this area.The book is the result of a cooperative effortamong a set of authors that includes both academ-ics and policymakers, with a range of specialtiescovering renewable technology, environmentaland natural resource economics, energy econom-ics and industrial organization, and with specificknowledge of the energy sector, electric powermarkets, and renewables policy in Europe, theUnited States, and globally.

Although many books currently exist on cli-mate change-related topics, most of these addressdifferent issues than this one. Many are focused onthe scientific issues around global climate change

and its implications in such areas as agriculture,migration, or public health (see, e.g., Kruger2006; Tester et al. 2005). Others focus on policyresponses to climate change, often from a politicalscience or international relations perspective (see,e.g., Dodds et al. 2009; Giddens 2009; Kalicki andGoldwyn 2005; Mitchell 2010; Stern 2009).Although a number of important texts have beenwritten on the economics of global climate (e.g.,Helm and Hepburn 2009), only a few address therole of renewable energy as part of climate changemitigation and the challenges that renewable gen-eration poses for electricity markets (e.g., Grubb,Jamasb and Pollitt 2008).

This book differs from others on renewablesin breadth and depth. It deals with technologyissues, wide policy questions (such as the impactof renewables support on the climate change andsecurity of supply goals of most governments),and practical implementation issues (e.g., theimplications for system design, market perform-ance, and policy design of massive deployment ofrenewable energy). It also describes therenewables policies and experiences of the UnitedStates, the European Union, the United King-dom, Germany, and Spain and examines the les-sons that can be extracted from them.

The book is organized into four parts. In PartI, Technology, Godfrey Boyle reviews in Chapter2 each of the technologies covered in the bookfrom both an economic and an engineering view-point, considering the impact of each on systemoperation, as well as its contribution tosustainability and economic efficiency.

Part II, Renewables, Climate Change, andEnergy Policy, focuses on the high-level energypolicy concerns that underlie renewables policy.In Chapter 3, Erin Mansur discusses how wide-spread concerns about climate change andpolicymakers’ and voters’ preferences will affectthe development of renewable energy use. BoazMoselle in Chapter 4 looks at the other maindriver of renewables policy, the desire to developindigenous sources of energy to meet security ofsupply concerns. He analyzes the main security ofsupply issues that drive the European debate,notably the question of reliance on Russian gasimports. Next, in Chapter 5, Kenneth Gillingham

2 Boaz Moselle, Jorge Padilla, and Richard Schmalensee

and James Sweeney delve into the economicmotivation for renewable energy policies by high-lighting the classes of relevant market failures. Keypolicy instruments are evaluated in the context ofdifferent market structures, with the goal ofinforming future policies in the renewable energysector. Finally, José Goldemberg provides inChapter 6 a broad economic analysis of the rela-tionships among four important instruments thathave become common in renewable energypolicy: subsidies to renewable energy, emissionstrading, the promotion of energy efficiency, andthe Clean Development Mechanism developedunder the Kyoto Protocol.

Part III, Renewable Generation and ElectricPower Markets, focuses one level down, lookingat the implications of renewable generation forelectric power markets. In Chapter 7, WilliamHogan discusses the difficulties associated withdeploying renewable electricity in large-scalewholesale electricity markets. He considers suchissues as system design, transmission and distribu-tion, and the development of smart grids, withspecial emphasis placed on the problems of inte-grating wind power into wholesale market design.Chapter 8, by Richard Green, considers the regu-latory issues that will arise from the desire to inte-grate large-scale renewable energy deploymentinto existing electricity markets and discusses howthese issues can be resolved. James Bushnell inChapter 9 studies the intersection of two impor-tant trends: the restructuring of electricity marketsand the growth of environmental regulation. Heprovides an empirical study that elucidates howthe increasing penetration of intermittent renew-able generation will alter the economics of invest-ment in conventional thermal generation. Finally,in Chapter 10, Christian von Hirschhausenaddresses the development of a ‘supergrid’, whichis currently being debated in the context of har-nessing renewable energy. Using one of the moreadvanced concepts for a ‘supergrid’ in theEuropean/North African region, the discussionalso highlights the practical obstacles in the way ofthe realization of such a project.

Part IV, National Experiences, offers insightsfrom the experiences of various countries,with individual chapters on the United States,

the European Union, the United Kingdom,Germany, and Spain. The aim here is to capturelessons both from the different policies and policyinstruments that have been applied and from dif-ferences in the policy concerns and debates acrossthe various countries. In Chapter 11, RichardSchmalensee analyzes the development of renew-able energy policies and deployment in theUnited States. He compares and contrasts thegrowth of renewable electricity deployment inTexas and California, highlighting the ability ofwell-tailored policy to advance the growth ofrenewable energy use, and discusses the difficultiesof integrating wind power into the existing mar-ket framework. Christopher Jones in Chapter 12provides an in-depth analysis of the emergence ofthe European Union’s renewables policy, in thecontext of its wider climate change and energysecurity concerns. He describes the policy toolsadopted and analyzes some of the key issues aris-ing, such as the relationship with the EU Emis-sions Trading Scheme and the role of tradingwithin the renewables promotion framework.

The remaining case studies concern three EUmember states. Michael Pollitt in Chapter 13examines the United Kingdom’s renewableenergy policy in the context of its overall decar-bonization and energy policies. He explores theshortcomings of the UK renewables policy to dateand suggests policy changes that would better suitthe country’s institutional and resource endow-ments. In Chapter 14, Hannes Weigt and FlorianLeuthold discuss the development of renewableenergy policy and deployment in Germany. Inparticular, they examine the implications of cur-rent policy for market design and future develop-ment, with a particular focus on the growingshare of wind energy. Finally, Chapter 15, by LuisAgosti and Jorge Padilla, analyzes the regulatoryschemes and market structure that have made pos-sible the rapid growth of the Spanish renewablessector in recent years, as well as the consequencesof this growth for the Spanish gas and power mar-kets. The chapter focuses in particular on theimpact of the large-scale deployment of windgeneration on system operation and balancing,wholesale prices, competition, and investmentincentives.

Toward a Low-Carbon Future in Electricity? 3

It is now widely accepted that a low-carbonworld is technically feasible, and that renewablesare likely to have a significant long-term role toplay in the future energy system. The relevantquestion today is not whether renewables shouldbe supported, but rather how to harness thesetechnologies so that they can effectively and effi-ciently contribute to the ultimate goal of mitigat-ing harmful climate change.

ReferencesDodds, Felix, Andrew Higham and Richard Sherman.

Foreword by Achim Steiner. 2009. Climate Changeand Energy Insecurity. London: Earthscan.

Giddens, Anthony. 2009. The Politics of Climate Change.Cambridge, UK: Polity Press.

Grubb, Michael, Tooraj Jamasb, and Michael Pollitt,eds. 2008. Delivering a Low Carbon Electricity System.Cambridge, UK: Cambridge University Press.

Helm, Dieter, and Cameron Hepburn, eds. 2009. TheEconomics and Politics of Climate Change. Oxford:Oxford University Press.

Kalicki, Jan H., and David L. Goldwyn, eds. 2005.Energy & Security. Baltimore: Johns Hopkins Univer-sity Press.

Kruger, Paul. 2006. Alternative Energy Resources: TheQuest for Sustainable Energy. Hoboken, NJ: Wiley-Blackwell.

Mitchell, Catherine. 2010. The Political Economy of Sus-tainable Energy. Basingstoke, UK: PalgraveMacmillan.

Stern, Nicholas. 2009. A Blueprint for a Safer Planet.London: The Bodley Head Ltd.

Tester, Jefferson W., Elisabeth M. Drake, Michael J.Driscoll, Michael W. Golay and William A. Peters2005. Sustainable Energy: Choosing Among Options.Cambridge, MA: MIT Press.

4 Boaz Moselle, Jorge Padilla, and Richard Schmalensee

Part I

Technology

2

Renewable Energy Technologies forElectricity GenerationGodfrey Boyle

It is tempting to imagine that the 17th-centurypoet and mystic William Blake was thinking of

renewable energy when, in The Marriage of Heavenand Hell, he declared that “Energy is EternalDelight” (Blake 1790). Renewable energy is usu-ally described in more prosaic terms, however.According to the International Energy Agency’sdefinition, “Renewable energy is derived fromnatural processes that are replenished constantly”(IEA 2002). Similarly, Twidell and Weir (1986)define it as “energy obtained from the continuousor repetitive currents of energy recurring in thenatural environment.”

How, then, should “renewable energy tech-nologies” be defined? The following working defi-nition may suffice: renewable energy technologiesare those technologies that enable constantlyreplenished renewable energy flows to be har-nessed to produce power in forms useful tohumanity on a sustainable basis.

The sun is the source of the vast majority ofthe power that drives the abundant and variedsources of renewable energy that are available tohumanity on the earth. The total quantity of solarpower incident on our planet is approximatelyfour orders of magnitude greater than our currentrate of use of fossil and nuclear fuels.

Solar radiation can be used directly, to provideheating, lighting, and hot water in buildings and

to generate electricity. The sun also powers theworld’s weather systems and is thus the indirectsource of hydro, wind, and wave power. It alsodrives the process of photosynthesis in plants andso is the energy source underlying biofuels in theirvarious forms.

Two other terms that should be defined hereat the outset are “energy” and “power.” Energy isdefined as the capacity to do work: that is, to movean object against a resisting force. In everyday lan-guage, the word “power” is often used as a syno-nym for “energy,” but this is not strictly correct.Power is defined as the rate of doing work—thatis, the rate at which energy is converted from oneform to another or transmitted from one place toanother.The main unit of measurement of energyis the joule (J), and the main unit of measurementof power is the watt (W), which is defined as arate of one joule per second.

Solar-Based RenewableEnergy Sources

The Solar Resource

Solar energy makes an enormous but largelyunrecorded contribution to our energy needs. It is

the sun’s radiant energy, aided by the atmosphericgreenhouse effect, that maintains the earth’s sur-face at a temperature warm enough to supporthuman life.

The sun has a surface temperature of 6,000degrees Celsius (°C), maintained by continuousnuclear fusion reactions between hydrogen atomswithin its interior.This is a relatively slow process,and the sun should continue to supply power foranother five billion years. The sun radiates hugequantities of energy into the surrounding space,and the tiny fraction intercepted by the earth’satmosphere 150 million kilometers (km) away isnonetheless equivalent to about 15,000 timeshumanity’s present rate of use of fossil and nuclearfuels. Even though approximately one-third ofthe intercepted energy is reflected away by theatmosphere before reaching the earth’s surface,this still means that a continuous and virtuallyinexhaustible flow of power, amounting to some9,000 times our current rate of consumption ofconventional fossil and nuclear fuels, is available inprinciple to human civilization.

Direct Uses of Solar Energy

Solar energy, when it enters our buildings, warmsand illuminates them to a significant extent.When buildings are designed to take full advan-tage of the sun’s radiation, their needs for addi-tional heating and artificial lighting can bereduced. The sun’s heat can also be harnessed byusing solar collectors to produce hot water forwashing and, in some circumstances, space heat-ing in buildings. Such collectors are in widespreaduse in sunny countries, such as Israel and Greece,but are also quite widely used in less sunny places,such as Austria.

Solar radiation can also be concentrated bymirrors or lenses to provide high-temperatureheat to drive heat engines that generate electricity.Such solar thermal-electric power stations are inoperation in several countries. Solar radiation canalso be converted directly into electricity usingphotovoltaic (PV) panels, usually mounted on theroofs or facades of buildings.

Indirect Uses of Solar Energy

Solar radiation can be converted to useful energyindirectly, via other energy forms. Sunlight falls ina more perpendicular direction in tropical regionsand more obliquely at high latitudes, heating thetropics to a greater degree than polar regions.Theresult is a massive heat flow towards the poles,carried by currents in the oceans and the atmos-phere. The energy in such currents can be har-nessed, for example by wind turbines.

Where winds blow over long stretches ofocean, they create waves, and a variety of devicescan be used to extract that energy.

Biofuels are another indirect manifestation ofsolar energy. Through photosynthesis in plants,solar radiation converts water and atmosphericcarbon dioxide into carbohydrates, which formthe basis of more complex molecules that consti-tute biofuels such as wood or ethanol.

A large fraction of the solar radiation reachingthe earth’s surface is absorbed by the oceans,warming them and adding water vapor to the air.This water vapor condenses as rain to feed rivers,in which dams and turbines can be placed toextract some of the energy of the flowing water,creating hydropower.

Nonsolar RenewablesTwo other sources of renewable energy, tidal andgeothermal energy, are not dependent on solarradiation. Tidal energy is sometimes confusedwith wave energy, but its origins are quite differ-ent. It arises from the gravitational pull of themoon (plus a small contribution from the sun) onthe world’s oceans. The source of geothermalenergy is heat from within the earth causedmainly by the decay of radioactive materialswithin the earth’s core.

Renewable ElectricityGenerating TechnologiesThis chapter concentrates on the use ofrenewables for electricity production, looking ini-

8 Godfrey Boyle

tially at solar, wind, and wave power, then turningto biofuels, hydro, tidal, and geothermal energy. Itthen analyzes the costs and economics of renew-able electricity, and finally examines the chal-lenges involved in integrating renewable electri-city sources into current and future electricitysupply systems.1

Solar Thermal Electricity

Greek legend relates that in 212 BC, Archimedesordered his warriors to use their polished shieldsto concentrate the rays of the sun onto Romanships besieging Syracuse, setting fire to them. AsEverett observes: “Although long derided asmyth, Greek navy experiments in 1973 showedthat 60 men each armed with a mirror 1m by1.5m could indeed ignite a wooden boat at 50m”(Everett, 2004, 51).

Solar thermal electricity generating systemsessentially use the sun’s energy, in either direct orconcentrated form, to drive a heat engine, whichin turn drives an electrical generator. Heatengines (a steam engine, for example) convertheat to mechanical work by exploiting the tem-perature difference between a hot heat source anda cold heat sink, usually the ambient environ-ment. The efficiency of a heat engine increaseswith increasing temperature difference betweensource and sink, so it is useful to rank the varioustechnologies in order of their operating tempera-ture.

Low-Temperature Solar Thermal-ElectricTechnologies

Solar PondsA solar pond is a very large pond filled with saltwater and used to collect solar energy. The toplayer has a low salt content, and the bottom layerhas a high salt content. Between these is an inter-mediate layer with a varying salt concentrationgradient, designed to prevent natural convectionof heat from the bottom to the top. If the toplayers of the pond are translucent, the bottomlayer can absorb substantial quantities of solarenergy and reach relatively high temperatures,around 90°C in sunny regions. As the temperature

at the top of the pond will remain at ambientlevels, around 30°C, the difference in these tem-peratures can be used to power an organicRankine cycle (ORC) heat engine, which in turncan power a generator to produce electricity.

An ORC engine operates on a cycle similarto that used in conventional steam turbines,except that the working fluid is not water (turnedto steam) but an organic fluid, such as pentane orbutane, which evaporates at lower temperatures.The vapor then drives a turbine. This cycle ena-bles power to be produced from much lower-temperature sources than those used with steamturbines, though also with much reduced effi-ciency.

The low efficiency means that solar pondsneed to be large to generate significant quantitiesof electrical power. (They are perhaps more suitedto low-temperature heat production.) They alsooperate effectively only at low latitudes becausethe absorbing area is horizontal. In addition, themaintenance costs of replenishing the surfacewater and maintaining the salt gradient are signifi-cant.

Ocean Thermal Energy Conversion (OTEC)In warm tropical regions of the world’s oceans,the temperature of water at the surface can exceed25°C, whereas at depths of around 1,000 meters,the water temperature can be below 5°C. Thistemperature differential can be used to drive aturbine employing an organic Rankine cycle heatengine, which in turn can power a generator toproduce electricity, in what is called ocean ther-mal energy conversion (OTEC).

The potential resource of solar energy storedin the surface layers of the world’s oceans andavailable to be tapped by OTEC systems is huge.Such systems should also be able to producepower on a continuous basis. Between the 1970sand the 1990s, several experimental plants weretested in Hawaii, and another in Japan. However,the engineering challenges involved in pumpinghuge quantities of water through extremely long,large-diameter pipes from the ocean depths to thesurface are formidable, and considerable energy isrequired simply to pump the water.

Renewable EnergyTechnologies for Electricity Generation 9

The U.S. firm Lockheed Martin is currentlydeveloping plans for a new OTEC system takingadvantage of modern materials and technology togenerate power on a scale of approximately 100megawatts (MW). Given adequate funding, thecompany foresees that a pilot 10 MW plant couldbe operating off Hawaii by 2013 (James 2009;Lockheed Martin n.d.).

Solar ChimneyThe solar chimney concept involves heating largevolumes of air in an extremely large “greenhouse”and allowing the warmed air to rise through avery tall chimney. The stream of rising air powersa wind turbine at the base of the chimney.

Such systems need to be very large if they areto generate significant quantities of power. In a50-kilowatt experimental project at Manzanaresin Spain, the greenhouse-type solar collector was240 meters in diameter and the chimney was 195meters high. Compared with concentrating solarcollectors, the energy conversion efficiency ofsolar chimney systems is low, and the land areathey occupy is correspondingly greater. They canproduce power using diffuse as well as direct solarradiation, however, and some energy storage canbe provided though heating the floor of thegreenhouse (see Schlaich 1995).

Medium- and High-Temperature SolarElectric Technologies

Parabolic Trough ConcentratorsConcentrating solar systems are normally practicalonly in relatively cloud-free regions of the worldwhere the majority of solar radiation is direct.This is because the mirrors or lenses they use can-not concentrate the sun’s indirect, diffuse radia-tion.

Parabolic trough concentrators (PTCs) incor-porate trough-shaped parabolic mirrors that trackthe sun along a single axis, concentrating the sun’srays onto a tube at the focus containing a heattransfer fluid, usually a synthetic oil. The absorbertube is normally surrounded by a glass tube toreduce heat loss by re-radiation. The oil is thenpassed through a heat exchanger to produce high-temperature steam. This powers a turbine, which

in turn drives an electrical generator. Relativelyhigh temperatures (around 400°C) can beattained, enabling fairly high energy conversionefficiencies (about 20%) to be achieved.

The world’s longest-operating concentratingsolar power plants are the nine Solar Energy Gen-erating Systems (SEGS) established by the LuzEnergy Corporation in California’s MojaveDesert between 1984 and 1990. They use para-bolic trough concentrators occupying severalsquare kilometers of land, with a combined out-put of 350 MW. The plants were designed to helpmeet California’s summer peak air-conditioningload. At night or in low-sun conditions, a naturalgas boiler can be used to supply auxiliary steam tothe turbines, enabling them to continue to supplypower. Another, more recent, large parabolictrough concentrating solar system in the UnitedStates is the 64 MW Solar One, installed in theNevada desert by Acciona Energy in 2007. TheIsraeli company Solel has signed a power purchaseagreement with the California utility Pacific Gasand Electric to install some 553 MW of parabolictrough concentrating solar capacity in the MojaveDesert by 2011 (Hopwood 2009).

Europe’s largest parabolic trough concentrat-ing solar electric plants are the Andasol installa-tions in southern Spain, each rated at 50 MW.Thefirst is in commercial operation, the secondentered its testing phase in 2009, and the third isdue to be commissioned in 2011.They are locatedon the Guadix plateau in the province of Granada,an area with clear skies and very high annual solarradiation levels. Andasol 1 has 312 rows of collec-tors, totaling more than 500,000 square meters inarea, and has an estimated annual output of 180GWh. It includes a molten salt-based heat storagesystem that allows power production to continuefor up to 7.5 hours after sunset.

A recent development in concentrator tech-nology involves the use of lower-cost linearFresnel reflectors instead of parabolic mirrors.These consist of flat or slightly curved mirrorsarrayed in long rows and aligned so that they focusthe sun’s rays onto a long tube, with the aid of asmall secondary reflector. Water in the tube isturned directly into high-pressure, high-temperature steam, which can be used to power a

10 Godfrey Boyle

turbine. An experimental plant was set up inAlmeria, Spain, in 2007 by researchers from theFraunhofer Institute for Solar Energy Systems,based in Freiburg, Germany. The German com-pany Novatec Biosol commissioned PE1, a 1.4MW plant using Fresnel technology, in Spain inMay 2009 and has commenced work on a larger,30 MW plant.

Power TowersAn early example of the power tower approachfrom the 1980s and 1990s was the large installa-tion at Barstow, California, now decommissioned.This used a field of tracking mirrors, or heliostats,to reflect the sun’s rays onto a boiler at the top of acentral tower.The first Solar One plant used high-temperature synthetic oils to transfer the heat to aboiler to raise steam. The later phase of theproject, Solar Two, included high-temperatureheat storage using molten salt, allowing electricityto be produced on a more continuous basis.

More recently, Abengoa Solar’s 20 MW PS20power tower near Seville in Spain (an uprated ver-sion of the company’s earlier 10 MW PS10design, installed nearby) became operational insummer 2009. The PS20 uses 1,255 movableheliostats, each 120 square meters in area, to focusthe sun’s rays onto a receiver and heat exchangerat the top of a 162-meter tower. The high-temperature solar heat produces steam to drive aturbine, which in turn powers a generator to pro-duce electricity. Some of the steam is stored ininsulated tanks, enabling generation to continueat night (Marsh 2009). The company is planningto construct larger power towers of up to 50 MWcapacity as part of a solar electricity generatingcomplex totaling 300 MW and incorporating avariety of solar electricity technologies, includingparabolic trough and dish Stirling systems(described in the next section).

In the United States, Brightsource Energy,using its Luz Power Tower 550 system, aims toconstruct a number of power towers, each of 100MW capacity, in projects for the utilities SouthernCalifornia Edison and Pacific Gas and Electric.The aim is to have some 900 MW operational by2013 (Hopwood 2009). Major concentrating

solar power (CSP) plants are also planned in Israel,the United Arab Emirates, and other Middle East-ern countries (Renew 2009).

Dish Stirling Solar ConcentratorsAnother approach to concentrating solar power isto place a suitable small engine, such as a Stirlingengine, at the focus of a dish-type solar concen-trator. Dish Stirling solar systems can operate ataround 1,000°C and deliver high energy conver-sion efficiencies, approaching 30%. The dishtracks the sun on two axes: azimuth and elevation.Practically, each individual unit can be con-structed only on a relatively small scale, around 10to 15 MW.

System Integration and Economics

The output of solar thermal-electric power plantsis dependent on the solar input, which varies on adaily and seasonal basis. It is well matched to peakelectric air-conditioning loads in many areas.Many systems incorporate thermal storage, andsome include an auxiliary fossil-fueled plant, ena-bling their output to be dispatchable in a similarmanner to conventional generating plants.

Electricity from concentrating solar thermal-electric systems is currently estimated to costaround 9 to 12 cents per kilowatt-hour (kWh),about two to three times the cost of conventionalelectricity (using normal accounting conventions,though these exclude the external costs of con-ventional generation). But with technologicalimprovements and quantity production, CSPcosts are expected to fall to about 6 cents/kWhwithin a decade (Marsh 2009).

Solar Photovoltaic Electricity

Physical Principles

The silicon solar photovoltaic (PV) cell is perhapsthe ideal energy conversion device. Its “fuel”input, solar energy, is free; silicon is the secondmost abundant material in the earth’s crust; itsoutput is that most useful of energy forms, elec-tricity; and, with no moving parts, it hasextremely low operation and maintenancerequirements.

Renewable EnergyTechnologies for Electricity Generation 11

Photovoltaic cells consist, in essence, of ajunction between thin layers of two different typesof semiconductor, known as p-type (positive) andn-type (negative). Semiconductors are materialswhose electrical properties are intermediatebetween those of conductors, which offer littleresistance to the flow of electric current, and insu-lators, which inhibit the flow of electricity. Thesemiconductors are usually made from silicon,although PV cells can be made from other mate-rials. N-type semiconductors are made from crys-talline silicon that has been “doped” with tinyquantities of an impurity, usually phosphorus, insuch a way that the doped material possesses asurplus of free electrons. Because electrons possessa negative electrical charge, silicon doped in thisway is known as an n-type semiconductor. P-typesemiconductors are doped with very smallamounts of a different impurity, usually boron,which causes the material to have a deficit of freeelectrons. These missing electrons are called“holes.” Because the absence of a negativelycharged electron can be considered equivalent toa positively charged particle, silicon doped in thisway is known as a p-type semiconductor.

A p–n junction can be created by joiningthese dissimilar semiconductors. This sets up anelectric field in the region of the junction, whichwill cause negatively charged particles to move inone direction and positively charged particles tomove in the opposite direction. When photons oflight of a suitable wavelength fall within the p–njunction, they can transfer their energy to some ofthe electrons in the material, causing an electriccurrent to flow across the junction.

Photovoltaic Materials and Technologies

Crystalline SiliconThe most efficient silicon solar cells are madefrom extremely pure monocrystalline silicon—that is, silicon with a single, continuous crystallattice structure with virtually no defects or impu-rities. Monocrystalline silicon is usually grownfrom a small seed crystal that is slowly pulled outof a molten mass of polycrystalline silicon, in asophisticated process developed originally for the

electronics industry. Polycrystalline silicon essen-tially consists of small grains of monocrystallinesilicon.

Solar cell wafers can be made directly frompolycrystalline silicon in various ways. Theseinclude the controlled casting of moltenpolycrystalline silicon into cube-shaped ingots,which are then cut into thin, square wafers andfabricated into complete cells in the same way asmonocrystalline cells.

Polycrystalline PV cells are easier and cheaperto manufacture than their monocrystalline coun-terparts but are less efficient. Commercially avail-able polycrystalline PV modules, sometimes called“multicrystalline” or “semicrystalline,” can attainenergy conversion efficiencies of around 14%,whereas monocrystalline module efficiencies canexceed 17%.

Gallium ArsenideAnother material suitable for PV is galliumarsenide (GaAs), a so-called compound semicon-ductor. It has a crystal structure similar to that ofsilicon but consisting of alternating gallium andarsenic atoms. In principle, it is highly suitable foruse in PV applications because it has a high lightabsorption coefficient, so only a thin layer ofmaterial is required. GaAs cells are more efficientthan those made from monocrystalline silicon.They can also operate at relatively high tempera-tures without substantial reduction in efficiency,which makes them well suited to use in concen-trating PV systems.

But GaAs cells are more expensive than siliconcells, partly because the production process is notso well developed and partly because gallium andarsenic are not abundant materials.They are oftenused when very high efficiency is required,regardless of cost, as in many space applications.

Thin-Film Silicon PVSolar cells can also be made from very thin filmsof silicon, in a form known as amorphous silicon(a-Si), in which the silicon atoms are much lessordered than in the crystalline forms describedabove. Amorphous silicon cells are much cheaperto produce than those made from crystalline sili-con. This form of silicon is also a better absorber

12 Godfrey Boyle

of light, so thinner, and therefore cheaper, filmscan be used.There are advantages in the manufac-turing process as well: it operates at a much lowertemperature than that for crystalline silicon, so lessenergy is required; it is suited to continuous pro-duction; and it allows quite large areas of cells tobe deposited onto a wide variety of both rigid andflexible substrates, including steel, glass, and plas-tics.

But a-Si cells are currently much less efficientthan their single-crystal or polycrystalline siliconcounterparts. Commercially available a-Si mod-ules achieve stable efficiencies in the range of 4%to 8%. A-Si cells are already widely used as powersources for a variety of consumer products, such ascalculators, where the requirement is for low costrather than high efficiency.

Other Thin-Film PV TechnologiesAmong the many other thin film technologies,some of the most attractive are those based oncompound semiconductors, in particular, copperindium diselenide (CuInSe2, usually abbreviatedto CIS), copper indium gallium diselenide(CIGS), and cadmium telluride (CdTe). Modulesbased on all of these technologies are in produc-tion from various manufacturers. Thin-film CIGScells have attained the highest laboratoryefficiencies of all thin-film devices, around 17%,and CIGS modules with stable efficiencies over10% are commercially available.

Cadmium telluride modules can be madeusing a relatively simple and inexpensive process,and efficiencies over 10% are claimed. However,the modules contain cadmium, a highly toxic sub-stance, so stringent precautions need to be takenduring manufacture, use, and eventual recycling.

Other thin-film and innovative technologiesentering production or in development includemultijunction PV cells, in which different p–njunctions are “tuned” to absorb light from differ-ent parts of the solar spectrum, and photo-electrochemical cells using dye-sensitized layers oftitanium dioxide. Still at the R&D stage are third-generation photovoltaic systems based onnanotechnology or using organic materials.

Concentrating PV Systems

The energy output of PV cells can be increased byusing mirrors or lenses to concentrate the incom-ing solar radiation onto the cells—an approachsimilar to that described above in the section onparabolic trough concentrators. The concentra-tion ratio can vary from as little as two to severalhundred or even several thousand times.The con-centrating system must have an aperture equal tothat of an equivalent flat plate array to collect thesame amount of incoming solar energy. In con-centrating PV systems, the cells are often cooled,either passively or actively, to prevent overheating.

Systems with the highest concentration ratiosuse sensors, motors, and controls to allow them totrack the sun on two axes—azimuth (horizontalorientation) and elevation (vertical tilt)—ensuringthat the cells always receive the maximum amountof solar radiation. Systems with lower concentra-tion ratios track the sun on a single axis and canhave simpler tracking mechanisms.

PV Systems for Remote Power

In many developing countries, electricity gridsare often nonexistent or rudimentary, particularlyin rural areas, and all forms of energy are usuallyvery expensive. Here PV systems, usually incor-porating battery storage, can be highly competi-tive with other forms of energy supply, and theiruse is growing rapidly.

Grid-Connected, Building-Integrated PV

In most parts of the developed world, grid elec-tricity is easily accessible and can provide a con-venient backup to PV or other fluctuating renew-able energy supplies. In these grid-connected PVsystems, a grid-commutated inverter, or synchro-nous inverter, transforms the DC power from thePV arrays into AC power at a voltage and fre-quency that can be accepted by the grid, while“debit” and “credit” meters measure the amountsof power bought from or sold to the utility.

PV arrays can be built into the roofs of houses.They can also be integrated into the roofs andwalls of commercial, institutional, and industrialbuildings, replacing some of the conventional wall

Renewable EnergyTechnologies for Electricity Generation 13

cladding or roofing materials that would other-wise have been needed and reducing the net costsof the PV system. In the case of some prestigiousoffice buildings, the cost of conventional claddingmaterials can exceed the cost of cladding with PV.

Commercial and industrial buildings normallyare occupied mainly during daylight hours; thiscorrelates well with the availability of solar radia-tion.Thus the power generated by PV can signifi-cantly reduce an organization’s need to purchasepower from the grid at the retail price and insome countries can be sold to the grid at a pre-mium price (discussed in the section on Costs andEconomics of Renewables below).

Large, Grid-Connected PV Power Plants

Large, centralized PV power systems, many atmultimegawatt scale, have been built to supplypower for local or regional electricity grids in anumber of countries, including Germany, Swit-zerland, Italy, and the United States. Comparedwith building-integrated PV systems, large stand-alone PV plants can take advantage of economiesof scale in purchasing and installing large numbersof PV modules and associated equipment, andthey can be located on sites that are optimal interms of solar radiation. On the other hand, theelectricity they produce is not used on-site andhas to be distributed by the grid, involving trans-mission losses.

Large plants also require substantial areas ofland, which has to be purchased or leased, but insome cases low-value “waste” land, such as thatalongside roads or railways, can be used. The landcan also often be used for other purposes as well asPV generation. In the 1.7 MW installation atSonnen in Germany, for example, the PV arrayshave been mounted at least one meter above theground, minimizing shading to the vegetationbeneath and allowing sheep to graze beneath thepanels. It would also, in principle, be possible tohave other forms of renewable energy generation,such as wind turbines, alongside a large PV sys-tem.

Large PV power plants are more economicallyattractive in those regions of the world that havesubstantially greater annual total solar radiation,

and clearer skies, than northern Europe. In suchregions, the majority of the radiation is direct,making tracking and concentrating systems effec-tive and further increasing the annual energy out-put. The price of electricity from such PV instal-lations is likely to be less than half of that fromcomparable non-tracking installations.

Kurokawa et al. (2007) give a detailed analysisof the feasibility and economics of very large-scalephotovoltaic power plants in various regions ofthe world, including the Mediterranean, MiddleEast, Gobi Desert, and Oceania. Their report alsoexamines the conditions under which very large-scale PV installations in deserts could contributeto sustainable community development in thesurrounding regions and includes a comparison ofthe relative merits of concentrating solar thermaland concentrating PV systems.

System Integration of Grid-Connected PV

In northern European countries, most PV powerwould be produced in summer, when electricitydemand is relatively low; much less would be pro-duced in winter, when demand is high. But inmany other regions, PV supply correlates wellwith summer daytime air-conditioning demand.PV power is quite reliable during daylight hoursin climates with mainly clear skies, such as Cali-fornia and southern Spain; in more cloudy coun-tries, it can be intermittent at times.

Cost of Energy from PV Systems

The Photovoltaic Power Systems (PVPS) taskforce of the International Energy Agency (IEA)regularly tracks the prices of PV modules and sys-tems for a wide range of IEA member countries.System prices include both modules and the asso-ciated balance-of-system (BoS) costs—those ofthe array support structure, cabling, switching,inverters, and metering—plus the cost of connec-tion to the grid.The trends in module and systemprices (capital costs, adjusted for inflation) for1997–2007 are shown in Figure 2.1.

As can be seen, PV module prices in thelowest-cost country (country 3, not identified butprobably Germany) in 2007 were around $4 perinstalled watt, while system prices were around $6

14 Godfrey Boyle

per watt.The operation and maintenance (O&M)costs of PV installations are very low, and thus themain driver of the cost per unit of electricity pro-duced is the capital cost. As the cost of electricityfrom PV systems continues to reduce and the costof grid electricity from conventional fossil andnuclear sources continues to increase, many ana-lysts expect the two to converge in the nearfuture, leading to grid parity for PV power (seeFigure 2.2).

Reducing the Costs of PV Power

The European Photovoltaic Industries Associationsuggests that electricity from PV could become “amainstream power source in Europe by 2020”(EPIA 2009, 1).The report outlines three possiblescenarios for PV contributions to the electricitysystem by that year. In the “Baseline” scenario,few changes have been made to the existing elec-tricity system, and by 2020, PV contributes only4% of European electricity consumption. The“Accelerated Growth” scenario envisages minorchanges to the electricity system, better coopera-tion from utilities, and optimized PV supplychains, with the result that PV provides 6% ofEurope’s electricity by 2020. In the “ParadigmShift” scenario, PV provides 12% of Europe’selectricity by 2020; this involves rapid and wide-spread adoption of “smart grid” and storage tech-

nologies, coupled with further improvementsacross the PV supply chain.

Looking beyond Europe to the potential forworld-scale PV development, EPIA andGreenpeace International (2008) point out thatthe projections in their earlier Solar Generationreports (2001–2007) consistently underestimatedthe actual growth in world PV production (seeTable 2.1). Growth in recent years has greatlyexceeded expectations. The EPIA/Greenpeaceprojections for future world PV growth, shown inTable 2.2, therefore may not be as optimistic asthey might seem. They suggest that by 2030, PVcould be supplying some 2,600 terawatt-hours(TWh). This is 8.9% of the total world electricitydemand forecast by the International EnergyAgency in its Reference Scenario for that year; or13.8% of world electricity demand if the energyefficiency improvements envisaged in Green-peace’s own Energy Revolution scenario (2005) areimplemented.

16

14

12

Pric

e of

PV

mod

ules

and

sys

tem

s(c

onst

ant U

S$/

W)

10

8

6

4

2

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

–

Country #1 modulesCountry #1 systemsCountry #2 modulesCountry #2 systemsCountry #3 modulesCountry #3 systems

Source: IEA 2008

Figure 2.1. Evolution of price of PV modules andsystems in selected IEA reporting countries, 1997–2007, allowing for the effects of inflation

1990

1.0[�/kWh]

0.8

0.6

0.4

0.2

2000

900 h/a*0.44 �/kWh

1800 h/a*:0.22 �/kWh

2010 2020 2030 2040

PhotovoltaicsUtility peak powerUtility bulk power

Note: the black band indicates that the market support programs willbe necessary until about 2020 in some markets.

*h/a: Hours of sun per annum900 h/a corresponds to northem countries of Europe1800 h/a corresponds to southern countries of Europe

Source: EPIA/Greenpeace, 2008, 41

Figure 2.2. Projected convergence of utility pricesand PV generation costs

Renewable EnergyTechnologies for Electricity Generation 15

Energy Balance of PV Systems

A common misconception about PV cells is thatalmost as much energy is used in their manufac-ture as they generate during their lifetime. Thismay have been true in the early days of PV, but arecent study (Alsema and Niewlaar 2000) foundthe energy payback time of PV modules, includ-ing frames and support structures, to be betweentwo and five years in European conditions, andstated that with future improvements, this shouldreduce to one and a half to two years. The use ofmaterials with low embodied energy (such as

wood) in PV array support structures can alsoimprove the overall energy payback time of PVsystems.

Wind Power

When solar radiation enters the earth’s atmos-phere, because of the curvature of the earth, itwarms different regions of the atmosphere to dif-fering extents—most at the equator and least atthe poles. Because air tends to flow from warmerto cooler regions, this causes what we call winds.

Table 2.1. Annual PV capacity installed/predicted: market results versus Solar Generation (SG) scenariopredictions since 2001 (in MW)

Year 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010

Market result 334 439 594 1,052 1,320 1,467 2,392

SG I 2001 331 408 518 659 839 1,060 1,340 1,700 2,150 2,810

SG II 2004 986 1,283 1,675 2,190 2,877 3,634

SG III 2006 1,883 2,540 3,420 4,630 5,550

SG IV 2007 2,179 3,129 4,339 6,650

SG V 2008 4,175 6,160 6,950

Source: EPIA/Greenpeace 2008Note: The 2008 market result was 5,568 MW installed, again greatly exceeding the SG V 2008 projection of 4,175 MW

Table 2.2. Projected solar PV electricity output and detailed associated projections for 2030 in the SG V scenario

Global solar electricity output in 2030

8.9% of global electricity demand from PVa

13.8% of global electricity demand from PVb

Detailed projections for 2030

PV systems cumulative capacity 1,864 GW

Electricity production 2,646 TWh

Grid-connected consumers 1,250 million

Off-grid consumers 3,216 million

Employment potential 10 million jobs

Market value €454 billion ($618 billion) per annum

Cost of solar electricity 7–13 euro cents (10–18 U.S. cents) per kWh depending on location

Cumulative CO2 savings 8,963 million metric tons of CO2

Source: EPIA/Greenpeace 2008aDemand forecast from IEA Reference ScenariobDemand forecast from Greenpeace Energy (R)evolution scenario

16 Godfrey Boyle

It is these air flows that are harnessed in windmillsand wind turbines to produce power. Windpower, in the form of traditional windmills usedfor grinding corn or pumping water, has been inuse for centuries. But in the past few decades, theuse of wind turbines for electricity generation hasbeen growing rapidly. This discussion will con-centrate on modern, horizontal-axis wind tur-bines, installed on land and offshore.

It can be shown, using basic physical princi-ples, that the power in a stream of flowing air—that is, in the wind—is proportional to the cubeof the wind velocity. It can also be shown that thepower produced by a wind turbine is proportionalto the area swept by the blades, which for a con-ventional horizontal-axis turbine is proportionalto the square of the rotor diameter.

A typical modern wind turbine starts produ-cing power at wind speeds of around 5 meters persecond (m/s). As wind speed increases, the poweroutput increases steadily until it levels off at therated power of the turbine, its maximum poweroutput level.The wind speed at which it does thisis called the “rated wind speed.” The turbine isdesigned to continue producing power at thisconstant level at wind speeds above this rated leveluntil it reaches an upper limit, the “shutdownwind speed.” At this speed, the turbine is designedto stop rotating to avoid damage from excessiveforces.

If a wind turbine could produce power con-tinuously at its rated capacity for a year, its annualcapacity factor would be 1.0. Clearly this isimpossible, as the wind does not blow all the time.In practice, annual capacity factors range fromabout 0.2 to 0.4, depending on the frequency dis-tribution and magnitude of the wind speeds at thesite where a turbine is located.

European Wind Resources and Potential

A detailed study published by the European Envi-ronment Agency (EEA 2009) confirms that thewind energy resource available on land and off-shore in the 27 European Union (EU) countries isenormous. The raw technical potential resource,excluding environmental and economic con-straints, amounts to some 20 times Europe’s pro-

jected electricity needs by 2020. As Table 2.3shows, even when environmental and economicconsiderations are taken into account, the eco-nomically competitive potential, onshore and off-shore, still amounts to three times the projected2020 electricity demand and seven times that for2030.

At the end of 2008, the EU’s installed windcapacity was 65 GW, producing around 142TWhof electricity, just over 4% of Europe’s needs (EEA2009). The EU has set a target to produce 20% ofits primary energy from renewable sources by2020, which will involve a substantial contribu-tion from wind-generated electricity. Details ofwind energy developments in Germany andSpain, the EU countries with the largest installedwind generating capacity, are discussed in Chap-ters 14 and 15.

U.S. Wind Capacity and Potential

According to the Global Wind Energy Council:“In 2008, the U.S. wind energy industry broughtonline over 8,500 megawatts (MW) of new windpower capacity, increasing the nation’s cumulativetotal by 50% to over 25,300 MW.The new instal-lations place the U.S. on a trajectory to generate20% of the nation’s electricity by 2030 from windenergy as long as the industry continues to garnerlong-term policy support” (GWEC 2009).

A U.S. Department of Energy study (NREL2008) shows how the United States could rapidlyexpand its wind energy capacity to enable theachievement of this contribution. Regarding theeconomics of such a major expansion in capacitycompared with a “No New Wind” scenario, itconcluded:

Compared to other generation sources, the 20%Wind Scenario entails higher initial capital costs(to install wind capacity and associated transmis-sion infrastructure) in many areas, yet offers lowerongoing energy costs than conventional powerplants for operations, maintenance and fuel.Given the optimistic cost and performanceassumptions of wind and conventional energysources … the 20%Wind Scenario could requirean incremental investment of as little as $43 bil-lion NPV (Net Present Value) more than the

Renewable EnergyTechnologies for Electricity Generation 17

base-case No New Wind Scenario. This wouldrepresent less than 0.06 cent (6 one-hundredthsof 1 cent) per kilowatt-hour of total generation by2030, or roughly 50 cents per month per house-hold. (NREL 2008)

The projections of the costs the Department ofEnergy made for its 20% Wind Scenario versusthe No Wind Scenario, as described above, areshown in Figure 2.3.

World Wind Capacity and Potential

By the end of 2008, the total wind generatingcapacity installed worldwide was 120 GW. During2008, about 27 GW of capacity was installed, val-ued at some $36 billion, and the average cumula-tive capacity growth rate during the preceding

20% wind No new wind

Wind 0&M costs

Transmission costs

Wind capital costs

Fuel costs

Capital 0&M costs

Conventional 0&M costs

$3,000

$2,500

$2,000

$1,500

$1,000

Bill

ions

of 2

006

dolla

rs

$500

$0

Source: NREL 2008

Figure 2.3. Incremental investment costs of a20% wind contribution to U.S. electricity demandin 2030

Table 2.3. Projected technically available, environmentally constrained, and economically competitive potential forwind energy in 27 EU countries in 2020 and 2030

Year TWh Share of 2020 and2030 demanda

Technical potential Onshore 2020 45,000 11–13

Onshore 2030 45,000 10–11

Offshore 2020 25,000 6–7

Offshore 2030 30,000 7

Total 2020 70,000 17–20

Total 2030 75,000 17–18

Constrained potential Onshore 2020 39,000 10–11

Onshore 2030 39,000 9

Offshore 2020 2,800 0.7–0.8

Offshore 2030 3,500 0.8

Total 2020 41,800 10–12

Total 2030 42,500 10

Economically competitive potential Onshoreb 2020 9,600 2–3

Onshoreb 2030 27,000 6

Offshore 2020 2,600 0.6–0.7

Offshore 2030 3,400 0.6–0.8

Total 2020 12,200 3

Total 2030 30,400 7

Source: EEA 2009aEuropean Commission projections for energy demand in 2020 and 2030 (European Commission 2008a, b) are based on two scenarios:“business as usual” (4,078 TWh in 2020 and 4,408 TWh in 2030) and “EC Proposal with RES trading” (3,537 TWh in 2020 and 4,279 TWhin 2030). The figures here represent the wind capacity relative to these two scenarios; e.g., onshore capacity of 45,000 TWh in 2020 is 11 to12.7 times the size of projected demand.bThese figures do not exclude Natura 2000 areas

18 Godfrey Boyle

decade was 30% per annum. The wind industrynow employs some 400,000 people worldwide(GWEC 2009).

According to a major study by the GWECcalled Wind Force 12, by 2020, the world windindustry could be providing some 3,054 TWh ofelectricity, around 12% of world demand. Thiswould involve an installation rate of about 158GW per annum, culminating in a total installedcapacity of 1,254 GW by 2020. The investmentrequired would be approximately €80 billion($109 billion) per annum, involving the creationof 2.3 million job-years of employment. Thestudy estimates that the generation cost of elec-tricity from wind in 2020 could reduce to 2.45euro cents (U.S. 3.34 cents) per kWh (GWEC2009). This may appear to be a best-case scenariobut cost reductions of this order could be feasible,given the economies achievable through verylarge-scale production coupled with increases inturbine size and further improvements in turbinetechnologies.

Offshore Wind

Figure 2.4 shows the recent buildup of offshorewind generating capacity in Europe. InitiallyDenmark was the leading country, but since 2008the United Kingdom has taken the lead, withsome 1,000 MW installed by 2009. The UnitedKingdom has ambitious plans to install up to 29

GW of offshore capacity by 2020, though thismay be constrained (Carbon Trust 2008; DECC2009).

Wave Power

When winds blow over the world’s oceans, theycause waves. The power in such waves as theygradually build up over very long distances can begreat, as anyone watching that power eventuallybeing dissipated on a beach will appreciate. Awide variety of technologies for harnessing thepower of waves have been developed over the pastfew decades (see Duckers 2004). One example isthe oscillating water column (OWC). Here therise and fall of the waves inside an enclosed cham-ber alternately blows and sucks air through an airturbine, which is coupled to a generator to pro-duce electricity. Another promising wave energyconversion system, currently undergoing tests offPortugal, is the Pelamis (sea snake), developed bythe Scottish company Ocean Power Delivery. It isoriented head-on to the wave front and consists ofa series of long, floating steel tubes connected byarticulating hydraulic joints. The varying pressurein the hydraulic joints, as the passing waves movethe cylinders relative to one another, powershydraulic motors, which in turn drive electricalgenerators.

1,000 2,500

2,000

1,500

1,000

100

0

900

800

700

600

500

400

300

200

100

001 02 03 04 05 06 07 08 090089 90 91 92 93 94 95 96 97 98 99

UKSwedenOther (Finland, Belgium, China)

DenmarkNetherlandsGermany

Cumulative global capacity

Ann

ual c

apac

ity (

MW

)

Cum

ulat

ive

capa

city

(M

W)

Source: BWEA 2009

Figure 2.4. Annual offshore wind capacity buildup in the United Kingdom, Sweden, Denmark, the Nether-lands, and Germany, plus total cumulative installed capacity, 1989–2009

Renewable EnergyTechnologies for Electricity Generation 19

Biofuel

Since prehistoric times, human beings have har-nessed the power of fire by burning wood to cre-ate warmth and light and to cook food. Wood iscreated by photosynthesis in the leaves of plants.Photosynthesis is a process powered by solarenergy, in which atmospheric carbon dioxide andwater are converted into carbohydrates (com-pounds of carbon, oxygen, and hydrogen) in theplant’s leaves and stems. These, in the form ofwood or other biomass, can be used as fuels—called “biofuels.”

Wood is still widely used as a fuel in manyparts of the developing world. In some countries,other biofuels, such as animal dung, are also used.These traditional biofuels are estimated to supplyaround 11% of world primary energy. If the for-ests that provide wood fuel are replanted at thesame rate as they are cut down, then wood fueluse is in principle renewable. When forests aremanaged sustainably in this way, the CO2

absorbed in growing replacement trees shouldequal the CO2 given off when the original treesare burned.

A significant contribution to supplies now comesfrom modern bioenergy power plants. These fea-ture the clean, high-efficiency combustion ofstraw, forestry wastes, or wood chips from treesgrown in special plantations.The heat produced isused either directly or for electricity generation,and often for both purposes. Municipal wastes, alarge proportion of which are biological in origin,are also widely used for heat or electricity produc-tion. Liquid wastes can also be processed inanaerobic digesters to produce methane, whichcan then fuel an internal combustion engine toproduce electricity and heat. As a recent IEAreport observes:

Technologies for producing heat and power frombiomass are already well-developed and fullycommercialised, as are first-generation routes tobiofuels for transport.A wide range of additionalconversion technologies are under development,offering prospects of improved efficiencies, lowercosts and improved environmental performance.However, expansion of bioenergy also poses some

challenges.The potential competition for land andfor raw material with other biomass uses must becarefully managed.The productivity of food andbiomass feedstocks needs to be increased byimproved agricultural practices. Bioenergy mustbecome increasingly competitive with other energysources. Logistics and infrastructure issues must beaddressed, and there is need for further technologi-cal innovation leading to more efficient andcleaner conversion of a more diverse range offeedstocks. (IEA 2009, 2)

Hydroelectricity

Another energy source that has been harnessed byhumanity for many centuries is the power offlowing water, which has been used for millingcorn, pumping, and driving machinery. Theoriginal source of hydroelectric power is solarenergy, which warms the world’s oceans, causingwater to evaporate from them. In the atmosphere,this forms clouds of moisture, which eventuallyfalls back to earth in the form of precipitation.Rain flows down through mountains into streamsand rivers, where its flow can be harnessed usingwater wheels or turbines to generate power. Dur-ing the 20th century, hydropower has grown tobecome one of the world’s principal electricitysources. In 2008, it contributed just over 6% ofworld primary energy (BP 2009).

Tidal Power

The physical principles underlying tidal energyare different from those underlying hydropower,but the technologies used in tidal “barrages” aresimilar to those employed in hydroelectric plants.Tides are caused chiefly by the gravitational pullof the moon on the oceans, although the sun’sgravity also plays a minor role. The output fromtidal energy systems is variable but highly predict-able.

The principal technology for harnessing tidalenergy essentially involves building a low dam, orbarrage, across the estuary of a suitable river. Thebarrage includes inlets that allow the rising sealevels to build up behind it. When the tide hasreached maximum height, the inlets are closed,

20 Godfrey Boyle

and the impounded water is allowed to flow backto the sea in a controlled manner, via a turbine-generator system similar to that used in hydroelec-tric schemes.

The world’s largest tidal energy scheme is atLa Rance, France, with a capacity of 240 MW.There are a few other, smaller tidal plants in vari-ous countries, including Canada, Russia, andChina. The United Kingdom has one of theworld’s best potential sites for a tidal energyscheme, in the Severn Estuary. If built, the largestof the various proposed Severn schemes would berated at around 8,600 MW, much larger than anyother single UK power plant, and could provideabout 5% of current UK electricity demand (seeDECC 2009; SDC 2007).

Another, newer tidal energy technologyinvolves the use of underwater turbines, ratherlike submerged wind turbines, to harness thestrong tidal and oceanic currents that flow in cer-tain coastal regions. In the United Kingdom, a 10kW prototype tidal current turbine was tested atLoch Linne, Scotland, in 1994; a larger, 300 kWprototype was tested off the Devon coast in 2002;and another 300 kW prototype is currently opera-tional in Strangford Lough, Northern Ireland.The technology is still under development, but itsprospects are promising.

Geothermal Energy

Geothermal energy is not derived from solarradiation. Its source is the earth’s internal heat,which originates mainly from the decay of long-lived radioactive elements far below the surface.The most useful geothermal resources occurwhere underground bodies of water, or aquifers,can collect this heat, in areas where volcanic ortectonic activity brings it close to the surface.Theresulting hot water, or in some cases steam, is usedfor electricity generation where its temperature issufficiently high to make this feasible, such as inItaly, Iceland, New Zealand, the United States,and the Philippines, and for direct heating inmany other countries.

If geothermal heat is extracted in a particularlocation at a rate that does not exceed the rate atwhich it is being replenished from deep within

the earth, it is a renewable energy source. But inmany cases, this is not so: the geothermal heat isin effect being “mined” and will be depletedlocally in perhaps a few years or decades.

In regions of the world where geothermalaquifers are not readily accessible, it may never-theless be possible to harness the earth’sgeothermal heat using enhanced geothermal, or“hot dry rock,” technologies. These involve drill-ing down to considerable depths to access the hotrocks, then fracturing them using high-pressurewater. This creates a form of heat exchanger intowhich cold water from the surface can be pumpedand from which hot water can be extracted forheating or electricity production. The EU cur-rently has a pilot project to develop this techniqueat Soulz, France (Dettmer 2009).

Costs and Economicsof RenewablesThe cost of energy generated by an energy sourcecan be considered to be composed of three mainelements: the capital costs, operation and mainte-nance (O&M) costs, and fuel costs. In the case ofmost renewable energy sources, the “fuel” is free(except for biofuels, which need to be grown andharvested), so the main cost elements are the capi-tal and O&M costs.

The capital cost of a renewable energy plant isnormally repaid over a period of years (oftenshorter than the plant lifetime), and interest ischarged to give the capital provider a satisfactoryrate of return on the investment. Private investorsusually expect a higher rate of return and repay-ment over a shorter period than do public sectorinvestors. To these regular repayments of capitalcosts, plus interest, must be added the O&M costs,which vary widely among different types of plant,to arrive at the total annual cost of operating theplant. This can then be divided by the annualenergy output (say, in kWh) to give an averageproduction cost of energy from the plant (say, incents per kWh). However, this is not usually thecost paid by the final energy customer. The finalretail price will include the costs of distribution

Renewable EnergyTechnologies for Electricity Generation 21

and marketing, plus the profits made by the vari-ous intermediaries in the distribution chain fromproducer to consumer.

Calculations of the estimated future capitalcosts of renewable energy plant, or of the energythey generate, are often adjusted for the effects ofinflation and are thus expressed in terms of costsor prices in a given year—for example, in 2009$.In a world of floating exchange rates, if cost calcu-lations involve a conversion from one currency toanother, they will be valid for only a limited time.In calculating the future capital costs of renewableenergy equipment produced in substantial quanti-ties, such as wind turbines or solar PV panels,analysts often assume that a reduction in unit costswill occur in the future as a result of the econo-mies gained through volume production—sometimes called the “learning curve” effect.

It will be obvious from the above considera-tions that the wide variety of ways in which costsand prices can be calculated can give rise to con-fusion.

Resource-Cost Curves for Renewables

An insight into the likely costs of electricity fromrenewable sources can be gained by examiningresource-cost curves. These show the quantity of

the resource estimated to be available fromselected renewables at various levels of cost. Fig-ure 2.5 is a recent example (DECC 2009). Thischart covers costs for heat and transport fuel, aswell as for electricity. From this figure, it can beseen that 25% of UK electricity could be pro-duced from renewables by 2020 for approximately£60 ($90) per MWh. Producing 28% of electric-ity from renewables would increase this to about£70 ($105) per MWh, and further increasing theproportion to 32% would cost around £100($150) per MWh.

The bar at the far right in the figure shows thecost of supplying 2% of electricity from small-scale renewable sources, estimated at about £150($225) per MWh.The price of electricity genera-tion from small-scale sources varies widely, as canbe seen from Table 2.4, which shows the pricesproposed under new feed-in tariffs for small- andmedium-scale electricity generation (under 5MW) in the United Kingdom from 2010. Thehighest price is for domestic-scale photovoltaicsunder 4 kW, at 36.5 pence (54.8 cents) per kWh;the lowest is for medium-scale biomass, hydro,and wind plants, at 4.5 pence (6.8 cents) perkWh. These tariffs are designed to provide a rea-sonable rate of return to investors in the various

0 28 45 62 72 82 116

131

142

145

157

168

174

TWh

Tota

l res

ourc

e co

st £

/MW

h, 2

008

5.5%

Hea

t

8.4%

Hea

t

12%

Hea

t

32%

Ren

ewab

le

10%

Hea

t

8% T

rans

port

10%

Tra

nspo

rt

12%

Tra

nspo

rt

24%

Ren

ewab

le e

lect

ricity

28%

Ren

ewab

leel

ectr

icity

elec

tric

ity

29%

Ren

ewab

le e

lect

ricity160

140

120

100

80

60

40

20

0

–20

2% s

mal

l-sca

le e

lect

ricity

Sources: UK Department of Energy and Climate Change (DECC) and Department for Transport (DfT) internal analysis, based on Redpoint/Trilemma 2009; NERA/AEA 2009; and Element/Pöyry 2009

Figure 2.5. Marginal resource costs for different levels of renewable generation by sector in the UnitedKingdom in 2020

22 Godfrey Boyle

technologies and should therefore give a reason-ably good indication of the current costs/prices ofelectricity from small- to medium-scale plants inUK conditions.

The “Portfolio Effect” and the Economicsof Renewables

Proponents of wind power and other forms ofrenewable energy have recently been pointing outthat because their “fuel” is free, this has a signifi-cant effect on their economics—one that is notusually taken into account in conventional analy-ses. When wind electricity, for example, is avail-able to the grid, its effective marginal cost to theelectricity system is near zero; thus it should dis-

place more costly power from conventional plants.When wind power is contributing significantquantities of energy, it can also reduce the spotprice of fossil fuels. Because it requires no fuel,wind energy if used significantly also reduces anation’s vulnerability to imported fossil-fuel priceincreases, with beneficial effects on GDP, andbecause of its fixed cost, it reduces the system riskin a nation’s generation portfolio. (For furtherdetails, see Auwerbuch 2009; Bruce 2009).

Some of the key economic and technologicalcharacteristics of the renewable electricity gener-ating technologies described above are summa-rized in Table 2.5: their cost drivers (the relativeroles of capital, O&M, and fuel costs); their levelsof maturity; the potential for scale economies; and

Table 2.4. Prices paid under proposed UK feed-in tariffs for electricity from 2010 and annual digression ratesfrom 2010 onward

Technology Scale Proposed initial tariff(p/kWh)

Annual digression(%)

Anaerobic digestion Electricity only 9.0 0

Anaerobic digestion CHP 11.5 0

Biomass < 50 kW 9.0 0

Biomass 50 kW–5 MW 4.5 0

Biomass CHP 9.0 0

Hydro < 10 kW 17.0 0

Hydro 10–100 kW 12.0 0

Hydro 100 kW–1 MW 8.5 0

Hydro 1–5 MW 4.5 0

PV < 4 kW (new build) 31.0 7

PV < 4 kW (retrofit) 36.5 7

PV 4–10 kW 31.0 7

PV 10–100 kW 28.0 7

PV 100 kW–5 MW 26.0 7

PV Stand-alone system 26.0 7

Wind < 1.5 kW 30.5 4

Wind 1.5–15 kW 23.0 3

Wind 15–50 kW 20.5 3

Wind 50–250 kW 18.0 0

Wind 250–500 kW 16.0 0

Wind 500 kW–5 MW 4.5 0

Existing microgeneratorstransferred from renewablesobligation

9.0 N/A

Source: DECC 2009Note: p/kWh = pence per kilowatt-hour; 1 pence = 1.5 cents (U.S.) as of this writing

Renewable EnergyTechnologies for Electricity Generation 23

the potential for near-term technologicaladvances. The factors that are likely to lead toreductions in the cost of electricity fromrenewables vary widely among technologies: insome areas, significant potential exists for techno-logical breakthroughs that could result in majorcost reductions; in other, more mature areas,steady incremental improvements will graduallyreduce costs. In addition, many renewable tech-nologies benefit from economies of scale, includ-ing the economies of physical scale, such as thoseachievable from building larger wind turbines,and the economies of series production, whichare achievable irrespective of physical scale.

Integration of Renewables intoElectricity SystemsSome renewable electricity sources, such asbioenergy, hydroelectricity, and geothermal

power, are dispatchable—in other words, likeconventional power plants, their output can,within limits, normally be made available quicklyto electricity system operators in response todemand.The outputs of other renewables, such asthose based on wind, wave, and solar power, arenot so readily dispatchable: they are weather-dependent and therefore variable—unless theyincorporate storage, as is the case with some solarthermal-electric systems. Wind, wave, and solarpower are sometimes described as intermittent,but a more accurate term for their output is vari-able. Tidal power is in an intermediate category:its output, though variable, is highly predictable.

But the variability of these renewable sourcesis not the major problem that some have sug-gested. National electricity systems are alreadydesigned to cope with the major fluctuations inboth demand and supply that occur on varioustimescales, from minutes to years. Introducing avariable renewable source introduces an additional

Table 2.5. Summary of economic characteristics of renewable electricity generation technologies

Technology Cost drivers Maturity level Potential for scaleeconomies

Potential for near-term technologicaladvances

Solar ponds Capital, O&M High Low Low

Ocean thermal Capital, O&M Med Med Med

Solar chimney Capital Low Med Low

Concentrating solar: trough Capital, O&M Med Med Med

Concentrating solar: powertower

Capital, O&M Med Med Med

Concentrating solar: dishStirling

Capital, O&M Med High Med

PV: crystalline Capital Med High Med

PV: thin film Capital Med High High

Concentrating PV Capital, O&M Med Med Med

Wind onshore Capital High Med Low

Wind offshore Capital, O&M Med Med Med

Wave Capital, O&M Low Med Med

Hydro (large) Capital High Low Low

Biofuel electricity Capital, O&M, fuel High? Med Med

Tidal barrage Capital High Low Low

Tidal stream Capital, O&M Low Med Med

Geo aquifers Capital High Low Low

Geo HDR Capital Low Low Med

Source: Author’s estimates

24 Godfrey Boyle

element of uncertainty into the system, but this ismanageable by system operators in a manner simi-lar to the way in which the uncertainty in bothdemand and conventional supplies is already man-aged. With low levels of penetration of variablerenewables on the system, the additional level ofuncertainty is small; it becomes larger as the pen-etration level increases.

For simplicity, this discussion will concentrateon the integration of wind power into electricitysystems; the integration of other renewables, suchas solar and wave power, is amenable to similarsolutions. The topic of renewable electricity inte-gration is covered in much greater detail in Boyle(2007).

In existing national electricity systems, variouskinds of backup power supplies are already pro-vided. In essence, these consist of short- and long-term supplies. Short-term backup supplies areprovided to maintain frequency and voltage stabil-ity on the network. These include fossil-fueledspinning reserve plants, operating at below theirrated power but able to increase output at veryshort notice; hydropower plants, includingpumped storage schemes, the output of which canbe made available quickly; and large diesel or gasturbine generators, which are able to be broughtonline within a few minutes.

Long-term backup supplies are those requiredto ensure security of supply in the event of majordisruptions, such as the simultaneous failure ofseveral large power stations. They include olderfossil-fueled plants, usually less efficient thanmodern plants and therefore infrequently used,that can be brought back into operation if neces-sary within a few hours or days. In addition, manynational systems have interconnectors to the elec-tricity systems of nearby countries; these can becalled on to provide additional power if needed.

How much would the additional backup sup-plies needed for substantial levels of wind contri-bution cost? A recent UK Energy Research Cen-tre study (UKERC 2006) reviewed a wide rangeof evidence from around the world and concludedthat for a 10% to 20% contribution of wind to theUK system, the additional cost would be £2.5 to£3 ($3.76 to $4.50) per MWh, rising to £3 to£3.5 ($4.50 to $5.26) per MWh for a 20% to 45%

wind contribution.These costs are relatively smallin relation to total generation costs.

Wind Forecasting

The cost of providing additional backup to copewith the additional variability of substantial quan-tities of wind generation can be significantlyreduced through the use of wind forecastingmethods. Figure 2.6 shows an example of a day-ahead wind forecast for one month in Germany,using techniques developed at the University ofKassel. These employ a combination of tech-niques: numerical weather predictions; analysis ofphysical processes; statistical time series analysis;and learning techniques based on artificial neuralnetworks (Lange et al 2007). As can be seen, thedifference between forecast and monitored poweroutput is relatively small. This enables backupplants to be scheduled efficiently and at minimumcost.

Storage

It is often asserted that energy storage is needed tocope with the variability of wind power and otherfluctuating renewable energy sources, though thisis not necessarily the case.Various large-scale stor-age technologies are available, ranging frompumped hydro schemes to compressed-air storageto vanadium redox (reduction-oxidation) batterystorage, but their costs are high. However, elec-tricity storage in large, integrated national elec-tricity supply systems is desirable only if it costsless than providing more conventional backupsupplies—or if storage can provide additionalbenefits to the system operator or to users, as maybe the case with battery storage in electric vehi-cles.

Electricity storage in electric vehicle (EV)batteries may become increasingly attractive assuch vehicles enter widespread use in order toreduce urban air pollution and emissions ofgreenhouse gases. EV batteries would be chargedmainly at night, when electricity systems nor-mally have spare capacity, and could readily absorbany surpluses of power from wind or otherrenewables. The batteries would be discharged

Renewable EnergyTechnologies for Electricity Generation 25

when the vehicles are driven during the day,which should not make substantial demands onthe electricity system, though some daytimecharging would be inevitable.

Intercontinental Integration ofRenewable Electricity: The EuropeanSupergrid and Desertec

The large-scale spatial dispersion of wind genera-tion can markedly reduce the variability of theresource. The European supergrid is a proposal tocreate an electrical network linking large offshorewind farms in high-wind regions of Europe toneighboring countries and beyond. The linkswould use high-voltage direct current (HVDC)technology, which enables power to be transmit-ted over long distances with low losses. At thepoint of connection to a country’s electricity net-work, the power would be converted to conven-tional alternating current (AC).

The proponents of the supergrid claim it ischeaper to generate electricity in high-wind areasand transmit it to the countries where it is neededthan to generate it at lower-wind sites in areaswith poorer resources. They also point out that asupergrid would facilitate a Europe-wide marketfor electricity (Hurley et al 2007). The govern-

ments of Germany, France, Belgium, the Nether-lands, Luxembourg, Denmark, Sweden, Ireland,and the United Kingdom are developing plans, tobe announced at the end of 2010, to build the firstphase of the supergrid in the North Sea region.

Even greater smoothing effects may beachievable if wind power sources can be inte-grated over a still larger area. Gregor Czisch of theUniversity of Kassel has modeled a series ofdetailed scenarios in which he has combined themonthly wind power production of good windsites in the EU and Norway with sites in North-ern Russia, western Siberia, southern Morocco,and Mauritania. When the combined output iscompared with the average monthly electricitydemand in the EU and Norway, supply anddemand are quite well matched, except for a smallshortfall around September–October (see Czisch2006; Murray 2009).

A similarly ambitious proposal, the DesertecIndustrial Initiative, has been put forward by theTrans-Mediterranean Renewable Energy Coop-eration (TREC) and German Aerospace Center(DLR). The proposal, backed by 12 major Euro-pean companies including Munich Re, theworld’s largest reinsurance company, involveslinking a wide range of renewable electricitysources right across Europe, the Middle East, and

Pow

er (G

W)

16

14

12

10

8

6

4

2

01.01 5.01 9.01 14.01 19.01 23.01 27.01 31.01

Forecasted powerMonitored power

Source: Lange et al (2007)

Figure 2.6 Day-ahead forecast of wind power output compared with monitored output, for one month inGermany

26 Godfrey Boyle

North Africa (EUMENA), again using HVDClinks to minimize power losses in long-distanceelectricity transmission. The system would even-tually combine the outputs of wind, biofuel,hydro, concentrating solar, photovoltaic, andgeothermal power plants. The combined outputwould be sufficient to meet the total electricityneeds of the entire area (see Desertec 2009;D-MTEC 2009). These proposals have recentlybeen endorsed in principle by the EuropeanUnion in its Mediterranean Solar Plan (Ferrero-Waldner 2009).

ConclusionsRenewable sources are in principle capable ofsupplying all of humanity’s energy requirementsmany times over. Many relatively new technolo-gies for renewable electricity generation, such assolar and wind power, are maturing and growingrapidly. Some others, such as hydroelectricity, arealready mature. Renewables are already makingsignificant inputs into the electricity supplies ofsome countries and have the potential to providemuch more power worldwide in comingdecades—provided that appropriate policy sup-port measures are implemented (for examples, seeREN21 2007, 2009).

The current costs of power from renewablesare often higher than from conventional sources,but the costs of the former are steadily falling andthose of the latter are likely to rise. In addition,the costs of conventional fossil fuels do not takeinto account the external costs to society of theiruse, such as their contribution to global climatechange. If all such costs were included, mostrenewable energy sources would already be cost-competitive (see Sovakool and Watts 2009).Moreover, substantial future cost reductions mayoccur for many renewables as a result of mass pro-duction and improvements in technology.

Note1. The subject of renewable energy is covered in

much greater detail in Boyle (2004).

References and FurtherInformationAEE (Spanish Wind Association). 2009. Wind Power

2008: Sector’s Yearbook, Analysis and Data. Madrid:Spanish Wind Association.

Alsema, E.A., and E. Niewlaar. 2000. Energy Viabilityof Photovoltaic Systems. Energy Policy 28: 999–1010.

Auwerbuch, S. 2009. Value of Wind Compared to GasGeneration: A Risk-Adjusted Approach. In The Eco-nomics of Wind Energy, edited by Krohn, S.,Morthorst, P.E., and Auwerbuch, S. Brussels: Euro-pean Wind Energy Association, 115–122, Section5.1.

Blake, W. 1790. The Marriage of Heaven and Hell.Various editions: electronic edition available atwww.gailgastfield.com/mhh/mhh.html (accessed24 March 2010).

BMU (Federal Ministry for the Environment, NatureConservation and Nuclear Safety). 2009a. Develop-ment of Electricity Generation from Renewable Energy.Bonn, Germany: BMU.

———. 2009b. Renewable Energy Sources in Figures: Sta-tus, June 2008. Bonn, Germany: BMU.

Boyle, G., ed. 2004. Renewable Energy: Power for a Sus-tainable Future. 2nd ed. Oxford. Oxford UniversityPress.

———. 2007. Renewable Electricity and the Grid: TheChallenge ofVariability. London: Earthscan.

BP (British Petroleum). 2009. BP Statistical Review ofWorld Energy, June 2009. London: BP.

Bruce, A. 2009. The Value of Wind: How MoreRenewables Means Lower Electricity Bills. RealPower (April–June): 9–11.

BWEA. 2009. Offshore Wind: Staying on Track. London:British Wind Energy, 8pp.

Carbon Trust. 2008. OffshoreWind Power: Big Challenge,Big Opportunity. London: Carbon Trust.

Czisch, G. 2006. Least-Cost European/TranseuropeanElectricity Supply Entirely with Renewable Energies.Kassel, Germany: ISET.

DECC (Department of Energy and Climate Change).2009. Severn Tidal Power Consultation: GovernmentResponse. London: DECC.

Desertec. 2009 12 Companies Plan Establishment of aDesertec Industrial Initiative.www.desertec.org/ (accessed January 15, 2010).

Dettmer, R. 2009. Return to Rock. Engineering andTechnology 4 (16): 48–50.

D-MTEC (Department of Management, Technologyand Economics). 2009. TheVision of a Super Grid and

Renewable EnergyTechnologies for Electricity Generation 27

Its Impact on the Market for Renewable Energy. Zurich:Swiss Federal Institute of Technology.

Duckers, L. 2004. Wave Energy. In Renewable Energy:Power for a Sustainable Future, 2nd ed., edited by G.Boyle. Oxford: Oxford University Press, 298–340.

EEA (European Environment Agency). 2009. Europe’sOnshore and OffshoreWind Energy Potential:An Assess-ment of Environmental and Economic Constraints. Tech-nical report 6/2009. Copenhagen: EEA.

Element/Pöyry. 2009. Design of a Feed-In Target forSub-5 MW Electricity in GB: Final Report. London:Pöyry, 118pp.

EPIA (European Photovoltaic Industries Association).2009. Set for 2020: Executive Summary. Brussels:EPIA.

EPIA/Greenpeace. 2008. Solar Generation V—2008:Solar Electricity for Over One Billion People and TwoMillion Jobs by 2020. Brussels: EPIA.

European Commission (EC). 2008a. European Energyand Transport Trends to 2030 (update 2007). http://ec.europa.eu/dgs/energy_transport/figures/trends_2030_update_2007/energy_transport_trends_2030_update_2007_en.pdf

———. 2008b. Model Based Analysis of the 2008 EUPolicy Package on Climate Change and Renewables.Brussels: Report to DG ENV.

Everett, R. 2004. Solar Thermal Energy. In RenewableEnergy: Power for a Sustainable Future, 2nd ed., editedby G. Boyle. Oxford: Oxford University Press, 51.

EWEA (European Wind Energy Association). 2009.Wind Energy:The Facts: Executive Summary. Brussels:EWEA.

Ferrero-Waldner, B. 2009. The Mediterranean SolarPlan: A Necessity, Not an Option. Speech 09/60 byEU commissioner for External Relations and Euro-pean Neighbourhood Policy, European Union Sus-tainable Energy Week, February 2009, Brussels.

Greenpeace. 2005. Energy Revolution:A Sustainable Path-way to a Clean Energy Future for Europe. Brussels,Greenpeace International.

GWEC (Global Wind Energy Council). 2009. GlobalWind 2008 Report. Brussels: GWEC.

Hopwood, D. 2009. The New Hot Ticket. EngineeringandTechnology 15 (Sept. 12–25): 48–51.

Hurley, B. Hughes, P. and Giebel, G. 2007. ReliablePower, Wind Variability and Offshore Grids inEurope. In Renewable Electricity and the Grid: TheChallenge of Variability, edited by G. Boyle. London:Earthscan, 181–199.

IEA (International Energy Agency). 2002. RenewableEnergy: Into the Mainstream. Paris: IEA.

———. 2008. Trends in Photovoltaic Applications: SurveyReport of Selected IEA Countries between 1992 and2007. Paris: IEA.

———. 2009. Bioenergy—A Sustainable and ReliableEnergy Source:A review of Status and Prospects, ExecutiveSummary. Paris: IEA.

James, T. 2009. Energy from the Deep. Engineering andTechnology 4 (14): 50–53.

Knies, G., Timm, G., Führ, F., Wolff, G., Schön, M.,Straub, M., Nokraschy, H. and Steinmetz, O. (eds).2009. An Overview of the DESERTEC Concept.DESERTEC Red Paper. Berlin: DESERTECFoundation.

Kurokawa, K., Komoto, K., van der Vleuten, P. andFaiman, D. (eds). 2007. Energy from the Desert. Lon-don: Earthscan/International Energy Agency.

Lange, B., Rohrig, K., Schlögl, F., Cali, U. and Jursa,R. 2007. Wind Power Forecasting. In RenewableElectricity and the Grid: The Challenge of Variability,edited by G. Boyle. London: Earthscan, 95–120.

Lockheed Martin. No date. Ocean Thermal EnergyConversion (OTEC). www.lockheedmartin.com/products/OTEC/(accessed 26 March 2010).

Marsh, G. 2009. Power Stations: Can Solar Join the BigHitters? Renewable Energy Focus, (May–June): 50–58.

Murray, J. 2009. Wind Holds Key to European SuperGrid. Interview with Dr Gregor Czisch. Real Power18 (Oct.—Dec.): 31–32.

NERA/AEA (NERA Economic Consulting/AEATechnology). 2009. The UK Supply Curve for Renew-able Heat. Report for the Department of Energy andClimate Change, July 2009, Ref: URN 09D/689(DECC). London: NERA Economic Consulting.

NREL (National Renewable Energy Laboratory).2008. 20% Wind Energy by 2030. NREL ReportDOE/GO-102008–2567, July. National RenewableEnergy Laboratory.

Redpoint/Trilemma. 2009. Implementation of the EU2020 RenewablesTarget in the UK Electricity Sector: ROReform. Report for the Department of Energy andClimate Change, June 2009, URN 09D/702. Lon-don: Redpoint Energy Ltd.

REN21 (Renewable Energy Policy Network for the21st Century). 2007. Renewables Global Status Report.Paris: REN21.

———. 2009. Renewables Global Status Report: 2009Update. Paris: REN21.

Renew. 2009. CSP Moves Ahead to 3GW. Renew 180:12.

Schlaich, J. 1995. The Solar Chimney: Electricity from theSun. Stuttgart: Edition Axel Menges.

28 Godfrey Boyle

SDC (Sustainable Development Commission). 2007.Turning the Tide: Tidal Power in the UK. London:SDC.

Sovakool, B., and C. Watts. 2009. Going CompletelyRenewable: Is It Possible (Let Alone Desirable)? Elec-tricity Journal 22 (4): 95–111.

Twidell, J., and A. Weir. 1986. Renewable EnergyResources. London: E & FN Spon.

UKERC (UK Energy Research Centre). 2006. TheCosts and Impacts of Intermittency. London: UKERC.

Renewable EnergyTechnologies for Electricity Generation 29

Part II

Renewables, Climate Change, andEnergy Policy

3

Renewables Forecasts in aLow-Carbon World:A Brief OverviewErin T. Mansur

This chapter examines how concerns forclimate—coupled with policymakers’ and

voters’ other preferences—are expected to affectthe growth of renewables, particularly in theUnited States and European Union (EU). Itbegins with a review of the recent growth inrenewables and a discussion of current carbonpolicies, then looks at predictions of levels ofinvestment in renewables in a low-carbon world.Within the next 20 years, some authors predictthat 20% to 40% of electricity will come fromrenewable resources, in an attempt to mitigate theimpacts of climate change. Given the wide rangeof assumptions regarding carbon policies, eco-nomic growth, the responsiveness of the economyto carbon policy, innovation in renewables, andother modeling approaches, it is hard to general-ize among these “black box” models. Their com-plexity and continuous updating make a detailedpeer review process necessary, though this isbeyond the scope of this chapter.

Under continued dependence on fossil fuelsin a business-as-usual scenario, climate scientistspredict significant increases in average tempera-ture and sea level, as well as many other climaticresponses. Climate change is expected toadversely affect many economic sectors, includingagriculture, forestry, insurance, health, tourism,and energy. Furthermore, it is likely to have many

notable nonmarket effects, such as the loss of floraand fauna and an increase in human morbidityand mortality. The Intergovernmental Panel onClimate Change (IPCC 2007) and Stern (2006)review the science and economic consequences ofclimate change.

These concerns may be addressed throughseveral mechanisms for mitigating climate change.Broadly, we may change how we produce and useelectricity, space heating, transportation, andother energy applications; we may sequestergreenhouse gases (for example, by reducing defor-estation or increasing reforestation); or we mayeven consider climate geoengineering.1 Withinthe electricity sector, methods for reducinggreenhouse gases, such as carbon dioxide (CO2),include both demand-side (conservation andenergy efficiency) and supply-side options. Sup-ply options include switching from coal to lesscarbon-intensive conventional technologies suchas natural gas, large hydroelectric, and nuclearpower; continuing to use coal but reducing thecarbon emissions with abatement technologiessuch as carbon capture and storage (CCS); or, thefocus of this book, turning to alternative renew-able sources of electricity, including wind, solar,geothermal, and small hydropower.

From an economic perspective, greenhousegases are global, stock externalities. (Chapter 5

covers this topic in more depth.) In brief, anexternality is a cost (or benefit) incurred by othersnot involved in a market transaction. Globalexternalities are those where the location of thesource of the pollutant is irrelevant to the dam-ages. For example, a coal-fired power plant inChina may be the source of released carbon diox-ide emissions, but the effects of these emissionswill be experienced all over the world. Further-more, these effects will be the same as had thecarbon dioxide emissions come from natural gaspower plants in, say, England. Stock pollutants arelong-lived, and their damages from current emis-sions may continue for years to come. For exam-ple, carbon dioxide emitted today will remain inthe upper atmosphere for 100 years or more(IPCC 2007).2 For stock pollutants, the optimalpolicy is clear in theory: each ton of emissionsreleased is charged the net present value of allfuture damages that it causes.3 Specifically, a pol-lution tax can achieve the efficient outcome. Acap-and-trade regulation can also be efficient ifthe permit price equals this tax.

A pollution tax on greenhouse gases wouldgive firms and consumers an incentive to changebehavior in many ways: switching to less carbon-intensive technologies, conserving electricity,driving less, and so forth. Firms investing inrenewables would be just one of many responses.Are renewables the best option for addressing cli-mate change? In order to answer this question,one would need to know the relative marginalcosts, both private and external, incurred by eachoption that reduces greenhouse gas emissions.

Whether or not producing electricity fromrenewable resources is the cheapest way to reducegreenhouse gas emissions, many governmentagencies and other authors predict that renewableswill be a major contributor to a low-carbonfuture. Investing in renewables is a popularresponse to climate concerns, in part, perhaps,because of the other externalities that it addresses.First, other energy sources have additional nega-tive externalities such as conventional air pollu-tion, water pollution, and nuclear waste. Notethat if regulation results in marginal social costsequaling marginal social benefits of these otherenergy sources, these externalities will be inter-

nalized and no further regulation is required. Sec-ond, positive technological spillovers may resultfrom investing in renewables from which otherindustries, or firms in the same industry, benefit.4

Learning benefits within a firm, which mayaccount for the largest gains, are not positiveexternalities. Finally, subsidizing technology isseemingly more politically palatable than taxingfirms and consumers for polluting.

This chapter provides an overview of currentelectricity sources and energy and climate policiesaffecting the recent growth in investment inrenewables, then looks at the direction of futurecarbon policies.This is followed by a review of theliterature on predictions of investment inrenewables under various low-carbon scenarios.

Current Energy SourcesElectricity production is dominated by fossil fuels.In the United States, coal accounts for more than50% of electricity production, and natural gasaccounts for another 20% (EIA 2009b). Conven-tional carbon-free technologies include nuclearpower, which produces 20% of U.S. electricity,and hydroelectric power, which provides 6%.TheEuropean Union is less dependent on fossil fuels:here, only 54% of power comes from coal, oil, ornatural gas (Eurostat 2009). Nuclear (28%) andhydropower (10%) are larger in the EU than inthe United States.

Nonhydropower renewables, such as wind,solar, geothermal, tidal, and biomass, account forless than 7% of electricity produced in either theUnited States or the EU. In 2008, the UnitedStates produced 4,111 terawatt-hours (TWh) ofelectricity.5 Of that, only 3% of was fromnonhydropower renewables, including wind (52TWh), wood (39 TWh), and other renewablesources (33 TWh) (EIA 2009b). In 2007, theEuropean Union (EU-27) produced 6.5% of itselectricity from wind (104 TWh), wood (52TWh), and other renewables (55 TWh) (Eurostat2009). Although the United States produces over20% more total electricity than the EU, the Euro-pean Union produced more power from renew-able sources than did the United States.

34 ErinT. Mansur

While renewables are still a small share of totalelectricity production, they have seen enormousgrowth in the past two decades. In 1990, theOrganisation for Economic Co-operation andDevelopment (OECD) countries had a totalcapacity of 2.4 gigawatts (GW) of wind powerand 0.5 GW of solar, including both photovoltaics(PV) and thermal (IEA 2004).6 By 2008, however,wind capacity had reached 17 GW in Spain alone(WWEA 2009). Table 3.1 reports installed wind

capacity for the 10 countries with the most windcapacity as well as worldwide totals for 2007 and2008.

Worldwide, wind capacity reached 121 GWin 2008, an increase of 29% over the previous year(WWEA 2009).This annual growth rate is typicalof this industry over the past decade. Figure 3.1plots annual worldwide capacity and growth ratefrom 1997 to 2008.

140,000

120,000

100,000

80,000

60,000

Cap

acity

(M

W)

Per

cent

age

grow

th (

%)

40,000

20,000

01996 1998 2000 2002 2004

Year

Capacity (MW) Annual growth

2006 2008 20100

510

15

20

25

30

35

40

45

Source: WWEA 2009

Figure 3.1. Wind power worldwide capacity and annual growth rate, 1997–2008 (in MW)

Table 3.1. Total installed capacity of wind energy in 2007 and 2008 (in GW)

Country 2007 2008 Change Growth rate

U.S. 16.8 25.2 8.4 50%

Germany 22.2 23.9 1.7 7%

Spain 15.1 16.7 1.6 11%

China 5.9 12.2 6.3 107%

India 7.9 9.6 1.7 22%

Italy 2.7 3.7 1.0 37%

France 2.5 3.4 0.9 39%

UK 2.4 3.3 0.9 38%

Denmark 3.1 3.2 0.0 1%

Portugal 2.1 2.9 0.7 34%

Rest of world 13.1 17.1 4.0 30%

Total 93.9 121.2 27.3 29%

Source: WWEA 2009

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 35

This impressive growth rate has also beenexhibited in other renewables as well. As recentlyas 2003, worldwide capacity in solar PV was only1.3 GW, but as of 2008, it had reached 13 GW.This tenfold increase can be attributed primarilyto Germany and Spain, which account for two-thirds of the world’s solar PV capacity. Overall,total renewables (excluding large hydropower) arenow at 280 GW (REN21 2009).Table 3.2 reportsthe capacity of various renewable resources bycountry.

Historically, the two major impediments torenewables have been costs and intermittency.Thecost of wind power has dropped substantially overthe past decade (IEA 2004). Now, with the pro-duction tax credit (discussed below), wind powercosts less than 5 cents per kWh in the UnitedStates and can compete with the cost of buildingnew fossil-fuel-fired power plants (Wiser andBolinger 2008). Photovoltaic solar power is anorder of magnitude more expensive: the averageinstalled cost (excluding direct financial incentivesor tax credits) in 2007 was $7.60 per watt in theUnited States (Wiser et al. 2009), or about 30 to50 cents per kWh, depending on the interest rateand capacity factor. These costs have fallen sub-stantially over the past few decades. Even fromsummer 2008 to summer 2009, they fell another40% with the entry of China into the market

(Galbraith 2009). Although dramatically cheaperthan in the past, solar PV remains an expensiveoption for addressing climate change (Borenstein2008). In comparison, the average installed cost ofwind in 2007 was $1.70 per watt in the UnitedStates (Wiser and Bolinger 2008). The IPCC(2007) projects that the costs of renewables willcontinue to fall, and that by 2030, renewables(including wind, solar PV, and biomass) will seecapital costs under $1.20 per watt. (See Chapter 2for further discussion on renewable technologiesand their costs.)

Recent Renewables andClimate PolicyThe United States has been setting policies topromote investment in renewables for more thanthree decades. In 1978, the Public Utility Regu-latory Policies Act first provided subsidies forinvesting in renewables. In some states, such asCalifornia, this led to substantial investments. (Fora review of the history of U.S. renewables policies,see Chapter 11 of this book, as well as Martinot etal. 2005.)

The growth in U.S. renewables has beendriven mainly by federal tax credit incentives.

Table 3.2. Capacity of renewable resources by type and country in 2008 (in GW)

WorldDevelopingcountries EU-27 China U.S. Germany Spain India Japan

Wind 121 24 65 12.2 25.2 23.9 16.8 9.6 1.9

Smallhydropower 85 65 12 60 3 1.7 1.8 2 3.5

Biomass 52 25 15 3.6 8 3 0.4 1.5 0.1

Solar PV 13 0.1 9.5 0.1 0.7 5.4 3.3 0 2

Geothermal 10 4.8 0.8 0 3 0 0 0 0.5

Solarthermal 0.5 0 0.1 0 0.4 0 0.1 0 0

Ocean(tidal) 0.3 0 0.3 0 0 0 0 0 0

Totalrenewablesª

280 119 96 76 40 34 22 13 8

Source: REN21 2009ªExcluding large hydropower

36 ErinT. Mansur

With the passage of the American Recovery andReinvestment Act of 2009 (ARRA; see U.S.Congress 2009), wind, small hydropower,geothermal, biomass, municipal solid waste, andmarine energy sources may be eligible to earn asubsidy of $21/MWh under the production taxcredit (PTC). Many of these renewable resourcescan forgo the PTC and become eligible for the30% investment tax credit (ITC).7 As a thirdoption, a Department of Treasury cash grant canbe selected in place of the PTC or ITC. Metcalf(2009b) provides greater detail on the tax incen-tives for renewables. U.S. wind investment is quiteelastic to changes in the user cost of capital(Metcalf 2009a): the federal PTC has been a sub-stantial driver of wind investment over the pasttwo decades. How these incentives are expectedto affect renewable investment is addressed in thenext section.

A second major driver of renewable invest-ment in the United States is the set of state-levelrenewable portfolio standards (RPSs, discussed inChapter 11). To date, 28 states and the District ofColumbia have implemented mandatory pro-grams that require utilities to purchase a certainpercentage of their power from renewablesources; the definition of the technologies that

comply with the regulation differs from state tostate (see DSIRE 2009). Markets for renewableenergy credits (RECs), or green tags, have arisenwithin some states. These allow firms to complywith the regulations by either directly investing inrenewables or buying credits from other compli-ant sources.

In Europe, some countries have chosenrenewable standards as a policy instrument toachieve the targets in the Renewable ElectricityDirective (2001/77/EC). As in many U.S. RECmarkets, several European countries have imple-mented tradable certificates in order to reachthese goals (European Commission 2009; Nielsenand Jeppesen 2003). Figure 3.2 reports the shareof renewable electricity by country for largehydropower and other renewables in 2007. Nota-bly, over half the electricity generated in Swedenis from large-scale hydropower or otherrenewables, and more than a quarter of Denmark’spower is from nonhydro renewables. The figurealso shows the 2010 objectives. Although thesetargets are greater than the share of electricitygeneration from renewable resources in 2007 formany countries, some, such as Denmark, have justreached the target, and Germany and Hungaryhave even exceeded it.

Source: Author’s calculations based on European Commission 2009 and Eurostat 2009

Figure 3.2. Share of renewable electricity by EU country in 2007 and targets for 2010

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 37

In order to comply with these standards, manycountries have used feed-in tariffs that havespurred investment in renewables. Many of thesesubsidies are on the order of hundreds of euros perMWh. For example, some solar power producersin Spain receive €340 ($463) per MWh underReal Decreto 1578/2008. In 2008, Germany’ssolar feed-in tariff was even greater: some solarproducers received €574 ($782) per MWh underthe Erneuerbare-Energien-Gesetz.8 As a result,Spain and Germany have seen dramatic growth inrenewable investments (see Tables 3.1 and 3.2).(Chapters 13 to 15 discuss the EU experience,focusing on the United Kingdom, Germany, andSpain.)

In addition to renewable standards and feed-intariffs, the EU has implemented a multinational,multi-industry cap-and-trade policy for green-house gases. The EU Emission Trading System(ETS) began in 2005 and regulates about 2 billionmetric tons of CO2 each year. With permit pricesfluctuating between €10 and €30 ($13 and $41) aton, this market is valued at around €37 billion($50 billion) per year.9 (Chapter 6 further dis-cusses emissions trading.)

In addition to costs, another major issue thatrenewable sources face is intermittency. Somerenewable resources, such as wind and solarpower, produce only some of the time and cannotbe depended on to produce reliably. Intermit-tency will not be completely solved until a cheap,reliable mechanism of storing power is developed(NREL 2008). However, at a level of investmentthat is predicted over the next few decades(between 20% and 30% of electricity generation),geographic averaging and natural gas capacity thatserves as backup reserve are likely to dampen theissue of intermittency. This may require newinvestment in natural gas power, especially in areaswhere renewables primarily produce off-peak(Campbell 2009). Complete dependence on cur-rent nonhydroelectric renewable technologiesseems questionable, however, because of intermit-tency issues.10

A related issue concerns the development ofgrids. In the United States, many of the locationsthat have the most potential for renewables are farfrom population centers where electricity

demand is greatest. Vajjhala et al. (2008) discusshow state RPSs would provide significantly differ-ent incentives for improving the transmission net-work than would a federal standard, which wouldallow for greater cost savings but would requiremore of a transmission build-out. State policieswould be less cost-effective, as renewables wouldbe built in many states that have high costs, butwould require less investment in new transmis-sion. (Chapters 7, 9, and 11 further discuss theimportant issues of intermittency and designingthe transmission grid.)

Direction of Future PolicyRecently, carbon policy has become a centraltopic in both the United States and the EU. In theUnited States, while regional policies have beenimplemented in the Northeast, such as theRegional Greenhouse Gas Initiative (RGGI), ordiscussed in the West, such as California AssemblyBill 32 (AB 32), the focus has turned to a nationalcap-and-trade policy. In particular, Congress isconsidering the American Clean Energy andSecurity Act of 2009 (H.R. 2454), also known asthe Waxman–Markey bill (U.S. House Commit-tee 2009). If it were to pass, this bill would be thefirst U.S. national climate policy and would regu-late multiple industries.11 Its goal is to reducegreenhouse gas emissions by 17% of 2005 levelsby 2020 and over 80% of 2005 levels by 2050,mainly through a national cap-and-trade system.

In addition to the tradable permit system, theWaxman–Markey bill contains a renewable elec-tricity standard (RES) that would require that15% of electricity purchased in the United Statescome from renewable resources (including solar,wind, biomass, landfill gas, and geothermal) by2020.12 This is in line with the White Houseagenda of 10% renewables by 2012 and 25% by2025 (Heal 2009). Finally, the United Statesrecently passed a short-term stimulus bill to helpwith economic recovery, the American Recoveryand Reinvestment Act of 2009 (ARRA).

In Europe, both carbon policies and goals ofrenewable shares are expected to continue. TheETS is currently in its second phase, from 2008 to

38 ErinT. Mansur

2012. In June 2009, the new ETS Directive(2009/29/EC) committed the EU to reducinggreenhouse gas emissions by 20% of 1990 levelsby 2020 (see European Parliament and Council2009 for more details). Furthermore, Directive2009/28/EC states that 20% of the EU’s totalenergy consumption must be from renewableresources by 2020. Note that this is a much largercommitment than in the United States, whichseeks to have 20% of electricity production fromrenewables by 2020. In the United States, and inthe rest of the world, about 40% of total energyconsumption is used to generate electricity. So20% of electricity generation in the United Statesis only 8% of total energy consumption.

These policies focus on the near future. Inorder to attain a low-carbon world, however,some believe we will need an even greater invest-ment in renewables.The next section of this chap-ter examines the level of investment in renewablesthat is expected if we have carbon policies eitherlike those described above or that call for evenmore dramatically reduced emissions. The focusremains on the United States and EU, but somemodels examine worldwide levels of investmentin renewables.

Forecasts of Renewables in aLow-Carbon WorldIn examining the policies outlined above, as wellas other possible policies, many authors haveattempted to forecast what level of investment inrenewables would be required in order to attain alow-carbon world. This section outlines the find-ings of some of the major studies in this area,including those by the International EnergyAgency (IEA), European Commission, U.S.Department of Energy (DOE) and EnvironmentalProtection Agency (EPA), FORRES 2020Project, and a few others. For each model, whenpossible, both the expected capacity of and gen-eration from renewables excluding large-scalehydropower, as well as total renewables, arereported. Each model has thousands of assump-tions regarding economic growth, resource avail-

ability, technology, prices, and so forth. A carefulcomparison of the models requires an in-depthexamination of each assumption and themodeling approach taken. However, the intenthere is to give a brief overview of the findings sothe reader can gain an appreciation of the range ofestimates in the literature, rather than to provide ameta-analysis or pick a preferred model. Therange of models discussed presents perspectives ofworldwide, U.S., and EU investment over thenext 20 to 40 years.

A Global Perspective

The IEA (2007) uses its World Energy Model(WEM) to examine long-run forecasts of emis-sions, energy supply and demand, and energysources under various scenarios. In its annualreport, the World Energy Outlook, IEA gives threescenarios: a reference or business-as-usual sce-nario; an alternative policy scenario that wouldinclude reducing greenhouse gas emissions by19% relative to the reference case by 2030; and a 450stabilization scenario. This last scenario is a low-carbon world in which carbon dioxide equivalentconcentrations would be stabilized at levels ran-ging from 445 to 490 parts per million (ppm)(IPCC 2007). This would require emissions ofcarbon dioxide equivalent (CO2e) to peak in 2015and fall by up to 85% of 2000 levels by 2050.13 Forthe year 2030, the reference scenario implies 42gigatons (Gt, or billion metric tons) of CO2 emis-sions, the alternative policy scenario implies 34Gt, and the stabilization scenario implies 23 Gt.

In a low-carbon world, nonhydropowerrenewable resources would double their expectedgeneration in 2030 relative to a business-as-usualscenario (see Table 3.3). This would include add-ing over 1,400TWh per year from wind and solarpower. Even in the reference case, the amount ofwind power is expected to increase fivefold fromcurrent levels.14 Overall, the IEA predicts that40% of electricity could be produced from renew-able resources, about half of which would be fromnonhydropower resources.

The U.S. Department of Energy also pub-lishes a global perspective of renewable invest-ments in its International Energy Outlook. In the

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 39

business-as-usual scenario, the DOE predicts apossible 6,724 TWh of renewables (21% of totalgeneration) by 2030, including 1,951 TWh ofnonhydroelectric generation (6.1% of total gen-eration). Of this, most of the nonhydropower willbe in the OECD countries (1,417 TWh).

A European Perspective

The European Commission’s World Energy Tech-nology Outlook—2050 (WETO) forecasts threecases of world energy out to 2050: a referencecase, carbon constraint case, and hydrogen case(European Commission 2006). The report uses aworldwide simulation model of various energysectors, known as Prospective Outlook on Long-term Energy Systems (POLES), to look at bothEU and global energy issues. The POLES modelaccounts for world population and GDP growth,as well as technological advancement lowering thecosts of energy.15 Table 3.4 reports WETO’s pre-dicted electricity generation by fuel type for eachscenario in 2030 and 2050.

In the reference model, WETO forecasts thattotal renewables will account for 21% of electri-city production by 2030 and 25% by 2050. By2050, solar and wind are expected to producemore than half of the electricity generated fromrenewables, with wind (6,433 TWh per year),particularly offshore, becoming even larger thanhydropower (4,853 TWh per year) worldwide.

This represents a 25-fold increase in wind powerfrom 2008 levels. The increase to 1,493 TWhfrom solar is even larger in magnitude (the UnitedStates and EU produced about 4.5TWh last year).Nuclear power is also predicted to increase rap-idly. The reference case includes modest carbonpolicy akin to what is currently in place in thosecountries with such policy. For this case, theWETO reports substantial investment inrenewables for the world as a whole. Nonetheless,carbon dioxide emissions are projected to be morethan double today’s levels by 2050 in this scenario.

The reference case is also reported by region.In Europe, WETO predicts even greater renew-able penetration: 26% of electricity productioncould be from renewables in both 2030 and 2050.Solar, biomass, and wind power will account forapproximately two-thirds of that generation.North America is similar, with an expected 20%of all electricity coming from solar, wind, orbiomass.

The carbon constraint case seeks to stabilizecarbon dioxide concentrations, though at levelsgreater than in the IEA study: the WETO caseaims to stabilize CO2e at 650 ppm. In 2030,WETO predicts 8,823 TWh of renewable pro-duction, about half from hydropower and a quar-ter each from wind and biomass. In 2050, wind ispredicted to be dominant, accounting for 7,336TWh of the 17,439TWh that are renewable.This

Table 3.3. IEA world energy outlook predictions, showing generation by energy source (in TWh)

Source 2005

2030referencecase

2030stabilizationcase Change Growth rate

Hydro 2,922 4,842 6,608 1,766 36%

Biomass 231 840 2,056 1,216 145%

Wind 111 1,287 2,464 1,177 91%

Geothermal 52 173 219 46 27%

Solar 3 161 406 245 152%

Tidal (wave) 1 12 28 16 133%

Nonhydro renewables 398 2,473 5,173 2,700 109%

Total renewables 3,321 7,315 11,781 4,466 61%

Total generation 18,197 35,384 29,300 –6,084 –17%

Percent renewable 18% 21% 40%

Source: IEA 2007, Table 5.6

40 ErinT. Mansur

implies that renewables will account for 30% oftotal electricity production, with wind (12.7%),solar (4.0%), and biomass (4.6%) all playingimportant roles. North America and Europe areexpected to have similar levels of nonhydrorenewable penetration: 14.2 and 9.8% wind, 6.3and 6.7% solar, and 7.2 and 4.1% biomass forNorth America and Europe, respectively.

The hydrogen case continues with carbonpolicies to stabilize greenhouse gas concentrationsbut also assumes substantial breakthroughs in

hydrogen technology. In particular, breakthroughsin the performance and cost of the distributionand consumption sectors of the hydrogeneconomy are identified as being of greatest impor-tance.This scenario changes the mix of fossil fuelsto nuclear power and changes transportation fuels:hydrogen provides 13% of final energy consump-tion (primarily for transport) in this case, com-pared with the 2% share in the reference case.

Table 3.4. Electricity generation by fuel type for the three WETO scenarios (in TWh)

Panel A: year 2030 World Europe North America

Energy sourceRefer-ence Carbon

Hydro-gen

Refer-ence Carbon

Hydro-gen

Refer-ence Carbon

Hydro-gen

Coal 12,689 9,114 8,205 1,551 969 794 2,944 1,208 1,491

Gas 8,760 9,438 8,851 1,319 1,545 1,540 1,836 2,281 2,186

Biomass 1,372 1,684 1,644 258 315 311 417 469 474

Nuclear 6,328 6,449 8,834 1,447 1,432 1,597 1,088 967 981

Hydro andgeothermal 4,148 4,284 4,226 697 706 702 711 732 730

Solar 91 213 183 17 29 28 13 68 55

Wind 1,880 2,642 2,417 545 608 604 428 728 727

Hydrogen 39 56 44 9 15 12 4 9 7

Total electricity 36,295 34,587 35,039 5,932 5,673 5,642 7,560 6,548 6,714

Percentrenewable 21% 26% 24% 26% 29% 29% 21% 30% 30%

Panel B: year 2050 World Europe North America

Energy sourceRefer-ence Carbon

Hydro-gen

Refer-ence Carbon

Hydro-gen

Refer-ence Carbon

Hydro-gen

Coal 19,066 9,016 9,371 1,860 781 633 3,976 1,415 1,525

Gas 9,072 9,640 8,959 1,337 1,492 1,465 1,408 1,841 1,738

Biomass 2,246 2,649 2,584 328 361 360 598 680 685

Nuclear 14,866 19,862 21,426 2,931 3,612 3,942 2,014 2,509 2,474

Hydro andgeothermal 4,853 5,128 4,998 738 746 743 764 784 782

Solar 1,493 2,326 2,058 344 593 591 308 590 523

Wind 6,433 7,336 6,799 817 859 838 1,115 1,337 1,352

Hydrogen 811 1,477 898 190 321 243 73 196 120

Total electricity 60,040 57,812 57,377 8,608 8,803 8,845 10,337 9,407 9,233

Percentrenewable 25% 30% 29% 26% 29% 29% 27% 36% 36%

Source: European Commission 2006

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 41

With respect to renewables, however, the hydro-gen scenario is qualitatively similar to the carbonconstraint case.

A U.S. Perspective

The DOE’s Energy Information Administration(EIA) publishes the Annual Energy Outlook, whichreports long-run projections of U.S. energy sup-ply and demand. The EIA (2009a) uses theNational Energy Modeling System (NEMS) toforecast scenarios relevant to investments inrenewables. The reference case for 2030 predictsan increase in renewable capacity in the UnitedStates of 57 GW (+/– 10 GW, depending on sce-nario’s assumptions regarding costs).This accountsfor about 22% of all new capacity constructedover the next 22 years. Regardless of the cost sce-nario, the EIA expects 730 TWh of renewablegeneration in 2030. Nonhydropower resourcesare predicted to account for about 14% of U.S.total electricity generation (or 430 TWh). Costsdo not change predictions much because the staterenewable portfolio standards are assumed to bind.The mix of renewables does change, however: theEIA forecasts greater use of wind power relative tobiomass if capital costs fall.

The EIA compares its reference case withsome other studies. Namely, it notes that IHSGlobal Insight (IHSGI 2008) predicts similar lev-els of growth in renewables, with slightly moreconservative estimates from Energy VenturesAnalysis (EVA 2008). In the reference case, by2015, the EIA predicts generation from hydro-electric and other renewables, including netimports, to be 555 TWh, while IHSGI predicts537 TWh and EVA predicts 420 TWh, justslightly more than the 374 TWh generated in2007. By 2030, the predictions increase to 758TWh (EIA), 864 TWh (IHSGI), and 535 TWh(EVA). These correspond to 15%, 17%, and 11%of total generation for each model, respectively.EIA’s projections of capacity are also in line withalternative models. In addition to IHSGI andEVA, the Institute of Energy Economics and theRational Use of Energy (IER 2008) also forecastsrenewable capacity for 2015 and 2030. The totalcapacity for hydroelectric and other renewables

was 131 GW in 2007. In 2015, it is expected toincrease to 157 GW (EIA), 160 GW (IHSGI),115 GW (EVA), and 208 GW (IER). By 2030,the capacity is predicted to be 191, 232, 128, and312 GW, respectively. In summary, the EIA refer-ence case seems consistent with several other pre-dictions by industry and academia.

The EIA predicts several counterfactuals rel-evant to a low-carbon future relative to a refer-ence case that predated the recent stimulus bill,ARRA. The first counterfactual is to look at theeffect of the bill, which extends two importantsubsidies: the PTC and the ITC. Figure 3.3 com-pares annual investment in nonhydropowerrenewables from 2006 to 2030 under the baselinereference case and under the ARRA. In 2012, thepolicy is expected to increase nonhydropowerrenewables from 195 to 310TWh, a 59% increase.By 2030, total renewables are expected to accountfor 15% of electricity generation in the UnitedStates (two-thirds of which are fromnonhydroelectric resources).

The increase in renewable capacity from theARRA is even more substantial. By 2015, anadditional 40 GW of renewable capacity isexpected, mostly from wind (35 GW). The EIAanticipates that by 2030, wind power capacity (68GW) will nearly rival that of conventionalhydropower (78 GW) in the United States. Onlysome of the increase in renewables that is attrib-uted to the ARRA is the result of simply extend-ing the PTC.

In a simulation that extends the PTC through2019, the EIA predicts that wind power wouldincrease by 19 GW (while biomass, municipalsolid waste, and geothermal power would beunchanged) by 2020 relative to the reference sce-nario. Although an increase of 19 GW may seemlarge, especially given that the United States hadonly 25 GW wind power capacity in 2008(WWEA 2009), the EIA expects that the ARRAwill increase renewable power by more than 33GW in 2020, with wind accounting for all of thegain.The ARRA allows developers to convert thePTC into federal grants, thereby avoiding a majorhurdle: namely, that only those with significanttax liability can benefit from the tax credit.

42 ErinT. Mansur

In a third scenario, the EIA examines a low-carbon case: the Lieberman–Warner bill, § 2191(LW110). The EIA expected that the bill wouldresult in the electric power sector’s emissions fall-ing by more than 50% relative to 2007 levels. Fur-thermore, the bill would have increased renewablegeneration in 2030 from 730 to 1,063 TWh (or22% of total electric generation). Renewablecapacity would nearly double from the referencecase (57 to 103 GW). Nuclear power andadvanced coal with carbon capture and storage(CCS) would also increase substantially: from13/1 GW of nuclear/CCS capacity in the refer-ence case to 47/99 GW in the LW110 case.

A Forecast of Proposed U.S. Carbon Policy

In a recent report, EPA (2009a) analyzed theWaxman–Markey bill. Although a federal carbonpolicy will almost surely differ from this exact bill,it is useful to examine nonetheless, as it is the mostrecent proposal as of this writing. This bill’s cap-and-trade policy would result in greater costs forfossil-fuel electricity generation, makingrenewables more competitive. EPA estimated theinitial cost of polluting to be around $13 per met-ric ton of CO2e. From the cap-and-trade policyalone, EPA finds that investment in primaryenergy that has either no or low carbon

emissions—namely renewables, nuclear power,and advanced coal with CCS—would rise for thebusiness-as-usual scenario of 14% of primaryenergy to 18% by 2020, 26% by 2030, and 38% by2050.

Using the detailed Integrated PlanningModel, EPA accounts for both the cap-and-tradeprogram and the RES. Table 3.5 reports theexpected investment in various sources of electricpower under the reference case as well as underthe Waxman–Markey bill. The reference case isbased on the EIA’s Annual Energy Outlook 2009(EIA 2009a). Relative to the reference case, theclimate bill would increase electricity generatedby renewables (excluding hydropower) by only3% to 5%. Including hydropower, the EPA modelpredicts approximately zero change in generationfrom renewables in 2025.This status quo is due toa reduction in total electricity consumption, inpart because of higher electricity prices but alsobecause of economic incentives to conserve. Oneway utilities can comply with the RES is throughenergy efficiency programs.

Even though EPA does not predict that thebill will result in substantial incremental investmentin renewables, the reference case includes 364TWh produced by renewables in 2025. In con-trast, current production from renewables is 124TWh (EIA 2009b). In other words, based on cur-

600

500

400

300

Gen

erat

ion

(TW

h pe

r ye

ar)

Cap

acity

(G

W)

200

100

–2005 2010 2015 2020 2025 2030

20

–

40

60

80

100

120

Baseline TWhARRA TWhBaseline GWARRA GW

Source: EIA 2009a

Figure 3.3. Expected electricity generation and capacity from nonhydropower renewable sources withand without ARRA, 2006–2030

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 43

rent policies, EPA expects renewables to tripleover the next 17 years. In the reference case,nonhydroelectric renewables would account for8% of total electricity generation (or 9% under theWaxman–Markey bill). Including hydropower,this increases to 14% (16% with the bill). A finalnote regarding the Waxman–Markey bill: accord-ing to EPA’s model, the bill is expected to causesubstantial growth in “clean” coal (advanced coalwith CCS), on the order of 400% to 1,200%,relative to the reference case.

In addition to the EPA study, the MIT JointProgram on the Science and Policy of GlobalChange has analyzed the bill (Paltsev et al. 2009).This analysis finds that annualized costs of the billwould be on the order of $400 per householdwith a permit price starting at just over $20 perton of CO2 in 2015 and reaching $38 by 2030.

The bill is expected to increase renewables onlyslightly: the share of power from nonhydroelectricrenewables would increase from 10% to 11%.

A Forecast of EU Renewables Policy

In a report financed by the European Commis-sion’s Directorate-General of Energy and Trans-port, Ragwitz et al. (2005) examined how invest-ment in renewables might proceed in the Euro-pean Union through 2020. The authors use theGreen-X model from the FORRES 2020 project,which assumes learning by doing and scale econo-mies.

For the year 2020, their model suggests thatthe reference case could have 385 TWh fromwind, 8.8 TWh from PV solar, and 607 TWh oftotal renewables (excluding large-scale hydro-

Table 3.5 EPA analysis of the Waxman–Markey Bill, showing generation by energy source, 2020 and 2025(in TWh)

Panel A: year 2020

Energy source Reference case H.R. 2454 Change Growth rate

Coal 2,222 1,940 –282 –13%

Advanced coal (w/ CCS) 14 71 57 407%

Natural gas/oil 703 486 –217 –31%

Nuclear 816 816 0 0%

Hydropower 290 285 –5 –2%

Other renewables 334 351 17 5%

Total renewables 624 636 12 2%

Total electricity 4,379 3,949 –430 –10%

Percent renewable 14% 16%

Panel B: year 2025

Energy source Reference case H.R. 2454 Change Growth rate

Coal 2,312 1,851 –461 –20%

Advanced coal (w/ CCS) 14 184 170 1,214%

Natural gas/oil 788 544 –244 –31%

Nuclear 837 820 –17 –2%

Hydropower 292 282 –10 –3%

Other renewables 364 375 11 3%

Total renewables 656 657 1 0%

Total electricity 4,607 4,056 –551 –12%

Percent renewable 14% 16%

Source: EPA 2009a

44 ErinT. Mansur

power). For comparison, the EU produced only104TWh from wind, 3.8TWh from solar PV, and212 TWh from nonhydropower renewables in2007 (Eurostat 2009). As a share of overall grosselectricity consumption in the EU countries, thisis an increase from 6.5% in 2007 (16% ifhydropower is included) to 15% (22%) in 2020.

The authors also examine a policy scenariowhereby “best practices” of currently availablestrategies are applied across all 27 EU countries.16

In this policy case, wind power is expected toproduce 461 TWh, another 17.9 TWh from solarPV, and 928 TWh from total renewables (exclud-ing large-scale hydropower), or 26% of expectedtotal electricity generation. With hydropower, thisincreases to 1,234 TWh (34%). The authors con-clude that following currently available best prac-tices, renewables could account for a third of allelectricity generation in the EU.

Other Models

The research of some academics, trade groups,and nonprofit organizations also has examined thequestion of what could be the long-run level ofinvestment in renewables. Palmer and Burtraw(2005) use the RFF Haiku model of the U.S.electricity industry to analyze the effects ofrenewable portfolio standards (RPSs) and renew-able energy production tax credits. In the refer-ence case, the authors predict that in 2020, 151TWh of nonhydropower could be generated, oronly 3.1% of total generation (9.5% if hydropoweris included). Under a general renewable produc-tion tax credit, which is set at current levels butapplies to a much broader set of renewables,Palmer and Burtraw find that 729 TWh would beproduced. This would result in 15% of electricitycoming from nonhydropower renewables (with43% of it from wind power). Total renewables,including hydropower, are predicted to accountfor 21% of electricity supply. Palmer and Burtrawlook at several possible RPSs for 2020: 5%, 10%,15%, and 20%. In each case, the standard is bind-ing for nonhydropower renewables, andhydropower accounts for an additional 6.5% oftotal electricity generation. For the low-levelRPS, nonhydropower renewables are dominated

by geothermal. However, a 20% RPS wouldrequire substantial investment in wind power,accounting for about half of all nonhydropowerrenewable generation (468 of 948 TWh).

Many other studies have examined this ques-tion. For example, the Intergovernmental Panelon Climate Change (IPCC 2007) notes thatrenewable electricity, including hydropower, was18% of world electricity supply in 2005, and thatwith a carbon price of $50 per ton of CO2e, itcould reach a 30% to 35% market share by 2030.The Union of Concerned Scientists (Cleetus et al2009) suggests a blueprint to reduce U.S. carbondioxide emissions by 84% below 2005 levels by2030. Through energy efficiency policies, thestudy suggests that electricity consumption couldbe reduced by 35% relative to the reference case.An RPS would result in wind, solar, geothermal,and biomass providing 40% of the total electricity,or about 21% of overall energy supply. TheWWEA (2009) expects 1,500 GW of installedwind capacity producing 12% of electricity gen-eration by 2020. The Energy Watch Group(Rechsteiner 2008) predicts that by 2025, theinvestment could be even greater, with 7,500 GWinstalled wind capacity producing 16,400 TWh ayear. Furthermore, the study anticipates that totalrenewables could account for more than 50% ofnew power installations worldwide by 2019 andall of it by 2022. By 2037, it suggests thatnonrenewable generation could be completelyphased out.

ConclusionsWithin 20 years, some authors predict that 20% to40% of electricity could be produced from renew-able resources. Some predict even higher levels.Table 3.6 draws together the medley of forecastsoutlined above to show the range of model pre-dictions as well as some important regional andpolicy differences.

Returning to an economist’s perspective, onewould need to know the relative net marginalsocial benefits for each type of renewableresource, at each location and point in time, inorder to determine whether a 20%, 30%, 40%, or

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 45

even 50% market share for renewable electricity isoptimal. This chapter focused, however, not onwhat should be the level of investment inrenewables, but rather on what modelers predictcould be the level of investment, based on eitherpolicies that are currently being discussed—suchas the third phase of the EU Emission TradingSystem, EU Directive 2009/28/EC on renew-ables, or the U.S. Waxman–Markey bill—orfuture policy objectives, such as stabilizing carbondioxide equivalent concentrations at 450 or 650ppm.

By reviewing major studies on this topic, thischapter has provided a range of estimates of levelsof investment in wind power, solar PV,nonhydropower renewables, and total renewablesfor the world overall, the EU, and the UnitedStates. The range of levels of investment reflectsmany differences regarding the assumptions in themodels, the policies being analyzed, and themethodologies used by the researchers.This chap-ter has shown that although opinions are diverseon what might happen over the next 20 to 40years, studies employing a variety of models and arange of assumptions regarding carbon policyforecast the likelihood of a significant increase inthe role of renewables in meeting electricitydemand.

Notes1. Geoengineering includes methods such as

stratospheric sulfur aerosols, which manage solarradiation. These methods do not reduce green-house gases, but rather limit their effects on globalwarming. Some also use this term to include car-bon sequestration.

2. Not all greenhouse gases are as long-lived. Meth-ane, for example, has a chemical lifetime in theatmosphere of approximately 12 years.Technically,a stock pollutant never expires; for example, lead isa true stock pollutant. However, the economicinsights of stock pollutants help in thinking aboutregulating greenhouse gases.

3. More precisely, the optimal level of pollution at apoint in time will be where present value marginalbenefits of pollution (the avoided compliancecosts) are equal to the present value marginal costsof pollution (the marginal damages). For furtherdiscussion, see Tietenberg (2006).

4. Fischer and Newell (2008) show that a policy pro-moting renewables may be more efficient than acarbon tax if these positive technology spilloversare large relative to the climate externalities. How-ever, they find for a range of parameter values, aprice on carbon is preferred even when there areknowledge spillovers.

5. One terawatt-hour is 1 million megawatt-hours(MWh) or 1 billion kilowatt-hours (kWh). One

Table 3.6. A summary of model predictions for 2030

Generation percentage

Geographic scope Modelingorganization

Carbon policy Nonhydrorenewables

Allrenewables

Worldwide IEA Business as usual 7% 21%

450 stabilization 18% 40%

Worldwide EC Business as usual 11% 21%

Carbon constraint 14% 26%

Hydrogen 12% 24%

Worldwide IPCC $50/ton price — 30%–35%

Europe EC Business as usual 15% 26%

Carbon constraint 17% 29%

Hydrogen 17% 29%

U.S. EIA ARRA 9% 15%

Lieberman–Warner — 22%

EPA Waxman–Markey 16% 26%

Sources: IEA 2009; European Commission 2006; EPA 2009b, 27

46 ErinT. Mansur

TWh approximately equals the annual consump-tion of about 89,000 U.S. residential customers.

6. Electricity generation capacity is measured inwatts (W), kilowatts (kW, 1,000 W), megawatts(MW, 1 million W), or gigawatts (GW, 1 billionW).

7. Some renewables, including solar, wind andgeothermal energy, first became eligible for taxcredits with the passage of the 1978 Energy TaxAct.

8. The EEG was further amended: as of January2009, PV solar subsidies have been reduced to€330–€430 ($449–$586) per MWh.

9. The price on August 24, 2009 was €15.34($20.89) (Point Carbon n.d.).

10. Some express concern that predictions of substan-tial levels of investment in the future are not feasi-ble, given the issue of intermittency. In a recentpaper, Heal (2009) reviews the literature on theeconomics of renewables. He shares these con-cerns and concludes that nuclear and CCS arelikely to be important sources of electricity in alow-carbon future.

11. It is unlikely that the Waxman–Markey bill willpass into law as it was written when the Housepassed it in the summer of 2009. As of this writing,Congress has not passed any climate bill.

12. The requirement is that 20% be from renewablesor energy efficiency, but at least 75% of thatamount must be through renewables. This con-straint is expected not to be binding.

13. CO2e is a measure for greenhouse gases wherebyall gases are normalized by their global warmingpotential.

14. In 2008, wind power produced 260 TWh(WWEA 2009); in the reference case, the IEA(2007) predicts a possible 1,287 TWh by 2030.

15. The POLES model was developed by the CentreNational de la Recherche Scientifique (CNRS)and is maintained by CNRS, UPMF University,Enerdata, and the Institute for Prospective Tech-nological Studies. See Enerdata (2006).

16. Best practices are defined as those “strategies thathave proven to be most effective in the past inimplementing a maximum share of RES”(Ragwitz et al. 2005).

ReferencesBorenstein, Severin. 2008. The Market Value and Cost

of Solar Photovoltaic Electricity Production. Work-

ing paper 176. Berkeley, CA: University of Califor-nia Energy Institute, Center for the Study of EnergyMarkets.

Campbell, Arthur. 2009. Government Support forIntermittent Renewable Generation Technologies.Working paper. Cambridge, MA: MassachusettsInstitute of Technology. econ-www.mit.edu/files/3563 (accessed August 17, 2009).

DSIRE (Database of State Incentives for Renewablesand Efficiency). 2009. Project of the North CarolinaSolar Center and Interstate Renewable EnergyCouncil funded by the U.S. Department of Energy’sOffice of Energy Efficiency and Renewable. Energy.www.dsireusa.org/ (accessed February 20, 2010).

EIA (U.S. Energy Information Administration). 2009a.Annual Energy Outlook 2009:With Projections to 2030.DOE/EIA-0383(2009). www.eia.doe.gov/oiaf/aeo/pdf/0383(2009).pdf (accessed August 19, 2009).

———. 2009b. Short-Term Energy Outlook—August2009. www.eia.doe.gov/pub/forecasting/steooldsteos/aug09.pdf (accessed August 29, 2009).

Enerdata. 2006. POLES Model: Prospective Outlook onLong-term Energy Systems: A World Energy Model.www.eie.gov.tr/turkce/en_tasarrufu/uetm/twinning/sunular/hafta_02/5_POLES_description.pdf (accessed August 29, 2009).

EPA (U.S. Environmental Protection Agency). 2009a.EPA Analysis of the American Clean Energy and SecurityAct of 2009 HR 2454 in the 111th Congress.www.epa.gov/climatechange/economics/economicanalyses.html (accessed August 20, 2009).

———. 2009b. EPA Preliminary Analysis of theWaxman-Markey Discussion Draft: The American Clean Energyand Security Act of 2009 in the 111th Congress.www.epa.gov/climatechange/economics/pdfs/WM-Analysis.pdf (accessed February 26, 2010).

European Commission. 2006. World Energy TechnologyOutlook—2050. Directorate-General for ResearchEUR 22038. ec.europa.eu/research/energy/pdf/weto-h2_en.pdf (accessed February 26, 2010).

———. 2009. The Renewable Energy Progress Report.Commission Staff Working Document SEC(2009)503 final. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2009:0192:FIN:EN:PDF(accessed August 19, 2009).

European Parliament and Council. 2009. Directive2009/29/EC of the European Parliament and of theCouncil of 23 April 2009. eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=OJ:L:2009:140:0063:0087: EN:PDF (accessed Feb-ruary 26, 2010).

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 47

Eurostat (European Commission Eurostat). 2009.Energy: Yearly Statistics 2007.

EVA (Energy Ventures Analysis). 2008. FUELCAST:Long-Term Outlook.

Fischer, Carolyn, and Richard G. Newell. 2008. Envi-ronmental and Technology Policies for ClimateMitigation. Journal of Environmental Economics andManagement 55 (2): 142–162.

Galbraith, Kate. 2009. More Sun for Less: Solar PanelsDrop in Price. NewYorkTimes, Aug. 26, B1.

Heal, Geoffrey. 2009. The Economics of RenewableEnergy. Working paper 15081. Cambridge, MA:National Bureau of Economic Research.

IEA (International Energy Agency). 2004. RenewableEnergy: Market and Policy Trends in IEA Countries.Paris: OECD Publishing.

———. 2007. World Energy Outlook 2007. Paris:OECD Publishing.

———. 2009. World Energy Outlook 2009 Fact Sheet.www.worldenergyoutlook.org/docs/weo2009/fact_sheets_WEO_2009.pdf (accessed February 26,2010).

IER (Institute of Energy Economics and the RationalUse of Energy). 2008. TIAM Global Energy SystemModel. Stuttgart: University of Stuttgart.

IHSGI (IHS Global Insight). 2008. Global PetroleumOutlook. Lexington, MA: IHSGI.

IPCC (Intergovernmental Panel on Climate Change).2007. Fourth Assessment Report. www.ipcc.ch/publications_and_data/publications_and_data_reports.htm#1 (accessedAugust 10, 2009).

Martinot, Eric, Ryan Wiser, and Jan Hamrin. 2005.Renewable Energy Policies and Markets in theUnited States. Working paper. San Francisco: Centerfor Resource Solutions.

Metcalf, Gilbert. 2009a. Investment in Energy Infra-structure and the Tax Code. Working paper.Medford, MA: Tufts University.

———. 2009b. Tax Policies for Low-Carbon Tech-nologies. Working paper w15054. Cambridge, MA:National Bureau of Economic Research.

Nielsen, Lene, andTim Jeppesen. 2003.Tradable GreenCertificates in Selected European Countries: Over-view and Assessment. Energy Policy 31 (1): 3–14.

NREL (National Renewable Energy Laboratory).2008. 20% Wind Energy by 2030. Increasing WindEnergy’s Contribution to U.S. Electricity Supply.Technical paper. Golden, CO: NREL.

Palmer, Karen, and Dallas Burtraw. 2005. Cost-Effectiveness of Renewable Electricity Policies.Energy Economics 27: 873–894.

Paltsev, S., J.M. Reilly, H.D. Jacoby, and J.F. Morris.2009. Appendix C: Cost of Climate Policy and theWaxman-Markey American Clean Energy andSecurity Act of 2009 (HR 2454). In The Cost ofClimate Policy in the United States, edited by S. Paltsev,J.M. Reilly, H.D. Jacoby and J.F. Morris. MIT JointProgram Report Series No. 173, 1–21.

Point Carbon. No date. Point Carbon home page.www.pointcarbon.com (accessed February 26,2010).

Ragwitz, Mario, Joachim Schleich, Claus Huber,Gustav Resch, Thomas Faber, Monique Voogt,Rogier Coenraads, Hans Cleijne, and Peter Bodo.2005. Analysis of the EU Renewable EnergySources’ Evolution up to 2020 (FORRES 2020).www.emu.ee/orb.aw/class=file/action=preview/id=254843/FORRES_FINAL_REPORT.pdf 3563(accessed August 25, 2009).

Rechsteiner, Rudolf. 2008. Wind Power in Context: AClean Revolution in the Energy Sector. Basel, Switzer-land: Energy Watch Group and Ludwig-Boelkow-Foundation.

REN21 (Renewable Energy Policy Network for the21st Century). 2009. Renewables Global Status Report:2009 Update. Paris: REN21 Secretariat.

Stern, Nicholas. 2006. The Economics of Climate Change:The Stern Review. Cambridge, UK: Cambridge Uni-versity Press.

Tietenberg, Thomas. 2006. Environmental and NaturalResource Economics. 7th ed. Reading, MA: Addison-Wesley Longman.

U.S. Congress. 2009. American Recovery and Reinvest-ment Act of 2009. 111th Cong., 1st sess.frwebgate.access.gpo.gov/cgi-bin/getdoc.cgi?dbname=111_cong_bills&docid=f:h1enr.pdf (accessed August 25, 2009).

U.S. Congress House Committee on Energy and Com-merce. 2009. American Clean Energy and Security Actof 2009. 111th Cong., 1st sess., HR 2454.energycommerce.house.gov/Press_111/20090515/hr2454.pdf (accessed August 20, 2009).

Vajjhala, Shalini, Anthony Paul, Richard Sweeney, andKaren Palmer. 2008. Green Corridors: LinkingInterregional Transmission Expansion and Renew-able Energy Policies. Discussion paper 08–06. Wash-ington, DC: Resources for the Future.

Wiser, Ryan, and Mark Bolinger. 2008. Annual Reporton U.S.Wind Power Installation, Cost, and PerformanceTrends: 2007. LBNL-275E. Berkeley, CA: LawrenceBerkeley National Laboratory.

Wiser, Ryan, Galen Barbose, and Carla Peterman.2009. Tracking the Sun: The Installed Cost of

48 ErinT. Mansur

Photovoltaics in the U.S. from 1998–2007. LBNL-1516E. Berkeley, CA: Lawrence Berkeley NationalLaboratory.

WWEA (World Wind Energy Association). 2009.WorldWind Energy Report, 2008.

www.wwindea.org/home/images/stories/worldwindenergyreport2008_s.pdf (accessed August24, 2009).

Renewables Forecasts in a Low-CarbonWorld: A Brief Overview 49

4

Renewable Generation and Securityof SupplyBoaz Moselle

A key question this book seeks to address iswhat justification exists for the promotion of

renewable generation over other forms of low-carbon generation—in other words, for policiesthat specifically promote renewable generationrather than a technology-neutral approach such asa carbon tax or cap-and-trade mechanism. Simpleeconomics suggests that the latter approach wouldbe more effective in achieving carbon reductionsat lowest cost, through competition between dif-ferent carbon abatement mechanisms and tech-nologies (e.g., renewable energy, nuclear, carboncapture and storage, reductions in non-generationsectors, energy efficiency).

In the European Union (EU), one of the mostcommon responses is that renewable generationmerits specific support because it enhances secu-rity of supply by reducing dependence onimported fuels. Concerns about import depend-ence refer particularly (though not exclusively) todependence on natural gas imports from Russiaand Algeria, which many observers view as poten-tially unreliable because of political instability and,in the case of Russia, a willingness to use energysupplies as a tool of geopolitics.1 This concern hasbeen greatly enhanced by interruptions in recentwinters to the flow of gas from Russia into theEU via Ukraine, as a result of disputes betweenRussia and Ukraine.

At the same time, a commonly voiced con-cern with renewable generation is that it willendanger security of supply by leading to exces-sive dependence on intermittent sources such aswind and solar power. To some commentators,this argues against the promotion of renewablegeneration. To others, it implies the need for sig-nificant changes in power market design to ensurethat sufficient backup capacity is available overvarious time frames.

This chapter therefore focuses on these twoquestions, examining to what extent security ofsupply concerns related to import dependencewarrant additional support for renewable genera-tion relative to other forms of low-carbon tech-nology, and to what extent security of supplyconcerns related to intermittency undermine thecase for supporting renewables at all or necessitatemajor changes in market design.

The focus is on the EU, where renewablesdeployment is most prominent on the policyagenda and is explicitly linked to security of sup-ply by policymakers. However, many of theconclusions—in particular, those relating tointermittency—can be applied to other jurisdic-tions as well.

The chapter begins by examining the issue ofimport dependence. It assesses the extent of theproblem and analyzes whether there are market or

other failures that warrant intervention, and if so,whether the promotion of renewable generationis the most efficient form of intervention toaddress the problem. It then focuses on the prob-lems posed by intermittency, again assessing theproblem and analyzing the case for policy inter-vention and the most appropriate form that inter-vention might take.

EU Dependence onImported FuelsThe need to reduce dependence on importedfuels is used to justify a range of EU policies,including not only the promotion of renewables,but also the promotion of energy efficiency andthe provision by some national governments ofsubsidies to domestic coal production. In the pastdecade, these themes have been developed innumerous policy documents and pieces of legisla-tion, including the European Commission’s 2000Green Paper on security of supply, the 2002Regulation on State Aid in the coal sector, the2008 Energy Security and Solidarity Plan, and the2009 Renewables Directive (European Commis-sion 2000; Regulation 1407/2002; EuropeanCommission 2008a; Directive 2009/28/EC).

This section therefore presents evidence onthe extent of EU import dependence and the fac-tors that have most given rise to concern withrespect to power generation: the large and grow-ing dependence on Russian gas imports and theeffect of supply interruptions in recent winters. Italso assesses the extent to which import depend-

ence is a problem for the main fuels used forpower generation and whether the promotion ofrenewable generation is the most appropriatepolicy response to any such problem.

Current and Projected Levels of EUImport Dependence

As Table 4.1 illustrates, the EU imports a largeproportion of its primary energy sources, includ-ing the main fuels used for power generation. In2006, around 80% of electricity was generatedfrom coal (29%), gas (21%), and nuclear sources(30%) (European Commission 2008b).2

Imports are very significant for natural gas,which, as explained below, is the main source ofconcern among policy makers. The EU holds just1.6% of the world’s gas reserves and currentlyimports 58% of its natural gas demand, mainlyfrom four countries: Russia, Norway, Algeria, andNigeria.3 Gas supplies 24% of total energydemand and 21% of electricity generation (Euro-pean Commission 2008b). Gas import depend-ence is set to increase, as EU indigenous produc-tion is forecast to decline rapidly in the comingdecade, from 176 million tons of oil equivalent(Mtoe) in 2010 to 131 Mtoe in 2019 (IEA 2009).4

European Commission analysis forecasts netimports of natural gas increasing from 257 Mtoein 2005 (58% of total consumption) to 390 Mtoein 2020 (77% of total consumption) under abusiness-as-usual scenario, without taking intoaccount the impact of the new energy policyadopted in 2009 (see European Commission2008b, Annex 2).5

Table 4.1. EU import dependence, 2005

EU primary energydemand (Mtoe)

EU primaryproduction (Mtoe)

Net imports(Mtoe)

Import dependence(percentage)

Oil 666 133 533 80.0%

Natural gas 445 188 257 57.8%

Solid fuel 320 196 127 39.7%

Renewables 123 122 1 0.8%

Nuclear/uranium

257 8 249 97.0%

Sources: European Commission 2008b, 65; Euratom 2008Note: Mtoe = million tons of oil equivalent

52 Boaz Moselle

Winter Supply Interruptions

The heavy dependence of the EU on Russian gashas been brought home to the public andpolicymakers alike in recent years by interruptionsto the supply of Russian gas at the start of thecalendar year, arising from disputes between Rus-sia and Ukraine. A number of such disputes haveoccurred since the breakup of the Soviet Union asa result of continuing difficulties in agreeing onthe details of a new gas transit and supply regime,as well as deeper underlying differences.The mostserious of these interruptions occurred at thebeginning of 2006 and 2009. In January 2006, gassupplies to the EU were interrupted for one day;in January 2009, the interruption lasted 16 days(European Commission 2009a).

Ukraine’s Role as a Gas Consumerand Transit Country

Ukraine is both a significant consumer of gas anda key transit country. Its daily consumption inwinter is about 300 million cubic meters per day(mcm/day), and another 300 to 350 mcm/day ofgas passes through Ukraine to the EU (EuropeanCommission 2009a). Imports from Russia viaUkraine constitute around 80% of EU imports ofgas from Russia and about 20% of total gasdemand in the EU (European Commission2009a). The Ukrainian gas sector features below-cost pricing for domestic and government cus-tomers, and chronic underinvestment in its oiland gas sector, including the gas pipeline infra-structure (Chow and Elkind 2009).

Disputes between Ukraine and Russia overgas supplies, transit, and payment for gas havebeen a feature of this market since the early 1990s.Ukrainian inability to pay for the huge volumes ofgas contracted (despite the very low prices Russiagave Ukraine) led to high levels of debt andunpaid bills on a continuous basis for many years(Stern 2005). The disputes remained unresolveddespite a series of agreements covering the gasvolumes and prices, the price of gas transit acrossUkraine, and the level of debt owed to Gazpromby the Ukrainian gas company Naftokhaz, which

were characterized by low gas prices for Ukraineand low transit charges for delivery of Russian gasto Europe.6

In March 2005, Russia claimed that Ukrainewas not paying for gas and was diverting gasintended for transit to the EU (BBC 2006). OnJanuary 1, 2006, Russia retaliated by cutting offgas supplies passing through Ukrainian territory.7

Russia and Ukraine reached a preliminary agree-ment on January 4, and the supply was restored.The agreement provided for an increase in thenominal price of gas but did not provide anagreed pricing formula for future years or a tran-sition period to higher prices.The new agreementwas set to expire on December 31, 2008.

The 2009 crisis began on January 1, whenGazprom cut off suppliers (again, it stopped sup-plying gas for Ukrainian consumption while thesupply of gas that was theoretically to be transitedthrough for European consumption continued).Initially disruption of supply to the EU was onlyminor, but by January 7, all supplies from Russiato the EU were cut, and supplies were notresumed until January 20. This was the most seri-ous gas supply crisis ever to hit the EU, deprivingit of 20% of its total gas supply (European Com-mission 2009a). Within days of the supply disrup-tion, 12 countries were affected. They respondedby drawing on storage, importing additional LNGsupplies, and fuel-switching by the use of fuel oiland coal. Increased supplies were sourced fromRussia via Belarus and Turkey, as well as fromNorway and Libya. Gazprom is estimated to havelost sales of $2 billion (European Commission2009a).

Is Import Dependence Really a Problem?

Reliance on imported fuels is not, per se, a causefor concern. For policy intervention to be justi-fied on security-of-supply grounds, a number ofconditions must be satisfied, including that:

• The reliance on imports creates a genuinesecurity-of-supply risk. This is unlikely to bethe case for a fuel that can be imported easilyfrom a number of different countries that arepolitically stable, friendly, and geographicallydiverse.

Renewable Generation and Security of Supply 53

• There is good reason to think that the normalmarket response will not efficiently addressany security-of-supply risks and that policyintervention can be expected to do better.

The first of these conditions is assessed below foreach of the main fuels used for generation: coal,uranium, and natural gas. This is followed by adiscussion of the potential for market or otherfailures that might justify intervention.

Coal

Globally, coal is much more abundant than oil ornatural gas. There are proven coal reserves of 826billion tons of coal, with a proven reserve-to-production ratio of 122 years (BP 2009).8 Coalreserves are available in almost every country, withrecoverable reserves in around 70 countries. Sixcountries together account for about 80% of coalreserves, as shown in Figure 4.1.

Given that world coal reserves are spreadacross a politically and geographically diverse setof countries, large in number and including someof Europe’s closest political allies, the prospect ofsignificant supply interruption seems relativelyremote. It is therefore implausible to argue thatdependence on coal imports is a significant threatto EU security of supply.

Uranium

The earth has 5.5 million metric tons of identifieduranium resources, distributed widely around the

world, as Figure 4.2 illustrates. At the current rateof consumption, this would constitute about 100years’ worth of supply.

Uranium’s extraordinarily high energy densitymakes it practical to maintain large stockpiles(Euratom 2008), reducing the risks associatedwith a short-term interruption in supply.This fac-tor and the diverse range of supply sources suggestthat dependence on uranium imports is not a sig-nificant security-of-supply risk for Europe,despite the high level of import dependence, una-voidable given that Europe has less than 2% of theworld’s identified uranium resources (EuropeanCommission 2008b).

Gas

The picture for natural gas is very different thanthat for coal or uranium. Prima facie there is goodreason to consider that the EU’s import depend-ence does represent a potential threat to securityof supply. As noted earlier, the EU imports morethan half of its gas, of which a large proportioncomes from Algeria and Russia, and gas importsare predicted to increase in coming years (Euro-pean Commission 2007) as output continues todecline in the main EU producing nations.

Dependence on Algerian and Russian gas is ofconcern because of the absence or weakness ofdemocratic institutions and transparent govern-ance arrangements in these countries. Algeria hasexperienced recent civil war and ranks poorly oninternational league tables in terms of democracy

South Africa3.7%India 7.1%

Australia 9.2%

China 13.9%

Other 18.2%

United States 28.9%

Russian Federation19.0%

Figure 4.1. World coal reserves

Australia 23%

Kazakhstan15%

Russia 10%

South Africa8%

Canada 8%

United States6%

Niger 5%

Namibia 5%Brazil 5%

Figure 4.2. World uranium resources

54 Boaz Moselle

and civil rights (World Audit 2010). Russia alsohas a low rank, and the poor climate for businessinvestment raises questions as to whether newinvestment required to maintain and increase gassupply will be forthcoming. There is also a ques-tion as to how far the supply of gas is a commer-cial decision versus an instrument to exercise geo-political influence. This means that the extent towhich the supply of gas will respond to increaseddemand is unclear.

In addition, analysts have noted that Russianeeds to replace declining fields with new pro-duction from the Yamal Peninsula and offshorefields and to refurbish a large, aging high-pressurepipeline network (Stern 2005). As mentionedabove, there is also a need to invest in the Ukrain-ian pipeline network or construct new pipelinesto maintain transit capacity to Europe.

Underlying these concerns is the absence ofrealistic alternative sources of natural gas. Relativeto other fuels, the ability to bring gas from differ-ent sources is inherently limited by the morecostly, capital-intensive and inflexible means oftransportation required, in the form of long-distance pipelines or liquefied natural gas (LNG).Moreover, while gas remains abundant at globallevel, with world proven reserves as of 2007 stand-ing at some 177 trillion cubic meters (tcm),9

equivalent to some 60 years of consumption atcurrent rates (BP 2009), those reserves are con-centrated in a small number of countries, asshown in Figure 4.3. Of these, just three coun-tries, Russia, Iran, and Qatar, hold about 53% ofthe total.

There is an unknown potential for Europeandomestic gas supply to be boosted by unconven-tional or shale gas. In the United States, substan-tial discoveries of unconventional gas have beenmade.10 However, estimates of the potential forunconventional gas in Europe are lower. Onestudy estimates that Europe has 29 tcm, whereasthe United States has around 233 tcm of uncon-ventional gas (Holditch 2007).11 Moreover, theability to extract the resources will depend onenvironmental consents and the cost of extractingunconventional gas in Europe.

In conclusion, it seems that gas importdependence is a potentially significant security-

of-supply risk for the EU. It is possible that acombination of LNG imports and the arrival ofunconventional gas (either in Europe or in theUnited States but “liberating” LNG flows thatcould come to the EU) will mitigate the problem.It is also possible that the risk is overestimatedbecause of the mutual dependence between theEU and its suppliers: revenue from gas sales is ofgreat importance to both Russia and Algeria, andindeed, they have been known to express concernabout “security of demand” from the EU, mirror-ing the EU’s concerns about security of supply(see, e.g.,Yenikeyeff 2006). Neither of these pos-sibilities can be viewed as certain, however, andthe risk is therefore a real one, albeit difficult toassess or quantify.

The problem is particularly acute for easternEurope. Estonia, Latvia, Lithuania, Bulgaria,Slovakia, and Finland are completely dependenton Russia for gas imports, while Greece, Hun-gary, and Austria are more than 80% dependent(European Commission 2008b). Among the sevennew eastern European member states, depend-ence on Russian gas imports averages about 77%(European Commission 2009a). Eastern Euro-pean commentators point to the experience inLithuania—where oil supplies from Russia to theMazeikiu refinery were halted because, it isclaimed, Russia objected to its sale to a Polishrefiner, PKN Orlen—as a sign of the potentialrisks they face (Geropoulos 2007). The politicaltemperature is clearly at its highest with regard toeastern Europe, given Russian resentment at itsloss of influence there since the breakup of theSoviet Union.

Algeria 2.4%

Russia 23.4%

Iran 16.0%Qatar 13.8%

Turkmenistan4.3%

Saudi Arabia4.10%

United States3.6%

Nigeria 2.8%United Arab

Emirates 3.5%

Venezuela 2.6%

Figure 4.3. World gas reserves: top 10 countries

Renewable Generation and Security of Supply 55

The Case for Policy Intervention

Gas import dependence is therefore an under-standable source of concern for Europeanpolicymakers. However, it does not automaticallyfollow that policy intervention is warranted. Mar-kets already provide strong incentives for marketparticipants to appropriately ensure against unreli-able supplies. Contracts between suppliers andconsumers generally oblige suppliers to deliverenergy, and suppliers that choose to contract withless reliable sources (in other words, are too relianton gas from Russia and Algeria) will face what-ever penalties their contracts contain. These pen-alties are negotiated on a bilateral basis and there-fore represent accurately the costs to consumers ofloss of supply, or the trade-offs consumers arewilling to make between price and security ofsupply (for example, a consumer may be willingto sign a contract that has no penalties in the eventof supply failure, such as through the operation ofa force majeure clause, but in that case the supplieraccepts a lower price in return). A similar logicapplies for consumers that choose to rely onshort-term contracts or spot markets: they acceptthe higher level of risk in return for greater flex-ibility or an expected lower price.

The key question therefore is whether theseincentives are sufficient to provide an efficient12

level of security—or, more accurately, whetherthey provide a more efficient level of security thancan be expected from policy intervention, bearingin mind that real-world policy interventions andreal-world markets are both inherently imperfectcompared to any theoretical ideal.

In that context, a number of problems couldundermine the ability of these market-basedincentives to give an efficient outcome. Theseinclude some market failures that typically providethe theoretical justification for policy interven-tions, but also other issues that are arguably moreimportant from both normative and positive per-spectives (i.e., they should be taken more seri-ously, and they have a bigger impact on policyoutcomes).

First, it may be that consumers (individualsand firms) are not good at making judgments ofthis kind, and that governments could make better

judgments and use them to implement betterpolicies. A case may therefore exist for interven-tion on essentially paternalistic grounds.

The second problem is the issue of politicallymotivated supply interruptions. Arguably, the riskof supply interruption by a hostile state actor isgreater the more disruptive the effect of the inter-ruption.13 Thus, although ensuring against lowrainfall in a hydro-dominated power system doesnot make rain more likely to fall, ensuring againstgas supply interruptions in the EU actuallyreduces the threat of interruption, because if suchinterruptions are relatively painless, then a hostilestate gets little strategic benefit from interruptingor threatening to interrupt supplies. If so, thenindividual investments in supply security (e.g.,booking more gas storage or installing dual fuelcapability at gas-fired power stations) create apositive externality, and as with any such external-ity, there will be an incentive to free ride: con-sumers will spend less than is socially optimal,because they face all the costs but only a small partof the benefits (the so-called “tragedy of the com-mons”). Moving down the chain, it follows thatsuppliers will not face appropriate incentives toensure security, and the market will under-provide security.

Third, experience shows that in conditions ofenergy scarcity, regulatory or political interven-tion will almost certainly prevent the market fromfunctioning efficiently.The prospect of such inter-vention will therefore undermine investmentincentives. For example, a private investor mightconsider investing in a gas storage facility even ifthe market already appears well supplied with gasstorage, on the basis that it would offer a very highreturn in the low-probability event that a majorgas shortage leads to prolonged spikes in spot gasprices. Experience in Great Britain suggests thatthese spikes could involve prices many times ashigh as under normal conditions,14 implyingspectacular returns to anyone holding gas in stor-age.

In reality, however, the investor will be awarethat many regulators or governments havearrangements in place that suspend the pricemechanism in such emergencies. Such an investorwill also be aware that even if those mechanisms

56 Boaz Moselle

are not yet in place, they could be introduced atshort notice, and moreover, in the absence ofprice controls, they would be likely to suffer ret-ribution if they were judged to have profiteered or“price-gouged” during a crisis.15

Conversely, market participants will also beaware that emergency arrangements typicallyinvolve the imposition of “shared pain” rules,which undermine private incentives to ensureagainst scarcity of supply. For example, scarce gassupplies might be allocated to all suppliers on apro rata basis related to the size of their customerload. Such an outcome would do nothing toreward the supplier who had purchased gas from amore reliable source.

Examples of these two tendencies—nonreliance on the price mechanism and a“shared pain” approach—can be found in the gas“emergency cash-out” arrangements in GreatBritain, which suspend the market-based deter-mination of prices for the duration of the emer-gency (Ofgem 2006). The European Commis-sion’s proposed new legislation on gas security ofsupply provides for a variety of non-market-basedmeasures including compulsory demand reduc-tion and forced fuel switching (European Com-mission 2009c).

A fourth problem is a more conventional mar-ket failure: because security of supply is to someextent a public good, markets will tend toundersupply it.16 Specifically, the issue arises fornatural gas and electricity because they are trans-mitted via networks used by many consumers,and most individual consumers cannot beremotely interrupted when supplies are tight.17

Domestic and commercial consumers generallydo not have real-time metering and are notexposed to higher spot prices when supplies arescarce. There is therefore no incentive for indi-vidual consumers to ensure against supply risks,for example by paying more to purchase from asupplier that has more gas in storage.18

In particular, with electricity the absence ofremote disconnection means that in the event of ablackout, all consumers will lose supply in an area,even though in general it would be possible to seta price—if all consumers were exposed to real-time prices—that would allow demand to match

supply. An individual consumer therefore has noincentive to purchase energy from more reliablesources, as the higher levels of security are spreadacross all consumers. Purchasing from a more reli-able source creates a positive externality butalmost no private benefit (see Joskow 2007).Again, this gives rise to free riding andunderprovision of security by the market.

Assessment

How material these problems are for security ofsupply is a difficult empirical question, both inabsolute terms and because assessment should beagainst a counterfactual that is based on a realisticassessment of the likely intervention that thepolicy process would give rise to. Nonetheless,some simple observations are in order.

First, the paternalistic argument that consum-ers are unlikely to make wise decisions is at leastplausible. Extensive research in psychology andbehavioral economics shows that human beingshave particular difficulty making decisions involv-ing low-probability events that are well outsidetheir normal range of experience (Tversky andKahneman 1992). However, the claim that gov-ernment intervention will lead to better outcomesis more contentious (apart from any other consid-eration, governments are made up of humanbeings subject to the same biases as others).

Second, the argument concerning politicallymotivated interruptions is also at least plausible.The key question for EU policymakers to answeris how likely Russia is to interrupt gas supplies forpolitical reasons. On the one hand, instances havealready occurred where Russia has cut off oil sup-plies to an EU member state for essentially politi-cal reasons. On the other hand, as noted earlier,profits from Gazprom are of great importance toRussia and members of its political elite, creatinga relationship of mutual dependence betweenRussia and the EU.

Third, the combination of regulatory andmarket failures described above seems to implythat all but the very largest consumers are cut offfrom any of the upside from investing in enhancedsecurity of supply.

Renewable Generation and Security of Supply 57

Finally, whatever the objective merits, it is alsoclear that governments are increasingly set onintervention in this area, at EU and national lev-els. In the context of this book, it is thereforeappropriate to ask whether, assuming thatpolicymakers are intent on intervention, the pro-motion of renewable generation is the best formof intervention to address the EU’s concernsabout import dependence and its impact on secu-rity of supply.

Is Promotion of Renewablesthe Right Intervention?

It is natural to expect that the promotion ofrenewable generation will reduce gas consump-tion and hence gas import dependence, by substi-tuting away from gas-fired generation. Analysiscarried out for the European Commission is con-sistent with this logic. Table 4.2 shows the pre-dicted effects of the EU’s new 20-20-20 energypolicy adopted in 2008, whose main componentsare commitments to achieve several goals by theyear 2020: reduce demand, with an indicative tar-get of 20% reduction in energy consumption rela-tive to business as usual; increase the use of renew-

able energy, with a binding target of 20% of finalenergy consumption; and reduce greenhouse gasemissions, with a 20% reduction relative to 1990levels.

As the table shows, the combination of meas-ures is predicted to reduce gas imports by about aquarter, relative to business-as-usual. However,this analysis also raises a number of questions.

First, it is clear that much of the reduction ingas consumption reflects the impact of energyefficiency measures that reduce total energy con-sumption, rather than the displacement of gas-fired generation by renewables. Indeed, one cansee from the table (see the “Impact of new policy”rows), that the predicted reduction in gas con-sumption induced by the new energy policy (sec-ond column) is much larger than the inducedincrease in renewable energy (last column).

Second, although the analysis does not allowone to separate out these two effects, it is likelythat the impact of renewables on gas-firedgeneration is materially affected by the need forcontinued use of gas to provide flexible backupfor intermittent renewables. Recent analysis byCapros et al. (2008) suggests that the new energypolicy will reduce coal-fired generation signifi-

Table 4.2. Energy consumption and import dependence by 2020

Gasimports(Mtoe)

Gas con-sumption(Mtoe)

Gas imports/consumption(%)

Solidsimports(Mtoe)

Solids con-sumption(Mtoe)

Solids imports/consumption(%)

Renewableenergyproduction(Mtoe)

2005 257 445 57.8% 127 320 39.7% 122

2020 (oil $61/bbl)

Business as usual 390 505 77.2% 200 342 58.5% 193

New policy(20-20-20)

291 399 72.9% 108 216 50.0% 247

Impact of newpolicy

–99 –106 –4.3% –92 –126 –8.5% 54

2020 (oil $100/bbl)

Business as usual 330 443 74.5% 194 340 57.1% 213

New policy(20-20-20)

245 345 71.0% 124 253 49.0% 250

Impact of newpolicy

–85 –98 –3.5% –70 –87 –8.0% 37

Source: European Commission 2008b, 65

58 Boaz Moselle

cantly more than gas-fired, both because of theimpact of carbon prices and because renewablepower requires extensive support by flexiblereserve power, supplied mainly by gas units.Indeed, the analysis in Table 4.2 also shows theimpact of the new policy on coal (solids) to beequal to or larger than the impact on gas.

Third, it is unclear which gas sources are mostlikely to be affected by the reduction in gasimports. If the main effect of the policy is to dis-place imports of LNG from relatively friendlysources, then the effect on security of supply issmall. However, this is likely to be the case. LNGis often viewed as the marginal source of gas,because of the relatively high cost of bringingLNG to the EU, and also because Algerian andRussian producers are to some extent captive sup-pliers, given the high cost of attempting to diver-sify away from their European customer base.19

Finally, the analysis described above suffersfrom a more fundamental flaw, in that thebusiness-as-usual counterfactual is arguably some-thing of a straw man. A more interesting counter-factual would be a scenario with a policy thatinvolves the promotion of all forms of low-carbonenergy on a technology-neutral basis: a carbon taxor cap-and-trade scheme (in this context, the EUETS with a tighter cap) and no policies aimedspecifically at promoting the large-scale deploy-ment of renewables.20

The effect of such a policy would be to pro-mote some combination of energy efficiencymeasures, nuclear power, coal-fired generationwith carbon capture and storage (CCS), andrenewables. The noteworthy point here is that ofthose four classes of technology, renewablegeneration—at least in the forms of wind, solar, orwave power—may well be the least suited toenhancing security of supply, because as notedearlier many renewable generation technologiesare intermittent and will likely be associated withcontinued extensive use of gas-fired generation as“backup”.21 It is therefore likely that they will dis-place less gas-fired output than equivalentamounts of nuclear power or coal-fired genera-tion (or investments in energy efficiency).Although increased use of nuclear power or coal-fired generation would probably entail increased

imports of uranium or coal, I have argued abovethat no significant security-of-supply issue shouldarise from such imports.

In conclusion, therefore, a policy that pro-motes low-carbon generation in general wouldprobably be more effective in addressing gasimport dependency and enhancing security ofsupply than the current policies that specificallypromote renewable generation.

IntermittencySome of the most prominent forms of renewablegeneration—in particular wind but also solar andwave power—are variable in output, with thelevel of production determined by exogenous fac-tors such as wind speed, and also unpredictable toa lesser or greater degree. “Intermittency” is theterm generally used to refer to this combinationof variability and relative unpredictability.

Two concerns arise from the intermittentnature of renewable generation. A short-run con-cern is the impact on “system balancing”—ensuring that supply and demand of power arematched on a second-by-second basis. A long-runconcern is whether a liberalized power market canbe relied on to produce enough investment tomeet the much greater need for backupgeneration—flexible capacity that will be usedprimarily when demand is high and wind outputis low, and whose overall utilization will thereforebe comparatively low.

System Balancing

The basic physics of electric power systemsrequires that production and consumption22 arematched on a second-by-second basis. In anypower system, a system operator (SO) is responsi-ble for continuously ensuring this balancing. TheSO has short-term control of certain generatingassets, which it uses close to and in real time tocorrect any difference between the amounts ofelectricity supplied to the system and the amountbeing consumed.

Small deviations from perfect balance takeplace continuously and result in fluctuations in the

Renewable Generation and Security of Supply 59

frequency of AC power. Certain generating unitsare configured to react automatically and instanta-neously to these deviations. This so-called “pri-mary reserve” acts as a first line of defense againstimbalances. In case of larger deviations, after theimmediate response of the primary reserve, gen-erators providing the so-called “secondaryreserve” increase or reduce injections within sec-onds, following the instructions of a centraldevice in a process known as automatic genera-tion control. Secondary reserve is a scarceresource, because it is provided by units with spe-cific technical capabilities. As soon as possible,therefore, typically with a lag of minutes, injec-tions by units providing so called “tertiaryreserve” are increased or decreased, following theinstructions of the system operator, and secondaryreserve capacity is restored to the pre-deviationlevel.

In a liberalized market, the SO generally con-tracts with generators, and sometimes large con-sumers, to procure these services.23 The nature ofthe reserve contracts varies, but for the purpose ofthis chapter, it is sufficient to note that the SO willpay plants to be available to provide balancingservices, as well as for the provision of the serviceswhen called on.

Clearly the task of system balancing becomesmore difficult the greater the changes in the levelsof output, especially if those changes areunpredicted or occur with only very shortnotice.24 The prospect of high levels of penetra-tion of intermittent generation therefore gives riseto concern that the job of system balancing willbecome more costly and less certain of success:the SO will have to purchase more balancingservices, and if it fails to purchase enough, it couldfind itself overwhelmed by unexpectedly volatileshifts in output from intermittent generation,endangering security of supply.25

System Stability Implications

From a system stability perspective (i.e., in termsof the risk of supply disruptions), these concernsare probably exaggerated. The more technicalaspects of the system-balancing challenges posedby intermittent generation are addressed in Chap-

ter 2 and references therein. In brief, it is clearthat significant advances have been made in theability to forecast wind speeds and the outputfrom wind generation, such that while high levelsof penetration of wind generation may add to thecost of system operation, they need not under-mine system stability. Current evidence suggeststhis is the case at least for penetration up to 20%(i.e., with up to 20% of electric power being gen-erated by wind).

The issue is at present less clear for otherintermittent sources, and in climates with cloudyskies, solar photovoltaic (PV) power may presentgreater challenges, as cloud cover means the vari-ability in output can occur over seconds ratherthan hours (although geographic dispersion willmitigate this to some extent). Nonetheless, Boyleargues in Chapter 2 that they can probably bedealt with in similar fashion (for more details, seealso Boyle 2007).

The findings of a very comprehensive surveypaper by Gross et al. are consistent with this con-clusion: “none of the 200+ studies [we have]reviewed suggest that introducing significant lev-els of intermittent renewable energy generationon to the British electricity system must lead toreduced reliability of electricity supply” (2006,iv).

However, these conclusions do assume thatadvances in forecasting will be effectively incor-porated into system operation procedures. Chap-ter 11 notes the example ofTexas, where a much-discussed emergency occurred in 2008 followinga rapid reduction in wind output. The reductionhad been predicted by commercially availableforecasts, but the SO had not purchased thoseforecasts.

Cost Implications

The same survey by Gross et al. (2006) alsoanalyzes the cost implications of intermittency,looking at how much additional reserve capacityis likely to be required and how much thisis likely to cost. The authors conclude that “forpenetrations of intermittent renewables up to 20%of electricity supply, additional system balancingreserves due to short term (hourly) fluctuations in

60 Boaz Moselle

wind generation amount to about 5–10% ofinstalled wind capacity. Globally, most studies esti-mate that the associated costs are less than£5/MWh of intermittent output, in some casessubstantially less.” Of course, an additional cost of£5 ($7.50) per MWh is a material issue,26 but thatforms part of a larger set of questions about thecost-effectiveness of renewable generation and isnot really a security-of-supply issue.

All of this analysis assumes, however, that thenecessary reserve will be there for the SO to callon. This naturally leads back to the question ofinvestment incentives.

Investment in Backup Generation

Given the difficulty of storing electricity and thelimited potential for shifting demand across time,the use of intermittent generation means that alarge set of backup generation is required toensure that demand can be met at times when theintermittent sources have low availability becauseof a lack of wind, sunshine, and so on. This needfor spare capacity is not unique to systems withintermittent generation: no type of generation isavailable with 100% certainty, and conventionalunits also close down for planned and unplannedmaintenance. Nevertheless, large-scale penetra-tion of intermittent generation gives rise to amuch higher requirement.

The size of this requirement will clearlydepend on the level of penetration of intermittentgeneration, the technologies involved, the specificelectricity system, relevant physical features (e.g.,the geographic and temporal distribution ofwind), and many other factors. This has been theobject of many engineering studies. For the pur-pose of synthesis, it is convenient to summarizeany such study in terms of its estimated “capacitycredit,” which measures how much conventionalthermal generation is displaced by a unit of inter-mittent generation. So, for example, a capacitycredit of 20% means that adding 100 megawatts(MW) of intermittent generation would allowone to retire 20 MW of conventional generationwhile maintaining the same overall level of systemsecurity.

A comprehensive survey of these studies canbe found in Gross et al. (2006), whose summaryof the estimates of the capacity credit from 19 ofthe studies is shown in Figure 4.4.

Clearly a capacity credit in the range impliedby these studies would add significantly to thetotal capital costs of the system. With regard tosecurity of supply, however, the concern is that aliberalized market will not have sufficient invest-ment to provide the required level of generationcapacity.

40

35

30

25

20

15

10

5

0 5 10 15 20 25 30 35 40Intermittent generation penetration level (% of total system energy)

Cap

acity

cre

dit (

% o

f ins

talle

din

term

itten

t gen

erat

ion

capa

city

)

160

249

244250

24674

241243 238

83 204

248

242

247

515

79240

121 83

Source: Gross et al. 2006, 43

Note: the shaded area refers to UK studies

Figure 4.4. Capacity credit values

Renewable Generation and Security of Supply 61

Starting Point: Excess Capacity

In the short run, there may be little issue with theavailability of reserves and peaking generationmore generally, because as new intermittentcapacity is added to the network, the existingconventional capacity remains available. Althoughgenerators could choose to retire this capacity, theincentives to do so are relatively weak, because theinvestment is already sunk and profits from opera-tion need cover only annual fixed costs (such astransmission charges or taxes levied annually) tomake it worth keeping the plant open.

Experience to date in Germany and Spain isconsistent with these arguments. Sensfuß et al.(2008) note that for Germany, “the developmentof renewable electricity generation has had nomajor impact on investments into new generationcapacity up to the year 2006. One reason is thatthe period after the liberalization of the electricitymarket was characterized by excess capacity and asubsequent decommissioning of power plants,”while “most of the decommissioned capacity wasdecommissioned for economic reasons such aslow efficiency of the plant, need for repairs orinefficient use of expensive fuels such as oil andgas.” Chapter 15 in this book describes the evolu-tion of Spanish capacity, characterized by highlevels of excess capacity due in large part to therapid expansion of renewable generation, and asyet without any consequent retirement ormothballing of plants.

In some markets, however, this initial over-hang of excess capacity might erode relativelyquickly, for a number of reasons:

• If there is too much capacity, then pricesmight fall to a level that induces plantmothballing or early retirement. Chapter 15indicates that this situation may be develop-ing in Spain.

• Generators with zonal or regional marketpower might have an incentive to retire someof these plants so as to raise peak prices andthe price of reserve.

• Incentives for early retirement may be exac-erbated by the costs of refurbishments,including those necessary to meet the

requirements of new environmental legisla-tion. So, for example, in the EU, the cost ofadding “scrubbers” to coal-fired units by2015, to comply with the Large CombustionPlant Directive, would have to be recoveredthrough future profits, and this could be dif-ficult if utilization is expected to be very low.

Investment Incentives in Energy-Only Markets

In the long run, however, there is a real questionas to whether energy markets will deliver theneeded investment. This question falls into awider debate as to the ability of liberalized energymarkets to provide sufficient levels of investment.The issue has been extensively discussed in aca-demic and policy circles for some years (Cramtonand Stoft 2005; Stoft 2002). This chapter can dono more than briefly sketch out the main posi-tions taken.

The issue relates specifically to “energy-only”markets, where generators’ only sources of rev-enue are the sale of electricity and the provision ofreserves, as described earlier in this chapter.27

Theoretical models suggest that although genera-tors in a competitive energy-only power marketcan earn sufficient operating profits to cover theircost of capital (i.e., the variable profit from sellingpower can provide a sufficient return on invest-ment), the requirements for that to happen arerather stringent and may not be met in practice inmost real-world power markets.

The problem arises because if such a market iscompetitive, then the spot price of electricity willapproximate the marginal cost of the most costlygenerator being called on—the “system marginalcost”—at any point in time except hours whendemand (strictly speaking, demand for energy andoperating reserves) exceeds available capacity. Inthose hours, it is possible for price to exceed sys-tem marginal cost, for example if it is set by price-responsive demand. The difference between priceand system marginal cost is referred to as a “scar-city rent”.

For peaking plants (the plants with the highestmarginal cost on the system), these scarcity rentsare the only way to create a return on capital. It is

62 Boaz Moselle

possible to show that at least in theory, scarcityrents are also necessary for plants with lower mar-ginal cost, if they are to earn a sufficient return tocover their sunk costs. It is therefore necessarythat prices in those hours be sufficiently high toprovide an appropriate return on capital, i.e., onethat will provide the right incentive for new gen-eration in peaking plants.

Prices at times of scarcity should generally beset either by demand-side response or by actionsof the system operator in its procurement of oper-ating reserves (Hogan 2005). If those mechanismswork appropriately, then it can be shown that intheory, the level of scarcity rents will be efficient,in the sense of ensuring that generators earn theircost of capital and have appropriate incentives fornew investment.

However, this outcome depends on the pres-ence of flexible scarcity pricing mechanisms andthe absence of market or regulatory imperfectionsthat limit demand-side response or distort systemoperator decisions. In practice, such imperfectionsare endemic:

• The development of mechanisms to allowdemand-side participation has been generallyrather slow in most electricity markets, limit-ing the potential for demand-side response toset prices at times of scarcity.

• The protocols followed by many systemoperators at times of scarcity do not lead tothe appropriate level of scarcity pricing.28

• The potential for prices to depart from gen-erators’ marginal costs at times of scarcity isoften limited by administrative measures con-straining prices, to mitigate market power orfor other reasons. For example, many cen-trally organized markets (“pools”) haveexplicit price caps in place,29 and some limitthe offer prices as a part of the ex ante marketpower mitigation process. In many markets,regulators monitor prices and perform expost investigations of price spikes, with achilling effect on scarcity pricing.

It is therefore argued that in practice, imperfec-tions in energy-only markets will lead to

underinvestment, particularly in peaking capacity.This issue is commonly referred to as the “missingmoney” problem.

On the other hand, proponents of a moremarket-oriented approach argue or assert that inthe absence of price caps, the market can in prac-tice be expected to provide sufficient capacity,theoretical models notwithstanding. In GreatBritain, this view underlies the existing marketdesign, the so-called New Electricity TradingArrangements (NETA), which does not have anyprice caps in place. Practical experience in thedecade since NETA was put in place is somewhatambiguous. Despite many claims of imminentcrisis, the lights have stayed on. However, this hasbeen achieved with very little new investment ingeneration, indicating that the system may haveenjoyed an overhang of excess capacity from thepreceding decade.

In sum, there are theoretical reasons to believethat in the absence of some form of capacitymechanism, a competitive energy-only marketwithout efficient scarcity pricing mechanisms mayunderdeliver on investment in reserve capacity(i.e., in flexible units that will experience lowaverage utilization). Although the materiality ofthose concerns is open to debate, it is clear thatany problems would be exacerbated considerablyby the much greater need for such units thatcomes with high levels of penetration by renew-able generation. Moreover, in practice, concernsabout underinvestment are likely to be wellfounded because of the combination of explicitprice caps and the implicit threat or shadow offuture price regulation in most or all liberalizedmarkets. Even in Great Britain, which untilrecently was viewed as the paragon of energymarket liberalization, reregulation is now beingopenly discussed (see, e.g., Ofgem 2010).

The example of Great Britain also illustrates adeeper problem with investment incentives in thecontext of current energy policy, of which policytoward renewables forms only a part. The natureof the policy response to climate change, particu-larly in the EU, means that all forms of investmentin new generation capacity are heavily influencedby government intervention. Thus renewables,nuclear, and CCS each attract technology-specific

Renewable Generation and Security of Supply 63

forms of support, whereas environmental regula-tion in the form of the EU Emissions TradingScheme as well as non-climate-related measures30

affect the relative and absolute returns on differenttechnologies. Investments are thus arguably sub-ject to very high levels of political risk, and it is byno means clear that markets are able to assess andbear these risks.

In conclusion, therefore, the possibility thatcompetitive liberalized markets will struggle toprovide sufficient peaking capacity to accommo-date large amounts of intermittent generation is avery real one, for a variety of reasons. The biggestfactor undermining investment incentives is thehigh level of uncertainty and political risk, whichaffects all generation investments to a lesser orgreater degree, except for projects that can rely onexplicit and iron-clad government guarantees.

The design of wholesale power markets mayneed to change to reflect these concerns, by pro-viding stronger and more reliable incentives forinvestment, such as in the form of capacity pay-ments or similar mechanisms. Capacity paymentsare widely used in the United States and have hadsome application in Europe (in Spain, for exam-ple, as well as in England and Wales in the 1990s)(Perekhodtsev and Blumsack 2009). These arepayments to generators that are additional to therevenue they receive from the sale of energy. Dif-ferent countries have taken alternative approachestoward implementing such a mechanism. InEurope, the approach generally has been for thetransmission system operator (TSO) to make pay-ments to generation on the basis of its availabilityto generate, recovering those payments as a sur-charge on transmission tariffs. In the UnitedStates, regulators have tended to place obligationson demand-side participants to contract forwardfor capacity via organized “capacity markets”.Thelevel of the obligation then determines thedemand for capacity in those markets, and thatcombines with supply to determine a capacityprice. In either case, the details of design (includ-ing, for example, determining the appropriatelevel at which to place the price, or the quantityassociated with an obligation) are extensive andpotentially challenging (Harvey 2005). For thepurposes of this chapter, it is sufficient to note that

high penetration of intermittent generationmeans that a number of EU regulators are likely tobe addressing those challenges in the comingyears.

Finally I note one caveat: the picture may besomewhat different in countries where generationinvestment decisions are more naturally influ-enced by informal ties between industry and gov-ernment. For example, in Germany neededinvestments may take place as the outcome ofinformal (or at least non-contractual) agreementsbetween government and industry, rather thanbeing either a pure market outcome or oneinduced by regulatory mechanisms such as cap-acity payments.

ConclusionsDependence on imported gas gives rise to real,albeit hard-to-quantify, security-of-supply issuesfor the EU, because of geopolitical concernsaround both Russia and Algeria. Those problemsare particularly acute for many of the new EUmember states in eastern Europe, where depend-ence is highest and relations with Russia are moststrained. Dependence on imported coal and ura-nium does not give rise to such concerns, becauseof the number, diversity, and friendliness ofpotential sources.

Market outcomes may not provide an efficientlevel of protection against the security-of-supplyrisks associated with gas import dependence,because of a variety of market and regulatory fail-ures. However, the promotion of renewable gen-eration is not the best policy response. Increasedpromotion of all forms of low-carbon energy(including energy efficiency) would appear to beat least as effective in enhancing security of supply,at lower overall cost.

Security-of-supply concerns related to inter-mittency and its impact on system operation andgrid stability are exaggerated. In particular, recentimprovements in wind forecasting mean that evenrather high levels of penetration for wind genera-tion can safely be accommodated by an efficientlyrun and appropriately regulated system operator.The impact is one of cost rather than a threat to

64 Boaz Moselle

stability. High levels of penetration of intermittentgeneration do, however, raise real questions aboutmarket design and security of supply—in particu-lar, whether existing energy-only markets willprovide strong enough incentives for the invest-ment needed in peaking generation to cope withperiods where high demand coincides with lowintermittent output.

In principle, market mechanisms are sufficientto ensure the right levels of investment. In prac-tice, however, the absence in most EU powermarkets of appropriate mechanisms for scarcitypricing, combined with very high levels of regu-latory uncertainty and risk, suggests that theremay be a need for some form of enhanced incen-tive such as capacity payments that reward genera-tion for availability, except in markets where thelevel of investment is strongly influenced byimplicit regulation and consensus-based decision-making involving industry, government, andother stakeholders.

AcknowledgmentsThanks to Luis Agosti, David Black, GodfreyBoyle, Toby Brown, Guido Cervigni, DmitriPerekhodtsev, and Dick Schmalensee for manyhelpful suggestions and input. All errors and omis-sions are mine.

Notes

1. A parallel argument is made in the EU and theUnited States about the benefit of renewable fuelsin reducing the risks arising from dependence onimported oil for transportation. The focus of thisbook, however, is on renewable generation.

2. The bulk of the remainder was from renewables(14%).

3. Russia provides 42% of the EU’s gas imports,Norway 24%, Algeria 18%, and Nigeria 5%.

4. A more recent forecast is even more dramatic,showing a fall from 166 Mtoe in 2010 to113 Mtoe in 2019 (ENTSOG 2009); all figures

converted from billion cubic meters (bcm) toMtoe using 1 bcm = 0.90 Mtoe (www.bp.com/conversionfactors.jsp).

5. The 390 Mtoe figure assumes an oil price of $61per barrel (bbl). A second business-as-usual sce-nario has an oil price of $100/bbl and forecasts netimports of 330 Mtoe (75% of total consumption).

6. Some observers note that the actual price paid byUkraine is higher than the contracted pricebecause of an agreement due to arrangements toprovide free gas in exchange for delivery of gasinto Ukraine. However, even allowing for theadditional cost, the price remains well belowEuropean levels. See Chow and Elkind 2009.

7. In principle, Gazprom did not cut off supplies tothe EU; it reduced the level of flows by theamount of gas that previously would have beenintended for Ukraine, while continuing to flowgas for transit across Ukraine to the EU. However,it was easily predictable that Ukraine would con-tinue to consume gas, with the effect of reducingtransit flows significantly.

8. Corresponding figures are 42 years for oil and 60years for gas (BP 2009).

9. The corresponding figure for 1987 was 70 tcm.10. The U.S. Energy Information Agency (EIA 2008)

has reported increases in the level of proven gasreserves as a result of the development of uncon-ventional gas resources. The Potential Gas Com-mittee (2009) reported an increase in reserves,(including proven, possible and speculativereserves) in 2008 to the highest level in its 44-yearhistory.

11. Figures converted from trillion cubic feet (tcf) totcm using 1 tcm = 35.3 tcf (www.bp.com/conversionfactors.jsp).

12. Here “efficiency” refers to the trade-off betweencost and risk. Arrangements are efficient if theadditional cost of investing to increase securityoutweighs the additional benefit (and the savingfrom spending less does not justify the increasedlevel of risk).

13. So either supply interruptions become moreprobable/frequent, or society pays a higher priceto avoid them, in the form, for example, of highernational security costs or unwanted changes inforeign policy to appease the potential interrupter.This argument has been used in the past to justifythe requirements for strategic oil storage.

14. For example, in 2005, a combination of factorstemporarily reducing supplies to the United King-

Renewable Generation and Security of Supply 65

dom led to price spikes of up to 500% betweenFebruary 23 and March 11 (Trade and IndustryCommittee 2005).

15. Some U.S. states even have legislation specificallyprohibiting price-gouging. For example, FloridaStatute 501.160 states that during a state of emer-gency, it is unlawful to sell “essential commodi-ties” for an amount that grossly exceeds the aver-age price for such commodities during the pre-ceding 30 days.

16. A public good is a good that is non-rivalrous andnon-excludable. This means that consumption ofthe good by one individual does not reduce avail-ability of the good for consumption by others, andthat no one can be effectively excluded from usingthe good.

17. In other words, it is not possible for the systemoperator to cut off supply to individual consumers,other than very large consumers (who often have“interruptible supply” contracts that allow forsuch actions).

18. Given metering technologies currently in place, itis not possible to create such an incentive. Forexample, gas meters typically record only cumula-tive consumption, and unless they were read on adaily basis—which would clearly be impossiblycostly—there would be no way to know howmuch has been consumed by an individual cus-tomer on a day when supplies were particularlyscarce.

19. This is a general analysis; individual import con-tracts can vary significantly.

20. Another interesting counterfactual, and one thatin principle should be the starting point for thedesign of any intervention, would be to use taxa-tion to correct for any security-of-supply exter-nalities. In theory, this might lead to different lev-els of taxation applied to gas from differentsources, with Russian gas probably incurring thehighest tax. In practice, this could create difficul-ties with World Trade Organization (WTO) rules,and it would also raise difficult questions about thequantitative assessment of the size of the external-ity. A more realistic approach would be to tax allgas. However, this would also be politically diffi-cult because of the aversion of key member states(notably the United Kingdom) to EU-level taxes,and because the United Kingdom and the Neth-erlands are both major gas producers.

21. This is not to assert that intermittency per se is asecurity-of-supply risk (see next section), butmerely to observe that all else being equal, inter-

mittent generation will displace less gas-fired gen-eration than will non-intermittent. Clearly thiswould not apply to hydro generation or tobiomass. The potential for new hydro is relativelylimited, however, and intermittent sources (in par-ticular wind) are forecast to be the dominant formof new installed renewable generation capacity inthe coming decade at least.

22. Including consumption in the form of losses aris-ing from transmission and distribution.

23. With the exception of primary reserve, whoseprovision is typically an obligation placed byadministrative means on generators connected tothe system.

24. An important distinction must be made herebetween wind and solar photovoltaic (PV) power.Wind variability occurs over a matter of hours andis relatively amenable to forecasting. Except in cli-mates with cloudless skies, solar PV can vary overseconds and is therefore more difficult to forecast.

25. This account has greatly simplified the complexi-ties of running an electric power system. As well asthe need to match total supply with total demand,the system has a number of other technicalrequirements, including so-called “voltage regula-tion”, and the need to respect transmission con-straints. The latter task in particular is likely tobecome more costly and challenging with theaddition of large amounts of new intermittentgeneration, as discussed in a number of the casestudy chapters in this book.

26. As of February 2010, this is about €5.70 ($7.75)per MWh.

27. This is in contrast to markets where generatorsalso receive payments for being available to gener-ate, via “capacity payment” mechanisms or cap-acity requirements and auctions, as discussed laterin this chapter.

28. Mechanisms allowing the scarcity of operatingreserves to set the price in the energy market arenot in place in most EU markets. Such mecha-nisms require an advanced level of integrationbetween the markets for energy and the marketsfor reserves, as well as between the spot and bal-ancing markets. Market designs allowing suchintegration can be seen mostly in the UnitedStates. See, e.g., Kranz et al. 2003.

29. For example, the markets in Alberta and Ontariohave price caps of C$1,000 ($979) and C$2,000($1,958) per MWh (Adib et al, 2008), and Texas(ERCOT) has a price caps of $2,250 (ERCOT2008). In Europe, Nordpool caps the day-ahead

66 Boaz Moselle

price at €2,000 ($2,720) per MWh (Nordpool2008). As discussed in Chapter 15 of this book,Spain has a very low cap of €180 ($245) perMWh, but it is not an energy-only market, asgenerators also receive capacity payments.

30. Notably the Large Combustion Plant Directive(Directive 2001/80/EC) and the Industrial Emis-sions Directive (still under negotiation in theEuropean Parliament at the time of this writing).

ReferencesAdib, P., E. Schubert, and S. Oren. 2008. Resource

Adequacy: Alternate Perspectives and DivergentPaths. In Competitive Electricity Markets: Design, Imple-mentation, Performance, edited by F. P. Sioshansi.Oxford, UK: Elsevier, 327–362

BBC. 2006. Ukraine Gas Row Hits EU Supplies. January1.

Boyle, Godfrey, ed. 2004. Renewable Energy: Power for aSustainable Future. 2nd ed. Oxford: Oxford Univer-sity Press/Open University.

———. 2007. Renewable Electricity and the Grid: theChallenge ofVariability, London: Earthscan

BP. 2009. Statistical Review ofWorld Energy. London: BP.June.

Bushnell, James. 2005. Electricity Resource Adequacy:Matching Policies and Goals. CSEM Working paper146. www.ucei.berkeley.edu/PDF/csemwp146.pdf(accessed February 26, 2010).

Capros, P., L. Mantzos,V. Papandreou, N.Tasios. 2008.Model-Based Analysis of the 2008 EU Policy Pack-age on Climate Change and Renewables. E3MLab/NTUA. ec.europa.eu/environment/climat/pdf/climat_action/analysis.pdf (accessed February 26,2010).

Chow, Edward, and Jonathan Elkind. 2009. Where EastMeets West: European Gas and Ukrainian Reality.Washington Quarterly 32 (1) (January): 77–92.

CNE (National Energy Commission). 2010.Información Estadística sobre las Ventas de Energíadel Régimen Especial (Statistical Information onSpecial Regime Sales, January). www.cne.es/cne/Publicaciones?id_nodo=143&accion=1&soloUltimo=si&sIdCat=10&keyword=&auditoria=F(accessed in January 2010).

Cramton, Paul and Steven Stoft. 2005. A CapacityMarket That Makes Sense. Electricity Journal 18 (7):43–54.

EIA (U.S. Energy Information Administration). 2008.U.S. Crude Oil, Natural Gas, and Natural Gas Liq-

uids Proved Reserves. www.eia.doe.gov/oil_gas/natural_gas/data_publications/crude_oil_natural_gas_reserves/cr.html (accessedFebruary 26, 2010).

ENTSOG (European Network of Transmission SystemOperators for Gas). 2009. European Ten Year Net-work Development Plan 2010–2019.www.entsog.eu/download/ENTSOG_TYNDR_MAIN_23dec2009.pdf(accessed February 26, 2010).

ERCOT (Electric Reliability Council of Texas). 2008.www.ercot.com/news/press_releases/2008/pr_print_1_173287_173287 (accessed January 6,2010).

Euratom. 2008. Annual Report. Luxembourg: EuropeanUnion.

European Commission. 2000. Towards a EuropeanStrategy for the Security of Energy Supply.COM/2000/0769 final. http://europa.eu/legislation_summaries/energy/external_dimension_enlargement/l27037_en.htm(accessed February 26, 2010).

———. (2007). European Energy and Transport Trends to2030 – Update 2007. Luxembourg: European Com-mission.

———. 2008a. EU Security and Solidarity ActionPlan. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2008:0781:FIN:EN:HTML(accessed February 26, 2010).

———. 2008b. Second Strategic Energy Review:Europe’s Current and Future Energy Position.http://europa.eu/energy/strategies/2008/2008_11_ser2_en.htm (accessed February 26, 2010).

———. 2009a. The January 2009 Gas Supply Disrup-tion to the EU: An Assessment. Commission staffworking document. http://ec.europa.eu/energy/strategies/2009/doc/2009_ser2_documents_travail_service_part1_ver2.pdf(accessed February 26, 2010).

———. 2009b. Proposal for Regulation to SafeguardSecurity of Gas Supply. http://ec.europa.eu/energy/security/gas/new_proposals_en.htm (accessedFebruary 26, 2010).

———. 2009c. Proposal for Regulation to SafeguardSecurity of Gas Supply and Repeal Directive2004/67/EC. http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=CELEX:52009PC0363:EN:NOT(accessed February 26, 2010).

Geropoulos, Kostis. 2007. Druzhba: The Not-So-Friendly Russian Oil Pipeline. www.neurope.eu/

Renewable Generation and Security of Supply 67

articles/Druzhba-The-notsofriendly-Russian-oil-pipeline/73863.php (accessed February 25, 2010).

Gross, Robert, Philip Heptonstall, Dennis Anderson,Tim Green, Matthew Leach, and Jim Skea. 2006.The Costs and Impacts of Intermittency:An Assessment ofthe Evidence on the Costs and Impacts of IntermittentGeneration on the British Electricity Network. A reportof the Technology and Policy Assessment Functionof the UK Energy Research Centre, with financialsupport from the CarbonTrust. London: UK EnergyResearch Centre.

Harvey, Scott. 2005. ICAP Systems in the Northeast:Trends and Lessons, California Independent SystemOperator, September 19.

Hogan, William. 2005. On an “Energy Only” ElectricityMarket Design for Resource Adequacy. Cambridge, MA:John F. Kennedy School of Government, HarvardUniversity.

Holditch, Stephen. 2007. NPC Global Oil and GasStudy. Topic Paper 29. National Petroleum Council,Washington. www.npchardtruthsreport.org/topic_papers.php (accessed February 26, 2010).

IEA (International Energy Agency). 2009.World EnergyOutlook 2009. Paris: IEA.

Joskow, Paul. 2007. Competitive Electricity Marketsand Investment in New Generating Capacity. In TheNew Energy Paradigm, edited by Dieter Helm.Oxford: Oxford University Press, 76–122.

Kranz, Bradley, Robert Pike and Eric Hirst. 2003.Intergrated Electricity Markets in NewYork. Electric-ity Journal 15 (2): 64–65.

Nordpool. 2008. Adjustment of Elspot Technical PriceCeiling in NOK and SEK. No. 98/2008.www.nordpoolspot.com/Market_Information/Exchange-information/No9808-Adjustment-of-Elspot-technical-price-ceiling-in-NOK-and-SEK/(accessed February 19, 2010).

Ofgem. 2006. Emergency Arrangements—Open Let-ter. www.ofgem.gov.uk/Pages/MoreInformation.aspx?docid=5&refer=

MARKETS/WHLMKTS/COMPANDEFF/GASEMERG (accessed February 26, 2010).

———. 2010. Project Discovery: Options for DeliverySecure and Sustainable Energy Supplies. London:Ofgem.

Perekhodtsev, Dmitri and Seth Blumsack. 2009.Wholesale Electricity Markets and Generators’Incentives: An International Review. In InternationalHandbook on the Economics of Energy, edited by JoanneHunt and Lester Evans. Northampton, MA: EdwardElgar, 722–768.

Potential Gas Committee. 2009. Potential Gas Com-mittee Reports Unprecedented Increase in Magni-tude of U.S. Natural Gas Resource Base. ColoradoSchool of Mines. www.mines.edu/Potential-Gas-Committee-reports-unprecedented-increase-in-magnitude-of-U.S.-natural-gas-resource-base(accessed February 26, 2010).

Sensfuß, Frank, Mario Ragwitz, and MassimoGenoese. 2008. The Merit-Order Effect: A DetailedAnalysis of the Price Effect of Renewable ElectricityGeneration on Spot Market Prices in Germany.Energy Policy 36: 3086–94.

Stern, Jonathan. 2005. The Future of Russian Gas andGazprom. Oxford: Oxford Institute of Energy Stud-ies.

Stoft, Steven. 2002. Power System Economics: DesigningMarkets for Electricity. NewYork: IEEE/Wiley.

Trade and Industry Committee, House of Commons.2005. Security of Gas Supply: First Report of Session2005–06. Vol. 2, Written Evidence. London: TSO(The Stationery Office).

Tversky, A. and D. Kahneman. 1992. ‘Advance in Pros-pect Theory – Cumulative Representations ofUncertainty’, Journal of Risk and Uncertainty, 5: 297–323.

World Audit. 2010. Freedom House Annual SurveyPolitical Rights Checklist. www.worldaudit.org/polrights.htm (accessed March 28, 2010).

Yenikeyeff, Shamil Midkhatovich. 2006. The G8 andRussia: Security of Supply vs. Security of Demand?Oxford: Oxford Institute for Energy Studies.

68 Boaz Moselle

5

Market Failure and the Structureof ExternalitiesKenneth Gillingham and James Sweeney

Policy interest in renewable energy technolo-gies has been gathering momentum for the

past several decades, and increased incentives andfunding for renewable energy are often describedas the panacea for a variety of issues ranging fromenvironmental quality to national security togreen job creation. Sizable policies and programshave been implemented worldwide to encouragea transition from fossil-based electricity genera-tion to renewable electricity generation, and inparticular to fledgling green technologies such aswind, solar, and biofuels.

The United States has a long history of policyactivity in promoting renewables, including state-level programs, such as the California Solar Initia-tive, which provides rebates for solar photovoltaicpurchases, as well as federal programs, such as taxincentives for wind. Even in the recent stimuluspackage, the American Recovery and Reinvest-ment Act of 2009, $6 billion was allocated forrenewable energy and electric transmission tech-nology loan guarantees (U.S. Congress 2009).(See Chapter 11 for further discussion of the U.S.experience.) Moreover, such policies are notrestricted to the developed world. For example,China promulgated a National RenewableEnergy Law in 2005 that provides tax and otherincentives for renewable energy and has suc-ceeded in creating a burgeoning wind industry(Cherni and Kentish 2007).

Advocates of strong policy incentives forrenewable energy in the United States use a vari-ety of arguments to justify policy action, such asending the “addiction” to foreign oil, addressingglobal climate change, or creating new technolo-gies to increase U.S. competitiveness. However,articulation of these goals leaves open the ques-tion of whether renewable energy policy is themost sensible means to reach these goals, or evenwhether renewable energy policy helps meetthese goals. Furthermore, many different policyinstruments are possible, so one must evaluatewhat makes a particular policy preferable overothers.

Economic theory can provide guidance andmore rigorous motivation for renewable energypolicy, relying on analysis of the ways privatelyoptimal choices deviate from economically effi-cient choices. These deviations are described asmarket failures and, in some cases, behavioral fail-ures.1 Economic theory indicates that policymeasures to mitigate these deviations can improvenet social welfare, as long as the cost of imple-menting the policy is less than the gains if thedeviations can be successfully mitigated.

Under this perspective, policy analysisinvolves identifying market failures and choosingappropriate policy instruments for each. While analmost unlimited number of different possible

policy instruments can be envisioned, an analysisof relevant market failures allows us to identifywhich instruments are most likely to improveeconomic efficiency.This endeavor is complicatedby the complexity of some market failures, whichmay vary intertemporally or geographically.

This chapter explores these issues in the con-text of renewable energy, with a particular focuson renewable energy used for electricity genera-tion. It first sets the stage with a brief backgroundon the fundamental issues inherent in renewableenergy. Next, it elaborates on the concepts ofcompetitive markets and resource use, and howthe deviations found in reality from the assump-tions of perfect markets may result in market fail-ures. This leads naturally to articulating the classesof possible deviations from perfect markets. A dis-cussion follows of the use of policy instruments tohelp mitigate or correct for these market failures,with a particular focus on how the structure of thefailure influences the appropriate policy approach.

Fundamental Issues inRenewable EnergyRenewable energy, including wind, solar, hydro,geothermal, wave, and tidal, offers the possibilityof a large, continuous supply of energy in perpe-tuity. Analysis of the natural energy flows in theworld shows that they provide usable energymany orders of magnitude greater than the entirehuman use of energy (Hermann 2006). Forexample, the amount of sunlight reaching theearth is more than 10,000 times greater than thetotal human direct use of energy, and the amountof energy embodied in wind is at least 4 timesgreater (Archer and Jacobson 2005; Da Rosa2005; EIA 2008). In principle, renewable energyoffers the possibility of a virtually unlimited sup-ply of energy forever.

In contrast, most of the energy sources we relyon heavily today, such as oil, natural gas, coal, anduranium, are depletable resources that are presenton the earth as finite stocks. As such, eventuallythese stocks will be extracted to the point thatthey will not be economical to use, because of

either the availability of a substitute energy sourceor scarcity of the resource. The greater the rate ofuse relative to the size of the resource stock, theshorter the time until this ultimate depletion canbe expected.

These simple facts about the nature ofdepletable and renewable resources point to aseemingly obvious conclusion: the United Statesand the rest of the world will eventually have tomake a transition to alternative or renewablesources of energy. However, the knowledge thatthe world will ultimately transition back torenewable resources is not sufficient reason forpolicies to promote those resources. Such transi-tions will happen regardless of policy, simply as aresult of market incentives.

The fundamental question is whether marketswill lead the United States and the rest of theworld to make these transitions at the appropriatespeed and to the appropriate renewable resourceconversions, when viewed from a social perspec-tive. If not, then the question becomes, why not?And if markets will not motivate transitions at theappropriate speed or to the appropriate renewablesupplies, the question becomes whether policyinterventions can address these market failures soas to make the transitions closer to the sociallyoptimal.

The question of why not may seem clear tothose who follow the policy debates. Environ-mental and national security concerns are fore-most on the list of rationales for speeding up thetransition from depletable fossil fuels to renewableenergy. Recently there have also been claims thatpromoting new renewable technologies couldallow the United States, or any country, tobecome more competitive on world markets orcould create jobs.

But much national debate often combinesthese rationales and fails to differentiate amongthe various policy options, renewable technolo-gies, and time patterns of impacts. The rest of thechapter explores these issues in greater detail inorder to disentangle and clarify the arguments forrenewable energy policy.

70 Kenneth Gillingham and James Sweeney

Resource Use and Deviationsfrom Perfectly FunctioningMarketsWelfare economic theory provides a frameworkfor evaluating policies to speed the transition torenewable energy. A well-established result fromwelfare economic theory is that absent market orbehavioral failures, the unfettered market out-come is economically efficient.2 Market failurescan be defined as deviations from perfect marketsdue to some element of the functioning of themarket structure, whereas behavioral failures aresystematic departures of human choice from thechoice that would be theoretically optimal.3

A key result for analysis of renewable energy isthat if the underlying assumptions hold, then thedecentralized market decisions would lead to aneconomically efficient use of both depletable andrenewable resources at any given time. Moreover,the socially optimal rate of transition fromdepletable energy supplies to renewable energycan be achieved as a result of decentralized marketdecisions, under the standard assumptions thatrational expectations of future prices guide thedecisions of both consumers and firms (Heal1993).

Although markets are not perfect, the conceptof perfectly competitive markets provides abenchmark for evaluation of actual markets. Iden-tification of market imperfections allows us toevaluate how actual markets deviate from the idealcompetitive markets and thus from the economi-cally efficient markets. Hence with economic effi-ciency as a policy goal, we can motivate policyaction based on deviations from perfectly com-petitive markets—as long as the cost of imple-menting the policy is less than the benefits fromcorrecting the deviation.4

For renewable energy, market failures aremore relevant than behavioral failures, as mostenergy investment decisions are made by firmsrather than individuals, so some of the keydecisionmaking biases pointed out in thebehavioral economics literature are likely to playless of a role. However, behavioral failures mayinfluence consumer choice for distributed genera-

tion renewable energy (e.g., residential solar pho-tovoltaic investments) and energy efficiency deci-sions.5 These could imply an underuse of distrib-uted generation renewable energy—or an overuseof all energy sources (including renewables) ifenergy efficiency is underprovided.

Both market failures and behavioral failurescan be distinguished from market barriers, whichcan be defined as any disincentives to the use oradoption of a good (Jaffe et al. 2004). Market bar-riers include market failures and behavioral fail-ures, but they also may include a variety of otherdisincentives. For example, high technology costsfor renewable energy technologies can bedescribed as a market barrier but may not be amarket failure or behavioral failure. Importantly,only market barriers that are also market orbehavioral failures provide a rationale based oneconomic efficiency for market interventions.

Similarly, pecuniary externalities may occur inthe renewable energy setting and also do not leadto economic inefficiency. A pecuniary externalityis a cost or benefit imposed by one party onanother party that operates through the changingof prices, rather than real resource effects. Forinstance, if food prices increase because ofincreased demand for biofuels, this could reducethe welfare of food purchasers. However, the foodgrowers and processors may be better off. In thissense, pecuniary externalities may lead to wealthredistribution but do not affect economic effi-ciency.

Nature of Deviations fromPerfectly Functioning MarketsIt is a useful to consider deviations from perfectlyfunctioning markets based on whether the marketfailure is atemporal or intertemporal.

Atemporal deviations are those for which theexternality consequences are based primarily onthe rate of flow of the externality. For example, anexternality associated with air emissions maydepend primarily on the rate at which the emis-sions are released into the atmosphere over aperiod of hours, days, weeks, or months. Such

Market Failure and the Structure of Externalities 71

externalities can be described statically. They maychange over time, but the deviation has economicconsequences that depend primarily on theamount of emissions released over a short timeperiod (e.g., hours, days, weeks, or months).These may have consequences that are immediateor occur over very long time periods.

Intertemporal deviations are those for whichthe externality consequences are based primarilyon a stock that changes over time depending onthe flow of the externality. The flows lead to achange in the stock over a relatively long periodof time, typically measured in years, decades, orcenturies. The stock can be of a pollutant (e.g.,carbon dioxide) or of something economic (e.g.,the stock of knowledge or of photovoltaicsinstalled on buildings). If the flow of the external-ity is larger (smaller) than the natural decline rateof the stock, the stock increases (decreases) overtime. Intertemporal externalities can best bedescribed dynamically, for it is the stock (e.g., car-bon dioxide), rather than the flow, that leads tothe consequences (e.g., global climate change).

For some environmental pollutants (e.g.,smog), the natural decline of the stock is rapid—perhaps over the course of hours, days, weeks, ormonths. For these pollutants, the stock leads tothe damages, and the stock is entirely determinedby the flow over this short time frame. These canbe treated as atemporal deviations, as the dynamicnature of the externality is less important withsuch a rapid natural decline rate.

For atemporal externalities, the appropriatemagnitude of the intervention depends primarilyon current conditions. Thus, because conditionscan change over time, the appropriate magnitudecould increase, decrease, or stay constant overtime. For intertemporal externalities, the appro-priate magnitude of the intervention dependsmore on the conditions prevailing over manyfuture years than on current conditions or those atone time. As time passes, the appropriate magni-tude of the intervention changes but, more pre-dictably, based on the stock adjustment process.Therefore, the appropriate price or magnitude ofthe intervention will have a somewhat predictabletime pattern.

Atemporal (Flow-Based) Deviations fromEconomic Efficiency

Atemporal deviations from economic efficiencyfall into several categories: labor market supply–demand imbalances, environmental externalities,national security externalities, information mar-ket failures, regulatory failures, market power,too-high discount rates for private decisions,imperfect foresight, and economies of scale.

Labor Market Supply–Demand Imbalances

Unemployment represents a situation in whichthe supply of labor exceeds demand at the prevail-ing wage structure, perhaps because of legal andinstitutional frictions slowing the adjustment ofthe wage structure. In the United States, suchunemployment does not occur very often, typi-cally only during recessions. At times of fullemployment,6 abstracting from the distortionaryimpacts of income or labor taxes,7 the social costof labor (i.e., the opportunity cost and other costsof that labor to the employee) would be equal tothe price of labor (i.e., the wage an employer mustpay for additional labor), and hence there is noroom to improve economic efficiency throughgreen jobs programs.

With unemployment, however, the price oflabor exceeds the social cost of that labor. Thisdifference represents a potential net economicefficiency gain, and thus any activity that employsadditional workers may improve economic effi-ciency. For example, if an additional amount ofsome economic activity produced no net profit,and therefore would not be privately undertaken,the net social economic gain would be equal tothe differential between the price of labor and itssocial cost.

With unemployment, the opportunity cost(and other cost) of labor to the person beingemployed could be expected to vary substantiallyacross individuals. Some unemployed persons mayuse their free time productively to perform workat home or improve skills, so that the opportunitycost of labor might be only slightly below thewage. Others may not be able to make such pro-ductive use of their time, so that the opportunity

72 Kenneth Gillingham and James Sweeney

cost might be virtually zero, significantly belowthe wage. Thus the potential net social gain fromadditional employment could range from nearlythe entire wage to zero.

Little evidence exists to suggest that additionalemployment in renewable energy can providelarger net social gains than any other industry,including the fossil-fuel industry. Moreover, suchgains must be seen as transient possibilities in aneconomy such as that of the United States, whichregularly is near full employment.

Environmental Externalities

Environmental externalities are the underlyingmotivation for much of the interest in renewableenergy. The discussion here focuses on generalissues in environmental externalities; specificissues inherent in intertemporal environmentalexternalities are addressed below in the sectiontitled “Stock-Based Environmental Externalities”.Combustion of fossil fuels emits a variety of airpollutants, which are not priced without a policyintervention. Air pollutants from fossil-fuel com-bustion include nitrogen oxides, sulfur dioxide,particulates, and carbon dioxide. Some of thesepollutants present a health hazard, either directly,as in the case of particulates, or indirectly, as in thecase of ground-level ozone formed from high lev-els of nitrogen oxides and other chemicals.

When harmful fossil-fuel emissions are notpriced, the unregulated market will overuse fossilfuels and underuse substitutes, such as renewableenergy resources. Similarly, if the emissions arenot priced, firms will have no incentive to findtechnologies or processes to reduce the emissionsor mitigate the external costs. The evidence forenvironmental externalities from fossil-fuel emis-sions is strong, even if estimating the precise mag-nitude of the externality for any given pollutantmay not be trivial.

In some cases, there may also be significantenvironmental externalities from renewableenergy production, such as hydroelectric facilitiesthat produce methane and carbon dioxide emis-sions from submerged vegetation, or greenhousegas emissions and nitrogen fertilizer runoff fromthe production of ethanol biofuels. In many other

cases, these environmental externalities are rela-tively small. Whether renewable energy resourcesare underused or overused relative to economi-cally efficient levels depends on which of the twoenvironmental externalities is greater: those fromfossil fuels or from the renewable energyresources. In most, but by no means all, cases, theexternalities from the fossil fuels are greater,implying that the market will underproviderenewable energy.

Unpriced environmental externalities fromeither fossil fuel or renewable energy use wouldimply either an overuse of energy in general or anunderuse of potential energy efficiency improve-ments.

National Security Externalities

Oil production around the world is highly geo-graphically concentrated, with the bulk of the oilreserves in the hands of national oil companies inunstable regions or countries of the world, such asthe Middle East, Nigeria, Russia, and Venezuela.Oil-importing countries, such as the UnitedStates, European nations, and China, have seenlarge security risks associated with these oilimports. In response, they have laid out substantialdiplomatic and military expenditures in theseregions, at least partly in order to ensure a steadysupply of oil. If increases in oil use lead to addi-tional security risks, these risks represent an exter-nality associated with oil use. Moreover, if theadditional security risks are met with increases indiplomatic and military expenditures, then theseadded expenditures can be used as an approximatemonetary measure of the externalities.

However, it appears unlikely that a modestincrease or decrease in oil demand will influencethese expenditures due to the lumpiness of theexpenditures, even though the increases in oil usecould lead to additional security risks. Conversely,long-term large changes in oil demand mayreduce national security risks and the correspond-ing military and diplomatic expenditures.

In many countries around the world, such asthose in Europe, the use of natural gas may havenational security externalities because of similarissues. Quantifying the national security exter-

Market Failure and the Structure of Externalities 73

nalities associated with oil or natural gas con-sumption is more fraught with difficulties thandoing so with environmental externalities, yetsome analysts have suggested that the magnitudemay be substantial (Bohi and Toman 1996). Oth-ers are more sanguine and believe that globalenergy markets can substantially buffer nationalsecurity risks.

In the U.S. context, natural gas and somerenewable energy resources, such as biofuels, aresubstitutes for oil with few or no energy securityexternalities and thus would be underused relativeto the economically efficient level. Improving theenergy efficiency of vehicles and furnaces is also asubstitute for oil and would also be underused.Most renewable energy resources produce elec-tricity, so until electric vehicles are a viable large-scale substitute for conventional vehicles fueled byrefined oil products, national security externalitiesapply only indirectly to such renewable energyresources. However, these national security exter-nalities, although indirect, can be important. Forexample, the production of electricity fromrenewables could lead to reductions in natural gasused for electricity production. This reductionwould lead to more availability of natural gas forother purposes, such as heating, where it couldsubstitute for oil in some locations. For biofuels,national security externalities are of foremostconsideration. Moreover, in the European con-text, renewable energy directly substitutes fornatural gas.

Information Market Failures

Information market failures relate most directly tothe adoption of distributed generation renewableenergy by households, such as solar photovoltaicsystems or microgeneration wind turbines. Ifhouseholds have limited information about theeffectiveness and benefits of distributed genera-tion renewable energy, an information marketfailure may occur. In a perfectly functioning mar-ket, one would expect profit-maximizing firms toundertake marketing campaigns to inform poten-tial customers. However, for nascent technologiesthat are just beginning to diffuse into the market,economic theory suggests that additional infor-

mation can play an important role (Young 2010).Information market failures are closely related tobehavioral failures. Reducing information marketfailures would also be expected to reducebehavioral failures associated with heuristicdecisionmaking.

Imperfect foresight by either firms or con-sumers (or investors in the stock market whoinfluence firms) suggests an inability to predictfuture conditions accurately, which may lead to anunderestimate or overestimate of how energyprices may rise in the future. If firms systemati-cally under- or overestimate future energy prices,then there may be an underinvestment oroverinvestment in research and development(R&D) for renewable energy technologies relativeto the economically efficient level.

Although it certainly seems plausible thatfirms have imperfect foresight, it is less plausible tobelieve that this imperfect foresight will systemati-cally lead to an underestimate of future energyprices, rather than random deviations that aresometimes underestimates and other times over-estimates. Even if firms have imperfect foresight,as long as the firms’ estimates of future prices arenot systematically biased, then on average invest-ment in renewable energy technologies wouldstill follow the economically efficient path. In thissituation, errors leading to overinvestment wouldbe balanced by those leading to underinvestment.At present, there is little evidence either for oragainst the hypothesis that firms systematicallyunderestimate future price increases.

Another information market failure is theclassic principal-agent or split-incentive problem,which may influence renewable energy adoptionin two ways. First, in many cases for rental prop-erties, landlords make the decision about whetherto invest in distributed generation renewableenergy, while tenants pay the energy bills (Jaffeand Stavins 1994; Murtishaw and Sathaye 2006).Second, if landlords are not compensated for theirinvestment decisions with higher rents, then theywould tend to underinvest in distributed genera-tion renewable energy. This market failure hasbeen most carefully examined in the context ofenergy efficiency and heating (e.g., see Levinsonand Niemann 2004), but the extent to which this

74 Kenneth Gillingham and James Sweeney

market failure is important for renewable energyhas not yet been empirically examined.

Finally, there may be a principal-agent prob-lem relating to managerial incentives. In manycases, managers have their compensation tied tothe current stock price, rather than the long-termperformance of the company (Rappaport 1978).However, investors may have difficulty distin-guishing between managerial decisions that boostshort-term profits at the expense of long-termprofits and those that boost both short- and long-term profits. In the context of renewable energy,the emphasis on short-term performance maylead to underinvestment in R&D for renewableenergy technologies, for the benefits of develop-ing such technologies are likely to be receivedover the long term, while the costs are borne inthe short term. Of course, this issue may occur inany industry and is not unique to renewableenergy resources.

Regulatory Failures

In some cases, the regulatory structure can createperverse incentives. For example, average costpricing of electricity implies that consumers oftenface a price of electricity that does not reflect themarginal cost of providing electricity at any giventime. This may influence the adoption of distrib-uted generation renewables, such as residentialsolar photovoltaic (PV) systems. In many loca-tions, electricity output from a solar PV unit tendsto be higher during the day, corresponding totimes of high electricity demand. To the extentthat the solar PV output is correlated with highwholesale electricity prices, consumers and firmsdeciding whether to install a new solar PV unitwill undervalue solar PV absent tariffs thataccount for the time variation. Borenstein (2008)quantifies this effect in California, finding thatsolar is currently undervalued by 0% to 20%under the current regulatory framework, and thatthis could rise to 30% to 50% if the electricitysystem were managed with more reliance onprice-responsive demand and peaking prices,because solar output would be concentrated attimes with even higher value.

Too-High Discount Rates

In some cases, the discount rate for private invest-ment decisions may be higher than the social dis-count rate for investments with a similar risk pro-file. For example, the corporate income tax dis-torts incentives for firms to invest, effectivelyimplying that they require a higher rate of returnon investments than they would otherwise. Alter-natively, credit limitations may also occasionallylead to a higher rate of return required for invest-ments. These credit limitations may be due tomacroeconomic problems, such as the recentliquidity crisis in the United States, or individuallimitations on the firm involved in the renewableenergy investment. Individual credit limitationsmay also apply in cases where consumers areinterested in installing distributed or off-grid gen-eration.

Discount rates that are too high may lead totwo effects. First, if firms investing in renewableenergy technologies have distorted discount rates,this could lead to underinvestment in renewableenergy resources relative to the economically effi-cient level. Second, if discount rates are too highfor firms extracting depletable resources, such asfossil fuels, then the fuels are extracted too rapidly,leading to prices that are lower than economicallyefficient. Because the depletable resource wouldbe depleted too rapidly, the transition to renew-able energy technologies may then be hastenedrelative to the efficient transition. However,investment in renewables may be second best, inthat it would still be optimal to invest more, con-ditional on the too-rapid extraction of depletableresources.

This phenomenon is applicable not only toenergy-related investments, but also to invest-ments throughout the economy. Thus this issueprovides reasons for changing incentives forinvestment throughout the economy, but it doesnot provide a particular reason for shifting invest-ments from other parts of the economy to renew-able energy, unless evidence suggested that highdiscount rates are particularly important forrenewable energy. However, we are aware of noevidence that could give a sense of the magnitudeof this distortion.

Market Failure and the Structure of Externalities 75

Economies of Scale

Economies of scale, particularly increasing returnsto scale, refers to a situation where the averagecost of producing a unit decreases as the rate ofoutput at any given time increases, resulting froma nonconvexity in the production function for anynumber of reasons, including fixed costs. Thisissue may inefficiently result in a zero-outputequilibrium only when we have market-scaleincreasing returns, where the slope of the averagecost function is more negative than the slope ofthe demand function, and the firm cannot over-come the nonconvexity on its own.

Market-scale increasing returns refer to anonconvex production function at output levelscomparable with market demand. Figure 5.1graphically illustrates the second condition. If thequantity produced is small (e.g., quantity a), thenno profit-seeking firm would be willing to pro-duce the product, but if production could beincreased to some level above the crossing point(e.g., at the quantity b), then it would be profit-able for the firm to produce: price would exceedaverage cost.

Usually a firm could overcome the situationin Figure 5.1 on its own simply by selling at a lowprice. Even if this is a risky endeavor, it is notlikely that all firms would ignore this opportunity.However, firms may not be able to take advantage

of the opportunity because of capital constraintsor a simultaneous coordination problem.

Capital constraints may be a problem only ifthe aggregate investment required is extremelylarge; otherwise, it is likely that some firm couldbe expected to raise the necessary capital. Capitalconstraints facing an economy, as occurred in the2008–2009 recession, could limit such capitalinvestments for an entire economy. Because suchevents tend to be transient, however, these con-straints at most could be expected to delay theinvestments.

Often economies of scale are accompanied bya “chicken-and-egg” problem, wherein multipleactors must simultaneously invest and ramp upproduction in order to commercialize a new tech-nology.This may be most relevant in technologiesthat require a new infrastructure, such ashydrogen-fueled vehicles, which may or may notuse renewable energy depending on the hydrogengeneration source. Such possibilities requireinterindustry cooperation and thus may greatlydelay investments. Similar chicken-and-egg prob-lems have been overcome in the past, as with per-sonal computers, operating systems, and applica-tion software or automobiles, gasoline, service sta-tions, and roads, but these problems greatlycomplicate investments.

It should be noted that the equilibrium thatwould occur with market-scale increasing returnswould unlikely be a workable competitive equi-librium, but rather a single-firm monopolisticequilibrium. In fact, the situation of market-scaleincreasing returns is often referred to as a “naturalmonopoly.” This situation raises the possibility ofmarket power.

Market Power

Uncompetitive behavior may influence the adop-tion of renewable energy technologies in severalways. First, market power in substitutes forrenewable energy can influence the provision ofrenewable energy through two channels. Firmseffectively exercising market power in substitutesfor renewable energy (e.g., at times the OPECcartel) would raise the price of energy above theeconomically efficient level, making investment in

Price

Demand

Average cost

Quantityba

Figure 5.1. Economies of scale: slope of averagecost function is more negative than slope ofdemand function

76 Kenneth Gillingham and James Sweeney

renewable energy more profitable and leading toan overinvestment in renewable energy. On theother hand, firms that have market power in sub-stitutes for renewable energy may have an incen-tive to buy out fledgling renewable energy tech-nologies to reduce competitive pressures—leadingto a possible underprovision of renewable energyresources if that purchasing firm “buries” therenewable technology. However, the prospect ofbeing bought by a competitor could provide astrong incentive for a new firm to be created withthe explicit intention of selling itself to a largercompany. Which effect dominates and whetherthere is market power in substitutes for renewableenergy can be determined only empirically.

Market power may also influence the adop-tion of renewable energy resources by influencingthe rate and direction of technological change. Ifless competition exists in a market, firms are morelikely to be able to fully capture the benefits oftheir innovations, so incentives to innovate arehigher (e.g., see Blundell et al. 1999; and Nickell1996). Conversely, if more competition exists,firms may have an incentive to try to “escape”competition by investing in innovations that allowthem to differentiate their product or find a pat-entable product. Some evidence suggests that therelationship between competition and innovationmay be an inverted U-shaped curve, with a posi-tive relationship at low levels of competition and anegative relationship at higher levels of competi-tion (Aghion et al. 2005; Scherer 1967).This rela-tionship likely holds in all industries, not just therenewable energy industry.

Finally, in some cases, vertically integratedutilities may effectively exercise market power byfavoring their own electricity generation facilitiesover other small generation facilities, includingrenewable energy facilities.This was a concern forthe implementation of renewables when utilitiesinvested mostly in nonrenewable energy, but utili-ties now typically invest in renewable energyalong with conventional generation plants.8

Intertemporal (Stock-Based) Deviations

An important intertemporal deviation may occurwith the existence of stock-based environmental

externalities. A second intertemporal deviationmay occur if an imperfect capture of the stock ofknowledge is created as a result of current actions,leading to underinvestment or underproductionof those activities that lead to growth of theknowledge stock. These can occur with knowl-edge generation processes, such as learning bydoing or research and development; market diffu-sion of a new technology; or network externali-ties. Intuitively, when others can capture some ofthe benefits from the choice made by a firm orconsumer, the uncaptured benefits will be sociallyvaluable but will not be taken into account by thefirm or consumer.

Stock-Based Environmental Externalities

As discussed above, some environmental exter-nalities have consequences based on the stock ofthe pollutant, rather than the flow, and the stockadjusts only slowly over time. For such environ-mental externalities, the intertemporal nature ofthe damages from the stock imposes additionalstructure on the time pattern of deviations.

Particularly relevant to renewable energy sup-plies are carbon dioxide and other greenhousegases. For CO2, every additional metric ton emit-ted remains in the stock for more than a century.Thus emitting a ton today would have roughly thesame cumulative impacts as emitting a ton in 20years. This implies that, absent changes in theregulatory environment, the magnitude of thedeviation for emissions now will be the same asthe magnitude of the deviation for emissions 20years from now. Economic efficiency implies thata society should be almost indifferent betweenemitting a ton of CO2 now, 20 years from now, orany year in between .9 As will be discussed belowin the section titled “Policies for Stock-BasedEnvironmental Externalities: Carbon Dioxide”, itis this relationship that imposes a structure on thetime pattern of efficient policy responses.

Similar issues arise for toxic metals releasedinto the waterways, radioactive nuclear waste,mercury in waterways and oceans, sequestrationof carbon dioxide in the deep oceans, and rainfor-est land degradation.

Market Failure and the Structure of Externalities 77

Imperfect Capture of Future Payoffs fromCurrent Actions

R&DWhen firms invest in increasing the stock ofknowledge by spending funds on R&D, they maynot be able to perfectly capture all of the knowl-edge gained from their investment. For example,successful R&D (e.g., creating a new class of solarphotovoltaic cells) by a particular firm could beexpected to result in some of the new knowledgebeing broadly shared, through trade magazines,reverse engineering by its competitors, or techni-cal knowledge employees bring with them as theychange employment among competitive firms. Inaddition, patent protection for new inventionsand innovations has a limited time frame (20 yearsin the United States), so after the patent lapses,other firms may also benefit directly from theinvention or innovation.

Fundamentally, R&D spillovers can bethought of as an issue of imperfect property rightsin the stock of knowledge: other firms can sharethat stock without compensating the original firmthat enhanced the knowledge stock.To the extentthat those spillover benefits occur, the social rateof return from investment in R&D is greater thanthe firm’s private rate of return from investment inR&D. Indeed, although estimates differ by sector,there appears to be substantial empirical evidencethat the social rate of return is several times that ofthe private rate of return. For example, in theUnited States, the social rate of return is estimatedin the range of 30% to 70% per year, while theprivate rate of return is 6% to 15% per year(Nordhaus 2002). However, the magnitude of theR&D spillovers depend on the stage in the devel-opment of a new technology, with more funda-mental research having significantly greater R&Dspillovers than later-stage commercializationresearch (Nordhaus 2009).

Evidence of high social returns to R&D isfound not just in the renewable energy sector, butthroughout the economy.Thus, to the extent thatsome R&D in renewable energy technologiescomes at the expense of R&D in other sectorswith a high social rate of return, the opportunity

cost of renewable energy R&D may be quite high(Pizer and Popp 2008). Empirical work suggeststhat additional R&D investment in renewableenergy will at least partly displace R&D in othersectors. Popp (2006) finds that approximately halfof the energy R&D spending in the 1970s and1980s displaced, or crowded out, R&D in othersectors. Part of the rationale for this may be thatyears of training are required to become a compe-tent research scientist or engineer, and thereforethe supply of research scientists and engineers is,at least in the short term, relatively inelastic. In thelonger term, crowding out is less likely to be anissue, as universities train more scientists and engi-neers.

Learning by doingA similar intertemporal market imperfection dueto a knowledge stock spillover may also occur ifthere is a significant learning-by-doing (LBD)effect that cannot be captured by the firm. LBDhas a long history in economics, dating back toArrow (1962). The basic idea behind LBD is thatthe cost of producing a good declines with thecumulative production of the good, correspond-ing to the firm “learning” about how to producethe good better.10 One interpretation is that withLBD, the cost is dependent on the stock ofknowledge, which is proxied by the stock ofcumulative past production. In the standard modelof LBD, the firm today bears the up-front cost ofproducing an additional unit and thereby alsoincreasing the knowledge stock, while all firms inthe industry benefit from the increased stock ofknowledge, leading to reduced costs in the futurefor all firms—an intertemporal spillover.

Importantly, LBD alone does not necessarilyimply a market failure. In some situations, onecould imagine that all knowledge leading to costreductions could be used only by the single firmmaking the decision. In this special case, there areno spillovers, and the firm would have the incen-tive to produce optimally, weighing the up-frontcost of learning against the benefits of the costreductions in the future as it would any invest-ment decision.11

Outside of this special case, the existence ofLBD can represent an externality with the poten-

78 Kenneth Gillingham and James Sweeney

tial to be an important market imperfection inrenewable energy markets. There is little or noempirical evidence on the degree of spilloversfrom LBD, but ample evidence exists that the costof several important renewable energy technolo-gies tends to decline as cumulative productionincreases (Jamasb 2007). This evidence alone doesnot prove the existence of a market failure, forother factors may also be able to explain the costdecreases (e.g., R&D or even time-dependentautonomous cost decreases).

The magnitude of a LBD market failure isspecific to each technology, and each has to beassessed on a case-by-case basis. Moreover, muchlike R&D spillovers, LBD spillovers are notunique to renewable energy technologies, butmay also be present in any number of fledglingtechnologies as they diffuse into the market.Hence both R&D and LBD spillovers can beconsidered broader innovation market failures thatlead to underinvestment in or underproduction ofcertain renewable energy resources.

Network externalitiesNetwork externalities occur when the utility anindividual user derives from a product increaseswith the number of other users of that product.The externality stems from the spillover one user’sconsumption of the product has on others, so thatthe magnitude of the externality is a function ofthe total number of adoptions of the product.Often quoted examples of network externalitiesinclude the introduction of the “QWERTY”typewriter keyboard, telephone, and fax machine(David 1985).

An important caveat about network externali-ties is that the externality may already be internal-ized. For example, the owner of the network mayrecognize the network effects and take them intoaccount in his or her decisionmaking (Liebowitzand Margolis 1994). Alternatively, in some cases,the recipients of the network spillover may be ableto compensate the provider. For example, for net-work effects in home computer adoption, thenew adopter might take the previous adopter tolunch as thanks for teaching him or her how touse the computer (Goolsbee and Klenow 2002).When the externality is already internalized, net-

work externalities are more appropriately titled“network effects” or “peer effects” and do notlead to market failures (Liebowitz and Margolis1994).

In the context of renewable energy, networkexternalities may play a role in the adoption ofdistributed generation. This may come about ifconsumers believe that installing renewableenergy systems on their homes sends a message totheir neighbors that they are environmentallyconscious—and that more installations in theneighborhood will increase this “image motiva-tion” or “snob effect.” Evidence for this effect hasbeen shown in Sacramento for solar panels(Lessem and Vaughn 2009). Little evidence isavailable to indicate whether this is truly a net-work externality or just network effects in distrib-uted generation renewable energy.

Policy InstrumentsEach of the failures described above can providemotivation for policy to correct the failure, but itis not always a simple task to appropriately matchthe policy to the failure.Table 5.1 lists some of themore common classes of policy instruments avail-able to address failures relevant to renewableenergy. This table is meant to be illustrative, asthere exist an almost uncountable variety of dif-ferent policy instruments.

How do we choose among the policy instru-ments? Economic theory along with carefulanalysis can provide guidance. First, both theoryand evidence indicate that multiple market fail-ures will likely require multiple interventions, so asensible policy goal involves matching the mostappropriate intervention to the failure (Aldy et al.2009; Goulder and Schneider 1999). In somecases, several policy instruments can address, orpartly address, a given market failure. In thesecases, if economic efficiency is the goal, the com-bination of policy instruments that provides thegreatest net benefits should be chosen. In addi-tion, many of the market failures relevant torenewable energy are broader market failures thatmay apply to a wide range of markets or tech-nologies. Therefore, economic efficiency would

Market Failure and the Structure of Externalities 79

be further enhanced if the interventions to addressthese market failures were not focused solely onrenewable energy.

Several concerns warrant careful attention inthis matching process. First, we care about howeffective the intervention will be at actually cor-recting the market failure. Second, the benefitsfrom the intervention must be weighed againstthe costs of implementing the policy, includingboth government administrative costs and indi-vidual compliance costs—taking into account therisk of poor policy design or implementation. Inaddition, careful consideration of any equity ordistributional consequences of the intervention isimportant, both for ethical reasons and for gainingthe political support for passage of the policy.

Uncertainty about the magnitude of the mar-ket failure and the effectiveness of the interven-tions is another important concern. In some cases,

potential damages from a market failure may belarge enough that the most sensible intervention isdirect regulation, so that we can be certain therisk is mitigated. For example, if a toxin is deemedto have sufficiently high damages with a highenough probability, it may be sensible for the gov-ernment to simply ban it. A comprehensive analy-sis of the costs and benefits of different policyoptions that explicitly includes uncertainty, in thiscase could be expected to reveal the need for sim-ply banning the toxin.

Finally, the temporal structure of the marketfailures may have a profound influence on thetemporal structure of optimal intervention. Eco-nomic theory suggests that not only should anintervention be matched to the failure, but alsothe temporal pattern of the intervention shouldbe matched to the temporal pattern of the failure.For example, the optimal correction for failures

Table 5.1. Some potential policy instruments

Direct regulation Command-and-control methods (e.g., requiring firms to generate electri-city from renewable energy resources)

Direct government-sponsored R&D Government funding for scientists and engineers working on improvingdifferent renewable energy technologies, support for national laboratories,funding research prizes such as “X prizes”

R&D tax incentives Subsidies for private renewable energy technology R&D

Instruments to correct market prices:excise taxes, cap-and-trade, subsidies

“Get prices right” by adding to the cost of goods (e.g., through a tax or apermit price) or reducing the cost of goods (e.g., through a subsidy)

Feed-in tariffs Require electric utilities to purchase electricity from other generators (oftensmall renewable energy generators) at a specified price

Information programs Education campaigns and required labels

Product standards Require firms to improve their product characteristics to meet a specifiedgoal (e.g., efficiency of solar PV cell or energy efficiency of lighting)

Marketable marketwide standards:renewable portfolio standards, low-carbon fuelstandards, corporate average fuel, economystandards

Require firms (e.g., utilities) to meet a specified standard (e.g., produce aspecified amount of electricity from renewables) or purchase permits orcertificates from other firms that overcomply with the standard

Transparency rules Require firms to provide more information about their current conditionsto investors

Macroeconomic policy Fiscal or monetary policies to stabilize the economy and provide liquidity tomarkets to reduce credit constraints

Corporate taxation reform Adjusting the corporate income tax to improve corporate incentives

Competition policy/laws Reduce the exercise of market power through antitrust action

Restructured regulation Reduce regulatory failures and loopholes in regulations that allow for mar-ket power

Intellectual property laws Laws to encourage innovation by allowing innovators to appropriate thebenefits of their work

80 Kenneth Gillingham and James Sweeney

that decrease in magnitude and eventually vanishover time would be a transient intervention.

Table 5.2 summarizes the matching, with thevarious market failures listed as rows and policyinstruments from Table 5.1 as the columns. Forthose instruments that appear to be potentiallywell matched to the market failures, the letters Pand T indicate whether the instrument could beexpected to be permanent or transient. Of course,the particular circumstances of each market failureand the potential policy must be assessed. Somepotential policies may be useful only under lim-ited circumstances, and the evidence for somemarket failures in renewable energy is muchweaker than others. Moreover, some of the policy

options listed may be reasonably well matchedwith a market failure but may be second best toother policy options.

Policy Instruments for Atemporal(Flow-Based) Deviations

Atemporal deviations lend themselves to policyinterventions that vary, perhaps greatly, withchanging external conditions. If the underlyingmarket deviation is a continuing problem, thenthe policy interventions can be expected to have arelatively permanent nature. If the deviation istransient, the appropriate policy interventionwould likewise be transient.

Dre

ct re

gua�

on

Dre

ct g

over

nmen

t-sp

onso

red

R&D

Com

pe��

ons,

suc

h as

X p

rze

R&D

tax

ncen

�ves

Exc

se ta

xes

Prod

uc�o

n su

bsd

es

Feed

-n ta

rffs

nfor

ma�

on p

rogr

ams

Prod

uct s

tand

ards

Cap-

and-

trad

e

Mar

keta

be

mar

ketw

de s

tand

ards

Tran

spar

ency

rues

Mac

roec

onom

c po

cy

Corp

orat

e ta

xa�o

n re

form

Com

pe��

on p

ocy

/aw

s

Rest

ruct

ured

regu

a�on

nte

ectu

a p

rope

rty

awLabor supply/demand imbalances T TEnvironmental externali�es P P P P P P P P P P P PNa�onal security externali�es P P P P P P P P P P P PInforma�on market failures P P PRegulatory failures PToo high discount rates P P PImperfect foresight P Economies of scale T T PMarket power P P PR&D spillovers P P P P PLearning by doing T TNetwork externali�es T P

Table 5.2. Sources of market failure and some illustrative potential policy instruments

Note: P indicates permanent change or instrument; T indicates transient instrument

Market Failure and the Structure of Externalities 81

Policies for Labor MarketSupply–Demand Imbalances

Labor market unemployment in well-functioningdeveloped economies can be expected to be atransient problem, associated with economicrecession. Typically, policies are crafted at thenational or international level and focus oneconomywide transient monetary or fiscal policiesthat are terminated when the economy returns tofull employment.12 However, it is often assertedthat subsidizing new renewable technologies isadvantageous because it would create jobs.

In theory, in order to align private incentiveswith socially optimal incentives, the labor cost ofproviding renewable energy could be subsidizedby the difference between the market price oflabor and the social cost of that labor. Thus inorder to improve economic efficiency, such alabor subsidy must vary sharply over the course ofthe business cycle; be zero during times of fullemployment; and differ across employees,depending on the options facing the unemployedperson. This set of conditions may make itextremely difficult to implement such a policy.

Moreover, and perhaps most important,unemployment is an economywide phenomenon,so an equally valid argument could be made forsubsidizing labor throughout the economy,including in the fossil fuels sector. Thus the “cre-ating jobs” argument does not clearly justify tar-geting the labor subsidies to the renewable energyindustry, unless a particularly large deviationexisted between the social cost of the labor andthe market price relative to the rest of theeconomy.

Policies for Environmental Externalities

Table 5.2 lists a wide variety of different policyinstruments to address environmental externali-ties.The most straightforward of these is to simplyprice the environmental externality, following thetheory first developed by Pigou (Baumol 1972).In doing so, firms and consumers will take intoaccount the externality in their decisions of howmuch to produce and consume. The price couldbe imposed directly as a pollution tax or pollution

fee, with the optimal tax set at the magnitude ofthe externality. Or a cap-and-trade system couldimpose a marketwide limit on the emissions, inwhich case trading of the allowances under a cap-and-trade system would lead to a market-clearingprice for the allowances. The cap should be set sothat the resulting permit price is equal to the mag-nitude of the externality.13 The magnitude of theexternality can be estimated based on damageestimates from scientific and economic literature.

In some cases, the risk from particularly severepollutants (e.g., possibly some criterion air pollut-ants) may be sufficiently high that the marginaldamage associated with the release of the pollut-ant would always exceed the economic costs ofreducing that pollution—implying that directregulation could be an economically efficientpolicy. Direct regulation would entail the govern-ment setting strict limits on the amount of thesevere pollutant emitted or, in some cases, possiblyeven banning emission of the pollutant entirely.

Environmental externalities from renewableenergy can be treated the same way as environ-mental externalities from fossil-fuel combustion.For most renewables, the environmental exter-nalities are small, so the appropriate tax would besmall. A few, such as corn-based ethanol and palmoil biodiesel, may have significant emissions ofsome pollutants, and the damages from theseshould be added to the price of the resource.

A second tax or subsidy approach to address-ing environmental externalities more closely fol-lows the policies in many countries. Rather thanputting a price on both fossil fuel and renewableenergy generation corresponding to the magni-tude of each externality, the same cost differentialcould be maintained by subsidizing low-emittingresources and not subsidizing (or taxing) high-emitting resources. However, this approach wouldhave the unintended consequence of makingenergy use less expensive than its actual socialcost, because the external costs would remainunpriced. An additional subsidy on energy effi-ciency investment can correct for the overuse ofenergy, removing the primary distortion inenergy markets. Unfortunately, this may still leadto a distortion through an overinvestment in thesubsidized energy-efficient technologies, because

82 Kenneth Gillingham and James Sweeney

the optimal choice may have involved moreenergy conservation and less investment inenergy-efficient technologies.14 In addition, sucha combination of subsidies may lead to furtherdistortions in factor markets, such as markets forthe inputs in the production of energy efficiencyequipment. Thus the economic theory suggeststhat the first-best approach to addressing environ-mental external damages is through taxes or per-mits, and the subsidy approach outlined above canbe considered a second-best approach to be pur-sued if the first-best approach is not politicallyfeasible.

Other approaches rely on the idea that if firmsmust clearly disclose their environmental impacts,they will be motivated to reduce those impacts,and consumers will be motivated to shift theirpurchases away from damaging products andtoward those that are environmentally benign.Information programs designed to publicize theenvironmentally damaging product or transpar-ency rules designed to document and communi-cate the environmental damages are prompted bythis idea. Enterprise software available from com-panies such as Hara Software15 have made it pos-sible to document and broadly communicate car-bon dioxide and other environmental impacts in atransparent manner.

Policies for National Security Externalities

Each of the policy instruments available forresponding to environmental externalities is alsoavailable for responding to national security exter-nalities. Again, the first-best policy interventionworks by getting prices right. By pricing theexternal costs imposed by the consumption of thefuel, firm and consumer decisions will take intoaccount the externality, resulting in an economi-cally efficient outcome. In this case, getting theprice right inherently involves taking into accountthe full external effect, such as the externality thatone country’s spending an extra dollar on defensecauses other countries to spend more on defense.With a correct price on the fuel, firms and con-sumers will substitute other energy resources thatdo not lead to national security risks, such as coalor renewable energy, or will find ways of reducingenergy use.

Just as for environmental externalities, a sec-ond approach would be based on maintaining aprice differential between fuels with high and lownational security external costs. This approachwould face the same issues: overuse of fuel withhigh external costs and overuse of energy in gen-eral. Policies to subsidize energy efficiency wouldhelp but may come at the cost of distortionsthrough overinvestment in energy efficiency oroverconsumption in some factor markets.

Other policy instruments may also improveeconomic efficiency by reducing consumption ofoil, such as product standards (e.g., for fueleconomy), but these approaches inherently lead toadditional economic distortions and thus are alsonot a first-best approach. For example, fueleconomy standards lower the effective cost permile of driving and thus induce more driving, areaction known as the rebound effect. The addi-tional driving may increase the use of oil, redu-cing the energy security (and environmental)benefits and at the same time increasing the exter-nal costs from accidents and congestion.

Policies for Information Market Failures

Information market failures stem from a variety ofsources, and some may be very difficult to address.Those that lead to an underinvestment in distrib-uted generation renewable energy by householdsmay be addressed through information programsto raise awareness. Similarly, consumers typicallycannot readily obtain information about theirinstantaneous use of electricity; they normallyreceive only a monthly bill for their total electri-city use. Programs to provide households withfeedback on the price and usage of electricity(e.g., “smart” meters or in-home dashboards todisplay instantaneous energy use) can help con-sumers make more informed choices relating touse of energy. Both feedback and informationprograms may also reduce behavioral failures, pos-sibly providing an additional benefit.

For interventions to address imperfect infor-mation, such as imperfect foresight for firms, theintervener—presumably a government agency—would have to possess better information. In situ-ations in which a government agency has superior

Market Failure and the Structure of Externalities 83

knowledge, such as of future probable energyprices, an obvious intervention is for the agencyto share that information. The Energy Informa-tion Administration of the U.S. Department ofEnergy provides exactly such data and projectionsaccessible to anyone. In fact, given the ability for agovernment agency to share informationbroadly,16 and at low cost, it is very unlikely thatimperfect foresight about future energy condi-tions would provide a strong case for other gov-ernmental interventions.

In cases when information is particularly diffi-cult to process or a principal-agent issue exists,consumers may be unable to make informed deci-sions, suggesting that the government canimprove economic efficiency by using its superiorinformation-processing ability to make sensiblechoices.This reasoning underlies appliance energyefficiency standards and may perhaps pertain todistributed generation renewables in limited cases.

If managerial incentives are misalignedbecause of the imperfect knowledge of stock mar-ket investors, accounting and information rules topromote transparency and a clearer flow of infor-mation may be warranted. These accounting andinformation rules are not specific to renewableenergy and may also improve economic efficiencyin general.

In addition, if the managers of some firms takea short-term perspective and underinvest inrenewable energy, then other firms with a longer-term perspective may invest more to take advan-tage of the long-term profit opportunities. Ifother firms with a longer-term perspective do notstep in, then there may be motivation for publicsupport for R&D, either through public R&D orsubsidies for private R&D.This may not be a verylikely outcome, but it could occur in the presenceof behavioral failures on the part of stock marketinvestors that lead to a systematic bias towardrewarding short-term performance.

Policies for Regulatory Failures

Policy interventions to reduce regulatory failuresinvolve simply changing the regulatory structureto reduce perverse incentives. For example, toimprove on average cost pricing of electricity,

real-time pricing (RTP) of electricity at thewholesale level could be expanded to the retaillevel.17 However, the benefits of RTP or time-of-use (TOU) pricing would have to be weighedagainst the technology and implementation costs,however.

Policies for Too-High Discount Rates

If the discount rate is too high because of thecorporate income tax, then the failure here is aregulatory failure, and the appropriate responsewould be a tax reform. One tax reform thatwould alleviate this issue would be to allow for theexpensing of capital investments. Other optionsinclude accelerated depreciation for investments,tax credits for research and development, or theelimination of the corporate income taxentirely.18 These issues are not particular torenewable energy development, however, and adeeper examination of tax reform is beyond thescope of this chapter.

If the discount rates are too high because ofcredit limitations, then the appropriate govern-ment response involves macroeconomic policyactions, primarily by the central bank. Both taxreforms and macroeconomic policy actions areeconomywide policy actions that may affectrenewable energy, but they are unlikely to have adisproportionate effect on the renewable energysector in particular.

Policies for Imperfect Foresight

If the evidence is sufficiently strong that a system-atic bias exists as a result of imperfect foresight,this would imply a variety of government inter-ventions designed to provide information aboutpossible future states of the world in order toimprove long-term decisionmaking. Governmentinformation programs that involve data collectionand possibly forecasting reports may help alleviatethe systematic bias by improving firms’ ability topredict future conditions. Increasing regulatoryconsistency by governments implementing clear,predictable long-term renewable energy policiescould also help improve long-term decision-making by firms.

84 Kenneth Gillingham and James Sweeney

Policies for Economies of Scale

Although economies of scale are not likely to playa significant role for renewable energy in general,it may play a role in specific areas. One approachto address economies of scale would be a tempo-rary direct subsidy sufficient to induce firms toproduce at the higher level. Once a sufficientlyhigh level of production is achieved such that thepositive competitive equilibrium can be reached,the subsidy can be removed. As indicated above,because in many cases firms can individually over-come problems of economies of scale, it isunlikely that such approaches are in fact needed.

Policies for Market Power

For market power relating to the possibility offirms buying out competing technologies, possi-bly including renewable energy technologies,enforcement of antitrust laws is likely to be themost effective intervention. In some cases, verticaldisintegration may be warranted to ensure a com-petitive market. Direct government subsidies forprivate R&D investment, coupled with limita-tions on the sale of the subsidized company, areanother possible alternative to address marketpower.

For market power motivating utilities to favortheir own generation over that from outside sup-pliers, a feed-in tariff or equivalent policy mayincrease economic efficiency if the price is setappropriately.The appropriate price would be thewholesale market price for electricity, adjusted forrisk and intermittency. Such a price would pre-vent utilities from favoring their own generation,but it also would prevent any distortions from aprice that does not correspond with the market.19

As an alternative to a feed-in tariff, regulatorscan restructure utilities to ensure that they do notfavor their own generation over that from outsidesuppliers. Alternatively, careful oversight of utili-ties by public utility commissions can also helpaddress market power.

Policies for Intertemporal(Stock-Based) Deviations

Intertemporal deviations are those in which theexternal costs are based primarily on a stock that

changes over time. Individuals influence thesestocks only indirectly, by altering the flows into orout of the stock. But once the flow is determined,those individuals have no further control of thestock. For that reason, policy instruments cannotbe directed toward the stock but must be directedtoward influencing the flow. For economic effi-ciency, the strength of the incentives must beguided by the intertemporal nature of the stockexternality and can be determined from the dis-counted net present value of the entire flow offuture impacts.

Policies for Stock-Based EnvironmentalExternalities: Carbon Dioxide

Carbon dioxide is perhaps the most importantstock-based environmental externality, and there-fore the following discussion focuses on this pol-lutant, but a similar result would hold for anystock-based pollutant.

In any given year, a firm can alter the amountof carbon dioxide it releases into the atmosphere,but once the carbon dioxide is released, the firmhas no further control. That additional carbondioxide remains in the atmosphere, increasing thestock of carbon dioxide for the next century. Theeconomically efficient carbon price in any givenyear (e.g., 2010) can be determined by taking thedamages each subsequent year and discountingthem back to the chosen year using the socialdiscount rate. In this sense, the optimal carbonprice is still the magnitude of the external cost,just as with atemporal environmental externali-ties.

The calculated optimal carbon price differs byyear (e.g., in 2010 compared with 2020) in threeways. First, and most important, the damages arediscounted back further at the earlier date, imply-ing that the carbon price is lower at the earliertime.20 Second, some damages occur during thetime between the two dates. Depending on thedamage function, this difference may be small(perhaps as it is between 2010 and 2020) and maynot change the increase over time of the optimalcarbon price.Third, a natural rate of decline of thestock occurs from dissipation of emissions in the

Market Failure and the Structure of Externalities 85

atmosphere. This fact would slightly increase therate of growth of the optimal carbon price.

The magnitude of the damages from carbondioxide is controversial, and estimates willimprove with increased scientific knowledge. Dif-ferent time patterns of damages would lead todifferent time patterns of the carbon price, basedon the three points discussed above. For example,if the damages from an additional metric ton ofcarbon dioxide grow (in real terms) at the socialdiscount rate, then the optimal carbon price willalso grow (in real terms) at approximately thesocial discount rate. Under the unlikely assump-tion that the incremental damages are constant inreal terms into the future, we could find a nearlyconstant (in real terms) economically efficientcarbon price.

Policies for Imperfect Capture of FuturePayoffs from Current Actions

R&DWhen a market failure occurs as a result of R&Dspillovers from imperfect property rights inknowledge generation, several possible govern-ment interventions might increase economic effi-ciency. The government could directly subsidizeprivate R&D to bring the private rate of returnfrom R&D closer to the social rate of return, anexample of getting the prices right. Such a subsidywould continue as long as a deviation existsbetween the private and social rate of return, andperhaps indefinitely. The economically efficientsubsidy would be set equal to the present dis-counted value of the spillovers from R&D.Importantly, R&D spillovers are likely to exist inmore than just the renewables sector, so an appro-priate policy would also provide the subsidy toprivate R&D in these other sectors.

The government could also directly fundR&D in sectors where spillovers are particularlyhigh. For example, the U.S. government directlyfunds research in renewable energy in nationallaboratories, universities, and some research insti-tutes. Theoretically, public R&D can improveeconomic efficiency if it is focused on researchareas where the social rate of return is sufficientlyhigh relative to the private rate of return. In these

cases, little R&D would have been undertaken byfirms relative to the economically efficientamount, so public R&D complements privateR&D. On the other hand, public R&D cancrowd out the private, depending on the nature ofthe R&D. For example, pure science public R&Dwould be much less likely to crowd out privateR&D than would demonstration projects. Theempirical evidence on public R&D is not clear-cut. David et al. (2000) review the empirical evi-dence on whether public R&D complements orcrowds out private R&D and find an ambiguousresult, suggesting that the result is situation-dependent and underscoring the importance ofthe nature of the R&D in the social rate of returnof public investment.

Intellectual property law plays a key role inhow well firms can capture the rents from theirinnovative activity. Determining the direction inwhich to change intellectual property law is notsimple. If intellectual property law is tightened(e.g., by increasing the length of time patents holdforce), then two opposing effects would exist.Firms could capture more of the benefits of R&Dand thus would have a greater incentive to investin it. But fewer spillover benefits may result fromthe R&D activity, so the social rate of return fromthe activity would be lower. Little empirical evi-dence is available to suggest that either a tighten-ing or loosening of intellectual property lawwould increase economic efficiency.

LBDIf learning by doing occurs in the production of anew technology (e.g., solar photovoltaic installa-tions), then the act of producing increases thefirm’s stock of cumulative experience and thusleads to declines in future costs. The stock ofcumulative experience grows when insights fromprevious production by that or another firmallows it to improve its production techniques.The stock also may decline if some of these tech-niques are forgotten. Theories of LBD oftenproxy all of these complex dynamics by postulat-ing that the cost of future production for all firmsat any time will be a function of the cumulativestock of experience from production in the mar-ket. But the market failure can be thought of in a

86 Kenneth Gillingham and James Sweeney

more general sense as a spillover from the stock ofany single firm’s cumulative experience from pro-duction to other firms.

Once a firm chooses how much to produce atany given time (i.e., the flow into the stock ofexperience), it subsequently has no further con-trol over the stock of experience and its impact onfuture costs. Thus economically efficient policiesfor LBD must focus on the quantity produced,while taking into account the fact that experienceis a stock. The most straightforward policy instru-ment to address LBD spillovers is a subsidy. Theeconomically efficient per-unit subsidy equals thediscounted present value of all future cost reduc-tions resulting from the additional production thatcannot be captured by the individual firm.

However, a second element to the economictheory behind LBD also may affect the economi-cally efficient policy: the spillovers from LBD arepostulated to decline along with the costs. Forexample, in the standard formulation, as illus-trated in Figure 5.2, the percentage cost decreasedepends on the percentage increase in the stock,so every additional unit has a progressivelydecreasing percentage impact on costs. Conse-quently, as cost decreases with a greater stock, agiven percentage cost decrease leads to a smallerabsolute cost decrease. These factors togetherimply that the LBD externality—and hence the

appropriate magnitude of the intervention—willbe declining over time. Acting in the oppositedirection, if the sales are growing rapidly, the costreduction is applied to a larger amount of produc-tion, reducing the rate of decline of the interven-tion.

Thus optimal subsidies for LBD will likely betransient and decline over time as LBD runs itscourse. The speed at which the subsidies arephased out will depend on the particular technol-ogy and may require adjustment if different con-ditions arise than were initially expected. In oneexample, the optimal solar PV subsidies for Cali-fornia calculated under the baseline assumptionsin van Benthem et al. (2008) follow a decliningpath and are phased out over 15 years.

Network externalitiesNetwork externalities may play a role in theadoption of distributed generation renewableenergy. If it can be demonstrated that a networkexternality truly exists, rather than networkeffects, then one approach to correct for thisexternality would be a temporary production sub-sidy (Goolsbee and Klenow 2002). Once a prod-uct has taken over (nearly) the entire market, therewould be no room for further spillovers and thusno need for the subsidy policy.

Installationsin 2010Discountfuturecumulativebenefitsto 2010 Installations

in 2015Discountfuturecumulativebenefitsto 2015

Installationsin 2020Discountfuturecumulativebenefitsto 2020

2010 2012 2014 2016 2018 2020Year of benefit

2022 2024 2026 2028 2030

Figure 5.2. Illustrative incremental benefits from additional cumulative installations: LBD

Market Failure and the Structure of Externalities 87

ConclusionsRenewable energy has an immense potential toserve our energy needs, and in the long run, atransition from depletable fossil fuel resources torenewable energy is inevitable. This chapter hasdelved into reasons why policymakers should beinterested in policies to promote renewableenergy, pointing to a variety of market failuresthat may lead to a divergence between the optimaltransition to renewables and the observed transi-tion. Economic theory suggests that we canimprove economic efficiency by matching thepolicy instrument to the market failure.

The structure and nature of each market fail-ure have important ramifications for the appropri-ate policy actions to correct for the market failureand move closer to an optimal transition torenewable energy. The discussion above distin-guished between atemporal market failures andintertemporal (i.e., stock-based) market failures.In either case, the economically efficient policyaction matches the temporal pattern of the marketfailure.This implies a temporary policy (e.g., LBDspillovers) in some cases and a permanent policy(e.g., R&D spillovers) in others.

Renewable energy policy is likely to requireseveral different policy instruments to address thevarious kinds of market failures. When the marketfailures are closely related, a single policy instru-ment can address, or partly address, more thanone market failure. For instance, provision ofinformation about low-cost or low-effort oppor-tunities to save energy and help preserve the envi-ronment may reduce the informational marketfailure and also influence consumers to partlyinternalize the environmental externalities(Bennear and Stavins 2007).

For renewable energy, the most importantmarket failures, with the strongest empirical evi-dence, appear to be environmental externalities,innovation market failures, national security mar-ket failures, and regulatory failures. Only a few ofthe market failures identified in this chapter areunique to renewable energy. Environmentalexternalities due to fossil-fuel use are the mostimportant of these, but if policy action is alreadyunder way to correct for externalities from fossil-

fuel emissions, then we must look to other marketfailures for motivation for renewable energypolicy. As these other market failures often applyto other parts of the economy, addressing themmay entail policy actions that extend muchbeyond renewable energy.

Political feasibility is a final consideration withimportant ramifications for renewable energypolicy. In some cases, the first-best policyapproach may not be politically feasible. Asecond-best approach may involve multipleinstruments, even in cases when the first-bestapproach involved only a single instrument. Forexample, rather than a single tax to internalizeenvironmental externalities, the same price differ-ential can be achieved by combining a smaller tax(or no tax) on fossil fuels with a subsidy forrenewable energy. Similarly, a cap-and-trade sys-tem may not be politically feasible because ofuncertainty about how high the costs of abate-ment might be, so a more viable option might beto use two instruments in a hybrid cap-and-tradeand tax system, commonly known as a cap-and-trade with a safety valve (Jacoby and Ellerman2004; McKibbin and Wilcoxen 2002; Pizer 2002).

In other cases, the only politically feasibleoptions for addressing the market failures relevantto renewable energy are not the first-choiceinstruments, but rather second-choice instru-ments that address the market failures indirectly.Renewable portfolio standards (RPSs) are one ofthe most prominent examples of a policy instru-ment that only indirectly addresses the marketfailures relevant to renewable energy. By setting arequirement on the amount of renewable energyin each utility’s electricity generation mix, anRPS adds an implicit subsidy on renewableenergy, with the magnitude of the subsidy directlyrelated to the stringency of the cap. If the RPS iscarefully set, this implicit subsidy could act justlike an appropriately set actual subsidy—leadingto the second-best outcome described above.However, finding this appropriate level for anRPS may be exceedingly difficult and subject tointense political disputes.

A sensible set of policies to address the marketfailures relevant to renewable energy has thepotential to greatly improve economic effi-

88 Kenneth Gillingham and James Sweeney

ciency—and at the same time would have otherbenefits, such as improving air quality and miti-gating the risk of catastrophic global climatechange. Much future work remains to betterquantify the most relevant market failures and fur-ther improve our understanding of how todevelop policies to best address these market fail-ures.

AcknowledgmentsThe authors thank Richard Schmalensee, BoazMoselle, Arthur van Benthem, Douglas Hannah,and Jonathan Leaver for very helpful comments.Any remaining errors are the sole responsibility ofthe authors.

Notes

1. The concept of behavioral failures stems frombehavioral economics and is quite new to environ-mental economics. See Shogren andTaylor (2008)and Gillingham et al. (2009) for recent reviewsdiscussing the concept in the context of environ-mental economics.

2. Economic theory defines “economically efficient”in technical terms as an allocation of resourceswhere no potential Pareto improvement exists,which refers to a reallocation of resources thatbenefits at least one individual and imposes nocosts on any others. Note that economic effi-ciency is a distinct concept from the equity orfairness of an allocation of resources.

3. It is still theoretically unclear how to disentanglesystematic biases in decisionmaking from inherentpreferences, but behavioral welfare analysis is anarea of active theoretical development and mayeventually shed light on this issue. See, e.g.,Bernheim and Rangel (2009).

4. Important equity or fairness concerns may also beinvolved. This chapter focuses on economic effi-ciency as a policy goal, while noting that equityconsiderations, in theory, can often be dealt withthrough lump-sum transfers of wealth that do notdistort incentives or through modifications of theincome tax rates. If the policy goal is reducing

global inequity, other distributional policies arelikely to be more effective than renewable energypolicy.

5. It is important to note that unless a behavioralfailure is a systematic, rather than random, depar-ture of observed choice from a theoretical opti-mum, it may be very difficult to formulate poli-cies. If the systematic departure is in a consistentdirection, the intervention can work in the oppo-site direction to correct this deviation. But ran-dom deviations would require an interventioncontingent on the deviation. For example, poorinformation about the operating characteristics ofdistributed photovoltaics could lead some peopleto install these devices even though they wouldultimately come to regret the decision and otherpeople not to install them even though the deviceswould have turned out to be beneficial. In suchcircumstances, development and dissemination ofinformation about photovoltaic operating charac-teristics for alternative locations could improvesuch decisions. For the most part, however, policyoptions designed to compensate for randomdeviations would be difficult to formulate andeffectively implement.

6. “Full employment” for a well-functioning devel-oped economy refers here to “at the natural rate ofunemployment.” Some unemployment will alwaysexist as a result of transitions between jobs andmismatches of available and needed skills.

7. Personal income taxes or labor taxes, such as theU.S. Social Security tax, provide incentives toreduce the supply of labor, so that the marginalsocial value of labor exceeds the value of that laborto the worker. However, issues of the distortionsassociated with and the reform of such tax systemsgo well beyond the scope of this chapter.

8. This remains a concern for the overall economicefficiency of investment, even when it does notdistort the mix of renewables versus nonrenewabletechnologies.

9. “Almost indifferent” because the cumulativeimpacts of emitting a ton now may be somewhatdifferent from the impacts of emitting a ton in 20years and because the regulatory environmentcould change in that period.

10. LBD is closely related to economies of scale,except that learning by doing has a distinctly dif-ferent intertemporal relationship, where costsdecline as a function of cumulative production,and increasing returns to scale implies that averagecosts decline with production at a given time.

Market Failure and the Structure of Externalities 89

11. If the knowledge leading to cost reductions by onefirm could have been used by other firms, but thatfirm somehow manages to keep all the knowledgeprivate, even if it does not use it (so it is effectively“wasted knowledge”), there would still be a mar-ket failure in that some of the potential benefits ofthe learning would not be captured by anyone.

12. As noted above, this does not deal with the labormarket problems associated with income taxes orlabor taxes that provide incentives to reduce thesupply of labor.

13. A substantial literature addresses the trade-offsbetween a tax and a cap-and-trade system, par-ticularly relating to policymaking under uncer-tainty. For a recent review discussing these issues,see Aldy et al. (2009).

14. If behavioral failures have caused an under-investment in energy efficiency, then there maynot be an overinvestment in energy efficiencyfrom the subsidy.

15. Information available at www.hara.com.16. In some cases, release of information is not possi-

ble or is undesirable, such as when it involvesweapons programs or other programs closelyrelated to national security.

17. In order to be effective, RTP would have to becomplemented with real-time feedback on theelectricity price at the current time.

18. Note that reducing taxes in some areas mayrequire increasing taxes in others, so a full analysisshould examine the relative distortions from eachof the different taxes.

19. For example, in the United States, the Public Util-ity Regulatory Policies Act (PURPA) of 1978required electric utilities to buy power from small-scale nonutility producers at the avoided cost rate,which is the cost the utility would incur were it toacquire the electricity from other sources. Choos-ing the appropriate price turned out to be remark-ably problematic.

20. For instance, the 2020 efficient tax would begreater than the 2010 efficient tax by a factor of (1+ r)10, where r is the annual social discount rate. Ata 5% discount rate, the 2020 carbon tax would be63% greater than the 2010 carbon tax.

ReferencesAghion, P., N. Bloom, R. Blundell, R. Griffith, and P.

Howitt. 2005. Competition and Innovation: AnInverted-U Relationship. Quarterly Journal of Eco-nomics 120: 701–728.

Aldy, J., A. Krupnick, R. Newell, I. Parry, and W. Pizer.2009. Designing Climate Mitigation Policy. Discus-sion paper 08–16. Washington, DC: Resources forthe Future.

Archer, C., and M. Jacobson. 2005. Evaluation of Glo-bal Wind Power. Journal of Geophysical Research D110: 1–20.

Arrow, K. 1962. The Economic Implications of Learn-ing by Doing. Review of Economic Studies 29: 155–173.

Baumol, W. 1972. On Taxation and the Control ofExternalities. American Economic Review 62: 307–322.

Bennear, L., and R. Stavins. 2007. Second-BestTheoryand the Use of Multiple Policy Instruments. Environ-mental and Resource Economics 37: 111–129.

Bernheim, D., and A. Rangel. 2009. Beyond RevealedPreference: Choice-Theoretic Foundations forBehavioral Welfare Economics. Quarterly Journal ofEconomics 124: 51–104.

Blundell, R., R. Griffith, and J. Van Reenen. 1999.Market Share, Market Value and Innovation in aPanel of British Manufacturing Firms. Review of Eco-nomic Studies 66: 529–554.

Bohi, D., and M. Toman. 1996. Economics of EnergySecurity. Norwell, MA: Kluwer Academic Publish-ers.

Borenstein, S. 2008. The Market Value and the Cost ofSolar Photovoltaic Electricity Production. CSEMWorking paper 176. Berkeley, CA: University ofCalifornia Energy Institute.

Cherni, J., and J. Kentish. 2007. Renewable EnergyPolicy and Electricity Market Reforms in China.Energy Policy 35: 3616–3629.

Da Rosa, A. 2005. Fundamentals of Renewable EnergyProcesses. Amsterdam; Elsevier Academic Press.

David, P. 1985. Clio and the Economics of QWERTY.American Economic Review 75: 332–337.

David, P., B. Hall, and A.Toole. 2000. Is Public R&D aComplement or Substitute for Private R&D?Research Policy 29: 497–529.

EIA (Energy Information Administration). 2008. Inter-national Energy Annual 2006. Washington, DC:Department of Energy, EIA.

Gillingham, K., R. Newell, and K. Palmer. 2009.Energy Efficiency Economics and Policy. AnnualReview of Resource Economics 1: 597–619.

Goolsbee, A., and P. Klenow. 2002. Evidence onLearning and Network Externalities in the Diffusionof Home Computers. Journal of Law and Economics35: 317–343.

90 Kenneth Gillingham and James Sweeney

Goulder, L., and S. Schneider. 1999. Induced Techno-logical Change and the Attractiveness of CO2 Emis-sions Abatement Policies. Resource and Energy Eco-nomics 21: 211–253.

Heal, G. 1993. The Optimal Use of ExhaustibleResources. In Handbook of Natural Resource andEnergy Economics, Vol. 3, edited by A. Kneese and J.Sweeney. Amsterdam: Elsevier.

Hermann, W. 2006. Quantifying Global ExergyResources. Energy 31: 1685–1702.

Jacoby, H., and A. D. Ellerman. 2004. The Safety Valveand Climate Policy. Energy Policy 32: 481–491.

Jaffe, A., R. Newell, and R. Stavins. 2004. The Eco-nomics of Energy Efficiency. In Encyclopedia ofEnergy, edited by C. Cleveland. Amsterdam: Elsevier,79–90.

Jaffe, A., and R. Stavins. 1994.The Energy Paradox andthe Diffusion of Conservation Technology. Resourceand Energy Economics 16: 91–122.

Jamasb, T. 2007. Technical Change Theory and Learn-ing Curves: Patterns of Progress in Electricity Gen-eration Technologies. Energy Journal 28: 51–71.

Lessem, N., and R.Vaughn. 2009. Image Motivation inGreen Consumption. Working paper. Los Angeles:University of California–Los Angeles.

Levinson, A., and S. Niemann. 2004. Energy Use byApartmentTenants When Landlords Pay for Utilities.Resource and Energy Economics 26: 51–75.

Liebowitz, S., and S. Margolis. 1994. Network Exter-nality: An Uncommon Tragedy. Journal of EconomicPerspectives 8: 133–150.

McKibbin, W., and P. Wilcoxen. 2002. The Role ofEconomics in Climate Change Policy. Journal of Eco-nomic Perspectives 16: 107–129.

Murtishaw, S., and J. Sathaye. 2006. Quantifying theEffect of the Principal-Agent Problem on US Residential

Use. LBNL-59773. Berkeley, CA: LawrenceBerkeley National Laboratory.

Nickell, S. 1996. Competition and Corporate Perform-ance. Journal of Political Economy 104: 724–746.

Nordhaus, W. 2002. Modeling Induced Innovation inClimate Change Policy. In Technological Change andthe Environment, edited by A. Grubler, N.Nakicenovic, and W. Nordhaus. Washington, DC:Resources for the Future Press.

Nordhaus, W. 2009. Designing a Friendly Space forTechnological Change to Slow Global Warming.Snowmass Conference on Technologies to CombatGlobal Warming. August 2009, Snowmass, CO.

Pizer, W. 2002. Combining Price and Quantity Con-trols to Mitigate Global Climate Change. Journal ofPublic Economics 85: 409–434.

Pizer, W., and D. Popp. 2008. Endogenizing Techno-logical Change: Matching Empirical Evidence toModeling Needs. Energy Economics 30: 2754–2770.

Popp, D. 2006. R&D Subsidies and Climate Policy: IsThere a“Free Lunch?” Climatic Change 77: 311–341.

Rappaport, A. 1978. Executive Incentives vs. Corpo-rate Growth. Harvard Business Review 56: 81–88.

Scherer, F. 1967. Market Structure and the Employ-ment of Scientists and Engineers. American EconomicReview 58: 524–531.

Shogren, J., and L. Taylor. 2008. On Behavioral-Environmental Economics. Review of EnvironmentalEconomics and Policy 2: 26–44.

U.S. Congress. 2009. American Recovery and Reinvest-ment Act of 2009. 111th Cong., 1st sess.

van Benthem, A., K. Gillingham, and J., Sweeney.2008. Learning-by-Doing and the Optimal SolarPolicy in California. Energy Journal 29: 131–151.

Young, P. 2010. Innovation Diffusion in Heterogene-ous Populations: Contagion, Social Influence, andSocial Learning. American Economic Review 99: 1899–1924.

Market Failure and the Structure of Externalities 91

6

Renewable Energy, EnergyEfficiency, and Emissions TradingJosé Goldemberg

The present energy system is based primarilyon the use of fossil fuels. In 2008, total pri-

mary energy consumption was 516 exajoules(EJ)—corresponding to 12.32 billion tons equiva-lent of petroleum—of which coal, oil and gas rep-resented 81% (Figure 6.1). The remaining 19%originated in traditional biomass (8%), usedmainly in developing countries; nuclear (6%);large hydropower (2%), which is a renewable

source of energy; and “new renewables” (3%),which include modern biomass, geothermal,solar, wind, small hydro, and marine energy.1

Large hydropower and new renewables togetherare referred to as “modern renewables.”

The 20th century marked the emergence offossil fuels, which represented less than 20% in themiddle of the 19th century. Until then,renewables—mainly fuelwood—supplied the bulk

Oil33.34%

Gas20.42%

Nuclear6.03%

New renewables 2.96%

Modern renewables 5.03%

Coal27.29%

Traditional biomass7.90%

Modern biomass1.83%

Large hydro2.06%

Small hydro0.24%

Marine0.00%

Solar0.11%

Wind0.16%

Geothermal0.64%

Modern biomass EJ %

Bioethanol 1.7 0.32Biodiesel 0.5 0.09

Bioelectricity 3.6 0.70

Heat 3.7 0.71

Sources: Author’s elaboration on the basis of UNDP 2000, 2004; REN21 2009

Figure 6.1. World primary energy supply shares of 516 exajoules (EJ)

of the energy needed for all purposes rangingfrom home heating to cooking to manufacturing(Figure 6.2).

These energy sources allowed the industrial-ized countries to achieve a high level of develop-ment benefiting at least one-third of the world’spopulation. Fossil fuels were cheap and abundantthroughout most of the 20th century, which dis-couraged efforts to optimize energy systems andreduce energy losses. The continued heavy reli-ance on fossil fuels today as the main source ofenergy is creating serious problems, which arebecoming increasingly evident as time goes by:

• Reserves of fossil fuels are finite and by theend of the 20th century were showing signsof exhaustion.

• Impurities of fossil fuels, such as sulfur oxidesand particulate matter, became a main sourceof pollution at the local and regional levels.

• The combustion of fossil fuels has the un-avoidable result of producing carbon dioxide(CO2), which is changing the composition ofthe atmosphere, as well as other greenhousegases (GHGs). Before the industrial age, atthe end of the 17th century, the amount ofCO2 in the atmosphere was 0.027%, or 270parts per million (ppm), but it has beengrowing steadily and reached 0.0338%, or

338 ppm, in 2008. Such increase is the mainsource of global warming and associated cli-mate changes.

Figure 6.3 shows the evolution of CO2 emissionsresulting from the present energy system. Carbondioxide emissions from Annex B countries (of theClimate Convention) have stabilized since 1990but are growing rapidly in developing (non-Annex B) countries (of the Climate Convention)at a rate of approximately 4% per year; this isreflected in world emissions, which are growingroughly 600 million tons of CO2 per year.

Carbon dioxide emissions are the dominantcomponent of greenhouse gas emissions, but in2006, they represented only 69.6% of total emis-sions.The remaining 30.4% were methane (CH4),nitrous oxide (N2O), and fluorinated gases withhigh global warming potential (GWP), which aresulfur hexafluoride (SF6), hydrofluorocarbons(HFCs), and perfluorocarbons (PFCs) (Figure6.4).

GHG emissions are usually expressed in CO2

equivalent. Total emissions in 2005 were approxi-mately 45 gigatons (Gt) of CO2 equivalent. Toreduce CO2 and other GHG emissions thusbecame one of the most urgent tasks we facetoday. There are two approaches to this problem:use energy more efficiently, consequently emit-

100

80

60

Per

cent

age

40

20

0

1850 1900

Year

Traditionalrenewables

Nuclear Modern biomass

Hydro

Oil

Gas

Coal

1950 2000

Source: UNDP 2000

FIGURE 6.2. World energy consumption, 1850–2000

94 José Goldemberg

ting less CO2 and extending the life of fossil-fuelreserves; and increase the contribution of renew-able energy in the world energy matrix.

National governments, as well as some sectorsof the productive system, such as industry, trans-portation, and residential, can adopt these solu-tions to varying degrees. In industrialized coun-tries, which have already reached a high level ofenergy consumption per capita, energy efficiencycan be implemented more easily, but renewableenergy can also play a significant role. In develop-ing countries, where energy consumption percapita is low and growth in energy services isinevitable, clean and efficient technologies andrenewable energy can be incorporated early in theprocess of development, following a different paththan that taken in the past by today’s industrializedcountries.

Against this background, this chapter discussesthe potential of energy efficiency, renewableenergy, and emissions trading schemes in achiev-ing the objective of reducing greenhouse gasemissions.

Renewable EnergyTable 6.1 lists the energy from various types ofrenewable sources used in the world at the end of2008, as well their yearly growth rates.Traditionalbiomass is left out of this table because it is usedmainly in rural areas as cooking fuel or charcoal inways that are frequently nonrenewable, leading todeforestation and soil degradation.

Some of the renewable energy sources aredeveloping rapidly, with impressive growth ratesof 38% per year for photovoltaic (PV) and 25%per year for wind (see Figure 6.5).

In 2008, renewables (including large hydro)represented approximately 5% of the world’s totalprimary energy consumption, but they are grow-ing at a rate of 6.3% per year, whereas total pri-mary energy supply is growing at a smaller rate ofapproximately 2% per year. Taking into accountthe appropriate efficiency and capacity factors,2

the numbers in Table 6.1 can be converted intothe total primary energy contribution fromrenewables, as shown in Table 6.2 and Figure 6.6.

An extrapolation of the contribution ofrenewables up to 2030 on the basis of the rates ofgrowth in the last 10 years is shown in Figure 6.7.

To give an idea of the effort that would beneeded to curb CO2 emissions up to 2050, the

Annex B countries (0.05% per year)

World

Non-Annex B countries (4.07% per year)

35.00

30.00

25.00

20.00

15.00

CO

2 em

issi

ons

(mill

ion

tons

CO

2)

10.00

5.00

–1988 1990 1992 1994 1996 1998 2000 2002 2004 2006

Year

Source: CDIAC 2009

Figure 6.3. World fossil-fuel CO2 emissions and growth rate, 1990–2005

Renewable Energy, Energy Efficiency, and EmissionsTrading 95

International Energy Agency (IEA) recently pro-duced two scenarios of what would be required interms of renewables in the electricity sector. Inone, called the ACT Scenario, global CO2 emis-sions would be returned to current levels by 2050;in the other, the BLUE Scenario, CO2 emissionswould be reduced by 50% from current levels.The results are shown in Table 6.3.

In the IEA scenarios, nuclear energy and coal-and gas-fired thermal power plants with carboncapture and storage (CCS) are included. Thesenumbers are very large but give an idea of theeffort required to prevent catastrophic climatechange.

The main policy instruments used to acceler-ate the introduction of renewables in the energysystems of a number of countries are feed-in tariffsand renewable portfolio standards (RPSs).Feed-in tariffs are a policy adopted by govern-ments to accelerate the introduction of renewableenergy sources in their matrixes. They obligateutilities to purchase the output of renewable gen-erators at a fixed price, which is set high enoughto ensure that the renewable generation in ques-

tion is profitable.The costs of the feed-in tariff aregenerally passed on to consumers via regulatedcharges. Feed-in tariffs have been used in Ger-many and Spain (see Chapters 14 and 15), and inboth countries, they have been associated withrapid growth in renewable generation. Under the

High GWPgases

N2O

CO2

CH4

22.9%

7.1%

69.6%

Source: UNFCCC 2009a

Figure 6.4. Contributions of greenhouse gases toglobal warming in 2006

Table 6.1. Electricity, heat, and transport fuel from renewable sources

Year

Growth rate per yearElectricity (GW)

end 1998 a end 2001 b end 2006 c end 2008 d

Large hydropower 640 690 774 860 3.00%

Wind power 10 23 74 121 28.32%

Small hydropower 23 25 73 85 13.96%

Biomass power 40 40 45 52 2.66%

Solar PV 0.5 11 7.8 13 38.52%

Geothermal 8 8 9.5 10 2.26%

Solar thermal power 0.4 0.4 0.4 0.5 2.26%

Ocean (tidal) power 0.3 0.3 0.3 0.3 0.00%

Hot water/heating (GWth) Growth rate per year

Biomass heating 200 210 235 250 2.26%

Solar collectors 18 57 105 145 23.20%

Geothermal 11 11 33 50 16.35%

Transport fuels (billion liters/year) Growth rate per year

Ethanol 18 18 39 57 12.22%

Biodiesel 0 1.2 6 12 38.95%

Sources: aUNDP 2000; bUNDP 2004; cREN21 2007; dREN21 2009Note: GW = gigawatts; GWth = gigawatts-thermal

96 José Goldemberg

German system, the rates paid for new contractsdecline annually, forcing the green energy sectorto innovate.

Renewable portfolio standards (RPSs) requireenergy suppliers to purchase a proportion of theirelectricity (typically 10% to 20%) from renewable

Biodiesel

PV gridWind electricity

Solar low-temperature heatGeothermal heat

Small hydropowerBioethanol

Geothermal electricity

Biomass electricity

Biomass heatPV off-grid

Solar thermal electricity

0% 10% 20% 30% 40% 50% 60% 70%Source: REN21 2009

Figure 6.5. Annual growth rates of renewable energy capacity, 1998–2008

Table 6.2. Primary energy production from new renewables

Energy production (EJ) Growth rate

Source/technology 1998 2001 2004 2008Modern biomass energy Total 6.033 6.369 7.001 9.422 4.56%

Bioethanol 0.450 0.45 0.763 1.675 14.05%

Biodiesel 0.000 0.045 0.079 0.45 38.95%

Electricity 2.618 2.782 2.921 3.616 3.28%

Heat 2.965 3.092 3.238 3.681 2.19%

Geothermal Total 1.800 2.106 2.604 3.285 6.20%

Electricity 1.656 1.908 2.147 2.385 3.72%

Heat 0.144 0.198 0.457 0.900 20.11%

Small hydropower Total 0.324 0.36 0.878 1.224 14.22%

Wind electricity Total 0.065 0.155 0.32 0.814 28.76%

Solar Total 0.056 0.212 0.280 0.573 26.18%

Low-temp heat 0.050 0.205 0.270 0.522 26.44%

PV grid 0.002 0.004 0.007 0.047 37.12%

PV off-grid 0.000 0.000

Thermal electricity 0.004 0.003 0.003 0.004 0.00%

Marine energy Total 0.009 0.009 0.009 0.009 0.00%

Tidal 0.009 0.009 0.009 0.009 0.00%

Total new renewables Total 8.29 9.21 11.09 15.33 6.34%

Large hydropower Total 8.98 9.09 9.29 10.12 1.20%

Total renewables 17.27 18.30 20.38 25.45 3.95%

Sources: UNDP 2000, 2004; REN 21 2007, 2009

Renewable Energy, Energy Efficiency, and EmissionsTrading 97

energy sources. This enables renewable generatorsto charge higher prices for their electricity. Inpractice, the policy is implemented by havingrenewable energy generators earn certificates forevery unit of electricity they produce. They thensell these along with their electricity to supplycompanies. As discussed in Chapters 13 and 11 ofthis book, RPS mechanisms have been used in theUnited Kingdom and in many U.S. states.

Renewable energies are being introduced in asignificant way in many countries, particularly inEurope in the form of “distributed generation”(Bayod-Rújula 2009), which is mostly renewable

and seems to be the approach that will be used ona large scale in the future (Figure 6.8).3

Energy EfficiencyThe amount of energy required to provide theenergy services needed depends on the efficiencywith which the energy is produced, delivered, andused. Gains in energy efficiency are usually meas-ured by indicators, one of which is called energyintensity (I) and defined as the energy (E) neces-

80

70

60

50

40

30En

erg

y [E

J]

Year

20

10

08.98 9.09 9.29

10.12 11.68

Total renewables New renewables32.04

43.72

13.16

59.25

72.41

15.33

25.45

11.09

20.38

9.21

18.30

8.29

17.27

1998 2000 2002 2004 2006 2008 2010 2012 2014 2016 2018 2020 2022 2024 2026 2028 2030

Large hydropower

Sources: UNDP 2000, 2004; REN 21 2007, 2009

Figure 6.7. Projection of renewable energy production up to 2030

Sources: Author’s elaboration on the basis of UNDP 2000, 2004; REN 21 2007, 2009

Figure 6.6. Contribution of renewables to the total energy supply (EJ)

98 José Goldemberg

sary per unit of gross domestic product (GDP): I= E/GDP. Reductions in the energy intensity ofeconomies over time indicate that the sameamount of GDP is obtained with a smaller energyinput, as shown in Figure 6.9.

Figure 6.10 shows the impressive improve-ments in energy efficiency made in the OECDcountries in the period 1973–1998; without thesegains, energy consumption would be 49% higherthan it was in 1998, which means a decline inenergy use of 2.3% per year. In terms of CO2

emissions for the OECD countries, this means areduction of roughly 350 million metric tons ofCO2 per year.

Such decline can be attributed to a combina-tion of the following structural factors:

• improvements in efficiency of use of materi-als and manufactured goods in industrializedand transition countries, such as increasedrecycling, substitution away from energy-

Austri

a

Belgium

Denm

ark

Fran

ce

Germ

any

Greec

e

Irelan

dIta

ly

Luxe

mbo

urg

Nethe

rland

s

Portu

gal

Spain

Sweden UK

Czech

Rep

ublic

Estonia

Hunga

ry

Latvi

a

Lithu

ania

Poland

Roman

ia

Slovak

ia

Sloven

ia

Bulgar

ia

Finlan

d0

5

10

15

20

25

30

35

40

45

50

DG

sha

re (

% e

nerg

y pr

oduc

tion

in 2

004)

Source: Cossent et al. 2009

Figure 6.8. Energy produced from distributed generation (DG) in European countries

Table 6.3. Additional rate of investment in power generation technologies needed in the electricity sector everyyear until 2050

ACT scenario BLUE scenario Normalized plant size

Coal-fired with CCS 30 35 500 MW plants

Gas-fired with CCS 1 20 500 MW plants

Nuclear 24 32 1,000 MW

Hydro 1/5 of Canada’s hydropower capacity

Biomass plants 30 100 50 MW plants

Wind: inshore 2,900 14,000 4 MW wind turbines

Wind: offshore 775 3,750 4 MW wind turbines

Geothermal 50 130 100 MW plants

Solar PV 115 215 Million solar panels

Solar CSD 45 80 250 MW plants

Source: OECD/IEA 2008a, 2008b

Renewable Energy, Energy Efficiency, and EmissionsTrading 99

intensive materials, improved material effi-ciency, and intensified use of durable andinvestment goods;

• shifts in the mix of activities undertaken inthe economy toward services and types ofindustrial production that are less energy-intensive; and

• saturation effects, whereby societies areapproaching natural limits on the number ofcertain goods in the residential and transpor-tation sectors (e.g., refrigerators, televisionsets, cars) that they can absorb, resulting inenergy usage growing at a slower rate thanGDP. (UNDP 2004)

Because more than 80% of the energy used in theworld today comes from fossil fuels, the reductionin energy intensity is reflected in a reduction incarbon intensity (I = CO2/GDP), defined as theamount of CO2 emitted per unit of GDP (in par-ity purchasing power) as shown in Figure 6.11.

As can be seen, the OECD countries haveexperienced a steady decline in the carbon inten-sity. Non-OECD countries also had a decline, butcarbon intensity stabilized after the year 2000.Figure 6.12 shows the energy intensities for dif-ferent industrial sectors, indicating that some ofthem have been declining more rapidly than oth-ers. This clearly points out the most interestingsectors from the viewpoint of optimizing energysavings.

Over the next 20 years, there is significantscope for improvements in energy efficiency,which may largely be attained by replacement ofthe existing capital stock. It is estimated thatwithin the next two decades, energy efficiency inindustrialized countries could be cost-effectivelyreduced by 25% to 35%, whereas in transitionaleconomies, reductions of more than 40% couldbe achieved. Developing countries, which typi-cally have high economic growth and dated capi-

Source: OECD/IEA 2008a, 2008bNote: RoW = Rest of the world

Figure 6.9. Evolution of energy intensity in anumber of countries and regions, 1990–2005

Table 6.4. The carbon market: Volumes and values for emissions trading, 2006–2007

Allowances 2006 2007

Volume(MtCO2e)

Value($ million)

Volume(MtCO2e)

Value($ million)

EU ETS 1,104 24,436 2,061 50,097

New South Wales 20 225 25 224

Chicago Climate Exchange 10 38 23 72

UK ETS N/A N/A

Total 1,134 24,699 2,109 50,393

Source: World Bank 2008Notes: The New South Wales scheme and the Chicago Climate Exchange, for voluntary trading of emissions, were set before the EUscheme; MtCO2e = metric tons of carbon dioxide equivalent.

100 José Goldemberg

tal and vehicle stocks, could achieve cost-effectiveimprovement of 30% to more than 45%

Global energy intensity could decline by asmuch as 2.5% per year as a result of a combinationof structural changes and efficiency improve-ments. The extent to which this will be achieveddepends on the effectiveness of governmentpolicy, adaptation of attitudes and behaviors, andthe level of entrepreneurial activity in energy con-servation and material efficiency.

Policymakers play a central role in drivingenergy efficiency improvements by setting stand-

ards, such as building codes; market-based incen-tives, including certificate markets; and adequatepayment systems (World Bank 2008) for energythat are applied by well-informed consumers,planners, and decisionmakers.

Emissions TradingThe discussions above have given brief overviewsof two of the main classes of technology that are

160

140

120Hypothetical energy use without savings

Actual energy use

49%Additional energy usewithout intensity declines =energy savings100

80

60

40

20

Year

Ene

rgy

(EJ)

0

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

Source: OECD/IEA 2004

Figure 6.10. Energy savings in the OECD countries, 1973–1998

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0Kilo

gram

s C

O2/

US

$ us

ing

2000

pric

es a

nd P

PP

1970 1975 1980 1985 1990 1995 2000 2005

Year

WorldNon-OECD totalOECD total

Source: OECD/IEA 2007Note: PPP is the purchasing power parity, adjusting nominal GDP figures according to the living conditions of each country

Figure 6.11. Carbon intensity: CO2 emissions per GDP (PPP), 1970–2005

Renewable Energy, Energy Efficiency, and EmissionsTrading 101

available to reduce GHG emissions: improve-ments in energy efficiency and the increased useof renewable energy. Emissions trading, fre-quently known as cap-and-trade, is a technology-neutral tool to incentivize efficient emissionsreduction by companies and countries. Severalemissions trading schemes are currently in placeand generally function as follows:

• Setting of a cap. A government or interna-tional body sets a limit on the total amount ofpollutants that can be emitted in a specificperiod.

• Allocation of allowances. Participants in thescheme are allocated allowances to pollute acertain quantity. Allowances can be allocatedon a number of bases, such as historical emis-sions, or through an auction of rights. Thetotal number of allowances allocated cannotexceed the cap set by the governing body.

• Emissions trading. Scheme participants thatemit more than their allocations must pur-chase additional allowances, while those thathave an excess number of allowances may sellthese. Participants that can reduce their emis-sions at low cost have an incentive to do so,and sell allowances, whereas those that facehigh costs of abatement that exceed the costsof allowances must pay to pollute more.

One such scheme was implemented in the UnitedStates under the 1990 Clean Air Act to curb emis-sions of sulfur dioxide (SO2; a local pollutant, nota greenhouse gas). By 2007, the scheme suc-ceeded in reducing annual SO2 emissions toapproximately 50% of 1980 levels (EPA 2009).

Another emissions trading scheme was imple-mented at the international level by the KyotoProtocol, which came into force in 2005. Thisbinds the participants to a cap-and-trade systemcovering the six major greenhouse gases. Thescheme covers only developed nations, in recog-nition of the fact that they are principally respon-sible for the majority of the past global emissions.Notably, the United States, which is responsiblefor a significant proportion of world emissions, isthe only developed nation that signed theUNFCCC that is not participating in the scheme.

The participants agreed to an average overallreduction in emissions of approximately 5% rela-tive to 1990 levels by the end of 2012 (UNFCCC2009b). During the five-year compliance periodbetween 2008 and 2012, participants can earn andtrade credits through three market-based mecha-nisms: international emissions trading (IET), jointimplementation (JI) projects, and Clean Develop-ment Mechanism (CDM). IET involves trading ofallowances among countries as described above; JIand CDM are means for countries to earn allow-

Paper & pulp

Nonmetallicminerals

Chemicals

Othermanufacturing

Primary metals

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

50%

1973

fina

l ene

rgy

inte

nsity

= 1

00%

60%

70%

80%

90%

100%

110%

Source: UNDP 2004

Figure 6.12. The evolution of industrial subsector energy intensities, 1974–1998

102 José Goldemberg

ances by reducing emissions in other countries, asexplained later in this chapter. The second com-mitment period of the Kyoto Protocol, togetherwith a long-term cooperative action under theUNFCCC, were discussed at the end of 2009 inthe 15th Conference of the Parties to the ClimateConvention in Copenhagen but no general agree-ment of the type reached in Kyoto was achieved.It was decided instead that the industrializedcountries, including the United States, commit toimplement individually or jointly quantifiedeconomy wide emissions targets for 2020 to besubmitted to Secretariat of the Convention byJanuary 31, 2010. The developing countrieswould present voluntary non-mandatory pledgeswhich will be subject to measurement, reportingand verification.

The European UnionEmissions Trading System

Under the Kyoto Protocol, the European Union(EU) committed to an 8% reduction of emissionsrelative to the 1990 level by 2012 (EU 2003).4

The responsibility for meeting this commitment isshared among member states, based on agreednational allocations. Pursuant to this, the EU hasset a target for a reduction in emissions of GHGsof at least 20% of 1990 levels by 2020. Central toits strategy for achieving this target is the EUEmissions Trading System (ETS), the largest mul-tinational emissions trading scheme in the world.

Under the ETS, member states agree tolegally binding national emissions targets with theEuropean Commission. Each country allocates itsnational emissions target among the companies inthose emissions-producing sectors of theeconomy covered by the scheme. Currently, theseallowances are distributed to companies by mem-ber governments free of charge. In the first phaseof operation, the scheme included more than10,000 industrial plants, comprising steel facto-ries, power plants, oil refineries, paper mills, andglass and cement installations, which accountedfor almost half of EU CO2 emissions, 2.4 billiontons of CO2 equivalent (CO2e) (European Coun-cil 2009). The emissions allowances could betraded to allow abatement to occur at the lowest

cost.The total number of allowances is reduced ineach year of the scheme in order to achieve thedesired overall reductions in emissions.

The scheme, which initially covered the 15then-existing member states, began its first tradingperiod on January 1, 2005. The first tradingperiod lasted until the end of 2007 and was pri-marily intended to ensure that the ETS was func-tional by the start of the first commitment periodunder the Kyoto Protocol in 2008. The secondtrading period commenced in January 2008 andwill end in December 2012, coinciding with theend of the first commitment period of the KyotoProtocol. Table 6.4 shows the volume of CO2

equivalent and volumes (in dollars) which wereobject of transaction in 2006 and 2007.

As shown in Figure 6.13, the price of allow-ances fell dramatically twice during the firstperiod. The first fall occurred following the peakin April 2006, when it was discovered that somecountries had overallocated allowances to indus-try, creating a substantial excess supply of allow-ances. Subsequently, prices faced another dra-matic decline to approximately €1.20 ($1.63) perton in March 2007 and later fell to around 10 eurocents (14 U.S. cents) in September 2007, becauseexisting allowances expired at the end of the firstphase and could not be carried over into subse-quent periods. The initial overallocation of allow-ances dampened the incentives to abate and cre-ated a windfall for many industrial companies.The subsequent collapse in the price of allow-ances reduced the incentives for abatement.

The collapse of prices at the end of the firstphase did not directly affect prices for contractsfor 2008, the first year of the second phase. Mar-ket participants knew already in 2007 that thisphase would be more stringent in regard to thecap and less lenient toward allowances, whichexplains the high prices for 2008 allowances.

In April 2009, the European Commissionannounced a number of changes to the schemethat will be effective from the third tradingperiod, commencing in January 2013 (EuropeanCouncil 2009):

• Member states will introduce auctioning ofallowances, which will be introduced gradu-

Renewable Energy, Energy Efficiency, and EmissionsTrading 103

ally into most sectors, with plans to auctionall allowances by 2020. Following the pri-mary auction, allowances can be tradedamong companies as before.

• Nitrous oxides and perfluorocarbons will beadded to the list of GHGs covered by thescheme.

• The scheme will be extended to include theaviation industry.

• The number of credits that can be earnedthrough CDM and JI activity will be limitedin order to motivate member states to reduceemissions domestically.

The proposed caps for the third trading periodtarget an overall reduction of greenhouse gases forthe sector of 21% in 2020 relative to 2005 emis-sions (European Council 2009). In addition, thethird trading period will be both more economi-cally efficient and environmentally effective. Itwill be more efficient because trading periods willbe longer (eight years instead of five) and have asubstantial increase in the amount of auctioning(from less than 4% in the second phase to morethan half in the third). Environmental effective-ness will be guaranteed by a robust and annuallydeclining emissions cap (21% reduction in 2020compared with 2005) and a centralized allocation

process within the European Commission. Figure6.14 shows the cap on emissions set by the EU aswell as actual (and future) emissions.

A robust secondary market for carbon certifi-cates exists through which investors bank on thefuture value of the ETS certificates changingmany times. The ETS does not include transport,however, and thus this action is limited to theindustrial process and energy sectors.

Joint Implementation

Joint implementation (JI) is one of the flexibilitymechanisms included in the Kyoto Protocol as ameans for developed nations with emissionsreduction commitments (Annex I countries) tomeet their obligations at the lowest possible cost.5

JI allows an Annex I country to reduce its domes-tic emissions reduction requirements by under-taking projects to reduce emissions in anotherAnnex I country.This allows Annex I countries toreduce their cost of emissions reduction.

A key requirement for a JI project is that itmust provide an emissions reduction, or a removalof carbon from the atmosphere that is demonstra-bly additional to what otherwise would haveoccurred. This is known as “additionality.” If aproject can be shown to be additional and meet

35.00

30.00

25.00

20.00

15.00

10.00

5.00

0.00Jan

2006Mar2006

May2006

Jul2006

Sep2006

Nov2006

Jan2007

Mar2007

May2007

Jul2007

Sep2007

Nov2007

Pric

e pe

r to

n C

O2e

( )

Source: ECX 2009Note: €1 = $1.36 as of this writing

Figure 6.13. Carbon price, January 2006–November 2007 (settlement price of December 2007 ECX futures)

104 José Goldemberg

other requirements set out by the UNFCCC, theproject owner can earn credits called emissionreduction units (ERUs) from the host country.Each ERU represents an emissions reductionequivalent to 1 ton of CO2 and can be set againstthe domestic emissions reduction requirement.

After a long preparatory process, JI projectsbegan to take shape, and as of March 2010, 207projects had been submitted. If all were imple-mented, they would lead to an emissions reduc-tion of 363,355 million tons of CO2 equivalent inthe period 2008–2012. The great majority of theprojects are in the Russian Federation and easternEuropean countries.

So far very few ERUs have been issued.

Clean Development Mechanism

The Clean Development Mechanism (CDM) isanother flexibility mechanism included in theKyoto Protocol.6 Under the CDM, Annex Icountries can reduce their domestic emissionsreduction requirements by undertaking projectsto reduce emissions in developing countries. Thisallows Annex I countries to reduce their cost ofemissions reduction, and global emissions to bereduced at a much lower global cost, by financingemissions reduction projects in developing coun-tries where costs are lower than in industrializedcountries. Furthermore, this encourages technol-ogy transfer and stimulates investment in sustain-able development in less developed countries.

The CDM is administered by a designatednational authority (DNA) in the country wherethe project takes place and is supervised by theCDM executive board at the international level.For projects to be approved, they must demon-strate additionality and meet the other require-ments set forth by the UNFCCC. If a project isapproved, the project owner is granted a certifiedemission reduction (CER), each of which repre-sents an emissions reduction equivalent to 1 ton ofCO2. This can be counted toward the domesticemissions reduction target.

By June 1, 2009, 4,417 projects had been sub-mitted; if all implemented, these would corre-spond to 2,931,813 million tons of CO2e, repre-senting roughly 1% of the total necessary effort tocurb GHG emissions until 2050. (Roughly 75%of the CDM projects are in China.) Table 6.6shows the number of CDM projects of each type,along with the 2012 CERs granted.

The distribution of CDM projects from dif-ferent sectors are shown in Figure 6.15.

In contrast to emissions trading schemes,which are actively traded in the stock market, JIand CDM are project-based transactions. Annualvolumes for 2006–2007 are given in Table 6.7.

Figure 6.16 gives estimates of the totalamount of emissions transactions of the EU ETS,CDM, and JI for the years 2002 to 2008.

3,000

Em

issi

on

s (t

on

s)

2,500

2,000

1,500

1,000

500

2,159 2,183 2,130 2,080 2,0612,203

02005 2006 2007 2008 2009 2010

Historic Projected Cap

Source: New Energy Finance 2009a

Figure 6.14. Cap and ETS emissions in the European Union

Renewable Energy, Energy Efficiency, and EmissionsTrading 105

Table 6.5. Distribution of JI projects by type

Type Number 2012 ERUs (000)

HFCs, PFCs & N2O reduction 31 13% 96,223 27%

CH4 reduction & cement & coal mine/bed 115 49% 166,302 46%

Renewables 78 33% 32,058 9%

Energy efficiecncy 54 23% 71,053 20%

Fuel switch 9 4% 10,896 3%

Afforestation & reforestation 1 0% 410 0%

Total 233 100% 362,355 100%

Source: UNEP RISO 2010

Table 6.6. CDM projects grouped by type

Type Number 2012 CERs (000)

HFCs, PFCs & N2O reduction 106 2.1% 741,442 26%

Renewables 2,959 60% 980,433 35%

CH4 reduction & cement & coal mine/bed 989 20% 576,142 20%

Supply-side energy efficiency 540 11% 375,976 11%

Fuel switch 111 2.2% 172,894 6.1%

Demand-side energy efficiency 190 3.8% 25,420 0.9%

Afforestation & reforestation 52 1.0% 15,224 0.5%

Transport 21 0.4% 8,076 0.3%

Total 4,968 100% 2,835,607 100%

Source: UNEP RISO 2010

Table 6.7. Volumes and values for CDM and JI, 2006–2007

2006 2007

Volume(MtCO2e)

Value($ million)

Volume(MtCO2e)

Value($ million)

Project-based transactions

Primary CDM 537 5,804 551 7,426

Secondary CDM 25 445 240 5,451

JI 16 141 41 499

Other compliance and voluntarytransactions

33 146 42 265

Total 611 6,536 874 13,641

Source: World Bank 2008

106 José Goldemberg

Stimulus PackagesA significant amount of the stimulus packagesadopted by a number of governments to face thefinancial crisis of 2007–2008 consists of invest-ments in green activities. These amount to 6% ofthe total recovery packages announced by govern-ments. Figure 6.17 shows how much 12 differentcountries invested in green stimuli. The estimatedamounts announced by these economies total$184.9 billion invested in green stimuli. China

and the United States are the leaders, in nominalterms, of the green stimuli activities, earmarking$68.7 billion and $66.6 billion, respectively.

The sector breakdown in Figure 6.18 showsthat energy efficiency remains at the heart of thelow-carbon fiscal stimuli. Accounting for as muchas 36% of the total $184.9 billion, this sector willreceive a boost of some $65.7 billion globally,mainly via building efficiency projects. In addi-tion, $7.9 billion has been announced for researchand development in energy efficiency.The secondmajor winner is electricity grid infrastructure.More than $48.7 billion has been earmarked forgrid development and upgrade, accounting forsome 26% of the total funds.

The U.S. Department of Energy has alreadydisbursed $41.9 million in grants for fuel cellenergy projects. Furthermore, $101.5 million hasbeen directed to wind energy research, anddetailed plans have been disclosed on $2.4 billionto be spent on carbon capture and storage and $4billion for grid upgrades. Out of a total of $2billion to support energy science research, thebreakdown of almost $1.3 billion have also beenconfirmed, and only some $725 million remainsto be allocated.

Source: UNEP RISO 2010

Figure 6.15. Distribution of CDM projects by type

EU ETS Primary CDM Secondary CDM Other140

120

100

80

60

40

20

02004

$1bn

2005

$11bn

$31bn

8

2006

24

6

2007

50

75

$64bn

$118bn

146

94

2008

$

bn

Source: New Energy Finance 2009b

Figure 6.16. Carbon market estimates, 2002–2008

Renewable Energy, Energy Efficiency, and EmissionsTrading 107

Notes

1. “Traditional biomass” is used to denote locallycollected and often unprocessed biomass-basedfuels, such as crop residues, wood, and animaldung, most of which are used noncommercially.Sometimes these lead to deforestation and there-fore are not renewable. “Modern biomass” refersto biomass produced in a sustainable way and usedfor electricity generation, heat production, andtransportation (liquid fuels). International EnergyAgency (IEA) statistics lump together traditionaland modern biomass under the heading of “com-bustible renewables and waste” (CRW).

2. Average conversion efficiency: heat, 85%; biomasselectricity, 22%; combined heat and power, 80%;geothermal electricity, 10%; all others, 100%.

3. The main sources of distributed generation arereciprocating engines (internal combustionengines), gas turbines, microturbines, fuel cells,photovoltaic systems, thermoelectric solar plants,wind energy conversion systems (WECSs),biomass to energy power plants, and small hydro-electric power plants (SHPPs).

4. Article 17 of the Kyoto Protocol.The Conferenceof the Parties shall define the relevant principles,modalities, rules and guidelines, in particular forverification, reporting and accountability for

China

U.S.

EU27

Japan

S Korea

Spain

Germany

Australia

UK

France

Brazil

Canada

68.7

66.6

11.3

8.0

7.7

7.6

3.7

3.4

2.7

2.5

1.9

0.8

Source: ECX 2009

Figure 6.17. Investments in green stimuli ($ billion)

Grid26%

Efficiency35%

All renewables19%

Unspecified5%

R&D12%

Transportation3%

Source: New Energy Finance 2009b

Figure 6.18. Sector breakdown of global green stimuli

108 José Goldemberg

emissions trading. The parties included in AnnexB may participate in emissions trading for the pur-poses of fulfilling their commitments under Article3. Any such trading shall be supplemental todomestic actions for the purpose of meeting quan-tified emissions limitation and reduction commit-ments under that Article.

5. Article 6 of the Kyoto Protocol. For the purposeof meeting its commitments under Article 3, anyParty included in Annex I may transfer to, oracquire from, any other such Party emissionsreduction units resulting from projects aimed atreducing anthropogenic emissions by sources orenhancing anthropogenic removals by sinks ofgreenhouse gases in any sector of the economy,provided that: (a) any such project has the approvalof the Parties involved; (b) any such project pro-vides a reduction in emissions by sources, or anenhancement of removals by sinks, that is addi-tional to any that would otherwise occur; (c) itdoes not acquire any emission reduction units if itis not in compliance with its obligations underArticles 5 and 7; and (d) the acquisition of emis-sion reduction units shall be supplemental todomestic actions for the purposes of meetingcommitments under Article 3.

6. Article 12 of the Kyoto Protocol. (1) A CleanDevelopment Mechanism (CDM) is herebydefined. (2)The purpose of the clean developmentmechanism shall be to assist Parties not included inAnnex I in achieving sustainable development andin contributing to the ultimate objective of theConvention and to assist Parties included in AnnexI in achieving compliance with their quantifiedemission limitation and reduction commitmentsunder Article 3. (3) Under the clean developmentmechanism: (a) Parties not included in Annex Iwill benefit from project activities resulting in cer-tified emission reductions; and (b) Parties includedin Annex I may use the certified emission reduc-tions accruing from such project activities to con-tribute to compliance with part of their quantifiedemission limitation and reduction commitmentsunder Article 3, as determined by the Conferenceof the Parties to this Protocol.

ReferencesBayod-Rújula, Angel A. 2009. Future Development of

the Electricity Systems with Distributed Generation.Energy 34: 377–383.

CDIAC (Carbon Dioxide Information Analysis Center,Oak Ridge National Laboratory). 2009. Fossil-FuelCO2 emissions, cdiac.ornl.gov (accessed January 102010).

Cossent, Rafael, Tomás Gómez, and Pablo Frías. 2009.Towards a Future with Large Penetration of Distrib-uted Generation: Is the Current Regulation of Elec-tricity Distribution Ready? Regulatory Recom-mendations under a European Perspective. EnergyPolicy 37: 1145–1155.

ECX (European Climate Exchange). 2009. ECX EUAFutures Prices. www.ecx.eu/ (accessed December21, 2009).

EPA (U.S. Environmental Protection Agency). 2009.Acid Rain Program 2007 Progress Report.www.epa.gov/airmarkets/progress/arp07.html(accessed January 10, 2010).

European Council. 2009. Directive 2009/29/EC of theEuropean Council. Press release. ec.europa.eu/environment/climat/emission/ets_post2012_en.htm (accessed December 21,2009).

EU (European Union). 2003. Kyoto Protocol.europa.eu/rapid/pressReleasesAction.do?reference=MEMO/03/154&format=HTML&aged=0&language=EN&guiLanguage=en (accessedDecember 15, 2009).

New Energy Finance. 2009a. EU ETS – Analyst Reac-tion (April 3, 2009). Available atcarbon.newenergyfinance.com/ (accessed April 3,2009).

———. 2009b. Carbon Industry Intelligence ResearchNote. October 2008. Available atcarbon.newenergyfinance.com/ (accessed May 22,2009).

OECD/IEA (Organisation for Economic Co-operation and Development/International EnergyAgency). 2004. Oil Crises and Climate Challenges:30 Years of Energy Use in IEA Countries.www.iea.org/textbase/nppdf/free/2004/30years.pdf(accessed December 21, 2010).

———. 2007. CO2 Emissions from Fuel Combustion,1971–2005. Paris: OECD Publishing.

———. 2008a. Energy Technology Perspectives. In Supportof the G8 Plan of Action. Scenarios & Strategies to 2050.Paris: OECD/IEA Publishing.

———. 2008b. WorldwideTrends in Energy Use and Effi-ciency. Key Insights from IEA Indicator Analysis. In Sup-port of the G8 Plan ofAction. Paris: OECD Publishing.

REN21 (Renewable Energy Policy Network for the21st Century). 2007. Renewables 2007: Global Sta-

Renewable Energy, Energy Efficiency, and EmissionsTrading 109

tus Report. www.ren21.net/pdf/RE2007_Global_Status_Report.pdf (accessedDecember 10, 2010).

———. 2009. Renewables Global Status Report: 2009Update. www.ren21.net/pdf/RE_GSR_2009_Update.pdf (accessed December10, 2010).

UNDP (United Nations Development Programme).2000. World Energy Assessment: Energy and the Chal-lenge of Sustainability. NewYork: UNDP.

———. 2004. World Energy Assessment: Overview 2004Update. NewYork: UNDP

UNEP RISO. (2010). CDM/JI Pipeline Analysis andDatabase http://cdmpipeline.org/cdm-projects-type.htm (accessed January 11, 2010).

UNFCCC (United Nations Framework Conventionon Climate Change). 2009a. Greenhouse Gas Inven-tory Data. unfccc.int/ghg_data/items/3800.php(accessed January 11, 2010).

———. 2009b. Kyoto Protocol. unfccc.int/kyoto_protocol/items/2830.php (accessed January11, 2010).

World Bank. 2008. State and Trends of the CarbonMarket 2008. siteresources.worldbank.org/NEWS/Resources/State&Trendsformatted06May10pm.pdf(accessed January 11, 2010).

110 José Goldemberg

Part III

Renewable Generation and ElectricPower Markets

7

Electricity Wholesale Market Designin a Low-Carbon FutureWilliam W. Hogan

Policy to convert the energy system to a low-carbon configuration involves a central focus

on the electricity sector. Electricity generation is amajor use of fossil fuels and therefore accounts fora large fraction of the emissions of carbon dioxide;reducing carbon emissions in other sectors, suchas transportation, may involve increased electrifi-cation; and electricity can be and has been gener-ated without carbon emissions. In addition, elec-tricity generating facilities burning fossil fuels arelarge and easy to identify and thus a natural focusfor regulation and policy to reduce carbon emis-sions.

Designing policy for a low-carbon electricitysector depends on the nature of the technologyand institutions in that sector. In many countries,the electric power system is already involved in amajor restructuring process. Critical technologi-cal and institutional features of that reform processcenter on wholesale market design elements thathave important implications for expansion of reli-ance on low-carbon energy sources. The whole-sale market covers bulk generation, dispatch, andtransport of electric power for ultimate delivery toretail customers. Although the details and jurisdic-tional rules differ in the United States and theEuropean Union (EU), a common distinctive fea-ture of the wholesale market is the interactionthrough the interconnected transmission grids

that transcend states and countries. The comple-mentary rules for retail supply or distribution ofelectricity are important, but the emphasis here ison electricity wholesale market design in a low-carbon future. The main theme emphasizes thestrong interaction between market design andpolicies for achieving a low-carbon future.

Electricity System FundamentalsAll electricity systems share certain fundamentalcharacteristics. Efficient power generation ben-efits from substantial economies of scale. Theresult of more than a century of evolution ofsophisticated power systems is a fleet of relativelylarge generating plants.

To enjoy the benefits of these economies ofscale, traditional power plants have not beenlocated near the ultimate load. Moving powerfrom the generating plants to the ultimate con-sumer requires high-voltage transmission lines.Because of the nature of electricity transmission,the higher the voltage, the lower the required cur-rent and the resulting transmission losses. Trans-formers convert the relatively low voltage at thepoint of generation to allow for high-voltagetransmission to another set of transformers that

lower the voltage for retail distribution to facto-ries, businesses, and homes.

Typically the local retail distribution system issimpler in its design than the transmission system.In particular, the distribution network is moredecentralized or radial in the sense that electricalpower problems in one distribution network, theprincipal cause of power outages, can be isolatedfrom the rest of the system.

The interconnected transmission network isanother story. The transmission network initiallydeveloped organically as individual companiessought to connect their generators to their loads.One of the features of the electric power system isthe speed of adjustment when a change occurs inload, generation, or transmission connections.With present technology and little storage, to afirst approximation, electricity is the limiting caseof same-time production. The power producedand consumed must maintain instantaneous bal-ance throughout the system, and any disruption ofthe transmission grid results in instantaneousredistribution of the flow of power in the system.The system is not controllable in the same ways aswe imagine for a network of pipelines with valvesand relatively slow-moving fluids. Power flows atthe speed of light.

Companies building bulk transmission facili-ties consequently relied on redundant pathways inthe transmission grid to ensure reliability. In addi-tion, the power system required standby generat-ing capacity that could be called on immediatelyin order to replace a lost generating unit. Asneighboring electricity companies would beunlikely to face disruptions at the same instant,redundant transmission and generation in one sys-tem could help other systems and lower the totalcost of maintaining reliability. This confluence oftechnology and economics led to the gradualdevelopment of interconnected alternating cur-rent (AC) transmission grids linking loads tothousands of generators through tens of thousandsof individual transmission links.

Throughout such an interconnected grid, allof the generators and load must be synchronized.The resulting ensemble has been called the “lar-gest machine in the world” (Amin 2000, 264).For example, the North American continent from

Mexico to Canada is dominated by only threesynchronous grids in the Eastern, Western, andTexas Interconnections, operating with limitedexchange through direct current (DC) intercon-nections between them. (Alaska and Quebec havetheir own separated grids as well.) Similarly, alarge synchronized grid covers 24 Europeancountries (ENTSO-E 2008).

The complexity of a large synchronized gridis well known to electricity system operators butnot most customers. Traditionally, the electricitycompany was vertically integrated from genera-tion through transmission to distribution. Cus-tomers faced a monopoly that provided the powerand charged a price only loosely coupled to actualcontemporaneous costs. The company made itsown investment plans for generation and wires,often through joint investments with other verti-cally integrated monopolies. Control of the inter-connected transmission grid was handled throughan array of bilateral and multilateral agreementsamong the members of the “club,” but with strictexclusion of potential new competitors. Theseclub agreements worked reasonably well,although below the surface, difficult problemsoften occurred at the seams between regions orwhen one member of the group was “leaning”too much on the system.

The club agreements involved coordinatedplanning and expansion of the transmission grid,always an opaque and arcane process. It was diffi-cult to define the resulting rights to transmitpower and determine how these rights would beallocated in the interconnected grid, and real-time dispatch of power plants had to be coordi-nated in order to maintain the grid’s reliability.

The rules and incentives for operation of thissystem were complicated by the nature of elec-tricity regulation and its focus on cost recovery.The emphasis of regulation or government own-ership was on control of the average costs, withrelatively little attention paid to the role of mar-ginal costs and the connection to incentives thatcould or should govern operations and efficientinvestment. This history would be a major hurdleto overcome in the development of electricitymarket reforms.

114 WilliamW. Hogan

Economic efficiency always played a second-ary role to maintaining system reliability, keepingthe lights on. Gold plating the system a littlemight produce some inconvenient questions onoccasion, but major blackouts or even relativelyminor but frequent local supply interruptionscould limit a career and were deemed unaccept-able. The common approach for the wholesalepower system was, and continues to be, treatingreliability as a constraint and economic efficiencyas a goal to be sought subject to that constraint.(Because of the lower availability of redundancyand greater expense of reliability, a more nuancedtrade-off is seen at the retail distribution level.)

In real time, the variable cost of generatingpower ranges from virtually zero for run-of-the-river hydropower, to moderate for nuclear andcoal plants, to expensive for natural gas and oilplants. With load changing rapidly over the dayand total generation at the trough during thenight at as little as 30% of the peak during the day,it makes sense to choose the cheapest generatorsavailable to produce at any time (MonitoringAnalytics 2009).

The basic idea is to operate as closely as possi-ble to the security-constrained (respecting all themany reliability constraints on the grid) economicdispatch that provides the necessary ancillary serv-ices and minimizes the cost of meeting a givenelectric load at any moment and over the courseof the day. Under the traditional electricitymodel, the aggregate cost of this economic dis-patch would be recovered through preset cus-tomer charges that muted or completely ignoredthe dramatic changes in opportunity costs overthe day and season. As a result, the traditional sys-tem often provided adequate cost recovery but didlittle to reveal or communicate incentives to guideeither operations or investment. Rather, invest-ment decisions relied on and defaulted to theexpert assessments of the established electricitycompanies and left few opportunities for innova-tion by new entrants.

Although the operating decisions were nei-ther open nor transparent, a reasonable case canbe made that traditional reliance on the frame-work of security-constrained economic dispatchproduced efficient operating decisions and cer-

tainly contributed to the reliability of the system.The signals sent for investment and innovationwere an entirely different matter, however. Thesystem depended on the planning decisions of theprotected monopolies, and the lack of transpar-ency coupled with major construction delays andcost increases precipitated deep concern withinvestment choices.

Electricity Market DesignAs long as the natural benefits of economies ofscale produced steady reductions in the real andnominal average price of electricity, as was truefor decades, the vertically integrated model withfranchise monopolies worked well. Rapid growthin electricity demand hid mistakes. Average costswere declining, and the big concern was to investrapidly enough to stay ahead of growing demand.However, this system began to break down whenthe returns to increasing scale declined, electricitygrowth patterns changed, and mistakes were madethat resulted in large cost overruns. The long-running good news that new investment wouldlower average electricity costs was soon replacedwith the bad news that new investment was moreexpensive and would raise average costs (Hogan2009).

Electricity Market Reform

These changes accompanied a general interest inreducing reliance on regulation and turning moreto market forces to guide investment and opera-tions. For example, in the 1980s, the UnitedStates eliminated regulation of natural gas produc-tion, and a great boom of activity and improvedeconomic efficiency ensued. Policymakers in theUnited Kingdom and Chile turned attention tosimilar reforms of electricity systems. In theUnited States, successful experiments with lim-ited introduction of generation investment bycompanies outside the monopoly club had dem-onstrated that new entrants could be accommo-dated without compromising efficiency or reli-ability. In 1992, the United States adopted the

ElectricityWholesale Market Design in a Low-Carbon Future 115

Energy Policy Act (EPAct92), which includedprovisions for open access to the transmissiongrid.

Authorities in Australia, Argentina, most ofEurope, New Zealand, the United States, andelsewhere were experimenting with new struc-tures for organizing the electricity system. Manyrecognized or assumed that economies of scale ingeneration had largely been exhausted, and newand more loosely regulated entrants were just ascapable of building and operating successfulpower plants as had been the incumbent electricutilities. This precipitated the notion that verticalseparation of generation, transmission, low-voltage distribution, and supply would be at leastpossible and perhaps desirable. Existing genera-tors, perhaps divided and divested from existingelectric companies, could compete against eachother and against new entrants to construct powerplants and sell electricity to others, including finalcustomers. Local distribution facilities wouldcontinue to operate more or less under the oldregulated monopoly model, albeit often as awires-only business with suppliers acquiring elec-tricity from generators and selling on to final cus-tomers.

A key requirement for this vertical separationand competition in generation would be openaccess to the common integrated transmissiongrid and associated unbundling of critical services.The best form of organization for the commoninterconnected grid with the unbundling of serv-ices was not immediately apparent. In the UnitedStates, the emphasis was soon on transmissionopen access and nondiscrimination in prices andservices for all participants in the electricity mar-ket. As it turned out, these seemingly innocuousrequirements had profound implications for thenature of electricity market design.

The details of the story and the alternativemarket approaches are many, and an extensive lit-erature exists on alternative market designs(Sioshansi 2008).The purpose here is not to recitethese details, but to emphasize certain featuresthat are salient for the discussion of a low-carbonelectricity system that relies heavily on technolo-gies such as solar, wind, and demand-side man-

agement, which differ substantially from thelegacy portfolio of fossil-fuel generation.

Alternative Market Designs

For this purpose, electricity system market designscan be divided into three categories. First are thelegacy systems that continue to be organizedaround the model of vertical integration andclosed transmission access.1 This would includemany of the Southeast regulated utilities andNorthwest municipal and public power systems inthe United States. Second are the organizedcountertrade markets with aggregate zonal trans-mission systems that support sometimes separatetrading platforms and scheduling procedures thatare only partially integrated with system opera-tions. The countertrade refers to the activities ofthe transmission system operator, who must con-duct offsetting transactions that undo the sched-ules of market participants to bring the net sched-ules in line with the capabilities of the transmis-sion system. This describes many systems,including most of the electricity markets in theEU. Third are the organized spot markets underindependent system operators (ISOs) that coordi-nate dispatch but do not own the grid and inte-grate real-time trading and dispatch with systemoperations. This integration includes the fulllocational granularity of the actual grid for deter-mining energy flows and locational prices. Thiswould include the multistate regional transmissionorganizations (RTOs) or equivalent single-stateISOs that together cover approximately two-thirds of the U.S. load.

To a significant degree, the vertically inte-grated systems are similar to those of the past, andwe can think about how regulated utilities wouldbe likely to adapt to a low-carbon future. Thecountertrade systems include many market inno-vations and allow a great deal of flexibility in trad-ing arrangements, operations, and investmentincentives. At their core, however, these marketsare incompatible with the requirements of systemoperations. The aggregate trading arrangementsignore many of the relevant constraints on thegrid. The associated nominal schedules are thusoften not collectively feasible, and countertrade is

116 WilliamW. Hogan

necessary to undo what the nominal market hasdone. Countertrade, or some other systemredispatch intervention like it, is necessary when-ever the market schedules do not reflect systemoperation reality. Furthermore, workablecountertrade market designs and interventions arenecessarily discriminatory in selectively choosingand paying for redispatch of some generators butnot others similarly situated, as shown by experi-ence in the United States (Hogan 2002). Thiscountertrade design has implications for the treat-ment of renewables and other low-carbon energysources. In order to understand the implications,it is convenient first to sketch out the essentialfeatures of the RTOs and organized spot marketsthat integrate market design and system opera-tions.

Integrated Locational ElectricityMarket Design

The U.S. RTOs have regular market monitoringorganizations that prepare extensive evaluations ofthe design and operation of their markets (seeHEPG, various dates). The critical design crite-rion is to structure the spot market products andservices as much as possible to conform to actualsystem operations, and then price those productsand services to reflect the marginal costs thatwould drive market clearing competitive prices.

The initial focus is on the spot market. Theforward markets play a critical role, but forwardmarket participants will look ahead to the antici-pated operation of spot markets. Thus efficientdesign of forward markets depends in significantmeasure on a good design for spot markets. Fur-ther, a focus on competitive market design is notintended to overlook the possibility of the exist-ence and exercise of market power. Rather, agood competitive spot market design greatly sim-plifies possibly necessary regulatory interventionto mitigate market power. For electricity systems,a necessary condition for good market design isgood design of the spot market.

When closely tied to system operations, goodspot market design can be integrated with bilateralor multilateral trading.The organized spot marketdoes not preclude decentralized trading. Transac-

tions through the spot market can be gross, with-out separate bilateral transactions and schedules,or net, covering only imbalances relative to bilat-eral transactions and schedules.The differences arelargely semantic, despite heated arguments in thepast for the merits of gross or net market design.

The principal market design problem thatarises in all electricity systems follows from thenature of the interconnected transmission grid. Inaddition to being large, interconnected, fastresponding, and contingency constrained, elec-tricity transmission grids have an unusual featurethat follows from the need for fast response. Theredundancies and contingency constraints arisebecause it is not possible to use valves to controlthe flow of power. To a good first approximation,once the pattern of load and generation is set, theflow of power through the grid is determined byphysical laws that distribute the power over thesystem to (more or less) minimize losses. It turnsout that power flows to varying degrees on everyavailable path between generation and load. Thismeans that control of power plant dispatch andcontrol of transmission flows are two sides of thesame coin. Changing the dispatch changes thepower flows, and changing the power flowsrequires changing the dispatch.

This technological fact has profound implica-tions. For example, it means that it is impossibleto operate the system with only decentralizeddecisions about generation and load. There mustbe central coordination of everything and centralcontrol of enough of the dispatch to meet therequirements of system operations. The parallelflow of power means that any trading system mustconfront material externalities as the real flows forone transaction interfere with the real flows ofother transactions. In the United States, the elec-tric utilities struggled for decades to design aworkable system of transmission flow accountingthat would support physical transmissions rights,and never succeeded in working around theinconvenient fact that power flows everywhere(Lankford et al. 1996).

The design solution to this otherwise intract-able problem is to recognize that the frameworkof security-constrained economic dispatch,already familiar to system operators, provides the

ElectricityWholesale Market Design in a Low-Carbon Future 117

foundation for a coordinated spot market(Schweppe et al. 1988). This spot market inher-ently respects the requirements of system opera-tions and provides a convenient connection toelectricity markets and forward trading.

Recognizing that there must be a systemoperator, this design asks for bids and offers forpower at the physical locations, schedulesbetween locations, or schedules with bids andoffers for deviations from the schedules. Thus thestructure of the information provided to the sys-tem operator is the same as under a vertically inte-grated system, only now the engineering esti-mates of loads and costs are replaced by the bidsand offers for power.

As before, the system operator selects thechanging pattern of economic dispatch to meetload and provide ancillary services subject to the(many) security constraints and the complex flowsof power on the transmission grid. In economicterms, the independent system operator internal-izes the major transmission externalities so thattransmission and transaction schedules can bepoint-to-point and be silent on the path betweenthe points.

The quantity result of the security-constrained economic dispatch for any interval isa schedule of generation and load at each electri-cal location. Hand in hand with the quantity dis-patch is a set of market-clearing prices for settle-ment purposes that capture the system marginalcost of meeting increased load or decreased gen-eration at each location. The term of art islocational marginal pricing (LMP). Theselocational marginal prices provide an immediatedefinition of the appropriate spot price of trans-mission between any two locations: a straightfor-ward arbitrage argument dictates that the com-petitive spot price of transmission between anytwo locations is the difference in the energy spotprices at the locations.

When there is no congestion in the system,the market clearing prices would be the same asthe normal single market clearing price of simplermodels, except for differences in marginal trans-mission losses. But when transmission constraintsare binding, the locational prices differ accordingto the marginal impacts of activity at each location

on system congestion. These congestion-induceddifferences can be surprisingly large and some-times counterintuitive. Simultaneous marketclearing prices can often be well above the mar-ginal cost of the most expensive generation run-ning at some locations and well below marginalcost or even negative elsewhere (Hogan 1999).This is a material effect of the complicated electri-cal interactions made visible in the spot marketbut previously hidden in system operations.

Once this security-constrained economic dis-patch framework is adopted, with accompanyinglocational prices applied in market settlements, itis a simple matter for market participants toschedule or otherwise arrange transactions. Forexample, a bilateral transaction can be scheduledbetween any two locations, for a charge at thedifference in locational prices. Any imbalances inactual delivery would be paid for at the relevantlocational prices. With security-constrained eco-nomic dispatch and locational prices, the imbal-ance market merges with the spot market. All par-ties can participate, so there is open access. And allparties face the same scheduling rules and result-ing prices, so there is nondiscrimination. No pro-vision exists for countertrading through marketpositions taken by the system operator, becausecountertrades are not needed to bring the spotmarket into conformance with system operations.This results in less concern over conflicts of inter-est for the system operator. By construction,security-constrained economic dispatch is con-sistent with system operations and the spot mar-ket.

In addition, the existence of the spot marketand locational prices provides the foundation forthe financial transmission right (FTR) (Hogan1992). In the simplest form, an FTR calls for pay-ing to the right holder the (congestion) differencebetween the corresponding locational prices.Absent grid outages, the FTR is a long-terminstrument that provides a perfect hedge for theshort-term congestion fluctuations in transmis-sion spot prices. With a matching FTR betweengeneration and load, a supplier can sign a fixedprice contract for delivered power and be sure thatthe system will always allow fulfillment of thecontract through either physical delivery or offset-

118 WilliamW. Hogan

ting purchases and sales with no incrementalcharges for volatile transmission congestion. Thisworkable financial transmission right provides thefunctionality sought without success in unwork-able systems of physical transmission rights.

Concerns early on were that this integratedlocational model could not work well with verymany locations. The United States went throughthe agony of forcing aggregations across zones toreduce the granularity of the market representa-tion. Zonal aggregation subject to rules of openaccess and nondiscrimination was tried and aban-doned in the PJM system, a large RTO that coversthe Mid-Atlantic states in the United States, thecorresponding New England RTO, the CaliforniaRTO, and theTexas RTO (Hogan 2002; PotomacEconomics 2009). This zonal aggregation inher-ently suppressed the detail of real system opera-tions and thereby lost the efficient pricing signals,which in turn required complex rules to bridgethe constructed gap between the spot market andsystem operations. Painful experience made itclear that the simplest system is to go as far aspossible to match temporal and locational granu-larity with real operations and the associated sys-tem opportunity costs. The integrated locationaldesign with very many locations works. Forexample, in the Mid-Atlantic states, PJM updatesthe coordinated dispatch and prices every fiveminutes for approximately 8,000 locations. Thisgranularity means that geographically proximatelocations that are electrically distant can have verydifferent prices, and these prices reinforce ratherthan contravene the imperatives of efficient andreliable system operations. One of the regularreports from system operators newly adopting thismodel is that it greatly simplifies and improvesdispatch operations, because market participantshave strong economic incentives to follow dis-patch instructions.

Some contend that a more geographicallyaggregated system would be easier and cheaper toimplement.This might be correct if an alternativewere available that could avoid the need to attendto all the detail in real operations.2 No such alter-native is available, however. If the system operatoris applying the basics of security-constrained eco-nomic dispatch, then all the essential information

on transmission lines and multiple locations mustalready be available. The distinctive feature of theintegrated locational market design is that itmakes the locational marginal opportunity costsvisible and applies these as the prices used in thesettlements system.The incremental costs of usingthe full granularity are modest, or even negativewhen considering the costs of imposing rules andprocedures to overcome the effects of artificialaggregation.

Aggregate Regional CountertradeMarket Design

The countertrade markets differ from the inte-grated locational markets described above in avariety of ways (Boucher and Smeers 2002). Atypical feature is an attempt to separate spot mar-ket trading from the actual details of system opera-tions. The energy product is defined according toinputs and outputs at some level of regional aggre-gation. Trading reaches an equilibrium in theseaggregate products, and this is turned into a set ofschedules for implementation by the transmissionsystem operator. The nominal equilibrium sched-ules determined in the market are seldom if everstrictly feasible for implementation on the actualgrid, because they ignore many or all of the con-straints on grid operations. The transmission sys-tem operator has available essentially two broadstrategies for bridging the gap between the nomi-nal market equilibrium and physical system:applying either ex ante limits or ex postcountertrading.

An ex ante limit would be to define differentregional aggregations for the energy products andpretend there is a simple pipeline between the tworegions. Transfers through the pipeline would belimited to a conservative estimate of the realcapacity, and transfers from anywhere in the firstzone to anywhere in the second zone would be(artificially) treated as having the same effect. Theconservative capacity cushion would help limitproblems caused by trades that would appearidentical in the pipeline model but would havedifferent impacts on capacity in the real system.This approach ensures underutilization of the gridby all parties that must use this mechanism for

ElectricityWholesale Market Design in a Low-Carbon Future 119

trading.The same conservative assumptions createopportunities for transmission owners to exploittransmission capacity that is in effect withheldfrom the rest of the market.

Other forms of ex ante rules required by thedisconnect between market design and systemoperation may include restrictions on the leadtime for changing energy schedules. Physicaloperation of plants and control mechanisms oftenallow for changes up to minutes before real time,but formal market schedules may foreclosechanges an hour or more ahead of operations.This scheduling lag could be particularly impor-tant for highly variable energy resources.

Despite the conservative capacity restrictionsand other ex ante rules, market schedules may stillviolate the physical limits of the system. Eventhough generation and load may balance withinthe boundaries of an aggregate zone used todefine a common energy product, for instance,intrazonal power flows may violate transmissionconstraints. There is no way to set an ex ante rulethat maintains the aggregation of the zone andreflects the impacts of the intrazonal limits on thetransmission system.The trade-determined sched-ules cannot be honored without some redispatchof the system, and this provides a need forcountertrade. Typically the initial schedules arehonored, but the system operator arranges a gen-eration redispatch that involves the equivalent ofpurchases and sales of energy at different locationsin the system and creates counterflows that relievethe transmission constraints. In principle and inpractice, this countertrade redispatch may involvethe very trades and generators that created thepotential overload. In effect, a generator may bepaid in the first instance to provide energy as partof the nominal market equilibrium schedule, andthen be paid not to generate the same energy inorder to accommodate the counterflow andrelieve the transmission constraints.

As might be expected, the incentive effects ofthese aggregation and countertrade arrangementscan be perverse in the extreme. The zonal aggre-gation models in the PJM system, New England,California, and Texas all faced these perverseincentives, which precipitated abandonment ofthe zonal aggregation and countertrading marketdesign.

Other markets have implemented counter-trade systems that work much better. For instance,the EU universally employs aggregate regions fordefining energy products, and it uses pipelinemodels with conservative transmission capacityassumptions. All these workable systems find itnecessary to dispense with principles of nondis-crimination. The transmission system operator isable to make onetime arrangements forcountertrades that relieve transmission constraintsand balance the system without having to offerthe same deal to every other participant in themarket. This reduces the ability of generators toextract payments for not running. Although nec-essary, this selective participation in the market bythe transmission operator creates its own set ofincentive problems.

The cost allocation mechanisms for thesecountertrade systems are varied. In some cases,such as in the United Kingdom, the transmissionsystem operator absorbs the costs under a pricecap regime and therefore internalizes incentives tominimize the countertrade costs. However, thissame system requires the operator to take on thesole responsibility for transmission expansion, asthe aggregate price signals in the market do notprovide sufficient incentive or information toguide detailed transmission investments.

Market Design for a Low-Carbon Future

In the United States, installations of wind energyare disproportionately found in the RTO marketsbecause of the greater ease of integration (EPSA2008). As would be apparent from the summary ofkey features presented here, there are net benefitsto moving away from either vertical integration orregional aggregation with countertrade andtoward the integrated locational market design.Similar conclusions can be found in other impor-tant reviews looking toward a low-carbon futurein the United States (Helman et al. 2010; Joskow2008) and United Kingdom (Grubb et al. 2008).As summarized by the International EnergyAgency in its review of market experience acrossits member countries: “Locational marginal pri-cing (LMP) is the electricity spot pricing modelthat serves as the benchmark for market design—

120 WilliamW. Hogan

the textbook ideal that should be the target forpolicy makers. A trading arrangement based onLMP takes all relevant generation and transmis-sion costs appropriately into account and hencesupports optimal investments” (IEA 2007, 16).

Connection Standards

The rules for generator grid interconnection playa major role in affecting the integration of newcapacity. In the case of the integrated locationalmarket design, there is fundamental simplicity intheory that can be reasonably approximated inreality. In the other market designs, system con-nection rules confront the challenge of designingprotocols to overcome the perverse incentivescreated in the market.

The simplicity of the integrated locationalmarket design is illustrated in the rules for “energyresource” interconnection in the PJM system(PJM 2009). An investor that has a site with therequired permits and wants to build a new facilityto sell energy into the PJM market has to meetonly a minimal set of requirements. Primarilythese are technical constraints to ensure that theconnection to the grid will satisfy electrical char-acteristics so that connection and charging of theline, with no material power generation, do notcompromise the rest of the system. In effect, thegenerator interface is an extension cord to the restof the grid that has to meet certain standards, andthe generator has to pay for the equipment andextension line.

This form of interconnection allows the gen-erator to produce, buy, and sell energy in the spotmarket at the locational marginal price at its pointof interconnection. There is no guarantee aboutthe price or profitability of the energy produced,nor any guarantee that access to the grid at thepoint of connection provides transmission rightsto hedge the cost of delivery to other locations inthe grid. When the transmission system is con-strained, the locational marginal price at the pointof connection might be very low, and the genera-tor may choose not to run. The generator canhave the system operator make these decisions inreal time by simply offering to produce with a

minimum acceptable price. There would be noneed to arrange advance schedules or bilateraltransactions.

Although the generator could participate inforward markets, arrange long-term contracts,and purchase FTRs to hedge transmission costs,these are not necessary conditions of the connec-tion rule, and immersion in trading would notchange the incentives for the price-taking genera-tor to make simple offers into the real-time mar-ket to allow the system operator to take theenergy when available and when the locationalprice is at least equal to the minimum bid.

As discussed below, in PJM, most generatorschoose a more complicated form of interconnec-tion in order to participate in the capacity marketthat operates on top of the energy market. How-ever, the simple ideal of the energy-only inter-connection is available and is used by some gen-erators.

The countertrade markets face a quite differ-ent situation when considering the rules for con-necting new generators. The generators do notsee the costs of transmission constraints obscuredby zonal aggregation of energy product defini-tions. Hence they may choose to locate wherecosts are actually higher but are imposed on some-one else. Furthermore, the typical countertrademarket design carries with it an explicit orimplied obligation to accept the generators’energy output and deliver it to some loads. Ineffect, in the terms of the integrated locationalmarket design, new generators acquire a bundle oftransmission rights by virtue of their interconnec-tion. The connection of new generators mayimply that the transmission owner must makesometimes substantial transmission investment tosupport these new rights. Similarly, for all genera-tors, but particularly for highly variable wind andsolar plants, there must be increased operatingreserves or other controls needed to supportenergy delivery. The transmission system operatorwill have to make these investments or imposeconstraints on the interconnection.

This gap between the incentives for the newentrant and the real costs of interconnection cre-ates a need for rules and regulations to overcomethe incentive effects. For example, new generators

ElectricityWholesale Market Design in a Low-Carbon Future 121

may be required to make so-called deep transmis-sion investments elsewhere in the network inorder to accommodate their new productionwithout undoing the implicit rights of other gen-erators. Alternatively, the costs of reliability andtransmission upgrades to support the expanded setof transmission rights are imposed on the trans-mission system operator or somehow collected insocialized charges applied to all customers.

Smart Grids and Smart Pricing

The notion of the smart grid typically implies amix of control technologies, information systems,and information distribution that allows for muchfiner control of all the devices connected in thelargest machine in the world. This would includemore flexible controls in the transmission grid,better monitoring of grid and distribution flows,real-time status and fault monitoring, automaticmeter reading, and importantly, real-time auto-mation of end-use devices. Signals from the cen-tral system operator would communicate throughthe network to change the settings on transform-ers deeply embedded in the grid or trigger cus-tomers’ settings to turn down the air-conditionersin their homes.

The evolution of the smart grid will be par-ticularly relevant in the development of dispersedgeneration, end-use efficiency, and load manage-ment. An important question is what informationis provided and who decides how to use it. Con-sumers are unlikely to surrender their ability tocontrol the temperature in the house or to decidewhether to run a factory. However, it is easy toimagine widespread adoption of smart devicesthat implement instructions as to when to curtailusage and when to run as needed.

For these smart systems to be successful, con-sumers must receive a signal that aligns theirincentives to behave in a way that reflects theopportunity costs in the system.This is a challengein the vertically integrated and aggregatedcountertrade systems, because under these marketdesigns, prices facing the consumer do not reflectthe granular detail of system operations. Com-mand and control under the vertically integratedmodel and countertrading under the aggregated

market models can in principle provide an effi-cient dispatch for generators and respect the trans-mission constraints, but they do not by themselvesalign the incentives of customers.

By contrast, the integrated locational marketdesign provides a straightforward alignment ofincentives. With a retail rate design that passesthrough the locational price of energy, the smartdevice would have smart prices that aligned withthe incentives of the consumer and the system’scosts. Some distortions would remain to theextent that distribution or other customer servicefixed costs are collected in variable rates, but thedistortions would be small given the large changesin energy prices that would or should be thenorm during peak periods.

The smart grid needs smart pricing. Marketmodels that suppress price information, especiallyin prices of final delivery to the consumer, wouldrequire new and sometimes complex rules toundo the consequent perverse incentives. By con-trast, the integrated locational market designintentionally links the market and operations inorder to provide the right price signals.

Resource Adequacy

A central function for vertically integrated electricutilities has always been to invest in anticipation ofelectricity demand. The club members of verti-cally integrated companies developed rules forinvesting while sharing the responsibility to main-tain reliability. In practice, this amounts to aresponsibility to build or contract for energy gen-eration capacity adequate to provide a specifiedplanning reserve margin (e.g., 15% of estimatedpeak load) that would ensure sufficient capacity inthe event of normal outages for maintenance andsome temporary equipment failures.

This obligation continues for the verticallyintegrated utilities. However, the situation hasbecome more complicated in the market modelswhere generation has been separated from trans-mission, distribution, and final supply to the cus-tomers. The success of these models in achievingresource adequacy is neither certain nor resolved.The early days of restructuring, at least in theUnited States, saw a surge of new construction of

122 WilliamW. Hogan

more efficient generating plants. This createdexcess capacity. When coupled with bad marketdesign and rising oil and natural gas prices, thiscapacity was not economic, and the result was amajor contraction in investment and significantfinancial losses for some generators. In one sense,this was the intended consequence of electricityrestructuring in shifting the burden of risks to vol-untary investors rather than imposing them oncustomers. The excess investment was not a goodthing, even if a good deal of the problem wascaused by the incentives of bad market designrather than the reliance on markets per se.

The result has been a concern that in thefuture, investment in generating capacity in elec-tricity markets would not be adequate. This con-cern has been strongly reinforced by the regulardemonstration that the returns to investors in theenergy and ancillary services prices have not beenadequate to support new investment.3 Although itis the anticipated returns that should govern, notthe realized returns in a market with excess cap-acity, the low realized returns are sobering. Theconsensus analysis is that energy prices in manyelectricity markets have been too low, and theyare especially low during periods of capacity scar-city, such as during the peak hours of the day.

The response to this concern has been varied.In some markets, notably New York, PJM, andNew England, a “capacity” product and an associ-ated capacity market have been created. Theessence of the capacity market has been to con-struct a forecast of future loads and investments,and then conduct an auction for existing and newgenerators, or demand response, to offer capacityto meet those requirements. The full descriptionof how to define “capacity” and ensure its futuredeliverability on the grid is another task. How-ever, it is clear that this capacity construct is anattempt to re-create something like the responsi-bility of the vertically integrated utility, but nowunder the management of the independent systemoperator.

The difficulties of defining and measuringcapacity for new low-carbon technologies such aswind, solar, and load management are particularlyrelevant. As discussed below, the variable availabil-ity of renewable resources and the problem of

defining baseline load make it hard to map theminto the equivalent savings in conventional fossil-fuel generation capacity. In part, the problems areinherent in capacity markets where the product isnot easy to define or measure, but the difficultiesare exacerbated by the substantial differences inthe technologies.

At the other end of the spectrum has been therecognition that inadequate scarcity pricing is theroot cause of the deficient incentive for invest-ment.4 Markets such as in Australia andTexas havemechanisms to allow prices to rise during times ofscarcity on the assumption that this will be suffi-cient to meet the needs of resource adequacy.

Nothing fundamental requires a choicebetween improved scarcity pricing and adoptingwell-designed capacity markets. Better scarcitypricing would provide benefits in improvedoperations as well as market incentives for invest-ment. Furthermore, better scarcity pricing inmarkets would complement the designs of cap-acity markets to account for actual or expectedscarcity payments (Hogan 2006). The design ofbetter scarcity pricing systems is an active agendaitem in the RTO markets in the United States(ISO-NE 2006; MISO 2009; NYISO 2008).

The concern with resource adequacy may bemore acute in the United States than in the EUmarket design models. All of the U.S. modelsinclude procedures to monitor and mitigate theexercise of market power, primarily through theapplication of offer caps. These procedures, cou-pled with many other market design nuances,accumulate to depress energy market prices. Bycomparison, the EU market is more opaque inthis dimension, more inclined to structural inter-ventions such as divestiture, and less inclined todirectly intervene in the bidding process. Theincentive to exercise market power is material,and the opportunity may be greater in the EUcontext. The net result is a more mixed pictureabout the adequacy of incentives for new genera-tion investment. Furthermore, at the moment,the more aggressive policies in some EU countriesfor introducing renewable energy sources, par-ticularly wind, have produced excess capacity andraised total system costs, but the increase in capac-

ElectricityWholesale Market Design in a Low-Carbon Future 123

ity depresses energy prices as they appear in thenominal market (Traber and Kemfert 2009).

Infrastructure Investment

Providing generation resource adequacy is onlypart of the necessary development of the electri-city system. Other necessary infrastructure devel-opment includes transmission grid expansion anddeployment of system upgrades in the form ofbetter information, smarter controls, and smartermetering. In the vertically integrated electric sys-tem, these investment decisions were and remainprimarily the responsibility of the electricitymonopoly, subject to regulatory approval. Therestructured markets, however, present a differentmix of choices, and in many respects responsibil-ity for infrastructure investment is unresolved.

A principal purpose of development of elec-tricity markets was to provide better incentivesand opportunities for decentralized decisions witha corresponding allocation of risk and reward forinvestments in generation plants and end-use effi-ciency. In an idealized electricity market struc-ture, energy market prices would be sufficientdrivers of generation and end-use investment. Asdiscussed above in the section titled ResourceAdequacy, the less-than-ideal reality has not yetachieved this goal for generation, and similarproblems arise in providing the incentives forend-use investment.

The case for transmission and related infra-structure investment is even more of a challenge.The transmission grid and associated dispatchfunctions of the system operator continue as natu-ral monopolies. Even in the idealized version ofelectricity market design, it could be difficult tofashion a regime that relied on market-basedtransmission infrastructure investment (Hogan2009). Accordingly, it is widely (but not univer-sally) assumed that some residual monopoly—atransmission owner, system operator, or hybridspecial-purpose organization—must make infra-structure investment decisions, cause the expan-sion to be implemented, and then use the com-pulsion through regulation to make the erstwhileunwilling beneficiaries pay.

If the beneficiaries were willing to pay, theycould in principle organize to make the transmis-sion investment without the coercive power ofregulation. In some cases, market participants domake their own decisions to invest and acquireincremental transmission rights (Hogan 2009). Inother cases, however, it may well be that no coali-tion of beneficiaries could be assembled to under-take desirable transmission investment. In thesecases, a transmission company, the system opera-tor, and the regulator must cooperate to imple-ment investment plans.

Designing a hybrid system that could accom-modate both market-based and regulatory-driveninfrastructure investment is a major challenge forliberalized electricity markets. In some regions,the assumption is that such a hybrid system isimpossible and only the transmission companycan or will make investments. Other regions haveseen a struggle to find a workable compromise. Aninteresting case is the hybrid transmission expan-sion framework adopted under the New YorkIndependent System Operator (NYISO), mod-eled after the transmission investment experiencein Argentina. In essence, the NYISO modelallows for decentralized investment when thebeneficiaries voluntarily assume the responsibilityto pay. In cases where the dispersion of benefitsmakes it difficult or impossible to assemble a coa-lition of the beneficiaries, the system operatorand regulators can go forward subject to asupermajority endorsement where the beneficiar-ies vote and pay for the investment in proportionto the estimated benefits (Budhraja et al. 2008;NYISO 2007). An important feature of bothinvestment avenues is that beneficiaries pay andthe costs are not socialized, which provides goodincentives and is compatible with the frameworkfor investment in generation and end use withoutcreating conflicting subsidy regimes.

The ability to develop an adequate hybridinfrastructure investment framework depends onthe rest of the market design. In systems withbroad energy regions with a common nominalenergy product and a single price, the details oftransmission congestion are hidden and marketsignals are weak (between regions) or absent(within regions). In these cases, market partici-

124 WilliamW. Hogan

pants have no incentive to invest, and a transmis-sion company monopoly may be the only work-able system. For example, in England and Wales,the single transmission owner National Grid isresponsible for designing, implementing, andcharging for transmission investment under thesupervision of the regulator. The United States,on the other hand, has a mixture of evolvinginfrastructure investment schemes. New Englandfollows close to a central-planning and full-costsocialization system, whereas New York, PJM,Texas, the Midwest, and California have varioushybrid models in place or in flux.

The details of each of these systems can becomplicated, but a general principle connects tothe underlying market design. By construction,the more aggregated the market design, the lessarticulated the investment incentives for transmis-sion. Within aggregated zones, market partici-pants have no incentive to invest because trans-mission constraints are assumed away. Betweenaggregate zones, transmission investment incen-tives may exist, but the price information in themarket would only dimly reflect the underlyingreality of the transmission interactions. Only inthe case of the fully locational integrated marketdesigns could energy prices furnish accurateinformation about transmission congestion thatprovides a framework for awarding feasible trans-mission rights. But these market designs result inefficient infrastructure investment decisions onlywhen the investment itself has relatively limitedimpact on market prices. Larger investments thatcould materially change expected market pricesmay then require a hybrid framework like theNewYork model.

Green Energy ResourcesGreen energy resources include a diverse array oftechnologies. Nuclear power and large-scalehydro provide energy without producing materialcarbon emissions, and their impacts on the elec-trical system are well understood. Newer low- orzero-carbon generation technologies such aswind, solar thermal, solar photovoltaic, tidal sys-tem, biomass, and fossil fuels with carbon capture

and storage offer a variety of nontraditional oper-ating characteristics that could affect electricitysystems and markets. Energy efficiency and loadmanagement investments could materially alterthe load profile and interact strongly with theportfolio of generation and transmission assets.

Desirable policy for pursuing low-carbontechnologies depends in large part on the associ-ated choices in market design. In some cases, suchas nuclear and hydropower, the technology isfamiliar and well understood. In others, such ascarbon capture and storage (CCS) or expandedcofiring of biomass, although costs may be differ-ent, nothing is obviously fundamentally differentfrom existing fossil-fuel generation that couldalter market design. In still other cases, such aswind or solar, the operating characteristics are sig-nificantly different. At a small scale, almost anynew generation technology can be absorbed, butlarge-scale integration of the newer low-carbontechnologies is more complex. And some antici-pated changes, such as the development of smartgrids, smart appliances, and electrification of thetransportation sector, could require or facilitatemajor changes in system operations and marketdesign.

Uncertainty and Technology Choice

If you know what to do, do it!5 If we knew whichtechnologies to embrace, which infrastructureinvestments to support, and how opportunitieswould evolve, then the low-carbon investmentproblem would be simple, and the vertically inte-grated monopoly electric utility model wouldhave much to recommend it. Regulators wouldcontinue to oversee the investment choices andset rates for cost recovery. With little uncertainty,the task would be relatively easy, risks would below, and the lack of efficient incentives for marketparticipants would be less problematic.

Unfortunately, we do not know exactly whatto do. By the late 1980s, partly as a result ofchanges in technology, vertically integrated elec-tric utilities around the world were saddled withhigh costs and unwanted assets. The resultingpressure led to reform of electricity markets tochange the incentives, locus of decisions, and allo-

ElectricityWholesale Market Design in a Low-Carbon Future 125

cation of risks. Part of the animating idea was topromote innovation, entry, and flexibility in theelectricity system. New technologies and newoperating procedures would arise as a result ofmarket incentives and decentralized decisions.Incentives would be better aligned to supportinnovation in both end use and generation facili-ties, and innovation would increase. Rewardswould go with the risks, and the market partici-pants would do a better job of euthanizing badinvestments as well as capitalizing on what works.

The more innovative the idea, the less likely itwould be adopted under the regulated verticallyintegrated organization. Regulation works betterwhen the choices are highly standardized and easyto evaluate. But when the product includes manymoving parts or involves a great deal of uncer-tainty, it can be much harder to write down therules. In these cases, business judgments can bemade and the uncertainty priced, with the associ-ated risks and rewards. For example, consider thecase of EnerNOC, an innovative aggregator ofdemand response services, and its demand prod-uct, which provides operating reserves in marketssuch as PJM (EnerNOC n.d.).The de facto prod-uct provided to PJM behaves essentially like reli-able standby generating reserve that can be calledon to meet gross load or reduce net load at a set oflocations.The actual service, hidden from view ofthe system operators, is a complicated array ofcontracts and operating agreements with endusers that allow a variety of interventions withvarious technologies to reduce or temporarilyshift electricity load. These contracts are negoti-ated privately, and voluntarily, to share risks andrewards. There is no need for regulators toapprove the complex and varied terms and condi-tions of the individual contracts. EnerNOC takesa reasonable business risk, constrained in itsreturns through the possibility of competitionfrom others in providing similar services. PJM getsinnovative access to cost-effective operatingreserves that fit naturally under the market design.The market design stimulated a market response,but the rules did not prescribe the result.

This broad motivation for electricity marketreform is greatly reinforced by the challenges ofdeveloping a low-carbon electricity system. Car-

bon emissions have a long duration in the atmos-phere and will affect the climate for centuries aftertheir sources have been retired, and generatingplants may operate for many decades. Hence weare interested in impacts over very long timescales,and we are hoping for and expecting dramaticinnovations in technology and operating proce-dures. Over this timescale, enormous uncertaintyexists about what will work and how the systemwill operate in practice. The biggest surprisewould be if there were no major surprises.

In this setting, it is even more important toengage the creativity and innovations of the manycurrent and potential market participants ratherthan rely on the wisdom of a few central planners.If we knew what to do, the central planners couldexecute. But because we do not know exactlywhat to do, we need an electricity market designthat facilitates learning.

In the face of great uncertainty, the signals andincentives should be as technology-neutral as pos-sible. Rather than prescribe or constrain thechoices of technology, market operations andprices should reflect the real costs incurred.To theextent possible, externalities should be internal-ized. For example, rather than mandate technolo-gies, we should put a price on carbon that shouldaffect technology choices.

Among the three market models, the verti-cally integrated model is most like central plan-ning, and the integrated locational market designis most attuned to reflecting real costs associatedwith system operations. The aggregated counter-trade models fall somewhere in between, withmore opportunities for decentralized investmentdecisions but more muted incentives to guidethose investments.

The need to support innovation, to pursuetechnologies, investments, and consumptionchoices not yet envisioned by anyone—includingby the central planner—provides a powerful moti-vation for getting the signals and incentivesaligned to support decentralized decisions. Bycontrast, the more the electricity market designrelies on socialized costs, muted incentives, andregulatory mandates, the more the central plannerassumes the burden for making investment deci-sions and applying regulatory compulsion torecover the costs.

126 WilliamW. Hogan

Investment and Operating Costs

There would not be much policy concern withlow-carbon technologies if they did not appear tobe more expensive than fossil-based alternatives.With a price on carbon emissions, the cost differ-ential between fossil-fuel plants and low-carbonalternatives may be overcome. However, materialdifferences in costs may remain for nascent tech-nologies entering the market.

The response to any remaining cost differen-tial would depend on the nature of the highercosts and the structure of the market design. If aprice is put on carbon, but the price is not suffi-cient to induce adoption of a particular technol-ogy, there could be several explanations. Forexample, it may be that the technology is simplytoo expensive compared with other alternatives.In this case, no real problem exists, and the failureto embrace the particular low-carbon investmentis the correct solution.

Another possibility is that the price of carbonis too low.The best response would be to raise theprice of carbon. Absent an adequate carbon price,if a policy decision is made to invest in low-carbon technologies that are not cost-competitive, the nature of the cost differentialwould have different impacts under alternativemarket designs.

An alternative case would be that the price ofcarbon is appropriate, the current cost of the newlow-carbon technology makes it not competitive,but learning by doing could lower the costsenough to make the low-carbon technologycompetitive. Again, the impacts will be differentunder alternative market designs.

A market design based on a vertically inte-grated monopoly will face familiar issues of tech-nology adoption and investment in all three cases.The regulator would oversee investment choicesand provide rates for cost recovery. Low-carbontechnologies could be adopted as a matter ofpolicy, despite their costs.The particular nature ofthe cost differentials would not be a first-orderconcern.

In the other market models, relying more onprices and other market incentives, the nature ofthe cost differential would matter more. With

higher costs, adoption will require some form ofsubsidy. If the higher cost is in the form of highinvestment costs, tax credits or other capital subsi-dies could be offered in many ways. Once thesubsidy is provided, the new generation plant orend-use investment can participate in the ongoingelectricity market on the same basis as otherinvestments.

If the source of the higher cost is a higheroperating cost, the interaction with the marketdesign may be more immediate. Consider, forexample, an end-use technology such as loadmanagement to reduce consumption when theavoided cost of electricity is very high. If the mar-ket design does not provide the appropriate pricesignals that capture the avoided cost in the system,often over relative short intervals, enough incen-tive may not exist to pursue load managementeven if the up-front investment cost is low or sub-sidized. Here both the temporal and locationalgranularity of prices would be important. Similarchallenges would arise for generation investmentssuch as pump storage, flywheel storage, orpondage hydro, which could smooth out effectiveload curves but face high effective operating costs.If the market design does not offer sufficientgranularity of price determination, then othercost-effective low-carbon technologies would betoo expensive to operate or would require regula-tory mandates for adoption.

Intermittency and Reliability

Of the many technical characteristics of certainlow-carbon technologies, the feature that drawsthe most attention is intermittency of the supplyand the implications for reliability and meetingsystem load. The canonical concern is for windand solar photovoltaic (PV) panels. Once con-structed, the variable operating cost of these gen-eration plants is effectively zero. But if the winddoes not blow or the sun does not shine, nopower is generated.

This intermittency gives rise to two relatedbut different problems. The first is the availabilityof the power when it is needed. With traditionalnuclear, hydro, and fossil-fuel plants, requiredmaintenance can be scheduled to minimize the

ElectricityWholesale Market Design in a Low-Carbon Future 127

impact on peak hours, and the random outagesthat cannot be controlled have a relatively lowprobability. By contrast, the implied effectivecapacity factor of 30% or below for individualrenewable facilities is equivalent to random out-age probabilities of 70% or more. The correlationof these outages with peak and near-peak loadswould be important, but the basic intermittency isa hurdle for some low-carbon technologies.

Another aspect of variable availability of low-carbon technologies is the speed at which theycome on or, more important, go off the system.This “ramp rate” can present material operationalchallenges for the system. For small-scale penetra-tion of the low-carbon technologies, not muchmore need be done than include this variability aspart of the natural variability of residual load. Buton a large scale, the challenge would be differentdepending on the dispatchability of the plant.

A common answer is that the individual vari-ability of solar or wind facilities is less of a prob-lem when regional diversification of sources issufficient where winds speeds are not correlated.It is true that the correlation reduces across widerareas (Holttinen et al. 2009).6 However, exploit-ing this portfolio effect depends importantly onthe configuration of the transmission grid as dis-cussed below under Green Infrastructure Devel-opment.

In principle, for the integrated locational mar-ket design, variability in production does notpresent much of a design problem. Security-constrained economic dispatch would imposelimits on the output of the plants and reduceprices for energy the more that capacity has to bereserved to meet the potential changes in produc-tion. For example, this might involve restrictingoutput from wind facilities when the wind isblowing hard to keep the combined output of thewind farm below the level that the system couldhandle if the wind suddenly drops. The net effectwould be to reduce the locational price of energyand increase the locational price of operatingreserves. Most of the cost of variability wouldthen fall on the wind supplier, with the associatedincentives to invest in storage or other backupoptions.

This appeal to the theory of integratedlocational market design contradicts the commonassumption that windmills should generate when-ever the wind is blowing. Although this is notcorrect as an economic matter, because at timesfree energy is too expensive, widespread concernexists that investment in low-carbon technologiesshould be combined with other investments toensure that the power can flow when it is avail-able.

The market design implications of the inter-mittency of supply are familiar. The more granu-larity, over time and location, in the prices anddispatch, the easier it is to integrate low-carbontechnologies.The better the prices reflect avoidedcosts, the easier it would be to integrate suchtechnologies, especially those with high operatingcosts.

The greater variability of supply for key low-carbon technologies and increased end-useresponse become more and more important thegreater the investment in smart grids. As discussedabove, smart grids require smart prices. Smarterprices would better reflect scarcity conditions,locational differences, and dynamic response. Inthe absence of sufficiently granular prices andquantity definition, incentives for low-carboninvestment diverge from the real costs of systemoperations. The details will be different in differ-ent markets, but the greater the gap, the more thereliance on regulation and central mandates tosupport low-carbon technologies.

Green Infrastructure Investment

The places where the wind blows and the sunshines are not always where people congregate.Hence, from a strictly technical perspective, someof the best locations for siting low-carbon tech-nologies like wind and solar energy are far fromloads. West Texas has a great deal of wind, but theload is in Dallas and Houston. The result is a sub-stantial need for transmission investment. InTexas,a policy decision has been made to move aheadwith transmission expansion to enable greater uti-lization of the wind resources and socialize thecost through additional charges to all customers.7

In other regions, the justification for addi-tional transmission and the associated cost alloca-

128 WilliamW. Hogan

tion are more problematic. In the case of NewYork, the large potential wind resources are in thenorthwestern region of the state, and the majorload center is around NewYork City. Controversyhas arisen about whether to build new transmis-sion and who should pay. The integratedlocational pricing design sends clear signals aboutthe potential benefits of expanded transmission,but so far the benefits have not been seen as worththe costs. Under the innovative New York trans-mission investment rules, the beneficiaries mustpay and must approve in advance according tosupermajority voting. In the case of the proposedNew York Regional Intertie, designed to exploitupstate wind, the developers withdrew the planon the grounds that it would not receive sufficientsupport from the putative beneficiaries to obtainapproval (Puga and Lesser 2009).

One of the challenges for transmission invest-ment is the intermittency of wind and solarenergy production. A standard argument is thatdiversification of sources of supply across largeregions where wind and cloud cover are not per-fectly correlated would provide a sufficient port-folio effect to mitigate the intermittency. Whatthis argument must confront is the level of trans-mission investment required. For the diversifica-tion argument to hold fully, transmission capacitymust be sufficient to accommodate the maximumoutput of all of the sources. In the case of dedi-cated long-distance lines to bring power to load,each line must be sized to be fully utilized only afraction of the time. Otherwise the full diversifi-cation benefit would not be realized. This surgecapacity investment in transmission, and plannedunderutilization, could materially add to the costsof interconnection for remote low-carbon energytechnologies (NAS Committee 2009).

The more the market design adheres to thegranularity of time and location, the greater willbe the transparency of incentives for transmissioninvestment. The more the market design aggre-gates across time and location, the more hiddenwill be the benefits of transmission. The higherthe degree of aggregation, the greater will be thetendency to socialize the cost of transmission.

Movement toward the integrated locationalmarket design facilitates a beneficiary-pays system

of transmission expansion and supports other fea-tures of market investment decisions. Without abeneficiary-pays protocol, there is no principledanswer to why the subsidies to transmissionshould not be extended to demand and produc-tion alternatives that are partial substitutes fortransmission expansion. A beneficiary-pays proto-col for transmission is an important long-termcomponent of market design. As ready examplesinTexas and New England demonstrate, however,the use of an integrated locational market designfor the energy market is no guarantee that trans-mission investment will be market-driven and notsocialized.

Green Policy Mandates

Many policies have arisen or might arise to sup-port development and deployment of low-carbontechnologies. In some cases, these policies wouldcomplement broader market decisions. In othercases, the efforts to internalize carbon externalitiesmight conflict with market design features.

Carbon Emission Caps

By far the most important component of an effi-cient and effective policy to internalize the cli-mate impacts of carbon emissions is to put a priceon carbon, as has been done in the EU and islikely in the United States. For all the usual rea-sons, the carbon pricing regime should, to theextent possible, be economywide and global in itscoverage (Stavins 2007). The details of the policychoice of a carbon tax versus a cap-and-trade pro-posal are important and raise many other issues.

A carbon cap with trading of emission permitsis widely accepted as being more politically feasi-ble because of the greater opacity of the revenuetransfers inherent in allocation of the permits. Aprincipal feature of a good allocation design is tobreak any connection between actual energy pro-duction and the aggregate allocation of the per-mits.8 In this event, free permit allocations involvea transfer of wealth but provide no reduction inthe cost of carbon emissions on the margin.

A key element of a workable policy will bethe credibility of the carbon regime and the expo-

ElectricityWholesale Market Design in a Low-Carbon Future 129

sure to future political risk (Neuhoff et al. 2009).Major investments in expensive low-carbon tech-nologies would be profitable only if sustainablepublic policy ensures a significant implicit orexplicit price on carbon emissions that is expectedto remain high enough for long enough torecover the substantial investment.

Sustainability of public policy will depend inpart on compatibility with the basic marketdesign. For example, an economywide cap oncarbon emissions would be largely technology-neutral and would integrate well with efficientmarket operations. Investors in a particular tech-nology would not need to worry as much thattheir specific venture would be subject to idiosyn-cratic subsidy changes. In order to change theeconomics of the individual investment, the car-bon cap would need to be changed for all emis-sions, and this is less likely than a change in atargeted subsidy. By contrast, the boom-and-bustcycle of solar deployment in the United States,driven by the dependence on periodicreauthorization of targeted tax credits for solarinvestments, illustrates the risk of developing asustained commitment to investment based onspecial-purpose support (NAS Committee 2009).Comprehensive carbon emission caps and goodmarket design would reinforce the sustainabilityof the system.

Research, Development, and Demonstration

The usual arguments imply that major spilloverbenefits should accrue to research, development,and even early demonstration (RD&D) efforts todiscover and refine better low-carbon technolo-gies. Instituting policies to put a price on carbonwould reinforce incentives for private RD&D, butthere is no reason to believe that the full externali-ties of the public-good aspect of RD&D wouldbe overcome by putting a price on carbon.Mounting an RD&D program to support energyinnovation (Anadon and Holdren 2009) has majorchallenges. However, the challenges are not muchaffected by different wholesale electricity marketdesigns.

Infant Industries and Learning by Doing

Commercialization beyond RD&D and learningby research interact with deployment and learningby doing. A major argument in favor of publicinvestment to support early adoption of low-carbon technologies is the ability to capture thebenefits of early investment, which accumulate toreduce the going-forward cost of the technology.The externality arises when the benefits of infor-mation about the success of the technology andthe improved understanding of how to build andoperate cannot be captured by the investor. Thuslearning by doing is a type of infant industry argu-ment, where early support is needed to launch thetechnology, but once mature, the technology isable to compete on its own without further sub-sidy.

An interesting feature of the application oflearning-by-doing analysis to energy technologiesis the substantial nature of the putative benefits.High learning rates generate such substantialreductions in total electricity generation cost thatmost of the benefit comes from improved effi-ciency of the electricity system, and carbon emis-sion reduction benefits are a small component ofthe gain. In some cases, if the claimed learningrates are true, it would pay to provide public sup-port for the technology even if carbon emissionsposed no cost (van Benthem et al. 2008).

This argument about large potential costreductions might apply to wind and solar genera-tion technologies that are alternatives to fossil-fuelgeneration but cannot apply to CCS technolo-gies. Although substantial learning may come inCCS design and deployment, the remaining costswill always be in addition to the cost of fossil-fuelgeneration. Even if CCS were effectively free, itwould not be cheaper than using the fossil-fuelplant without CCS. Hence all the benefits forCCS come from the carbon reduction.

Policies to support learning by doing shouldbe structured to recognize the distinction betweensupporting learning about the technologythrough early cumulative investment, and sup-porting the technology by offering sustained sub-sidies with little or no further learning. Quantita-tive targets for low-carbon technology investment

130 WilliamW. Hogan

may have little connection with the public ben-efits of learning. For instance, the case of the Cali-fornia solar rooftop initiative is instructive.Launched with a promise to install solar PV on amillion rooftops by 2015, the idea was to bothreduce carbon emissions and capture the benefitsof learning by doing. However, whereas theassumed learning rate is an input, the necessaryscale of the investment should be an output of theanalysis and not a constraint on the policy. In theCalifornia case, further analysis revealed both theimplied trajectory of the optimal subsidies andinvestment profiles as a function of the assumedlearning rate. A central conclusion was that theoptimal subsidies would be associated with invest-ment at only about a quarter of the scale of thenominal million-rooftop target. Higher learningrates needed less initial subsidy, and lower learningrates might not produce enough benefit to justifyany material subsidy (van Benthem et al. 2008).

Proper design of support policies to capturethe benefits of learning by doing interact withmarket design in straightforward ways. The moretransparent the market design and the clearer themarket signals, the easier to achieve the benefits ofthe learning and know whether the benefits areappearing. In the energy market, the better theprice signals, to include scarcity costs, the betterthe opportunity for private investment to drivethe benefits of lower costs.

Renewable Energy Standards

The competition between incentives and targetsappears most starkly in the case of renewableenergy standards (also known as renewable port-folio standards, or RPSs). Motivated by the mar-ket model of the vertically integrated electric util-ity, with a single buyer, the idea is to promotepenetration of low-carbon technologies by settingmandates for the percentage of electricity genera-tion that must be provided by defined categoriesof renewable or low-carbon energy sources. Suchgoals supported by feed-in tariffs have beenadopted as targets throughout the EU and asquantity mandates in many states in the UnitedStates. Experience with this policy instrument

presents a number of difficulties, some of whichinteract with the nature of the market design.

The standards often lack a principled basis. Acap on carbon emissions, where the emissions ofcarbon can be measured and controlled, istechnology-neutral. By contrast, renewableenergy standards require selecting the acceptabletechnologies. The result has been a substantialdegree of political bargaining, with variationacross jurisdictions revealing the fundamentaldisagreements about goals and means. Forinstance, in the United States, state programsdefine a wide array of more than a dozen possiblerenewable technologies, yet “states agree on onlythree technologies: biomass conversion, solarphotovoltaic and wind” (Michaels 2008, 10).

The renewable energy standards also sufferfrom confusion about objectives and a poor map-ping between policy and goals. Requiring mini-mum penetration standards to support earlydevelopment of expensive or risky renewabletechnologies is akin to the quantity target andsubject to the comments above about the mis-match between learning-by-doing justificationand quantitative targets.

Or consider the justification that the technol-ogy would never be competitive on its own, andthe minimum standard is intended to reduce car-bon emissions. Although this approach may havean appeal in the absence of a carbon cap, the argu-ment no longer applies in the presence of a cap. Assoon as there is a binding cap on carbon emis-sions, a renewable electricity standard (RES) pro-duces no incremental carbon emission benefits. Ifthe cap is binding, the effect of an RES is torequire more expensive investments that reducecarbon emissions and therefore lower the cost foreveryone else and every other technology inmeeting the carbon cap. Hence the proposed 33%RES by 2020 in California should raise costs ofcontrol in the state, to the benefit of the rest of thecountry.9 But once a cap-and-trade system is inplace, carbon emissions would be the same aswithout the RES. In effect, the RES lowers themarket price for carbon emission permits andraises the social cost of reducing carbon emissionsbut does not lower carbon emissions.10

ElectricityWholesale Market Design in a Low-Carbon Future 131

In market designs without vertical integra-tion, an RES presents the challenge of how toimplement and enforce the technology constraint.One approach is to create RES certificates andrequire all buyers or generators to redeem a cer-tain percentage of RES generation credits or pay atax (RECS n.d.).The price of the credits is then afactor in the price of electricity and alters theeconomic dispatch. However, the connectionwith carbon emissions is imperfect, as low-carbontechnologies such as nuclear are often excludedfrom RES credits. Thus the preferred cap-and-trade system would have parallel trading systemsto control emissions and promote renewability.There is nothing inherently infeasible about hav-ing multiple permits systems. For example, thecarbon permit would be layered on top of tradablesulfur dioxide emission permits. But the addedcomplexity does raise a question about the incre-mental benefits of RES mandates.

Green Uneconomic Dispatch

Another approach to promoting green energyresources would be to mandate priority for low-carbon technologies to be selected first in energydispatch (Dubash 2002). This idea presents morecomplications than might first appear, because ofthe nature of the transmission system interactionwith electricity market design.

The complex interaction of power flows inthe transmission system can produce cases whereactions at a location to reduce energy use orincrease emission-free electricity production canactually increase carbon emissions in the overallsystem (Rudkevich 2009). Meeting electricityload at certain locations creates counterflow in thegrid, which reduces congestion. Reducing thatcounterflow can create a requirement for amultiplant redispatch that increases overall carbonemissions. The technical condition is similar tothe effects of transmission counterflow, which candrive locational energy prices negative even whenall generation offers are positive. Hence the sys-tem impacts on carbon emissions are not alwayseasily predictable.

This might support a call for a green energydispatch protocol in place of economic dispatch,

but the interaction with energy market designwould be disruptive. In the case of the verticallyintegrated monopoly, there would be less of aproblem. By assumption, the monopoly caninternalize all the untoward effects in the produc-tion and delivery chain and ignore the internaltransfer prices.

By contrast, the market structures now basedon bid and offers in a framework of security-constrained economic dispatch would confront asubstantial disconnect between the economicprices applied in the settlements system and theopportunity costs of changes in the dispatch. Asmuch experience has shown, whenever marketprices used in settlements systems diverge materi-ally from the costs of real system operations,arbitrage opportunities and temptations are cre-ated that could undermine the dispatch and themarket.

It would be possible to impose a constraint onemissions in the dispatch (over very short inter-vals), and then apply the usual principles of eco-nomic dispatch subject to the emissions cap. Thiswould produce consistent market prices withscarcity components reflecting the cost of carbonemissions implied by the constraint. In effect, thiswould replace an economywide cap on emissions,allowing implicit trading across time and space,with a new cap-and-tax system with short-termcaps and implicit taxing at the implied variablecarbon price. Absent a formal electricity-onlycap-and-trade system that allowed intertemporalbalancing, the system operator would have todecide on the limit for virtually every dispatchperiod. It would be reasonable to anticipate thatthe implicit (very) short-term carbon price wouldbe highly volatile.

A far simpler approach would be if the cap onaggregate emissions and the implied carbon pricein the economywide cap-and-trade system weresufficient to internalize the cost of carbon. Thenthe carbon emission permit costs, along with allother costs for fuels and other emissions, would betreated the same in the security-constrained eco-nomic dispatch. In the integrated locational mar-ket design, the associated locational marginalprices would include the effects of carbon emis-sions. There would be no additional need for an

132 WilliamW. Hogan

incremental green dispatch that could break theconnection between market design and realoperations.

Feed-In Tariffs

Green energy mandates that require system opera-tors to accept energy from approved technologiesat approved fixed prices, known as feed-in tariffs,have been widely adopted in the EU. Not surpris-ingly, with high enough prices, such mandateshave been highly successful at increasing the mar-ket share of approved sources such as wind andsolar installations (Muñoz et al. 2007). However,the structure of the feed-in tariff creates collateraldamage similar to the effects of green uneco-nomic dispatch, and it is not clear whether themodel is consistent with developing sustainablegrowth of low-carbon investments.

Set aside the issue that the new technologiesare expensive and the expense is becoming a mat-ter of policy concern in the EU.11 The feed-intariff interacts poorly with the rest of the marketdesign in terms of system operations, price signals,and infrastructure investment.

The problem with system operations appearsbecause the system operator must take the energyinput, even when the resulting power flow mightforeclose other generation alternatives, except inemergency conditions. In the case of wind, forexample, the variable cost of wind energy may bezero, but this does not mean that it is always cost-effective to accept the wind energy. Conditionsoccur on the grid where the value of the windenergy is negative. In effect, the feed-in tariffforces wind energy onto the grid and drives theimplicit (in the EU) or explicit (in the U.S.RTOs) locational price of energy substantiallybelow zero. This is a regular event in PJM, forexample.

Because the mandate is to take the greenenergy, the burden of investing in transmissionconnections and upgrades falls on someone else,either on the transmission company or by somemethod of cost socialization through the systemoperator. Besides the obvious total cost inefficien-cies, this cost socialization distorts the choicesamong renewable energy resources.

The simple alternative to the feed-in tariffwould be some mix of investment tax credit ordirect subsidies that reduce the cost of investment(but not operations). In order to provide incen-tives for operations, collection of the subsidycould be conditioned on offering the capacity inthe real-time market, whereupon the usual prin-ciples of security-constrained economic dispatchwould apply.

Note that conditioning the subsidy on offer-ing the energy to the market is not the same thingas conditioning the subsidy on the energy actuallyproduced. System operators have good experi-ence in discriminating good-faith offers of energyand need not actually dispatch the plant wheneverit is offered to verify its bona fides. But if thesubsidy is conditional on producing the energy,the effect would be to convert the fixed invest-ment subsidy into a variable opportunity cost. Inthis case, the wind or solar plant could profitablyenter a negative bid (bounded below by minus thevariable subsidy value) to ensure it was dispatched.This would distort prices and exacerbate theproblem of the technology subsidies sending thewrong signals to the market.

Transmission Investment

A major challenge remains in defining the appro-priate protocols for transmission investment. Thegeneral problems of constructing a satisfactoryhybrid system, as outlined above, are magnified bythe perception of many that a major expansion oftransmission capacity is required to build thelong-distance connectors between loads and themajor potential sources of renewable energy(NAS Committee 2009).

The principal implication for green energymandates would be to decide on the degree towhich transmission costs are considered in thecompetition among green energy resources.Which major expansions of transmission systemsare necessary is far from obvious. Solar rooftopinstallations may require relatively little, and cer-tainly different, transmission investments than dis-tant wind or solar installations. Renewable energysources are partial substitutes for each other, andthere is no reason to believe that every renewable

ElectricityWholesale Market Design in a Low-Carbon Future 133

option should be pursued. It follows that themodel for transmission investment and cost allo-cation could have a major impact on the choice ofrenewable technologies. If transmission invest-ment cost is socialized, for everything or evenonly for renewable sources, then distant wind willlook cheaper and nearby rooftop solar moreexpensive, even when the opposite ranking maybe true.

An obvious principle to apply to green trans-mission investment, as for any transmission invest-ment, would be to adhere to a beneficiary-payssystem (Hogan 2009). If the requirement to payfor transmission makes a preferred renewabletechnology more expensive, then this can beaddressed directly in the balance of subsidies. Inthe presence of a carbon cap, with the capaccepted as the proper balance of carbon costs andbenefits, no good reason exists to treat greentransmission investment any differently than othertransmission investment.

ConclusionsPolicy supporting low-carbon technologies andend-use choices in the electricity sector interactswith electricity market design. Great uncertaintyremains about the appropriate investments, andthe consensus is that achieving a low-carbonfuture will require innovation and invention. Suc-cess is more likely if appropriate price signals andincentives appear to drive decentralized decisionsof the many participants in electricity markets.Placing a price on carbon is a critical step. Gettingthe resulting incentives right will be easier thecloser the electricity market design reflects thereality of electricity system operations. The expe-rience with electricity market alternatives pointsto the integrated locational electricity marketdesign as the only workable model that is compat-ible with open transmission access and nondis-crimination.

Notes1. In the U.S. case, vertically integrated utilities are

subject to certain open-access provisions, but these

open-access rules lack the critical ingredient ofnondiscriminatory access to economic dispatch.

2. Likewise, administrative aggregation withoutchanging the underlying grid does not increasecompetition or reduce market power.

3. PJM reports that on average between 1999 and2008, a combustion turbine would have earned43% of its fixed charges in energy net revenue(Monitoring Analytics 2009).

4. The causes of inadequate scarcity pricing aremany, including but not limited to regulatory offercaps on supply (Hogan 2005).

5. “If you know what to do, do it! If you don’t knowwhat to do, decision analysis can help you decide.”Aphorism from Professor Ron Howard of theDepartment of Management Science and Engi-neering at Stanford University.

6. This portfolio effect does not eliminate cases suchas a night when the sun is down nor necessarilywhen even a large area can be becalmed andremove all the wind generation (NERC 2009).

7. Texas Utilities Code § 36.053 (d), available athttp://codes.lp.findlaw.com/txstatutes/UT/2/B/36/B/36.053 (accessed February 20, 2010).

8. Initial designs in the EU included allocations fornew power plants as a set-aside under the total cap.This creates the incentive to build new plants andraises the cost for existing plants by reducing thenet cap for the existing fleet.

9. California Executive Order S-14–08, November17, 2008, available at www.gov.ca.gov/executive-order/11072/ (accessed February 20, 2010).

10. The argument is slightly different for a carbon tax,where the price of carbon is fixed. Then an RESadds to costs by requiring use of more expensivelow-carbon technologies, but it does reduce car-bon emissions.

11. According to ENDS Europe (2009): “The newGerman coalition government will cut feed in tar-iffs for solar power by more than the planned 8%reduction for 2010, according to a 128-pagepolicy document published on Friday [October23, 2009]. The tariffs will now be reviewed everythree years.”

ReferencesAmin, M. 2000. National Infrastructures as Complex

Interactive Networks. In Automation, Control, and

134 WilliamW. Hogan

Complexity: An Integrated Approach, edited by TariqSamad and John Weyrauch. New York, John Wileyand Sons, 263–286.

Anadon, Laura Diaz, and John P. Holdren. 2009. Policyfor Energy Technology Innovation. In Acting inTimeon Energy Policy, edited by Kelly Sims Gallagher.Washington, DC: Brookings Institution Press,89–127.

Boucher, Jacqueline, andYves Smeers. 2002. Towards aCommon European Electricity Market Paths in theRight Direction Still Far from an Effective Design.Journal of Network Industries 3 (4): 375–424.

Budhraja, Vikram, John Ballance, James Dyer, FredMobasheri, Alison Silverstein, and Joseph H. Eto.2008. Transmission Benefit Quantification, CostAllocation and Cost Recovery. Sacramento, CA:California Energy Commission, Energy Researchand Development Division.

Dubash, Navroz K. 2002. The Changing Global Con-text for Electricity Reform. In Power Politics: Equityand Environment in Electricity Reform, edited by N.K.Dubash. Washington, DC: World Resources Insti-tute, 11–30.

ENDS Europe (Environmental Data Services Europe).2009. New German Coalition Confirms Solar Sub-sidy Cuts. www.endseurope.com/22460 (accessedFebruary 20, 2010).

EnerNOC (Energy Network Operations Center). Nodate. EnerNOC: Get More from Energy.www.enernoc.com/ (accessed February 20, 2010).

ENTSO-E (European Network ofTransmission SystemOperators for Electricity). 2008. Regional GroupContinental Europe, StatisticalYearbook. Brussels, Euro-pean Network ofTransmission System Operators forElectricity.

EPSA (Electric Power Supply Association). 2008. EPSAStatement on Renewable Energy in OrganizedElectricity Markets. Press release. Washington, DC:EPSA.

Grubb, Michael, Tooraj Jamasb, and Michael G. Pollitt.2008. A Low-Carbon Electricity Sector for the UK:What Can Be Done and How Much Will It Cost? InDelivering a Low-Carbon Electricity System, edited byMichael Grubb, Tooraj Jamasb, and Michael G.Pollitt. Cambridge, UK: Cambridge UniversityPress, 462–497.

Helman, Udi, Harry Singh, and Paul Sotkeiwicz. 2010.RTOs, Regional Electricity Markets, and ClimatePolicy. In Generating Electricity in a Carbon-ConstrainedWorld, edited by Fereidoon P. Sioshansi. Burlington,MA: Academic Press, 527–563.

HEPG (Harvard Electricity Policy Group). Variousdates. RTO annual reports. www.hks.harvard.edu/hepg/rlib_rp_RTO_ISO_reports.html (accessedFebruary 20, 2010).

Hogan, William W. 1992. Contract Networks for Elec-tric Power Transmission. Journal of Regulatory Eco-nomics 4: 211–242.

———. 1999. Getting the Prices Right in PJM: Analy-sis and Summary: April 1998 through March 1999,the First Anniversary of Full Locational Pricing.Cambridge, MA: Harvard University.www.hks.harvard.edu/fs/whogan/pjm0999.pdf(accessed April 16, 2010).

———. 2002. Electricity Market Restructuring:Reforms of Reforms. Journal of Regulatory Economics21 (1): 103–132.

———. 2005. On an “Energy Only” Electricity Mar-ket Design for Resource Adequacy. Cambridge,MA: Center for Business and Government, John F.Kennedy School of Government, Harvard Univer-sity. www.hks.harvard.edu/fs/whogan/Hogan_Energy_Only_092305.pdf (accessed April16, 2010).

———. 2006. Resource Adequacy Mandates and Scar-city Pricing (“Belts and Suspenders”). Commentssubmitted to the Federal Energy Regulatory Com-mission. Docket Nos. ER05-1410-000 and EL05-148-000. www.hks.harvard.edu/fs/whogan/Hogan_PJM_Energy_Market_022306.pdf (accessedApril 16, 2010).

———. 2009. Electricity Market Structure and Infra-structure. In Acting inTime on Energy Policy, edited byKelly Sims Gallagher. Washington, DC: BrookingsInstitution Press, 128–161.

Holttinen, Hannele, Peter Meibom, Antje Orths,Bernhard Lange, Mark O’Malley, John Olav Tande,Ana Estanqueiro, Emilio Gomez, Lennart Söder,Goran Strbac, J. Charles Smith, and Frans van Hulle.2009. Impacts of Large Amounts of Wind Power onDesign and Operation of Power Systems, Results ofIEA Collaboration. 8th International Workshop onLarge Scale Integration of Wind Power into PowerSystems as Well as onTransmission Networks of Off-shore Wind Farms. October 2009, Bremen, Ger-many.

IEA (International Energy Agency). 2007. TacklingInvestment Challenges in Power Generation in IEACountries: Energy Market Experience. Paris: OECDPublishing.

ISO-NE (ISO New England). 2006. FERC ElectricTariff No. 3, Market Rule I, Section III.2.7.

ElectricityWholesale Market Design in a Low-Carbon Future 135

Joskow, Paul. 2008. Challenges for Creating a Compre-hensive National Electricity Policy. Speech given tothe National Press Club. September 2008, Washing-ton, DC. www.hks.harvard.edu/hepg/Papers/Joskow_Natl_Energy_Policy.pdf (accessed February20, 2010).

Lankford, Craig B., James D. McCalley, and NarinderK. Saini. 1996. Bibliography on Transmission AccessIssues. IEEE Transactions on Power Systems 11 (1):30–40.

Michaels, Robert J. 2008. A Federal Renewable Elec-tricity Requirement, What’s Not to Like? PolicyAnalysis No. 627. Washington, DC: Cato Institute.

MISO (Midwest ISO). 2009. FERC Electric Tariff,Vol. No. 1, Schedule 28.

Monitoring Analytics. 2009. 2008 State of the MarketReport for PJM. Vol. 2, Detailed Analysis.www.monitoringanalytics.com.

Muñoz, Miquel, Volker Oschmann, and J. DavidTábara. 2007. Harmonization of Renewable Elec-tricity Feed-In Laws in the European Union. EnergyPolicy 35: 3104–3114.

NAS Committee (Committee on America’s EnergyFuture). 2009. National America’s Energy Future:Tech-nology and Transformation. Washington, DC: NationalAcademy of Science.

NERC (North American Electric Reliability Corpora-tion). 2009. Special Report:Accommodating High LevelsofVariable Generation. Princeton: NERC.

Neuhoff, Karsten, Sam Fankhauser, EmmanuelGuerin, Jean Charles Hourcade, Helen Jackson,Ranjita Rajan, and John Ward. 2009. StructuringInternational Financial Support to Support Domes-tic Climate Change Mitigation in DevelopingCountries. Cambridge, UK: Climate Strategies,www.climatestrategies.org.

NYISO (New York Independent System Operator).2007. Order No. 890 Transmission Planning Com-pliance Filing. Docket No. OA08-13-000. CoverLetter Submitted to Federal Energy RegulatoryCommission, December 7, 14–15.

———. 2008. Ancillary Service Manual, Vol. 3.11,Draft, 6-19–6-22.

PJM 2009. PJM Open Access Transmission Tariff,Subpart A: Interconnection Procedures. 6th Rev.Vol. No. 1, Para. 36.1.

Potomac Economics. 2009. 2008 State of the MarketReport for the ERCOT Wholesale Electricity Markets.www.hks.harvard.edu/hepg/Papers/RTO-ISO_reports/2008_ERCOT_SOM_REPORT_Final.pdf(accessed April 16, 2010).

Puga, J. Nicolas, and Jonathan Lesser. 2009. PublicPolicy and Private Interests: WhyTransmission Plan-ning and Cost-Allocation Methods Continue to Sti-fle Renewable Energy Policy Goals. Electricity Jour-nal, 22 (10) (December): 7–19.

RECS (RECS International Association). No date.Renewable Energy Certificate System.www.recs.org (accessed February 20, 2010).

Rudkevich, Aleksandr. 2009. Economics of CO2 Emis-sions in Power Systems. Boston: CRA International.www.hks.harvard.edu/hepg/ (accessed February 20,2010).

Schweppe, F. C., M. C. Caramanis, R. D. Tabors, andR. E. Bohn. 1988. Spot Pricing of Electricity. Norwell,MA: Kluwer Academic Publishers.

Sioshansi, Fereidoon P. 2008. Competitive Electricity Mar-kets: Design, Implementation, Performance. Oxford:Elsevier.

Stavins, Robert N. 2007. A U.S. Cap-and-Trade Sys-tem to Address Global Climate Change. HamiltonProject discussion paper 2007–13. Washington, DC:Brookings Institution.

Traber, Thure, and Claudia Kemfert. 2009. Gone withthe Wind? Electricity Prices and Incentives to Investin Complementary Thermal Power Plants underIncreasing Wind Energy Supply. Berlin: GermanInstitute for Economic Research.

van Benthem, Arthur, Kenneth Gillingham, and JamesSweeney 2008. Learning-by-Doing and the OptimalSolar Policy in California. Energy Journal 29 (3): 131–152.

136 WilliamW. Hogan

8

Energy Regulation in aLow-Carbon WorldRichard Green

Increasing the use of renewable energy will havesignificant impacts on the operations of energy

utilities, the companies that deliver electricity andgas to consumers using networks of wires andpipes. As discussed in Chapter 2, many renewableenergy technologies are very different from thetraditional large-scale, centralized technologiesthat utilities have relied on for more than 100years. Electricity transmission and distributionsystem operators will have to become used to tak-ing power from many smaller generators, distrib-uted around the system, while larger-scale renew-able generators are often sited far away from themajor loads. Other renewable electricity may beself-generated by consumers and firms, and elec-tric utilities would then provide a residual serviceor perhaps install and maintain the self-generators.Gas utilities may find that the demand for theirproduct falls significantly as consumers switch torenewable sources of heat.

If the utilities’ operations are changing, shouldthis affect their regulation? Some principles ofregulation, such as the need for transparency andstability, will surely remain unchanged. Whatabout the way in which these principles are putinto practice? Should the scope of regulation, theactivities subject to an economic regulator, bebroadened (or narrowed) as the amount of renew-able generation grows? Should the regulator’s

objectives, or the factors it takes into accountwhen deciding how to achieve those objectives,change? How are its decisions likely to change?

This chapter begins with a discussion of thescope of economic regulation and its interactionwith energy policy. It then presents four models ofeconomic regulation for energy utilities: a retailcompetition model, used in the Nordic countries,the United Kingdom, and some U.S. states (suchas Texas), in which only the natural monopoly indelivery (transmission and distribution) is regu-lated; a wholesale competition model, in whichunregulated generators provide power, but theprices paid by (at least some) retail customers areregulated (as in France, Spain, and some otherU.S. states); a single buyer model, in which aregulated entity is responsible for procuringpower through competitive tenders (as in somedeveloping countries); and an integrated firmsmodel, in which regulators have to oversee mostof the industry’s value chain (as in those parts ofthe United States that have not adopted liberaliza-tion policies).

The chapter subsequently goes into moredetail about how the increase in renewable energywill affect the electricity industry, then draws outsome general implications of these changes foreconomic regulation, which will apply regardlessof the specific model chosen.This is followed by a

discussion of lessons that are specific to the par-ticular tasks facing regulators, which vary accord-ing to the regulatory model chosen. Although theelectricity industry will see the biggest changes,the chapter ends with a look at possible impactson the gas industry and lessons for its regulation.

The Scope of EconomicRegulation and Energy PolicyThis discussion of economic regulation concen-trates on government or government agencyregulation of the prices and service standardsoffered by an industry that is perceived as consti-tuting a natural monopoly. It is thus distinct fromcompetition policy, which applies to industrieswhere competition is believed to be possible;health and safety standards; and environmentalregulation.

Traditionally, the concerns of an economicregulator have been to protect consumers fromhigh prices and protect investors in the utilitiesfrom low prices.This does not mean that consum-ers and investors have diametrically opposinginterests, however. In the short term, it might bepossible to reduce the utilities’ prices without illeffects, but in the long term, this is likely to makethe companies unsustainable, unable to raise thefunds needed to expand or even maintain theircapital equipment. Spiller and Viana Martorell(1996) document what happened when somegovernments in Latin America reduced electricityprices below the level of long-term viability: theeventual outcome was nationalization. Onlywhen new systems of regulation were created,with sufficient protection for investors, was it pos-sible to reprivatize the companies.

From this perspective, the regulator can beseen as balancing two objectives, both of whichare actually in the interests of consumers. One isto minimize the price of energy. The other is toensure a secure supply of energy, and the need togive the utilities sufficient funds can be seen ascontributing toward this objective.To achieve thisobjective, it is necessary to go beyond simply pro-viding them with adequate funds, and beyond the

scope of regulation, as for example the security ofsupply of imported energy is a foreign policy con-cern for many countries. Nonetheless, if theindustry’s capital stock suffers because the utilitieslack resources, the country’s security of energysupply will quickly fall.

An emphasis on security of supply may be anunusual way of framing a regulatory issue, but thisis a key objective of energy policy for govern-ments. The U.S. Department of Energy has fivestrategic themes, the first of which is energy secu-rity, defined as “promoting America’s energysecurity through reliable, clean and affordableenergy” (DoE, 2006, 7). The European Union’s2006 Green Paper on Energy set three core objec-tives: sustainability, competitiveness, and securityof supply.The United Kingdom has four goals forenergy policy, including goals related to the envi-ronment (“to put the UK on a path to cut ourcarbon dioxide emissions by some 60% by about2050, with real progress by 2020”), security ofsupply (“to maintain reliable energy supplies”),and the cost of energy (“to ensure that everyhome is adequately and affordably heated”) (DTI2003).1

Energy policy thus commonly adds a thirdgoal—achieving a low environmental impact—tothe two that have traditionally concerned regula-tors: achieving low prices while ensuring suffi-cient investment to maintain security of supply.These three goals may often appear to conflict andare frequently portrayed as the corners of a trian-gle. Diversifying energy sources can increase secu-rity of supply, but if it involves raising the share ofhigher-cost resources, it will act against the objec-tive of keeping prices down. Adding pollutioncontrol equipment will reduce environmentalimpacts and raise prices. In some countries, coal isseen as a more secure primary fuel than oil andgas, but the carbon dioxide (and local pollutant)emissions from burning coal, and hence its envi-ronmental impact, are higher. At other times,however, a new development, and particularly anew technology, may further all three goals atonce. The introduction of combined cycle gasturbine generators in the United Kingdom in the1990s raised security of supply by diversifyingaway from coal (which then dominated the indus-

138 Richard Green

try’s fuel mix),2 reduced the environmentalimpact from sulfur emissions (and also from car-bon dioxide, which was not then a priority forenvironmental policy), and was also cheaper thanbuilding coal-fired stations when new investmentwas required (Newbery 1999).3 Increasing theefficiency with which energy is used will tend toreduce the cost and environmental impact of agiven level of energy services—that is, heat andlight, rather than gas or electricity. This generallyleads to lower overall levels of energy consump-tion, which will increase security of supply bycutting the amount of primary energy required.4

Models of RegulationThe way an industry is regulated has to reflect itsstructure. In the past, this did not in itself causeany variation in practice, because practically all ofthe energy utilities were either vertically inte-grated monopolies or part of a chain of monopo-lies. Several stages can be distinguished: electricitygeneration (or the production of gas) is the first,followed by high-voltage or high-pressure trans-mission over potentially long distances. Distribu-tion at lower voltages or pressures takes the elec-tricity or gas to the consumer and was tradition-ally defined to include the sales transaction as well.(Full energy liberalization, however, is based onthe idea that retailing, called “supply” in theUnited Kingdom, can be separated from distribu-tion.) Gas utilities might buy their product fromoil companies, and some industrial self-generatorswould sell surplus power to the electricity grid,but the rest of the industry was monopolized.Regulators accordingly concerned themselveswith the final prices charged to end consumersand the intermediate prices used when a regionalor national utility sold gas or power to a smallerfirm for local distribution.

This started to change when competition wasintroduced to the energy utilities. Once it wasrealized that the natural monopoly at the heart ofeach industry covers only its network, then othercompanies could be allowed to use that network,whether they were producing energy or selling itto consumers.The first experiments were for lim-

ited competition in electricity generation in theUnited States, where the Public Utilities Regula-tory Policies Act of 1978 required utilities to buypower from certain kinds of rival generators attheir avoided cost of generation, and in Chile,where the 1982 Electricity Act led to the creationof several regional spot markets (Pollitt 2004).5

The privatization of the gas industry in GreatBritain went further, as independent companieswere meant to compete to sell gas to large con-sumers, and the privatization of electricityincluded a timetable that would culminate inevery consumer in the country having a choice ofelectricity supplier (retailer).

Many other countries followed suit, andthe European Union (EU) adopted a pro-liberalization policy in 1996 and committed itselfin 2003 to introducing full retail competition by2007 at the latest. Despite this commitment, notall member states have achieved, or perhaps donot wish to achieve, the same level of competitionin practice. Similarly, in the United States, 13states and the District of Columbia have imple-mented retail competition, and others participatein regional transmission organizations that runorganized wholesale markets, but nearly half donot (EIA 2009).6 This means that moves toincrease the share of renewable energy need to bestudied in the context of a range of regulatorysystems.

These systems differ in terms of the activitiesthat the regulators have to carry out. Do theyneed to regulate the prices paid by consumers?Are separate prices incurred for using the networkthat need to be regulated, or are the costs of thenetworks simply bundled into the consumers’prices? Are the costs of generation subject to regu-lation?

Figure 8.1 shows four different models ofregulation defined below for the purposes of dis-cussion in this chapter: retail competition, whole-sale competition, single buyer, and integratedfirms.

The first, retail competition, is the minimalistmodel of regulation, in which the regulator’sactivities are largely confined to the naturalmonopoly of the network of pipes and wires andthe system operation. For this model to work

Energy Regulation in a Low-CarbonWorld 139

well, effective competition must exist both inelectricity generation or gas production and inretailing. If retail competition is effective, thenprices to end consumers do not need to be regu-lated, as retailers will be forced to set prices closeto their costs of buying energy, paying to use thenetwork and dealing with their customers. Simi-larly, prices in the wholesale market can generallybe set by the normal processes of competition.The caveat is needed because of the special fea-tures of electricity, a nonstorable product thatmust be delivered through a transmission networksubject to complicated and interacting physicallimitations. The demand for electricity is veryinelastic in the short term, not just because it can-not be stored, but also because most customers donot see the real-time price of power. These fea-tures make it possible for even companies withlow market shares to exercise market power atparticular times or in particular parts of the coun-try, and thus a system of wholesale market moni-toring is often needed to ensure that this is notabused. Even so, this market monitoring shouldbe seen as a special form of competition policy,rather than of economic regulation.

The regulator’s role under the retail competi-tion model is to ensure that the prices for usingthe network are appropriate, and that all compa-nies have equal access to it. If the network owneris still integrated into the other stages of the indus-try, this can be hard to ensure, as an unscrupulousoperator can limit its competitors’ access to thenetwork in many ways.7 The best solution to thisproblem is to separate the role of system operatorfrom any commercial user of the network, either

by banning network owners from the other stagesof the industry or by creating an independent sys-tem operator (ISO) to control the network assetsowned by companies active in production, gen-eration, or retailing. An ISO has the advantagethat it can coordinate operations over a wider areathan a single company, which will be useful if thehistoric pattern of asset ownership has created apatchwork of systems that would be inefficientlysmall if operated separately. At the same time, anISO may not be as effective in pushing through aprogram of capital investment in the networks ofindependent firms as a single network companywould be.

It is worth noting that the problem of dis-crimination among network users has historicallybeen much less important at the distribution levelthan in transmission. Energy sources (power sta-tions or gas fields) are usually connected to thetransmission system rather than to distribution.Even when a large power station was connectedto the distribution system, it was usually the trans-mission operators that controlled its ability tooperate. At the retail level, it was practicallyimpossible for a distribution system operator todiscriminate among customers of different retail-ers in terms of the quality of energy supply, as allthe customers connected to the same segment ofthe network would receive their supply under thesame physical conditions. The regulatory require-ments for keeping distribution networks separatehave therefore typically been weaker than withtransmission, particularly for smaller utilities.8

The second model, wholesale competition,has effective competition in generation, basedaround an active wholesale market, but competi-tion in retailing is limited or nonexistent, so finalprices to consumers are regulated, as well as thecharges for using the networks. This might be atransitional phase; if competition has recentlybeen introduced and is becoming stronger, butmost customers are still with the original incum-bent retailers, then some protection can be appro-priate.9 Some EU countries have so little compe-tition to serve domestic customers that this transi-tional phase may appear permanent, even thoughthe market is officially open. It is also possible as apermanent policy to restrict competition to the

Retailcompetition

Generation

Competitive

Competitive

Tendered

Regulated

Transmission/distribution

Regulatedprices

Regulatedprices

Regulatedprices

Regulatedprices

Regulatedprices

Regulated

Retail

Competitive

Wholesalecompetition

Single buyer

Integratedfirms

Figure 8.1. Models of regulation

140 Richard Green

wholesale market only, if policymakers are uncer-tain of the benefits of retail competition and donot wish to incur the transaction costs of imple-menting it. The wholesale market needs to haveenough separate buyers to avoid a monopsony, buta market covering several regional or subregionalretail monopolies will meet this condition. In theUnited States, Vermont and West Virginia partici-pate in regional wholesale markets (ISO NewEngland and PJM, respectively) without openingtheir retail markets to competition.

A third model, which has been applied in anumber of developing countries, is to have a sin-gle buyer. This is a regulated entity that is respon-sible for procuring electricity from generators, butthrough contracts rather than a wholesale market.The industry may be too small for adequate com-petition among generators, or policymakers maydistrust the market’s ability to deliver enoughcapacity, particularly in a system facing highdemand growth.The use of long-term contracts isimportant, as electricity generation has very highsunk costs, and a producer selling to a singleorganization without the protection of a contractrisks being held up for a lower price once itsinvestment is complete. When the EuropeanCommission designed a single buyer model, itmade it equivalent in its effects to limited retailcompetition, but most countries that actually usethis model combine it with monopolies in retail-ing.

This means that prices to end consumers mustbe regulated, as with the wholesale competitionmodel. It is likely, however, that prices no longerneed to be regulated for transmission and distribu-tion, for these prices may not exist. Generatorswill have agreed to the terms for connecting tothe network as part of their long-term contracts,and as this would be done before an investmentdecision is made, the contracts should provide suf-ficient protection to both parties. If the generatorsknow that they will have to make payments to usethe network, they will seek higher prices from thesingle buyer.

Although there may not be formal prices forusing the network, the regulator will still have tooversee its revenue requirements, ensuring thatthe amount recovered from consumers is reason-

able. Figure 8.1 thus describes this part of theindustry as regulated without specifying that it hasregulated prices.The cost of generation should bekept down by competitive tendering, and thenpassed on to consumers (just as adequate compe-tition in a wholesale market should keep downthe amount that regulated retailers could chargein the second model).

The fourth and final model of regulation isone of integrated firms.This does not mean that avertically integrated monopoly is responsible forevery stage of gas or electricity supply. It may wellbe that vertical separation exists between the utili-ties responsible for energy production (or import)and transmission, and those responsible for distri-bution and retailing. The German electricityindustry has a three-layer system of intercon-nected, regional, and municipal utilities, with theinterconnected utilities owning almost all of thetransmission system and much of the generation,but regional and municipal utilities deliveringenergy to about two-thirds of customers (Schulz1995).The model is also compatible with a utilitythat buys some of its gas or power from thirdparties, but without the separation implied by aformal single buyer. Independent power produ-cers that sell power as their main businesses wouldneed the protection of long-term contracts, butindustrial self-generators with occasional sur-pluses of power could sell these under short-termdeals.

In this model, generation is now regulated, asare the networks.The regulator must oversee finalprices for all consumers, and there may be nointermediate prices to regulate, although if a largeutility sells gas or electricity to a smaller one, theregulator may also need to oversee this transactionto ensure that excessive costs are not passed on.

The main challenge for regulators—under anymodel—is to ensure that prices are high enoughfor the utility to recover efficiently incurred costs,but not so high that consumers are exploited.Theregulator can generally measure what a utility’scosts actually are (assuming that competentaccountants are available, and with a caveat con-cerning the cost of capital), but it is a differentmatter to measure what the costs could be. Whenjust one firm is being regulated, the regulator can

Energy Regulation in a Low-CarbonWorld 141

only extrapolate from past trends, adjust for pre-dicted changes in the future, commission effi-ciency studies from consultants, and perhaps makecomparisons with firms in other jurisdictions,hoping that their environments are sufficientlysimilar to make the comparisons valid. When theregulator must deal with several firms, it is possi-ble to make comparisons among them on theprinciple of yardstick regulation (Shleifer 1985). Ifthe firms face sufficiently similar operating envi-ronments, then there should be no reason why allfirms could not reach the level of the best per-formers, and their prices should be set to reflectthis. Good performers could expect to receivemore than the cost of capital, whereas firms withhigher costs would not be able to do so until theycut those costs. In practice, some of the differ-ences in operating environments will be signifi-cant, and regulators will have to adjust for these(Jamasb and Pollitt 2007; Pollitt 2005). The diffi-culty of doing so is why competition, where pos-sible, generally gives better incentives to firmsthan does even the most enlightened regulation.

Utilities in a Low-Carbon WorldHow will the move to a lower-carbon energy sys-tem affect the energy utilities? Many countries, aswell as individual U.S. states, have adopted targetsfor renewable energy, and the European Unionnow has a legally binding target for 2020 that 20%of its final energy consumption shall come fromrenewable sources, as discussed in Chapter 12.Given the resources available for renewable heatand transport, and their costs, the proportion ofrenewable electricity will need to be well above20%.

Humans have been using renewable genera-tion for as long as the electricity supply industryhas existed—the world’s first public electricitysupply, installed in 1881 in the English village ofGodalming, was hydroelectric (Hannah 1979). Anumber of countries, both developed and devel-oping, get much or all of their electricity fromhydropower. Most developed countries, however,have developed most of their available large-scalehydro resources, prohibiting some potential

developments for environmental reasons, thoughopportunities still exist to develop small-scaleschemes. Further increases in renewable energytherefore are likely to come largely from burningbiomass (including waste) and from wind power.In some countries, solar photovoltaic power maybe a sensible option, though not in northerly, fre-quently cloudy regions like the UK; other coun-tries are surrounded by seas containing significantenergy in their currents, tides, and waves. Devicesto harvest this energy are being developed but arestill mostly in the prototype stage.10

Most renewable generation will cost morethan conventional alternatives, excluding exter-nalities, unless the price of fossil fuels rises dra-matically. In most cases, this is because of highcapital costs—several renewable technologies gettheir “fuel” for free. Biomass generation is themain exception for which fuel costs are a signifi-cant part of the whole. Large-scale hydroelectricgeneration was generally a low-cost energysource, although few sites remain available, andwind turbines at particularly good, windy loca-tions onshore appeared quite competitive withfossil-fueled power stations at the high energy andcarbon prices of early 2008. These high costsmean that companies will not be willing to buildmost types of renewable generators unlessrequired or they are offered financial support.Chapter 6 and the various country studies discussthe mechanisms that can be used to achieve this.

The EU target for renewable energy leads to awide range of national targets, related to eachcountry’s income and existing level of renewableenergy, and the mix of renewable sources usedwill vary among countries. Table 8.1 gives a set ofestimates prepared for the European Commission,giving a mix of sources predicted to minimize thecost of meeting the 20% target for renewableenergy.

The use of biomass and hydroelectricity willincrease, but their percentage share of renewablegeneration would fall. Solar and marine energywould increase from very low levels at present. Byfar the greatest increase would come from windgeneration, which might provide more than one-sixth of Europe’s electricity in 2020.11

142 Richard Green

Table 8.1. EU electricity generation from renewablesources

Generation in 2006 Potential for 2020,Green-X model,least-cost scenario

TWh Percentage TWh Percent-age

Biomass 90 17% 186 14%

Geothermal 6 1% 7 1%

Hydro 344 66% 398 30%

Solar 2 0.1% 62 5%

Marine 1 0% 124 9%

Wind 82 16% 545 41%

Total 525 100% 1,322 100%

Source: European Commission 2007Note: TWh = terawatt-hours

This level of wind generation will lead to sig-nificant challenges in dealing with intermittency.Large-scale wind generators reach full outputonly at wind speeds between around 12 and 25meters per second (m/s), shutting down at higherspeeds and losing power rapidly when the wind islow.12 A well-sited turbine might have an averageoutput equal to 30% of its rated capacity over theyear, but the actual output will vary sharply overtime, and meaningful predictions are impossiblemore than a few days in advance. In regions whereclouds are rare, the main factor affecting the out-put from solar power will be the earth’s predict-able rotation (although in regions with variablecloud cover, this will have a significant effect onsolar output), and tidal energy can be predictedyears in advance but will sometimes be highest atoff-peak times. Biomass generators can run whenthe user chooses, but if these users are householdsor small businesses with combined heat and powerunits, they may choose a heat-led operating pat-tern that does not necessarily produce power atthe times when it would be most useful for theelectricity industry (Hawkes and Leach 2007).

Intermittent outputs from renewable genera-tors cause two main problems. First, the systemoperator will need to keep more plants in reservefor times when the level of renewable output sud-denly falls. This particularly applies to wind out-

put, as most other renewables are more predict-able over the relevant timescales (of a fewhours).13 Gross et al. (2006) have estimated thatthese reserves would cost £2 to £3 ($3.05 to$4.58) per megawatt-hour (MWh) of intermittentrenewable energy if the United Kingdom were toget 20% of its electricity from wind. Note that thesystem operators always have to keep some plantsin reserve in order to respond instantly if a powerstation fails, but it has traditionally been sufficientto set this reserve equal to the size of the largestunit running on the system—the worst case for asingle failure. If the wind drops over a large area,many separate wind turbines, with a greater col-lective capacity, could be affected. The level ofreserve generation (or demand reduction poten-tial) needed to cover this risk naturally grows withthe amount of wind generation, and the systemreserve requirement is the greater of that requiredby wind stations and by conventional plants. Therules of the system—either those of the wholesalemarket or the regulations of the system operator,depending on where the responsibility for provid-ing these reserves lies—will have to ensure thatthese can be paid for.

The second problem is that the wind or tidemay not be available at the time of peak electricitydemand, which is what determines the totalamount of capacity needed on the system. Not allof the capacity of a conventional plant can becounted against this need, as it may be unavailableat the critical time, but it is generally sensible toassume that more than 80% will be. The risk thatmany wind generators will be becalmed at thesame time is greater, however, and the effectivecontribution of a portfolio of wind power stationsto ensuring ability to meet peak demand has beenestimated at only 20% (or less) of their total cap-acity.14 This means that 10 gigawatts (GW) ofadditional wind power could replace only 1 to 2GW of conventional stations, in terms of planningto meet the peak demand. The energy producedby the wind stations could actually replace theoutput of around 3.5 GW of conventional plants,if the time of delivery were irrelevant.15 In otherwords, between 1.5 and 2.5 GW of conventionalplants would be needed per 10 GW of wind,because we cannot know whether wind stations

Energy Regulation in a Low-CarbonWorld 143

will produce when we most want them to. Grosset al. (2006) predict that this would cost £3 to £5($4.58 to $7.63) per MWh of wind energy in UKconditions. The wider the geographic dispersionof the wind stations, the lower this penalty willbe—as long as this dispersion reduces the correla-tion among their outputs and enough transmis-sion is available.

Note that both of these problems are far lesssevere in a system with, or connected to, largeamounts of hydro generation with reservoir stor-age. The output from hydro generators canrespond inversely to the output from wind, stor-ing energy in the form of water until it is needed.The cost of short-term reserve to cover variationsin wind output is very low, as it costs almost noth-ing to keep a hydro station ready to increase out-put at short notice. At most hydro stations, thereservoir is not large enough to keep the genera-tor running throughout the year, so the systemcan provide enough energy over the year as awhole only if it has more than enough capacity tomeet the peak demand. In this context, the riskthat a wind generator will not be available at thepeak time is less important.

In a system with little hydro generation, theinevitable consequence of increasing the variableoutput from wind generators is that the outputfrom conventional generators will also tend to bemore volatile, except to the extent that demand-side response takes up the slack. Sometimes theoutput from wind stations will rise at the sametime as demand, reducing the increase requiredfrom conventional stations; at other times,demand and renewable outputs will move inopposite directions, increasing the strain on thesystem. There will also be some stations that areneeded for the times when high demand levelscoincide with low wind speeds but do not runvery often. It will be important to ensure thatthese generators can cover their costs.

Demand response can reduce the need forpeak capacity, cutting the industry’s costs(Borenstein 2005). Some response can be madewithout consumers noticing, such as delaying thestart of water heaters or refrigerator motors whenprices are particularly high, or responding to someother signal such as system frequency. Some

industrial customers have long been used toscheduling energy-intensive processes away fromthe times when power is most expensive, althoughthis rescheduling is likely to involve some costs.The likely growth of electric vehicles offers agood opportunity to increase the “unnoticed”element of demand response, simply by ensuringthat the vehicles mainly charge themselves duringlow-demand periods overnight, although it willbe important to ensure that the vehicle is suffi-ciently charged by the time it is needed. Electricvehicles have a significant advantage over mostother forms of energy storage, in that the batteryis needed in any case, and this kind of use by theelectricity industry is essentially a free good. Carbatteries could also be discharged when theindustry is particularly short of power, but impos-ing additional cycles of deep discharging andrecharging would shorten the battery’s life. Giventhe cost of batteries, this would be an expensivesource of electricity.

Moving on to network issues, in many coun-tries significant investments will be needed toaccommodate an increase in renewable generation(Pollitt and Bialek 2008). Many of the best sitesfor renewable generators are remote from thecenters of demand, and transmission lines will beneeded to move the electricity to where it isneeded. Most renewable generators are on a muchsmaller scale than conventional centralized powerstations, so many separate schemes, each with itsown grid connection, may be needed to replace asingle large conventional station. Because therenewable schemes are generally smaller, theirconnections will often be to the distribution net-works rather than the transmission system. Thiswill certainly be the case where self-generation orcommunity-owned small-scale generation is con-cerned.

Connecting large amounts of generation tothe distribution networks will require significantchanges in their operating practices. Broadlyspeaking, transmission networks have always beenoperated to accept two-way flows across everyline in the system, but distribution systems havegenerally been designed to take power in onedirection only—from the connection with thetransmission system to the customer.16 If power

144 Richard Green

always flows in one direction, the design of thesystem is simpler, and protection equipmentagainst power surges is needed on only one side ofthe substation it is intended to protect. If powercould suddenly flow from a generator located onthe customer side of the substation, then extraprotection equipment is needed, and it will benecessary to manage the flows on the distributionnetwork in a much more active manner. Thisrequires additional equipment to monitor condi-tions in real time and operating practices thatrespond to them. Local generators and loads mayhave to change behavior in order to keep powerflows within acceptable limits. Greater monitor-ing, however, can also change those limits. Forexample, windy conditions allow the heat frompower flowing through the network to dissipatemore easily, increasing the safe operating limit atthe very times when the output from nearby windfarms will be at its greatest (Douglas 2005).

Finally, it is worth mentioning another changethat is not directly connected to the growth inrenewable energy but is clearly linked to the movetoward a low-carbon energy system. This is theadoption of carbon trading, first in the EuropeanUnion, and now in parts of the United States.17

The need to buy permits raises the marginal costof generation from fossil fuels, even if the genera-tor receives those permits for free.18 This will tendto raise the wholesale price of power produced bya given capital stock; in a fully competitive mar-ket, the price would rise by the price of emissionsmultiplied by the carbon content of the marginalplant.

In the longer term, the level and compositionof the industry’s capital stock may change becauseof carbon trading. First, if generators receive somepermits free of charge, contingent on the exist-ence of a power plant, then the value of thosepermits can be offset against the cost of keepingthe station open. This may allow the station toremain open when its revenues from the electri-city market alone would not justify this. Thiscould lead to a market equilibrium with a greateramount of capacity, which would tend to depressthe price of power, partly offsetting the impact ofthe carbon price (Green 2005).

A more significant change to the capital stockshould be that compared with high-carbon powerstations, low-carbon generators become relativelymore attractive. Natural gas will gain relative tocoal, and there is already talk of nuclear powerenjoying a renaissance in the United States andUnited Kingdom. Many countries are promotingdemonstration schemes for carbon capture andsequestration and, once the technology is proven,could compel generators to fit this to new plantsand perhaps even retrofit it to existing ones. If thecarbon price rose to a sufficiently high level andwas expected to remain there, carbon capturewould be economic without government com-pulsion. The price within the ETS, however, hasnever reached this level, let alone stayed there longenough to make such investments attractive.

This also applies to renewable generators. Ifthey receive a wholesale price that has beenincreased by the cost of carbon emissions, this willreduce their need for additional support. Likelylevels of carbon prices, however, are unlikely tomake most renewable generators competitivewith gas-fired or nuclear generators, at least giventhe costs expected in the United Kingdom(DECC 2009a).

General Implicationsfor RegulationHow do these trends affect the task of regulation?First, it is clear that the cost of energy is going torise. Most renewable energy costs more than con-ventional sources at current fuel prices, evenbefore considering the costs of investment in thenetworks and changed operating requirements. Iffossil-fuel prices were to rise significantly, then amove to renewable energy would not lead to afurther increase, and might in extreme casesreduce costs, but this would be little comfort toenergy consumers. The EU has estimated thatmeeting its 20% target for renewable energy willcost an extra €24 billion to €31 billion ($32.7billion to $42 billion) in 2020, against a predictedenergy bill of around €350 billion ($476.7 billion)(European Commission 2007). This assumes an

Energy Regulation in a Low-CarbonWorld 145

oil price of $48 per barrel; if the oil price rises to$78 per barrel, energy costs will be much higher,but the additional cost of renewable energy willfall to at most €11 billion ($15 billion) and wouldbe zero in one scenario. Some of these costs maybe borne by taxpayers, but the remainder willhave to come from energy consumers, and regu-lators will need to allow the utilities to collectthese—as with any other efficiently incurred costincrease. Where parts of the industry have beenliberalized, the regulators will not determineprices directly, but regulators that have oversightof these competitive areas must not intervene tosuppress necessary increases.

Second, the electricity industry will need a lotof capital investment, both in renewable genera-tion (many technologies are capital-intensive) andin changes to the network to accommodate it.This will affect the weight that regulators shouldplace on the different cost components when set-ting incentives for network companies and forgenerators, if they are regulated. A regulated firm’scosts can be broken down into its cash operatingcosts (labor, materials, and bought-in services),the return on its new investments, and the returnon its existing assets. The firm is assumed able tocontrol both its operating costs and its investmentin new assets, but its existing capital stock is, natu-rally, fixed. The required return on both new andexisting assets depends on the cost of capital,which is affected by the perceived risk of the busi-ness.

High-powered incentives to reduce costs willincrease the company’s risk, as success or failure todo so will feed through to the company’s profits.Standard models of financial economies, such asthe capital asset pricing model, suggest thatdiversifiable risks should not affect a company’scost of capital. Nonetheless, the level of incentivesmight have an impact through two routes. First, ifthe company’s profits become more sensitive tothe overall state of the economy, and are positivelycorrelated with it, then its beta will rise, increas-ing the cost of its equity. Second, if favorable taxtreatment reduces the cost of debt and a companywith risky cash flows has to reduce its gearing, thiswill increase its weighted average cost of capital.

An optimal regulatory scheme would balancethe expected savings from strengthening theincentive on the volume of operating costs, or ofinvestment, against the cost of having to pay ahigher expected rate of return. In other words,although it is possible to give the firm very high-powered incentives to reduce both its operatingcosts and its investment (such as a pure price capregime does), this would be undesirable if itcaused too great an increase in the cost of capital.If the industry’s cost structure shifts away fromoperating costs, then optimal regulation wouldimply paying more attention to the rate of returnon capital and less to the level of operating costs.An increase in the volume of investment willmean that more weight should be placed onincentives for efficient investment than on theimpact of a higher cost of capital on payments forexisting assets. In other words, incentives for effi-cient operation might be relaxed, and those forefficient investment should be strengthened.

One way of doing this in practice is to offerthe firm a menu of regulatory contracts (LaffontandTirole 1993; Sappington 1982). In the UnitedKingdom, the regulator has given distributioncompanies such menus for their investments overthe next price control period (Pollitt and Bialek2008). The regulator uses consultants to assess thenecessary level of investment, but the firm canpropose a higher base figure. The firm’s allowedrevenues are then linked to this proposal and itsoutturn expenditure. Whatever the outturn levelof expenditure, the firm’s revenues are highest ifthis was the level that it actually proposed, provid-ing an incentive for truthful cost revelation. Theallowed revenues increase by less than a unit foreach unit increase in the firm’s actual expenditure,however, so the firm has to bear a share of any costoverrun and keeps a share of any savings. Thisprovides an incentive to keep actual costs down,and one that is greater the tougher the target thatthe firm accepts. Accepting a tough target thusallows the firm the chance to benefit from high-powered incentives. A firm that accepts and meetsthe regulator’s target will be allowed revenues ofmore than the amount it spends, which gives amargin of error in case the regulator’s target is tootough. Under Ofgem’s scheme, which differs

146 Richard Green

from some versions of the menu of contractsapproach, the allowed revenues also increase byless than a pound for each pound that the firmproposes to spend above the regulator’s target.Thegreater the firm’s proposal, the less revenue thefirm will be allowed for any outturn spendingbelow this proposal.19 Overall, the scheme givesthe firm an incentive to reveal the truth about itsexpected capital needs and to keep its costs down,while allowing it to pass part of any genuine over-run on to consumers.

A final general implication is that the numberof small electricity generators that wish to injectenergy into the network will increase. Any energyproducer connected to the network must be sub-ject to the network operator’s technical standardsand the industry’s commercial arrangements—purely bilateral transactions are impossible in aninterconnected system. Energy regulators areoften involved in overseeing the details of suchstandards and arrangements to ensure that thesedo not discriminate against particular networkusers. If we wish to further the development ofmicrogeneration (and the high costs of manymicrogenerators should be taken into account inthat decision), then the regulators will have toensure that the technical standards are not tooonerous and the commercial arrangements aresuitable for nonspecialists.

Accommodating these generators will behelped by a variety of so-called smart grid tech-nologies. Traditionally, system operators have hadlittle real-time information on the actual state ofthe distribution network, which forces them torun it in a relatively conservative manner. Moreinformation can be used to increase the allowablepower flows. Furthermore, some small-scale gen-erators and loads can be dispatched, responding toinstructions based on the state of the system—both the overall balance between generation anddemand and conditions on the local distributionnetwork. Regulators will have to ensure thatcompanies are rewarded for innovations such asthese, and that the boundaries among generation,retailing, and networks do not act as barriers.Thisis potentially a disadvantage of the retail competi-tion model, which sometimes severs the relation-ship between a customer and its network operator

(although in some countries, the customer con-tinues to receive a bill for network services).

The need to give incentives for innovation hasemerged as a key theme of the British regulatorOfgem’s 2008–2010 review of network regula-tion, known as RPI-X@20.20 Spending onresearch and development by the distribution net-work operators in Great Britain has risen dramati-cally since a separate revenue stream to fund it wascreated in 2005 (Ofgem 2009). One key issue isthat spending may have to be incurred well beforecustomers see any benefit from it, and regulatorsmust allow companies to fund this.

Implications for theRegulator’s TasksHow will the general implications of an increasingshare of renewable energy change what regulatorsneed to do? The following analysis is brokendown by the major tasks regulators must under-take: regulating the wires, regulating generationrevenues, and regulating retail prices.

Regulating the Wires

In both the retail and wholesale competitionmodels, the regulator has to oversee the prices,and other terms, for companies wishing to use thenetwork; in the retail competition model, this isthe only area of ex ante regulation. This mayinclude the rules of the wholesale market (dis-cussed in Chapter 7). A liquid multilateral marketwith prices that reflect the changing state of sup-ply, demand, and the transmission system will befar more effective than a bilateral market withprices that ignore the fact that generators in somelocations may be unable to get their power toconsumers. The transmission system operatorsalready have to undo some trades of this kind tomake the European wholesale markets workproperly, and the scale of the task will increasewith the level of renewable generation.21

Price regulation has two key aspects. One isthe average level of prices, and the other is theirdistribution among customers. As discussed above,

Energy Regulation in a Low-CarbonWorld 147

the average level of prices is linked to the indus-try’s expected costs, and a key aspect of these is thelevel of investment. Investment in the electricitynetworks will need to rise, increasing costs andhence prices. Strong incentives to minimizeinvestment costs might keep prices down butcould jeopardize investment in renewable energyif insufficient grid capacity is available. For anextreme example, new generators inquiring aboutconnecting to the transmission system in Scotlandwere told that they might have to wait more than10 years for capacity to become available (Houseof Lords 2008a, para. 133).

The primary cause of such delays is the longlead time for transmission investment, typicallygreater than that for generation. In part, this timedepends on the process of obtaining planning per-mission for new transmission lines, which hasbecome a lengthy one in many countries. Some-times it will not be possible to develop a line intime to meet an unanticipated need for capacity,and the regulator (and would-be generators) willhave to accept this—although reforms to speedthe planning process would make this less com-mon.

The impact of this lead time is worsenedbecause it will not always be clear in advanceexactly where capacity will be needed, or howmuch, even if the general requirement is clear. Ifthe network owner is worried that it will bepenalized for spending that turns out to be unnec-essary, it will not wish to start the detailed invest-ment process until the needs are clear and, prefer-ably, contracts are signed with users. This is likelyto impose significant delays, which would beunnecessary in cases where it had earlier becomeclear that some investment was likely to berequired. Regulators should allow companies torecover the cost of making plans, and even seekingplanning permission, for new lines that appearlikely to be needed but are later abandoned ascircumstances change. To reduce the risk that toomany unwanted lines are planned, wider consul-tation should be done on network investmentplans. Pollitt (2008) calls for a process of “con-structive engagement” already used abroad and inthe regulation of UK airports, where users andthe airport owner could jointly agree on a pro-

gram of capital investment. Littlechild (2008)describes the negotiations used to decide ontransmission investments in Chile, where trans-mission users were able to insist on a cheaperalternative to the line originally proposed by thetransmission utility. A balance is needed betweenmaking a timely start to investment and notauthorizing projects with insufficient demand.The California Independent System Operator’sLocation Constrained Resource InterconnectionPolicy requires financial commitments fromrenewable generators to use at least 60% of a newline’s capacity, but then incorporates the line’scosts in its general transmission tariff.22 A moreprescriptive approach would imply the creation(or empowerment) of some kind of national plan-ning agency to coordinate generation and net-work investment and act as a “guiding mind,” asOfgem puts it in its RPI-X@20 review.

It is also important to ensure that the invest-ment needs in the grid are not increased by inflex-ible procedures. Traditionally, transmission com-panies have been reluctant to connect generatorsunless they are certain they can almost alwaysaccept their output. This is a sensible policy forconventional generators, as, if the investment isdesirable, they can expect to be needed to meetpeak demand, and preventing them from generat-ing could risk power cuts. Wind generators, incontrast, may not generate at the time of the peak;if they do, some other stations, which need to beavailable, will not be running. The transmissionsystem does not need to accept every generator’soutput simultaneously, and investment rulesshould be altered accordingly. It is important,however, to allocate access to the gridefficiently—something achieved automatically ina wholesale market based on locational marginalprices.

The second aspect of price regulation is thedistribution of payments among network users. Inparticular, how should they vary over time andspace? A partial answer is that if the wholesalemarket adopts nodal marginal pricing, then cost-reflective variations will occur over time andspace, and these will be passed on to networkusers.The prices at nodes with net generation are,on average, below those with net demand,

148 Richard Green

because the costs of transmission are incurred aspower flows from generation to demand nodes.The transmission operator can therefore “buy lowand sell high,” retaining the surplus. The problemis that other charges will have to be added, forthese net revenues from nodal marginal pricesgenerally recover just a small part of the total costof transmission—often only around 20% (Pérez-Arriaga et al. 1995). It is possible to derive anoptimality condition for transmission investment,under which the marginal cost of expandingtransmission capacity should just equal the mar-ginal benefit of doing so, which is given by thedifference in prices at each end of the line. Withconstant returns to scale and perfectly divisiblecapital, this suggests that nodal pricing wouldcompletely recover the costs of an optimal system.Perez-Arriaga et al. suggest that this result doesnot hold because of economies of scale andindivisibilities, because it is often sensible to builda line that is currently too large, in order toaccommodate future growth, and because chan-ging patterns of generation and demand meanthat the inherited transmission system will inevi-tably be suboptimal.This suggests that the missingrevenue should not be seen as related to identifi-able costs of transmission, but as a lump sum thathas to be recovered with the minimum of distor-tion.

Given this, how should the remaining net-work costs be recovered across time? Should gen-erators (and customers) pay whenever they aresending power to (or taking power from) the grid?Should they pay based on the maximum cost thatthey could impose on the grid or based on thecost they impose when the load on the grid isheaviest? Given that capital costs form a high pro-portion of the total, this would argue for someform of peak load pricing, rather than recoveringcosts evenly over time. Peak load pricing is basedon charging users at the time when the load onthe grid is heaviest, as capacity built to cope withthese times is effectively “free” at times with lowerloads.

The issue is that with this kind of peak loadpricing, renewable generators would tend to facemuch lower transmission payments per kW ofcapacity than would conventional generators, but

those payments would be variable over time. Insome years, a renewable generator would be ableto generate at the time of the system peak andwould be assessed for transmission charges on thebasis of this. In other years, the system peak wouldcome when the local wind was low, and the gen-erator would have little output and would paylittle for transmission.This is unlikely to be politi-cally acceptable and could also produce perverseincentives not to generate at the system peak—transmission charges per MW can be an order ofmagnitude above all but the very highest electri-city prices, and if just one or two hours of genera-tion creates a liability to those charges, it will notbe worthwhile.23

This suggests that transmission charges shouldbe based on either the capacity of a power stationor the energy it generates over the year, perhapsweighted according to the time of production. Ifthe charges are capacity-related, intermittentrenewable generators will pay much more perMWh actually generated than conventional sta-tions able to run at base load. Conventional gen-erators that are needed to meet peak demands, butrarely run at other times, will be in a similar situ-ation. Both may claim this to be discrimination.Furthermore, they will need some revenue sourceto cover these costs if they are to remain availablein a long-term equilibrium. For peaking genera-tors, this is likely to be reflected either in theprices for peak power or in a capacity market, ifone exists. For renewable generators, less able torely on peak prices or capacity markets, theremainder will have to come from their supportscheme. Charging intermittent renewable genera-tors on the basis of their capacity, but adjusted fortheir relatively low expected output, might be apolitically acceptable way of reducing the cost ofthis—effectively, using a hybrid scheme.

If transmission charges are energy-related, theimpact on peak prices and the cost of supportingrenewables will be lower, though electricity priceswill be slightly higher throughout the year as gen-erators pass through this additional marginal cost.Furthermore, wholesale prices may not accuratelyreflect the marginal cost of expanding transmis-sion to accommodate more generation (even withnodal marginal prices, indivisibilities might mean

Energy Regulation in a Low-CarbonWorld 149

that the system has to be expanded to the pointwhere the price differences at each end of the lineare quite low), or wholesale prices may notinclude any transmission component. In this case,the transmission charges should signal the cost ofexpanding the system, and capacity-relatedcharges will send a more accurate signal thanenergy-based ones.

Similar considerations apply to the decisionabout how network charges should vary overspace. The overall system of transmission charges(including transmission-related elements in anywholesale market) should signal the costs of usingthe transmission system and ensure that the net-work companies recover their efficiently incurredcosts (Green 1997). If other parts of the pricingsystem send adequate locational signals, so thatnetwork charges are purely to recover a lump sumof missing revenue, then a uniform national orsystemwide charge is appropriate. If generatorsand customers do not otherwise face charges thatreflect the spatial variation in transmission costs,however, then transmission charges should varyover space.

The implications of this may depend on fac-tors specific to each market. In Great Britain, forexample, transmission charges are highest in Scot-land (which is bad news for wind farms sitedthere) and lowest in southwest England (whichwill be good news for marine energy schemeslocated off the coast of that region). In general,however, it seems likely that renewable generatorslocated in rural areas will face higher transmissioncharges than conventional generators closer to theloads. Once again, this will affect the cost of sup-porting renewable energy. The higher the trans-mission charges paid by the marginal generator(the one whose needs just determine the level ofsupport required), the greater the support it willneed. Unless the support scheme separates outand pays each generator’s individual transmissioncharge, this will raise the rents obtained by gen-erators in more favorable locations.

Transmission charges may not, in fact, havemuch impact on the locational choices made byrenewable generators, which need to choose sitesthat maximize their output. A 1% difference inthe load factor achieved by a British wind farm,

for example, would largely offset the difference intransmission charges between the north and southof Scotland or central and southwest England.24

In this case, minimizing the rents obtained bywell-sited generators might be more importantthan sending precise signals. However, the impacton conventional generators must also be takeninto account. These have more flexibility tochoose their sites and should be sent signals oftransmission costs—either through spatially vary-ing wholesale prices or through transmissioncharges.

The final issue involved in regulating the wiresis how to regulate the transmission system opera-tor, as distinct from the transmission owner. Thetransmission owner receives payments for havingbuilt the transmission wires and keeping them ingood repair; the transmission operator is responsi-ble for managing their daily use to move power.Sometimes both functions are combined in a sin-gle company, but an independent system operatorcan be used when it is most effective to run asingle market over several transmission owners’grids or when regulators want to ensure that thegrid is operated independently of generationinterests, without ordering an integrated com-pany to divest any assets.

The system operator will need to buy operat-ing reserves and, in markets without nodal pri-cing, will resolve congestion by buying and sellingpower on either side of a constraint. As theamount of renewable energy increases, so willthese costs. This implies that the benefits fromgiving the system operator strong incentives tocontrol them are likely to rise. It may be difficultto give strong financial incentives, relative to thecosts involved, to an independent system opera-tor, for organizations like these have few assets tocushion them against unforeseen events thatwould lead to a bad outturn, and hence perform-ance penalties, even in a single year. If significanttransmission assets are owned by an independentcompany that is also the system operator, how-ever, then the revenue stream for these assets pro-vides a suitable cushion against which tostrengthen the system operator’s incentives.25 Theincrease in renewable energy therefore providesan additional argument for establishing independ-

150 Richard Green

ent transmission companies, as opposed to inde-pendent system operators.

Regulating Generation Revenues

In the model of integrated firms, the regulator isresponsible for overseeing each stage of the indus-try’s value chain, from generation onward. It is nolonger necessary to oversee a transmission or dis-tribution tariff, as there would be no independentgenerators or retailers to use it, but the issues sur-rounding investment in the networks remain.Theone simplifying factor is that the regulated firmcan coordinate investment in generation and thenetworks internally, rather than rely on the pricemechanism to persuade generators to favor sitesthat are helpful to the transmission company.

Against this, the regulator now has someresponsibility for the level of investment in gen-eration. As discussed earlier, most forms of renew-able generation are good substitutes for conven-tional generators in providing energy (measuredin MWh of power at some point during the year)but are poor at providing capacity (measured inMW of ability to provide electricity when itis really needed). Regulators will have to allowutilities to add renewable generation to theirportfolios without retiring many conventionalstations. If the conventional fleet is aging, newconstruction of both renewable and conventionalpower stations may be needed. The regulator willthen have to accept that some of the conventionalcapacity will rarely be required—it is needed toensure that peak demand can be met when thiscoincides with low availability of renewable gen-erators, and this will not happen every year. Toomuch pressure to demonstrate that reserve stationsare “used and useful” will make utilities reluctantto keep this reserve at a suitable level.

Regulators could also face challenges of a dif-ferent kind if they face a rapid rise in renewableoutput when conventional capacity was alreadyadequate. In this case, some of that capacity trulywould be surplus to requirements. Should theregulator eliminate it from the rate base so thatconsumers do not have to pay for assets they arenot using? The regulatory compact adopted inmany jurisdictions that have liberalized the energy

industries has been to allow utilities to recover thecost of such stranded assets, in the interest of long-term investment incentives. This would also beappropriate in this case but will increase theapparent cost of renewable energy. With a liberal-ized wholesale market, the cost of stranded assetsis absorbed by their owners, but the risk thatsomething like this would happen should havebeen taken into account when those owners cal-culated their cost of capital, which would nor-mally be higher than in a regulated system.

The cost of generation will also be affected bythe cost of carbon in countries with a system ofcarbon trading. If generators have to buy all theirpermits, then this will be a cost of business thatthe regulator should include in their allowablerevenues. If generators receive some permits forfree, however, then the regulated firm’s revenuerequirements should take this into account, mut-ing the increase in price. Politically, this may be anadvantage, but it implies that the impact of carbonpricing will have to come entirely from the supplyside, rather than from both supply- and demand-side responses.

It is worth pointing out that where generationis traded in a wholesale market, regulators areoften responsible for market monitoring, eventhough this is really an aspect of competitionpolicy rather than of economic regulation. Mar-ket monitors need to know when to intervene tostop the abuse of market power, which typicallymanifests itself as high prices. The challenge inelectricity markets is that some periods of highprices may be essential if generators are to recovertheir fixed costs,26 and market monitors must notintervene to suppress these necessary peaks. A sys-tem with a high proportion of intermittentrenewable generators may find that the residualload on thermal generators becomes much moreuneven, with periods of zero (or negative) mar-ginal costs. If the market is so competitive thatprices can rise above marginal cost and pay forcapacity only in the few hours when demand ishigh enough to create scarcity rents, prices atthese times will have to be extremely high. Anoligopoly with moderate price-cost margins for alarger number of hours might produce a morestable outcome. In either case, the need to differ-

Energy Regulation in a Low-CarbonWorld 151

entiate between high and excessive prices, andperhaps respond to political pressure to keepprices down, will make the task of the marketmonitor more demanding.

Regulating Retail Prices

In three of the models described earlier, the regu-lator is responsible for overseeing the pricescharged to at least some energy consumers, eitherbecause there is no retail competition (althoughthere may be wholesale competition) or becausepolicymakers judge that it is not yet sufficientlyeffective to protect consumers. The combinationof regulation and a partially competitive marketoffers particular challenges, as a danger exists thatif the regulator puts too low a cap on the incum-bent’s prices, then entrants will not be able to wincustomers, and competition will never develop tothe point where it becomes self-sustaining(Joskow 2008).

If regulators wish to develop retail competi-tion, and the costs of renewable energy are signifi-cant, then it is important that they be sharedacross customers regardless of their choice ofretailers. If the costs were small, it might be possi-ble to assign them all to the incumbent retailer, onthe basis that this would do no more than partiallyoffset the advantages of incumbency; most EUcountries have taken a similar approach to the costof universal postal service obligations. The likelyexcess costs of renewable energy, however, are toogreat for this approach to be workable.This meansthat the costs must be recovered either across allenergy retailers or via distribution tariffs, whichmust also be nondiscriminatory across retailers.

One of the standard mechanisms for support-ing renewable energy does this automatically. If allretailers are obligated to obtain a set proportion oftheir power from renewable generators, usuallyadministered by a system of tradable certificates,then this should be neutral across competitors. If ahigh proportion of the renewable generators ableto provide eligible power are integrated withincumbent retailers, then this could form a barrierto entry in retailing;27 however, the rapid increasein renewable energy that most countries are tar-

geting should allow plenty of room for entry ingeneration to relax this constraint.

The other standard mechanism is to payrenewable generators a feed-in tariff that offersthem a premium price for their power. Someorganization has to make these payments andrecoup the cost of doing so. A distribution com-pany might appear to be well placed to do this.First, most renewable generators will be con-nected to its system, which gives it the knowledgeof which generators are eligible for support andwhat they produce. Second, it already bills eitherconsumers or their retailers for distribution serv-ices in a nondiscriminatory manner and can addthe cost of the feed-in tariff to these bills. In asystem with more than one distribution utility,however, there will almost certainly be a mis-match between each utility’s share of renewableenergy sources (and hence payments) and energyconsumption (and hence receipts). On the basisthat the cost of supporting renewable energyshould be shared over a wider area, side paymentsamong distributors would be needed. This meansthat the scheme would have to be overseen by agovernment agency or the regulator’s office toensure that the correct payments were made. Itmight be best for this agency to actually calculatethose payments, using information provided bythe distribution companies.

These points also apply in the retail competi-tion model, because that depends on retailers fa-cing similar obligations. They would not apply ifno retail competition existed, because then nodistortion could arise. In the absence of retailcompetition, the regulator has more ability toinfluence which customers actually have to payfor renewable energy. If regulators have a particu-lar duty to protect selected groups of customers,this raises the issue of whether the price risesshould be skewed toward others. The problemwith this approach is that it can be difficult toensure that such help is actually given to thosewho need it most. Keeping down tariffs for thosewho use very little electricity may benefit theowners of second homes, for example. Many“fuel-poor” households have relatively highenergy consumption, often because their homesare poorly insulated.28

152 Richard Green

The distribution of price increases amongcustomer classes is also relevant here: commercialand (particularly) industrial customers facinginternational competition might claim that theyshould not be exposed to unnecessary price rises,suggesting that domestic customers should face ahigh proportion of the support costs. For mostindustrial and commercial customers, energy costsare not significant;Yago et al. (2008) found that a30% increase in the price of electricity would raisecosts by less than 1% in sectors making up 90%of the UK economy. The remaining sectors willbe more affected, however, and some of thoseare exposed to international competition—particularly aluminum smelting. Should regula-tors try to design tariffs that will shield suchenergy-intensive industries from price increases?The general thrust of liberalization policy hasbeen to reduce such political interference in mar-kets, as benefits often go to the best lobbyistsrather than the most deserving causes, and alignprices to costs.

The Natural Gas IndustryWhat about the gas industry?There is some scopeto replace natural gas with biogas from plant oranimal waste, but otherwise, the impact of renew-able energy on the gas industry will come fromthe demand side rather than the supply side. Inthe medium term, natural gas is likely to befavored over coal as a fuel for electricity genera-tion. This is both because of its lower carbonemissions and because gas turbines can be moreflexible in operation, and better able to respond tointermittent renewable sources, than large coal-fired units. Although the shift to renewableswould crowd out fossil fuels, a large enough moveaway from coal could still cause an increase in gasdemand, as could the increased use of gas-firedcombined heat and power to increase energy effi-ciency. The first would require large quantities ofgas to be delivered to a few sites, whereas thesecond would increase the quantity sent to manyseparate sites around the country. In both cases,additional capacity in the network might berequired.

When it comes to the growth of renewableenergy, however, this could lead to a reduction inend users’ demand for natural gas. The use ofrenewable heat sources—biomass and heatpumps29—could double as part of the EU’s driveto obtain 20% of its energy from renewablesources, displacing natural gas and other fossilfuels (European Commission 2007). In the longterm, the need to reduce CO2 emissions by 80%or more is likely to be incompatible with burningnatural gas without carbon capture, which willprobably be impractical for small-scale heatingsystems.

Gas industry regulators may thus face thechallenging task of managing both growth anddecline. If the demand for gas for electricity gen-eration continues to grow, then the issues involvedwill be similar to those facing electricity networkswhere more investment is needed. However, thenetworks involved will mainly be the high-pressure pipelines, delivering large amounts of gasto large generators. If renewable heat and renew-able sources of electricity reduce the demand forgas by end users, then the utilities will have to facethis, even though the capital (and some operating)costs of their distribution networks will not bedeclining. This would imply a rising per-unitprice, just as the price of natural gas may also beincreasing. The process would not be sudden, anda switch to biogas might boost demand, but thepotential exists for a vicious circle of rising pricesand falling demand. In extremis, regulators orgovernments might need to consider giving gasutilities additional support to ensure that they canstill manage their distribution networks safelywhile reducing sales.

ConclusionsThis chapter has suggested specific ways regulatorswill need to respond if the energy utilities are toabsorb large amounts of renewable energy. Whatabout the broader questions posed at the outset?

Should regulators’ objectives change, or thefactors they take into account when deciding howto achieve those objectives? Regulators doubtlesswill have to consider their impact on the environ-

Energy Regulation in a Low-CarbonWorld 153

ment when making decisions. If they did not,they would find it hard to justify the investmentsneeded to accommodate renewable electricitygeneration, for example. Should protecting theenvironment be seen as an objective to placealongside consumer and producer protection or aconstraint on how the regulator goes about itsother duties? A mathematical problem can be setup to give the same results whether environmentalprotection is included as something to maximizeor a constraint that must be met, if the constraintis set at the right level. In the real world, however,regulators would probably have to act in differentways if they were to give environmental protec-tion the same priority as their other duties. We donot want them to damage the environment, buttheir relative expertise is not in this area, and itwould be preferable if economic regulators werenot designing environmental policy. It wouldseem best for regulation to continue to prioritizeconsumer protection and ensuring that firms canfinance their activities, but to do so in light ofgovernment environmental policies, and promot-ing their achievement.30

Should the scope of regulation be broadenedor narrowed? Regulators will need to make dif-ferent decisions on specific issues, but there willbe no step changes in the kinds of questions theymust address. The bulk of the evidence so far isthat well-designed liberalization programs offersuperior performance to the traditional utilitymodel (Joskow 2008).

Would the growth of renewable energy over-turn this result? Integrated firms can coordinatethe development of the network to accommodaterenewable generation. The single buyer modelallows the policymaker to choose a mix of low-carbon generators in a coordinated manner tomeet its targets. Wholesale market monitoring toguard against the exploitation of market powerwill be more challenging in an industry with ahigh proportion of intermittent generation withvery low marginal costs. Retail competition mayact as a barrier to the development of smart gridtechnologies that coordinate devices for consum-ers and on the network. Competitive models thatraise the cost of capital will be more costly for anindustry that is becoming more capital-intensive.

These are significant arguments, all suggestingthat liberalization will be less successful at dealingwith the challenges of renewable energy.They arenot likely conclusive, however. The great advan-tage of competition is that it can allow newentrants to spot opportunities that incumbentshave not exploited. If we want many consumersto generate part of their own electricity, then amodel with free entry for competing retailers thatcould provide and maintain generating equipmenton a customer’s premises could further this. Eventhough the retail competition model often seversthe direct link between the customer and distri-bution network operator, it should be possible tofind a business model through which retailers orothers that install a smart grid application on acustomer’s premises can share in the resulting ben-efits. Regulators may have to help the industrydevelop the essential information protocols andcompatibility standards to make this possible, butcompetitive environments are generally moreinnovative than monopolies.

As the level of renewable generation increases,less of the market may be truly competitive. Gov-ernments may reserve part of the market forrenewable generators, either implicitly or explic-itly. Other support schemes may be needed topersuade companies to invest in carbon captureand storage or nuclear power. The danger wouldbe that if the market for unsupported generationbecame a small rump, competition within itwould no longer act as sufficient discipline on theparticipating firms, and the prices set there wouldlose their ability to send meaningful signals.

Helm (2008) has suggested that a “utility”should be created to tender for these stations,financing them in a way that minimizes the cost ofcapital. This is effectively a variant of the singlebuyer model. If sufficient competition existsamong developers to win the tenders and buildthe stations, it might achieve the best of bothworlds. The danger, however, is that the singlebuyer would end up favoring particular technolo-gies that turned out to be unsuccessful, and thatthere would be no checks and balances to preventthis. It would be better to establish a long-termcarbon price and use this to provide incentives forinvestment in low-carbon generation on the level

154 Richard Green

playing field of a competitive wholesale market.Similarly, efforts to provide a “guiding mind” tocoordinate investment in generation and the net-works should be based on constructive engage-ment among the relevant parties, sharing infor-mation on their plans and needs, rather than onprescriptive central planning.

At the same time, competitive models doinvolve greater risks, and thus a higher cost ofcapital, than the alternatives. In a more capital-intensive industry, the cost of capital will make agreater contribution to consumer prices, but theimportance of choosing an efficient amount ofcapital will also rise—particularly during a periodof high investment, such as most countries arenow facing. Again, competitive mechanisms gen-erally lead to better investment decisions thanregulatory ones; even when investors haveindulged in excessive plant building, this has notbeen at the cost of electricity consumers.

Therefore, although the detailed decisionseconomic regulators need to make will change,and they will have to take more account of theenvironment as a constraint on what they expectcompanies to do, the shift to renewable energywill not change the fundamental tasks or nature ofeconomic regulation. We may need to adapt ourexisting tools but can continue to use them tobuild a low-carbon energy system.

AcknowledgmentsThanks go to the editors, Chris Hemsley andAndrew Quinn, and participants at the Centre forCompetition and Regulation Policy Workshopheld at Aston Business School in July 2009, forhelpful comments. The author also thanks BoazMoselle for directing him to the relevant part ofthe California ISO’s website.

Notes

1. The fourth goal, “to promote competitive marketsin the UK and beyond, helping to raise the rate ofsustainable economic growth and to improve our

productivity,” is also related to the price of energy,although one might argue that a competitive mar-ket is a means to an end rather than an end initself.

2. Although an indigenous source of supply wouldnormally be seen as secure, miners’ strikes causedmajor power cuts in the 1970s and threatenedthem in 1984–1985, thus diversification awayfrom coal reduced the risk of a recurrence.

3. In practice, much of the new investment was madebefore additional capacity was actually needed,because market power raised prices and madeentry attractive, and it also may have raised theindustry’s costs in the short term.

4. The demand for energy services is likely toincrease because of their lower effective price (theso-called direct rebound effect—as comfortbecomes cheaper, people like to be more comfort-able) and because the money saved in one areatypically will be spent on other goods and servicesthat also consume energy (an indirect effect).Sorrell (2007) reviews the evidence for this andconcludes that in developed countries, the directrebound effect is between 10% and 30% of theenergy saving per unit of energy services.The evi-dence on indirect effects is inconclusive, but theymay be quite large (more than 50% of the initiallypredicted energy saving).

5. The prices in these spot markets, and for longer-term sales, were calculated based on estimatedmarginal costs, rather than freely submitted priceoffers.

6. The map of regional transmission organizations,which do not correspond neatly with stateboundaries, shows 20 states with little or no cov-erage. Even in those states, however, some bilateraltrading may occur.

7. For example, the network operator might denyaccess on spurious grounds of system security (i.e.,claiming that the network would be vulnerable tofailure if power flows from the entrant’s areaincreased), fail to invest in new infrastructure thatwould facilitate access, or institute a “marginsqueeze” through setting excessive prices forancillary services.

8. The EU’s second electricity and gas directives, forinstance, allow member states to choose whetherto exempt utilities serving fewer than 100,000 cus-tomers from the requirements for legal separationbetween distribution and retailing that are other-wise compulsory (European Commission 2003a,b).

Energy Regulation in a Low-CarbonWorld 155

9. In Great Britain, the incumbent retailers facedprice controls when selling to domestic customersuntil 2002, even though all customers had a choiceof retailers by mid-1999. In contrast, price con-trols for medium-size commercial and industrialelectricity customers were lifted as soon as thismarket was opened to competition in 1994, in thebelief that these buyers were more sophisticatedand could protect themselves by shopping around.

10. An exception is the tidal barrage at La Rance,France, which is effectively a hydroelectricscheme—a very mature technology—built acrossan estuary rather than an inland river. A similarscheme across the Severn Estuary in the UnitedKingdom could generate 5% of the country’s elec-tricity (DECC 2009b) at the cost of significantchanges to the local ecosystem and £21 billion($32 billion).

11. While wind energy would be 41% of renewablegeneration in this scenario, 43% of electricity gen-eration would be renewable, giving an overallpenetration of more than 17%.

12. The range can vary among turbine designs. Forexample, GE’s 3.6-megawatt (MW) offshore windturbine does not need to cut out until the windreaches 27 m/s but needs a speed of 14 m/s toreach full output, whereas its 1.5 MW 1.5xleonshore design can reach full power at a speed of11.5 m/s but cuts out after 10 minutes averaging20 m/s.

13. Individual photovoltaic units could have volatileoutputs on a partly cloudy day, producing muchless when a cloud moves overhead.

14. Gross et al. (2006) give a range from 19.1% to 26%for Great Britain, if it were to get 20% of itsenergy from wind—somewhat less than is nowplanned.The UK government uses a range of 10%to 20% for the larger quantities expected (Houseof Lords 2008b, 484), and E.ON UK, an inte-grated power and gas company, has suggested thatthe figure should be less than 10% (ibid., 107).

15. This assumes that the wind generators have a loadfactor of 28% and the conventional generators oneof 80%, which is rather low but produces roundnumbers in the example: both sets of stationswould produce an average output of 2.8 GW.

16. Some power stations have been connected at dis-tribution voltages, particularly at the highervoltages close to the transmission system, but on alimited scale compared with the likely growth ofdistribution-connected renewable generation.

17. Federal legislation for a nationwide scheme isbeing debated at the time of writing (November2009).

18. A generator with a surplus of permits, albeit free,still faces the opportunity cost of not selling thespare ones.

19. If the firm spends more than it proposed, however,it would have been better off had it proposed ahigher level of spending.

20. Because Great Britain follows the model of retailcompetition, formal regulation covers only thetransmission and distribution networks, and thisreview does not deal with issues related togeneration or retailing, even though many net-work innovations will also affect these sectors.For documents related to the review, seewww.ofgem.gov.uk/Networks/rpix20/Pages/RPIX20.aspx.

21. This is because many renewable generators will belocated far from demand, and their low load fac-tors mean that it would not be economic to buildenough transmission capacity to accommodate alltheir output at the rare times when they are pro-ducing at close to full output, leading to anincrease in transmission constraints.

22. For more information, see the California ISO’swebsite at www.caiso.com/1816/1816d22953ec0.html.

23. If this behavior allowed a generator to avoid build-ing the extra transmission, and the generator couldstill meet the peak demand, it would be efficient.

24. These transmission charges vary by around £7.00($10.68) per kW per year, whereas a forwardwholesale price for 2010 (at the end of June 2009)was £51.25 ($78.17) per MWh, and a renewablegenerator would also receive Renewables Obliga-tion Certificates that were selling for around£52.00 ($79.32) per MWh (UK Powerfocus 2009).A 1% change in load factor leads to an increase inoutput of 88 MWh (for a 1 MW unit), and hencean extra £4.70 ($7.17) per kW per year inincome.

25. A state-owned transmission company potentiallyhas the best financial cushion of all, but it may wellbe less receptive to financial incentives than aprivate-sector organization.

26. The alternative paradigm uses a capacity market torecover generators’ fixed costs, sometimes coupledwith payments linked to spot prices that should actto restrain those prices (Cramton and Stoft 2008).

27. If a high level of integration between retailing andrenewable generation extended to the generation

156 Richard Green

sector overall, then the barrier to entry in retailingwould be severe. If several regional incumbentsexist within a larger market, however, effectivecompetition among them is possible, though notinevitable.

28. In the United Kingdom, fuel-poor households aredefined as those that need to spend more than 10%of their income to achieve acceptable standards ofheat and light.

29. Heat pumps basically use refrigerator technologyin reverse to extract heat from the environment,chilling a fluid so that it is colder than conditionsoutside (or underground) and can absorb heatfrom its surroundings. Compressing the fluid thenraises its temperature to a point where the heat canbe extracted and used. The gross energy deliveredby a heat pump can be four times that used forpumping, in which case 75% of the energy deliv-ered would be counted as renewable by the EU.

30. In countries where the regulator is constrained toclosely follow specific laws, this implies that thoselaws should give the regulator tasks that achievethese same aims.

ReferencesBorenstein, S. 2005. The Long-Run Efficiency of

Real-Time Electricity Pricing. Energy Journal 26 (3):93–116.

Cramton, P., and S. Stoft. 2008. Forward ReliabilityMarkets: Less Risk, Less Market Power, More Effi-ciency. Utilities Policy 16 (3): 194–201.

DECC (Department of Energy and Climate Change).2009a. Consultation on Renewable Electricity FinancialIncentives 2009. London: DECC.

———. 2009b. The UK Renewable Energy Strategy. Cm7686. London: Stationery Office.

DoE (Department of Energy). 2006. U.S. Department ofEnergy Strategic Plan. Washington, DC: Departmentof Energy.

Douglas, J. 2005. In the Zone. Power Engineering Inter-national 13 (12): 45–47

DTI (Department of Trade and Industry). 2003. OurEnergy Future: Creating a Low Carbon Economy. Cm5761. London: Stationery Office.

EIA (Energy Information Administration). 2009. USIndependent System Operators. www.eia.doe.gov/cneaf/electricity/page/channel/fig8.html (accessedMarch 22, 2010).

European Commission. 2003a. Directive 2003/54/ECof the European Parliament and of the Council of 26 June

2003 concerning Common Rules for the Internal Marketin Electricity and Repealing Directive 96/92/EC. Brus-sels: Commission of the European Communities.

———. 2003b. Directive 2003/55/EC of the EuropeanParliament and of the Council of 26 June 2003 concerningCommon Rules for the Internal Market in Natural Gasand Repealing Directive 98/30/EC. Brussels: Com-mission of the European Communities.

———. 2007. Renewable Energy Road Map: RenewableEnergies in the 21st Century: Building a More Sustain-able Future: Impact Assessment.Accompanying Documentto the Communication from the Commission to the Coun-cil and the European Parliament. SEC (2006) 1719/3.Brussels: Commission of the European Communi-ties.

Green, R. J. 1997. Electricity Transmission Pricing: AnInternational Comparison Utilities Policy 6 (3): 177–184.

———. 2005. Electricity and Markets Oxford Review ofEconomic Policy 21 (1): 67–87.

Gross, R., P. Heptonstall, D. Anderson,T. C. Green, M.Leach, and J. Skea (2006) The Costs and Impacts ofIntermittency: AnAssessment of the Evidence on the Costsand Impacts of Intermittent Generation on the BritishElectricity Network. London: Imperial College.

Hannah, L. 1979. Electricity before Nationalisation. Lon-don: MacMillan.

Hawkes, A., and M. Leach. 2007. Cost-Effective Oper-ating Strategy for Residential Micro-combined Heatand Power. Energy 32 (5): 711–723.

Helm, D. R. 2008. Renewables—Time for a Rethink?www.dieterhelm.co.uk/node/488 (accessedMarch 22, 2010).

House of Lords. 2008a. The Economics of RenewableEnergy, Economic Affairs Select Committee Fourth Reportof Session 2007–8. HL195-I of 2007–8. London: Sta-tionery Office.

———. 2008b. The Economics of Renewable Energy, Eco-nomic Affairs Select Committee Fourth Report of Session2007–8.Vol. 2, Evidence. HL195-II of 2007–8. Lon-don: Stationery Office.

Jamasb, T., and M. G. Pollitt. 2007. Incentive Regula-tion of Electricity Distribution Networks: Lessons ofExperience from Britain. Energy Policy 35 (12):6163–6187.

Joskow, P. L. 2008. Lessons Learned from ElectricityMarket Liberalization. Energy Journal 29, SpecialIssue in Honor of David Newbery, 9–42.

Laffont, J.-J., and J.Tirole. 1993. ATheory of Incentives inProcurement and Regulation. Cambridge: MIT Press.

Energy Regulation in a Low-CarbonWorld 157

Littlechild, S. C. (2008) Some Applied Economics ofUtility Regulation. Energy Journal 29, Special Issue inHonor of David Newbery, 43–62.

Newbery, D. M. (1999) Privatization, Restructuring andRegulation of Network Utilities. Cambridge, MA: MITPress.

Ofgem. 2009. Innovation in Energy Networks: Is MoreNeeded and How Can This Be Stimulated? RegulatingEnergy Networks for the Future: RPI-X@20. Workingpaper 2. London: Office of Gas and Electricity Mar-kets.

Pérez-Arriaga, I. J., F. J. Rubio, J. F. Puerta, J. Arceluz,and J. Marin. 1995. Marginal Pricing of Transmis-sion Services: An Analysis of Cost Recovery. IEEETransactions on Power Systems 10 (1): 546–553.

Pollitt, M. G. 2004. Electricity Reform in Chile: Les-sons for Developing Countries. Journal of NetworkIndustries 5 (3–4): 221–262.

———. 2005. The Role of Efficiency Estimates inRegulatory Price Reviews: Ofgem’s Approach toBenchmarking Electricity Networks. Utilities Policy13 (4): 279–288.

———. 2008. The Future of Electricity (and Gas)Regulation in a Low-Carbon Policy World. EnergyJournal 29, Special Issue in Honor of DavidNewbery, 63–94.

Pollitt, M. G., and J. Bialek. 2008. Electricity NetworkInvestment and Regulation for a Low-Carbonfuture. In Delivering a Low Carbon Electricity System:Technologies, Economics and Policy, edited by M.

Grubb, T. Jamasb and M. Pollitt. Cambridge: Cam-bridge University Press, 183–206.

Sappington, D. 1982. Optimal Regulation of Researchand Development under Imperfect Information. BellJournal of Economics 13 (2): 354–368.

Schulz, W. 1995. Restructuring the Electricity Market:A German View. In Competition in the Electricity Sup-ply Industry: Experience from Europe and the UnitedStates, edited by O. J. Olsen. Copenhagen: DJØFPublishing.

Shleifer, A. 1985. A Theory of Yardstick Competition.RAND Journal of Economics 16: 319–327.

Sorrell, S. 2007. The Rebound Effect:An Assessment of theEvidence for Economy-wide Energy Savings fromImproved Energy Efficiency. London: UK EnergyResearch Centre.

Spiller, P.T., and L.Viana Martorell. 1996. How ShouldIt Be Done? Electricity Regulation in Argentina,Brazil, Uruguay and Chile. In International Compari-sons of Electricity Regulation, edited by R. J. Gilbertand E. P. Kahn. Cambridge: Cambridge UniversityPress: 82–125.

UK Powerfocus. 2009. Month at a Glance. UK Powerfocus109: 1.

Yago, M., J. P. Atkins, K. R. Bhattarai, R. J. Green, andS. D.Trotter. 2008. Modelling the Economic Impactof Low-Carbon Electricity. In Delivering a Low Car-bon Electricity System:Technologies,Economics and Policy,edited by M. Grubb,T. Jamasb, and M. Pollitt. Cam-bridge: Cambridge University Press, 394–413.

158 Richard Green

9

Building Blocks: Investment inRenewable and NonrenewableTechnologiesJames Bushnell

Within a span of 20 years, the electric powerindustry has become the central focus of

two extraordinary policy trends, each one signifi-cant enough to fundamentally reshape the indus-try. One of these trends is liberalization, a termthat has come to encompass both privatizationand regulatory restructuring. Beginning with thevisions articulated in such works as Joskow andSchmalansee (1985) and Schweppe et al. (1988),the restructuring movement in electricity can beviewed as an extension of the trend toward marketliberalization that had previously transformed theairline, communications, and natural gas indus-tries. The generation sector of the industry hasundergone a sporadic but inexorable transitionfrom economic regulation under cost-of-serviceprinciples to an environment in which marketsheavily influence, if not dominate, the remunera-tion and investment decisions of firms.

The second trend to engulf the electricityindustry has been the growth of the environmen-tal movement. More specifically, the growingalarm over the threat of global climate change,and the more recent engagement of policymakersin combating it, is likely to dominatedecisionmaking in the power industry over thenext several decades. Electricity and heat produc-tion are responsible for 40% of CO2 emissions inthe United States and about 31% worldwide(Stern 2006).

Although not obvious at first glance, thesetwo trends, restructuring and environmentalregulation, share many common ideological roots.In the United States, the growing stringency in airquality regulation was accompanied by anincreased acceptance of market-based environ-mental regulations. These include cap-and-trademechanisms, such as the program put in place tolimit SO2 emissions under the 1990 amendmentsto the Clean Air Act (Ellerman et al. 2000).Regulators were also interested in experimentingwith market-based incentives to promote alterna-tive energy sources. Many trace the birth of theU.S. independent power industry to the passage ofthe Public Utilities Regulatory Policies Act(PURPA) in 1978.The PURPA legislation estab-lished mandates for the purchase of energy pro-duced by qualifying small and renewable sourcesof generation (Joskow 1997; Kahn 1988).Although inspired largely by environmental andenergy security goals, the largest impact ofPURPA is arguably in the resulting demonstra-tion of the viability of smaller-capacity generationtechnologies and the nonutility generator businessmodel.

One important aspect of the independentpower producer business model was the relativefreedom—and risk—allowed in investment ofnew facilities. Investments are based on market-

based long-term contracts and projections of mar-ket revenues, rather than regulatory findings ofneed and guaranteed cost recovery. The restruc-turing movement in the United States was led bystates with the worst track records in utility invest-ment (Ando and Palmer 1998). Although evi-dence suggests that operations have become moreefficient in these states (Wolfram 2005), restruc-turing was primarily intended to improve theincentives of firms to make prudent investments(Borenstein and Bushnell 2000). In some parts ofthe world, this general approach to investment hascome to dominate the industry. In many others,policymakers continue to search for the propertools for balancing market incentives with con-cerns over reliability and adequacy of investment(Joskow 2005; Oren 2005). One central aspect ofthis search concerns the design of wholesale elec-tricity markets and the payment streams they pro-vide to suppliers. Markets can differ greatly on theprimary sources of remuneration for generators,with some relying on energy and ancillary serv-ices markets, while others have establishedmechanisms for compensating suppliers for theirinstalled or available capacity (Bushnell 2005;Cramton and Stoft 2005).

This chapter studies the intersection of thesetwo trends as they come to dominate the eco-nomics of the industry. In particular, it examineshow the increasing penetration of intermittentrenewable generation can change the economiclandscape for merchant power investment in con-ventional thermal generation. Currently, renew-able generation earns revenues from a wide rangeof sources, from energy markets to governmenttax credits. The impact of renewable generationon the electricity markets in which they partici-pate has to date been relatively modest outside ofregions of high concentration such as west Texas.That will almost certainly change, however, asstate and federal policies considerably ramp up theamount of renewable generation throughout thecountry. This can have a profound impact onprices and the economics of supply for bothrenewable and nonrenewable generation.

An equilibrium model of generation invest-ment is developed, based on the long-standingprinciples of finding the optimal mix of capital-

intensive and higher-marginal-cost resources toserve a market with fluctuating demand. Thismodel is then applied to data on electricity mar-kets from several regions of the western UnitedStates to examine how the interaction of increas-ing wind capacity and electricity market designaffects the equilibrium mix of thermal capacityand the revenues earned by renewable suppliers.The chapter first provides a brief background onthis question, then describes the equilibrium con-ditions that form the stylized investment model.Next, it details the data and assumptions used inthe study. The final section contains the bulk ofthe results and analysis.

Background: Renewable Energyin Restructured ElectricityMarketsRenewable, or green, power is viewed by manypolicymakers as the key to combating greenhousegas emissions within the power sector. Explicitand implicit subsidies for renewable power con-tinue to grow. By the end of 2007, 25 U.S. statesand Washington, DC, had some form of renew-able portfolio standard (RPS), which requirespurchasers of wholesale electricity to procuresome percentage of their power from renewablesources (Wiser and Barbose 2008). The long-standing, but intermittent and precarious, pro-duction tax credit (PTC) for wind energy in theUnited States pays wind producers 2.1 cents/kWh for energy production. The AmericanRecovery and Reinvestment Act of 2009 con-tained several provisions favorable to renewablegeneration, including the extension of the PTCuntil 2012 and alternative investment tax creditsfor facilities constructed in 2009 and 2010 (Wiserand Bolinger 2009).

For the industry as a whole, the growth ofnonutility generation has coincided with theexpansion of renewable generation sources.This isnot the product of happenstance; from the passageof PURPA, various legislative purchasing man-dates and tax incentives played a dominant role in

160 James Bushnell

the growth of both renewable and nonutility gen-eration. To this day, the renewable industry isdominated by nonutility producers.1

The subsidization of renewable generation isexpanding in parallel with efforts to create cap-and-trade programs for CO2. This can be seen asantithetical to the spirit of a cap-and-trade pro-gram, where promoting flexibility in complianceoptions is a central ideal. Unlike the SO2 pro-gram, cap-and-trade is but one of a broad set ofpolicy tools being brought to bear against green-house gas (GHG) emissions. Some view this asundermining the strength of cap-and-trade regu-lation. The cap has less incremental impact ifmuch of the GHG reductions are alreadyaccounted for under various more directed meas-ures and regulations. In regions such as California,cap-and-trade is viewed more as a backstop thanas a bulwark in combating climate change.

Although policies that promote renewablegeneration sources are extremely popular withregulators, politicians, and the general public,their continued expansion to unprecedented lev-els does raise some concerns. One source of con-cern is cost. Although the technological frontiercontinues to advance, much controversy existsover the appropriate timing and form of policyintervention to promote renewable generation.Most accept that renewable generation would notbe a significant source of supply today if not forsome form of public support. The fact that theexternal cost of GHG emissions have not yet beenpriced into the investment decisions of fossil-based generation firms certainly provides justifica-tion for support of renewable power, but the pros-pect of regional and possibly national caps onCO2 emissions undermines that justification. Acommon argument for support of renewable gen-eration is the hope that expansion of supply willyield learning benefits, thereby lowering costs offuture supply. However, a market failure existsonly if that learning cannot be appropriated forprivate gain. Although the bulk of public supportfor renewable generation has taken the form ofproduction mandates or credits, it is not clearwhether commercialization is the point in thesupply chain where the problem of intellectualproperty is most acute. Further, the evidence to

date indicates that cost reductions in alternativeenergy sources can be driven as much byexogenous technology developments as by theexpansion of installed capacity (Nemet 2006).

The most commonly heard concern over therapid expansion of renewable electricity supply isover the fact that this supply is available onlyintermittently (NERC 2009). With the prospectof one-fifth or more of electrical energy comingfrom intermittent sources, many in the industryare confronting the fact that the traditional toolsfor planning for and providing reliable electricservice may prove inadequate. In fact, as discussedbelow, the traditional utility planning paradigmhas been disrupted by market liberalization overthe last 10 years.The industry has yet to settle on asingle framework to replace utility planning. Thelarge-scale addition of intermittent resources istherefore happening against a backdrop in whichthe mechanisms through which generators arecompensated are very much in flux.

Investment in RestructuredElectricity MarketsSince the onset of market liberalization, concernshave been raised that the newly formed marketregimes would fail to produce adequate invest-ment in generation capacity. Ironically, in manyparts of the world, it was the cost of excess cap-acity that provided the impetus for liberalization.The safety net of guaranteed capital cost recoveryin both publicly owned and rate-of-return regu-lated utilities had provided a high degree of reli-ability. Indeed, the reliability of electric supply inmost OECD countries is so high that it is oftentaken for granted. U.S. electricity consumers,unlike those in many developing countries, fullyexpect to be able to consume as much electricityas they need whenever they desire.

These high levels of reliability came at a highcost, however, particularly when combined withthe weak incentives for cost control provided bypublic ownership and regulation. Under the tradi-tional model, a utility and its regulators jointlyforecast a “need” for investment, and the regulator

Building Blocks: Investment in Renewable and NonrenewableTechnologies 161

would guarantee the recovery of costs undertakento meet that need. In the liberalized market, pri-vate firms no longer receive a guaranteed recoveryof their investments. One of the hopes for liberali-zation was that this market-based risk would leadto more prudent and cost-effective investmentdecisions. At the very least, it was observed, thecosts of overinvestment would be borne by inves-tors rather than ratepayers under the new marketregime. In many markets, this latter belief hasbeen supported by the fact that many of the firmsthat procured or expanded capacity in liberalizedmarkets experienced severe financial difficultiesduring the early part of this decade.

Whereas a transition away from paymentsbased on a cost-of-service framework is shared byall liberalized markets, the revenue streams thatreplaced these payments differ greatly. Many mar-kets focus the remuneration of generators on theprovision of energy and related services. In thejargon of the U.S. electricity industry, this con-ceptual framework has been referred to as an“energy-only” framework. The name, which issomewhat inaccurate, refers to the fact that con-tributions against fixed and sunk costs arise onlyfrom payments for the provision of either energyor associated operating reserve services. Althoughno market is fully unconstrained in this way, mar-kets such as those found in the United Kingdom,Australia,Texas, New Zealand, and Norway oper-ate under general energy-only principles to theextent that they have no or relatively high pricecaps and provide no other specific payments forthe supply of capacity.

In many markets, however, the revenues pro-vided from the provision of energy and ancillaryservices appear to be insufficient to cover the fixedcost of new entry (Joskow 2005). Myriad reasonscan be given for this, including the existence ofprice caps, the subtle but significant impact of thedecisions of system operators on market prices,and simply the overinvestment of capacity. Thisand other factors have led to a level of discomfortamong many policymakers over leaving invest-ment decisions entirely up to the market. There-fore, many electricity markets, including several inthe United States, provide payments for capacity“availability” that supplement revenues received

for the provision of energy and ancillary services.This feature is not unique to the United States, ascapacity payments played a significant role in theearly years of market liberalization in the UnitedKingdom and continue to be a significant factorin Spain and Colombia.

The details of these capacity payments vary,but the general common features that are repre-sented in the stylized model used here are a formalor informal constraint on energy prices combinedwith a fixed payment (here assumed to be in dol-lars per megawatt-year) based on installed cap-acity. The fixed payments can be scaled accordingto the expected or historic availability of genera-tion, a fact most significant for wind generationsources.

In many restructured markets, some form ofpayment for installed or available capacity is madeto producers as a supplement to the revenues theyearn through the sale of energy and ancillary serv-ices. These payments are not without controversy,however, as debate continues over how exactly tomeasure and remunerate the provision of “reli-able” capacity (Cramton 2003; Hogan 2005;Oren 2005).

One aspect of this debate is how to deal withunconventional sources of generation. Resources,such as hydroelectric facilities, that are energy-limited cannot produce at their full capacity all thetime. Many renewable resources can supply poweronly intermittently, and their supply is dependenton ambient conditions rather than under the con-trol of the operator. In general, the capacity pay-ments made to resources such as these are scaleddownward according to rough probabilistic meas-ures of their potential availability. As exploredbelow, the specification of such rules will interactwith the level of penetration of renewable genera-tion to shift the relative value of different types ofpayment streams for intermittent producers.

The power industry today therefore featurestwo contrasting models for financing new invest-ment: the energy-only model, which relies onperiodic, extremely high prices for energy andancillary services to provide the scarcity rents thatare applied to the recovery of capital costs; and thecapacity payment model, in which a large portionof the capital costs are recovered through capacity

162 James Bushnell

payments. Under energy-only markets, thechoice and profitability of specific generationsources will depend on the degree and timing ofhigh prices. Under capacity markets, the spotenergy prices are somewhat less critical, but thespecific implementation of capacity payments isvery important to the relative profitability of tech-nologies.

The large-scale deployment of intermittentresources can imply a major paradigm shift forboth investment models. Electric systems willlikely experience a massive addition of renewablegeneration capacity that is largely motivated bynonmarket considerations such as climate change.This will result in an influx of energy withextremely low marginal cost, but only duringsome time periods. As a result, the remaining needfor thermal generation capacity could look verydifferent than it would in the absence of therenewable capacity. In market terms, the levelsand patterns of energy prices could be quite dif-ferent with the addition of renewables. Themonths and hours that experience peak priceswill be driven as much by the availability of inter-mittent resources as by the fluctuations in end-usedemand.These questions are explored empiricallyin the following section.

Equilibrium Model of ElectricityInvestmentThis section uses a long-run equilibrium model ofinvestment to explore the ramifications of greatlyexpanded intermittent supply. A technical formu-lation of the model is provided in the appendix atthe end of this chapter.The model draws from theclassic framework of utility investment, whichapplies a mix of technologies of varying capitalintensity to satisfy fluctuating demand (see Kahn1988).This demand is often represented in a load-duration curve, which illustrates a cumulative dis-tribution of demand levels over some time period,such as one year. This basic model is expanded toincorporate elements of peak load pricing asarticulated in theory by Borenstein (2005). Themodel examines the mix and cost of technologies

that achieve the break-even point where annualenergy revenues for each technology equal theirannualized cost of capacity. As in Borenstein(2005), these values depend on prices rising abovethe marginal cost of the highest-cost technologieswhere, in effect, demand sets the market price.This process has come to be called “scarcity pri-cing” in wholesale electricity markets. Similar toLamont (2008), the model also incorporates inter-mittent resources. As described below, the windproduction profiles used here are based on specificprojections of wind production profiles, ratherthan the stylized correlation coefficients used byLamont.

The model here assumes perfect competition,essentially free entry into any generation technol-ogy in the markets, and also disregards concerns of“lumpiness” of capacity. Firms are free to installany combination of capacity sizes that satisfy dif-ferentiable equilibrium conditions. This greatlysimplifies computational concerns and, in light ofthe size of the markets being examined here, isnot an unreasonable assumption. As this is a long-term model, it also ignores operational constraintssuch as minimum run times, start costs, andramping rates. These are obviously importantconsiderations of operating an electricity systemthat will be affected by the expansion of intermit-tent technologies, but they are beyond the capa-bilities of the model used here.

The approach of the model is to examine theactual load profiles or hourly distributions ofdemand of certain markets, and then impose vary-ing levels of intermittent wind production onthose demand distributions. In other words, thewind investment is considered exogenous to theequilibrium investment model, having beenimplemented through nonmarket constraints suchas a renewable portfolio standard.The model thenderives the mix of thermal technologies thatwould be constructed to serve the resultingresidual demand that is left over after accountingfor wind production. The equilibrium resultingfrom an assumption of competitive entry and nolumpiness is equivalent to the optimal, or least-cost, set of technologies.The intuition behind theequilibrium constraints described in the appendixis straightforward. Firms will continue to con-

Building Blocks: Investment in Renewable and NonrenewableTechnologies 163

struct additional capacity in a given thermal tech-nology as long as the revenues implied by theresidual demand are sufficient to cover a levelizedcost of investment, as well as operating costs.

The empirical calculations are based on datataken from the Western Electricity CoordinatingCouncil (WECC) for the reference year 2007.These data are in turn subdivided into the fourWECC subregions: the California (CA) region;the Northwest Power Pool (NWPP) region; thesouthwest (AZNM) region, made up mostly ofArizona and New Mexico; and the RockyMountain Power Pool (RMPP) region (see Figure9.1).

The general approach is to ask how electricityload would have looked during 2007 under vari-ous levels of wind penetration. The model thensolves for the equilibrium investment mix of con-ventional technologies that optimally serves theresulting load shape.This section includes descrip-tions of the data sources and assumptions used inimplementing this calculation.

It is important to keep in mind that this is nota simulation of the incremental investment requiredgoing forward in these markets, but rather anexercise that examines how the long-run equilib-rium mix of generation and costs would change.Thus it is not meant to be predictive of theseactual markets, but uses these market data to

develop calculations for a range of possible repre-sentative markets. The market-based modelassumes that all regions are restructured (when infact, only California is currently even partiallyrestructured) and that the investment choices arestarting from a clean slate of no existing capacity.

One difficulty with simulating electricitymarkets in a high level of detail is that, althoughdata on most fossil-fuel-based generation units arequite extensive and reliable, far less data exist onthe activities of hydroelectric plants, renewablegeneration, and the substantial amount of powergenerated from combined heat and power (CHP)or cogeneration plants. When building a counter-factual re-creation of an electricity market, thesedata gaps make assumptions about the missingproduction necessary.

This chapter takes the approach of restrictingthe construction of a counterfactual market out-come to the portion of resources for whichdetailed data are available. In effect, it assumesthat, under the counterfactual assumptions ofwind penetration, the operations of nonmodeledgeneration plants would not have changed. Thetotal production from “clean” sources is unlikelyto change in the short run. The production ofelectricity missing from the data is driven by natu-ral resource availability (rain, wind, sun) or, in thecase of CHP, to nonelectricity production deci-sions. The economics of production are such thatthese sources are essentially producing all thepower they can. However, it is important to rec-ognize that this modeling approach assumes thatexisting unconventional sources will not changenot only how much they produce, but also whenthey produce it. This is a problematic assumptionin regions with substantial hydro resources, suchas the Pacific Northwest. Ideally, an investmentanalysis would involve a co-optimization ofhydro, wind, and thermal electric production.This is beyond the scope of this chapter. For thisreason, the results pertaining to the Pacific North-west region should be interpreted with this short-coming in mind.

In any event, the goal here is not to reproducethe electric system as it actually operated in 2007,but rather to assess how investment decisionswould play out if the industry were starting from a

NWPP

RMPP

CA

AZNM

Figure 9.1. The four WECC subregions

164 James Bushnell

completely clean slate and faced the residual (afterexisting unconventional generation) load shapesof 2007. The data used here are meant to conveyconditions present in representative electricitysystems, rather than completely reproduce a spe-cific system.

Demand Data

The primary data source for this discussion is theBASECASE dataset from Platts, which is in turnderived primarily from the continuous emissionsmonitoring system (CEMS) used by the U.S.Environmental Protection Agency (EPA) tomonitor the emissions of large stationary sources.2

Almost all large fossil-fired electricity generationsources are included in this dataset, althoughhydroelectric, renewable, and some small fossilgeneration sources are missing. The CEMSreports hourly data on several aspects of produc-tion and emissions. Hourly data on nuclear gen-eration plants are included with fossil generationdata in the BASECASE dataset. The model hereuses the hourly generation output and carbonemissions for available facilities.

These hourly output data are aggregated byfirm and region to develop the demand in thesimulation model. As described above, this is infact a residual demand: the demand that is leftafter applying the output from non-CEMS plants.Plant cost, capacity, availability characteristics, andregional fuel prices are then taken from the PlattsPOWERDAT dataset. These data are in turnderived from mandatory industry reporting to theEnergy Information Administration (EIA) andNorth American Electric Reliability Council(NERC).

These data are then combined to create ademand profile and supply functions for periodsin the simulation. Although hourly data are avail-able, for computational reasons these are aggre-gated into representative time periods. Each of thefour seasons has 50 such periods, yielding 200explicitly modeled time periods. The aggregationof hourly data was based on a sorting of the Cali-fornia residual demand. California aggregate pro-duction was sorted into 50 bins based on equalMW spreads between the minimum and maxi-

mum production levels observed in the 2007 sam-ple year. A time period in the simulation thereforeis based on the mean of the relevant market datafor all actual 2007 data that fall within the boundsof each bin. For example, every actual hour (ofwhich there were 14) during spring 2007 inwhich California CEMS production fell between7,040 and 7,243 MW were combined into a sin-gle representative hour for simulation purposes.

The number of season-hour observations ineach bin is therefore unbalanced; there are rela-tively few observations in the highest and lowestproduction levels, and more observations closer tothe median levels. The demand levels used in thesimulation are then based on the mean productionlevels observed in each bin. In order to calculateaggregate production and revenues, the resultingoutputs for each simulated demand level was mul-tiplied by the number of actual market hours usedto produce the input for that simulated demandlevel. Table 9.1 presents summary statistics ofCEMS load levels for each of the four WECCsubregions.

Table 9.1. Summary statistics of demand

CEMS load (MW)

Region Mean Min. Max. S.D.

CA 13,216 6,022 29,985 3,626

NWPP 15,334 9,670 18,884 2,400

AZNM 17,942 13,626 25,586 2,706

RMPP 6,986 5,531 9,141 723

Note: MW = megawatts; S.D. = standard deviation

Wind Generation Data

The wind generation profiles used in this chaptercome from WECC transmission planning studies.The WECC studied several scenarios for renew-able energy penetration (see Nickell 2008), withparticular focus on an assumption of 15% of totalWECC energy being provided from renewablesources. This modeling effort employed a datasetfrom the National Renewable Energy Laboratory(NREL) that provides 10-minute wind speedswith a high level of geographic resolutionthroughout the U.S. portion of the WECC sys-tem. The WECC study combines these wind

Building Blocks: Investment in Renewable and NonrenewableTechnologies 165

potential data with other local sources of informa-tion to construct projections of new wind devel-opment, as well as of hourly wind productionfrom those potential new sources.

This chapter draws on the hourly load profilesof the projected wind facilities from the WECCstudy and aggregates these profiles according tothe four WECC subregions described above.Because of the focus here on the investmentimpacts of wind penetration, this section looks atthe baseline level of estimated production used inthe WECC study and also a level that is doublethat used in the WECC study.The aggregate gen-eration levels are summarized in Table 9.2. As aportion of CEMS load, the new wind sourceswould account for about 15% of 2007 CEMSenergy, although these resources are not evenlydistributed across the WECC.3 The RMPP area,which includes the wind-rich areas of Wyoming,has a great deal of wind potential, whereas thedesert Southwest has much less.

When the projected additional wind produc-tion is combined with and assumed to displaceCEMS production, the result is a sharply shiftedresidual load profile that must be served by con-ventional generation sources. Figures 9.2 and 9.3illustrate the hourly CEMS load, both before and

after accounting for the additional wind resourcesfor the months of August and December.

The aggregate effects are well summarized byload duration curves presenting the cumulativedistribution of CEMS load and residual demandafter new wind sources. Figure 9.4 presents theseload duration curves for the four subregions.

The CEMS load profile in California is muchmore variable than in other regions, while CEMSload levels in the Pacific Northwest are relativelyconstant because of the abundance of hydroenergy in that region. In all cases, the increasingpenetration of wind resources makes the load pro-files steeper. This reflects the fact that wind pro-

Figure 9.2. CEMS load and wind production for August

Table 9.2. Aggregate generation levels

Hourly averages

Load Wind High wind

Region (MWh) (MWh) Share (MWh) Share

CA 13,216 1,866 14% 3,733 28%

NWPP 15,334 2,229 15% 4,458 29%

AZNM 17,942 1,445 8% 2,891 16%

RMPP 6,986 1,902 27% 3,804 54%

Totals 53,479 7,443 14% 14,885 28%

Note: MWh = megawatt-hours

166 James Bushnell

duction is not correlated with CEMS load.To theextent that more wind production is generated inlow CEMS load hours, the residual load becomesmore variable and the load duration curve steeper.This effect is most pronounced in the RMPPregion, where CEMS load was relatively constantbut which experiences the highest degree of windpenetration.

The market implications of Figure 9.4 are centralto the results of this chapter, so they deserve alittle further discussion. The increasing penetra-tion of wind resources in the WECC will create asurge of energy supply, much of which will beuncorrelated with end-use demand.The net resultis a residual load shape that is more “peaky.” Aswill be demonstrated in the results of the

Figure 9.3. CEMS load and wind production for December

Figure 9.4. Annual distribution of CEMS load net of wind

Building Blocks: Investment in Renewable and NonrenewableTechnologies 167

simulations, the optimal mix of resources to servethis profile of residual demand will be composedof a far greater share of low-capacity-cost, high-marginal-cost peaking resources.

Thermal Generation Cost Data

The model here examines the optimal possiblemix of generation technologies, assuming it startsfrom a clean slate, with no sunk (or stranded)investment decisions. This section examines theoptimal mix of three basic technologies that formthe backbone of most U.S. electric systems. Eachrepresents different levels of the trade-off betweencapital costs and marginal costs. Included are abase load pulverized coal technology, a midmeritcombined cycle gas turbine (CCGT) technology,and a peaking gas combustion turbine (CT). Thecosts of construction and operation for each ofthese technologies are taken from the EnergyInformation Administration’s 2007 Annual EnergyOutlook (EIA 2007).The basic cost characteristics,taken from the EIA study, are summarized inTable 9.3. To convert these costs to an annualizedfixed cost, a 15-year payback period and 10%financing cost are assumed. Fuel costs are takenfrom the EIA’s figures for 2007. The resultingaggregate (including operating and maintenance)costs are summarized in Table 9.4.

Analysis and ResultsUsing the data described in the previous section,the resulting optimal mix and level of generationcapacity are calculated for each of the four WECCsubregions. For the purposes of this study, each

region is treated as isolated from the others.This isequivalent to assuming that transmission flowsamong the regions do not change from their 2007levels. As described above, the four regions repre-sent a wide spectrum in terms of current demandfor generation and future wind potential.

Energy-Only Market

The results are first examined under an assump-tion that each region operates under an energy-only market paradigm, with no price cap and nocapacity payments. The analysis begins with theequilibrium energy prices in each market. Figures9.5 and 9.6 illustrate the hourly market-clearingenergy prices in each market for the final week ofAugust and first week of December. Note thatthese prices are plotted on a logarithmic scale,reflecting the highly volatile nature of equilibriumelectricity prices in an energy-only market.Although significant differences in energy pricesare difficult to detect in CA, the impact of windpenetration is clear on the pricing patterns inregions such as the NWPP and RMPP.

The changes in the residual demand profilesand the resulting equilibrium prices do have asignificant effect. Figure 9.7 summarizes the equi-librium investment levels under three wind sce-narios: wind at 2007 levels, wind at 14% ofCEMS load, and wind at 28% of CEMS load.

Several aspects of the results are reflected inFigure 9.7. First, the already volatile CA load pro-file implies an optimal mix of relatively little baseload generation compared with the other regions,whereas the very consistent load of the NWPPimplies an optimal mix that is heavily base load,

Table 9.3. Thermal generation costs from EIA

Totalover-night

FixedO&M

Vari-ableO&M

HR

cost($/kW) ($/kW)($/MWh)

Btu/kWh

Scrubbed new coal 2,058 27.53 4.59 9,200

CCGT 962 12.48 2.07 7,200

CT 670 12.11 3.57 10,800

Source: EIA 2007Notes: O&M = operating and maintenance costs; HR = Heat Rate

Table 9.4. Thermal generation costs used insimulations

Total Fuel costs Total

annual fixed marginal

cost($/kWy) ($/MMBtu)

cost($/MWh)

Coal 282.17 1.74 20.60

CCGT 136.57 7.06 52.90

CT 98.51 7.06 79.82

Note: kWy = kilowatt-year

168 James Bushnell

with no peaking resources at all under the baselinescenario. Second, the increasing penetration ofwind resources produces a clear shift of invest-ment toward less capital-intensive peakingresources in every market. This shift is most pro-nounced in the RMPP region, where wind pen-etration is the greatest as a percentage of baseline

CEMS load.Third, less thermal capacity is neededin every market as a reflection of the fact thatwind generation has lowered the residual demandrequired to be served by thermal sources. How-ever, the equilibrium thermal capacity require-ment is reduced only modestly by the entry ofnew wind capacity.

Figure 9.5. Energy market prices for August

25

125

500

25

125

500

8,500 8,600 8,700 8,800 8,500 8,600 8,700 8,800

AZNM CA

NWPP RMPP

2007 load With projected windWith high wind

Pric

e ($

/MW

h)

Hours of month

Graphs by region

Hourly prices by wind penetration for December

Figure 9.6. Energy market prices for December

Building Blocks: Investment in Renewable and NonrenewableTechnologies 169

These factors are summarized in Table 9.5.For each region, the aggregate equilibrium ther-mal capacity and assumed wind capacity are givenin the first two columns of figures. The assumedaverage capacity factor, taken from the wind pro-files from the WECC study is given in the nextcolumn, and the shares of thermal capacity thatare base load and peaking are given in the last twocolumns. Note that the large levels of new wind

capacity, those of more than 10 gigawatts (GW),result in reductions of equilibrium thermal capac-ity of only 1 to 2 GW.

Across the regions, the reduction in thermalcapacity averages about 15% of the new installedwind capacity, with relatively little variation acrossregions. It is important to mention again, how-ever, the strong assumption made here that hydrooutput, particularly in the NWPP region, would

4,000 12,000 20,000 4,000 12,000 20,000

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

AZNM CA

NWPP RMPP

Coal CCGTCT

Graphs by region

Thermal capacity by wind penetration level

Figure 9.7. Equilibrium capacity for energy-only market

Table 9.5. Equilibrium results for energy-only market

Thermal New wind Wind capacity Share Share

Capacity (MW) Capacity (MW) factor Coal CT

CA 23,308 NA NA 43% 44%

22,753 5,670 33% 36% 50%

22,442 11,340 33% 28% 55%

NWPP 14,472 NA NA 93% 0%

13,188 7,890 28% 81% 4%

12,237 15,780 28% 64% 10%

AZNM 20,276 NA NA 73% 11%

19,691 3,840 39% 68% 14%

19,141 7,680 39% 62% 17%

RMPP 6,751 NA NA 86% 7%

6,000 4,650 41% 61% 20%

5,374 9,300 41% 26% 37%

170 James Bushnell

not adjust to the new intermittent capacity. Bytaking advantage of the implicit storage potentialof the hydro resources, one would expect theequilibrium capacity needs in this region to bereduced quite a bit more than implied by thiscalculation.

Capacity Market Results

As in the appendix, the simulation of a capacitymarket requires two important parameters to bespecified. The first is the energy market price cap,set here to be $1,000/MWh. The second impor-tant parameter is the capacity market paymentmade to the generation sources. In order to calcu-late the capacity payment, the implied shortfallthat would be created by capping prices at$1,000/MWh is first estimated.4

In practice, capacity payments are intended toreplace the revenues necessary for investment thatare in principle denied to suppliers through eitherexplicit or implicit restraints on energy prices (seeJoskow 2005; Oren 2005). For this study, this wasaccomplished by calculating the total revenues ofpeaking generation sources under the energy-only scenarios described above. Next, a counter-factual level of income for a 1 MW peaking gen-erator that would have resulted from the sameinvestment levels is calculated, but with pricesearned by generators capped at $1,000. The dif-ference, sometimes known in industry jargon asthe “missing money” caused by price caps, can beexpressed as a dollars-per-kilowatt-year ($/kWy)value.This value was used as the capacity paymentin the second set of simulations. These paymentsare summarized in Table 9.6.

Table 9.6. Capacity payments resulting from $1,000price cap (in $/kWy)

No new 14% of 28% of

wind CEMS CEMS

CA 58.41 54.58 55.84

NWPP 0.00 0.00 0.00

AZNM 1.15 1.57 12.06

RMPP 0.00 2.53 24.80

It is worth noting that these values are quite a bitlower than those currently seen in U.S. electricitymarkets. One reason for this is that the investmentnumbers from the EIA represent generic invest-ment costs for the country, while capacity marketstend to operate in regions of the United States,such as California and New York, where invest-ment can be much more costly. Another moreimportant driver is that these equilibriumsimulations are allowing the price to rise abovethe marginal cost of a peaking plant more fre-quently than has been historically seen in thesemarkets. This is a reflection of the fact that themodel determines the equilibrium, break-evenlevel of capacity, whereas today’s markets tend tofeature more capacity than this level. In practice,today’s capacity markets do not attempt to differ-entiate among causes of revenue shortfalls; theyusually calculate net costs of entry based on his-toric energy prices.5 Therefore, the revenues lostto the price cap in this simulation produce lessmissing money than has been estimated from cur-rent capacity market proceedings.

Next, the above simulations are repeated,with two adjustments to the original model sum-marized by equations (9.4) and (9.5) in the appen-dix at the end of this chapter. The most strikingresults are naturally found in the hours that wereformerly those with prices significantly above$1,000. Figures 9.8 and 9.9 illustrate the changesto the peak hour price duration curves for the CAand RMPP regions because of both wind pen-etration and the capacity market policies. Thesefigures summarize the 150 highest price hours ineach market, in order from highest to lowest.Note that prices in these figures are on a logarith-mic scale because of the high volatility.

For the wind scenarios, the same hours areplotted. As is clear from these figures, the highestprice hours in the baseline simulations are notthose producing the highest prices as wind invest-ment increases. This reflects that fact that as windinvestment increases, prices are increasinglydriven by wind availability as well as total end-usedemand. This is particularly true for the RMPPregion, where the highest price hours under highlevels of wind investment rank below the 100thhighest price hours without the wind investment.

Building Blocks: Investment in Renewable and NonrenewableTechnologies 171

These figures also illustrate the impact that theprice cap has on these “scarcity price” hours. Ingeneral, the highest price hours are reduced to thecap levels, and price levels in most other hoursremain unchanged. The resulting impact of thesecapacity market elements on investment levels aresummarized in Table 9.7.

Revenues of Wind Resources

Because individual wind plants will have varyingprofiles across and within regions, it is difficult tomake general statements about the equilibriumrevenues of wind plants. Nevertheless, the earn-ings are estimated of a hypothetical 1 MW “port-

25,000

100,000

1,000,000

5,000,000

0 25 50 75 100 125 150 0 25 50 75 100 125 150

Capacity market Energy only

2007 load With projected windWith high wind

Pric

e ($

/MW

h)

Number of hours at or above load

Graphs by market

Cumulative prices by wind penetration

Figure 9.8. Highest 150 prices CA

25,000

100,000

1,000,000

5,000,000

0 25 50 75 100 125 150 0 25 50 75 100 125 150

Capacity market Energy only

2007 load With projected windWith high wind

Pric

e ($

/MW

h)

Number of hours at or above load

Graphs by market

Cumulative prices by wind penetration

Figure 9.9. Highest 150 prices RMPP

172 James Bushnell

folio” of plants that features the same productionprofile as the regional aggregate profile used toconstruct the residual demand.

In order to evaluate the revenues of intermit-tent resources under a capacity market paradigm,further assumptions are needed about how thereliable capacity of those resources, on which thecapacity payment is based, is measured. The mostbasic, and clearly overgenerous, method would beto assume that 100% of the installed capacity waseligible for capacity payments. Given the inter-mittent availability of wind resources, this is notthe usual approach. A more conventionalapproach is to discount the installed capacityaccording to a historical measure of the capacityfactor (average energy output divided by capacity)of either the specific unit or the class of technolo-gies from which the unit is drawn. Even thisapproach can overstate the “value” of capacity ifthe production profile of a generation unit isnegatively correlated with total system load. Athird approach, similar to one recently adopted forthe purposes of measuring wind and solar capacityin California, is to measure the production ofresources only during high demand hours, dis-carding production statistics for other hours.6

For purposes of comparison, revenues havebeen calculated under a capacity market in two

ways, roughly following the options outlinedabove. The first approach discounts the capacitypayment according to the annual capacity factor,derived from the wind profile data. The secondapproach calculates a capacity factor only forhours 14 through 17 of each day. The results forthe two alternative calculations of capacity factor(using annual average and peak hour average)were very similar, so only the revenues arereported, assuming capacity payments are basedon the peak hour average capacity factor.

Table 9.8 summarizes the revenues of thishypothetical average wind turbine for eachregion. These values are given in terms of $/kWy.By comparison, the peaking units are earning$95.82/kWy, while coal plants are earning$282.17/kWy.7 To the extent that actual costs ofnew wind facilities would exceed these equilib-rium investment revenue levels, the differencewould have to be captured in subsidies—eitherthrough the production tax credit or throughprice premiums paid by utilities in order to com-ply with their renewable portfolio standards. Asfurther reference, using the same assumptions andcost estimates from the EIA as were used to calcu-late thermal annual fixed costs, wind costs wouldbe roughly $231/kWy.

Table 9.7. Investment levels with a capacity market

Thermal New wind Wind capacity Share Share

Capacity (MW) Capacity (MW) factor coal CT

CA 23,421 N/A N/A 43% 44%

23,141 5,670 33% 36% 50%

22,817 11,340 33% 28% 55%

NWPP 14,472 N/A N/A 93% 0%

13,188 7,890 28% 81% 4%

12,237 15,780 28% 64% 10%

AZNM 20,282 N/A N/A 73% 11%

19,691 3,840 39% 68% 14%

19,168 7,680 39% 62% 17%

RMPP 6,751 N/A NA 86% 7%

6,001 4,650 41% 61% 20%

5,383 9,300 41% 26% 37%

Note: MW = megawatts.

Building Blocks: Investment in Renewable and NonrenewableTechnologies 173

Note that under the energy-only market, therevenues for an average profile wind plant declinein each region. This is because prices are beinginfluenced increasingly by wind availability, and aprofile that mirrors the system wind profile wouldbe producing during hours of glut and not pro-ducing during hours of wind production shortfall.If revenues are instead based on a combination ofcapped energy market revenues and capacity pay-ments, wind producers do a little better thanunder the energy-only paradigm. This is a muchstronger effect under the high wind penetrationscenarios.8 Revenues in the RMPP area are about5% higher with a capacity payment.This is in partbecause the capacity payment rewards productionduring high-demand hours, whereas the energy-only market rewards production during high-pricehours. As wind penetration increases, the high-price hours are relatively more focused on low-wind hours than on high-demand hours.

Impact of a Carbon Market

The last scenario examined is the application of aprice of CO2 onto the electricity sector. For thepurposes of this discussion, the source of the CO2

price could be either a cap-and-trade mechanismor a CO2 tax. Rather than try to calculate aclosed-loop equilibrium price for CO2, it isassumed that these regions participate in broader

CO2 markets with a price of $25/metric ton.Thisvalue is approximately the 2012 futures price forone ton in the European Union’s Emissions Trad-ing System (ETS) market for CO2. It is alsoassumed that the imposition of CO2 prices affectsonly marginal and not capital or fixed costs of anyof the thermal generation technologies.

The same approach as before can be used tocalculate the resulting equilibrium investment lev-els of thermal capacity and revenues for windfacilities. Only the equilibrium is calculated underthe energy-only market paradigm. Figure 9.10illustrates the equilibrium investment capacitiesunder the different wind scenarios. Note that coalis much less used in all markets and is driven outof the CA and RMPP markets completely underhigh wind penetration.

Table 9.9 summarizes the energy market rev-enues earned by the hypothetical wind plantunder various levels of wind penetration. Withthe $25/ton carbon price, wind revenues are sub-stantially higher overall. The degradation of theserevenues with increasing wind penetration is alsomore pronounced, however. With a price onCO2, revenues in the RMPP region are only 5%lower per kilowatt-year under high wind penetra-tion than they would be for the first megawatt ofwind capacity added to that region. This is con-trasted to the almost 15% decline in revenues forthe same comparison in the absence of a carbon

Table 9.8. Summary of the revenues of the hypothetical “average” wind turbine ($/kWy)

Energy-only

New wind New wind

Region CEMS load 14% of CEMS 28% of CEMS

CA 113.24 112.11 109.27

NWPP 123.49 106.48 105.17

AZNM 138.31 135.00 132.75

RMPP 158.39 142.03 135.17

Capacity Market

New Wind New Wind

Region CEMS load 14% of CEMS 28% of CEMS

CA 126.74 124.82 122.31

NWPP 123.49 106.48 105.17

AZNM 138.35 135.59 136.27

RMPP 158.39 143.02 144.36

174 James Bushnell

market. With carbon at $25/ton, wind resourcesare able to earn relatively more revenues duringeven off-peak hours when coal would be settingthe price. Increasing wind penetration leads tomore of these hours, but the differential betweenthese off-peak and on-peak hours is smaller thanin the absence of a carbon price.

Table 9.9. Wind revenues with carbon price at$25/ton (in $/kWy)

New wind New wind

Region CEMS load14% ofCEMS 28% of CEMS

CA 194.15 193.00 185.29

NWPP 188.06 173.46 172.81

AZNM 231.09 228.41 227.96

RMPP 251.85 246.10 238.87

Estimating the Cost of EnergyAvailability Profiles

Given that the equilibrium mix of generationresources can be quite different with high pen-etration of wind resources, it is natural to ask whatthe costs impacts of this changing mix might be.This question is addressed by comparing twohypothetical scenarios. First the total and average

costs of serving the residual demand (e.g., thatwhich is left over after the new energy is applied)are calculated under the assumption that newenergy appears in a manner consistent with thetwo wind penetration scenarios described above.In other words, the average cost of constructingand operating thermal plants to meet the demandthat is not met by the renewable production iscalculated. Second, the same amount of energy isused as in the wind scenarios, but instead assum-ing that it is applied as a base load supply source.In other words the “new” energy is distributedevenly across all hours.

The results of this calculation are summarizedin Table 9.10. The columns labeled “As wind”refer to the same wind distributions that havebeen applied to previous results, and those labeled“as base load” show the results for the evenly dis-tributed energy. Because the more volatile windprofiles require the construction of fewer baseload plants and less frequent operation of peakingplants, average costs are higher under the windprofiles. In California, average costs from the vari-ability of supply increase about 4% ($3/MWh)under 14% wind penetration and close to 9% ($7/MWh) with high wind penetration. In the high-penetration RMPP region, costs rise close to 25%under the high-wind-penetration scenario.

4,000 12,000 20,000 4,000 12,000 20,000

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

Wind at 28%

Wind at 14%

Levels of 2007

AZNM CA

NWPP RMPP

Coal CCGTCT

Graphs by region

Thermal Capacity by Wind Penetration Level

Figure 9.10. Equilibrium capacity investment with $25/ton CO2

Building Blocks: Investment in Renewable and NonrenewableTechnologies 175

ConclusionsThe increasing deployment of renewableresources whose intermittent production is deter-mined by natural forces will reshape the econom-ics of power generation in developed electricitymarkets. This chapter has presented calculationson what the optimal mix of major conventionalgeneration sources would be under variousassumptions of end-use demand and penetrationof wind generation. Data on actual demand forthermal generation in the western United Stateswere combined with highly detailed estimates ofproduction from new wind resources for thisregion of the country. The result is a load shapewith relatively higher spreads between peak andaverage demand for thermal production. As dem-onstrated in the equilibrium model, the amountof coal-fired base load production that would bean economical equilibrium investment steadilydeclines as wind penetration increases. The reli-ance on the low-capital-cost combustion turbinetechnology increases.

Another key change in the economics ofpower systems will come from the rising impor-tance of intermittent production as a driver ofmarket prices. As these simulations demonstrate,the availability (or lack) of wind resources will bean important contributor to market clearingprices. The normal relationship between end-usedemand levels and market prices becomes re-defined as wind resources grow to take a substan-tial share of the market. Implications of this are

that wind resources that are “typical,” in the sensethat their output is correlated with the bulk ofother wind resources, will earn less, and the totalcapacity of wind resources is ramped up. Theirproduction will be correlated with hours of sur-plus and therefore increasingly less correlated withprices. In the presence of a capacity market, thiseffect is more muted. This is because capacitymarkets—at least at present—award capacity pay-ments based on availability during high-demandperiods, rather than high-price periods. This toomay change, however, as the underlying econom-ics of the energy markets become more stronglyinfluenced by the ebbs and flows of intermittentgeneration.

Overall, increasing reliance on intermittentresources creates, or increases, costs in a fashionsimilar to that caused by fluctuating end-usedemand. In planning to serve a system whereconsumption fluctuates widely, firms must turn toresources that are more flexible, but also moreexpensive on an average cost basis. While theadded costs associated with fluctuating end-usedemand can be greatly mitigated by enablingprice-responsive consumption, the intermittencyof renewable supply is a fact of nature. Storagetechnologies can play a valuable role here, andestimates such as those developed in this chaptercan provide an indication of the potential value ofsuch storage options.

Although the analysis in this chapter wasgrounded in data taken from actual energy mar-kets, it is important to recognize the limitations of

Table 9.10. Impact of intermittency on average thermal costs ($/MWh)

New energy 14% of CEMS load New energy 28% of CEMS load

As base load As wind As base load As wind

CA 75.73 78.61 81.87 88.92

NWPP 57.70 59.20 59.02 63.56

AZNM 59.89 61.05 60.71 63.37

RMPP 57.71 63.09 62.28 85.11

With carbon at $25/ton

CA 100.26 102.78 106.46 107.18

NWPP 82.16 83.57 83.15 87.31

AZNM 84.82 85.76 85.56 87.61

RMPP 83.43 88.20 88.43 104.92

176 James Bushnell

this exercise. Two important elements of someelectricity markets are missing here, althoughtheir effects work in opposite directions. Theshort-term operational constraints of thermalgeneration units have not been modeled. Thepresence of such constraints would tend to favorthe nimble combustion turbine technology evenmore heavily. Also not modeled is the potentialreallocation of production from energy-limitedresources, namely hydroelectric. In the PacificNorthwest in particular, this will be an importantresource that can go a long way toward counter-acting the effects of intermittent generation. Infact, even with these limitations, the effects ofwind penetration in the NWPP region are rela-tively minor in contrast to the wind-rich buthydro-poor Rocky Mountain region. The poten-tial for increased trade among the regions also hasnot been modeled, although such trade will belimited by the availability of transmission capacity.

Of course, the real western United States isnot starting from scratch in building its investmentportfolios. Outside of California, coal-fired gen-eration is currently a mainstay of electric compa-nies west of the Mississippi. To the extent thatthese results portend changes in the economics ofthese technologies, they would affect the earningsof the owners of these technologies more than theactual mix of generation resources.

Appendix: An EquilibriumModel of GenerationInvestmentThis appendix describes the technical derivationof the equilibrium investment model employedfor the results presented above. Each conventionalgeneration technology, indexed by i, features amarginal cost ci and fixed cost of capacity Fi. Firmsinvest in capacity that serves a market withdemand that fluctuates over time periods t {(1…T ) with some degree of price elasticity.Demand at time period Qt(pt) is represented as:

Qt(pt) = at – f(pt)

where at is an additive shift of demand and f(pt) is afunction of market price pt.

The perfectly competitive firms in the modelcontinue to add production in any given hour,and capacity overall, as long as the revenues fromadding the production or capacity exceed thecosts. In equilibrium, therefore, production levelsin any hour will be set such that the marginal costof production equals the market price. This equi-librium point can be represented with the follow-ing complementarity condition:

qit ≥ 0 ⊥ pt – ci – ψit ≤ 0 ∀ i, t (9.1)

where ψit represents the equilibrium shadowvalue of the capacity of technology i and willnever be positive if price is below marginal costs ci.This is the shadow price on the constraint thatproduction be no greater than installed capacity, asreflected in the following condition:

ψit ≥ 0 ⊥ qi t – Ki ≤ 0 ∀ i, t (9.2)

Equation (9.1) is therefore equivalent to settingprice equal to marginal operating costs as long asproduction quantities are below the capacity con-straint Ki. The equilibrium level of investmentwill arise from the condition that the value of amarginal unit of capacity equals the cost of thatcapacity.

Ki ≥ 0 ⊥ Fi – Σtψit ≤ 0 ∀ i (9.3)

where Σtψit represents the cumulative value of anextra unit of capacity type i aggregated over alltime periods. Recall that this value is zero for agiven period if the capacity is unneeded in thatperiod, which in this model is equivalent to pricesfalling below the marginal cost of production oftechnology i.

The equilibrium level of investment and pro-duction can be found by simultaneously solvingfor the above three conditions. These conditionsform a complementarity problem (see Cottle etal. 1992) of size t × i. The following sectionsdescribe the data used in formulating the empiri-cal model.

Market Demand

Demand is represented with the partial-log func-tion:

Qrt (prt) = art – br ln(prt)

Building Blocks: Investment in Renewable and NonrenewableTechnologies 177

where r is used to denote the region.The value forb was set at 800 for CA, NWPP, and AZNM, anda value of 400 was used for the smaller RMPA.The price elasticity for this functional form ofdemand is equal to br /Qr, so at the mean observeddemand level (summarized inTable 9.1), elasticityis about 0.05 in each market. In other words, anonzero but still very modest level of demandresponse is assumed. One advantage of this func-tional form of demand is that its convexity implieslittle price response at levels around the marginalcosts of generation, but more response whenprices reach “scarcity” levels over $500/MWh.

Price Caps and Capacity Payments

The above modeling framework imposes two sig-nificant assumptions to reach its equilibrium.First, at least some degree of price response fromend-use demand is assumed. Second, equilibriumenergy prices are not constrained in any way andare allowed to rise in order to balance supply anddemand.

In modeling a stylized capacity market, theabove model is modified in several ways. First, theprice cap is represented with the addition of alarge capacity “fringe” technology with a mar-ginal cost of $1,000/MWh. In other words, inaddition to actual thermal technologies, i, there isan additional complementarity condition similarto equation (9.1) without the capacity constrainton production:

qCAPt ≥ 0 ⊥ cCAP – pt ≤ 0 ∀ t (9.4)

where qCAPt is positive only if the price cap levelcCAP is binding.The quantity qCAPt can be thoughtof as the energy shortfall caused by the price cap,to be dealt with through either demand rationingor out-of-market transactions. To allow for cap-acity payments, equation (9.3) is modified so thatthe annual fixed costs of entry equal energy mar-ket revenues plus the capacity payment:

Ki ≥ 0 ⊥ Fi – CAP_PAY – Σtψit ≤ 0 (9.5)

The capacity market equilibrium is therefore rep-resented by the simultaneous solution of condi-tions (9.1), (9.2), (9.4), and (9.5).

Revenues to Wind Producers

Earnings of the average wind profile are calculatedby multiplying the output of wind production bythe market price for each period. In other words,the energy market earnings of such a portfolio canbe expressed as

Σt prt * CFrt * Capacityr (9.6)

where CF refers to the capacity factor of wind inregion r at time t.

Notes

1. Private independent power producers (IPPs) own83% of cumulative wind capacity in the UnitedStates (Wiser and Bolinger 2009). The passage ofthe Emergency Economic Stabilization Act inNovember 2008 could constitute a major shift inthis trend. Among the act’s many provisions wasthe extension of investment tax credit (ITC) forcertain forms of renewable generation. The actalso allows, for the first time, utility companies totake advantage of the ITC, which had previouslybeen reserved only for nonutility producers.

2. The CEMS data are available at www.epa.gov/cems. The Platts datasets, POWERDAT andBASECASE, are available via paid subscriptionservice at www.platts.com.

3. The study assumes a 15% total renewable penetra-tion, but only half of that is estimated to comefrom wind. However, in 2007, about half of theexisting energy currently generated in the WECCcame from non-CEMS sources. So the wind por-tion of our residual demand profiles is roughly7.5% of total load in the base case, and 15% underthe assumption of doubled wind capacity.

4. The mechanisms and levels for limiting prices var-ies by market. Markets in the eastern U.S. techni-cally limit offer prices to $1,000/MWh. In theorymarket clearing prices can rise above this level, butthey have not in practice done so. One hypothesis(Joskow 2005) is that actions taken by operators topreserve reliability also coincidently limit pricesbelow “scarcity” levels required to recoup invest-ment costs.

5. An example of such a calculation for the NewYork ISO region can be found in NERA (2007).

6. The newly adopted California rule also uses anexceedance measure, rather than a capacity factor.

178 James Bushnell

This means that the capacity payment is based onthe percentage of peak hours in which productionexceeds a given threshold (e.g., 70%) of nameplatecapacity. Using the data in this discussion, thismeasure gave extreme results, so instead the focushere is on a measure of peak hour capacity factor.

7. Recall that these are their annual entry costs, andthe equilibrium conditions equilibrate net operat-ing costs with these annual fixed costs.

8. The $1,000/MWh price cap was almost neverbinding in the NWPP and AZNM regions; there-fore, the results for the energy-only market andcapacity markets are virtually the same.These mar-kets have no “missing money,” and thus no cap-acity payment was necessary even with the pricecap in place.

ReferencesAndo, Amy W., and Karen L. Palmer. 1998. Getting on

the Map: The Political Economy of State-LevelElectricity Restructuring. Working paper 98–19-REV. Washington, DC: Resources for the Future.

Borenstein, Severin. 2005. The Long-Run Efficiencyof Real-Time Electricity Pricing. Energy Journal 26(3): 93–116.

Borenstein Severin, and James B. Bushnell. 2000. Elec-tricity Restructuring: Deregulation or Reregu-lation? Regulation 23 (2).

Bushnell, James. 2005. Electricity Resource Adequacy:Matching Policies and Goals. Electricity Journal 18 (8):11–21.

Cottle, R. W., J.-S. Pang, and R. Stone. 1992. TheLinear Complementarity Problem. London: AcademicPress.

Cramton, Peter. 2003. Electricity Market Design: TheGood, the Bad, and the Ugly. Presented at the 36thAnnual Hawaii International Conference on SystemsScience. January 2003, Hawaii.

Cramton, Peter, and Steven Stoft. 2005. A CapacityMarket that Makes Sense. Electricity Journal 18 (7):43–54.

Ellerman, A. D, P. L. Joskow, Richard Schmalensee, J. P.Montero and E. Bailey 2000. Markets for Clean Air.Cambridge: Cambridge University Press.

EIA (Energy Information Administration). 2007.Annual Energy Outlook. Washington, DC.

Hogan, William H. 2005. On an “Energy Only” Mar-ket Design for Electricity Resource Adequacy.Working paper. Cambridge, MA: Harvard Univer-

sity. http://ksghome.harvard.edu/~whogan/Hogan_Energy_Only_092305.pdf (accessed12 October, 2009).

Joskow, Paul L. 1997. Restructuring, Competition andRegulatory Reform in the U.S. Electricity Sector.Journal of Economic Perspectives 11 (3): 119–138.

———. 2005. The Difficult Transition to CompetitiveElectricity Markets in the United States. In ElectricityDeregulation: Choices and Challenges, edited by JamesGriffin and Steve Puller. Chicago: University ofChicago Press, 31–97.

Joskow, Paul L., and Richard Schmalensee. 1985. Mar-kets for Power:An Analysis of Electrical Utility Deregula-tion. Cambridge, MA: MIT Press.

Kahn, Edward E. 1988. Electric Utility Planning andRegulation. Washington, DC: American Council foran Energy-Efficient Economy.

Lamont, Alan D. 2008. Assessing the Long-Term Sys-tem Value of Intermittent Electric Generation Tech-nologies. Energy Economics 30: 1208–1231.

Nemet, Gregory F. 2006. Beyond the Learning Curve:Factors Influencing Cost-Reductions in Photo-voltaics. Energy Policy 34: 3218–3232.

NERA (National Economic Research Associates).2007. Independent Study to Establish Parameters ofthe ICAP Demand Curve for the New York Inde-pendent System Operator. www.nyiso.com/public/webdocs/committees/ (accessed October 14, 2009).

NERC (National Electricity Reliability Council).2009. 2009 Long-Term Reliability Assessment.www.nerc.com/files/2009_LTRA.pdf (accessedOctober 14, 2009).

Nickell, Bradley M. 2008. TEPPC Renewable EnergyCases: Renewable Generation Information. SaltLake City: Western Electricity Coordinating Coun-cil.

Oren, Shmuel S. 2005. Ensuring Generation Adequacyin Competitive Electricity Markets. In ElectricityDeregulation: Choices and Challenges, edited by JamesGriffin and Steve Puller. Chicago: University ofChicago Press, 388–414.

Schweppe, Fred C., Michael C. Caramanis, Richard D.Tabors, and Roger E. Bohn. 1988. Spot Pricing ofElectricity. Boston: Kluwer Academic Publishers.

Stern, Nicholas. 2006. The Stern Review:The Economicsof Climate Change. London: HM Treasury.

Wiser, Ryan, and Galen Barbose. 2008. RenewablePortfolio Standards in the United States, Berkeley,CA: Lawrence Berkeley National Laboratory.http://eetd.lbl.gov/ea/ems/re-pubs-html (accessedSeptember 9, 2009).

Building Blocks: Investment in Renewable and NonrenewableTechnologies 179

Wiser, Ryan, and Mark Bolinger. 2009. 2008 WindTechnologies Market Report. DOE/GO-102009–2868.Washington, DC: U.S. Department of Energy.

Wolfram, Catherine. 2005. The Efficiency of Electri-city Generation in the U.S. after Restructuring. In

Electricity Deregulation: Choices and Challenges, editedby James Griffin and Steve Puller. Chicago: Univer-sity of Chicago Press, 227–255.

180 James Bushnell

10

Developing a SupergridChristian von Hirschhausen

Harnessing renewables effectively requirestransporting electricity generated by renew-

able energy to centers of demand. The difficultyof doing this is sometimes underestimated in thepolicy debate on renewable targets. Renewableenergy sources are “fuel from heaven” only if thederived products, mainly electricity, are deliveredto the customer’s locale on demand. In this con-text, the concept of a “supergrid,” or transmissionoverlay network, is being intensely debated onboth sides of the Atlantic. This concept has beenpropelled to center stage by politically bold state-ments about the future share of large-scale renew-able energy sources in a carbon-constrainedworld. In addition, China has recently started todevelop long-distance high-voltage direct currenttransmission to transport more hydroelectricityfrom the center of the country to demand loca-tions on its east coast. As the political debateaccelerates, however, the pros and cons ofsupergrids must be objectively assessed and thedifferent, sometimes competing, visions put into acomparative perspective.

This chapter discusses issues related to thedevelopment of supergrids that are currentlydebated in the context of harnessing renewableenergy. The term “supergrid” refers to the assimi-lation of a high-voltage direct and alternating cur-rent (HVDC and HVAC) network overlay with

the existing traditional alternating current (AC)network, with the objective of integrating large-scale renewable energy sources (L-RES) into thegrid. This chapter does not address the linkbetween renewables and existing low-voltage net-works, nor does it cover operational issues, bal-ancing, or other topics at the local low-voltagelevel. Instead, it highlights some of the key issues,conceptually and by using a variety of existingconcepts for potential supergrids. A distinction ismade between studies focusing more on the har-nessing aspect of renewables, such as the SolarGrand Plan, and those that include a more seriousnetwork component, such as the Joint Coordi-nated System Plan developed for a group ofNorth American system operators. An inverserelation exists: the more a supergrid study looks atthe obstacles to developing the appropriate trans-mission infrastructure, the less such studies tend toprefer supergrids, instead arguing in favor of moredecentralized, regional solutions.

If it is commonly accepted that networkdevelopment is important for harnessing L-RES,few researchers to date have looked closely at theengineering-economic aspects of expansionprojects on both sides of the Atlantic.The conceptof linking European electricity markets to renew-able energy potential in the Middle East andNorth African (MENA) region via HVDC trans-

mission is particularly interesting and is presentedas an in-depth case study. The European Eco-nomic Association–Middle East North Africa(EEA-MENA) supergrid interconnection plan for2050 is a variant of the Desertec project and theMediterranean Solar Plan (MSP), but with itsown distinctive features. Thus, in addition toincluding different amounts of concentrated solarpower (CSP) electricity imported from theMENA region to Europe by constructing asupergrid, the study also considers northernEuropean offshore wind and hydro resources andfinds that the consideration of a truly Europeanmarket for renewable electricity changes thetopology of the optimal supergrid significantly,tending toward more regional clustering. Alsohighlighted is an important distributional issuethat arises between “winners” of more transmis-sion lines, whose electricity prices fall, and “los-ers,” whose prices increase.

Next, several critical issues for the furtherdevelopment of supergrids are examined. Trans-mission investment has been problematic in boththe United States and Europe, and the develop-ment of a supergrid poses even more com-plex long-term challenges to industry andpolicymakers. For industry, a balance has to bestruck between lucrative investments in genera-tion and transmission assets and the economic,political,andregulatoryrisksrelatedthereto.Policy-makers, on the other hand, face the challenge ofhaving to design instruments and regulatoryframeworks to achieve very long-term objectivesgiven high uncertainty and widely divergingincentives between stakeholders, such as federalversus regional levels and producers versus con-sumers. Surprisingly little attention has beengiven to long-term planning mechanisms, a criti-cal element in such complex projects. Planningand regulatory issues are closely related to financ-ing issues, and the optimal organizational struc-ture for a supergrid has yet to be identified. Thediscussion also includes a political economy inter-pretation of the obstacles to overcome, amongother compensating mechanisms between poten-tial winners and losers.

The chapter therefore concludes on askeptical note. The findings that have been pre-

sented temper the enthusiasm that generallyaccompanies supergrid concepts: although theunderlying objectives for their development maybe well founded, the large number of obstaclesmakes a rapid emergence of supergrids unlikely.Instead, it seems more reasonable to focus on astepwise, bottom-up approach by which transmis-sion expansion follows a gradual, more traditionalpath.

Stylized Supergrid Projects inthe United States andEurope/North Africa

Typology of Supergrids

The concept of supergrids is closely related toattempts to tap distributed renewable energysources and thus pave the way to a low-carbonelectricity system. All studies on sustainableenergy systems acknowledge the need for large-scale high-voltage transmission. Jacobsen andDelucchi (2009) have sketched out “a path to sus-tainable energy by 2030 in which wind, water,and solar technologies could provide 100% of theworld’s energy, eliminating all fossil fuels.” Thisvision also includes a decarbonized transportationsector. The authors stress the importance of trans-mission: “each nation needs to be willing to investin a robust, long-distance transmission system thatcan carry large quantities of wind, water, and solartechnologies from remote regions, where it isoften greatest—such as the Great Plains for windand the desert South West for solar in the UnitedStates—to centers of consumption, typically cit-ies” (65).

Although a precise definition for the term“supergrid” does not exist, this chapter followsthe Jacobsen–Delucchi suggestion of what isrequired to make a large-scale renewables-basedenergy system workable in terms of transmissioncapacity. More generally, there is a commonunderstanding that supergrids have become syn-onymous with transcontinental electricity net-works, tying in the existing high-voltage grid.

182 Christian von Hirschhausen

Also, the term “transmission overlay” is used tocharacterize a portfolio of high-voltage transmis-sion additions to existing electricity networks thattogether serve and link entire regions and markets(Kaupa 2009; Midwest ISO et al. 2008).Supergrids are thus characterized by the follow-ing:

• flexibility in system balancing;• high capacity for bulk power transmission;

and• geographically long distances.

In addition to the prevailing alternating current(AC) technology in power transmission, HVDCor Flexible AC Transmission Systems (FACTS)1

technologies are increasingly used in supergrids.These hybrid systems can improve transmissioncapacity and system stability by bypassing heavilyloaded AC systems.The relevance of these featuresincreases with the prospective integration oflarge-scale renewable generation capacities thatproduce higher fluctuations in loads. Moreover,power generation becomes increasingly distrib-uted and a growing number of generation capaci-ties are located far from load centers, leading togreater transmission distances. In this context,HVDC offers two important advantages over ACtechnologies:

• On the one hand, higher bulk power capaci-ties can be realized per line. Thus fewer linesare necessary, an important factor in the pub-lic’s acceptance of new transmission projects,which is often low because of the not-in-my-backyard (NIMBY) mindset.

• On the other hand, HVDC lines have lowerpower losses over long distances, as shown inTable 10.1. Despite the higher initial invest-ment costs of HVDC and FACTS, thesecharacteristics make DC bulk power trans-mission (≥ 1,000 megawatts) more economi-cal than AC for transmission distances above600 kilometers (Claus et al. 2008). HVDCtransmission technology is rapidly evolving.It allows longer water crossings and does notrequire phase synchronization.2

Table 10.1. Characteristics of high-voltagetransmission technologies

Voltage 735 kVAC

500 kVDC

800 kVDC

Power losses per1,000 km line

6.7% 6.6% 3.5%

Transmission capacity 3 GW 3 GW 6.4 GW

Source: Siemens 2009, 4

The multiple visions of future supergrid develop-ments can be classified according to technical,economic, or institutional criteria. To structurethe subsequent discussion, electricity generationand transmission are differentiated as follows:

• generation: supergrid projects that rely onone source of renewable generation(singlesource) or integrate a variety of L-RES(multiple sources); and

• transmission: gradual extension of existingAC with additional local and interregionalpower lines or radical changes that include atransmission overlay with multigigawatt“highways” interconnecting entire conti-nents.

Table 10.2 classifies the major existing visions andconcepts for the United States and Europe basedon these criteria. The subsequent sections reviewstudies with different grades of detail and focus,ranging from conceptual ideas to more detailedcalculations, and outline some of the most rel-evant supergrid visions. This discussion followsthe structure in the table, starting with the U.S.projects and then looking at those in Europe andNorth Africa.

Supergrid Projects in the United States

Following are some of the major supergridprojects in the United States. For other U.S.supergrid projects, see the survey by Tierney(2008).

The Solar Grand Plan

Zweibel et al. (2008) and Fthenakis et al. (2009)developed a vision of a solar-based supergrid

Developing a Supergrid 183

across the United States, called the “Solar GrandPlan.” In essence, it is based on a combination ofphotovoltaic (PV) facilities spread across differentareas and concentrated solar power (CSP) systemsin the Southwest, where the highest average solarirradiation in the country can be harnessed (Fig-ure 10.1).

The Solar Grand Plan also includes a substan-tial amount of compressed-air energy storage(CAES) located close to demand. New HVDClines would connect the Southwest to the rest of

the country. The technologies, scale, and learningfor a project of this size are substantial; thus 2,940GW of PV should be developed through anational energy plan built around solar power.Increased thin-film module efficiency from 10%to 14% and reduced installation costs from $4 to$1.20 per watt (W) in 2050 would lead tolevelized electricity costs of $0.05 per kilowatt-hour (kWh). A corresponding development is alsoexpected for CSP, though at a somewhat smallerscale: cumulated 558 GW in 2050 will lead to

Average irradiation in kWh/m2/day:8 7 6 5 4 3 2

Source: Zweibel et al. 2008, 53

Figure 10.1. Solar radiation in the United States

Table 10.2. Studies for large-scale integration of renewable energy sources

Energy SourceGeographic approach

Continental Regional

Single Zweibel et al. (2008), A Solar Grand PlanTrieb et al. (2009), Characterisation of Solar

Electricity Import Corridors from MENA toEurope (Desertec, Mediterranean Solar Plan)

Airtricity (2006), European Offshore SupergridProposal

Midwest ISO et al. (2008), Joint Coordinated Sys-tem Plan 2008

Office for Metropolitan Architecture (OMA 2009),Masterplan Zeekracht

Multiple Trieb et al. (2006), Trans-Mediterranean Inter-connection for Concentrating Solar Power

AWEA and SEIA (2009),a Green PowerSuperhighways

Krapels (2009), Integrating 200,000 MWs ofRenewable Energy into the US Power Grid

Egerer et al. (2009), Sustainable Energy Networksfor Europe

aThis study builds on a wind-only study for the United States; see AEP 2007.

184 Christian von Hirschhausen

decreasing installation costs from $5.30 to$3.70/W. Thus electricity generation costs forCSP drop to $0.09/kWh. By 2050, solar powerwould provide 35% of total energy consumed inthe United States and 70% of electricity. If imple-mented, the Solar Grand Plan would substantiallyreduce fossil-fuel consumption, CO2 emissions,and U.S. import dependence.The plan’s total costssum up to $420 billion until 2050.3 Fthenakis etal. (2009) concentrate on the feasibility of thevision laid out by Zweibel et al. (2008). Theyrefine the proposal of a two-step implementationplan: Stage I (from the present to 2020) comprisesa 10-year solar deployment and incentive programthat includes guaranteed loans, a mandatory solarportfolio standard for electric utilities, and a solarfeed-in tariff scheme. Costs are estimated at $300billion; this amount would ensure a transition toself-sustained growth in the CSP and PV marketsafter 2020. Nevertheless, Stage II (from 2020 to2050) should include a commitment to an annualdeployment schedule to sustain growth forrenewable technologies.

It is evident that HVDC transmission is a cru-cial part of the Solar Grand Plan, as 90% of theplan’s solar energy production is located in theSouthwest. Transmission-wise, the Solar GrandPlan therefore proposes a radical approach: anentirely new HVDC power transmission back-bone. Interestingly, no major technical advancesare deemed necessary. During Stage I, the firstexpansion of the DC transmission system wouldoccur, extending the network via existing rights-of-way along interstate highway corridors, whichwould minimize land acquisition and regulatoryhurdles.4 In Stage II, the remaining extensionswould occur.The HVDC transmission companieswould not need to be subsidized, because “theywould finance construction of lines and converterstations just as they now finance AC lines, earningrevenues by delivering electricity” (Zweibel et al.2008, 57).5

However, three issues are not addressed by theauthors. First, the role of other L-RES, such aswind power in the Great Plains, as potential com-petitors to solar power is underrepresented. In thiscontext, it appears questionable that solar powerfrom the Southwest will serve the majority of the

U.S. electricity supply until 2050. Second,nationwide transmission planning is key, yetFthenakis et al. (2009) give a vision of HVDCtransmission only for the Southwest.The anatomyof their HVDC transmission overlay does not takeinto account storage sites, also a key element ofthe plan. Third, economic analyses of the welfareeffects as well as profit considerations have notbeen included so far.

Wind-Based Transmission Expansion Plan

The Joint Coordinated System Plan 2008(JCSP’08) contains a wind-based regional trans-mission expansion plan developed by the systemoperators Midwest ISO (MISO), Pennsylvania–New Jersey–Maryland Interconnection (PJM),Southwest Power Pool (SPP), Tennessee ValleyAuthority (TVA), Mid-Continent Area PowerPool (MAPP), and several key members of theSoutheastern Electric Reliability Council(SERC). The JCSP evaluates scenarios of a gridoverlay in the year 2024 for the Eastern Intercon-nection. An optimal transmission grid overlay hasbeen modeled for two scenarios of power genera-tion capacity developments: a reference scenarioand a 20% wind energy scenario. In both sce-narios, the transmission overlay includes 800 kilo-volt (kV) HVDC as well as 765 kV, 500 kV, and345 kV AC lines, but the transmission expansiondiffers. The reference scenario assumes that thepresent renewable portfolio standard (RPS)requirements are met in 2024 with local onshorewind resources. The wind scenario assumes a sig-nificant enlargement of wind power capacitiescontributing to 20% of the U.S. Eastern Intercon-nection energy use.6 The two scenarios’ mainassumptions and results are displayed inTable 10.3.

The reference scenario suggests primarily345 kV line extensions in the Northwest.7 Pre-liminary analyses show that the potential benefitsfrom reduced energy costs for consumers in theEast exceed incurred costs of approximately $50billion on an aggregate interregional level in com-parison with the present grid configuration. The20% wind energy scenario shows significantlymore high-voltage transmission capacity (see Fig-ure 10.2). Approximately 75% of the conceptual

Developing a Supergrid 185

transmission overlay consists of 765 kV AC or800 kV DC. Supported by a network of 345 and500 kV AC lines as well as several 800 kV DCfeeder lines, seven major HVDC lines will trans-

port electricity from renewables and thermalplants in the Midwest to the load centers in theEast and Southeast.

JCSP Option 3 3 1 Lines by Voltage kV

JCSP Option 3 3 1 Two Circuit lines by Voltage kV

JCSP Option 3 3 1 HVDC Feeder Lines by Vo tage kV

Existing Lines greater than 230 by voltage kV

230 (1)345 (48)500 (18)765 (27)800 HVDC (2)

800 to 801 (2)765 to 800 (34)500 to 765 (560)345 to 400 (1324)230 to 345 (4111)

400 and 800 HVDC

345 (4)500 (5)

Source: Midwest ISO et al. 2008, 9

Figure 10.2. JCSP’08 wind energy scenario conceptual transmission overlay

Table 10.3. JCSP’08 main scenario assumptions and results

Reference scenario Wind scenario

Percentage Percentage

New generationexpansion capacity(MW)

Wind 58,000 31% 229,000 67%

Base load steam 76,800 40% 37,200 11%

Gas CT 49,200 26% 69,600 20%

Gas CC 4,800 3% 4,800 1%

Other fossil 1,200 1% 1,200 0%

Total 190,000 100% 341,800 100%

Transmission overlay(miles)

HVAC 7,109 71% 6,898 48%

HVDC 2,870 29% 7,582 52%

Total 9,979 100% 14,480 100%

Transmission capitalcost (2024 mill-ion $)

Transmission overlay 42,159 72,825

Transmissionsubstations

6,401 7,074

Source: Midwest ISO et al. 2008, 6

186 Christian von Hirschhausen

Methodologically advanced, the high-renewables scenario for transmission planning canbe critically assessed with respect to the concen-tration on onshore wind. Canadian hydro poten-tial as well as offshore wind capacities should bemore strongly reflected in such a scenario todetermine future transmission expansion. Mid-west ISO et al. (2008) also recommend that futureanalyses should pay more attention to local anddecentralized energy generation scenarios ratherthan long-distance production and transmission;this would affect the type, location, and cost ofnecessary transmission infrastructure and couldchange the associated social benefits.

Integrating Wind, Solar, andBiomass Generation

AWEA and SEIA (2009) and Krapels (2009) areprominent examples of multisource studies, witha focus on wind and solar energy, as well as somegeothermal and biomass. In contrast to the SolarGrand Plan project, Krapels adopts a bottom-up,regional approach to transmission expansionbased on the belief that “the structure and govern-ance of the power industry in the United Statesfalls along regional lines, strongly suggestingefforts to meet the environmental goals should beregional” (2009, 4). The diversity of the states’interests, political alignments, and resources sug-gests that any federal policy should only definebroad objectives and allow the states and regionsto determine how to achieve the target.

Krapels’s analysis assumes that a national RPSof 20% would require about 200,000 MW ofwind, solar, and biomass generation capacities.This, in turn, would necessitate significant invest-ment in generation facilities and an overhaul ofthe transmission systems to take the renewableenergy to market and achieve the following:

1. on the East and West Coasts, connect nearbyterrestrial and offshore wind resources to thepopulation centers;

2. in the Midwest and Southwest, connect thehighest-quality wind to inland populationcenters on both sides of the ContinentalDivide;

3. in Texas, connect the wind and solar in theNorth and West to the population centers inthe Center and South; and

4. in the South, where wind resources aremeager, allow nuclear power plants to meetRPS and carbon targets (Krapels 2009, 4).

Krapels rejects the economic and political feasibil-ity of a coast-to-coast “electricity superhighway”and argues for a series of initiatives from coastalstates that would essentially confine the super-highway to a smaller area complemented with asystem enabling the coastal states to harness near-terrestrial and offshore wind (see Figure 10.3).This coastal complement would eliminate theneed for midwestern wind delivered via the pro-posed national electricity superhighway. Accord-ing to Krapels, regional transmission planning inthe United States has not led to the desired inter-state transmission expansion to increase competi-tion, improve security of supply, and lower aver-age consumer prices.8 These traditional mecha-nisms will not allow L-RES to be built at theenvisioned scale and thus must be addressed indetail. Despite the regional approach, institutionalhurdles remain the most challenging part of trans-mission expansion.

Supergrid Projects in Europe/North Africa

This section looks at some of the major supergridprojects in Europe and North Africa. SeveralEuropean Union-sponsored projects also assesssome supergrid proposals, such as SUSPLAN andthe SOLID-DER project. Other ongoing orrecent studies include WIND FORCE 12 byGreenpeace, Tradewind, and DENA 2.9

Submarine Wind Energy Superhighwaysfor Europe

In Europe, different projects for large-scale windintegration are currently on the agenda, based onoffshore transmission lines. The vision byAirtricity (2006), a wind power developer andoperator, integrates large offshore wind capacitiesalong the continent’s coastlines into the Europeanelectricity system via a submarine grid of HVDC

Developing a Supergrid 187

lines (see Figure 10.4).The connection of distrib-uted wind over long distances reduces windfarms’ power variability. It has been shown thatpower output correlation decreases as distancebetween wind farms increases (see, e.g., Sinden,2007), because wind patterns are regional andchange over longer distances as a result of areas ofhigh and low barometric pressure. The question,however, is to what extent wind power will proveto be a reliable and predictable source of energy(Airtricity 2006). Considering the demand side,the differences in lifestyles, time zones, and uses ofelectricity will lead to elongated peak demand andtherefore result in a greater capacity credit for thesupergrid than for offshore wind farms being con-nected to a single national electricity system.Moreover, the study argues that the HVDC sub-marine grid would resolve much of the complexi-ties associated with achieving a single Europeanelectricity market, as the supergrid would serve asan interconnector among national markets.

The study delivers neither capacity projec-tions nor cost estimations for the proposed pan-European submarine supergrid. Instead, it sug-gests an initial project in the North Sea as anucleus for further HVDC expansion.10

Airtricity’s (2006) calculation for this 10 GW ini-tial project deserves critical evaluation withrespect to two issues. First, the project’s econom-ics are strongly dependent on the assumed finance

structure, but the proposed debt interest rate of5.5% per annum seems rather low for a project ofthis size and scope. Hence it implicitly requiresfinancial guarantees by public institutions to lever-age private capital. Second, because a predictableyield for investors is necessary to attract invest-ment, it will likely require a fixed electricity pricefor wind power and preferred feed-in guarantees.Institutional and political barriers will remain themajor roadblocks to such supranational agree-ments.

Source: Krapels 2009, 4

Figure 10.3. Integrating 200,000 MW of renewable energy in the U.S. grid (exemplified for wind power)

Source: Airtricity 2006, 9

Figure 10.4. Airtricity’s European offshoresupergrid proposal

188 Christian von Hirschhausen

The North Sea Wind Energy Super Ring

A less ambitious and more focused project is theNorth Sea Wind Energy Super Ring. The idea isto develop a meshed transmission grid in theNorth Sea to better connect the grids of north-west Europe, the United Kingdom, and Scandi-navia. This interconnection should facilitateexploiting the potential of the North Sea’s windresources, as well as reducing the problem ofintermittent wind availability.

One of several concepts has been designed bythe Office for Metropolitan Architecture (OMA)for the Society for Nature and Environment(Natuur en Milieu 2009). The authors calculate atheoretical annual electricity generation potentialof 13,400 terawatt-hours (TWh) from windpower in the North Sea. The Master Plan for theNorth Sea develops the structure of a HVDC ringto harness the wind potential by connecting tothe load centers on the European continent andthe British Isles (Figure 10.5). Including sevencountries, the ring structure allows flexible dis-patch and integrates Scandinavian pump storagefacilities. Moreover, exploited submarine gas andoil fields in the North Sea could potentially serveas compressed air storage facilities. The meshedstructure increases security of supply in compari-son with single transmission lines. If the EnergySuper Ring were implemented, it is estimatedthat Europe could reach energy independencefrom Russia and the Gulf states in 2050 (Natuuren Milieu 2009).

Although this is an interesting approach, sev-eral questions remain concerning its implementa-tion according to OMA (2009).The plan includesno projections about the actual wind powercapacity needed, nor does it specify the necessarytransmission line capacity. It also does not con-sider the possible use of submarine gas and oilfields for future carbon capture and storage(CCS). Hence compressed air storage capacitycould be less than expected. Finally, the plan doesnot indicate the extent to which Scandinavianhydropower reservoirs are able to store windcapacity from the proposed project.

The Desertec Project

The Desertec concept, a mainly solar-based, radi-cal transmission expansion project, is being devel-

oped jointly by theTrans-Mediterranean Renew-able Energy Cooperation (TREC), a network ofscientists, politicians, businesspeople, and theGerman society of the Club of Rome.11 Thetechnical studies were carried out by the GermanAerospace Center (DLR) and financed by theGerman Ministry for the Environment (BMU)(see Trieb et al. 2005, 2006).12

The Desertec project focuses on the genera-tion of about 2,400 TWh per year of mainly CSPby 2050. This is based on generation cost reduc-tions for CSP, where electricity costs are expectedto drop from 9.8 euro cents (U.S. 13.3 cents) perkWh in 2015 to 5.5 euro cents (7.5 cents) perkWh in 2050.13 About 700 TWh of this solarelectricity will be supplied to the EuropeanUnion (EU) via a new HVDC overlay network.For this, Trieb et al. (2009, 5) assume an HVDCtransfer capacity of 20 × 5 GW cables. Investmentis in the same range as for the U.S. Solar GrandPlan, €400 billion ($545 billion) until 2050, ofwhich €45 billion ($61 billion) is dedicated to theHVDC overlay network. Figure 10.6 shows thestructural network that has been developed using

Source: OMA 2009, 9

Figure 10.5. The North Sea wind power transmis-sion ring

Developing a Supergrid 189

least-cost algorithms with several constraints. AnHVDC core connects 11 CSP production sites inthe MENA region with 27 European demandcenters, including corridors for exports of SaudiArabian CSP to destinations as far away as Milan,Paris, and even London.

The Desertec project is more explicit aboutthe institutional structure of the transmission linesthan its U.S. equivalent, the Solar Grand Plan. Itproposes the creation of an international transmis-sion association to be owned by the companiesparticipating in the construction and operation ofthe HVDC lines. The Nabucco Gas PipelineInternational GmbH, a project company set up todevelop a multinational natural gas pipeline fromTurkey to Austria, is considered a role model.Thetransmission project association would be respon-sible for planning, financing, constructing, oper-ating, and marketing of the supergrid. A closerelation between the national and interregionaltransmission system operators (TSOs) would needto be ensured either through direct ownership orcontractual arrangements.

The main criticism of the trans-Mediterranean supergrid as proposed in Trieb etal. (2009) is related to other L-RES deployment.The proposed HVDC network overlay does nottake wind and hydropower in the North Sea and

Scandinavia into account. Thus the authors implythat solar power from the MENA region cancompete with these renewables in the UnitedKingdom and the countries adjacent to the NorthSea. Considering the potential for other L-RES,the supergrid’s layout (as shown in Figure 10.6)appears questionable. In addition, the implemen-tation of the proposed HVDC grid overlay,stretching from Saudi Arabia to the United King-dom, seems challenging in terms of its institu-tional and political obstacles even under the pro-posed institutional design.

Evaluating the Supergrid Proposals

The concepts of the supergrids reviewed hereshow differing foci and grades of detail. Somestudies (AWEA and SEIA 2009; Zweibel et al.2008) concentrate on renewable energy capacitydevelopment and address transmission expansionfor integrating L-RES into the existing grid as anecessary side condition. Other studies (e.g.,Midwest ISO et al. 2008) explicitly model futuretransmission overlays for different generation sce-narios, accounting for consumer benefits and reli-ability issues.

While none of these studies is beyond criti-cism, all make one point clear: the supergrid

Source: Trieb et al. 2009, 124

Figure 10.6. The Desertec HVDC overlay network (physical map)

190 Christian von Hirschhausen

projects require reaching an operational level oftransmission planning and regulation. Surpris-ingly, in none of the numerous studies surveyedfor this chapter is there any reference to technol-ogy as an obstacle. There is also a large consensusthat transmission issues are critical, whereasexpanding the use of renewable energy sources assuch is less difficult. This holds even though thepure investment sums for tapping the renewableenergies significantly exceed the transmissioninvestments.

Few studies surveyed include an economicanalysis beyond some rough financial indicators,such as costs. In the United States, Midwest ISOet al. (2008) calculate consumer benefits underconstraints of security of supply. For Europe,Trieb et al. (2009) provide some cost predictionsbut do not address welfare considerations. Thenext section therefore explores these aspects in theanalysis of another specific project, the EEA-MENA supergrid, before the discussion turns tothe general issues faced by all of the supergridprojects.

Case Study: EEA-MENA 2050SupergridThis case study is an engineering-economicanalysis of the European–North AfricanSupergrid (EEA-MENA 2050). It proposes awelfare-optimal HVDC extension plan underassumed CSP capacity deployment in the MENAregion, using a network model of the Europeanelectricity market. The model maximizes thevalue of total welfare less the annuity for theselected transmission lines to determine optimaltransmission corridors. It also mirrors the evolu-tion of a trans-Mediterranean supergrid in dis-crete time steps of 10 years until the year 2050.14

Modeling Approach

The subsequent simulation, applies anengineering-economic DC load flow modelbased on ELMOD (Leuthold et al. 2008).ELMOD is a welfare maximization model with

technical constraints, including thermal limits,electricity losses, loop flows, and security con-straints. The applied model consists of 105regional zones, represented by nodes. The gridconsists of interzonal high-voltage lines con-nected to the nodes by auxiliary nodes and lines.Each main node has its own demand and genera-tion portfolio. The existing European electricitygrid, comprising the UCTE region, NORDEL,the United Kingdom transmission system opera-tor, and other UKTSO members, is modeled,with the model limited to the high-voltage level(132 to 750 kV).The existing AC grid includes allexisting interzonal links and high-voltage AC linesaccording to ENTSO-E (2008) and the otherEuropean TSOs. Concerning the future develop-ment of the AC grid, all projects planned by theEuropean system operators until 2030 areassumed to be completed (ENTSO-E 2009).15

Further AC grid expansion is not implementedbecause of the lack of available informationbeyond 2030. In addition to the AC grid, existingDC connections are implemented in the modelwith its technology-specific parameters.16

Electricity demand is characterized by a refer-ence demand and a reference price. While thereference demand is given by the scenarios, thereference price equals the costs of the marginalplant under base load conditions. A linear inversedemand function is determined for each regionwith an assumed price elasticity of 0.1. The CSP-exporting regions are modeled as simple exportnodes.

Generation capacities comprise 14 types ofplant according to three groups: fuel-fired(nuclear, lignite, hard coal, gas-fired combinedcycle gas turbines [CCGT], gas, oil, and biomass);renewable generation (run of river, wind on- andoffshore, PV, CSP, and geothermal plants); andstorage (hydropower reservoirs). As the fluctuatingcharacter of wind generation cannot be deter-mined by season or time of day, three cases—high,medium, and low—are included for each loadlevel. This leads to a total of 24 model cases, dis-tinguishing among seasons, day and night,demand, and wind availability. All model caseshave equal weights.17 An overview of the en-dogenous and exogenous model parameters is pre-sented in Table 10.4.

Developing a Supergrid 191

Scenarios

Given the high variance in any of the parameterestimations until 2050 (e.g., demand, technolo-gies, climate policy), the study is limited to abusiness-as-usual (BAU) scenario as the referencecase and a technological development (TD) sce-nario that captures the main characteristics of arenewable future as possible boundaries for theEEA-MENA supergrid. Figure 10.7 provides anoverview of the generation capacities for the BAUandTD scenarios for 2010–2050. Fossil-fuel price

developments are based on current market pricesand develop according to fuel-specific escalationrates ranging between 1% and 2.5% per annum (inreal terms). In both scenarios, slightly increasingelectricity demand is expected, from 3,500 TWh(2008) to 4,200 TWh (2050).

In the BAU scenario, coal, lignite, and nuclearremain almost on the same level as today except inGermany, where nuclear is phased out. Theappearance of L-RES is limited: only 60% ofEuropean offshore wind and no Scandinavianhydropower potentials are exploited, and CSP

NuclearLigniteCoalCCGTGasOilHydroReservoirWind/Onshore

Wind/OffshorePhotovoltaicCSPBiomassGeothermal

Technological development Business as usual[GW]

1,400

1,200

1,000

800

600

400

200

02010 2020 2030 2040 2050 2010 2020 2030 2040 2050

Source: Egerer et al. 2009, 42

Figure 10.7. Installed capacities for the BAU and TD scenarios, 2010–2050

Table 10.4. Exogenous and endogenous parameters of the model

Exogenous Zonal demand and generation capacityCSP generation expansion in MENA and number of HVDC linesMarginal costs for generation technologies including CO2 pricea

Extension of AC grid according to existing extension plans until 2030Scenario-based evaluation: influence of stronger integration of Scandinavia to Continental Europe via HVDC con-

nections

Endogenous Determination of the welfare optimal integration strategy for CSP by an assessment of connections from 3 MENAregions to 30 different demand centers in Europe

Allocation of yearly seasonal hydro storage (reservoirs) generation budget for balancing purposes (wind, demandlevels, etc.)

a Prices for emissions certificates are based on a review of price projections (see Egerer et al. 2009, 44). CO2 prices have been allocated togeneration costs for different technologies via technology-specific emission factors.

192 Christian von Hirschhausen

capacity in the MENA region does not exceed12 GW until 2050. This development assumesmoderate CO2 certificate prices, with €24 ($33)per metric ton in 2030 and €30 ($41) per ton in2050. In contrast, the TD scenario adopts a per-spective of an 80% reduction of CO2 emissions by2050 (base 1990). To realize these targets, anincrease of the CO2 price to €57 ($78) per ton in2030 and €100 ($136) per ton in 2050 is assumed.Also assumed are intensive technology R&D andthat the technical potentials described in terms ofharnessing renewables, particularly wind, areattained. Nuclear is phased out, and any potentialhydrogen developments are ignored. The deploy-ment of smart grids allows a higher share of localdecentralized renewables in addition to theL-RES that are modeled explicitly. Natural gasserves as backup capacity.

In addition to CSP integration in NorthAfrica, other L-RES potential in Europe isreflected, in theTD scenario, by two subscenariosfor the integration of wind and hydropowercapacities. The first subscenario, A, assumes re-inforcement of the grid in addition to themodeled network described above to transportthe vast amount of future offshore wind from theNorth Sea to demand centers in the United King-dom, France, and Germany. This is reflected inthe model by additional transmission in the coun-tries adjacent to the North Sea. In the secondsubscenario, B, the HVDC lines connectingScandinavia to the rest of Europe are expandedexogenously to a capacity of 8 GW per connec-tion in 2050. These lines are modeled in additionto the transmission capacity according tosubscenario A.Thus subscenario B can be consid-ered to be close to the concept of the North SeaEnergy Super Ring discussed above.18 These twosubscenarios allow an examination of the welfarebenefits of seasonal hydropower generation fromreservoirs as possible balancing power to offshorewind.

Results

Following are the results of the model. First, gen-eration capacities and prices for the BAU and TDscenarios are discussed. The second part proposes

welfare maximizing paths for grid expansion inthe EEA-MENA supergrid. Third, effects from astronger integration of the Scandinavian electri-city market are considered.

Generation and Prices

In the BAU scenario, coal remains a cheaper alter-native to natural gas. With the stronger growth ofthe CO2 certificate price in the TD scenario, coalloses its competitiveness against CCGT genera-tion. As the model structure allows only limitedpeak demand, the gas share in reality mightincrease as a result of the need for open-cycle gasturbines (OGT) to serve as suppliers for peakdemand with only a few full load hours. Whereasin the BAU scenario, 60% of the electricitydemand is supplied by fossil fuels, this valuedecreases to about 35% in the TD scenario. In2050, CSP delivers about two-thirds of the elec-tricity provided by nuclear today. Biomass as aCO2-neutral fuel becomes an important source ofelectricity too. Wind generation is only 25%higher than in the BAU scenario. Being lessaffected by high CO2 prices, CCGT becomes thedominant fossil plant in the TD scenario. Its sharedoes not decrease until 2040, as RES capacitiescontinue to grow.

Because of the stronger growth of the pricefor CO2 certificates, the TD scenario leads to astronger increase in the average electricity price,19

exceeding €120 ($163) per MWh in 2050, almostdouble the price in the BAU scenario. For theBAU scenario, the remaining CCGT is used onlyfor mid and peak load, where prices are higherthan the average electricity price. In the TD sce-nario, CCGT plants have lower marginal genera-tion costs and replace coal in the mid and baseload. The high electricity price can be explainedby more frequent use of backup OGT, whichhave higher marginal costs as a result of theirlower efficiency. Finally, the prices also indicatethe strong increase of biomass capacity in the TDscenario favored by the steep growth of the CO2

certificate price.

HVDC Grid Expansion

In the BAU scenario, a CSP capacity of 11.8 GWis installed in Morocco, Algeria,Tunisia, and Libya

Developing a Supergrid 193

by 2050. For this limited electricity generationcapacity, four transmission cables are built, and thecapacity is split in half between the Moroccan andTunisian export nodes.The first two lines are builtin 2030, and the other two are installed in 2050.In 2030, CSP electricity exports from Tunisia toItaly yield only 3.8 euro cents (U.S. 5.2 cents) perkWh cost coverage and, because of higher elec-tricity prices in Spain, 4.4 euro cents (6 cents) perkWh cost coverage for Moroccan CSP plants.Rising electricity prices in Spain until 2050 willlead to 6 euro cents (8.2 cents) per kWh costcoverage.20 These figures indicate that CSP willnot become competitive in the BAU scenario.

For the TD scenario, more lines—30 connec-tions until 2050 with a capacity of 4 GW each—need to be built to transmit the 113.8 GW of peakexport CSP generation from the MENA regionin 2050 (Table 10.5). The TD scenario assumesextensive CSP generation in Morocco, Tunisia,and the Middle East. For 2020, the first line isbuilt from Morocco to southern France; interest-ingly, it does not supply the Spanish market, as theIberian electricity market has large shares of lessexpensive CCGT compared with more expensivecoal from Italy and Germany that influences theelectricity price in France.The other two MENAexport nodes,Tunisia and the Middle East, chooseone of the closest demand nodes for the HVDCconnection because of lower line costs.

Although the model is welfare-maximizing,CSP’s profitability is crucial to potential investors.With the assumed price path for gas and coal,electricity price in the targeted sales market isaround 5 euro cents (6.8 cents) per kWh. Consid-ering transmission costs of roughly 1 to 2 eurocents (1.4 to 2.7 cents) per kWh, the amount leftfor generation is far from the break-even point ofabout 9 euro cents (12 cents) per kWh for CSP in

2020. Therefore, the first installations must eitherbe subsidized or be built in the expectation offuture profitability from rising electricity prices. Adecade later, CSP reaches its break-even point. In2030, CSP generators in the three exportingregions can sell electricity at prices of 7 to7.7 euro cents (9.5 to 10.5 cents) per kWh net oftransmission costs. By then, CSP plants built inthe MENA region are assumed to have levelizedgeneration costs of 7 euro cents (9.5 cents) perkWh.21 Continuously rising electricity priceswould lead to even higher profits over the lifetimeof the CSP facilities.Thus a rapid diffusion of CSPstarting in 2030 seems possible.

Under profitable generation conditions forCSP, transmission lines develop according to Fig-ure 10.8 until 2050. However, no transmission isrealized to any node of the reinforced northernEuropean grid. This result suggests that a moreregional integration of CSP generation capacity isthe preferred mechanism. CSP from the exportnode in Morocco is connected to Spain andFrance, Tunisia delivers most of its exported elec-tricity to Italy, and the CSP generation from theMiddle East is connected to the southeastern partof Europe. No lines connect to Germany, theBenelux, or the United Kingdom.22

Stronger Integration of the ScandinavianElectricity Market

The integration of the Scandinavian market isexamined via a comparison of subscenarios A andB in theTD scenario, assuming that Scandinavia isconnected to the United Kingdom and centralEurope by six large-scale HVDC cables. Repeat-ing the welfare optimization yields the followingresults:

Table 10.5. Number of HVDC transmission cables and CSP generation capacity in MENA (TD scenario)

2020 2030 2040 2050

Morocco 1 1.9 GW 2 7.8 GW 5 18.5 GW 9 32.7 GW

Tunisia 1 1.4 GW 4 5.7 GW 4 13.4 GW 6 23.6 GW

Middle East 1 3.5 GW 4 14.4 GW 9 33.9 GW 15 57.5 GW

Source: Egerer et al. 2009, 53

194 Christian von Hirschhausen

• The welfare increase for the exogenousexpansion scenario for HVDC connectionsoutweighs the associated cost. Therefore, anendogenous optimization could be evenmore profitable (this has not been modeledhere because of the increasing complexity).

• Electricity prices are converging: by 2050,prices in Scandinavia increase by 4 euro cents(5.4 cents) per kWh (price increase by 50%)and decrease in the rest of Europe by about1 euro cent (1.4 cents) per kWh.

• Lower prices in the entire European electri-city market (continental Europe) lower theprofits of CSP generation in MENA.

• The welfare optimal expansion paths forHVDC connections from MENA to Europeremain the same until 2050.

On the one hand, a stronger integration of north-ern Europe’s electricity markets has a positiveimpact on the total welfare by balancing windfrom the North Sea with reservoirs in Scandina-via. With the additional integration of the

Scandinavian electricity markets, the annual wel-fare gains increase strongly over time. In 2050,the annual welfare benefit exceeds €3 billion($4 billion), while the levelized annual cost of thenecessary connectors between Scandinavia andcontinental Europe is about half that amount.Therefore, a more detailed analysis of conceptslike the North Sea wind power transmission ringseems reasonable.

On the other hand, changes in welfare areaccompanied by a modified distribution of con-sumer and producer surplus. Thus prices increasein Scandinavia when a higher share of inexpensivehydro can be exported to the United Kingdomand central Europe. Consequently, Scandinavianconsumer surplus decreases while producer sur-plus increases. In the United Kingdom and centralEurope, lower electricity prices have the oppositeeffect. Figure 10.9 presents estimated zone-specific price differences from the reinforcedinterconnections to Scandinavia in 2050. Itbecomes obvious that customers in the Nordic

Wind & Hydro

Morocco

Tunisia

Middle East

Source: Egerer et al 2009, 57Note: Each line has a capacity of 4 GW unless a higher number indicates multiple lines.

Figure 10.8. HVDC grid in 2050 (TD scenario)

Developing a Supergrid 195

countries have good reasons to favor a low inte-gration of the Scandinavian electricity market,despite reduced welfare for the rest of Europe.

The integration of Scandinavia and the lowerprices in central and southern Europe are badnews for a rapid diffusion of CSP and, in fact,could postpone its development for years becauseof the lower prices in the zones connected toMENA by HVDC. It seems probable that an iso-lated analysis of CSP generation without consid-ering electricity system developments, such as theintegration of available potential for wind and bal-ancing hydro capacity, may result in misleadingforecasts. Thus a single focus on L-RES from theMENA region seems inappropriate.

Financial Requirements and Sources

The investment volume for the CSP generationcapacity in the TD scenario is about €390 bil-lion ($531 billion) until 2050.This amount can beallocated to each decade as shown in Figure10.10.23 The total investment of €63 billion ($86billion) for transmission expansion is derived fromthe HVDC lines for CSP integration according toFigure 10.8. The investment projections accord-ing to Figure 10.10 show that the project’s majorcosts are attributed to generation capacity invest-ment. All in all, the share of transmission invest-ment is only about 14% of the total investment

volume. Nevertheless, generation and transmis-sion have to be taken into account equally,because both stages in the value chain are inter-twined: as long as generation capacity has notbeen built up, transmission lines are not necessaryand vice versa. Not only because of this “chickenand egg” problem, it should also be determinedhow best to finance generation and transmissionexpansion.

Several public and private initiatives are underway, such as the Trans-Mediterranean EnergyCooperation/Desertec Initiative, Union for theMediterranean, and bilateral technology agree-ments. However, common institutional and regu-latory platforms among the EEA and MENAcountries need to be further developed to finance

Northern zones

Southern zones

–2 –1 0 1 2 3 4 5euro cents/kWh

Source: Egerer et al. 2009, 60

Figure 10.9. Electricity price changes with reinforced interconnectors to Scandinavia in 2050

200

150

100

50

02020 2030 2040 2050

Transmission investmentCSP investment

[bn ]

Source: Calculations based on Egerer et al. 2009

Figure 10.10. Required generation and transmis-sion investment for the EEA-MENA supergrid

196 Christian von Hirschhausen

such large transnational infrastructure projects.This includes coordinated funding schemes by theEEA member states’ national development banks,multilateral development banks (European Bankfor Reconstruction and Development, AfricanDevelopment Bank), the MENA states in closecooperation with the UNFCCC, the World BankGroup, private investors (private equity funds,venture capital firms, private-public partnershipentities), and other stakeholders. In this context,the importance of a stable regulatory frameworkas a prerequisite for leveraging private capitalshould be kept in mind.

Apart from institutions, funding could be pro-vided for an EEA-MENA supergrid by severalmechanisms:

• Extending transferability of certified emissionreductions (CERs) from clean developmentmechanism (CDM) projects within the EUEmission Trading Scheme (EU-ETS); a lessprobable alternative is the de facto integra-tion of MENA countries into the EU-ETSvia direct substitutability of European emis-sions allowances by CERs from MENAcountries.

• Establishing new, or actively exploiting,existing international technology transferprograms and funds, including the WorldBank’s and multilateral development banks’Climate Investment Funds and Clean Tech-nology Fund, bilateral clean energy and tech-nology funds, and future United NationsFramework Convention on Climate Change(UNFCCC) Multilateral Clean TechnologyFunds as proposed by the G77+ countries(see World Bank 2008; Seligsohn et al. 2009).

• Granting financial guarantees or extendingpolicy instruments for renewable energytechnology deployment (e.g., feed-in tariffs,tax incentives) by EEA countries to energyprojects in the MENA region.

• Setting a generous limit for the upcomingdetermination of allowed renewable energyimports from joint projects between EUmember states and third countries to meetthe national targets on the promotion of

renewable energies according to the recentEU Directive 2009/28/EC (EU 2009a, Arti-cle 9, 2009/28/EC).

Assessment

The EEA-MENA case study provides insightsinto the potential unfolding of a supergrid and therelated technical and economic aspects. Clearly,there may be economic benefits from developinga supergrid, such as the use of less carbon-intensive sources of electricity and higher reliabil-ity within the transmission network. The EEA-MENA case study showed welfare benefits of thegrid development starting in 2030, given assump-tions that do not seem to be outside the realm ofreality. Transmission expansion develops graduallyfrom a nucleus of point-to-point relations andspreads over the region over time. Taken as awhole, the project is profitable, but this ignoresrisk assessment and an analysis of real businesscases. Similar results might be expected from wel-fare assessments or cost–benefit analyses (CBA) ofthe other projects presented in this chapter.

The assumptions of the model and the choiceof technologies and regions considered have a sig-nificant effect on the results, and thus the resultsdiscussed here differ somewhat from thoseobtained in the Desertec study by Trieb et al.(2009). The model of the European electricitysystem applied here allows for a spatial and tem-poral resolution of demand, generation capacities,and electricity prices. In particular, applying amultisource approach (solar, wind, and hydro) in atruly pan-European context yields quite differentresults in terms of network design and prices.Thus the North Sea Energy Super Ring or anyother North Sea energy grid topology (“Seatec”)may become a “competitor” to the solar-foundedDesertec project. The results underline that theproject needs to be seen in the context of othertransmission expansion projects, capacity replace-ments, and evolving generation from L-RES.Compensation mechanisms need to be put inplace as a cooperation scheme between the losers(transit countries) and the winners from lowerelectricity prices. The above discussion hastouched on these institutional obstacles, includingfinancing problems.

Developing a Supergrid 197

One issue that needs further exploration is themultiple incentive structure of the actors involvedin the process. In particular, the North Africancountries may face conflicts of interest in consid-ering the supergrid option. Previous rent earners,such as electricity consumers in the Nordic coun-tries and the natural gas industry, may oppose theproject. The domestic nonenergy industry hasargued for the deployment of energy-intensiveindustries in the MENA region, using electricityfrom L-RES and exporting the products, such as“green aluminium”; this is expected to favor theemergence of new industries and employment.Also, instead of transmitting electricity physicallyto the EEA, natural gas could serve as a “cur-rency” for electricity from L-RES in MENAcountries with natural gas resources, such as Libyaand Algeria. Some actors have indeed argued thatthe EEA countries could pay a fixed price forCSP-generated electricity, which can be con-verted to an amount of gas with equivalent energycontent. Exporting gas through existing pipelinesand via LNG terminals might avoid the need forextensive electricity transmission expansion. Thusit is not clear whether the supergrid idea is thepreferred solution of all players involved.

Challenges for DevelopingSupergridsWhat are some general challenges and problemsrelated to the deployment of supergrids? In par-ticular, recent difficulties faced by transmissionexpansion projects on both sides of the Atlanticshow that a cost–benefit ratio below one or posi-tive welfare effects do not suffice to makesupergrids “fly.” Instead, questions regarding theinstitutional structure of a further supergrid archi-tecture abound. Many of them touch on the issueof institutional design, or what 2009 Nobel Prizewinner Oliver Williamson calls the choicebetween “markets and hierarchies,” referring towhich element of the value-added chain can bedeveloped with the use of market signals, and forwhich a more hierarchical or planning approach ismore appropriate. In fact, most institutions used

in transmission expansion might combine ele-ments of both markets and hierarchies, becominghybrid organizational forms.

Planning Issues

Any supergrid requires long-term planning tocreate the conditions under which large-scaletransmission investment can unfold. Transmissionplanning arrangements include, among others,the planning process itself, the implementationand mechanisms for cost recovery, the role ofmarkets (price signals) as a decision support, andtrade-offs between transmission and non-transmission investment (Moselle and Brown2007). Note that the supergrid planning processstretches well beyond conventional transmissionplanning procedures. Time horizons projectedforward are much longer, and system changesmust be considered. Moreover, the maximumspan of demand and generation forecasts used intraditional national planning procedures is 7 to 10years (e.g., in the United Kingdom), with themaximum of (less-detailed) forecasts of 20 years inAlberta (Moselle and Brown 2007).This is signifi-cantly shorter than the three- to four-decadehorizon adopted by most supergrids.

Clearly, the institutional framework for plan-ning and developing the supergrid is not yet avail-able in the United States and Europe.The UnitedStates is more advanced with respect tointerregional transmission development. How-ever, the Federal Energy Regulatory Commission(FERC) cannot impose the development of asupergrid on any of the regional transmissionorganizations (RTOs) or the individual independ-ent system operators (ISOs). Neither can it, forthe time being, consider RPS obligations as legiti-mate criteria for the development of new trans-mission (Krapels 2009).

In Europe, there is still less ground on whichthe planning process for a supergrid could bebased. Contrary to many restructured electricitysystems, continental Europe lacks a regional trans-mission development plan (Moselle and Brown2007). Transnational transmission expansion isonly beginning to be coordinated at the Europeanlevel. An emerging European regulatory agency,

198 Christian von Hirschhausen

the Agency for the Coordination of EnergyRegulators (ACER), is set to negotiate planningideas with its counterpart, the European Networkof Transmission System Operators for Electricity(ENTSO-E).24 Given the lack of institutionalsupport, it seems that without federal orEuropean-wide planning procedures and someimplementing power, the development of anysupergrid is unlikely. Whether any such planningprocesses can reasonably be implemented is yetanother open question.

Cost Allocation

Cost allocation and cost recovery are crucial inany transmission regime, but they become par-ticularly difficult in the supergrid context, wheremore than one ISO orTSO is involved. A restruc-tured market with open access requires some formof cost allocation mechanism for transmissionexpansion. Baldick et al. (2007) identify principlesto follow: they propose using a standard measure-ment for societal benefits that incorporates eco-nomic values and having in place an open, trans-parent, and inclusive process for regionally basedtransmission planning and analysis. New transmis-sion investment should be supported in federal orother wholesale rates, instead of in the retail ratebase, and free entry should be permitted in trans-mission investment.

There are two main approaches to allocatecosts. First, those who clearly benefit from theinvestment are identified and consequently pay forit; note that the question of who benefits is nottrivial either, particularly in a dynamic contextwith a changing network topology. Second, thecost of new transmission investment is socializedamong all users in a region or market (Baldick etal. 2007). Another key issue of cost allocation isthe ability of the institutional system to compen-sate parties who “lose” from transmission invest-ment. Besides the question of cost allocation,other issues, such as consumer concerns aboutprice development, can arise. Namely, partiesprotecting consumers in regions with low-costgeneration may fear that transmission investmentwill cause generation-related prices to equalizeover a larger geographic region and thus increasefor a specific region.25

Vertical or horizontal coordination, or both,would make cost allocation for a supergrid evenmore complex. In the United States, the dualpricing regime, a structure of transmission costrecovery that is affected by both state and federalratemaking practices, complicates the appropriatecost allocation. Currently, FERC, the federalregulator, must consider both the allocation ofcosts among different generators and load, and thereflection of those costs in retail rates. Obviously,this dual-pricing regime has a rather complicatingeffect. Europe has been slow to develop an inter-TSO compensation mechanism (Moselle 2008),which complicates any supergrid development. Inaddition, the atomization of European TSOsmakes successful cooperation more difficult. Thethree Scandinavian system operators have devel-oped a voluntary compensation mechanism(Nylund 2009). But even this relatively smalljointly managed grid illustrates the weakness ofvoluntary approaches to inter-TSO cooperationin Europe and seems to suffer from the absence oflegally binding arrangements. As the number ofparticipating TSOs increases, transaction costsincrease exponentially.

Market Design

Both gradual transmission expansion andsupergrids should deploy a market-oriented mar-ket design that is conducive to socially optionalinvestment decisions. A critical element thereof isbid-based, security-constrained economic dis-patch with locational marginal pricing (LMP), ornodal pricing, as proposed by Hogan (2008), toachieve welfare maximization and identify poten-tial transmission bottlenecks. In many instances,LMP is the most cost-efficient solution to trans-mission problems because the prices identify thetrue scarcities in the system. Thus longer-termplanning for a supergrid should incorporate thiswelfare-maximizing algorithm. In the UnitedStates, these principles are fulfilled in the morereform-oriented regions, most of which are alsoelaborating on supergrid proposals. Note thatwhile LMPs are useful in providing locational sig-nals for placing generation and load, their valuefor establishing transmission investments is more

Developing a Supergrid 199

doubtful; therefore, financial transmission rights(FTRs) based on LMP differences have not fullyresolved the issue of transmission investment, aswas initially hoped for.

In Europe, no electricity system has a nodalpricing system in place, not even the UnitedKingdom, which applies static G-components,with generators compensated for relieving net-work congestion or penalized for contributing toit. Continental Europe is still largely dominatedby uniform transmission pricing. Hence the con-ditions under which efficient supergrid invest-ment could occur are not really in place.

A critical issue in any intermittent-based sys-tem is the design of the balancing market andcreating incentives for investment in backupcapacity for intermittent generation fromrenewables. Most of the scenarios discussed earlyin this chapter have not gone the full way ofinstalling storage capacity for the intermittentlarge-scale renewables (wind and solar), primarilybecause of cost. In this case, the generators them-selves will need to analyze other technologies tocope with the intermittency of the renewablesportfolio.Thanks to their high ramping flexibility,relatively low capital costs, widely available fuel,and ease of operation, gas turbines are likely to bethe most economic backup technology.

Cost–Benefit Analyses

Even if a workable long-term planning process isin place, with cost recovery and price signalsdefined, an institutional design regarding theassessment of transmission expansion would berequired. Long-term planning can define the “bigpicture” of generation and transmission expan-sion, but it cannot provide an assessment ofwhether an individual line, or a certain part of thenetwork, should or should not be extended. Forthis, some form of cost–benefit comparison mightbe useful. As a rule of thumb, three approachescan be applied:

• Traditional net present value (NPV) assess-ment of individual transmission investments,by relating (expected) private revenues tocosts; this business perspective adheres to the

logic of an investor but ignores larger eco-nomic benefits and costs.

• Calculating cost changes by producers andconsumers as a result of the expansion oftransmission lines; this corresponds to a moreeconomic approach, as it approximates thewelfare effects of investments by the changein producer and consumer surpluses. Forexample, PJM values the economic benefitof additional transmission investment by0.7 Δ PC + 0.3 Δ LP, where Δ PC is thechange of production costs and Δ LP is thechange in load payments (Lin 2009).

• A comprehensive, sophisticated cost–benefitanalysis that also takes into account indirecteffects of transmission expansion, such asincreased system reliability or competition inone of the zones. For example, the CaliforniaISO (CAISO) uses a comprehensive trans-mission economic assessment methodology(TEAM) as a decision support tool for trans-mission planners, in which the annual ben-efits from an expansion include productioncost benefits, competitivity benefits, opera-tional benefits, generation investment costsavings, reduced losses, and emission benefits(California ISO 2004).26

All substantial public policy interventions of thetype required for the supergrid projects discussedrequire justification, and a sound, transparentcost–benefit analysis provides just that. Thus theCAISO-TEAM methodology might be devel-oped further to become applicable to longer-termsupergrid developments as well.This CBA frame-work could then be applied to the projects underdiscussion. This is particularly important forEurope, where the 10-year EU-wide plans are notyet systematically subject to cost–benefit analysis.The normative aspects (yes or no) of such analysishave a strong positive effect, increasing transpar-ency and accountability of the actors involved.

Regulatory Options

While the “optimal” regulation of transmissionexpansion is already a controversial issue, this iseven more the case for the regulatory framework

200 Christian von Hirschhausen

of any supergrid. The following issues merit con-sideration and await solutions:

• Regulated versus merchant transmissioninvestment: in general, the more meshed atransmission network, the more difficult it isto isolate the effects of a specific newly con-structed line, and hence the more difficult itis to apply merchant investment. Krapels(2009) suggests that HVDC lines are able toattract private merchant investment becauseof the point-to-point character of these lines.The empirical underpinning related theretois scarce, however: except for a few localtransmission lines, there is little independenttransmission expansion in the United Statesand Europe. It seems that the need for coord-inated, long-term planning and inter-TSOcoordination, information asymmetriesbetween the incumbent network owner/operator and the market entrant, and scaleand scope effects are arguments in favor ofregulated investment by the incumbent ISOs,RTOs, or TSOs. This does not exclude ten-dering of a large share of the activities thatwill be subject to market forces.

• Incentive versus cost-based regulation: opt-ing for a regulatory approach necessitatesdetermining the “power” of the incentivescheme, in other words the degree to whichthe network tariffs are fixed by the regulator,and thus does not correspond to someapproximation of costs (cost-based regula-tion). Whereas incentive regulation hasresulted in increased static allocative effi-ciency, its benefits in terms of dynamicinvestment efficiency are unclear. Although ageneral consensus exists that an incentive-based regulatory regime can theoreticallyinduce an “optimal” level of transmissionexpansion under certain conditions, the real-world implementation of these mechanismshas not followed suit. The Hogan-Rosellon-Vogelsang mechanism (Hogan et al. 2007) isan example of such a traditional, non-Bayesian incentive scheme. However, itassumes constant demand and has difficultiesaccommodating intermittent renewable elec-

tricity sources. Finally, the danger of regula-tory discretion and limited commitment, andthe potential instability of the political andinstitutional framework, would suggest areduction of the incentive power of the regu-latory regime for a supergrid.

Political Economy Arguments: AchievingConsensus and Compensating Losers

The case studies examined earlier suggest thattechnology per se is not a real obstacle. Also, a lackof long-term planning might be overcome at theoperational level, albeit at higher transaction costsand potential losses in efficiency. The real chal-lenge for the implementation of a supergrid, orany other megaproject of that type, is how toreach consensus among a very large group ofstakeholders, in this case generators, system own-ers, TSOs, consumers, and the decisionmakingentities at all levels of government.

In some settings, one can use force to obtainconsensus, such as in the case of large-scale infra-structure projects in the former Soviet Union.Another way of pushing megaprojects through isto dedicate unlimited financial resources, as in theU.S. space race in the 1960s. Closer to our field ofinfrastructure development, the top-downapproach by which the United States imple-mented its interstate highway program beginningin the 1950s can also be considered a historicexample, whatever the driving force behind it(e.g., transport or national defense) might havebeen. These examples do not directly apply tosupergrids, however, as centralized, top-downinterventionism is not an option in a democracy,financial means are very limited, and siting andimplementation of large-scale infrastructures aremuch more difficult today than 60 years ago.

Instead, the key to implementing such large-scale projects is to create a consensus, in particularabout the sharing of the gains and the compensa-tion of losers by the winners. It is not yet clearhow the welfare gains that any supergrid is likelyto achieve can be divided among the participatingparties to make the project become a reality. Atleast three levels need to be considered:

Developing a Supergrid 201

• On the producer side, and even assuming thattransmission can be treated as a cost block,allocating producer rent among the playersdepends on the behavioral assumptions—inother words, whether one assumes coopera-tive or noncooperative behavior—and theallocation of grandfathered rights. In addi-tion, established incumbent electricity pro-ducers may hesitate to invest in low-costcapacity abroad that might undermine theprice in their home markets as well (cannibal-izing effect).

• On the consumer side, there will be clearwinners and losers from the additional gen-eration and transmission capacities. As thecase study demonstrates, former locked-inlow-price consumers tend to be adverselyaffected, while consumers in former high-price zones clearly benefit.

• Last but not least, the sharing of rents amongdifferent federal levels is a potential source ofconflict as well. Welfare maximization overthe entire supergrid region is, in fact, anoversimplification, because U.S. states andEuropean nations are likely to influence deci-sions and seek to optimize only their ownwelfare. As in the planning process, thiswould speak in favor of centralization ofdecisionmaking with flexible side paymentsand compensation.

ConclusionsThe main message of this chapter is that harness-ing renewable energy to generate electricity isrelatively simple in comparison with transportingthat electricity over long distances to large-scaledemand centers. The idea of developing large-scale supergrids, or transmission overlay grids,may be appealing, judging by the mushroomingnumber of pilot studies. However, large-scale har-nessing of renewables may be limited by transmis-sion bottlenecks that will be difficult to overcomein the current market and institutional environ-ment:

• Few of the reviewed case studies deal withtransmission issues in a serious way, and those

that do tend to prefer decentralized, regionalnetwork integration to supergrids.

• The case study in this chapter, the EEA-MENA project, indicates that large-scaleCSP generation in MENA countries and itsHVDC transmission to Europe may increasewelfare and even become profitable undercertain conditions. However, a pan-European perspective is required, and it iscurrently unclear whether the institutionalarrangements necessary for the project,including adequate financing mechanisms,will be put in place in time.

• All supergrid projects also face several cross-cutting issues: the necessity of efficient long-term planning procedures, the simplificationof cost allocation issues, market-orientedpricing where possible, a comprehensivecost–benefit analysis to identify economicsolutions, and adequate regulatory incentives.

Potential investors face a multitude of market,policy, and regulatory risks; technological uncer-tainties; coordination problems; and other barri-ers, such as local resistance. The current lack ofbusiness cases for network expansion can be over-come only if a clear, stable, and equitable long-term policy framework is in place. Welfare ben-efits of network expansion projects should be dis-tributed in an incentive-compatible fashion.Hence an institutional analysis is required todetermine opportunities and challenges in thedevelopment of such scenarios.

Several issues that should be addressed in moredetail include timing, irreversibility of investment,and risk considerations. This chapter has onlytouched on issues of the appropriate regulatorysetting, the most critical being today’s policy offavoring national subsidies over federal regulation.On both sides of the Atlantic, transmission plan-ning competence and the ability to develop andinstall supergrids are limited. The most importantissues to resolve are those linked to rent sharing.Thus future research should investigate the distri-butional effects of different scenarios on producerand consumer rents in each of the regions con-cerned, as well as ponder the proper compensa-

202 Christian von Hirschhausen

tion mechanisms to overcome local resistance towelfare-improving projects.

AcknowledgmentsThis chapter is based on the research programElectricity Markets at the Dresden University ofTechnology (TU Dresden). The author thanksJonas Egerer and the study project 2050 for back-ground research, and Florian Leuthold, ChristinaBeestermöller, Johannes Herold, Robert Wand,and Hannes Weigt for comments and suggestions;thanks also to two referees for constructive com-ments.

Notes

1. According to IEEE (1997), FACTS are defined as“alternating current transmission systems incorpo-rating power electronic-based and other staticcontrollers to enhance controllability and increasepower transfer capability.” FACTS comprise anumber of technologies, which are used to stabi-lize the frequency, control the real and reactivepower flow, or compensate and manage the effec-tive power flow. Improving the static stability of amultiline transmission system, FACTS canincrease the power flow on existing AC lines.

2. HVDC can be connected to AC networks via aconverter station, using conventional line-commutated current source converters (CSCs), orself-commutated voltage source converters(VSCs). The latter are based on insulated gatebipolar transistors (IGBT) and can be adjustedwith respect to load flows. These components alsooffer black start capacity. Whereas traditionalHVDC was mainly limited to point-to-pointoperation,VSCs enable HVDC systems to work inmore complex structures. This development offersthe possibility to create entire HVDC transmissionoverlays as a supergrid.

3. The Solar Grand Plan even stretches to the end ofthis century: by 2100, the L-RES could generate100% of U.S. electricity and more than 90% oftotal U.S. energy (Zweibel et al. 2008).

4. “This backbone would reach major markets inPhoenix, Las Vegas, Los Angeles and San Diego tothe west, and San Antonio, Dallas, Houston, New

Orleans, Birmingham (Ala.), Tampa (Fla.), andAtlanta to the east” (Zweibel et al. 2008, 54).

5. For an average transmission distance of 1,500miles, one converter per line, and a capacity utili-zation factor of 27%, Fthenakis et al. (2009) esti-mate a levelized transmission cost of $0.024/kWhon average. Investment and operation costs fortransmission will be included in grid developers’royalties.

6. For a U.S.-wide study on large-scale wind integra-tion, see the transmission plan by American Elec-tric Power (AEP 2007) at the request of theAmerican Wind Energy Association, suggesting a765 kV AC overlay grid.

7. The Midwest and the East Coast are connected by765 kV AC lines and two 800 kV DC lines, withmajor reinforcements between Massachusetts andNew Jersey.The two HVDC lines are fed by addi-tional 400 and 800 kV DC lines. Moreover, two-circuit 500 kV AC lines build up the existing gridin Oklahoma and connect Arkansas and Alabamaas well as Maryland and NewYork. A new 345 kVtwo-circuit line will be built in NewYork.

8. This is also reflected in the introduction ofincentive-based rate treatments for transmission ininterstate commerce by FERC as a measure tostimulate interstate transmission expansion(Energy Policy Act 2005, Section 1241, andamendment to the Federal Power Act, Section219).

9. For more information on SUSPLAN (PLANningfor SUStainability), see www.susplan.eu/; for theSOLID-DER project (Coordination Action toConsolidate RTD Activities for Large-Scale Inte-gration of Renewable Energy into the EuropeanElectricity Market), see www.solid-der.org/; forWIND FORCE 12 by Greenpeace, seewww.ewea.org/fileadmin/ewea_documents/documents/publications/reports/wf12–2005.pdf;for Tradewind, see www.trade-wind.eu/; and forDENA 2, see www.dena.de/de/themen/thema-reg/projekte/projekt/netzstudie-ii.

10. The initial project is supposed to be connected tothe United Kingdom, the Netherlands, and Ger-many.

11. The overall objective of the Desertec Foundationis to make productive use of the largest source ofenergy, solar, as a basis for electricity consumptionand sustainable development around the globe,with a focus on desert countries. Besides theenergy problem, the Desertec consortium alsoaddresses issues of water supply, food security, andclimate change.

Developing a Supergrid 203

12. Since the summer of 2009, the project has beenaccelerated by the Desertec Industrial Initiative(DII), a consortium of utilities, equipment pro-ducers, banks, and insurance companies. At theEU level, there is also activity in the form of theUnion for the Mediterranean, which has devel-oped the Mediterranean Solar Plan (MSP). TheMSP is organized in three phases: building aninstitutional structure (ongoing); developing aMaster Plan Study (until 2010); and implementingthe plan until 2020. The following discussionrefers to the most recent study at the time of thiswriting, Trieb et al. (2009).

13. Being highly site-dependent, this calculation isbased on a solar direct normal irradiance (DNI) of2,700 kWh/m²/a and a CSP design for extensiveheat storage (solar multiple of 4).

14. For technical calculations on the EEA-MENAproject, we refer to an ongoing research project atthe Chair of Energy Economics and Public SectorManagement at Dresden University ofTechnology(see Egerer et al. 2009).

15. This includes the high-priority “projects of Euro-pean interest for inter-European electricity inter-connection, and linkage with neighboringregions”: according to theTrans-European EnergyProjects (TEN-E) and Priority InterconnectionPlan (PIP) (EC, 2006).

16. Compare Leuthold et al. (2008) and Egerer et al.(2009) for a detailed technical discussion.

17. For a detailed description, see Egerer et al. (2009,32).

18. For the spatial configuration of the additionallines, see Egerer et al. (2009, 46).

19. The average electricity price (system price) is cal-culated from the model’s demand-weighted zonalprices.

20. Cost coverage is calculated by subtracting trans-mission costs from electricity prices in the exportelectricity market.

21. For a detailed overview of prices, transmissioncosts, and cost coverage for generation, see Egereret al. (2009, 56).

22. In addition to zones with large wind capacity,zones with large hydropower capacity (e.g., Swit-zerland, Austria, and Serbia) are not served byCSP-generated electricity. Although these zoneshave seasonal storage capacity (reservoirs), for wel-fare purposes it seems preferable to substitutefossil-fueled base load capacity elsewhere.

23. These cost estimates are based on decreasing aver-age investment cost for CSP from €5,300 ($7,220)

per kW in 2020 to €3,944 ($5,373) per kW in2030 to €3,429 ($4,671) per kW in 2050 (Trieb etal. 2009, 83).

24. As a first step, Regulation 714/2009/EC (EU2009b) requires that ENTSO-E adopt a nonbind-ing communitywide 10-year network develop-ment plan every 2 years, which is then assessed bythe agency (ACER), consistent with the national10-year development plans provided by the mem-ber states.

25. Beyond the underlying principles, a concrete costallocation is the method of average participation(AP) (Pérez-Arriaga and Olmos 2009). AP cantrack the actual upstream and downstream flows tothe generators, and loads that can be associatedplausibly with them. Theoretically, the cost of the“used” fraction of each line can be apportioned togeneration and load in proportion to the aggregateeconomic benefits.

26. An essential component of the CAISO-TEAM isthat it implements a market simulation modelbased on dynamic supply bids and incorporating adetailed physical transmission modeling capabilityfor a reliability region. Besides, CAISO-TEAMincludes uncertainty and risks about the futurethat can partly be quantified. Apart from thesebenefits, cross-sectoral positive externalities com-prise simplified rights-of-way for the use of othernetwork infrastructure being built, once the land isassigned to transmission purposes (such as opticalfiber telecommunications), reflecting a positiveexternality of transmission expansion. Addition-ally, long-term resource cost advantages, synergieswith other transmission projects, fiscal benefitsfrom construction and taxes, and impacts on fuelmarkets should be taken into considerationaccording to Pfeifenberger and Newell (2007).

ReferencesAEP (American Electric Power). 2007. InterstateTrans-

mission Vision for Wind Integration.www.aep.com/about/i765project/docs/WindTransmissionVisionWhitePaper.pdf (accessedDecember 3, 2009).

Airtricity. 2006. European Offshore Supergrid Pro-posal: Creating a More Powerful Europe. www.trec-uk.org.uk/resources/airtricity_supergrid_V1.4.pdf(accessed December 2, 2009).

AWEA/SEIA (American Wind Energy Association/Solar Energy Industries Association). 2009. Green

204 Christian von Hirschhausen

Power Superhighways: Building a Path to America’sClean Energy Future. www.awea.org/GreenPowerSuperhighways.pdf (accessed December3, 2009).

Baldick, Ross, Ashley Brown, James Bushnell, SusanTierney, and Terry Winter. 2007. A National Perspec-tive on Allocating the Costs of New Transmission Invest-ment: Practice and Principles. Blue Ribbon Panel onCost Allocation. www.puc.nh.gov/Transmission%20Commission/120108%20Progress%20Report/Attachment%20M.pdf (accessed March 18, 2010).

California ISO (California Independent System Opera-tor). 2004. Transmission Economic AssessmentMethodology (TEAM). Folsom, CA: CaliforniaISO.

Claus, M., D. Retzmann, S. Sörangr, and K. Uecker.2008. Solutions for Smart and Super Grids withHVDC and FACTS. 17th Conference of the Elec-tric Power Supply Industry. October 2008, Macau.www.ptd.siemens.de/CEPSI08_Art.pdf (accessedDecember 1, 2009).

EC (European Commission). 2006. Decision No.1364/2006/EC of the European Parliament and ofthe Council of 6 September 2006 Laying DownGuidelines for Trans-European Energy Networksand Repealing Decision 96/391/EC and DecisionNo. 1229/2003/EC. Official Journal of the EuropeanUnion: L262/1-23.

Egerer, Jonas, Lucas Bückers, Gregor Drondorf,Clemens Gerbaulet, Paul Hörnicke, RüdigerSäurich, Claudia Schmidt, Simon Schumann, Sebas-tian Schwiedersky, Thorsten Spitzel, and AnjaThanheiser. 2009. Sustainable Energy Networks forEurope: The Integration of Large-Scale RenewableEnergy Sources until 2050. www.tu-dresden.de/wwbwleeg/publications/wp_em_35_Egerer_et_al_2050.pdf (accessed March1, 2010).

ENTSO-E (European Network ofTransmission SystemOperators for Electricity). 2008. ENTSO-EGridMap. www.entsoe.eu/index.php?id=77(accessed November 10, 2009).

———. 2009. UCTE Transmission Development Plan2009. www.entsoe.eu/fileadmin/user_upload/_library/publications/ce/otherreports/tdp09_reporr_ucte.pdf (accessed November 10,2009).

EU (European Union). 2009a. Directive 2009/28/ECof the European Parliament and of the Council of23 April 2009 on the Promotion of the Use ofEnergy from Renewable Sources and Amending and

Subsequently Repealing Directives 2001/77/ECand 2003/30/EC. Official Journal of the EuropeanUnion L140/16–62.

———. 2009b. Regulation No. 714/2009 of theEuropean Parliament and of the Council of 13 July2009, on Conditions for Access to the Network forCross-Border Exchanges in Electricity and Repeal-ing Regulation (EC) No 1228/2003. Official Journalof the European Union: L211/15–35.

Fthenakis,Vasili, James Mason, and Ken Zweibel. 2009.The Technical, Geographical, and Economic Feasi-bility for Solar Energy to Supply the Energy Needsof the US. Energy Policy 37 (2): 387–399.

Hogan, William. 2008. Electricity Market Design:Coordination, Pricing and Incentives. Presentationat the Toulouse Conference on Energy Economics.www.energypolicyblog.com/wp-content/uploads/2008/06/20080623_hogan.pdf (accessed December8, 2009).

Hogan, William, Juan Rosellon, and Ingo Vogelsang.2007. Toward a Combined Merchant-RegulatoryMechanism for Electricity Transmission Expansion.Paper presented at the IAEE European Conference.April 2007, Florence, Italy.

IEEE (Institute of Electrical and Electronics Engineers).1997. Proposed Terms and Conditions for FACTS.IEEE Transactions on Power Delivery 12 (4): 1848–1853.

Jacobsen, Mark Z., and Mark A. Delucchi. 2009. APath to Sustainable Energy by 2030. Scientific Ameri-can 299 (November): 58–65.

Kaupa, Heinz. 2009. Smart Grids and/or Super Grid?Presentation at the Alpbacher Technologiegesprächeconference. August 2009,Vienna. www.bmvit.gv.at/service/publikationen/innovation/downloads/3kaupa.pdf (accessed December 3, 2009).

Krapels, Edward. 2009. Integrating 200,000 MWs ofRenewable Energy in the US Power Grid.www.hks.harvard.edu/hepg (accessed November 3,2009).

Leuthold, Florian, Hannes Weigt, and Christian vonHirschhausen. 2008. ELMOD – A Model of theEuropean Electricity Market. Electricity Marketsworking papers WP-EM-00. www.tu-dresden.de/wwbwleeg/publications/wp_em_00_ELMOD.pdf(accessed November 13, 2009).

Lin, Jeremy. 2009. Market-BasedTransmission PlanningModel in PJM Electricity Market. Energy Markets2009. Paper presented at the EEM 2009 (6th Con-ference on European Electricity Markets). May2009, Leuven, Belgium.

Developing a Supergrid 205

Midwest ISO, PJM, SPP, TVA, and MAPP. 2008. JointCoordinated System Plan 2008: Economic Assess-ment. Report, Vol. 1. www.jcspstudy.org (accessedDecember 7, 2009).

Moselle, Boaz. 2008. Reforming TSOs: Using the“Third Package” Legislation to Promote Efficiencyand Accelerate Regional Integration in EU Whole-sale Power Markets. Electricity Journal 21 (8): 9–17.

Moselle, Boaz, and Toby Brown. 2007. InternationalReview ofTransmission Planning Arrangements. Brussels:Brattle Group.

Natuur en Milieu (Netherlands Society for Nature andEnvironment). 2009. Office for Metropolitan Archi-tecture Presents Master Plan Offshore Wind in theNorth Sea. Utrecht, Netherlands: Natuur en Milieu.

Nylund, H. 2009. Sharing the Costs of TransmissionExpansion: A Cooperative Game Theory ApproachApplied on the Nordic Electricity Market. Proceed-ings of the IAEE 10th European Conference. Sep-tember 2009, Vienna, Austria.

OMA (Office for Metropolitan Architecture). 2009.Masterplan Zeekracht. www.zeekracht.nl/sites/default/files/oma.pdf (accessed November 29,2009).

Pérez-Arriaga, Ignacio, and Luis Olmos. 2009. A Com-prehensive Approach for Computation and Implementationof Efficient Electricity Transmission Network Charges.Cambridge, MA: Center for Energy and Environ-mental Policy Research.

Pfeifenberger, Johannes, and Sam Newell. 2007. Evalu-ating the Economic Benefits of Transmission Invest-ments. Paper presented at EUCI’s Cost-EffectiveTransmission Technology Conference. May 2007,Nashville, TN.

Seligsohn, Deborah, Lutz Weischer, Shane Tomlinson,and Pelin Zorlu. 2009. Key Functions for aUNFCCCTechnology Institutional Structure: Iden-tifying Convergence in Country Submissions.Working paper. Washington, DC: World ResourcesInstitute. www.wri.org/climate/cop-15 (accessedDecember 3, 2009).

Siemens. 2009. Ultra HVDC Transmission System,Newsletter Issue 0910. www.trec-uk.org.uk/resources/siemens_uhvdc_october_2009.pdf(accessed December 1, 2009).

Sinden, Graham. 2007. Characteristics of the UK WindResource: Long-Term Patterns and Relationship toElectricity Demand. Energy Policy 35 (1): 112–127.

Tierney, Susan. 2008. A 21st Century “Interstate ElectricHighway System”: Connecting Consumers and DomesticClean Power Supplies. Boston: Analysis Group.

Trieb, Franz, Christoph Schillings, Stefan Kronshage,Uwe Klann, Peter Viebahn, Nadine May, ReginaWilde, Christian Paul, Malek Kabariti, AbdelazizBennouna, Hani El Nokraschy, Samir Hassan, LailaGeorgy Yussef, Tewfik Hasni, Nasir El Bassam, andHonorat Satoguina. 2005. Concentrating SolarPower for the Mediterranean Region. Final Report.Stuttgart: German Aerospace Center (DLR) Insti-tute of Technical Thermodynamics Section SystemsAnalysis and Technology Assessment.

Trieb, Franz, Christoph Schillings, Stefan Kronshage,Peter Viebahn, Nadine May, Christian Paul, MalekKabariti, Khaled M. Daoud, Abdelaziz Bennouna,Hani El Nokraschy, Samir Hassan, Laila GeorgyYussef, Tewfik Hasni, Nasir El Bassam, and HonoratSatoguina. 2006. Trans-Mediterranean Interconnec-tion for Concentrating Solar Power. Final Report.German Aerospace Center (DLR) Institute of Tech-nical Thermodynamics Section Systems Analysis andTechnology Assessment, Stuttgart. www.dlr.de/tt/Portaldata/41/Resources/dokumente/institut/system/projects/TRANS-CSP_Full_Report_Final.pdf (accessed December 1,2009).

Trieb, Franz, Marlene O’Sullivan, Thomas Pregger,Christoph Schillings, and Wolfram Krewitt. 2009.Characterisation of Solar Electricity Import Corri-dors from MENA to Europe. Stuttgart: REACCESSReport, German Aerospace Center (DLR) Instituteof Technical Thermodynamics Section SystemsAnalysis and Technology Assessment.www.dlr.de/tt/Portaldata/41/Resources/dokumente/institut/system/publications/Solar_imimpo_DLR_2009_07.pdf(accessed December 1, 2009).

World Bank. 2008. Illustrative Investment Programs forthe Clean Technology Fund.http://siteresources.worldbank.org/INTCC/Resources/Illustrative_Investment_program_May_15_2008.pdf (accessed November 26, 2009).

Zweibel, Ken, James Mason, and Vasilis Fthenakis.2008. Solar Grand Plan. Scientific American 298 (1):48–57.

206 Christian von Hirschhausen

Part IV

National Experiences

11

Renewable Electricity Generationin the United StatesRichard Schmalensee

Thomas A. Edison’s Pearl Street Station inNew York, the first permanent, commercial

electric generating plant, began operation on Sep-tember 4, 1882 (IEEE 2008b). Just 26 days later,the first commercial generating plant usingrenewable energy—a hydroelectric facility—began operation in Appleton, Wisconsin (IEEE2008a).The United States has considerable hydro-electric potential and moved aggressively, particu-larly in the 1930s, to exploit it. By 1949,hydropower accounted for just under a third ofU.S. electricity generation (EIA 2009b, Table 1.1).

Since then, however, the relative importanceof hydropower has waned, as potential dam siteswere of lower quality than those alreadyemployed, the performance of other generatingtechnologies improved, and the public becameincreasingly concerned about the environmentalimpacts of dams. In recent years, more attentionhas been given to the possible demolition ofhydroelectric dams than to their possible con-struction. Hydropower accounted for only about6% of U.S. electricity generation in 2007 (EIA2009b, Table 1.1).

Renewable generation technologies otherthan hydroelectricity, referred to in this chapter asnonhydro renewable (NHR) technologies, beganto attract significant attention from public andprivate decisionmakers in the United States and

abroad after the energy crises of the 1970s. Asenvironmental concerns, particularly those relatedto climate change, have become more important,support for these technologies has generallyincreased. In the United States the result has beena complicated saga of erratic and unfocused fed-eral policy and widely divergent state policies,with results that have not surprisingly varied con-siderably over time and among the states.

The chapter begins with a brief quantitativeoverview of the actual and potential importanceof nonhydro renewable energy over time in theUnited States, until recently a leader in NHRgeneration of electricity. Lately, other nationshave provided more effective support of thesetechnologies and accordingly have taken the leadin using them. Next, the chapter outlinesrationales and policy tools for supporting NHRsand examines policy at the federal level in theUnited States. It then considers state-level policiesand their effects, with brief discussions of experi-ences in two major states that have played verydifferent leadership roles in this area: Californiaand Texas. This is followed by a look at the mostrapidly growing NHR technology in the UnitedStates—wind—and some of the issues and con-cerns its growth has raised.

Nonhydro Renewablesin the United StatesBetween 1949 and 2008, both total U.S. energyconsumption and consumption derived fromNHRs grew at about 1.95% annually on average.In the first half of this period, from 1949 to 1978,total energy consumption grew at a 3.21% averageannual rate, while energy from NHRs grew onlyabout a third as fast—at a 1.06% annual rate.Thereafter, the growth of total energy consump-tion slowed dramatically to a 0.72% annual rate,while the growth of energy from NHRs acceler-ated to a 2.82% annual rate. Despite this impres-sive growth, however, NHRs have neveraccounted for more than 4.5% of total U.S.energy consumption (EIA 2009a, Table 1.3).

Figure 11.1 gives a breakdown of total energyconsumption from nonhydro renewables over the1978–2008 period by source. In the early years,the only important source in this category wasbiomass, mainly wood and wood waste, used togenerate heat rather than electricity. In recentyears, biofuels, chiefly ethanol, have become ofcomparable importance. Together with a small

contribution from what is termed “other solar”—the use of solar energy to produce heat, mainly towarm swimming pools—these three nonelectricuses of renewable energy are much more impor-tant than the use of NHRs to generate electricity.

Since the late 1970s, NHRs have been ofinterest to policymakers primarily because of theirperceived potential to displace fossil fuels (and insome jurisdictions, nuclear energy) in electricitygeneration. Despite this interest, however, and awide variety of policies aimed at encouraging theuse of NHRs, these technologies have accountedfor only 2% to 2.5% of total U.S. electricity gen-eration since 1989, as Figure 11.2 shows. For all ofthe 1990s, NHRs played a more important role ingenerating electricity in the United States than inEurope. But major European nations, particularlyGermany, were much more aggressive in promot-ing NHRs over most of this period, and the shareof these technologies in European electricity gen-eration has accordingly been rising. It is nowalmost double that in the United States.

Figure 11.3 shows the contributions of thevarious NHR technologies to electricity genera-tion since 1990. About 70% of biomass generation

6.0%

5.0%

4.0%

3.0%

Per

cent

age

Year

2.0%

1.0%

0.0%

1978

1979

1980

1981

1982

1983

1984

1985

1986

1987

1988

1989

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

Geothermal (electric) Wind (electric) Biomass (electric)Biofuels Other biomass Solar and other

Source: EIA 2009a

Figure 11.1. Nonhydro renewable energy consumption as a percentage of total U.S. energy consumption,1978–2008

210 Richard Schmalensee

is fueled by wood and wood waste; the remainderis fueled by biogenic municipal solid waste,landfill gas, and a variety of other substances.Between 1990 and 2007, geothermal generationdeclined slightly, and biomass-fueled generationgrew at only a 1.1% average annual rate. In thesedata, solar generation grew at an average annualrate of 3%, but from a tiny base. Because the U.S.Energy Information Agency (EIA) tracks genera-

tion only from solar installations with capacitiesabove 1 megawatt (MW), it seems likely that solargeneration at the end of this period was under-stated by at least 60%.1 Even correcting for thisbias, however, solar’s share remains tiny. Wind,which grew at an average annual rate of 15.9%,accounted for the bulk of NHR growth over thisperiod.

6.0%

5.0%

4.0%

3.0%

Per

cent

age

2.0%

1.0%

0.0%

Year19

8919

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

06

United States Europe

Source: EIA n.d.b

Figure 11.2. Share of nonhydro renewable electricity in total generation, 1989–2006

120,000

100,000

80,000

60,000

Mill

ions

of k

Wh

40,000

20,000

0

Year19

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

07

WindOther biomass

Geothermal

Wood & wood derivedSolar thermal & photovoltaic

Source: EIA 2009b, Data Table EIA-906

Figure 11.3. U.S. electricity generation from nonhydro renewable energy, 1990–2007

Renewable Electricity Generation in the United States 211

Many analysts contend that even with currenttechnology, nonhydro renewables have the poten-tial to play a much larger role in the United Statesthan they do at present. Table 11.1 comparesactual generation from nonhydro renewables inthe United States in 2007 with estimates by theInternational Energy Agency (IEA) of “total real-izable potential by 2020.”2 These estimates areintended to reflect both natural endowments (e.g.,average solar radiation) and the relative costs ofcurrent NHR technologies, but they are inevita-bly imprecise and should accordingly be treatedwith caution. It is worth noting, however, thateven though the IEA believes that solar genera-tion has not even reached 1% of its potential,despite decades of attention by policymakers, theagency also estimates that its ultimate potential ismuch less than either biomass or wind.

There has been a great deal of variation instate-level experience with NHRs. In 2007,NHRs accounted for 2.53% of total U.S. net gen-eration, but more than 5% in seven states and lessthan 1% in eleven states. This variation reflectsdifferences in both the potential for variousrenewable technologies and state-level policiestoward renewables. To shed quantitative light onthe relative importance of these two sources ofvariation would require plausible estimates ofstate-level, technology-specific potentials compa-rable to the IEA estimates in Table 11.1, but nosuch estimates appear to exist.3

Table 11.2 provides information on the sevenstates for which NHR generation accounted formore than 5% of total generation, as well as thetwo states not in this set that were in the top fivein terms of total NHR generation. These nine

Table 11.1. Actual and potential NHR generation in the United States

ResourceEstimated total realizablepotential by 2020 (TWh)

Actual 2007 grossgeneration (TWh)

Actual as a percentageof potential

Biomass 501.6 58.6 11.7

Wind 300.4 32.3 10.8

Solar 85.2 0.7 0.8

Geothermal 36.0 16.9 47.0

Tidal & wave 2.3 0.0 0.0

Total 925.5 108.4 11.7

Sources: Estimated potentials from IEA 2008a, 65; generation from IEA 2008c, 396Note: TWh = terawatt-hours

Table 11.2. Leading NHR generation states

2007 NHRgeneration

State Percent of state total TWh Main NHR technology or technologies

Maine 26.1 4.21 Wood/wood waste

California 11.8 24.85 Geothermal

Vermont 8.0 0.65 Wood/wood waste

Minnesota 7.2 3.93 Wind

Hawaii 6.6 0.75 Wind, geothermal

Iowa 5.8 2.91 Wind

Idaho 5.7 0.65 Wood/wood waste

Texas 2.5 10.29 Wind

Florida 1.9 4.30 Wood/wood waste, other biomass

Source: EIA 2009c

212 Richard Schmalensee

states accounted for 92% of U.S. NHR generationin 2007. For most states, one technology is thedominant contributor to NHR generation, buttwo are nearly tied in both Hawaii and Florida.The importance of wood and wood waste is clearhere, as it was in Figure 11.3, particularly in heav-ily forested states like Maine,Vermont, and Idaho.The so-called “wind belt,” extending northwardfromTexas to the Canadian border, is visible here,though some states in that belt are conspicuous bytheir absence. (The wind belt is discussed belowin the section on Wind Power in the UnitedStates.) It is also interesting to note theunimportance of solar power, even in states withabundant solar resources such as California,Hawaii, and Texas.

Federal Policies in Supportof NHRsThe U.S. federal government has long supportedresearch and development (R&D) aimed atadvancing NHR technologies and has morerecently moved to subsidize their deployment.Motivations for such support have varied over

time; energy security is less important now thanearlier, and environmental concerns, particularlythose associated with global climate change, havebecome more important in recent years.

Research and Development

Government financial support for basic researchand precommercial development aimed atadvancing NHR, or almost any other, technologycan be justified by the positive externalities thatknowledge spillovers produce. Despite this ration-ale and strong rhetorical support for NHRs, how-ever, the data reveal that U.S. policymakers havehistorically allocated more generous R&D sup-port to fossil-fuel and nuclear technologies, whichare generally much more mature than NHRs.Between 1978 and 2007, federally sponsoredR&D on renewable technologies amounted to17.8% of total energy-directed R&D, while39.3% was spent on nuclear technologies and32.1% on fossil-fuel technologies (EIA 2008a,40).4 Figure 11.4 graphs federal expenditure onrenewables during this period.

Not only has federal R&D support for NHRslagged behind that for more conventional tech-nologies, but it has also varied substantially over

30.0%

25.0%

20.0%

15.0%

Per

cent

age

10.0%

5.0%

0.0% 0

200

400

600

Mill

ions

of 2

007$

800

1,000

1,200

1,400

1,600

Year19

7819

7919

8019

8119

8219

8319

8419

8519

8619

8719

8819

8919

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

0720

08

R&D on renewables as a percentage of all energy R&DAbsolute R&D expenditure on renewables

Source: EIA 2008a

Figure 11.4. Federal expenditure on R&D in renewable technologies, 1978–2007

Renewable Electricity Generation in the United States 213

time in both relative and absolute terms, as Figure11.4 shows. R&D for NHRs, and for most otherenergy technologies, peaked in 1980, fueled byintense concerns over energy security and rapidlyrising oil prices. As oil prices receded, so didenergy-related R&D. Since the early 1980s, R&Dfunding in support of NHRs has been both mod-est and variable from year to year—hardly condu-cive to long-term, sustained efforts aimed atmajor breakthroughs.

Support for NHR Deployment:Conceptual Overview

Before examining actual federal and state policiesto support the deployment of existing NHRtechnologies, it is worth noting that such policiesare difficult to justify economically. In the pres-ence of a binding cap on carbon dioxide emis-sions, for instance, a subsidy to NHR deploymentwill have no impact on total emissions but willraise the total social cost of meeting the cap. Andalthough many claim that widespread deploymentof NHRs will lower their costs through learningby doing, they rarely note that firm-specific learn-ing that does not lower the costs of other firmsdoes not justify subsidies. Rigorous empirical sup-port for the importance or even existence of suchspillover benefits is lacking. Other arguments forsubsidizing NHR deployment, such as capitalmarket imperfections (which somehow permitlarge, risky investments in other sectors), infantindustry considerations (which logically shouldapply across the economy and have supportedpolicies with a terrible historical record), and jobcreation (which lacks rigorous support and runscounter to historical progress by favoring labor-intensive over labor-saving technologies), are evenless persuasive.

Nonetheless, subsidies of four basic sorts havebeen adopted in the United States and abroad:feed-in tariffs, output subsidies, investment subsi-dies, and output quotas. Feed-in tariffs, whichguarantee a predetermined, above-market pricefor power over a period of years, are the mostpopular policy device outside the United States(BMU 2007; Coenraads et al. 2008; EREC 2007;European Commission 2008; IEA 2009).5 They

provide strong incentives for minimizing costs andmaximizing production. Feed-in tariffs, however,generally do not provide stronger incentives forgenerating electricity when it is more valuable(e.g., by scheduling maintenance accordingly),and they provide an invisible subsidy by shifting allrisk related to the supply of and demand for elec-tricity to other market participants. An outputsubsidy, paid on top of market price, can eliminateboth of these shortcomings while retaining theother good incentive properties of a feed-in tariff.Output subsidies are not widely employed, how-ever, and like feed-in tariffs, they can provideincentives to operate NHR facilities even whenthe marginal value of their generation is negative(e.g., see the discussion of Texas in the section onState Policies and Experience).

Investment subsidies are not particularlyattractive economically, because they provideweaker incentives for reducing initial cost than dofeed-in tariffs or output subsidies. Nonetheless,governments in the United States and abroad thatpromote deployment of NHRs almost all useinvestment subsidies as part of their policy pack-ages. Finally, output quotas, known in the UnitedStates as renewable portfolio standards (RPSs),typically require load-serving entities to generateor procure a minimum fraction of energy fromNHRs. This approach is not as popular as feed-intariffs abroad, but it is very popular at the statelevel in the United States and is part of legislationbeing actively debated at the federal level.6 Notonly has the United States adopted a different mixof policies than most other wealthy nations, but italso has implemented those policies in ways thatsignificantly limit their efficiency and effective-ness.

Federal Support of NHR Deployment

Somewhat ironically, in light of subsequent devel-opments, the first federal initiative that supporteddeployment of NHRs did so almost unintention-ally and led to the establishment of generousfeed-in tariffs in several states.The Public UtilitiesRegulatory Policies Act of 1978 (PURPA) wasprimarily aimed at opening electric utilities tocompetition and increasing efficiency in electri-

214 Richard Schmalensee

city markets. PURPA required utilities to pur-chase electricity generated from certain defined“qualified facilities” at the utilities’ avoided costs.Qualified facilities could be either cogenerationfacilities, which produced both useful heat andelectricity, or certain small NHR generators.Because electric utilities at that time were almostall vertically integrated, avoided costs were to bedetermined by state regulators rather than marketprices, and regulators in some states (notably Cali-fornia, as discussed later in the section on StatePolicies and Experience) responded by establish-ing feed-in tariffs that were based on the expecta-tion of high and increasing generation costs. Ascosts of conventional generation in fact camedown, this system became unsupportable, and itwas largely dismantled by the early 1990s(Borenstein and Bushnell 2000, 48).

Since then, federal policy has promoted NHRdeployment primarily through favorable corpo-rate income tax provisions: accelerated deprecia-tion and tax credits for production and invest-ment. After 1986, most NHR generating assets,which had been depreciated over 15 years for taxpurposes, could be written off over 5 years. (Thelist of eligible NHR technologies was expandedin 2005 and 2008.) This increased the presentvalue of tax deductions for depreciation by aroundhalf.7

The Renewable Electricity Production TaxCredit (REPTC) was first established by theEnergy Policy Act of 1992.8 It provided for a cor-porate income tax credit of 2.1 cents per kilowatt-hour (kWh; 1.5 cents/kWh in 1993$, indexed forinflation) for generation using some technologiesand half that for others for (generally) the first 10years of operation. Favored NHR technologiesare currently wind, closed-loop biomass, andgeothermal; other eligible technologies includeopen-loop biomass, landfill gas, municipal solidwaste, and certain hydroelectric, marine, andhydrokinetic facilities. The legislation establishingthe REPTC also established a Renewable EnergyProduction Incentive (REPI) program, whichauthorized payments roughly equivalent to theproduction tax credit to entities such as state andlocal governments that were not corporateincome tax payers.

This output subsidy policy has not been con-sistently or predictably implemented over time.Payments actually made under the REPI must beappropriated annually and thus are far from cer-tain. Solar facilities were eligible for the REPTConly briefly—if they began operation in 2005.The REPTC expired at the end of 2001 and wasthen extended in March 2002. It then expired atthe end of 2003 and was not renewed until Octo-ber 2004, in legislation that extended it until theend of 2005. Legislation passed in 2005 extendedit through the end of 2007, legislation passed in2006 extended it through 2008, and laws passedin 2008 and 2009 revised and extended it through2012 for wind and 2013 for other technologies.Figure 11.5 shows a surge in installation of windcapacity during 2001 before the REPTC expired,followed by a drastic drop-off during 2002,reflecting the uncertain status of the REPTC untilMarch and the lag between project initiation andcompletion. Similarly, the unavailability of theREPTC during 2004 shows clearly in the figure.If investors cannot rely on a subsidy’s remaining inplace, that subsidy provides at most weak incen-tives for long-term investments in such things astechnology development and efficient productioncapacity.

The Energy Tax Act of 1978, which waspassed along with PURPA, established investmenttax credits for a variety of NHR technologies.These were modified several times in the ensuingyears (EIA 1999). Since 2005, the ResidentialRenewable Energy Tax Credit (RRETC) hasprovided personal income tax credits for up to30% of investments in solar electric systems, solarhot water systems, wind turbines, fuel cells, andgeothermal heat pumps. Also since 2005, theBusiness Energy Investment Tax Credit (BEITC)has provided a 30% corporate income tax creditfor investment in essentially all solar systemsexcept those used to heat swimming pools, as wellas for fuel cells and small wind turbines. It pro-vides a 10% investment tax credit for certain othertechnologies. Both of these provisions were ini-tially scheduled to expire at the end of 2008, butlegislation that year extended them to 2016.

For solar systems, the initial investmentaccounts for most of the life cycle cost, so a 30%

Renewable Electricity Generation in the United States 215

investment tax credit is a very large subsidyindeed. The Interstate Renewable Energy Coun-cil (Sherwood 2009) reports that the annualgrowth rate of photovoltaic capacity installeddoubled in 2006, and capacity installed in 2008was triple that in 2005. Similar dramatic growthoccurred in solar hot water and space heating. It isimportant to note, however, that an investmenttax credit is most valuable when it is less thancurrent taxable income. This condition is prob-ably satisfied for most homeowners who can seri-ously consider installing a solar system, but it isunlikely to be satisfied for any corporation spe-cializing in grid-scale solar power. The need forsuch firms to use joint ventures and other devicesto ensure that the full value of the investment taxcredit is received can add significant friction to theprocess of financing solar projects.

The REPTC and BEITC have been the mostimportant sources of support for renewablesdeployment. In FY2007, the reductions in taxrevenue caused by subsidized financing ofrenewables facilities under other programs was$100 million, compared with $690 million insuch tax expenditures for the REPTC andBEITC. But programs supporting fossil fuels wereconsiderably more costly, resulting in $2.7 billionin tax revenue reductions (GAO 2007, 37).

The American Recovery and ReinvestmentAct of 2009, generally known as the stimulus bill,allows taxpayers eligible for the BEITC or theREPTC for facilities entering service or, gener-ally, beginning construction in 2009 or 2010 toelect to receive a BEITC-equivalent cash grantinstead.The rationale is that tax credits are of lim-ited value during a period of unusually low cor-porate profits; but on the other hand, entitlementto a grant is of no value if Congress does notappropriate sufficient funds.

Several relatively small federal grant and loanguarantee programs also exist, each targeted atcertain classes of entities (e.g., municipal govern-ments) and technologies. These programs aremodified from time to time, and the actual fund-ing available is determined each year by theappropriations process. In FY2006, excludingenergy efficiency, the federal government madeonly $16.7 million in grants to support renewablesand guaranteed just $23.8 million in loans (GAO2007, 50).

As with R&D, it is interesting to comparefederal subsidies for the use of NHRs with subsi-dies for other generation technologies. Table 11.3shows EIA estimates of total subsidies and supportby technology for 2007, both in absolute dollarterms and per megawatt-hour (MWh) of genera-

5,900

4,900

3,900

2,900

1,900

Cap

acity

in m

egaw

atts

900

Year

–100

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

Source: EIA 2009b, Table EIA-860

Figure 11.5. Wind electricity capacity addition, 1994–2007

216 Richard Schmalensee

tion. In absolute terms, coal (especially clean coal)and nuclear power were the most heavily subsi-dized, while per unit of output, solar and windwere by far the biggest winners. It is interesting tonote that wind, in particular, received more than12 times as much support as “other NHR,” eventhough those technologies accounted for morethan twice as much generation.

State Policies and ExperienceThis section begins with an overview of statepolicies in support of NHR deployment and thenprovides brief discussions of the California andTexas experiences. It focuses initially on what isgenerally considered to be the most importantstate-level NHR policy and is certainly one of themost popular: renewable portfolio standards.9

Renewable Portfolio Standards

Renewable portfolio standards require that aminimum percentage of electricity generated orsold by a covered entity come from sources desig-nated as renewable. Compliance is usually meas-ured on an annual basis, and the required percent-age typically increases over time. Iowa enacted thefirst RPS in 1983, and Nevada adopted the sec-ond in 1998. Since then, the pace has increaseddramatically, and 29 states plus the District ofColumbia now have RPSs. These include all thestates listed in Table 11.2 except Vermont (whichhas a voluntary standard), Idaho, and Florida.

States with RPSs accounted for 62% of U.S. netelectricity generation in 2007. Table 11.4 liststhese states in order of their initial RPS adoptionand gives some of their most important currentfeatures. Five additional states have voluntarygoals for renewable energy; these are listed inTable 11.5.

While all RPS programs have the commongoal of increasing the share of renewableresources, they differ considerably along multipledimensions. One important difference is howcompliance is to be achieved. In states withorganized wholesale markets, entities that distrib-ute power are generally responsible for meetingRPS targets and are given considerable freedomto choose how to do so. In states that have regu-lated, vertically integrated utilities, regulatorsoversee contracting and utility procurement. InNew York and Illinois, a state agency has directresponsibility for the procurement of renewablesunder the RPS (Wiser and Barbose 2008). Somestates have legislated explicit per-MWh penaltiesfor noncompliance, whereas others allow thestate’s public utility commission to determine theappropriate penalty.

Differences also exist in the definition ofresources that are deemed renewable and whetherthe RPS is applicable solely to investor-ownedutilities or is extended to smaller retail supplierswith a lower target, as is the case in Colorado andOregon. In some instances, the size of a facility isan important determinant of whether its outputcounts toward RPS requirements. Maine, forinstance, requires facilities to be 100 MW or

Table 11.3. Subsidies and support to electricity production by technology, 2007

Fuel/end use2007 net generation

(million MWh)2007 subsidy

($ million)Subsidy($/MWh)

Coal & refined coal 2,018 3,010 1.49

Natural gas & oil 919 227 0.25

Nuclear 794 1,267 0.16

Hydroelectric 258 174 0.67

Solar 1 14 24.34

Wind 31 724 23.37

Other NHRa 70 59 0.84

Source: EIA 2008a, xviaIncludes biomass (and biofuels), landfill gas, municipal solid waste, and geothermal

Renewable Electricity Generation in the United States 217

Table 11.4. State renewable portfolio standards (RPSs)

StateFirst RPSadoption Current target Other requirements

Iowa 1983 105 MW

Nevada 1997 6% by 2005, 20% by 2015 5% of RPSs to be solar in each year

Massachusetts Nov. 1997

15% by 2020 (Class 1) and 1% each yearthereafter, 3.5% of sales each year startingin 2009(Class 2)a To be determined

Connecticut July 1998 27% by 2020 20% Class 1, 3% Class 2, 4% Class 3b

New Jersey Jan. 1999 22.5% by 2021 2.12% solar by 2021

Texas Sept. 1999 2,280 MW by 2007, 5,880 MW by 2015 500 MW from sources other than wind

Maine 1999 30% by 2000 (Class 2), 10% by 2017(Class 1)c

Hawaii 2001 20% by 2020

Wisconsin Oct. 2001 10% by 2015

California 2002 20% by 2010

Maryland May 2004 20% by 2022 (Tier 1), 2.5% 2006–2018(Tier 2)d 2% solar by 2022

Rhode Island June 2004 16% by 2020

New York Sept. 2004 24% by 2013 Not specific

Colorado Nov. 2004 20% by 2020 4% of RPSs to be solar in each year

Pennsylvania Nov. 2004 18% by 2020 (8% Tier 1, 10% Tier 2)e 0.5% solar by 2020

DC April 2005 20% by 2020 0.4% from solar by 2020

Montana April 2005 5% in 2008, 10% in 2010, 15% in 2015

Delaware July 2005 20% by 2019 2.005% photovoltaic by 2019

Arizona Nov. 2006 15% by 2025 30% of RPSs from distributed renewables after 2012

Washington Nov. 2006 15% by 2020

Minnesota Feb. 2007 25% by 2025, Xcel Energy 30% by 2020 Xcel Energy 25% of RPSs from wind in each year

New Hamp-shire

May 2007 23.8% by 2025 0.3% solar, 6.5% existing biomass, 1% existing smallhydro

Oregon June 2007 25% by 2025 Varies by utility

Illinois August 2007 25% by 2025 75% from wind

North Carolina Aug. 2007 12.5% by 2020 0.2% solar & thermal by 2018, 0.2% swine waste by2018, 900,000 MWh from poultry waste by 2014

New Mexico Aug. 2007 20% by 2020 4% solar, 4% wind, 2% geothermal & biomass, 0.6%distributed renewables

Michigan Oct. 2008 10% by 2015 Varies by utility

Missouri Nov. 2008 15% by 2021 0.3% solar by 2021

Ohio Jan. 2009 12.5% by 2025 0.5% solar by 2025

Kansas May 2009 20% by 2020

Source: DSIRE 2009a

aClass 1: (in-state, on-site) solar, wind, ocean thermal, wave and tidal, fuel cells, landfill gas, qualifying hydroelectric, qualifying biomass,geothermal; Class 2: (in-state, on-site) systems dating prior to December 1997 using the same technologies as Class 1bClass 1: solar, wind, fuel cell, landfill gas, small hydroelectric, wave, tidal and ocean thermal, sustainable biomass; Class 2: trash to energy,biomass not included in Class 1; Class 3: customer-sited cooling-heating-power systems, recent savings from conservation and loadmanagement, recycled energy from heat pipescClass 1: RPS mandate to provide 30% of sales through renewables; Class 2: portfolio goal to increase new renewable capacity by 10% by2017dTier 1: solar, wind, qualifying biomass, landfill gas, geothermal, wave, tidal and ocean thermal, small hydroelectric, fuel cells; Tier 2: trash toenergy, hydroelectric other than pump storageeTier 1: new and existing solar, wind, small hydro, geothermal, biomass, fuel cells, qualifying gas; Tier 2: new and existing waste coal, largehydro, waste to energy, distributed generation, demand-side management, certain biomass

218 Richard Schmalensee

smaller. In other states, the age of a facility deter-mines whether it is considered eligible. For exam-ple, in Massachusetts, only capacity installed after1997 is considered eligible. Hydroelectric facili-ties are generally eligible subject to the capacityconstraint that each state sets on facilities. Somestates, such as New Hampshire, are explicit intheir consideration of small hydroelectric facili-ties. A number of states have tiers or set-asides,often with different time frames or target levels.Fifteen states and the District of Columbia haveprovisions favoring solar power or distributedgeneration; nine have minimum solar require-ments of various sorts; and the others give extracredit for solar or distributed generation. Illinois,on the other hand, requires that 75% of renewablegeneration come from wind. Nine of the RPSjurisdictions give at least some credit for solar hotwater systems as displacing nonrenewable genera-tion, and Hawaii, Nevada, and North Carolinahave provisions that allow demand-side efficiencyto be used to meet a part of the RPS requirements(Wiser and Barbose 2008).

The most common mechanism for demon-strating compliance with RPSs is the purchase ofrenewable energy certificates (RECs) (Corey andSwezey 2007). Renewable generators sell powerat the market price and then also sell, in effect, a 1MWh REC for each MWh of electricity theyhave sold. Distribution utilities and others obligedto obtain a minimum percentage of their electri-city from NHRs demonstrate compliance by pur-chasing an appropriate number of RECs and sur-rendering these to the authorities. The ability totrade RECs ensures that costs are minimizedwithin the state, as there are economic incentivesto, in effect, produce the certificates using the

cheapest available NHR technology. (This regimedoes not, however, create any incentive to favortechnologies with large spillover benefits.)Because the potential for NHR generation differswidely among states, even in the absence of anationwide RPS, interstate trading of RECswould potentially be an important way of redu-cing the cost of meeting the states’ goals. Butunfortunately, state RPS programs differ in somany dimensions—including the precise defini-tion of an REC—that interstate trading is virtu-ally impossible. Indeed, some state RPS programsprohibit it altogether.

At the federal level, in 2005, the Senate passeda bill containing a national RPS that would haverequired 10% of electricity in the country to begenerated by renewables by 2020, but the bill diedin the House. In 2007, the House passed legisla-tion containing a national RPS of 15% by 2020;this bill died in the Senate. Most recently, theAmerican Clean Energy and Security Act of2009, or Waxman–Markey bill, passed by theHouse in June 2009, contains a national RPS withnationally tradable RECs. The bill’s standard,which could be met with a combination ofenergy efficiency savings and NHR generation,would start at 6% in 2012 and rise to 20% by2020. As of this writing, the fate of this provisionis yet to be determined.

A majority of state RPS programs have onlyrecently become operational—10 of them are lessthan three years old, and 19 are less than five yearsold. Furthermore, RPSs are just one of the manystate-level policies that have been adopted to pro-mote renewable energy. As a consequence, it isdifficult to make confident statements about theeffectiveness of RPSs in increasing NHR genera-

Table 11.5. State voluntary renewables goals

State Date goal adopted Current goal

North Dakota August 2007 10% by 2015

South Dakota February 2008 10% by 2015

Vermont March 2008 20% by 2017

Utah March 2008 20% by 2025

Virginia July 2009 15% by 2025

Source: DSIRE 2009a

Renewable Electricity Generation in the United States 219

tion, let alone assess their costs. A significantimpact is suggested by the fact that in 2007, stateswith RPS programs accounted for an overwhelm-ing 86% of new NHR generating capacity, ascompared with just 22% of all new generatingcapacity (EIA 2009b, Data Table EIA-860). In amultivariate statistical analysis, Menz and Vachon(2006) find that RPS programs effectively encour-age cumulative renewable energy investment andcapacity deployment. However, in a later analysisusing additional control variables, including theenvironmental orientation of each state’s legisla-tors and the size of each state’s agencies concernedwith natural resources, Carley (2009) finds thatthe adoption of RPS mandates does not effec-tively increase the share of renewable electric gen-eration.There are at least reasons to be concernedthat some of these programs may fail to meet theirgoals.

Other State Policies

In addition to RPSs of various shapes and sizes,state governments have adopted a wide variety ofother measures aimed at promoting NHR gen-eration. Table 11.6 provides some informationregarding their popularity. As with RPSs, no twostate policies for a particular type of incentive areidentical.

The IEA lists the three most important statepolicies promoting renewables as RPSs, publicbenefit funds, and tax incentives (IEA 2008a).Public benefit funds are generally financed by asmall surcharge on retail electric rates and are usedto support renewable energy in a wide variety ofways. They are projected to total $7.3 billion by2017 (DSIRE 2009a). All states except Arkansasoffer some subsidy for investment in NHR gen-eration, but the design and impact of tax benefits,rebates, grant, and subsidized bond or loan pro-grams vary enormously.

Beginning with Massachusetts and Wisconsinin 1982, 42 states and the District of Columbiahave established net metering policies, and theEnergy Policy Act of 2005 requires all utilities toprovide net metering to customers that request it.Net metering allows utility customers with someNHR generators, generally only small residentialor commercial installations, to sell electricity tothe distribution entity that serves them at theretail rate the customer pays for electricity, not thetypically much lower wholesale rate the distribu-tion entity pays for other power. In Massachusettsin 2007, for instance, retail rates averaged $0.152per kWh, while the average wholesale price in theNew England market was only $0.068 per kWh(EIA n.d.b; FERC 2009). A small part of this dif-ference reflects power losses in transmission anddistribution, but these losses average only about7% of net generation in the United States (see,e.g., EIA 2009b). Most of the wholesale–retaildifference arises simply because regulated pricesdo not reflect incremental costs: retail rates aregenerally set to recover the fixed costs of distribu-tion systems through a per-kWh charge added towholesale electricity rates rather than throughfixed charges of one sort or another.

Although net metering programs are popularin state capitals, they are not yet widely used.Only 48,280 utility customers participated in netmetering programs in 2007; 95% were residential,and 72% were in California (which established itsprogram in 1995). But participation did grow at a46% annual rate over the 2004–2007 period (EIA2007, 2008b).

Ten U.S. states—Connecticut, Delaware,Maine, Maryland, Massachusetts, New Hamp-

Table 11.6. Other state policies to promoteNHR generation

Type of incentiveNumber of

states

Personal tax: credits or other 21

Corporate tax: credits or other 23

Sales tax: exemption or deduction 25

Property tax: exemption or special assess-ment 32

Rebates programs 19

Grant programs 22

Subsidized bond or loan programs 34

Production incentives 9

Public benefit funds 18

Net metering 43

Source: DSIRE 2009aNote: The District of Columbia is counted as a state in this table

220 Richard Schmalensee

shire, New Jersey, New York, Rhode Island, andVermont—have recently signed the RegionalGreenhouse Gas Initiative (RGGI) agreement.10

The agreement obliges these states to cap theirtotal CO2 emissions from the electric power sec-tor through 2015, and then reduce them by 10%by 2018. Beginning at the start of 2009, electricutilities in these states have had to obtain and sur-render allowances equal to their CO2 emissions.These allowances are mainly auctioned by thegovernments of the 10 states involved. In princi-ple, this system caps utility CO2 emissions in theaffected region, but allowance prices have so farbeen quite low: $2 to $3 per ton of CO2, a frac-tion of prices in the European Union’s EmissionsTrading System for CO2. It thus seems unlikelythat this system has so far had much effect on thedeployment of nonhydro renewables.

Finally, it is interesting to note the relativeunpopularity of production incentives such asfeed-in tariffs and output subsidies in the UnitedStates. Perhaps the most important in this countryis the California feed-in tariff discussed below, butit is available only to small generators.

California: A Long History of Carrots

California was an important early leader amongU.S. states in promoting NHR generation.Despite generating only 5.5% of the nation’s elec-tricity in 1990, California accounted for 37% ofU.S. NHR generation that year. As Figure 11.6shows, however, California’s share of NHR gen-eration has declined over time as other states havemoved to promote renewables, but in 2007, Cali-fornia still accounted for just under 24% ofnational NHR generation.

California’s early high share of national NHRgeneration is mainly due to its response toPURPA. In 1983, the California Public UtilitiesCommission (CPUC) developed policies thatguaranteed qualifying facilities generous feed-inrates for a period of 10 years. These policies werebased on assumptions that oil prices (and thereforeavoided costs) would continue to rise from whatwere then already high levels (Hirsh 1999). Theresult was a boom in construction of small NHRgenerators and other qualifying facilities. Eventhough the CPUC suspended further contracts

40.0%

35.0%

30.0%

25.0%

20.0%

Per

cent

age

15.0%

10.0%

5.0%

0.0%

Year19

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

07

Share of nonhydro renewables in all California generationCalifornia's share of all U.S. nonhydro renewable generation

Source: EIA 2009b, Data Table EIA-906

Figure 11.6. Nonhydro renewables in California, 1990–2007

Renewable Electricity Generation in the United States 221

for power generation in 1985, qualifying facilitiesthat had already contracted with the CPUC werepermitted to sell power at high rates. By 1986,California had nearly 90% of global wind generat-ing capacity (DOE 2008, 6). Figure 11.7 showsthat the bulk of the resulting NHR generationwas powered by geothermal energy and biomass.It also shows that a number of these facilities shutdown when their PURPA contracts expiredbetween 1993 and 1995, though most continuedto operate.

In the ensuing years, California adopted awide variety of policies to promote solar andother NHR generation, relying more heavily onsubsidies of various sorts than on regulation ormandates, but Figures 11.6 and 11.7 show thatthese policies did not produce significant increasesin the NHR share of total generation or even theabsolute amount of NHR generation.11 In 1995,for instance, the CPUC proposed an RPS regimeto increase renewable generation, but the follow-ing year, the legislature adopted a production-based auction funded by surcharges on electricutility bills instead (Wiser et al. 1996). This pro-gram ran from 1998 to 2001, provided between$540 million and $640 million in funding eachyear, and supported 4,400 MW of existing and

1,600 MW of new renewable capacity (Ritscheland Smestad 2003; Wiser et al. 1996).12

Between 1998 and 2007, California’s Emerg-ing Renewables Program, funded by the ratepay-ers of California’s four largest investor-ownedutilities, has funded roughly 130 MW of capacityadditions from smaller wind, solar, and fuel cellfacilities (CEC 2009).The Self Generation Incen-tive Program (SGIP), funded at roughly $83 mil-lion each year, supports customer-generatedrenewable energy via wind, solar, and fuel cellsources. Between 2001 and 2007, the SGIPfunded over 300 MW of total capacity addi-tions.13

In 2006, California adopted the CaliforniaSolar Initiative (CSI), a 10-year $3.2 billion pro-gram to fund the development of 3,000 MW ofsolar capacity (DSIRE 2009c). Part of the CSI isthe New Solar Homes Partnership, a 10-year$400 million incentive-based program aimed atconstructing homes with solar energy systemswith a focus on energy efficiency (DSIRE2009d).

Beyond the big-ticket subsidy programsdescribed above, the state of California runs a slewof smaller, more localized funding programs forrenewable energy development. These include 5

30,000

25,000

20,000

15,000

Mill

ions

of k

Wh

10,000

5,000

0

Year19

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

07

WindBiomass, wood & wood derived

GeothermalSolar thermal & photovoltaic

Source: EIA 2009b, Data Table EIA-906

Figure 11.7. Electricity generation by nonhydro renewable sources in California, 1990–2007

222 Richard Schmalensee

utility-specific grant programs, 11 utility-specificloan programs, and 94 utility-specific rebate pro-grams and other smaller financial incentives. Andsolar systems are exempted from local propertytaxes.

Following eight earlier adopters, Californiaenacted an RPS in 2002. The California RPSrequires electric utilities to acquire 20% of theirelectricity from renewable sources by 2010. Morerecently, the state has adopted a nonbinding goalof 33% by 2020. Of the 7,000 MW of contractsfor renewable generation signed between 2002and 2007, 53% is wind, 23% solar, 12%geothermal, 7% biomass, and less than 1% smallhydro and ocean energy (Wiser and Barbose2008). It should be clear from the foregoing,however, that the RPS is but one in a large set ofpolicies aimed at promoting NHR generation inCalifornia.

In 2008, California established a feed-in tarifffor small renewable energy facilities with capaci-ties of 1.5 MW or less. It has the efficiency-enhancing feature that rates are to vary by time ofday. Facilities that sell power under this tariff arenot eligible for additional state incentives and pro-grams (DSIRE 2009b).

Texas: RPS-Driven Wind EnergyDevelopment

In contrast to California, Texas was slow toembrace NHR generation, and it does not seemto have made much use of explicit subsidies. In1990,Texas accounted for less than 2% of nationalNHR generation, even though it generated 9.3%of the nation’s electricity. Dramatic changes beganto occur in 1999, when Texas adopted an RPSprogram with binding obligations beginning in2002 and extending (as amended) to 2015. AsFigure 11.8 shows, the result was extraordinarygrowth in wind power, dwarfing the Californiagrowth shown in Figure 11.7. In 2007, Texasaccounted for 9.8% of U.S. electricity generationand an identical percentage of NHR generation.And in 2008, Texas accounted for 31% of newU.S. wind generation capacity, more new capacitythan any country except China and the UnitedStates (AWEA 2009). Texas currently has about8,800 MW of wind generating capacity (AWEA2009; ERCOT 2009a), which is about 12% oftotal capacity, and is projected to have 18,500MW of wind generating capacity by 2015(O’Grady 2009).

12,000

10,000

8,000

6,000

Mill

ions

of k

Wh

4,000

2,000

0

Year19

9019

9119

9219

9319

9419

9519

9619

9719

9819

9920

0020

0120

0220

0320

0420

0520

0620

07

Wind Biomass

Source: EIA 2009b, Data Table EIA-906Note: Solar-generated electricity accounted for up to 385,000 kWh of generation in 1990, but that number decreased to zero in 2001

Figure 11.8. Electricity generation by nonhydro renewable sources in Texas, 1990–2007

Renewable Electricity Generation in the United States 223

Observers who have analyzed Texas’s RPSattribute its strong impact to several factors. Atleast initially, the system was technology-neutral,14 so the most economical resource—wind in this case—could be used most intensively.As discussed below in the section on Wind Powerin the United States,Texas, along with a few otherstates, is blessed with excellent wind resources.This is particularly true of western Texas, wherewind speeds average 8 meters per second year-round, and wind power facilities can operate atcapacity factors of 40% or more (Langniss andWiser 2003), at the upper end of the 25%–40%range commonly cited for wind generators (see,e.g., AWEA 2009). The remote location of thesesites and the lack of transmission infrastructureinitially inhibited exploitation of their potential.Subsequent legislation required utilities toupgrade their transmission systems as necessary tomeet the state’s RPS goals and allowed them torecover the costs in retail rates (DSIRE 2009a;Langniss and Wiser 2003). As discussed furtherbelow, however, transmission from these sitesnonetheless remains a problem.15

It is also important that the Texas RPS pro-gram applies to most of the state’s retail electricityload in a way that permits these remote sites to befully exploited. While most RPS programs setgeneration requirements,Texas’s RPS sets require-ments for capacity and thus for capacity additionsyear by year. Owners of renewable generatingcapacity can sell both electricity and renewableenergy credits to any electricity retailer in theTexas market, and retailers can freely buy, sell, andbank these credits. A capacity conversion factor,now based on performance of renewable generat-ing units in the two prior years, is used to convertcapacity requirements into MWh requirements,which are allocated to retailers based on theirshares of statewide electricity sales. The relativelylarge, organized, competitive wholesale powermarket in Texas (Potomac Economics 2009) hasenabled an effective market for renewable energycredits.16

In addition, the structure of the Texas RPSenables renewable energy suppliers to sign long-term (10- to 25-year) contracts, which reducestheir risk and makes it easier to raise capital, thus

lowering costs. Because the law sets annualrequirements, each retailer can forecast how manycredits it will need in the future and can thusconfidently sign long-term contracts withrenewables suppliers.17

Finally, the Texas RPS legislation providesstrict, automatic penalties for noncompliance,while providing flexibility to electricity retailersthrough the ability to trade and bank renewableenergy credits (Langniss and Wiser 2003). TheCalifornia program, in contrast, seems to rely onthe general enforcement powers and discretion ofthe Public Utility Commission to deal with anyinstance of noncompliance.

Wind Power in theUnited StatesAccording to the IEA (2008a, 65–66), the poten-tial for wind-powered electricity generation inthe United States substantially exceeds that in anyother Organisation for Economic Co-operationand Development (OECD) nation, as well in Bra-zil, Russia, India, and China. In 2008, the U.S.Department of Energy published a detailed analy-sis of a scenario under which wind would accountfor 20% of U.S. electricity generation by 2030(DOE 2008). And, as noted earlier in the sectionon Nonhydro Renewables in the United States,wind-powered generation grew at a 15.9% aver-age annual rate during 1990–2007 and accountedfor the bulk of NHR growth over that period.That growth has not been geographically uni-form, however. Additionally, because wind, likesolar but unlike some other NHR technologies, isintermittent, that growth is raising new issues ofpower system design and operation.18

According to the American Wind EnergyAssociation, 12 states in the so-called Wind Belt,stretching northward from Texas to the Canadianborder, have 93% of the wind energy potential inthe United States (AWEA n.d.).19 Table 11.7 pro-vides information on these states, as well as 3 stateson the Pacific Coast that have substantial windgeneration despite having considerably less poten-tial than the Wind Belt states. Together, the 15

224 Richard Schmalensee

states listed in the table accounted for 30% of allU.S. electricity generation in 2007 and 92% ofwind-powered generation. The table also showswhen the states that have RPSs initially enactedthem.

Table 11.7 reveals a number of interesting pat-terns. First, all these states look very different fromthe rest of the country in terms of their potentialfor and use of wind generation. Second,Texas andCalifornia stand out in terms of total wind gen-eration, together accounting for just over a thirdof the U.S. total. Third, there is enormous varia-tion in the extent to which the potential for windgeneration is exploited, with the three PacificCoast states generating much larger percentages oftheir estimated potentials than any of the WindBelt states. In the case of California, this is clearlyattributable to very generous early subsidies forrenewable generation. Oregon and Washingtonare also well known as green states, but they werenot early adopters of RPSs, nor do they have sub-

sidy programs with California-like levels of gener-osity. Part of the reason for their relatively inten-sive exploitation of their wind potential may bethat California utilities have taken advantage oftheir ability to meet their RPS obligations by pur-chasing NHR power generated elsewhere.

Within this group of states, the share of totalgeneration accounted for by wind also shows sub-stantial variation. Here, though, neither Texas northe Pacific states are outliers. On the high side,Iowa and Minnesota are the most reliant on windgeneration. Iowa was an early RPS adopter, butMinnesota was not. On the low side, both interms of reliance on wind generation and exploi-tation of wind potential, Nebraska, Wyoming,Montana, and North Dakota stand out. Amongthese four, only Montana had an RPS in 2007.Perhaps more important, all are relatively small interms of population and total electricity genera-tion, and all are distant from major populationcenters, so that the wind integration problems dis-

Table 11.7. Leading wind generation states

2007 Wind generation

State/regionEstimated windpotential (TWh) TWh

% of est.potential

% of totalgeneration Initial RPS year

North Dakota 1,210 0.62 0.05 1.99

Texas 1,190 9.01 0.76 2.22 1999

Kansas 1,070 1.15 0.11 2.30 2009

South Dakota 1,030 0.15 0.01 2.44

Montana 1,020 0.50 0.05 1.71 2005

Nebraska 868 0.22 0.03 0.67

Wyoming 747 0.76 0.10 1.65

Oklahoma 725 1.85 0.26 2.54

Minnesota 657 2.64 0.40 4.84 2007

Iowa 551 2.76 0.50 5.65 1983

Colorado 481 1.29 0.27 2.40 2004

New Mexico 435 1.39 0.32 3.87 2007

Wind Belt total 9,984 22.33 0.22 2.58

California 59 5.59 9.47 2.65 2002

Oregon 49 2.44 2.56 2.28 2006

Washington 37 1.25 6.52 2.26 2007

Pacific total 161 9.27 6.39 2.49

Total Wind Belt + Pacific 10,129 31.60 0.31 2.55

Rest of U.S. 648 2.85 0.44 0.10

Sources: Estimated potentials are from AWEA n.d. and, for Washington and Oregon, personal communication with AWEA. Generation dataare from EIA 2009c, and RPS dates are from Table 11.4

Renewable Electricity Generation in the United States 225

cussed below are unusually difficult. More gener-ally, the Wind Belt suffers from the fact that only7% of the U.S. population lives in the top 10 statesfor wind potential (NERC 2009a).

As mentioned earlier, the rapid growth andsignificant penetration of wind power are raisingimportant issues in power system design andimplementation. Wind power is intermittent,meaning both that the output of wind-poweredgenerators is variable over time and that it isuncertain—it cannot be perfectly forecast. Theoutput of hydroelectric facilities is also variableand uncertain because rainfall cannot be perfectlyforecast, but these features manifest themselvesover timescales of months or years. The output ofwind facilities is variable and uncertain overtimescales of hours; the output of solar facilitiescan vary from minute to minute as clouds pass byoverhead. Because residential, commercial, andindustrial demands also vary over these timescales,at small levels of penetration wind power poses nonew issues. At large scales, however, an influentialindustry group has concluded that “reliably inte-grating high levels of variable resources [includingwind and solar] into the North American bulkpower system will require significant changes totraditional methods used for system planning andoperations” (NERC 2009a).20

Perhaps the most talked-about manifestationof the variability of wind power took place inTexas on February 26, 2008 (ERCOT 2008;Galbraith 2008; Grant et al. 2009, 51–52). As theevening electricity load was increasing, wind gen-eration dropped from over 1,700 MW to 300MW in a three-hour period because wind speedsdecreased. This was roughly equivalent in magni-tude to a single large fossil-fuel generating unitgoing offline—not a rare event—and it happenedgradually. (Moreover, commercial forecasts, notavailable to the grid operators, had predicted thefall in wind speeds.) This emergency, which wasexacerbated by the unforeseen unavailability ofsome fossil-fuel-fired capacity, was mainly han-dled by curtailing service to large industrial andcommercial users that had contracted for inter-ruptible power, reducing system loads by 1,100MW within a 10-minute period. Although windprovides only a few percent of generation inTexas

on average, winds in the western part of the stategenerally blow the strongest at night, whendemand is low. As a result, wind can provide morethan 10% of power in Texas on some occasions,making its variability a potentially more seriousissue. In fall 2008, the operators of the Texas gridincreased backup power requirements, particu-larly at night.

With wind generation in Texas expected tomore than double by 2015, wind’s variability isprojected to be a more serious concern going for-ward, and more flexible gas-fired capacity willlikely be needed. Since power prices tend to below at night, however, and the gas-fired units usedto provide backup power tend to have high mar-ginal costs, some observers worry that the Texasmarket may not provide sufficient backup capacity(see O’Grady 2009 and Chapter 9 of this vol-ume).21

Similar issues have confronted the BonnevillePower Administration (BPA), which operates thegrid in the Pacific Northwest with about 41,000MW of peak generating capacity. During one24-hour period in December 2008, BPA receivednear-zero levels of wind generation early and latein the day, punctuated by nearly 1,600 MWaround midday. Between January 5 and 14, 2009,wind output varied from 500 MW to 1,500MW—followed by two weeks of zero output. Inpart, this high degree of variability has arisenbecause most of the relevant capacity is geo-graphically concentrated along the Washington–Oregon border—just as much of Texas’s windgeneration capacity is concentrated in the westernportion—and thus wind conditions are highlycorrelated within the generating fleet (Riner2009, 8).22

As wind power gains in importance, it willbecome more important to enhance the flexibilityof the overall electric power system (Grant et al.2009; IEA 2008b; Milligan et al. 2009; NERC2009a). The basic methods for doing this are wellknown. One can, in principle, increase demandresponsiveness to supply conditions and go wellbeyond what was possible in the Texas emergencydiscussed above. A variety of institutional, regula-tory, and technological barriers make this far fromstraightforward, however. On the supply side, one

226 Richard Schmalensee

can add generating units that can increase ordecrease output rapidly (mainly gas-fired undercurrent technology) or add grid-level storage(generally too expensive with current technol-ogy). Alternatively, one can use transmission andoperational integration to create large power sys-tems, taking advantage of the fact that “the corre-lation between production from multiple windplants diminishes as those plants are geographi-cally farther apart” (Kirby and Milligan 2008, 3).It is often difficult to get permission to buildtransmission capacity in this country, however,and geographic averaging is inherently expensivein areas that are thinly populated and distant frommajor load centers, like many of the states in theWind Belt.

In addition, some utilities (and their regula-tors) have been reluctant for a variety of reasons tojoin the large integrated regional systems operatedby regional transmission operators (RTOs) andindependent system operators (ISOs) that cur-rently meet about two-thirds of U.S. electricitydemand (IRC 2007). The open, flexible whole-sale electricity markets operated in those systemshelp integrate wind generation economically andreliably. In light of these advantages, it is not sur-prising that as of the end of 2007, 74% of U.S.installed wind generating capacity was located inISO and RTO regions, even though those areashad only 44% of the nation’s wind energy poten-tial (Kirby and Milligan 2008). Nor is it surprisingthat Minnesota and Iowa, which have the highestlevels of wind energy penetration (see Table 11.7)and do not appear to have experienced seriousoperational problems, are part of the MidwestISO. So is North Dakota, but this state has notactively promoted renewable generation, and it isdistant from the major load centers in this ISO.On the other hand, Nebraska and Wyoming,which make conspicuously little use of theirabundant wind resources, are not in an organizedregional market, and neither is most of Montana.

Finally, a number of observers have com-mented on the fact that spot electricity prices inwestTexas are often negative, particularly at night,when the wind is strongest and demand is lowest(Lively 2009; Wang 2008; see also Lawhorn et al.2009, 86–87). This generally happens when the

power lines connecting the wind generators in thewest to the major load centers elsewhere in thestate are operating at capacity, and spot prices inthe rest of the state are positive. Negative priceswould induce conventional thermal generatingunits to shut down, unless the energy cost oframping up when demand rises would exceed thecosts of paying the grid operator to take power.Wind units do not generally incur ramp-up costs,and when there is an excess of power, it wouldcommonly be more efficient for wind units tocease production than for, say, a base load coalplant with significant ramp-up costs to do so.(There are currently no costs of CO2 emissions tofactor into this comparison in most of the UnitedStates.) But if a wind-powered generator remainsin operation, it earns both the federal REPTC of2.1 cents/kWh and the value of the renewableenergy credits to which its output entitles it. InTexas, it is apparently a better deal for wind gen-erators to pay the grid operator 4 cents/kWh totake their output than to shut down. In the case ofTexas, adding transmission capacity seems the bestway to deal with this problem, and the state plansto add more than 2,300 miles of new transmissioncapacity by 2015, about a 6% increase (ERCOT2009b; O’Grady 2009). More generally, outputsubsidies and feed-in tariffs can be expected toraise operating issues of this sort from time totime.

ConclusionsThough it is far from certain as this is written thatthe United States will soon adopt a cap-and-tradesystem to limit CO2 emissions from the combus-tion of fossil fuels, there seems a fair chance that itwill do so within, say, the next decade. Whetheror not this happens, and even though supportingthe deployment of NHR generation would makelittle economic sense if a CO2 cap were in place, itseems close to certain that, as in Europe, U.S.governments will nonetheless continue to supportboth the development and deployment of NHRtechnologies. And, absent a dramatic change inthe nature of support policies, it also seems closeto certain that wind will account for the domi-

Renewable Electricity Generation in the United States 227

nant share of new NHR generating capacity inthe United States for some years to come.

As the References section at the end of thischapter indicates, much attention is currentlybeing paid, in the United States and abroad, to thechallenges posed by large-scale integration ofwind and other intermittent generation technolo-gies. But it seems that almost no attention is beingdevoted to increasing the economic efficiency ofstate and federal policies that subsidize develop-ment and deployment of NHR technologies.Among the issues raised by the current U.S.regime and discussed above that deserve seriousthought are the following:

• Federal R&D support has tended to favorrelatively mature fossil and nuclear technolo-gies, and support for NHR technologies hasbeen far from steady over time.

• Channeling federal support for NHRdeployment primarily through the corporateincome tax disadvantages small firms withoutsubstantial income streams and complicatesproject financing generally.

• The variability of federal subsidy programsover time discourages investments in long-lived tangible and intangible capacity by pro-ducers of NHR generating equipment.

• Subsidies tied to the level of initial investmentprovide weak incentives for cost containmentor efficient and reliable operations.

• When transmission capacity is inadequate,output subsidies (and feed-in tariffs) can cre-ate perverse incentives to generate powerwhen it is not needed.

• The variation in state-level renewable port-folio standards almost entirely prevents inter-state trading, which would lower totalnational costs of meeting state-specific tar-gets.

• Although states have adopted a variety of dif-ferent RPS designs, little systematic analysishas been done of the performance implica-tions of these design differences.

• Federal policy toward the electric power sec-tor has ignored the evidence that organizedISO/RTO wholesale power markets facili-tate deployment of NHR generation.

• The myriad state and federal (and other)NHR programs and policies confront awould-be NHR generator with unproduc-tive complexity.

The last of these issues is not likely to go away, ascomplete preemption of state and local authorityin this area is hard to imagine. Nor would theideal way forward be simply to adopt some othernation’s policy regime. Other chapters in this vol-ume make it clear that no perfect solution to theproblem of efficiently subsidizing NHR deploy-ment has yet been implemented. But as in otherareas of public policy, careful economic analysiscould at least increase the returns from invest-ments in deployment subsidies.

AcknowledgmentsThe author is indebted to Dhiren Patki for superbresearch assistance and Boaz Moselle for excep-tionally useful comments on an earlier draft. Thischapter also benefited from the author’s participa-tion in the MIT Future of Solar Energy study, inwhich Kevin J. Huang provided excellent researchassistance.

Notes

1. According to the International Energy Agency(IEA 2008c), which bases its U.S. figures on datasubmitted by the EIA, solar-thermal installationsaccounted for 97.4% of 2007 U.S. solar-electricgeneration, whereas according to the InterstateRenewable Energy Council (Sherwood 2009),which measures all grid-connected solar electriccapacity, solar thermal units accounted for only49.2% of average solar generating capacity in 2007(taking the average of solar thermal and photo-voltaic (PV) capacities at the end of 2006 and theend of 2007). If PV units had the same capacityfactor as solar thermal units on average, total U.S.solar generation was 98% above the IEA figures. Itwas at least 60% above the IEA figures if PV unitshad at least 62% of the capacity factor of solarthermal units, and this seems a conservativeassumption.

228 Richard Schmalensee

2. Two clarifying observations are in order. First, theIEA’s data are based on reports submitted to it bynational statistical agencies such as the EIA, so theIEA understates U.S. solar generation just as theEIA does. Second, the IEA reports gross electri-city generation, including estimated in-plantlosses, whereas all EIA numbers are for net genera-tion. Net generation is about 5% below gross gen-eration for the United States as a whole.

3. Estimates of the state- and technology-specificpotentials have been published by the Union ofConcerned Scientists (UCS; see Deyette et al.2003). The UCS estimates did not take intoaccount costs, however, and they imply totalnational potential roughly 20 times as large as theIEA’s and with a very different pattern across tech-nologies. The UCS estimate of solar potential was3 times as large as that of biomass potential, andestimated wind potential was 19 times as large.

4. The remainder (10.9%) was devoted to end-usetechnology.

5. At least 16 of the 30 OECD nations have feed-intariffs, as do 7 of the 8 European Union (EU)member states that are not OECD members.

6. Only seven EU members employ this device,three of which also use feed-in tariffs. Twonon-EU OECD nations also employ this mecha-nism. (Coenraads et al. 2008; EREC 2007; Euro-pean Commission 2008; IEA 2009)

7. Conventional generation assets are depreciatedusing 150% declining balance over 15 years, withthe option to switch to straight-line depreciationat any point. Qualifying NHR generating assetscan be depreciated using 200% declining balanceover 5 years, with the same option. The figure inthe text was computed by assuming the switchoccurs when it is most profitable and employing a10% discount rate. For more details, see Metcalf(2009a, b).

8. Except where noted in the following discussion,all information on state and federal policies is fromDSIRE (2009a).

9. Rabe (2007) suggests that these policies are popu-lar because they are seen to rely on market forces.A more cynical assessment would be that theircosts are less visible than those of other subsidytypes.

10. A good source of information on this initiative isits website, www.rggi.org/home.

11. For an interesting discussion on this topic, seeTaylor (2008).

12. The surcharge has been estimated to have addedbetween 2% and 3% to the price of a kWh. Theproceeds were aimed at production credits forbiomass, wind, solar, geothermal, and smallhydroelectric facilities.

13. Like the Emerging Renewables Program, fundingfor solar and PV projects for the SGIP was takenover by the California Solar Initiative in 2007.

14. As noted above, Texas has given a premium fornonwind renewable generation since 2005(DSIRE 2009a).

15. Observers have also noted that the use of a zonalpricing system within Texas, rather than a nodalpricing system, contributed to inefficient use oftransmission capacity; see Potomac Economics(2009).

16. Connecticut is one example of a state where theRPS applies to such a small section of the electri-city market as to be rendered marginal (Langnissand Wiser 2003).

17. In contrast, in some states, considerable uncer-tainty exists around the size and scope of the RPS,as well as its end date. For instance, in Maine, theRPS is slated to be reviewed every five years, andin Connecticut and Massachusetts, the end date ofthe RPS is unclear.

18. These issues have also been encountered in othercountries with high levels of wind penetration.For an interesting discussion of the experience infour such countries, see Ackerman et al. (2009).

19. The AWEA’s total estimated U.S. potential is 36times the IEA estimate, indicating that the formerincludes sites that the latter does not deem eco-nomic by 2020. The AWEA numbers are usedhere on the assumption that they are correlatedwith potentials that would be estimated using theIEA’s assumptions and methods.

20. For illuminating recent discussions of wind inte-gration issues, see Grant et al. (2009) and Milliganet al. (2009).

21. A similar problem has recently been analyzedquantitatively in the German context; see Traberand Kemfert (2009).

22. For an interesting discussion and graphical depic-tion of wind variability, see NERC (2009b).

ReferencesAckerman, Thomas, Graeme Ancell, Lasse Diness

Borup, Peter Børre Ericksen, Bernhard Ernst, FrankGroome, Matthias Lange, Corinna Möhrlen, Antje

Renewable Electricity Generation in the United States 229

G. Orths, Jonathan O’Sullivan, and Miguel de laTorre. 2009. Where the Wind Blows. IEEE Power &Energy Magazine 7: 65–75.

AWEA (American Wind Energy Association). 2009.Annual Wind Industry Report: Year Ending 2008.www.awea.org/publications/reports/AWEA-Annual-Wind-Report-2009.pdf (accessed February22, 2010).

———. No date. Top 20 States with Wind EnergyResource Potential. www.awea.org/newsroom/pdf/Top_20_States_with_Wind_Energy_Potential.pdf(accessed February 22, 2010).

BMU (Federal Ministry for the Environment, NatureConservation and Nuclear Safety). 2007. EEG—The Renewable Energy Sources Act:The Success Story ofSustainable Policies for Germany. Berlin: BMU.www.gtai.com/uploads/media/EEG_Brochure_01.pdf (accessed September 3,2009).

Borenstein, Severin, and James Bushnell. 2000. Elec-tricity Restructuring: Deregulation orReregulation? Regulation 23: 46–52.

Carley, Sanya. 2009. State Renewable Energy Electri-city Policies: An Empirical Evaluation of Effective-ness. Energy Policy 37: 3071–3081.

CEC (California Energy Commission). 2009. Emer-ging Renewables Program. www.energy.ca.gov/renewables/emerging_renewables/index.html(accessed July 17, 2009).

Coenraads, R., G. Reece, C. Klessmann, M. Ragwitz,A. Held, G. Resch, and C. Panzer. 2008. RenewableEnergy Country Profiles: Final Version February 2008.Utrecht, Netherlands: EU Commission.

Corey, Karlynn, and Blair Swezey. 2007. RenewablePortfolio Standards in the States: Balancing Goalsand Rules, Electricity Journal 40: 21–32.

Deyette, Jeff, Steve Clemmer, and Deborah Donovan.2003. Plugging in Renewable Energy: Grading the States.Washington: Union of Concerned Scientists.

DOE (U.S. Department of Energy). 2008. 20% WindEnergy by 2030: Increasing Wind Energy’s Contri-bution to U.S. Electricity Supply. DOE/GO-102008–2567. Washington, DC: DOE.www.20percentwind.org/20percent_wind_energy_report_revOct08.pdf(accessed February 22, 2010).

DSIRE (Database for State Incentives for Renewablesand Efficiency). 2009a. Project of the North Caro-lina Solar Center and Interstate Renewable EnergyCouncil funded by the U.S. Department of Energy’sOffice of Energy Efficiency and Renewable Energy.www.dsireusa.org (accessed February 22, 2010).

———. 2009b. California Feed-In Tariff.www.dsireusa.org/incentives/incentive.cfm?Incentive_Code=CA167F&re=1&ee=1(accessed February 22, 2010).

———. 2009c. California Solar Initiative.www.dsireusa.org/incentives/incentive.cfm?Incentive_Code=CA134F&re=1&ee=1(accessed February 22, 2010).

———. 2009d. CEC: New Solar Homes Partnership.www.dsireusa.org/incentives/incentive.cfm?Incentive_Code=CA150F&state=CA&CurrentPageID=1&RE=1&EE=1 (accessedFebruary 22, 2010).

EIA (Energy Information Administration). 1999. Fed-eral Financial Interventions and Subsidies in Energy Mar-kets 1999: Primary Energy. Washington, DC: EIA.

———. 2007. Green Pricing and Net Metering Programs,2005. Washington, DC: EIA.

———. 2008a. Federal Financial Interventions and Subsi-dies in Energy Markets 2007. Washington, DC: EIA.

———. 2008b. Renewable Energy Annual 2007. Wash-ington, DC: EIA.

———. 2009a. Annual Energy Review 2008. Washing-ton, DC: GPO. www.eia.doe.gov/aer/overview.html (accessed February 22, 2010).

———. 2009b. Electric Power Annual 2007. Washing-ton, DC: EIA.

———. 2009c. State Renewable Energy 2007. Washing-ton, DC: EIA.

———. No date, a. Electricity. www.eia.doe.gov/fuelelectric.html (accessed November 14, 2009).

———. No date, b. International Energy Statistics.http://tonto.eia.doe.gov/cfapps/ipdbproject/IEDIndex3.cfm?tid=6&pid=29&aid=12 (accessedFebruary 22, 2010).

ERCOT (Electricity Reliability Council of Texas).2008. ERCOT Demand Response Program HelpsRestore Frequency FollowingTuesday Evening GridEvent. www.ercot.com/news/press_releases/2008/nr02-27-08 (accessed February 22, 2010).

———. 2009a. ERCOT Expects Adequate PowerSupply for Summer: Update. www.ercot.com/news/press_releases/2009/nr05-29-09 (accessedFebruary 22, 2010).

———. 2009b. ERCOT Quick Facts.www.ercot.com/content/news/presentations/2009/ERCOT_Quick_Facts_February_2009.pdf(accessed February 22, 2010).

EREC (European Renewable Energy Council). 2007.National Policy: Overview of EU Member States.www.erec.org/policy/national-policy.html (accessedSeptember 3, 2009).

230 Richard Schmalensee

European Commission. 2008. Commission Staff WorkingDocument: The Support of Electricity from RenewableSources. SEC(2008) 57. Brussels: EU Commission.http://ec.europa.eu/energy/climate_actions/doc/2008_res_working_document_en.pdf (accessedFebruary 22, 2010).

FERC (Federal Energy Regulatory Commission).2009. Electric Power Markets: New England (ISO-NE). www.ferc.gov/market-oversight/mkt-electric/new-england.asp#prices (accessed November 14,2009).

Galbraith, Kate. 2008. Texas Adjusts Its Grid for Wind.NewYorkTimes, Nov. 13.http://greeninc.blogs.nytimes.com/2008/11/13/texas-adjusts-its-grid-for-wind/ (accessed February 22,2010).

GAO (U.S. Government Accountability Office). 2007.Federal Electricity Subsidies: Information on ResearchFunding,Tax Expenditures, and Other Activities that Sup-port Electricity Production. Washington, DC: GAO.

Grant, William, Dave Edelson, John Dumas, JohnZack, Mark Ahlstrom, John Kehler, Pascal Storck,Jeff Lerner, Keith Parks, and Cathy Finley. 2009.Change in the Air. IEEE Power & Energy Magazine 7:47–58.

Hirsh, Richard F. 1999. PURPA:The Spur to Compe-tition and Utility Restructuring. Electricity Journal12: 60–72.

IEA (International Energy Agency). 2008a. DeployingRenewables: Principles for Effective Policies. Paris: IEA.

———. 2008b. Empowering Variable Renewables. Paris:IEA. www.iea.org/g8/2008/Empowering_Variable_Renewables.pdf (accessed February 22,2010).

———. 2008c. Renewables Information 2008. Paris:IEA.

———. 2009. Global Renewable Energy: Policies andMeasures. www.iea.org/textbase/pm/?mode=re(accessed September 3, 2009).

IEEE (IEEE Global History Network). 2008a. Mile-stones:Vulcan Street Plant, 1882. www.ieeeghn.org/wiki/index.php/Milestones:Vulcan_Street_Plant,_1882 (accessedSeptember 3, 2009).

———. 2008b. Pearl Street Station.www.ieeeghn.org/wiki/index.php/Pearl_Street_Station (accessed September 3, 2009).

IRC (ISO/RTO Council). 2007. Progress of Organ-ized Wholesale Electricity Markets in NorthAmerica: A Summary of 2006 Market Data from 10ISOs & RTOs. www.isorto.org/atf/cf/{5B4E85C6–7EAC-40A0–8DC3–

003829518EBD}/IRC_State_of_the_Markets_Report_103007.pdf(accessed February 22, 2010).

Kirby, B., and M. Milligan. 2008. Facilitating WindDevelopment:The Importance of Electric Industry Struc-ture. NREL Technical Report NREL/TP-500–43251. NREC.

Langniss, Ole, and Ryan Wiser. 2003. The RenewablesPortfolio Standard in Texas: An Early Assessment.Energy Policy 31: 527–535.

Lawhorn, John, Dale Osborn, Jay Caspary, BradleyNickell, Doug Larson, Warren Lasher, and ManzarEa. Rahman. 2009. The View from the Top. IEEEPower & Energy Magazine 7: 77–88.

Lively, Mark B. 2009. Renewable Electric Power—TooMuch of a Good Thing: Looking at ERCOT.USAEE Dialog 17 (2): 21–26.

Menz, Frederic C., and Stephan Vachon. 2006. TheEffectiveness of Different Policy Regimes for Pro-moting Wind Power: Experiences from the States.Energy Policy 34: 1786–1796.

Metcalf, Gilbert E. 2009a. Investment in Energy Infra-structure and the Tax Code. Tufts University, Depart-ment of Economics.

———. 2009b. Tax Policies for Low-Carbon Tech-nologies. NationalTax Journal 62: 519–533.

Milligan, Michael, Kevin Porter, Edgar DeMeo, PaulDenholm, Hannele Holttinen, Brendan Kirby,Nicholas Miller, Andrew Mills, Mark O’Malley,Matthew Schuerger, and Lennart Soder. 2009. WindPower Myths Debunked. IEEE Power & EnergyMagazine 7: 89–99.

NERC (North American Electric Reliability Corpora-tion. 2009a. Accommodating High Levels ofVariableGeneration. www.nerc.com/files/IVGTF_Report_041609.pdf (accessed February 22,2010).

———. 2009b. How BPA [Bonneville Power Admin-istration] Supports Wind Power in the PacificNorthwest. www.nerc.com/docs/pc/riccitf/BPA_supports_wind_power_for_the_Pacific_Northwest_Mar_2009.pdf (accessed February 22, 2010).

O’Grady, Eileen. 2009. Texas Power Suppliers FacingWind Challenges. www.reuters.com/article/idUSTRE59701B20091008 (accessed February 22,2010).

Potomac Economics. 2009. 2008 State of the MarketReport for the ERCOT Wholesale Electricity Mar-ket. www.puc.state.tx.us/wmo/documents/annual_reports/2008annualreport.pdf (accessed Feb-ruary 22, 2010).

Renewable Electricity Generation in the United States 231

Rabe, Barry. 2007. Race to the Top: The ExpandingRole of U.S. State Renewable Portfolio Standards.Sustainable Development Law and Policy 10: 10–16.

Riner, Bobette. 2009. BPA Struggling with Vagaries ofWind Power. Natural GasWeek, July 29.

Ritschel, Alexander, and Greg P. Smestad. 2003.Energy Subsidies in California’s Electricity MarketDeregulation. Energy Policy 31: 1379–1391.

Sherwood, Larry. 2009. U.S. Solar MarketTrends 2008.Interstate Renewable Energy Council (IREC).www.irecusa.org/fileadmin/user_upload/NationalOutreachDocs/SolarTrendsReports/IREC_Solar_Market_Trends_Report_2008.pdf(accessed March 19, 2010).

Taylor, Margaret. 2008. Beyond Technology-Push andDemand-Pull: Lessons from California’s Solar Policy.Energy Economics 30: 2829–2854.

Traber, Thure, and Claudia Kemfert. 2009. Gone withthe Wind? Electricity Market Prices and Incentives

to Invest in Thermal Power Plants under IncreasingWind Energy Supply. DIW Berlin Discussion PaperNo. 852. http://ssrn.com/abstract=1430905(accessed February 22, 2010).

Wang, Ucilia. 2008. Texas Wind Farms Paying Peopleto Take Power. Greentech Media, Dec. 10.www.greentechmedia.com/articles/print/texas-wind-farms-paying-people-to-take-power-5347/(accessed February 22, 2010).

Wiser, Ryan, and Galen Barbose. 2008. Renewable Port-folio Standards in the United States:A Status Report withData through 2007. Berkeley, CA: LawrenceBerkeley National Laboratory.

Wiser, Ryan, S. Picker, and C. Goldman. 1996. Califor-nia Renewable Energy Policy and ImplementationIssues—An Overview of Recent Regulatory and Legisla-tive Action. Berkeley, CA: Lawrence BerkeleyNational Laboratory.

232 Richard Schmalensee

12

The European Union’s Policy on theDevelopment of Renewable EnergyChristopher Jones

This chapter discusses policy in the EuropeanUnion (EU) toward renewable energy. It

begins with a general history of the developmentof EU policy, then looks more specifically at themain elements of the Renewable Energy Direc-tive adopted by the EU at the beginning of 2009.This is followed by a discussion of future energytargets, as well as issues and challenges the EU willface in meeting them. Finally, grid and energysecurity issues are examined.

The Gradual Development of anEU Policy on Renewable EnergyAlthough it is generally considered that the Euro-pean Union’s push toward developing an ambi-tious policy to support the growth of renewableenergy is recent, in fact this development began asearly as 1996, when, in a Green Paper on renew-able energy, the European Commission suggestedan indicative target for a share of renewableenergy of 12% for 2010, equivalent to a doublingof share in the EU at the time (European Com-mission 1997). It is interesting to note that thereasons why the commission proposed this targetfocused on the issue of energy security (the EUwas steadily becoming more dependent on

imported energy from Russia and the MiddleEast) at least as much as, if not more than, envi-ronmental issues.

This target was endorsed by the EuropeanCouncil in its resolution on the matter from June1997, stating that a target of 12% renewableenergy by 2010 “is ambitious and could give use-ful guidance for increased efforts at Communityas well as at Member State level” (EuropeanCouncil 1997). The European Parliament, for itspart, proposed a goal of 15% and also called on thecommission to submit specific measures, includ-ing the setting of targets per member state.

The next step was the commission’s WhitePaper of 1997,1 which argued that “an indicativetarget is a good policy tool, giving a clear politicalsignal and impetus for action.” At the time, how-ever, no attempts were made to divide the targetamong the member states, and the target was notgiven a legal status in community legislation. TheWhite Paper went no further than to suggest that“targets in each Member State could stimulate theefforts towards increased exploitation of the avail-able potential” and that it would be importantthat each define its own strategy and “within itpropose its own contribution towards the overall2010 objective” (European Commission 1997,10).

The White Paper led to the proposal and sub-sequent adoption on September 27, 2001, of the

Renewable Electricity Directive (2001/77/EC),which laid down indicative minimum targets forthe share of renewable energy in electricity gen-eration in each member state and introduced pro-visions that did the following:

• obliged member states to introduce guaran-tees of origin for renewable electricity, withsome (rather general) requirements regardingtheir format and operation;

• required member states to actively considerhow to reduce administrative burdens appli-cable to production plants for renewableelectricity;

• set forth rather general requirements regard-ing grid connection and operation in anattempt to eliminate any potential discrimi-nation that renewable electricity producersmight face in getting access to the grid,including a provision stating that memberstates “may also provide for priority access tothe grid system of electricity produced fromrenewable energy sources” and that “whendispatching generating installations, transmis-sion system operators shall give priority togenerating installations using renewableenergy sources insofar as the operation of thenational electricity system permits”; and

• laid down various reporting obligations onboth member states and the commission.

Although the text of the final directive seems toprovide very little in terms of concrete and legallybinding obligations on member states, a carefulreading of its Article 3(2) makes it clear that mem-ber states have to take “appropriate steps” toachieve the targets given in the Directive andpublish a report setting out how they intended toachieve this. Logically, therefore, if such a reportwas clearly inadequate in terms of the measuresproposed in order to meet the target, the memberstate would have failed to have met its obligationin 3(2). Indeed, the commission has commencedinvestigations into certain member states for fail-ing to comply with this provision of the direc-tive.2

However, it is clear that the burden of proofwas on the commission to demonstrate that the

member state’s plan could not have any reasonablechance of meeting its 2010 target. In fact, variouscommission reports indicated that the EU was notlikely to meet its overall target of 21% of the EU’stotal electricity supply coming from renewablesources by 2010. A result of 19% was consideredmore likely.

The second piece of legislation flowing fromthe White Paper was the 2003 “Biofuels” Direc-tive (2003/30/EC), which required memberstates to set targets for renewable energy in trans-port in 2005 and 2010, taking the reference valuesof 2% and 5.75% into account. The commissionbegan infringement proceedings against ninemember states that initially failed to properly fulfilthe first of these obligations. All subsequently settargets for 2005, 2010, or both that the commis-sion considered appropriately ambitious in rela-tion to the requirements of the directive. The textdoes not, however, require member states toachieve the targets they have set themselves. Aswith the Renewable Electricity Directive, com-mission reports have indicated that many memberstates, and thus the EU as a whole, are unlikely tomeet their targets by 2010. A result of 4.2% hasbeen considered more likely (European Commis-sion 2009a).

Following the period when these directiveswere adopted, the EU’s energy position changedconsiderably. Although groundbreaking at thetime, it rapidly became clear that these measureswere inadequate to meet the needs of Europe’senergy and climate policy for the 21st century.Three developments in particular can be identi-fied as leading to this change.

First, the price of oil increased from $23/barrel in 2001, when the Renewable ElectricityDirective was adopted, to between $50 and $65/barrel in the period 2005–2007, then to $126/barrel in June 2008. Second, it was becomingmore and more evident that the EU was becom-ing overwhelmingly dependent on imports, espe-cially of oil and gas, for its energy; and higherprices and volatility of imports were increasinglyviewed as a threat to the EU’s prosperity. Finally,the EU’s heads of state and government becameconvinced that climate change was such a threatthat it demanded immediate and determined

234 Christopher Jones

action by the EU, even though much of the rest ofthe world was failing to take equivalent action,notably in the case of the U.S. Bush administra-tion.

The European Council at Hampton Court inOctober 2005 discussed the need for the EU todevelop a new European energy policy to meetthree core objectives: sustainability, security ofenergy supply, and competitiveness. It invited thecommission to put forward a blueprint forchange. At that time, the price of oil had recentlydoubled over two years, and a gas dispute betweenRussia and the Ukraine looked increasingly likely.Indeed, such a dispute erupted shortly after thecouncil, in December 2005, leading to limitedinterruptions to the EU’s gas supplies. In the lightof these developments, the council invited thecommission to propose a new energy policy forEurope.

The commission presented its First StrategicEU Energy Review in January 2007, titled “AnEnergy Policy for Europe” (European Commis-sion 2007b). The key elements of this reviewappear in the appendix at the end of this chapter.The review explains the EU’s rationale for all sub-sequent changes in its energy policy and thendetails the concrete proposals, which were laterendorsed by the EU’s heads of state and govern-ment, as well as the European Parliament.

In essence, the review unequivocally acceptsthe need for the EU to act to limit the globaltemperature increase to 2°C compared withpreindustrial levels, and it identifies and acceptsthe existence of growing threats to the EU’senergy security. It considers that dealing withthese challenges could provide a major opportu-nity for the EU in terms of competitiveness, jobs,and growth, and then proposes the following stra-tegic energy objectives for the EU:

• an EU objective in international negotia-tions of 30% reduction in greenhouse gasemissions in developed countries by 2020compared to 1990. In addition, 2050 glo-bal GHG emissions must be reduced by upto 50% compared to 1990, implying reduc-tions in industrialised countries of 60–80%by 2050;

• an EU commitment now to achieve, in anyevent, at least a 20% reduction of green-house gases by 2020 compared to 1990.

In order to achieve these goals, the commissionproposed the revision of the emissions trading sys-tem to bring it into line with the 20% objective by2020 and to improve its functioning; the agree-ment of an objective for the EU to improve itsenergy efficiency by 20% by 2020 (meaning that,if achieved, the EU would reduce its energy con-sumption by some 13% compared with 2005 lev-els); and finally, that the EU should agree on a20% target for renewable energy in its total energymix (compared with about 8.5% in 2005). Theseproposals have become known as the “20-20-20”package.

It is worth reflecting on quite how ambitiousthis package of measures was and remains. First,20% greenhouse gas cut by 2020—a unilateralcommitment irrespective of action outside theEU—would be far from simple to achieve andwould require difficult actions. Second, the 20%renewable energy objective should be seen in lightof recent developments in the EU’s energy mix.Of the 8.5% renewable energy in the EU’s energymix at the time, some 7% came from biomass andhydro, established for purely commercial reasons.The additional 1.5% from wind, PV, solar, andother renewables had been added over the courseof a decade and meant that the EU had been byfar the world leader in installing new renewablecapacity. In essence, therefore, the 20% targetmeant that in just 11 years, the EU would committo installing more than 10 times the renewablecapacity that it had achieved over the previousdecade. Or to put it another way, every year overthe next decade, the EU would install the samerenewable energy capacity that it had achievedover the entire last decade. Furthermore, it wasproposed that this 20% target would not simply bea vague political commitment, but would beenforced thorough legally binding commitmentsplaced on every single EU member state.

The first significant step toward the agreementon the “new directive” referred to in the commis-sion’s Energy Review—the future RenewableEnergy Directive—was taken by the European

The European Union’s Policy on the Development of Renewable Energy 235

Council of March 2007, under the German presi-dency and a very determined Chancellor AngelaMerkel. The chancellor considered it vital for theEU’s credibility in ongoing attempts to promotean international agreement on climate changethat a real commitment to the commission’s pro-posals be reached at the council.

The council formally endorsed the commis-sion’s “20-20-20” proposal, including the princi-ple of legally binding targets for renewable energyfor each member state. It invited the commissionto submit concrete proposals to make the “20-20-20” approach a reality. In addition, it agreed withthe commission’s proposal that 10% of all trans-port fuels must come from renewable energysources by 2020 in every member state.3 It isprobably fair to say that the commission’sapproach regarding the overall renewable energytargets, asking the council to endorse the princi-ple and the overall EU target while deferring thequestion of how the target would later be sharedamong the member states, facilitated agreement.

The commission proposed a package of meas-ures on January 23, 2008, including a revision ofthe EU Emission Trading System (ETS) for theperiod 2013–2020 so that it would achieve cutscommensurate with the EU’s political commit-ments to reduce greenhouse gases by 2020 and toinclude the principle of auctioning allowancesrather than free allocation (with some excep-tions); legally binding greenhouse gas targets onmember states for non-ETS sectors; and a newRenewable Directive.

It is worth highlighting the fact that the EUhas taken the approach of having a specific targetfor renewable energy in addition to binding meas-ures on ETS and non-ETS sectors. It is true thatsome argued that separate renewable energy tar-gets were unnecessary, even counterproductive.Given the existence of an ETS regime that wouldcover the electricity sector, surely, it was argued,this should set the level of renewable energy com-pared with nuclear, carbon capture and storage(CCS), and so forth. Indeed, according to themodeling work undertaken by the commission inpreparing its 20-20-20 proposal, the ETS systemwould be expected to produce about 14% of

renewable energy in the EU’s energy mix by2020, without the need for further action.

However, the member states took the viewthat such an approach would not meet the EU’swider energy policy aims, for several reasons.First, the EU considered that it was in its interestsin terms of energy supply security to ensure that agreater proportion of its energy was generatedindigenously, via renewables. The EU already hadthe position of world leader in renewable energyinstallations, notably wind and PV, and this wasproviding considerable employment in the EUand creating an important export industry.According to Energy Commissioner AndrisPiebalgs at a technology conference in Stockholmin October 2009:

Sixty per cent of the world’s wind capacity wasinstalled in Europe at the end of 2007, andEuropean companies had a global market share of66 per cent of turbine sales … Our low-carbonenergy industry today, which exists quite simplybecause we have invested where otherswatched, has produced 1.4 million jobs, a figurethat could double by 2020, and exports of avalue of 3.7 b€ … We are exporting our 3, and5 MW windmills. These probably would nothave been developed in the EU were it not for ourdetermination to take a risk by supporting thesetechnologies before they were mature. We areexporting our leading-edge PV panel systems, ourbio-gas plants, our biofuel expertise, and our con-centrated solar technologies.

Second, there was real concern that the ETS sys-tem alone would not provide the necessary incen-tives to maintain a thriving renewable energyindustry in the EU, and many doubted that itwould in reality lead to even 14% of renewableenergy in the EU’s energy mix by 2020. Thesedoubts followed from the EU’s experience in sup-port schemes for renewable electricity to date.Three types of support schemes have been operat-ing in the EU:

• Tendering schemes, where a governmentdetermines how much renewable electricity(or any other form of renewable energy) it

236 Christopher Jones

wants and tenders for the company willing toconstruct and operate the capacity for thelowest subsidy.

• Green certificates, where a governmentplaces an obligation on electricity retailers toensure that a minimum percentage of theirsales are renewable in origin. They areobliged to acquire enough green certificatesto demonstrate that this minimum level isachieved, and producers of renewable elec-tricity are allocated certificates for their pro-duction, which they can then sell. The priceis thus set by the market, based on the cost ofthe marginal certificate necessary to meet therequired level of demand.

• Feed-in tariffs, where a government decideswhat level of subsidy should be given to thedifferent forms of renewable energy, and pro-ducers are paid this subsidy for each unit pro-duced (such as kilowatt-hours). It is usuallyelectricity distributors that are given an obli-gation to purchase and then resell to finalcustomers the renewable electricity producedin their catchment areas, and to pay thefeed-in tariff subsidy to the generator, finan-cing it through a customer charge on the dis-tribution tariff. Normally a mechanism existsto compensate distributors in areas that have agreater level of renewable energy than others.One advantage of such a system is that it ena-bles a government to give different subsidiesto different technologies, such as a lowerprice to onshore wind than to photovoltaics(PV) or offshore wind, eliminating windfallprofits that might result when prices are setby the marginal, most expensive unit of pro-duction.

The commission has carefully followed the rela-tive success of each system, as have many others.In principle, one would expect that the tenderingoption and green certificate system would lead tothe lowest prices because of the competitionamong producers that would result. However, thishas not proven to be the case. Perhaps because ofthe immature nature of green certificate marketsthat have operated to date in the EU, it appearsthat the uncertainty regarding future price levels

that is inherent in a green certificate system (giventhe speed of technological development, it is nor-mal to assume that the price will reduce overtime), the risk premium demanded under suchsystems appears to be high. In any event, greencertificate systems have typically led to higherprices than have feed-in tariffs. In addition, greencertificate and tendering systems have had prob-lems in catalyzing the actual construction of thelevel of renewable energy that has been desired,apparently because of the difficulty in attractingfinancing to what is inherently a more riskyinvestment than a feed-in tariff system. Green cer-tificate schemes have never been introduced at thesize necessary to create a truly liquid market,however, nor with the long-term certaintyneeded for them to develop optimally.

Another reason why a feed-in tariff may becheaper than a green certificate system stems fromthe ability to differentiate among the costs of dif-ferent forms of renewable energy in the feed-intariff system, which is not typically a feature of agreen certificate system.4 One could argue that agreen certificate system based on the price of themarginal unit will result in pushing producers intomore of the cheaper technologies, where theymake more profit, thereby bringing down theprice. However, the ability to move into thesecheaper technologies is constrained by planningand environmental constraints; it is this elementthat differentiates renewable energy from othermarkets, where competition will always push pro-ducers to install cheaper production techniques.

An additional concern for green certificatesystems results from the speed of increase indemand that will result from the new RenewableEnergy Directive. Supply for renewable energycan be“lumpy” because of the difficulty in gettingplanning permission for new, large projects.Thereis genuine concern that supply will have difficultyin keeping up with exponentially growingdemand, enabling the price for the price-settingmarginal (and scarce) unit to be very high. This isnot an inherent feature of the feed-in tariff sys-tem. One answer to this problem is to have sepa-rate green certificates for different forms ofrenewable electricity, so that each will reach itsown competitive price level. However, this would

The European Union’s Policy on the Development of Renewable Energy 237

require that the government decide how much ofeach type of renewable energy it desires whensetting the green certificate requirements on dis-tributors, which is not necessary under thefeed-in tariff system.

These issues were considered vital whendetermining whether the EU should rely on theETS mechanism to support the development ofrenewable electricity, or whether it should con-tinue its approach with separate EU and nationalrenewable targets. As the ETS system is economi-cally the same as a green certificate system, somequestions remain as to its effectiveness on animperfect renewable electricity market, and itbrings the same level of risk, in terms of both costand its ability to actually deliver the amount ofrenewable energy needed.

At the end of the day, it is these reasons thathave led the EU to establish a separate renewableenergy target, partly a rational response to theEU’s emerging energy security vulnerability, andpartly an industrial policy consideration based onthe beliefs that the global market for renewablefacilities would continue to expand greatly overthe coming decades, and that the continuedgrowth of installed capacity in the EU wouldunderpin its industry’s continued major role.Indeed, this objective has been underlined by theEU’s recent focus on increasing its support toemerging low-carbon energy technology with theStrategic Energy Technology Plan and the Finan-cing Communication (European Commission2009b). In the Financing Communication, thecommission has set out a series of TechnologyRoad Maps identifying the concrete researchprojects that will need to be undertaken in the EUup to 2020 and proposing a doubling of the cur-rent levels of public support to such research anddevelopment (R&D).

The European Council and Parliament agreedon the 20-20-20 package, with its formal adop-tion taking place on December 11–12, 2008, lessthan one year after its presentation by the com-mission.The Renewable Directive has the follow-ing key elements:

• A community decision to ensure that by2020, 20% of the EU’s total energy needs will

come from renewable energy sources. It isnotable that the directive refers to renewableenergy as a proportion of the EU’s totalenergy consumption and not just its electri-city production, as was the case in the exist-ing Renewable Electricity Directive.

• Division of this target among the EU’s 27 EUmember states in the form of separate, legallybinding minimum renewable energy targetson a state-by-state basis. It is this legally bind-ing nature that marked the biggest departurefrom earlier renewable energy legislation.

• A method to permit one member state toinvest in the production of renewable energyin another member state or, under differentrules, in a third country, with the resultantenergy counting toward the investing mem-ber state’s target.

• Rules to overcome administrative barriers tothe development of renewable energy andensure access to the grid, in particular forelectricity from renewable sources.

• An obligation on all member states to ensurethat at least 10% of their energy needs forroad transport are met through renewableenergy sources. This does not mean a 10%biofuel obligation, as alternatives might beused, such as cars powered by renewable elec-tricity or renewably produced hydrogen. It isclear, however, that much of the 10% willcome from biofuels. The directive prescribesminimum sustainability criteria that must berespected for any biofuels counting towardthe target, to ensure that they effectively con-tribute to a real and important reduction incarbon dioxide compared with burning gasor diesel fuel.

The Main Elements of theNew DirectiveFollowing is a short summary of the new Renew-able Energy Directive adopted by the EU at thebeginning of 2009.5

238 Christopher Jones

Legally Binding National Targets

In reaching agreement on the directive, one of themost difficult issues was the division of the 20%target among the individual member states. As thetarget was to be legally binding in nature, if amember state fails to meet it, the commissionmay, through an infringement procedure, bringthe country in question before the EuropeanCourt of Justice for failing to meet its EU lawobligations. Given the fact that meeting the targetwould imply the granting of subsidies by memberstates, the division of the 20% was of centralimportance in the negotiations.

The European Council conclusions ofDecember 11–126 specifically referred to the needto set the national targets “with a view to sharingefforts and benefits fairly and equitably among allMember States, taking into account differentnational circumstances, starting points andpotentials.” The commission considered two basicapproaches for sharing the target among memberstates: sharing the effort on the basis of memberstates’ national resource potential; or sharing onthe basis of a flat-rate increase in the share ofrenewable energy (measured in percentage points)in each member state weighted by gross domesticproduct (GDP) per capita.

Under the first approach, each member state’srenewable energy resources and costs were esti-mated, together with a forecast for total energyconsumption in 2020. On the basis of this, aneconomic modeling exercise was carried out todetermine the effort sharing. This would result inlower cost for the EU, because the targets wouldreflect the potential for producing renewableenergy in each member state.

Under the second approach, each memberstate would have to increase its share of renewableenergy by a fixed number of percentage points, towhich a modulated contribution defined accord-ing to a set of objective criteria, most importantlyGDP per capita, would be added.This would leadto targets that would reflect the economicstrength of the member states rather than thepotential for increasing the share of renewableenergy.

Although both approaches obviously havemerit, at the end of the day the commission con-cluded that the latter approach would be appro-priate, finally assigning targets partly on a pro ratabasis, all member states contributing equally toachieve half the target, and partly on a GDP percapita basis. This mixed approach was consideredto be a simple and fair distribution of the effort, inwhich the wealth of the different member stateswould be reflected through a GDP per capita cri-terion. In detail, this approach was applied by thecommission as follows:

• First, the existing share of renewable energyas a percentage of the total energy mix wascalculated for each member state for 2005.The share was then adjusted to reflect effortsalready made: those member states whosegrowth in renewable energy was more than2% over 2001–2005 received a reduction of athird of that growth from the 2005 base yearshare (referred to as an “early-starter bonus”).In total, the EU needs to increase the share ofrenewable energy in its energy mix by 11% tomeet the 20% by 2020 target. For the pro ratapart of the target, 5.5% was added to eachmember state’s existing share of renewableenergy in 2005.

• For the remaining 5.5% of the EU’s energymix to be met by renewable energy, this wasallocated on the basis of a contribution perEU citizen, weighted by a GDP per capitaindex to reflect different levels of wealthacross member states, and multiplied by eachone’s population.

• The next step was to add together these twoelements to derive the full renewable energyshare of total final energy consumption in2020.

• Finally, a cap was introduced to ensure thatno target would be 50% or more of the totalnational energy mix for any member state.

The targets per member state, as adopted by theEuropean Council and Parliament, are shown inTable 12.1.

The European Union’s Policy on the Development of Renewable Energy 239

Table 12.1. Member state targets

Share of energyfrom renewablesources in grossfinal consump-tion of energy,2005

Target for shareof energy fromrenewablesources in grossfinal consump-tion of energy,2020

Belgium 2.2% 13%

Bulgaria 9.4% 16%

Czech Republic 6.1% 13%

Denmark 17.0% 30%

Germany 5.8% 18%

Estonia 18.0% 25%

Ireland 3.1% 16%

Greece 6.9% 18%

Spain 8.7% 20%

France 10.3% 23%

Italy 5.2% 17%

Cyprus 2.9% 13%

Latvia 32.6% 40%

Lithuania 15.0% 23%

Luxembourg 0.9% 11%

Hungary 4.3% 13%

Malta 0.0% 10%

The Netherlands 2.4% 14%

Austria 23.3% 34%

Poland 7.2% 15%

Portugal 20.5% 31%

Romania 17.8% 24%

Slovenia 16.0% 25%

Slovak Republic 6.7% 14%

Finland 28.5% 38%

Sweden 39.8% 49%

United Kingdom 1.3% 15%

In addition to these targets, the directive also pro-vides requirements regarding the trajectory thatmember states should respect over time. This isimportant. As mentioned above, the member statetargets are legally binding. In other words, eachmember state accepts a community law obligationto ensure that its share of renewable energy meetsthe above levels by 2020. Failure to do so wouldresult in legal sanction by the European Court forfailure to have met its obligation. However, if thecommission could take action only for failure of amember state to meets it obligation at the end of

the 2020 period, it would be a rather theoreticalexercise, as it would be too late for the country inquestion to do anything to meet the target at thatstage.

While in many respects the ideal approachwould have been to have had both a legally bind-ing end target and a binding minimum trajectory,it was felt that this did not give member statesenough flexibility. Thus an indicative trajectorywas agreed on, whereby member states shouldachieve 20% of their target by 2012 (i.e., 20% ofthe difference between the level of renewableenergy in their energy mix in 2005 and their 2020target) and 30% by 2014. It is true that this mini-mum trajectory is rather flat, and it was agreedupon on the basis that expected cost reductionsfor renewable energy, particularly for PV and off-shore wind, mean that it makes sense to focusmost efforts during the period 2015–2020.Achieving 30% of the target by 2014, however,would put member states in a reasonable positionby that date to make the final target a realisticproposition.

Member states are not subject to infringementproceedings solely for not meeting the minimumindicative trajectory. Regardless, this does notmake the trajectory ineffective. In combinationwith another obligation contained in Article 3(2)of the directive, which requires member states totake “measures effectively designed to ensure thatthe share of energy from renewable energysources equals or exceeds that shown in theindicative trajectory,” the indicative trajectorybecomes important. The explicit connectionbetween these two elements—the indicative tra-jectory and the effectively designed measures—was designed to allow the commission to use thetrajectory as a series of benchmarks for individualmember states when assessing compliance withthe obligation to take measures effectivelydesigned to increase the share of renewableenergy.

The directive does not define what kind ofeffectively designed measures will be required. Itis obvious that they will need to include appropri-ate renewable energy support schemes, as well asmeasures when necessary on grid connection,planning, and the availability of adequate finance.7

240 Christopher Jones

It therefore seems clear that the requirementfor member states to adopt effectively designedmeasures, combined with the indicative trajec-tory, will be effective in ensuring that the aims ofthe directive will be met in practice. Memberstates are obliged to submit national action plansto the commission on a regular basis, the first byJune 2010. These will be produced according toan agreed standard format, explaining the meas-ures that member states have taken and intend toadopt in order to ensure that they will meet theirfinal targets as well as their indicative trajectories.Where these plans manifestly put a member stateon a path whereby it will be unrealistic for it toachieve its final target, one may reasonably assumethat the commission could already bring an actionagainst the country in question before the Euro-pean Court of Justice, on the grounds that themember state in question had infringed Article3(2) of the directive, as it had not taken “measureseffectively designed to ensure that the share ofenergy from renewable energy sources equals orexceeds that shown in the indicative trajectory.”

In conclusion, therefore, the directive setsambitious, legally binding, and effectivelyenforceable targets for member states that shouldbe achieved in practice.

Trade in Renewable Energy Credits

Renewable energy is a good. The EU has beenbuilt on four fundamental freedoms: trade, labor,capital, and services. This is based on the beliefthat competition and the law of comparativeadvantage will push the production of goods andservices in the EU to the point where it is eco-nomically most beneficial to produce them. Thelaw of comparative advantage applies to few prod-ucts more strongly than renewable energy. Itmakes as much sense to put PV panels on housesin northern Europe when roofs in southernEurope have none as to produce olives in green-houses in Finland.

Logically, the EU should support renewableenergy where it is the cheapest to produce it, as infact the Common Agricultural Policy in principlesupports the production of food in the EU whereit is most advantageous to do so in climatic andsoil terms.

Nonetheless, the manner in which thenational targets have been allocated, as explainedabove, was in the end not based on where theeconomic potential for renewable energy in theEU exists. Roughly speaking, much of the bestrenewable potential in the EU is concentrated inthe south (sun), at the northern periphery (off-shore wind), and in the eastern member states(biomass). Allocating targets solely on the basis ofeconomic renewable energy potential in manycases would have placed the highest burden on thepoorest countries. The final compromise, allocat-ing targets on a split pro rata–GDP per capitabasis, means that targets are allocated irrespectiveof economic potential, and often countries withthe highest targets have relatively limited renew-able economic potential.

In such circumstances, it can clearly be arguedthat trade in renewable certificates, or guaranteesof origin (which the directive requires memberstates to issue to renewable energy producers onrequest in a standardized manner to enable mutualrecognition), should be a basic principle underly-ing the directive. Member states would meet theirnational targets by offering support to generatorsregardless of where they are located. Generatorsand producers would then establish themselveswhere it would be the cheapest to produce,meaning that national subsidy schemes would beas limited as possible.

It is with this in mind that the commissionoriginally proposed elements of a free tradeapproach for renewable energy. During negotia-tions with the council and parliament, however,this was rejected in favor of a more restrictedapproach, for two reasons. First, the market forrenewable energy is special in that it relies on sub-sidies for its existence (at least at present; costs arecoming down and the ETS is increasing the costof fossil-fuel-based generation). Many memberstates consider that their citizens would be lesslikely to be willing to pay extra for renewableenergy if it is produced and consumed, and thejobs are created, elsewhere. And second, a freetrade approach will not permit, or will make verydifficult, the contemporaneous existence of dif-ferent types of support systems in the memberstates. As discussed above, three types of support

The European Union’s Policy on the Development of Renewable Energy 241

system for renewable energy have been developedin the EU: green certificates, tendering, andfeed-in tariffs.

During the negotiations on the directive,member states with a feed-in tariff expressed theconcern that the existence of free trade wouldcause their system to implode. They argued thatwhere other member states operate green certifi-cate systems, the prices in those countries wouldbe set by the marginal units, likely to be PV oroffshore wind. This would be higher than theprice offered in feed-in tariff countries to cheaperforms of renewable energy, such as onshore wind,large-scale PV, or small hydro. If renewable gen-erators from feed-in tariff countries would be freeto sell into green certificate systems, these low-cost producers would obviously choose to exporttheir energy to higher-cost support systems.Feed-in tariff countries thus would be unable tomaintain their approach of differentiating theirsupport depending on the actual cost of produ-cing the different forms of renewable energy.

These arguments were considered to be per-suasive, and in the final text of the directive, amore restrictive form of trade in renewable creditswas introduced. As mentioned above, it makessense to envisage the possibility of one memberstate acquiring green credits from renewableenergy produced in another country, particularlyas some member states have high targets and rela-tively little indigenous renewable energy poten-tial. However, the need to envisage the possibilityof trade was made subject to the right of memberstates to limit support to renewable energy to thatproduced and consumed within its territory, andto limit trade by national producers.

The final approach thus allows a member stateto prohibit generators of renewable electricityfrom selling their production into supportschemes of other member states. Nevertheless, if amember state has excess renewable energy poten-tial, it may agree with another country to sell thatpotential.This could be done, for example, by theexporting country establishing its own supportscheme such that the amount of renewable energygenerated and consumed domestically exceededits national target, selling the resultant statisticaltransfers to another country, which could then

account for the latter’s target. Alternatively,through bilateral agreement, companies from theimporting member state might produce and sellrenewable electricity in the exporting memberstate but be subsidized directly by the importingcountry’s support scheme, with statistical transfersaccounting for the importing country’s target.

Similar arrangements are also possible withnon-EU countries, whereby an EU member statemay invest in major new renewable projects incountries with major untapped potential, such assolar in deserts, wind in Northern Africa, orbiomass in the Ukraine. In this case, however, therenewable energy must come from a facility con-structed after the date of entry into force of thenew directive, and the resultant electricity must bephysically imported into the EU. These con-straints on third-country projects underline thatone of the main objectives of the RenewableEnergy Directive is security of supply. The pro-duction of renewable energy will displace EUimports only if it is actually consumed within theEU.

Notwithstanding the existence of thesemechanisms for enabling trade, initial expecta-tions are that most member states will, at least inthe short to medium term, choose to supportdomestically produced renewable energy, andtrade will be the exception rather than the rule.

Other Policy ConsiderationsSeveral other questions have arisen regarding EUpolicy toward renewable energy, or more preciselyones that will inevitably become increasinglyimportant as the new Directive is implemented.These include the issue of whether the tradingsystem provided by the new legislation is adequateto enable member states with limited competitiverenewable potential to really import credits fromother member states or third countries, and whatshould be the EU’s longer-term approach toensuring the appropriate level of renewableenergy in its overall energy mix in order to meetits objectives of competitiveness, sustainability andsecurity.

242 Christopher Jones

Should the EU Move Toward a TradingSystem That Is More Open?

With a 20% target for renewable energy in theEU’s overall energy mix, it is expected that 35% ofthe EU’s total electricity supply will come fromrenewable energy sources, compared with 16%today. As discussed in more detail below, by 2050,one can expect a significantly higher percentagethan this, probably 60% or more, in the context ofthe action that will be needed to achieve an 80%cut in EU greenhouse gas emissions. If the cur-rent situation remains, with renewable energybeing produced nationally with national supportschemes, then this major part of the EU’s electri-city supply will be removed from the competitiveinternal energy market,8 making it so distorted asto be incapable of operating reasonably efficiently.

Potential medium-term solutions to thisproblem that would not destroy the effectivenessof the support schemes for renewable energy thathave worked well for many years might include aregional, and then an EU-wide, feed-in tariff.There are strong grounds to argue that workshould now commence on the development of anEU-wide renewable support mechanism thatwould enable renewable energy to be tradedacross borders as an integral part of the internalenergy market. This work should have certainpreconditions from the start, however, to ensurethat it does not result in a wait-and-see approachto investments in the meantime. For example, aguarantee could effectively be given that anyfuture system would not affect the supportschemes applicable to existing investments priorto the introduction of any new EU-wideapproach; in other words, they should have guar-anteed “grandfathering” rights. In addition, anynew approach must be as effective as the existingmechanisms presently operating in the EU interms of ensuring that renewable energy grows tomeet the EU’s targets.

Should the EU’s Approach to Importsfrom Third Countries Evolve over Time?

Already today, under the Renewable EnergyDirective, renewable energy produced outside the

EU, but resulting from new facilities constructedas a result of EU investment, can count toward amember state’s national target if that electricity isphysically delivered into the EU. Moving toward2050, it makes sense to reexamine this approachin the medium term, given the significantlygreater level of renewable energy that will berequired by the EU; the enormous cheap renew-able energy potential in neighboring countriesand the cost of building interconnectors and thelosses that result when transporting electricityover long distances.

For example, a regional wider Europeanrenewable energy market agreement could bepursued, whereby neighboring countries wouldagree to make some form of commitment regard-ing renewable energy development, and thiswould open the possibility for the EU to buildfacilities and receive the renewable energy, andgreenhouse gas (GHG), credits. This in turnwould enable the EU to use some of the gas andoil that otherwise would have been used in theseneighboring countries for the parts of itseconomy difficult to replace with carbon-freeenergy services, such as planes, heavy goods vehi-cles, and maritime vessels, and provide importantinward investment in these countries. Such adevelopment, based on the model of the EnergyCommunity, under which the countries of south-east Europe become part of the EU’s internalenergy market, clearly will take time to negotiateand would be far from simple. Given the mutualadvantage that would result, however, it is anobjective worth pursuing.

Further Renewable Energy Targetsto 2030 and 2050

It is evident that the EU’s climate change objec-tives do not stop at a 20% greenhouse gas reduc-tion by 2020. Indeed, the EU has already acceptedthat in the context of an international deal onclimate change it would accept a 30% cut by2030. On many occasions, national EU primeministers, chancellors, and presidents, as well asthe commission president, have spoken of an 80%cut by 2050. Others have even considered a 95%cut for the EU by 2050. Similar issues apply with

The European Union’s Policy on the Development of Renewable Energy 243

respect to energy security; the EU’s policies onenergy efficiency and indigenous energy produc-tion will continue as a form of insurance policyagainst price volatility.

The practical consequences of an 80% to 95%cut in greenhouse gas emissions for the EU haveto be seen in the context of the expected globalgrowth in energy demand over the next 40 years.Given the anticipated 40% to 50% increase in glo-bal population during this period, and theimproving standards of living in the developingworld, energy use and greenhouse gas emissionsare widely predicted to increase by 40% to 60% by2050 compared with today’s levels on a business-as-usual scenario.The challenge of preventing thisgrowth in emissions in developing countries fromrendering irrelevant any efforts in the developedworld to reduce carbon emissions is huge. Invest-ments in these developing countries in energyefficiency and low-carbon energy will need to beas great as those in the developed world, if notgreater. Given this, it is clear that the overwhelm-ing bulk of any EU 80% reduction commitmentwill have to be realized within the EU. It may wellbe that the developed world will have to contrib-ute to making the needed cuts in the developingworld, but these will need to be in addition toreductions at home, not instead of them. Put sim-ply, there will be no “low-hanging fruits” in thedeveloping world that might allow the EU toachieve its own reduction target cheaply; it willhave to realize an 80% or more cut in Europe.

An 80% cut in the EU is a huge challenge.Today almost 20% of the EU’s greenhouse gasescome from agriculture (methane from livestock)and industrial emissions (nitrous oxide). Both ofthese are difficult and expensive to reduce sub-stantially, as are, potentially, emissions from airtransport, maritime, and large freight vehicles. Soto achieve an 80% cut, the EU will have to investheavily in energy efficiency, and then move to azero-carbon electricity system, a zero-carbonheating and cooling system for buildings, and azero-carbon road and rail transport system. Putsimply, the EU will have to have a more or lesszero-carbon energy system by 2050. There arelimited ways to achieve this: renewable energy,

nuclear power, and coal and gas power generationcombined with carbon capture and storage(CCS).9

How Might the EU Move to aZero-Carbon Energy System by 2050?

Today nuclear energy meets around 15% of theEU’s total energy demand. There are reasonablegrounds to believe that this level will be main-tained in the EU, perhaps even increased some-what, as some countries such as Italy are seriouslyconsidering nuclear power. It remains to be seen,however, whether this 15% level will be signifi-cantly increased. In any event, even were it to bedoubled, it would still fall far short of what isneeded.

Renewable energy currently accounts forslightly less than 9% of the EU’s energy needs.This will increase to 20% by 2020. Even with thishuge growth in renewable energy, and maintain-ing or somewhat increasing the levels of nuclearpower in the EU, a large gap will still remain.

According to most studies, CCS offers hugepotential for low-carbon energy production. Itwill be particularly important for countries withvast cheap coal reserves, such as China and SouthAfrica. Without effective CCS technologies, cli-mate change will be very difficult to address.Nearly all studies conclude that CCS will be farfrom a cheap technology, however, and thatrenewable energy will be at least as competitive,and probably more so, especially where goodrenewable energy potential exists (see, e.g., IEA2008). Furthermore, CCS will remove around80% of greenhouse gas emissions, not 100%.

In fact, the EU is blessed with excellentrenewable energy potential: onshore wind, off-shore wind in the North and Baltic Seas, solar andPV in the south, biomass particularly in the east.Many studies indicate that with large-scale invest-ment and a determined approach to energy effi-ciency, the EU could meet all its energy needs by2050 from renewable energy. In fact, after consid-ering the EU’s options in producing a near zero-carbon energy sector by 2050, one comes to theconclusion that it will certainly include a highproportion of renewable energy. It is difficult to

244 Christopher Jones

paint a realistic scenario where renewable energywill (or at least should) account for less than 50%of the EU’s energy mix by 2050, and 60% to 80%seems perfectly reasonable. Such predictions areless surprising when one considers the directionof the EU’s energy mix today. In 2009, newrenewable energy installed capacity surpassed thatfrom traditional forms of electricity, with windtaking by far the largest share. If this trend contin-ues, renewable energy will already meet around30% of the EU’s energy needs by 2020, way abovethe 20% target.

In the event, therefore, that the post-Copenhagen process reaches an agreement that issufficiently robust to permit the EU to formallyrevise its 2020 target to a 30% cut and accept an80% or more reduction figure for 2050, the ques-tion will clearly be posed whether the 20% 2020renewable energy target should be increased. Inthis respect, it is worth noting that the EU’s newETS Directive already provides for an adjustmentof the available carbon credits from 2013 onwardin the event that the EU accepts a 30% green-house gas cut, to bring the reduction trajectoryfor ETS into line with this new objective. In addi-tion, it is reasonable to expect that if the EUaccepts a binding 80%+ emissions cut for 2050, itwill equally revise the ETS Directive accordingly,providing its industry—and in particular, its elec-tricity industry—with the long-term certaintyand predictability that it needs to make the neces-sary investments to achieve such radical cuts.

Although the EU has never seriously debatedwhether renewable energy targets should followsuit for 2020 or a legally binding minimum targetshould be set for 2050 for renewable energy, goodgrounds exist for such a move. Important argu-ments can be made in favour of using the ETSmechanism only to ensure the necessary (80% to95%) reductions, widening its scope in terms ofthe industries and sectors covered, and permittingthe electricity industry to adapt in the most cost-effective manner, which would result in a greateror lesser proportion of nuclear, renewable energy,and fossil-fuel electricity production combinedwith CCS.

A number of important counterargumentsalso can be made, however. These were already

considered above but merit further discussionregarding the future. It is clear that from the per-spectives of energy policy, industrial policy (i.e., afocus on jobs in the EU), and energy security, solong as renewable energy is competitive comparedwith CCS or nuclear, there is merit in ensuring asignificant proportion of renewable energy in theEU’s mix. A concern, however, is that this signifi-cant proportion may not result from a simpleapplication of the ETS system, even if renewableenergy would be a relatively cheap source com-pared with CCS and even nuclear.

First, the ETS system is economically thesame as a green certificate system and thereforeresults in the same level of risk, in terms of bothcost and its ability to actually deliver the amountof renewable energy needed. Renewable projectsare typically small undertakings, as local knowl-edge is often necessary to secure planning permis-sion and the small scale of many projects excludeslarge undertakings. These small companies havedifficulty in attracting financing in a market whereprofitability depends on a support mechanismcharacterized by fluctuating prices and thus highlevels of risk. The result of an ETS-only systemmay be lost investment opportunities at these lev-els, with consequently more focus on large CCSprojects by major utilities that can secure finan-cing.

Second, the ETS system relies on—in fact, isbased on—the principle that it is operating on areasonably functioning market. Thus in principle,in determining how to react to increasing carbonprices, generators should be able to freely chooseamong nuclear, investing in energy efficiency,coal and gas with CCS, and renewable energy.However, the ability to construct most nuclearand large renewable-energy-based plants isdependent on planning and environmental per-mission for often highly controversial projects,such as onshore wind, large-scale solar PV, orhydropower, and most important, permission tobuild new overhead lines. To enable the choice tomove into renewable electricity to be made, gov-ernments will also have to undertake massiveinfrastructure investments, such as smart grids,offshore grids, and pumped storage, which willresult in increased transmission charges to cus-

The European Union’s Policy on the Development of Renewable Energy 245

tomers. These are difficult political decisions andtypically subject to very long delays, if approved atall. Indeed, the result of simply relying on theETS system may, perversely, be that because thereare no minimum renewable targets, it becomesmore politically expedient to take no action onpermitting, planning, and investment, forcinggenerators down the coal and CCS route, whichmay well be more expensive, produce less jobs inthe EU, and not have the same positive effects interms of promoting the EU’s energy independ-ence as would result from the cheaper renewableenergy.

These issues are very important for the devel-opment of renewable energy in the EU and for itsability to meet its climate change commitments.

Under any scenario, it is difficult to envisage afuture zero-carbon electricity industry that doesnot figure at minimum around 50% renewableelectricity. It will almost certainly be more. Inorder to ensure that member states concentrate onensuring the development of appropriate plan-ning laws, as well as the cross-border regulatoryrules and financial incentives and guarantees nec-essary to enable the development of largemulticounty projects in due time, there are there-fore strong grounds for agreeing to minimumbinding 2050 targets on a member state by mem-ber state basis for these (conservative) levels in theevent of an agreement resulting from the Copen-hagen process. However, this issue has hardly beendebated at the EU level.

Grid and Energy Security IssuesMeeting the 20% 2020 target will lead to a mas-sive change in the EU’s electricity system, notablyin terms of the number of generators (from hun-dreds to millions), the number of sites, the dis-tance of much of the generation from load (largerenewable generation facilities are often consider-able distances from major centers of demand), andthe difficulty of ensuring that the system is kept inbalance. Much renewable energy is intermittent;the wind does not blow or the sun does not shineaccording to schedule.

This gives rise to three major challenges. First,many new transmission lines will be needed toreflect the changing topography of EU genera-tion. The need for such investment is alreadybeing felt in the EU; when the wind blows unex-pectedly hard in Germany, which has much of theEU’s installed wind capacity, transmission opera-tors struggle to bring the energy generatedsecurely and effectively to demand centers. Sec-ond, with many millions of more or less predict-able generators, in order to have a cost-effectiveapproach to balancing, a smart grid is imperative.And third, many large-scale renewable projectswill be cross-border in nature, such as offshorewind parks. These will require common regula-tory and grid access rules to enable them to oper-ate effectively.

Regarding the first challenge, the need fornew overhead lines, it is already clear that thedifficulty in acquiring permission to build newhigh-voltage electricity lines threatens to be amajor bottleneck to the development of many ofthe most cost-effective forms of renewable elec-tricity, notably onshore wind farms, large-scalePV parks, and offshore wind farms. Almost bydefinition, these installations are situated awayfrom large population centers and industrialdemand, and new lines are thus essential. Alreadyin the EU, the failure to approve such overheadlines at the same time that renewable capacity hasbeen added has led to considerable grid problems,as capacity is built before the necessary lines areconstructed to efficiently carry it to demand.Underground lines are becoming more cost-effective, but they remain more expensive thanoverhead AC lines by a factor of 5 to 10. This hasto be addressed—it is not possible in the mediumterm to continue constructing large wind farmswithout, at the same time that the farms areapproved, agreeing to the construction of thenecessary grid reinforcements.

A new approach to infrastructure planning isforeseen in the third internal energy market pack-age, which sets the rules for the liberalization ofthe EU’s electricity market. Under this recentlyadopted legislative package, the new EuropeanAgency for the Cooperation of Energy Regula-tors (ACER) has to regularly adopt a 10-year roll-

246 Christopher Jones

ing infrastructure plan on the basis of a proposalprepared by the transmission companies through anew association, the European Network ofTrans-mission System Operators for Electricity(ENTSO-E). This has great potential to addressthis priority. It will, however, need to be followedup with real determination at the national politi-cal level to ensure that the necessary planning per-mission is granted sufficiently quickly for eitheroverhead lines or “undergrounding,” financedthrough transparent and expensive public subsi-dies, before permission is given to constructrenewable capacity. Any other approach is likely toresult in unacceptable grid security risks in themedium term.

One possible approach is to require nationalenergy regulators to approve new renewablecapacity projects above a given size (5 MW, forexample) before construction to ensure that thenecessary grid capacity exists. Some argue, how-ever, that this will simply result in an insurmount-able barrier to new large-scale renewable energyplants: because of the political difficulty inapproving new overhead transmission lines, plan-ning permission will not be forthcoming in anyevent, and projects will simply be blocked by theregulator’s veto. It is far better, goes the argument,to let the new renewable capacity build up untilnetwork security becomes a pressing and una-voidable issue; only then will planning permissionfor the overhead lines be granted. While such anapproach cannot be described as good govern-ance, it may, regrettably, be true. This is a realproblem, but pursuant to the subsidiarity princi-ple, planning permission is an issue that can beaddressed only at the member state level.10

In regard to the second issue, the need foradequate balancing, it is vital that a clear and pre-dictable framework be developed at the EU levelto ensure that the necessary capacity is availablewhen it is needed to deal with rapidly increasinglevels of renewable energy, not after. The EU hasmany potential sources of balancing capacity, suchas pumped storage, compressed air, and gas-firedgeneration including CCS. Over the next years,therefore, the EU should give priority to deter-mining a clear set of rules on issues such as whoseresponsibility it is to ensure that adequate balanc-

ing capacity exists—whether this should simplybe left to the market, a supplier of last resort, thenational energy regulator or the relevant transmis-sion company—and equally important, who pays.

In addition, one very important source of bal-ancing capacity will need to come from the devel-opment of a smart grid. In essence, this means theability of electricity distribution companies,through smart meters in every home and business,to control a number of sources of energy demand,such as refrigerators, freezers, and most impor-tant, electric car batteries. In the third internalenergy market package, member states haveaccepted an obligation to install such meters in allbuildings when it is cost-effective to do so. Asmart grid requires a great deal more than this,however, and its establishment must be a priorityof the commission, national energy regulators,transmission companies, distributors, and memberstates over the coming years.

Finally, the third issue with a new and com-prehensive regulatory framework for renewableenergy concerns common regulatory rules forgrid access regarding projects broader than a singlemember state. As projects such as the North SeaOffshore Ring further develop, this will need tobe another priority for the commission, nationalenergy regulators, and transmission companies.

ConclusionsThis short tour of the EU’s new renewable energypolicy is far from exhaustive, yet it does illustratesome important truths. Though the policy is boldand ambitious, it represents but the first stepsdown what will be a very long road. The clearestconclusion to be drawn at present is that there is apressing need to develop a clear, transparent, andpredictable regulatory regime at the EU level.Infrastructure and grid access rules must be putinto place rapidly to ensure that along with thegrowth in renewable energy, adequate infrastruc-ture is built to make the best use of availablecapacity. This will be a major challenge, but giventhe emergence of an increasingly mature regula-tory environment at the EU level with the crea-tion of ACER and ENTSO-E, there are good

The European Union’s Policy on the Development of Renewable Energy 247

grounds to believe that with determination, it willbe possible to meet this challenge.

Appendix: Key Excerpts from theFirst Strategic Energy ReviewFollowing are excerpts containing the key ele-ments from the First Strategic EU EnergyReview, titled “An Energy Policy for Europe,”presented by the European Commission in Janu-ary 2007 (European Commission 2007b). Unlessnoted otherwise, the European Commission wasthe source of all data except for the percentagegiven in the first sentence, which was from theEuropean Environment Agency. The review alsohad a footnote stating that the assumed dollarexchange rate was $1.25 per euro, and that it wasusing for comparison an oil price of $60 (in2007$) in 2030. First, in terms of the rationale forchange, the commission stated the following:

1.1. SustainabilityEnergy accounts for 80% of all greenhouse gas(GHG) emission in the EU; it is at the root of cli-mate change and most air pollution. The EU iscommitted to addressing this – by reducing EUand worldwide greenhouse gas emissions at aglobal level to a level that would limit the globaltemperature increase to 2°C compared to pre-industrial levels. However, current energy andtransport policies would mean EU CO2 emissionswould increase by around 5% by 2030 and globalemissions would rise by 55%.The present energypolicies within the EU are not sustainable.

1.2. Security of supplyEurope is becoming increasingly dependent onimported hydrocarbons. With “business as usual”the EU’s energy import dependence will jumpfrom 50% of total EU energy consumption todayto 65% in 2030. Reliance on imports of gas isexpected to increase from 57% to 84% by 2030,of oil from 82% to 93%.

This carries political and economic risks. Thepressure on global energy resources is intense.The International Energy Agency (IEA) expectsglobal demand for oil to grow by 41% by 2030.How supply will keep up with this demand is

unknown: the IEA in its 2006 World Energy Out-look stated that “the ability and willingness ofmajor oil and gas producers to step up investmentin order to meet rising global demand are particu-larly uncertain” [EIA/DOE 2006, 4]. The risk ofsupply failure is growing.

In addition, the mechanisms to ensure solidar-ity between Member States in the event of anenergy crisis are not yet in place and severalMember States are largely or completely depend-ent on one single gas supplier.

At the same time, EU electricity demand is, ona business as usual scenario, rising by some1.5% per year. Even with an effective energy effi-ciency policy, investment in generation alone overthe next 25 years will be necessary in the order of€900 billion. Predictability and effective internalgas and electricity markets are essential toenable the necessary long term investments totake place and for user prices to be competitive.These are not yet in place.

1.3. CompetitivenessThe EU is becoming increasingly exposed to theeffects of price volatility and price rises on inter-national energy markets and the consequencesof the progressive concentration of hydrocarbonsreserves in few hands. The potential effects aresignificant: if, for example, the oil price rose to 100$/barrel in 2030, the EU-27 energy total import billwould increase by around €170 billion, an annualincrease of €350 for every EU citizen.Very little ofthis wealth transfer would result in additional jobsin the EU.

Providing that the right policy and legislativeframeworks are in place, the Internal Energy Mar-ket could stimulate fair and competitive energyprices and energy savings, as well as higherinvestment. However, all the conditions to achievethis do not yet exist.This prevents EU citizens andthe EU economy from receiving the full benefits ofenergy liberalisation. A longer time horizon in thearea of carbon constraints is required in order topromote the necessary investments in the elec-tricity sector.

Boosting investment, in particular in energy effi-ciency and renewable energy should create jobs,promoting innovation and the knowledge-basedeconomy in the EU. The European Union isalready the global leader in renewable technolo-gies, accounting for a turnover of €20 billion andemploying 300,000 people [European Union

248 Christopher Jones

Committee 2008]. It has the potential to lead therapidly growing global market for low carbonenergy technologies. In wind energy, for example,EU companies have 60% of the world marketshare. Europe’s determination to lead the globalfight against climate change creates an opportu-nity for us to drive the global research agenda. Alloptions should be kept to ensure the developmentof emerging technologies.

At the same time, the social dimension ofEurope’s energy policy needs to be taken intoaccount throughout all stages of designing andimplementing the individual measures. While thispolicy should overall contribute to the growth andjobs in Europe on the long term, it may have asignificant impact on some internationally tradedproducts and processes in particular in the areaof energy-intensive industries.

The relevant parts of the review concerningrenewable energy and ETS are as follows:

2. A STRATEGIC OBJECTIVE TO GUIDEEUROPE’S ENERGY POLICY

The point of departure for a European energypolicy is threefold: combating climate change, lim-iting the EU’s external vulnerability to importedhydrocarbons, and promoting growth and jobs,thereby providing secure and affordable energy toconsumers.

In the light of the many submissions receivedduring the consultation period on its Green Paper[European Commission 2006b], in this StrategicEnergy Review the Commission proposes thatthe European Energy Policy be underpinned by:

• an EU objective in international negotia-tions of 30% reduction in greenhouse gasemissions in developed countries by 2020compared to 1990. In addition, 2050 glo-bal GHG emissions must be reduced byup to 50% compared to 1990, implyingreductions in industrialised countries of60–80% by 2050;

• an EU commitment now to achieve, in anyevent, at least a 20% reduction of green-house gases by 2020 compared to 1990.

These form a central part of the CommissionCommunication “Limiting Climate Change to 2°C– Policy Options for the EU and the world for 2020and beyond.”

Meeting the EU’s commitment to act now ongreenhouse gases should be at the centre of thenew European Energy Policy for three reasons: (i)CO2 emissions from energy make up 80% of EUGHG emissions, reducing emissions meansusing less energy and using more clean, locallyproduced energy, (ii) limiting the EU’s growingexposure to increased volatility and prices for oiland gas, and (iii) potentially bringing about amore competitive EU energy market, stimulatinginnovation technology and jobs.

Taken together, this strategic objective and theconcrete measures set out below to make it areality represent the core of a new EuropeanEnergy Policy …

3.3. A long-term commitment togreenhouse gases reduction and theEU Emissions Trading SystemThe EU traditionally favours the use of economicinstruments to internalise external costs as the[sic] allow the market to determine how to reactmost efficiently and with limited costs. More par-ticularly, in its Communication Limiting ClimateChange to 2°C – Policy Options for the EU andthe world for 2020 and beyond, the Commissionhas set out how the emissions trading mechanismis and must remain a key mechanism for stimulat-ing reductions in carbon emissions and how itcould be used as a basis for international effortsto fight climate change. The Commission isreviewing the EU ETS to ensure that emissionstrading reaches its full potential: this is critical tocreating the incentives to stimulate changes inhow Europe generates and uses its energy.

3.4. An ambitious programme ofenergy efficiency measures atCommunity, national, local andinternational levelFor Europe’s citizens, energy efficiency is themost immediate element in a European EnergyPolicy. Improved energy efficiency has the poten-tial to make the most decisive contribution toachieving sustainability, competitiveness andsecurity of supply.

On 19 October 2006 the Commission adoptedthe Energy Efficiency Action Plan [EuropeanCommission 2006a], containing measures thatwould put the EU well on the path to achieving akey goal of reducing its global primary energy use

The European Union’s Policy on the Development of Renewable Energy 249

by 20% by 2020. If successful, this would meanthat by 2020 the EU would use approximately13% less energy than today, saving €100 billionand around 780 millions tonnes of CO2 each year.However, this will require significant efforts both interms of behavioural change and additionalinvestment …

3.5. A longer term target forrenewable energyIn 1997, the European Union started workingtowards a target of a 12% share of renewableenergy in its overall mix by 2010, a doubling of1997 levels. Since then, renewable energy pro-duction has increased by 55%. Nevertheless theEU is set to fall short of its target. The share ofrenewable energy is unlikely to exceed 10% by2010. The main reason for the failure to reach theagreed targets for renewable energy – besidesthe higher costs of renewable energy sourcestoday compared to “traditional” energy sources –is the lack of a coherent and effective policyframework throughout the EU and a stable long-term vision. As a result, only a limited number ofMember States have made serious progress inthis area and the critical mass has not beenreached to shift niche renewables production intothe mainstream.

The EU needs a step change to provide a cred-ible long term vision of the future of renewableenergy in the EU, building on the existing instru-ments, notably the Renewable Electricity Direc-tive. This is essential to realise present targets[European Commission 2007c] and trigger furtherinvestment, innovation and jobs.The challenge forrenewables policy is to find the right balancebetween installing large scale renewable energycapacity today, and waiting until research lowerstheir cost tomorrow. Finding the right balancemeans taking the following factors into account:

• Using renewable energy today is gener-ally more expensive than using hydrocar-bons, but the gap is narrowing – particu-larly when the costs of climate change arefactored in;

• Economies of scale can reduce the costsfor renewables, but this needs majorinvestment today;

• Renewable energy helps to improve theEU’s security of energy supply by increas-ing the share of domestically produced

energy, diversifying the fuel mix and thesources of energy imports and increasingthe proportion of energy from politicallystable regions as well as creating newjobs in Europe;

• Renewable energies emit few or no green-house gases, and most of them bring sig-nificant air quality benefits.

In the light of the information received during thepublic consultation and the impact assessment,the Commission proposes in its RenewableEnergy Roadmap [European Commission 2007d]a binding target of increasing the level ofrenewable energy in the EU’s overall mix fromless than 7% today to 20% by 2020. Targetsbeyond 2020 would be assessed in the light oftechnological progress …

This 20% target is truly ambitious and willrequire major efforts by all Member States. Thecontribution of each Member State to achievingthe Union’s target will need to take into accountdifferent national circumstances and startingpoints, including the nature of their energy mix.Member State should have the flexibility to pro-mote the renewable energies most suited to theirspecific potential and priorities. The way in whichMember States will meet their targets should beset out in National Action Plans to be notified tothe Commission. The Plans should containsectoral targets and measures consistent withachieving the agreed overall national targets. Inpractice, in implementing their Plans MemberStates will need to set their own specific objec-tives for electricity, biofuels, heating and cooling,which would be verified by the Commission toensure that the overall target is being met. TheCommission will set out this architecture in a newrenewables legislative package in 2007.

A particular feature of this framework is theneed for a minimum and coordinated develop-ment of biofuels throughout the EU.While biofuelsare today and in the near future more expensivethan other forms of renewable energy, over thenext 15 years they are the only way to significantlyreduce oil dependence in the transport sector. Inits Renewable Energy Roadmap and BiofuelsProgress Report [European Commission 2007a],the Commission therefore proposes to set a bind-ing minimum target for biofuels of 10% of vehiclefuel by 2020 and to ensure that the biofuels usedare sustainable in nature, inside and outside theEU. The EU should engage third countries and

250 Christopher Jones

their producers to achieve this. In addition, the2007 renewables legislative package will includespecific measures to facilitate the market penetra-tion of both biofuels and heating & cooling fromrenewables. The Commission will also continueand intensify the use of renewable energythrough other policies and flanking measures withthe aim of creating a real internal market forrenewables in the EU.

Notes

1. European laws are adopted by codecision, whichmeans that that the council, a body composed ofrepresentatives of the EU’s 27 member states, hasto agree on new laws with the European Parlia-ment, comprising 785 directly elected membersfrom all EU countries. The council agrees on thetext mostly under a qualified majority voting sys-tem (which applied, for example, to the renewableenergy and electricity texts discussed in this chap-ter), meaning that larger countries get more votesthan smaller ones, taking into account inter aliatheir greater populations. However, it is importantto understand the role of the European Commis-sion, an administrative body of 27 headed by apresident (currently Manuel Barroso) with 26commissioners, thus one representative per mem-ber state. The commission has the sole right ofinitiative on new legislative proposals and is veryimportant when negotiating the final text with thecouncil and parliament.

2. The commission’s progress report on renewableenergy (European Commission 2009a) found thatCyprus, Finland, Latvia, Malta, Romania, andSlovenia were all showing low levels of recentgrowth and of progress toward the 2010 targets.

3. This was the only element of the commission’sproposals that, in fact, the council could formallyagree on; it concerned an equal target for allmember states. The other elements proposed anoverall EU target (20% renewable energy in theoverall EU energy mix) without giving a memberstate by member state breakdown of the target.

4. To illustrate this, on the basis of the figures out-lined in the IEA Energy Technology PerspectivesReport and taking the lowest figure in the rangefor each technology, the current cost of onshorewind is 4.6 euro cents (6.23 U.S. cents) perkilowatt-hour (kWh), offshore wind 5.3 (7.2),solar PV 21.2 (28.7), concentrated solar power 8.8

(11.92), and biomass 4.2 (5.7). Although onshorewind and biomass are among the cheapest, supplyis limited because of planning constraints and isunlikely to meet all the demand for renewableelectricity, certainly given the level of futuredemand resulting from the new RenewableEnergy Directive. Thus under a green certificatesystem, the overall price will be set by the cost ofthe marginal unit necessary to meet the level ofdemand set (or at least the price that the marginalunit is willing to bid into the price-setting mecha-nism, which in times of shortage can be consider-ably higher than the actual cost; see the discussionimmediately below). This may be photovoltaics,for example, which would mean that all other pro-ducers would benefit from a windfall profit, allbeing paid 21.2 euro cents (28.7 U.S. cents) perkWh.

5. For a more complete account of the directive, seeHodson et al. (forthcoming).

6. See www.consilium.europa.eu/ueDocs/cms_Data/docs/pressData/en/ec/104692.pdf.

7. Indeed, the directive recognizes that public sup-port is necessary to reach the community’s objec-tives for renewable energy.

8. The internal energy market is the name given tothe liberalized electricity market in the EU. It isbased on the principles that customers may pur-chase from any electricity supplier in the EU, andgenerators can establish themselves in any EUcountry and export to any customer regardless ofwhere they are situated in the EU. For a detailedanalysis of the market, see Jones (2010b) .

9. A more detailed analysis of this discussion can befound in Jones (2010a).

10. The subsidiarity principle states that action can betaken at the EU level only when it is essential toachieve a given (and recognized as a legitimate)objective of the EU Treaty and equivalent resultscannot be achieved adequately by action at thenational level. For further information on thisconcept, see European Commission (2001).

ReferencesEIA/DOE (Energy Information Administration/U.S.

Department of Energy). 2006. International EnergyOutlook. http://161.116.7.34/conferencies/viitrobada/INTE.%20ENEREG.%20AGC.%20resum%20angl%C3%A8s.pdf (accessed February 20, 2010).

The European Union’s Policy on the Development of Renewable Energy 251

European Commission. 1997. Communication fromthe Commission, Energy for the Future: RenewableSources of Energy, White Paper for a CommunityStrategy and Action Plan. http://europa.eu/documents/comm/white_papers/pdf/com97_599_en.pdf (accessed February 20, 2010).

———. 2001. White Paper on European Governance.http://eur-lex.europa.eu/LexUriServ/site/en/com/2001/com2001_0428en01.pdf (accessed Feb-ruary 20, 2010).

———. 2006a. Action Plan for Energy Efficiency:Realising the Potential. http://ec.europa.eu/energy/action_plan_energy_efficiency/doc/com_2006_0545_en.pdf (accessed February 20,2010).

———. 2006b. Green Paper: A European Strategy forSustainable, Competitive and Secure Energy. http://ec.europa.eu/energy/green-paper-energy/doc/2006_03_08_gp_document_en.pdf (accessed Febru-ary 20, 2010).

———. 2007a. Biofuels Progress Report. http://ec.europa.eu/energy/energy_policy/doc/07_biofuels_progress_report_en.pdf (accessed Feb-ruary 20, 2010).

———. 2007b. An Energy Policy for Europe. http://europa.eu/legislation_summaries/energy/european_energy_policy/l27067_en.htm (accessedFebruary 20, 2010).

———. 2007c. Follow-Up Actions of the GreenPaper: Report on Progress in Renewable Electricity.http://ec.europa.eu/energy/energy_policy/doc/06_progress_report_renewable_electricity_en.pdf(accessed February 20, 2010).

———. 2007d. Renewable Energy Roadmap:Renewable Energies in the 21st Century; Building aSustainable Future. http://ec.europa.eu/energy/

energy_policy/doc/03_renewable_energy_roadmap_en.pdf (accessedFebruary 20, 2010).

———. 2009a. Renewable Energy Progress Report.http://eur-lex.europa.eu/LexUriServ/LexUriServ.do?uri=COM:2009:0192:FIN:EN:PDF(accessed February 20, 2010).

———. 2009b. Strategic Energy Technology Plan(SET Plan). http://ec.europa.eu/energy/technology/set_plan/set_plan_en.htm (accessedFebruary 20, 2010).

European Council. 1997. Council Resolution of27 June 1997 on Renewable Sources of Energy. Offi-cial Journal C 210.

European Union Committee. 2008. The EU’s Targetfor Renewable Energy: 20% by 2020. Vol. 1, 27thReport of Session 2007–08.http://www.publications.parliament.uk/pa/ld200708/ldselect/ldeucom/175/175.pdf (accessedFebruary 20, 2010).

Hodson, Paul, Christopher Jones, and Hans van Steen,eds. Forthcoming. EU Energy Law. Vol. III, Book 1,Renewable Energy Law and Policy in the EuropeanUnion. Leuven, Belgium: Claeys & Casteels Publish-ing.

IEA (International Energy Agency). 2008. Energy Tech-nology Perspectives: Scenarios and Strategies to 2050.Paris: OECD Publishing.

Jones, Christopher. 2010a. A Zero Carbon EnergyPolicy for Europe: The Only Viable Solution. In EUEnergy Law.Vol III, Book 3,The European RenewableEnergy Yearbook. Christopher Jones (ed). Leuven,Belgium: Claeys & Casteels Publishing, 21–100.

Jones, Christopher, ed. 2010b. EU Energy Law. Vol. I,The Internal Energy Market. 3rd ed. Leuven, Belgium:Claeys & Casteels Publishing.

Piebalgs, A. 2009. http://europa.eu/rapid/pressReleasesAction.do?reference=SPEECH/09/488(accessed April 6, 2010).

252 Christopher Jones

13

UK Renewable Energy Policysince PrivatizationMichael G. Pollitt

This chapter reviews the progress with increas-ing renewable energy supply in the United

Kingdom since 1990, with a particular focus onrecent developments. This country is regarded asone where the considerable potential for renew-able energy,1 relative to other major Europeancountries, has failed to be realized. It is also fre-quently suggested that the United Kingdomneeds to change its policies to renewables to lookmore like those in Germany or Spain (e.g.,Mitchell 2007).

The aim of this chapter is to look at theUnited Kingdom’s renewable energy policy in thecontext of its overall decarbonization (i.e. carbonemissions reduction) and energy policies. Thechapter explores the precise nature of the failureof UK renewables policy and suggests policychanges that might be appropriate in light of thecountry’s institutional and resource endowments.The focus is on the electricity sector in terms ofboth renewable generation and, to a lesser extent,the facilitating role of electricity distribution andtransmission networks. The interactions amongthe UK’s electricity, heat, and transport sectorswithin the overall decarbonization policy contextare also examined.

The discussion suggests that the precise natureof the failure of UK policy is rather more to dowith societal preferences and the available mecha-

nisms for resolving social conflict than with eco-nomic incentive arrangements. Radical changesto current policy are required, but policymakersmust be careful that they are institutionally appro-priate to the United Kingdom. Calls to “just doit” with respect to delivery of larger quantities ofrenewables are economically irresponsible andhighly likely to backfire in terms of achievementof ultimate policy goals such as decarbonizationand energy security.

UK renewable energy policy exists in a widerenergy policy context. The country’s statedenergy policy can be summed up as aiming toachieve “secure, affordable and low-carbonenergy” (see DECC n.d.b). It therefore has threeidentifiable priorities: addressing climate change,providing energy security, and keeping energybills down.These policy objectives are naturally intension.The first two are expensive, whereas tack-ling the third entails keeping prices down, if notfor everyone, then for a significant minority ofpoor consumers. Between 1990 and 2003, resi-dential electricity prices fell significantly in realterms in the United Kingdom, by around 30% perunit, but have risen by around 40% from 2003 to2008 (QEP 2009). The number of householdsdefined as being in energy (or fuel) poverty,spending 10% or more of total expenditure onheating and power, has risen from a low of 2 mil-

lion in 2003 to 3.5 million in 2006 (BIS 2008),out of a total of around 25 million households(ONS 2007).This has put a strain on the ability ofricher consumers to simultaneously finance poorconsumers, via bill payments to company supportschemes (see Ofgem 2009b),2 and expensive poli-cies arising from climate change and energy secu-rity objectives. European Union (EU) directiveshave also provided significant shape to UK energypolicy, providing the basis for targets to 2020 forCO2 reduction and renewable electricity genera-tion share.

The chapter begins by reviewing the UnitedKingdom’s overall decarbonization policy andpotential for renewables, then its policy towardrenewables since 1990, with a particular focus onrecent developments. This is followed by anexamination of the evidence on the performanceof UK policy compared with that of other coun-tries. Next, a new institutional economics per-spective is used to discuss what sorts of policiesmight be right for the United Kingdom in thelight of the evidence. Finally, the chapter exam-ines the issue of overall policy toward decarbon-ization and the place of renewables within this.

Decarbonization Policy and thePotential for Renewable Energyin the United KingdomAn important context for the United Kingdom’srenewable energy policy is its overall decarboniza-tion policy. The country has one of the mostambitious decarbonization policies in the world,as embodied in the 2008 Climate Change Act(OPSI 2008a).3 This policy consists of a commit-ment to reducing net greenhouse gas (GHG)emissions by 80% by 2050 (from 1990 levels),with an intermediate target reduction of 26% by2020. This target is supported by five-year carbonbudgets, the first period being 2008–2012 inclu-sive.These budgets are formulated in the Office ofClimate Change and supported by a report fromthe independent Committee on Climate Change(CCC). Government ministers have a statutory

duty to introduce policies that support achieve-ment of the targets. The committee’s first report(CCC 2008) was published in December 2008.This gave indicative budgets for the periods 2008–2012, 2013–2017 and 2018–2022.The budget forany period beyond this must be set at least 12years ahead.

The report was then followed up by a signifi-cant discussion in the HM Treasury budget for2009 of policy measures aimed at supporting theachievement of the decarbonization targets in thelight of the report (see HM Treasury 2009).4 Theannounced measures included support for greenmanufacturing, improvements to the renewablesupport for offshore wind, increased funding forcombined heat and power, and a support mecha-nism for up to four carbon capture and storage(CCS) plants. The intention of the legislation isthat if the government were to fail to enact appro-priate policies to keep the United Kingdom ontrack to achieve its targets, this could result in legalaction against ministers by third parties, though itremains to be seen on what grounds any actionwould be likely to be successful, given the lessthan direct link between specific governmentpolicy and impact on a national GHG target.

For reference, in 2008, UK GHG emissionswere 623.8 metric tons of CO2e (CO2 equivalentunits), which is 20% below the 1990 baseline of779.9 tons (Defra 2008). This means the UnitedKingdom is the only major European country tohave already met and exceeded its 2012 Kyototarget for emissions reduction target, which was12.5% (see EEA 2006, Table 1). It is, however,worth pointing out that the UK target is the resultof negotiations within the EU to share out theKyoto-negotiated EU-wide target, and that thebaseline date of 1990 is very favorable to theUnited Kingdom. This is because it coincideswith the privatization of the UK power industry,leading to a “dash for gas,” which resulted in anunintended environmental windfall as dirtiercoal-fired plants were displaced from the system(see Newbery and Pollitt 1997). This favorablestarting place in which the United Kingdom findsitself is certainly a major factor in its relativeenthusiasm for decarbonization.5 The 2009 EURenewables Directive further commits the

254 Michael G. Pollitt

United Kingdom to a 15% target for renewablescontribution to total final energy consumption in2020 as part of the EU’s overall 20% renewablesby 2020 target. This further target is acknowl-edged and accepted in the CCC report. TheUnited Kingdom also has a specific annual targetfor the percentage of electricity from renewablesout to 2015 as part of its Renewables ObligationCertificate scheme, discussed later in this chapter.

The report suggests that by 2020, the share ofrenewables could be as much as 30% in total elec-tricity generation (CCC 2008, 208). It also dis-cusses the potential for the direct reduction ofemissions from buildings rather than via large-scale grid-connected electricity. This involves acombination of renewable heat and micro-generation. For residential buildings, it identifies apotential contribution of 14% reduction in heatemissions via a combination of biomass, solar hotwater, heat pumps, and biogas by 2020. In addi-tion, small contributions may be made by PV andother sources for microgeneration of electricity.

Recently, the newly created responsible gov-ernment ministry, the Department for Energy andClimate Change published its UK RenewableEnergy Strategy (DECC 2009f). In line with theCCC report, this suggested that more than 30% ofelectricity should be generated from renewablesby 2020, as well as 12% of heat and 10% of trans-port energy, in order to meet EU targets.

The United Kingdom’s commitment todecarbonization is likely to lead to a relativelytight domestic policy with strong pressure forpurchasing of renewable electricity and CO2 per-mits from abroad. In 2007, the country was a netpurchaser of CO2 permits to the tune of 26 tons,or 3% of its 1990 GHG level (Defra 2009). It alsopurchased energy via the interconnector withFrance (3% of total electricity delivered), whichmay have displaced higher-carbon energy in theUnited Kingdom, and was one of the largest netimporters of internationally traded bioenergy,mainly for cofiring in coal-fired power plants andfor blending in gasoline (DECC 2009b; Jungingeret al. 2008; Perry and Rosillo-Calle 2008). All ofthese have some scope for expansion in terms ofachieving the net decarbonization of the UKeconomy.

Given the ambitious targets for decarboniza-tion and renewable energy in the United King-dom, it seems highly likely that nationally thesetargets will be missed, certainly on renewables. Inthese circumstances, serious consideration will begiven to meeting the targets via net purchases ofCO2 or green energy certificates from abroad(e.g., funding CCS in China). Indeed, ifadditionality could be clearly established, thiswould seem to be a very sensible option given thatat the margin, such purchases would be muchcheaper than domestic alternatives.

A defining feature of the United Kingdom isthe considerable potential it has for renewableenergy relative to its demand. The country hassome of the best wind, tidal, and wave resourcesin Europe, as well as affording opportunities forbiomass and solar. The technical potential of eachof these resources is very great, but the estimatedeconomic potentials are given in Table 13.1. UKelectricity supplied in 2008 was 380 terawatt-hours (TWh) (DECC 2009b, Table 5.5).

In addition, it is worth mentioning that theUnited Kingdom has up to 1,000 years’ worth ofstorage capacity of CO2 in the North Sea andcurrently generates around 13% of its electricity

Table 13.1. Estimates of the likely economicpotentials for different renewable technologies in theUnited Kingdom

Technologycategory

Technologydetail

Annualpotential

Wind power Onshore 50 TWh

Offshore 100 TWh

Bioenergy Biomass 41 TWh

Geothermal Ground sourceheat pumps

8 TWh

Hydro Large scale 5 TWh

Small scale 10 TWh

PV Retrofitted andbuilding inte-grated

> 1 TWh

Marine Wave energy 33 TWh

Tidal barrage 50 TWh

Tidal stream 18 TWh

Total ~316 TWh

Source: Jamasb et al. 2008b, 81–82

UK Renewable Energy Policy since Privatization 255

from nuclear power (DECC 2009b). The UnitedKingdom has endowments of coal, oil, and gas(though all three are depleting).Thus carbon cap-ture and storage and nuclear power are likely tocompete with renewables to play a part in decar-bonization of the electricity sector.The country isalready committed to an auction for one demon-stration CCS plant and is reviewing designs for anew generation of nuclear power plants, with anannouncement in November 2009 on its pre-ferred sites for new building (see DECC 2009c).Electricity demand growth is increasing slowly, ataround 1% a year, and energy efficiencymeasures—such as the elimination of filamentlightbulbs starting in 2011 (DECC 2009e) and theintroduction of smart metering for all electricitycustomers by 2020—seem likely to moderatedemand growth.

MacKay (2008, 109) predicts the likely con-tribution of renewables to UK decarbonization inthe context of delivering the current level ofenergy consumption of 125 kilowatt-hours perday per person. He suggests that renewables con-tribution is likely to be only 18.3 kWh/day/person, made up of the following: hydro, 0.3;tidal, 3; offshore wind, 4; biomass, 4; solar PV, 2(+ 2 from solar hot water); and onshore wind, 3.Thus renewable energy would contribute around15% toward total decarbonization. MacKay’sanalysis is helpful in that it illustrates that a bigcontribution toward current electricity provisioncomes in the context of electricity being thesource of only about one-third of current emis-sions of GHGs.

The exact mix of different renewable tech-nologies, CCS fitted to coal- or gas-fired plants,nuclear, and demand reduction in the UK energymix will depend on the relative costs of the differ-ent technologies. Kannan (2009) shows theimpacts of different assumptions on the signifi-cance of CCS in UK decarbonization and hencethe implications for other sources of decarboniza-tion. Demand reduction technologies are thecheapest GHG abatement technology at themoment (see CCC 2008, 221), though demandreduction measures suffer from well-known insti-tutional barriers to adoption (Grubb and Wilde2008). Nuclear is probably the next cheapest.

Among the renewable technologies in the UnitedKingdom, onshore wind, biomass, and offshorewind are lowest-cost at scale to 2020. Table 13.2shows some cost sensitivities for 2005.

The table illustrates large uncertainty in thecosts of building new plants, even with establishedtechnologies. For wind, this reflects the impor-tance of exact location, which determines bothbuilding costs and the available wind.The range ofcosts illustrates substantial overlap under favorableversus unfavorable circumstances for any pair oftechnologies. However, it is important to pointout that this uncertainty over actual costs for cur-rent new building does call into question projec-tions of costs to 2020. For instance, Dale et al.(2004) assume onshore and offshore new buildingcosts of £650 and £1,000 ($975 and $1,500) perkW, respectively, in scenarios with 25% energyfrom wind. The most recent (albeit prerecession)wind parks are currently costing nearer to £1,000and £2,500 ($1,500 and $3,750) per kW (seeBlanco 2009; and Snyder and Kaiser 2009).This issomewhat concerning, given a return tomacroeconomic growth, for the likely projectedcosts of renewable scenarios to 2020, especiallygiven that the costs of electricity (which willinclude cumulative subsidy commitments torenewables) in 2020 will still likely reflect, tosome extent, the cumulative cost of all windcapacity installed since 2005.

Table 13.2. Examples of estimated costs oftechnologies for the United Kingdom in 2005

Technology Technologydetail

p/kWh

Nuclear Generation III 3.04–4.37

Gas CCGTa with CCS 3.65–6.78

Coal IGCCb with CCS 3.5–5.67

Wind Onshore 4.68–8.89

Offshore 5.62–13.3

Source: Jamasb 2008b, 75.Notes: The spread of estimates reflects ranges in the discountrate, capital cost, fuel and carbon prices, and other sensitivities;p/kWh = pence per kilowatt-hour, given in 2005 values; 1 pence= 1.5 cents (U.S.) as of this writing.aCCGT = combined cycle gas turbinebIGCC = integrated gasification combined cycle

256 Michael G. Pollitt

As Jamasb et al. (2008b) note, a key determinantof the relative attractiveness of different technolo-gies will be the degree of learning in costs, andthis depends on their current stage of develop-ment. Foxon et al. (2005) note that the variousrenewable technologies available to the UnitedKingdom are at different stages of development.Wind costs can be expected to fall as capacityincreases significantly around the world; however,the prospects for learning in hydro and tidal bar-rages are low, limiting their ultimate scope forexpansion. The additional costs of fitting CCS aredifficult to estimate because of a lack of informa-tion, while the scope for learning may be con-strained by the maturity of the different elementsof the CCS process (see Odenberger et al. 2008).This is in addition to the difficulty of reconcilingall the interested parties (Drake 2009). PV, tidalstream, and other marine technologies offer thegreatest potential for decreases from the currentcosts, given low current levels of output and theimplied scope for cost reduction.6

SKM (2008) provides estimates of the possiblecost of decarbonization of the electricity sector to

2020. Under their estimates, renewables provide34%, 41%, and 50% of electricity supply underthe lower, middle, and higher renewables sce-narios. Table 13.3 shows that renewables couldimpose significant total costs on the electricitysystem. The capital costs of connecting offshorewind in particular could involve up to £15 billion($22.5 billion) of expenditure, more than the totalcost of generation under a conventional scenario.The cost of balancing and intermittency couldrise by up to £7 ($10.50) per megawatt-hour(MWh), or 10% of total system costs. The UnitedKingdom may have the wind resources, but theywill have significant cost implications for the sys-tem, raising average electricity costs by up to 40%against baseline.

Policies toward Renewables inthe United KingdomThis section provides an overview of UKrenewables policy since the privatization of the

Table 13.3. Costs of electricity sector decarbonization to 2020 (2008 prices)

Renewables scenarios

Conventional Lower Middle Higher

New generation capacity (£ billion)

Renewable capacity 2.3 50.1 60.2 77.4

Nonrenewable capacity 14.9 12.6 12.3 12.0

Total 17.2 62.7 72.5 89.4

Network (£ billion)

Offshore wind connection 0.0 8.4 10.6 14.1

Onshore wind connection 0.1 1.0 1.2 1.4

Other reinforcement 0.8 0.8 0.8 0.8

Total 0.9 10.2 12.6 16.3

Total grid investment costs(generation + network, £ billion)

18.1 72.9 85.1 105.7

Marginal generation cost (£/MWh) 35.9 25.0 22.6 18.9

Cost per MWh produced (£/MWh)

Generation costs (fixed and variable) 46.8 51.9 52.6 54.5

Balancing and intermittency 1.7 6.3 7.2 8.7

Grid expansion for renewables 0.1 3.5 4.1 5.2

Total cost including network (£/MWh) 48.6 61.7 63.9 68.4

Source: SKM 2008, 8Note: £1 = $1.50 as of this writing

UK Renewable Energy Policy since Privatization 257

country’s electricity supply industry beginning in1990. Summarizing UK policy is not a straight-forward task because of the large range of govern-ment initiatives toward renewable energy and thegreat number of policy changes that have beenannounced in recent years, some of which haveyet to be implemented fully.7 Discovering theexact cost of renewable energy support is not easy,as evidenced by the fact that the best sources ofinformation are answers to parliamentary ques-tions rather than published annual statistics.This isparticularly true of the expenditure on individualtechnologies. The heroic efforts of Mitchell andConnor (2004), who reviewed UK renewablespolicy from 1990 to 2003, provided the inspira-tion for some of the presentation here.

In broad outline, there have been two mainsupport mechanisms for renewable electricity andheat generation since privatization in the UnitedKingdom: the Non-Fossil Fuel Obligation(NFFO), which ran from 1990 to 2002, and theRenewables Obligation (RO) scheme, whichbegan in 2002. During their period of operation,these have been the most significant forms ofrenewable energy support in the United Kingdomand were designed to work in parallel with liber-alized electricity and gas markets.

The assessment of renewable support policiesis complicated because there are two obviousmetrics of success: the amount of renewables real-ized relative to potential (quantity); and the totalcost of renewable energy support policy relative tothe amount of generation actually supported(suitably discounted). These two trade off, mean-ing that success in one is likely to be associatedwith less success in the other.

The Non-Fossil Fuel Obligation

The Non-Fossil Fuel Obligation (NFFO) wasoriginally designed as a way of financing the extracosts of nuclear power that became clear in therun-up to privatization. A non-fossil-fuel levy wasintroduced on final electricity prices to pay fornuclear decommissioning liabilities, and electri-city suppliers were forced to buy nuclear power athigher-than-market prices in auctions for non-fossil-fuel power run by the Non-Fossil Purchas-

ing Agency (NFPA).8 In order to avoid this beingseen as a discriminatory subsidy to the nuclearindustry, it was recast as a way of supporting non-fossil-fuel generation more generally, and a por-tion was allocated to support renewable energy(Mitchell and Connor 2004). The portion wassmall, but it provided a relatively significantamount of money to the industry at a time whengovernment expenditure on new technologieswas falling to a very low level, and the thenDepartment of Energy was closing. The moneywas allocated to new renewable projects via aseries of bidding rounds whereby renewableenergy projects bid for an (inflation-indexed) per-kilowatt-hour price for initially 8 and later 15years. Winning bids were selected by cost withineach technology category.

The result was a significant number of bids ineach of the auction rounds and falling bid costs ineach successive round.9 Connor (2003, 76)reports that in the five rounds of NFFO in Eng-land and Wales, onshore wind costs fell from 10pence (15.0 cents) per kWh in 1990 to 2.88 pence(4.3 cents) per kWh in 1998, with substantial fallsfor the other technology bands. Although NFFOwas successful in soliciting a large number ofcompetitive bids and in ensuring that any fundedprojects were cost-effective for electricity custom-ers, it failed rather spectacularly in one keyrespect: delivery of actual investment by the win-ning bidders.

Across the United Kingdom, between 1990and 1999, out of 302 awarded wind projects cov-ering 2,659 MW, only 75 projects were built,rated at 391 MW (Wong 2005). Spectacularly, notone of the 33 large wind projects awarded in thefifth round of NFFO in England and Wales wasever contracted. By contrast, out of 308 landfillgas projects awarded, 208 were operational in2008, with 458 MW of capacity out of 660 MWcontracted. For all the rounds of NFFO, out of933 awarded contracts, 477 were built, represent-ing 1,202 out of 3,639 MW (DECC 2009b, Table7.1.2). The primary cause for the failure was thatbidders were overoptimistic in their estimates ofthe actual delivery costs of the projects, oftenbecause the nature of the least-cost auction—withno assessment of likelihood of delivery—

258 Michael G. Pollitt

incentivized minimization of expenditure on pre-paring realistic bids (Mitchell and Connor 2004).

In reviewing the failure of the NFFO policy,it is important to remember the context in whichit operated. Renewables were then a very lowpriority for UK government policy, and it was aperiod of a rapid switch from coal- to gas-firedpower. Prices and pollution, in terms of quantitiesof CO2, SOx, and NOx, fell substantially. Thefocus on market-driven investments was good forenergy and carbon-efficient combined heat andpower (CHP) investment in the industrial andcommercial sectors (Bonilla 2006; Harvey 1994;Marshall 1993), which had struggled prior to pri-vatization (Jarvis 1986). UK privatization was asignificant policy success in economic terms,especially when the benefits to the environmentare considered (Newbery and Pollitt 1997).

The privatization and market liberalizationpolicies ensured that the United Kingdom wouldeasily meet its Kyoto targets for 2012 without anyfurther action, which was not the case for otherleading European countries. The mood at thetime was nicely summarized by a governmentminister for energy in 1988, Michael Spicer, whowrote that “privatisation of the electricity supplyindustry should boost the commercial prospectsfor these [green] technologies as a free market isestablished” (Elliott 1992, 266). Indeed, Friendsof the Earth was optimistic that the opening up ofthe residential energy market to competition in1998–1999 would give rise to demand for greentariffs and stimulate the production of greenenergy (Stanford 1998). It was only as the EUmoved toward substantial targets for renewablesthat it became clear that the United Kingdomneeded a policy that delivered large quantities ofrenewables.10 Nevertheless, significant lessons canbe learned from the NFFO experience.

Somewhat surprisingly, little quantitativeanalysis has been done on the bids that were suc-cessful under NFFO and the factors in their suc-cess and failure. Elliott (1992, 267) criticized theNFFO scheme as a “somewhat half-heartedhybrid market/interventionist system” that“would still leave short-term price and marketfactors to shape important long term strategicchoice concerning patterns of technological

development.” Institutional barriers emergedearly on as a critical factor in successful projectimplementation (McGowan 1991).

In particular, it became clear that projects hada problem with gaining the necessary consentsrequired to start building, known as “planningpermission” in the United Kingdom, and that alack of attention was given to proper environmen-tal impact assessments (Coles and Taylor 1993).Hull (1995) noted that in the early years, less thanhalf of all councils, the local government bodiesresponsible for consents, had planning guidancefor renewable energy projects, and more impor-tant, there was a lack of learning among councils.Calls came for clearer guidelines for the planningprocess to facilitate wind power (Roberts andWeightman 1994). Early industry views of thescheme were positive, recognizing that it did con-stitute a significant increase in expenditure overprevious levels (Porter and Steen 1996). However,the successive rounds of auctions were thoughtnot to provide assurance of continuity of supportfor renewables generally (Elliott 1994; Mitchell1995), and some worried that although they sup-ported near-market technologies, declines inR&D expenditure were bad for less advancedtechnologies such as marine (Elliott 1994).

The final years of NFFO, 1999–2001, coin-cided with a sharp decline in wholesale electricityprices as significant amounts of new gas-firedcapacity came into the market and competitionincreased within the initially duopolistic genera-tion sector (Evans and Green 2003). NFFO gen-erators had made overoptimistic bids, and theirsituation was exacerbated by the end of the com-pulsory wholesale power pool, which had guaran-teed the pool price to all generators, in March2001. It was replaced with a contract market and abalancing market. Imbalance between supply anddemand for an individual generator was nowmore likely to result in a financial penalty. Inter-mittent renewable generators were more likely toneed to participate in the balancing market to bal-ance their physical and contractual positions;because of the exogenous effects of weather, windgenerators have less capacity to match supply anddemand than fossil-fuel generators, who canadjust their spinning reserve. This is not necessar-

UK Renewable Energy Policy since Privatization 259

ily inefficient, however, as generators should beincentivized to solve the imbalance problem. Theimpact of this effect seems to have subsided afterone year of operation of the new arrangements,partly as a result of the arrival of a more generoussubsidy regime when Ofgem (the independentUK agency responsible for electricity and gasregulation) found little evidence of negativeimpact from the change to the trading system onrenewable generators (see Ofgem 2002).

The Renewables Obligation Scheme

The Renewables Obligation (RO) scheme, whichreplaced NFFO in 2002, uses a form of tradablegreen certificates (TGCs), known in the UK asRenewables Obligation Certificates (ROCs).Under this plan, the government set a minimumshare of electricity to be acquired by electricitysuppliers from renewable sources (larger hydro-electric schemes in operation before 2002 are

excluded). This share is steadily increasing from2002 to 2015 (see Table 13.4). Under the ROscheme, electricity suppliers must acquire thesecertificates in the prescribed target share ofrenewable generation for each annual period.They can do this by buying or earning ROCs,which are created when renewable generatorsgenerate electricity.This essentially splits the mar-ket into two parts, renewable and nonrenewable,with renewable generators getting a price for theROCs they create plus the wholesale price ofpower.11

The UK scheme has two important featuresintroduced at its inception, however. One is abuyout price (i.e., a penalty price) for ROCs ifnot enough are created by renewable generation.This price is specified for each trading period andeffectively caps the price that creators of ROCscan receive. The other is recycling of the revenuecollected from the buyout sales of ROCs. This

Table 13.4. RO targets and delivery against targets

Target renewableshare in GBa

% delivery in UK Nominalbuyout price(£/MWh)

Total costb

(£ million)

2002–2003 3.0 59% 30 282

2003–2004 4.3 56% 30.51 415.8

2004–2005 4.9 69% 31.59 497.9

2005–2006 5.5 76% 32.33 583

2006–2007 6.7 68% 33.24 719

2007–2008 7.9 64% 34.3 876.4

2008–2009 9.1 65% 35.36 1,024.6

2009–2010 9.7 37.19

2010–2011 10.4 + inflation thereafter

2011–2012 11.4

2012–2013 12.4

2013–2014 13.4

2014–2015 14.4

2015–2016 15.4 Estimated: ~1,733(2008–2009 prices) assumingno demand growth

Sources: OPSI, 2009; and Renewables Obligation annual reports from Ofgem various dates.Notes: From 2016, the share was fixed at 15.4% until 2027, now extended to 2037 for new projects; RO scheme cost is total cost includingrevenue recycling; £1 = $1.50 as of this writingaTarget share lower in Northern Ireland, but NI ROCs are tradable throughout UK. There is also a nominal distinction between Scottish ROCs(SROCs) and English and Welsh ROCs (ROCs), but these are tradable, and both are included in the GB target share.bWe report costs based on multiplying the buyout price by the actual ROC requirement. There appear to be small discrepancies in the actualreported payments and this figure in Renewables Obligation annual reports from Ofgem.

260 Michael G. Pollitt

takes the form of allocating the revenue back tothe creators of ROCs in proportion to thenumber they created.

The renewable energy industry was very posi-tive about the new incentive mechanism (Hill andHay 2004). So they should have been, because thescheme is very generous. Thus for example in2007–2008, the buyout (penalty) price was£34.30 ($51.45) per MWh, and only 64% of therequired ROCs were created by generators,meaning the buyout price was binding in the cer-tificate market. The total payment by supplierswas the target quantity of renewables multipliedby £34.30 ($51.45) per MWh. This meant that36% of the total ROC payment made by supplierswas available to be recycled and was divided pro-portionally among the generators who createdactual ROCs. Accordingly, for each ROC actuallypresented, the renewable generators received£34.30 plus £18.65 ($27.98) (i.e., an additional36/64 times £34.30 less costs of the scheme)Thissum is in addition to the wholesale cost of power.As the total cost to suppliers of the ROC schemewas £876 million ($1,314 million), this impliesthat consumers overpaid, relative to what wasnecessary to secure the renewable generationactually supplied, by at least the value of thebuyout revenue of around £315 million (36% of£876 million [$1,314 million], or 1% of the totalelectricity expenditure of £30.7 billion [$46 bil-lion] in 2008) (DUKES 2009).12 Interestingly, thegovernment collects the associated ROC pay-ments on the generation contracted under NFFOvia the NFFO fund, which creates a surplus abovethe payments to generators under that program;this surplus is estimated to be around £200 mil-lion ($300 million) per year (Tickell 2008).

The RO scheme is curious for two reasons.First, it relies on underdelivery to trigger themaximum subsidy amount. If the target numberof ROCs (or more) were presented, then theprice would drop to zero. Second, in the case ofunderdelivery, the maximum amount of subsidy ispaid to those actually creating ROCs. Thus thescheme assumes failure to meet the target andensures that a fixed total subsidy is paid, given this,regardless of how few ROCs are created.

The scheme is further complicated by theintroduction of “banding” starting on April 1,2009 (see Table 13.5). This changes the exchangerate to ROCs of some renewable generation:established technologies will get less than 1 ROCper MWh, newer more. This change breaks thelink between the total number of ROCs and theshare of renewable energy generation and willpresumably result in a reduced amount of elec-tricity being produced from renewables if thescheme is fully successful (if the share of high-exchange-rate technologies were to take off, as itmight with offshore wind). The Carbon Trust

Table 13.5. Banding of ROCs from April 1, 2009

Generation type ROCs per MWh

Landfill gas 0.25

Sewage gas0.5

Cofiring of biomass

Onshore wind

1.0

Hydro

Cofiring of energy crops

Energy from waste with CHP

Cofiring of biomass withCHP

Geopressure

Standard gasification

Standard pyrolysis

Offshore wind

1.5Biomass

Cofiring of energy cropswith CHP

Wave

2.0

Tidal stream

Advanced gasification

Advanced pyrolysis

Anaerobic digestion

Energy crops

Biomass with CHP

Energy crops with CHP

Solar photovoltaic

Geothermal

Tidal impoundment—tidalbarrage

Tidal impoundment—tidallagoon

Source: DECC 2009d

UK Renewable Energy Policy since Privatization 261

(2006) recommended the move to banding torecognize the different stages of development thatthe technologies had reached, and hence thehigher learning benefits associated with increasedfunding to earlier-stage technologies. Oxera(2005) points out the cost implications of allowingNFFO plants to earn ROCs once their NFFOcontracts expired (around £620 million [$930million]), giving those projects unexpected addi-tional subsidy. Oxera calculated that as much ashalf the payment via ROCs was in excess of thatrequired to ensure that the funded projects wentahead, and that existing landfill gas projects didnot require any ROCs to be economically viable.

The scheme, as shown in the table, impliesthat the subsidy to offshore wind could beincreased by £26.47 ($39.71) per MWh (50% ofthe 2007–2008 ROC revenue) and to tidal by£52.95 ($79.43) per MWh (100% of the 2007–2008 ROC revenue). In the 2009 budget, off-shore wind was subject to an emergencyrebanding provision, which saw the offshore windROC band go to 2 for 2009–2010 and 1.75 for2010–2011, now increased back to 2 from 2010–2014.

Policy Costs and Deliveryunder NFFO and RO

Table 13.6 summarizes the financial commitmentsmade under the NFFO and RO schemes, as wellas a reference amount for the amount of publicR&D expenditure reported to the InternationalEnergy Agency (IEA). The increased significanceof the RO scheme is evident.

While the RO scheme is the most significantelement of the United Kingdom’s expenditure onrenewables, it is not the only element. Table 13.7is a summary offered in a ministerial answer to aparliamentary select committee question. It isnoteworthy that significant additional amounts arestill being spent by the taxpayer on supportingearlier-stage technologies outside the CO2 priceand RO support mechanisms. However, the orderof magnitude of energy customer support forrenewables is of the order of £1.8 billion ($2.7billion) in 2008, in addition to £400 million($600 million) by the taxpayer. This level of sup-

port is up 47% in real terms from the figure esti-mated by Wordsworth and Grubb (2003) of £1.3billion ($1.95 million) in 2002–2003.13

As the above discussion of the progress withthe RO scheme has made clear, the developmentof electricity from renewables has been disap-pointing in terms of overall cost relative to deliv-ery, given the United Kingdom’s resource poten-tial and ambitious targets.Table 13.8 gives the fig-ures in terms of total electricity generation. Anumber of features stand out. First of all, electri-city from biomass in 2008 is larger than that fromwind. Hydro remains significant within the UKrenewable portfolio. Connor (2003) reported esti-mates from 2002 that suggested the United King-dom would meet only two-thirds of its target levelby 2010. This still seems likely. However, thestriking thing about the 2002 estimates is that forbiomass, offshore wind, and hydro, they seemlikely to be met or exceeded, though not byonshore wind. The United Kingdom is failing tomeet its projections for renewables as predicted,

Table 13.6. Financial support (£ million) forrenewables in the United Kingdom (nominal)

R&D RO NFFO

1990–1991 14.7 6.1

1991–1992 17.1 11.7

1992–1993 16.1 28.9

1993–1994 15.2 68.1

1994–1995 9.1 96.4

1995–1996 9.1 94.5

1996–1997 6.2 112.8

1997–1998 4.3 126.5

1998–1999 3.3 127

1999–2000 4.6 56.4

2000–2001 4.4 64.9

2001–2002 6.1 54.7

2002–2003 10.5 282.0 -

2003–2004 11.6 415.8 -

2004–2005 19.7 497.9 -

2005–2007 36.6 583.0 -

2006–2007 49.5 719.0 -

2007–2008 41.6 876.4 -

Sources: UK government renewable R&D budget data from IEA2009; Mitchell and Connor 2004, 1943

262 Michael G. Pollitt

but this is largely due to the failure to deliver thelong-expected increase in generation fromonshore wind.

Both NFFO and RO have stimulated electri-city from landfill gas and cofiring of biomass andmunicipal waste (with fossil fuels). These tech-nologies were near market in the early 1990s andhad good prospects at that time. Brown andMaunder (1994) discuss the United Kingdom’spotential for exploiting landfill gas, and Jamasb etal. (2008a) explore the prospects for waste toenergy, noting it has significant further potential,especially if CHP is involved. The use of biomassfor cofiring in coal-fired plants continues to beone of the most sensible uses of biomass, as it iswell proven that mixes of up to 10% biomassrequire little adjustment to existing plants(Thornley 2006). Small hydro projects have also

had some success, with a steady increase in hydrogeneration from these schemes.These projects useestablished technology and have benefited frommarket-based support mechanisms. Paish (2002)highlights around 400 MW of further potentialfor small-scale hydro in the United Kingdom.

There also have been promising developmentswith offshore wind in the United Kingdom,assuming the actual delivered costs can be keptdown. As of August 2009, offshore wind capacityis currently 598 MW, but an additional 1,246MW are under construction, and a further 3,613MW have been consented. This contrasts with3,730 MW of onshore wind capacity, with only930 MW under construction and 3,275MW con-sented (BWEA n.d.).

It seems likely, given the continuance of highlevels of support via banded ROCs, that offshore

Table 13.7. Support for renewable energy in 2007–2009

Scheme Description Cost Paid by

Renewables ObligationCertificates

Electricity suppliers must buy a proportion oftheir sales from renewable generators or pay abuyout charge

£874 million in 2007–2008

Electricityconsumers

EU Emissions TradingScheme

Renewable generators indirectly benefit from theincrease in electricity prices as other companiespass the cost of emissions permits into the priceof power

Perhaps £300 million in2008, given current per-mit prices

Electricityconsumers

Carbon EmissionsReduction Target

Energy companies must install low-carbon itemsin homes, which could include microgenerationfrom 2008

Total cost will be £1.5billion over 3 years,mostly spent on energyefficiency

Gas and electricityconsumers

Renewable TransportFuel Obligation

Fuel suppliers must supply a proportion ofbiofuels or pay a buyout charge

No more than £200 mil-lion in 2008–2009

Consumers

Climate Change Levy Electricity suppliers need not pay this tax (passedon to non-residential consumers) on electricityfrom renewable generators

£68 million to UK genera-tors, £30 million to gen-erators abroad in 2007–2008

Taxpayers, viareduced revenues

Lower fuel duty forbiofuels

The rate of fuel duty is 20 pence (30 cents) perliter below that for petrol and diesel

£100 million in 2007 Taxpayers, viareduced revenues

Environmental Transfor-mation Fund

Grants for technology development and deploy-ment, including subsidies for installing renewablegeneration, planting energy crops, and develop-ing biomass infrastructure.

£400 million over 3 yearsstarting in 2008–2009

Taxpayers

Research councils Grants for basic science research £30 million in 2007–2008 Taxpayers

Energy TechnologiesInstitute

Grants to accelerate development (after the basicscience is known) of renewables and otherenergy technologies

Allocation and eventualsize of budget not yetannounced

Taxpayers and spon-soring companies

Source: House of Lords 2008, Table 6 Note: £1 = $1.50 as of this writing

UK Renewable Energy Policy since Privatization 263

wind will overtake onshore wind generation,albeit on the back of very disappointing deliveryof onshore wind projects.

Looking at the success of the NFFO and ROschemes, NFFO did well on cost of the policy butnot as well on quantity of renewables delivered,whereas RO did better on quantity delivered butmuch less well on cost of the policy.

Other Renewables Policies

While the main support mechanisms have favoredwind and biomass, direct government funding hasalso helped the marine industry. A resurgence inresearch and demonstration funding in the last 10years has resulted in some positive developments

(see Mueller and Wallace 2008).The first 1.2 MWtidal stream plant was installed in 2008 (Riddell2008), and the industry is well placed internation-ally to exploit this and related marine technolo-gies (Elliot 2009). The UK government is cur-rently conducting another feasibility study of the8.5 GW Severn Barrage, which could generate5% of the country’s current electricity demand.This is the biggest of the United Kingdom’spotential tidal projects (Conway 1986), but costand environmental issues remain to be addressed(see DECC 2009f). However, a trial with asmaller scheme first, such as a barrage across theMersey, would seem sensible for learning thatmight benefit the much larger Severn project.

Table 13.8. Renewable electricity generation (GWh) in the United Kingdom, 1990–2008

1990 2000 2001 2002 2003 2004 2005 2006 2007 2008

Wind

Onshore wind 9 945 960 1,251 1,276 1,736 2,501 3,574 4,491 5,792

Offshore wind 0 1 5 5 10 199 403 651 783 1,305

Solar photovoltaics 0 1 2 3 3 4 8 11 14 17

Hydro:

Small scale 91 214 210 204 150 283 444 478 534 568

Large scale 5,080 4,871 3,845 4,584 2,987 4,561 4,478 4,115 4,554 4,600

Biofuels:

Landfill gas 139 2,188 2,507 2,679 3,276 4,004 4,290 4,424 4,677 4,757

Sewage sludgedigestion

316 367 363 368 394 440 470 456 496 564

Municipal solidwaste combus-tion

221 840 880 907 965 971 964 1,083 1,177 1,226

Cofiring withfossil fuels

286 602 1,022 2,533 2,528 1,956 1,613

Biomass 0 410 743 807 947 927 850 797 964 1,155

Total Biofuels and wastes 676 3,796 4,493 5,047 6,174 7,364 9,107 9,288 9,270 9,315

Total Renewables 5,857 9,828 9,516 11,093 10,600 14,147 16,940 18,136 19,646 21,597

Total Generation 319,701 377,069 384,778 387,506 398,209 393,867 398,313 398,823 397,044 389,649

% Total Renewables 1.83% 2.61% 2.47% 2.86% 2.66% 3.59% 4.25% 4.55% 4.95% 5.54%

Wind 0.00% 0.25% 0.25% 0.32% 0.32% 0.49% 0.73% 1.06% 1.33% 1.82%

Hydro 1.62% 1.35% 1.05% 1.24% 0.79% 1.23% 1.24% 1.15% 1.28% 1.33%

Biofuels 0.21% 1.01% 1.17% 1.30% 1.55% 1.87% 2.29% 2.33% 2.33% 2.39%

Source: Digest of UK Energy Statistics, various issues

264 Michael G. Pollitt

PV has relied on direct government supportfor installation programs that have involved only asmall number of installations, mainly funded viathe government’s Industry Department (DTI,then BERR) under the Low Carbon BuildingsFund. This funding has installed only a few hun-dred PV systems. The degree of satisfaction withthe technology among the recipients of fundinghas been positive (Faiers and Neame 2006), but alack of significant sums of money and properassessment of the learning from the policy hasbeen noted (Keirstead 2007). This is in spite of awell-regarded R&D plan for solar being put inplace in the 1990s (Stainforth et al. 1996) andwork showing that significant community instal-lations of solar would not pose any local gridproblems (Thomson and Infield 2007). The gov-ernment has made two very recent changes to itsrenewables policy, which are relevant to anyassessment of the need for reform of the currentarrangements (allowed for in primary legislation(OPSI 2008b)).

First, a feed-in tariff (FIT) for small-scale low-carbon generation commences in April 2010 (seewww.fitariffs.co.uk/). This will be for renewableelectricity generation up to 5 MW and fossil-fuelCHP up to 50 kW. Meant to encourage PV,small-scale wind (including microwind),microhydro, and micro-CHP, this policy respondsto industry concerns about the lack of ambition inmicrogeneration policy (Lupton 2008).

The second policy is a Renewable HeatIncentive (RHI) (see www.rhincentive.co.uk).This has the potential to be a significant policycovering all scales of production: household,community, and industrial. It is intended to drivethe share of renewable heat to 14% (though this isnot a firm target) up from 0.6%. It could cover airsource heat pumps, anaerobic digestion to pro-duce biogas for heat production, biomass heatgeneration and CHP, ground source heat pumps,liquid biofuels (but only when replacing oil-firedheating systems) and solar thermal heat and hotwater.

The scheme is not finalized at the time ofwriting and is due to commence in April 2011.

An Assessment of RenewablesPoliciesA 20-year view of UK renewables policy suggestsa failure to translate the country’s early resource-based promise into actual delivery of renewableenergy. It would be wrong to suggest widespreadpolicy failure, however. The United Kingdom ismaking progress on decarbonization and hasstrong and increasingly comprehensive policies inplace, covering electricity, heat, and transport (viapolicies toward electric vehicles and biofuels).

Two points are worth making at this stage.First, renewable energy policy remains an expen-sive gamble for all countries. Second, it is unclearwhat part particular renewable technologiesshould play in decarbonization to 2050.

As Helm (2002) has pointed out, a sensiblyhigh and stable price of carbon is the startingpoint for all economically feasible decarboniza-tion policies. In the absence of this, it is virtuallyimpossible to establish proper signals for maturetechnologies and near-market technologies,whose response to the proper price signal deter-mines how fast the country needs to accelerateless developed technologies. This is particularlytrue for nuclear, CCS, and demand reductioninvestments, many of which are being delayed bylow, volatile, and uncertain prices for carbon.TheUnited Kingdom, with its diversified energy sys-tem, exposure to world energy markets, andopenness to both nuclear and CCS, has keenly feltthe lack of a proper carbon price signal.

As Nelson (2008) discusses, the failure to set asufficiently tight cap on CO2 at the EU levelmakes UK renewables policy meaningless as apolicy for decarbonization. More renewable elec-tricity generation within the EU Emissions Trad-ing System (ETS) simply causes fuel switching inthe fossil plants from gas to coal, not to mentiondelaying nonrenewable low-carbon investmentsin CCS and nuclear. In this context, UKrenewables policy has been somewhat conserva-tive with respect to funding levels under NFFOand to renewable energy targets under the ROand, until recently, unwilling to pick winners. AsEikeland and Sæverud (2007) point out, however,

UK Renewable Energy Policy since Privatization 265

the ending of the United Kingdom’s status as anenergy exporter in 2003 and the associated rapiddecline in oil and gas reserves have led to areawakening of energy security concern as amajor driver of UK energy policy.This is likely toexplain substantially increased interest in deliver-ing more domestic renewable capacity.

Failure to deliver large quantities ofrenewables so far is not a particular issue, in thatdelay will probably mean lower costs of exploita-tion (resulting from learning by doing elsewhereand learning by research) when they are finallyexploited. The unfortunate aspect of the RO sys-tem is its failure to deliver cost-effectively therenewables that it has delivered. This has been aserious design flaw, and the inability of the UKgovernment to learn and correct the flaw does notbode well for any other long-term mechanism putin place to support renewables. Nevertheless,given the targets for delivery that exist within thescheme, it is clearly important to consider whythe scheme has not delivered the quantity ofrenewables intended.The failure of the scheme todeliver overall lies squarely with one particulartechnology: the failure to deliver sufficient quan-tities of onshore wind.

Onshore Wind and the Planning Problem

The standard reason given for the delivery failureis difficulties in getting new wind farms throughlocal planning processes. Whereas conventionalpower plants can easily be built on existing sitesand require national-level planning consents,wind farms are often small in terms of MWcapacity and require local planning permission ifless than 50 MW, which covers most onshoreinstallations.14 Evidence has consistently shownthat gaining planning permission is a seriousobstacle to the development of wind farms or,more precisely, that the costs of obtaining permis-sion are often prohibitive in terms of imposeddelays, negotiation costs, and planning restrictionson the precise nature of the final investment.

In the United Kingdom, local planning deci-sions typically involve an applicant, such as a windproject developer, making a planning application.This includes the submission of detailed plans and

an impact assessment to the relevant local govern-ment authority.The application is initially assessedby a local planning officer, who makes recom-mendations on the plans to the relevant group ofelected local councilors for the area, who in turnvote on the proposal. Plans would be available forpublic consultation, and objections could beraised during the review period. Planning applica-tions can be granted subject to conditions andobligations. This process might result in a numberof iterations in the plans. Should permission berefused, the applicant can appeal the decision, inwhich case a costly public inquiry would ensue.The relevant central government department alsohas the right to disallow a locally approved plan-ning application so objectors can appeal to therelevant government minister. At the nationallevel, plans need to be submitted to the relevantgovernment department for referral to the secre-tary of state for final decision. Objections can beraised to these plans according to the planningguidelines. This national-level process is beingstreamlined, as below.15

The average time for local and national plan-ning decisions on onshore wind in 2007 was 24months, with approval rates of 62% (Chamberlain2008, 21). For large projects, the Ministry ofDefence, National Air Traffic Control, and civilairports were major objectors. Attempts have beenmade since 2007 to obligate local councils to settarget levels of energy from renewables for newdevelopments. The 2008 Planning Act (see OPSI2008c) allows for setting up an InfrastructurePlanning Commission to decide on large onshorewind farms (greater than 50 MW) as well as largeoffshore projects (greater than 100 MW) (seeNAO 2008, 40–41, for a discussion).

The literature has dug more deeply into theplanning problem. Hedger (1995) highlights thatwind power development involves a clash of plan-ning cultures: land use versus energy supply. Thefirst is fundamentally local, participatory, and con-cerned with preserving rural landscapes; the sec-ond is fundamentally national, top-down, andconcerned with delivering technological solutionsto national energy supply requirements. Thesecultures were bound to clash in onshore windpower development.

266 Michael G. Pollitt

Mitchell and Connor (2004) stress that theemphasis on cost minimization, combined withthe tying of subsidies to particular locations andplans, meant that many successful NFFO bidsfailed to get through the planning process. Thiswas because the bidders were not able to invest inlocal engagement or respond to the outcome ofthe engagement process by modifying their pro-posals. Indeed, the competitive nature of NFFOmeant that often the bidders had to keep prospec-tive locations secret and did not engage in localconsultations prior to bidding. Toke (2005b)found that for the projects he examined from thethird through fifth NFFO rounds in England andWales, 47 were granted planning permission, 47refused planning permission, and 96 did not makeor complete an application.16

The main reasons given for planning objec-tions were visual amenity impairment and worriesabout noise (Eltham et al. 2008). These gave riseto concerns about economic effects on houseprices and tourism. The United Kingdom is adensely populated island, with many areas oflower population and high ground located innational parks or other places that attract tourists.Increasing numbers of residents or second-homebuyers have been moving to such areas for theirvisual amenities rather than employment reasons(see Strachan and Lal 2004 for a discussion of thedebate around tourism). The decline of employ-ment in farming and rural industries has reducedthe scope for arguments based on the smallnumber of permanent jobs that might be createdin the energy sector, because increasing percent-ages of people living in the countryside work innearby conurbations and are not looking foremployment in local industries.

Rural environmental protection and localcommunity action groups thus had strong incen-tives to organize opposition to individual windfarm projects, although in some cases tourismactually increased after wind turbines wereinstalled, and the noise from a modern turbinethat is 500 meters away is no more than in a quietbedroom (Strachan and Lal 2004). A number ofstudies (e.g., Eltham et al. 2008; Warren et al.2005) have shown that attitudes to wind farmsconsistently improve after construction, with

many people’s fears not being realized. It is alsotrue that in general, majority support exists fornew wind farms, but there are a significantnumber of both local and nonlocal objectors togiven schemes (Warren et al. 2005). This suggestsa social gap or democratic deficit at the local levelthat needs to be overcome (Bell et al. 2005) inorder to connect national policy delivery withlegitimate local concerns.

Rather surprisingly, little systematic study hasbeen done of success rates in individual localauthority areas or by individual developers orownership type. OnlyToke (2005b) has attempteda regression analysis, looking at planning permis-sion acceptance and refusal for wind projectsbased on a sample of 51 proposals. Among hisfindings is that if the local planning officers (whoprocess applications and make recommendationsto the local councilors who vote on the applica-tion) object, then projects are almost alwaysrefused, whereas if they accept a project, it is likelyto go through on appeal.Toke also finds that if theCampaign to Protect Rural England, which cam-paigns “for the beauty, tranquillity and diversity ofthe countryside” (CPRE n.d.), objects to aproject, it is likely to be opposed by the localparish council. One developer, Wind Prospect(2008), which has a joint venture with EDF, amajor energy company, to develop onshore windfarms in the United Kingdom, has invested heav-ily in local consultation and seems to have beenmore successful in gaining planning permission(see Toke 2005b). Active community involvementhas led to successful development in some cases,particularly when the community owns shares inthe wind farm, but these are small in capacityterms.17 However, under both NFFO and RO,there has been an unwillingness to actively involvecommunities in co-ownership of onshore winddevelopments, possibly because of the dominanceof large power companies within the UK windpower sector and the high transaction costs ofsuch engagement.

Overall, it is difficult to tell whether the fullcost of developing wind power onshore is actuallymuch higher than it would appear, given thesocial value of the UK countryside, or whether afeasible redistribution of the current benefits

UK Renewable Energy Policy since Privatization 267

toward potential local objectors would be enoughto solve the planning problem. Bergmann et al.(2008) use willingness-to-pay modeling of a sam-ple of rural and urban dwellers in Scotland. Whileboth groups value reduced environmental impactfrom power generation highly, the authors findthat urban dwellers are willing to pay more for anoffshore wind farm than for an equivalent largeonshore wind farm and value the rural employ-ment opportunities less than do rural people. Theactual construction costs of wind farms in theUnited Kingdom are difficult to come by, but theinformation that is available suggests thatsimulations of the likely penetration of newprojects are still based on optimistic assumptionsthat wind costs will be much cheaper than theycurrently are.18 High actual costs may therefore bea factor delaying investment. The achieved loadfactors for the whole UK wind portfolio in 2008were 27% for onshore and 30.4% for offshore(DECC, 2009b, 206) in contrast to higherassumptions in some calculations (e.g., Dale et al.2004, who assume 35% for both onshore and off-shore wind).

No doubt smaller, more local developmentswould facilitate reduced planning objections, butthey would come with their own higher costs.The move to FITs for such smaller developmentsshould help increase the number of such projects.However, in examining scenario rankings fromdifferent wind actors in northwest England,Mander (2008) finds that expansion of offshorewind was the only part of a wind strategy thatboth pro-wind and pro-countryside lobbies couldagree on, even if onshore wind became morecommunity-driven. Attempts to streamline theplanning process have been made, with significantreforms to the appeals process in 2003 (Toke2003), giving more power at the national level;nevertheless, there is clearly still an issue of gettingpermission. Attempts in 2005 to streamline theplanning process in Wales (under a devolvedadministration) have had mixed success (Cowell2007). The Welsh Assembly designated “strategicsearch areas,” which were assessed to be moresuitable for large wind farm developments andhence more likely to be approved on appeal.These proved controversial, with both pro- and

anti-wind lobbies. The wind developers wereunhappy that many proposed schemes lay outsidethe designated areas, and anti-wind groups wereunhappy with where some of the boundaries ofthe acceptable areas were drawn.

BiomassBiomass is likely to be the second-largest renew-able energy source out to 2020 in the UnitedKingdom. Biomass is frequently cited as a signifi-cant, albeit finite contribution to UK decarbon-ization (of the order of up to 5%) (seeTaylor 2008for a review). Biomass policy toward waste hasbeen largely successful because of the near-marketnature of the technology and its responsiveness toboth NFFO and RO subsidies. The direct burn-ing of biocrops has also been successful, given theemerging global market in tradable biomass fromcountries such as Brazil, Canada, and the UnitedStates (Junginger et al. 2008).

Government support for local biocrop plantshas proved problematic, however, given the tech-nological, planning, and economic constraints. Ahigh-profile project involving local biomass andnew technology failed as a result of financing con-cerns (Piterou et al. 2008), and it is difficult tojustify the use of local biocrops for anything otherthan direct burning in existing coal-fired powerstations in direct competition with internationallytraded biomass, which is usually produced moreefficiently abroad. Nevertheless, some focusgroup studies have suggested that there is publicsupport for the use of local biomass in small CHPplants and skepticism about the overall GHGimpact of the use of internationally tradedbiomass (see Upham et al. 2007).

It is not environmentally sensible to use localbiocrops to produce biofuel in the United King-dom. Local biocrops produce more GHG impactwhen directly burned to produce power and heat(Hammond et al. 2008). Indeed, in the longerrun, the current use of biofuels to blend withpetrol and diesel may be phased out as the vehiclefleet is electrified (for current use, see Bomb et al.2007). The difficulty of making a sensible indus-trial policy argument for a local crop-dedicated

268 Michael G. Pollitt

biomass power plant within a viable long-rundecarbonization strategy is helpfully discussed byvan der Horst (2005). Indeed, Slade et al. (2009)criticize UK bioenergy policy as being character-ized by lots of initiatives but with a lack of clarityas to precise objectives to be delivered. If thecountry were to rely on internationally tradedbiomass as its key input, this would require bettercertification as to the source of the biomass (vanDam et al. 2008).

Bioenergy, with its complicated supply chain,displacement impacts, and total production cyclesustainability impacts, requires proper pricing ofall its environmental effects, including GHGs andlocal pollutants, in order to calculate whether it isworthwhile (Elghali et al. 2007). The life cycleGHG impact of biocrops (i.e., the impact on theamount of carbon stored in the stock of growingcrops) is further complicated by the carbon stor-age impacts of increasing the area set aside forgrowing them (Cannell 2003).

UK Performance versus That ofOther CountriesThe discussion so far indicates that comparativeassessment of UK policy on renewable energywould not be straightforward. It is clear that theUnited Kingdom has pursued a successful decar-bonization strategy to date and that relative suc-cess has been achieved in several areas, both inresponding to price signals and in developing newtechnologies for deployment in the country. Theone area of failure is in deployment of onshorewind at least cost. The net environmental impactof this failure is currently zero, given that theUnited Kingdom is on course to meet its GHGreduction targets. Still, this environmental per-formance could have been delivered at lower cost.The excess costs of the current set of policies arehard to estimate, given the diversity of supportinstruments. However, a lower-end estimatewould be the amount of revenue recycling withinthe RO mechanism, as this overpayment seemslargely unnecessary to deliver the observed quan-tity of renewables connected to the electricity sys-

tem.19 This excess cost is significant and rising.Nevertheless, it remains small compared with thehigh cost of the renewable deployment strategiesof some other countries, such as Germany andSpain, which have not allowed them to meet theirGHG reduction targets.

It is fashionable to suggest that the root causeof the problem of underdelivery of onshore windis the use of a tradable green certificate (TGC)scheme rather than a FIT, as used in Germany andSpain (see, e.g., Butler and Neuhoff 2008;Jacobsson et al. 2009; Lipp 2007; Meyer 2003;Toke 2005a; Toke and Lauber 2007). A more bal-anced assessment by the International EnergyAgency (IEA 2006) of the UK renewable energypolicy points out that TGCs have worked well (atleast to the date of the IEA’s assessment) in anumber of jurisdictions, such as Texas, Sweden,Australia, and New Zealand. It is only in theUnited Kingdom where they seem to have mani-festly failed to deliver the intended capacity.

Two common theoretical arguments havebeen made for the superiority of FITs overTGCs.One is that by offering a fixed price per kWh todevelopers, this allows new renewables to befinanced more easily.The other is that FITs attractlarge quantities of renewables because these arenot limited to the most attractive sites.

The first argument is well put by Mitchell etal. (2006), who maintain that the UK RO schemeexposes renewables to price, volume, and balan-cing risks, rather than just volume risks as under aFIT. Although this clearly does impose costs, it isnot clear that it is suboptimal or that it explainsnondelivery against the United Kingdom’srenewables targets. Higher risk is relevant to non-delivery where development is small-scale and thedevelopers have little or no credit history; herethere may well be a significant market failure inthe market for external finance. However, it israther a weak argument when the ultimate devel-opers are mostly large multinational companiesmaking portfolio investments, and when mostROC credits are bought by the six multinationalsupply companies who dominate the UK market,each with generation interests and the option toinvest directly in renewable capacity.

UK Renewable Energy Policy since Privatization 269

The second argument makes less theoreticalsense, because it is not clear why developing themost attractive sites first is not desirable in anycase. The quantity of renewables forthcoming isclearly accelerated by offering initially highreturns, but offering a margin for renewables toattract investors is a function not of whether thesubsidy regime is a FIT or TGC, but of how largea quantity of renewables is required under eitherscheme. TGCs can set ambitious targets, as in theUnited Kingdom, and can deliver attractiveprices. Low prices for renewables are not a prob-lem with the ambitious RO targets.

In the end, the question becomes whether theUnited Kingdom would have delivered moreonshore wind capacity had there been a FIT forwind energy. For community schemes, the answeris quite possibly yes, because the uncertainty ofindividual project cash flows may well have beenan issue for funders. However, for larger schemeschiefly owned by multinational energy compa-nies, it is hard to say.The problem has clearly beenrelated to planning permission, and it is not obvi-ous how changing the funding regime improvesthe prospects for gaining planning permissionunless it is more generous and offers scope forproviding attractive payments to the local com-munity.

The literature seems to suggest that two morefundamental dimensions are of interest to explainthe differences in delivery of onshore wind amongthe United Kingdom, Germany, Spain, and Den-mark: land use constraints and local involvementin ownership, such as via local cooperatives orfarmers (see Table 13.9).

Local ownership, which is very high in Den-mark and also notable in Germany, is a determi-nant of successful strategic deployment in thesecountries (Szarka and Bluhdorn 2006; Toke2007). This is important because these two coun-tries face similar land use constraints to the UnitedKingdom. The development in Spain, however,has occurred with similar ownership of windassets by multinational companies, but in the con-text of very little land use constraint (Toke andStrachan 2006). Thus it seems clear that thesecountries have different institutional and physicalstarting points than the United Kingdom.

Econometric modeling by Soderholm andKlaassen (2007) of diffusion rates of wind poweracross Europe confirms that the United Kingdomhas lower diffusion (penetration) relative to othercountries, and that FITs do tend to be more suc-cessful in encouraging diffusion, but that a givenFIT would likely have less of an impact here thanin Germany.

What is clear is that the financial cost of windpower delivered onshore is unnecessarily high inthe United Kingdom. Butler and Neuhoff (2008,1856) show that while the NFFO schemes didresult in much lower support prices for wind inthe United Kingdom than in Germany, they werenot that much lower once adjusted for the qualityof the underlying wind resources. Under the RO,renewable support costs are estimated to havebeen twice as high in 2006 as they would havebeen under a German support tariff applied toUK wind resources (which would have beenlower than the actual tariff in Germany). Toke(2005a) shows that the RO scheme with revenue

Table 13.9. Differences among leading wind countries in Europe

1,000 mi2 landper millionpopulation,2009–2010

% onshore wind ownedby utilities/

corporations

% owned byfarmers

% owned bycooperatives

Wind cap-acity (MW),end 2008

UnitedKingdom

1.5 98 1 0.5 3,288

Germany 1.7 55 35 10 23,903

Spain 4.3 > 99 < 0.5 0 16,740

Denmark 2.9 12 63 25 3,160

Sources: Wikipedia, List of Countries and Dependencies by Population Density (accessed 26 March 2010); Wind Power 2009; Toke 2005a

270 Michael G. Pollitt

recycling was more expensive per kWh than theGerman FIT following reductions in the size ofthe FIT in Germany.

Looking at Spain, where large utilities havedominated in ownership of wind generation simi-larly to the situation in the United Kingdom,Stenzel and Frenzel (2008) note the positive reac-tion of incumbent Spanish companies to windpower development in Spain in contrast with thatin Germany. They highlight the importance ofcorporate self-interest in promoting wind powerdevelopment. Wind power in Germany devel-oped in spite of opposition from German utilities,which were forced to accommodate renewablesand bear the costs of connection to the grid. InSpain, this has led the corporate generators tosupport investment in better prediction of windspeeds at individual wind farm sites in order tobetter manage the grid. In Germany, however,significant costs have been imposed on the trans-mission system that are not reflected in the con-nection incentives of wind developers. This hasled to grid management issues in Germany, whichwill become more costly to deal with as windcapacity increases (Klessmann et al. 2008). It iseven possible to suggest that the continuation ofthe grip of incumbents on the German powermarket is in significant part because of the unwill-ingness of the German government to liberalizethe market fully, for fear of undermining the abil-ity of the incumbents to finance the significantreinforcement costs associated with renewablesexpansion.

In 2008, the United Kingdom had around13.2 GW in 195 projects that were in Great Brit-ain’s “GB Queue” (see Ofgem 2007a).These wereprojects that wished to connect to the power grid,but for which no firm connection right could beoffered, unlike under the German FIT, whererenewable capacity must be connected and paidfor generated power (see Swider et al. 2008). TheUK government has suggested that this is one ofthe barriers to the rollout of renewables (DECC2009e). This may explain some of the slow deliv-ery of renewable wind connection in the UnitedKingdom, but it certainly does not explain themost significant part of it. It is impossible to tellhow economically viable many of the projects in

the GB Queue are, and Ofgem has identified onlyaround 450 MW of wind capacity that needs tobe prioritized via accelerating transmission invest-ment (see Ofgem 2009a). It is also the case thatnew renewable connections should face the truecosts of connection to the grid and capacity, andthey should come onstream when it is at leastsystem cost, rather than only least generation cost.Nodal pricing would seem to be a more appropri-ate way of signaling this, rather than the “connectand manage” approach under FITs in Germany(see Pollitt and Bialek 2008).

The correct pricing for transmission capacityalso points to the need for the United Kingdomto look closely at the efficiency of utilization oftransmission assets and their operational criteria.The GB transmission system in general operatesunder an N-2 safety standard, wherein the systemmust be operated in such a way that if a major linkfails, it must be capable of handling anothersimilar-size failure.This gives rise to lower rates ofutilization of transmission grid assets than incountries with an N-1 safety standard and givesrise to less use of automatic voltage control equip-ment. This suggests that there is scope for operat-ing the assets much more smartly in the presenceof large-scale renewables. For instance, the nomi-nal rating of Scotland–England interconnectors isaround 7 GW, whereas the declared capacity is 2.2GW; this suggests that transmission constraintscould be made less in practice than they might beon paper. Ofgem’s recent LENS scenariomodeling (Ault et al. 2008) of the electricitytransmission and distribution networks suggeststhat a range of network sizes and capabilities arepossible by 2050, depending on how and wherenew generation capacity, including renewables,was connected.

Looking to other countries with TGCschemes, it is quite clear that Sweden, Australia,and New Zealand have avoided the problems ofoverpayment that characterize the UK ROscheme, and these jurisdictions have significantlyfewer land use constraints. Kelly (2007) discussesthe UK scheme in contrast to those of Australiaand New Zealand. The Australian scheme, com-plemented by an Office of the Renewable EnergyRegulator (see ORER n.d.), has much less ambi-

UK Renewable Energy Policy since Privatization 271

tious targets than the UK scheme but does nothave any revenue recycling. The New Zealandscheme has higher targets than Australia’s but isvoluntary.The Swedish scheme also does not haverevenue recycling and is combined with carbontaxes throughout the economy (see SwedishEnergy Agency n.d.). The United Kingdomwould do well to examine the overall carbonreduction incentives in Sweden.

Szarka (2006) raises an important issue aboutpolicy comparison across countries in the case ofrenewables, suggesting that policy should beaimed at paradigm change, not just installedcapacity. Clearly what matters is where the coun-try ends up in terms of decarbonization, and whatis required is radical change to the UK energysystem. He maintains that the real success of Ger-man policy has been to engage large numbers ofindividuals in taking action on climate change, asinvestors in local wind farms.This is an importantperspective, because it suggests that the real failureof UK policy is not gaining practical support forthe sorts of changes to the energy system that arerequired. Failure to focus on this aspect of theproblem has led to an ineffective policy onrenewables deployment, which will be moreexpensive than it need have been, due to a com-bination of underdelivery and overpayment.

Another issue is the stability of policy throughtime. A concern of UK policymakers in setting upthe RO scheme was to introduce stability in thesubsidy regime over a long period, in contrast tothe stop-start nature of NFFO. However,although stability is a desirable goal in itself, thishas been an excuse for not facing up to the seriousdeficiencies of the RO scheme. Little evidence isavailable to indicate that the United Kingdom hashad a less stable policy toward renewables thancountries with high penetration rates ofrenewables, such as Denmark, Germany, andSpain, where responses to incentives were rapidand significant changes have occurred to supportpolicy over time.

What Might Be Right for theUnited KingdomIf a problem exists with the delivery of onshorerenewable capacity in the United Kingdom, whatshould be done about it? Answering this questionrequires attention to the institutional context ofthe United Kingdom (following Rodrik 2008).The country’s policy context is a liberalized mar-ket for a relatively small island with concernsabout fuel poverty, global warming, and energysecurity. It is clear that what is needed is a policyconsistent with a liberalized energy market andwith environmental targets. By contrast, Germanyis much less committed to liberalized energy mar-kets. It also has much more of a focus on a greenindustrial policy aimed at promoting the manu-facture of wind turbines for export. Although theUnited Kingdom has paid lip service to this sortof objective, the reality is that only 4,000 jobs inthe country depend on the wind productionindustry; even in Germany, the figure is only38,000 (EWEA 2009). It is quite clear that for anindustry requiring around £1 billion ($1.5 bil-lion) of subsidy per year, this is not a cost-effectivejob creation scheme.

The focus should rather be on least-costachievement of environmental targets, which willbe much more important for the competitivenessof the UK economy and for incomes and employ-ment. The current RO scheme is clearly far toogenerous to existing onshore wind, and it doesnot guarantee cost-effectiveness for offshore windand marine energy. It is also important that theaim of long-run cost reduction for technologiesthat are currently not cost-effective be main-tained, and that these technologies compete withnuclear and CCS projects in a reasonable timeframe. An important starting point for this is thecreation of a single high and stable carbon pricethroughout the economy. This would immedi-ately give clear signals to nuclear and CCS andprovide the backstop technologies against whichcontinuing subsidies can be measured. It wouldalso provide the right incentives to biomass interms of cofiring, landfill gas, and waste.

272 Michael G. Pollitt

The principle of various levels of support fortechnologies at different stages of development isalso well established, and recent moves in UKpolicy to recognize this are sensible and impor-tant. What is needed is the right mix of R&Dsupport, competitions, and general supportmechanisms such as a FIT or TGC. It seems clearthat for small schemes, a FIT for small-scale windand small hydro does offer an attractive mecha-nism at the current low levels of development inthe United Kingdom, and hence moves in thisdirection are sensible, given high transaction costsin setting up such schemes and arranging finance.

For offshore wind, it would seem that aNFFO-style set of annual auctions would offerthe best way of keeping prices down. NFFOarrangements could be amended to ensure actualdelivery, with penalties for nondelivery. Indeed,given the scale of offshore wind’s potential and theproblem of finding a suitable level of support ini-tially, relative to other sources of renewables, thiswould seem to be a good way forward. Bids couldtake the form of contracts for differences (as sug-gested by Ofgem 2007c for the reform of the ROscheme), whereby bids would be for a fixed pricefor the electricity generated, which would be paidat that price minus a reference wholesale price,with the payments being levied across licensedsuppliers in proportion to their supply.This wouldincentivize efficient location decisions, as connec-tion and use of system charges would still beborne by the generators, and they would beincentivized to maximize the actual wholesaleprice they received in the market. It would also tiein with successful experience of the use of com-petitions for infrastructure delivery under the pri-vate finance initiative (Pollitt 2002). As with anyprocurement process that is repeated with (poten-tially) a smallish number of bidders over time, theauctions would have to be monitored for collu-sion among the bidders, but given the standardnature of the investments and transparency of thebidding strategies employed by the players, actualor tacit collusion would be easy to spot. Annualbid rounds would offer the chance to adjust quan-tities required and other details of the auction eas-ily over time to reflect learning.

For large-scale onshore wind, the RO mecha-nism could be made to work by removing therevenue recycling and adjusting the targetsaccording to the expected amount of capacityfrom offshore wind.This would essentially rewardonshore renewable generation with a fixed rev-enue supplement equal to the buyout price,assuming the target was not met or exceeded.However, it remains the case that all renewablecapacity should be expected to face the fullamount of transmission and distribution costsimposed on the system. This would encourageoptimal siting, local generation more generally,and proper competition between renewable sup-ply and demand reduction measures. Barthelmieet al. (2008) show that there would be benefits tolearning from Spain in terms of improving theshort-term forecasting of wind power availability.Improved forecasting might have increased theprice of wind power received by generators by theorder of 14% in 2003.20

In sum, the current revenue recycling withinthe RO mechanism is unnecessary and should bestopped. This is line with an early National AuditOffice report on the RO mechanism, whichwarned the government that it needed “to keep afirm grip of the Obligation’s cost relative to otherinstruments for reducing carbon dioxide” (NAO2005, 4). The system needs to be altered withrespect to offshore renewables in order to ensureleast-cost delivery of an initially very expensiverenewable energy source. Large onetime projectslike the Severn or Thames Barrage (associatedwith a new London airport), if deemed necessaryafter appropriate cost–benefit analysis, must beauctioned rather than financed within the ROmechanism.21 The RO scheme could further beamended to remove its all-or-nothing property byensuring that in the unlikely event that targetswere met or exceeded, the total amount of sub-sidy would be divided proportionately among allthose presenting ROCs. This would remove thecliff-edge effect on the renewable subsidy ofmeeting the target.22

What the history of UK renewables since1990 really tells us is that there are importantinstitutional barriers to expansion of renewablesonshore. These have to do with the lack of local

UK Renewable Energy Policy since Privatization 273

benefit from renewable projects that employ asmall number of people and have a significant per-ceived amenity impact. The key learning fromDenmark and Germany is that local populationsmust perceive such projects as being of positivebenefit to them rather than simply satisfying somedistant national policy objective, which they mayotherwise support. The United Kingdom mustdevelop local energy companies that are owned bylocal investors or local customers or councils if thepotential exploitation of local energy resources—wind, biomass, hydro, and other technologies—isto be realized. This is because virtually all renew-able electricity and heat technologies involve sig-nificant local impacts in terms of siting of indus-trial facilities close to residential areas.

For offshore renewables, getting costs downwill be the challenge. Costs need to decrease sig-nificantly in order for energy customers to bewilling to support large quantities of offshorerenewables. The current combination of capitalgrants and arbitrary ROC banding is not a satis-factory or sustainable way forward. Auctions fornew capacity would be institutionally compatiblewith the United Kingdom’s liberalized electricitymarket and offer the prospects of falling pricesover time. They would also tie in with the auc-tions to build, own, and operate offshore trans-mission lines to the new wind farms that Ofgem iscurrently implementing (see Ofgem 2007b).Under Ofgem’s offshore transmission auctionscheme, once an offshore wind farm has a firmcontract for connection to the onshore transmis-sion grid, an auction is triggered to build theinterconnection between the shore and the windfarm.

In the end, success in UK policy towardrenewable deployment, relative to other coun-tries, must be measured in terms of the net presentvalue of the amount of renewable electricity gen-erated scaled by the amount of subsidy. Althoughthis success metric will be difficult to measure atany point along the pathway, in the interim, suc-cess should be measured in terms of the extent towhich the maximum amount of renewable gen-eration (adjusted for technological maturity) isbeing supported for the current level of subsidy.UK policy clearly is not being successful, given

the large amount of relatively cheap unexploitedwind resources in the United Kingdom, in theface of overpayment to existing renewable gen-erators.

ConclusionsThe United Kingdom is struggling to develop acoherent set of policies for decarbonization fol-lowing its successful experience in liberalizingenergy markets. Various authors have suggestedthat the decarbonization policy is so ambitiousthat it demands radical institutional changes(Mitchell 2007; Pollitt 2008). However, little con-sensus has been reached on what form those insti-tutional changes should take.

What is clear is that solutions must targetleast-cost, or else the whole policy is likely to failas a result of the actual cost becoming prohibitive.On the path to this sort of ultimate policy failure,large amounts of resources are likely to be wasted,to little overall effect and for no benefit to the UKeconomy or the global climate.The United King-dom has had a long history of failed governmentintervention in the energy market and in indus-trial policy in general (Pollitt 2008). It must notcontinue this sort of tradition. It has, however,had good experience with the role of markets,undertaking basic R&D, and the use of marketmechanisms to deliver public goods. The countryhas also particular concerns about fuel poverty,which argues for a focus on keeping the costs ofrenewables policy down.

The United Kingdom agreed to an ambitiousrenewable generation target that was unnecessar-ily tough—in terms of the required speed ofincrease in the share—in the face of its EU CO2

targets, which could have been met in a muchmore straightforward way by a combination ofdemand reduction and a switch from coal- to gas-fired generation (see Grubb et al. 2008). Why thecountry got itself into this position is not appar-ent, but it clearly hoped that the EU ETS wouldbe much more effective than it has been in sup-porting decarbonization. Because of this, the EURenewables Directive has become more signifi-cant for the country than it needed to be.

274 Michael G. Pollitt

The United Kingdom also must resist calls tosee national renewables policy as anything otherthan a policy for delivering learning benefits onthe path to cost parity with established technolo-gies. An industrial policy based around theemployment or export potential of renewables isnot a sensible use of national economic resources.No doubt some benefit will accrue to the UnitedKingdom from exploiting its domestic renewablespotential, but this will arise naturally and shouldnot be an objective of policy. The British WindEnergy Association (BWEA) reports that theUnited Kingdom is a net exporter of small-scalewind turbines, the part of the market least affectedby government subsidy (see BWEA 2009). Thecountry needs to move to a more competitiveenergy market wherein smaller firms competewith large incumbents to supply power anddeliver national targets and the capacity to rapidlyadopt new lower-cost innovations exists. This isessential if incumbent costs are to be kept downand oligopoly pricing and excessive subsidyregimes are to be avoided.The 40 years from 2010to 2050 are very likely to see huge technologicaland lifestyle changes that will substantially changethe potential picture of the power, heat, and trans-port sectors (see Ault et al. 2008). The UnitedKingdom must have institutional arrangements toincentivize potentially drastic innovation withinthe renewables sector.

The country must incorporate the learningfrom both its NFFO and its RO experiences intofuture subsidy regimes.The evidence suggests thata reformed NFFO-type auction could be a sensi-ble way to deliver large offshore wind parksmostly built by big multinational utility compa-nies. Onshore, it is clear that there are legitimateland use issues with renewables, which can beaddressed only by smaller-scale projects for localpublic benefit. This policy is in line with some ofthe more decentralized scenarios of the futuredevelopment of electricity networks, and it wouldhave the added co-benefits of substantially re-inforcing the need for paradigm change at theindividual level and aiding behavioral changes thatwould support the optimal use of technologiesthat promote energy efficiency.

The United Kingdom also needs to signifi-cantly improve the quality of the information onwhich policy decisions are being made. There is asevere lack of analysis of the drivers of past policyoutcomes, partly as a result of the lack of informa-tion on the financial characteristics of individualprojects that have received subsidies. No study hasbeen done on the actual performance of renew-able projects in the United Kingdom. Foxon andPearson (2007) highlight the need for improve-ments to the process of energy policymaking,whereby analysis is properly used to evaluatepolicy, and policy is revised in the light of analysis.One particular area for improvement is in theconsistency of energy policy among heat, power,and transport fuel in terms of value of subsidies forcarbon reduction, entry barrier reduction, andpromoting learning.

The information available to potential, oftensmall-scale, developers could also be improvedwith significantly more use of geographical infor-mation system (GIS) mapping of potential renew-able energy sites and guidance on acceptabledesigns and siting rules. This would focus devel-oper efforts on sites much more likely to securelocal public support and obtain planning permis-sion. This sort of proactive approach to preparingthe ground for projects would seem to addresssome of the calls for more united governmentapproaches (e.g., Keirstead 2007) toward energypolicy in the United Kingdom. It also likelywould aid in resolving resource conflicts amonglocal community, leisure, defense, air traffic, andenergy interests.

Finally, a focus on renewables must not detractfrom the overriding policy aim of decarbonizationof the economy. This requires sensible carbonprices and the workings of the price mechanismwith regards to transmission and distributioncosts. In the end, it is only when locational costsand environmental externalities are properlypriced that any given renewables project, with itsparticular characteristics, can be evaluated amongthe myriad alternatives. Although the UK policiestoward renewables may currently be failing todeliver new capacity in sufficient quantity to hitlong-term renewables targets, it is by no means

UK Renewable Energy Policy since Privatization 275

clear that those countries that are doing better inthis regard are any nearer to achieving long-termdecarbonization.

AcknowledgmentsThe author acknowledges the ongoing intellec-tual support of the ESRC Electricity PolicyResearch Group. Bin Feng provided excellentresearch assistance. The comments of BoazMoselle, David Newbery, Jorge Padilla, DickSchmalensee, Steve Smith, and Jon Stern areacknowledged.

Notes

1. The definition of renewables used in this chapter isthat in the EU Renewables Directive(2009/28/EC): “ ‘energy from renewable sources’means energy from renewable non-fossil sources,namely wind, solar, aerothermal, geothermal,hydrothermal and ocean energy, hydropower,biomass, landfill gas, sewage treatment plant gasand biogases” (European Commission 2009, Arti-cle 2(a)).

2. This indicates that in August 2009, 8% of a typicalelectricity bill and 3% of a typical gas bill wasbeing charged to support environmental schemes,of which the most expensive were targeted towardlower-income consumers.

3. UK carbon targets are net of trading, and hencecan include carbon credits purchased from abroad.

4. HM Treasury (Her Majesty’s Treasury) is the UKMinistry of Finance.

5. It is worth noting that Germany also likes the1990 baseline date, as this coincides with the col-lapse of the Berlin Wall and the rapid decarboniza-tion of the former East Germany as a result ofindustrial decline and improved environmentalstandards.

6. See, e.g., DECC (2009a, 92), which shows pro-jected cost decreases for PV of 70% to 2050,against only 22% for coal-fired CCS.

7. NAO (2008, 17) reports 20 government policies,strategies, and reviews on energy between 1997and 2009, with 16 of those from 2003 onward.

8. Initially the levy was 10.6% in England and Wales,but it fell to 0.9% in 1998 when payments for

nuclear power ended. It was phased out in April2002, having been 0.3%. The levy rate in Scot-land, which was not used to fund nuclear liabili-ties, began at 0.5% in 1996 and reached a maxi-mum of 1.2% (Wikipedia, s.v. “Fossil Fuel Levy”).

9. England and Wales had five rounds of NFFO:NFFO-1, start date 1990, followed by NFFO-2,-3, -4, and -5 in 1992, 1995, 1997, and 1998.Scotland had three rounds: SRO-1, -2 and -3 in1995, 1997, and 1999. Northern Ireland had tworounds: NI-NFFO-1 and -2 in 1994 and 1995.See Wong (2005, 131). The last NFFO contract isdue to expire in 2018.

10. Under the 2001 EU Renewables Directive, theUnited Kingdom signed up to a 10% target forrenewable electricity generation, which is embod-ied in the successor scheme to NFFO (EuropeanCommission 2001).

11. Continuing NFFO contracts are funded via therevenue from the auction of ROCs (by the NFPA)associated with the contracts (see Ofgem 2004).

12. Assuming here that no one has invested in arenewable generation project that would beunprofitable without the “recycled” revenues. Theactual reported figure for recycled revenue is£307m (Ofgem, Renewables Obligation AnnualReport 2007–2008, 1).

13. UK inflation between September 2002 and Sep-tember 2008 was 15% (ONS 2009). The NationalAudit Office reported a figure of only £700 mil-lion ($1.05 billion) per annum for annual costs2003–2006 (NAO 2005, 35).

14. In May 2009, only eight operational schemesexisted with a capacity of 50 MW or moreonshore (DECC 2009b, 145–51).

15. For more details on the planning process in Eng-land, see DECC (2009 n.d.b).

16. In this vein, Upreti and van der Horst (2004) havean enlightening discussion of one NFFO biomassproject that, because it could not be modified assuggested by the local consultation process, even-tually had to be abandoned.

17. One of the few examples of significant capital rais-ing from the local community was the Baywindproject in Cumbria, which first raised £1.2 mil-lion ($1.8 million) to form a cooperative todevelop wind power (see www.baywind.co.uk).

18. Compare actual costs in Snyder and Kaiser (2009)and Blanco (2009) with cost simulation assump-tions in Dale et al. (2004) and Strbac et al. (2007).

19. This is because the recycled revenue component ishighly uncertain and unlikely to be a key part ofthe business case for a new renewables project.

276 Michael G. Pollitt

20. It might be considered odd that UK wind genera-tors have not done this already, given the financialincentive to do so, but this may be due to thecurrently low level of wind capacity, relative tosome of the fixed costs in setting up such a system.

21. See SDC (2007) on potential tidal projects in theUnited Kingdom.

22. The government has considered this issue but hasdecided not to do anything about it at themoment, given the gap between delivery andactual (or future) targets (DECC 2008).

ReferencesAult, G., D. Frame, N. Hughes, and N. Strachan. 2008.

Electricity Network Scenarios in Great Britain for 2050,Final Report for Ofgem’s LENS Project. Ref. No.157a/08. London: Ofgem.

Barthelmie, R. J., F. Murray, and S. C. Pryor. 2008.TheEconomic Benefit of Short-Term Forecasting forWind Energy in the UK Electricity Market. EnergyPolicy 36 (5): 1687–1696.

Bell, D., T. Gray, and C. Haggett. 2005. The “SocialGap” in Wind Farm Siting Decisions: Explanationsand Policy Responses. Environmental Politics 14 (4):460–477.

Bergmann, A., S. Colombo, and N. Hanley. 2008.Rural versus Urban Preferences for RenewableEnergy Developments. Ecological Economics 65 (3):616–625.

BIS (Department for Business, Innovation and Skills).2008. Fuel Poverty Statistics: Background Indicators,2008: Annex to Fuel Poverty Strategy Report, 2008.www.berr.gov.uk/files/file48037.pdf (accessed Feb-ruary 22, 2010).

Blanco, M. I. 2009. The Economics of Wind Energy.Renewable and Sustainable Energy Reviews 13 (6–7):1372–1382.

Bomb, C., K. McCormick, E. Deurwaarder, andT. Kåberger. 2007. Biofuels for Transport in Europe:Lessons from Germany and the UK. Energy Policy 35(4): 2256–2267.

Bonilla, D. 2006. Energy Prices, Production and theAdoption of Cogeneration in the UK and the Neth-erlands. Cambridge Working Papers in Economics:27–27. Cambridge, UK: University of Cambridge,Faculty of Economics (formerly DAE)

Brown, K. A., and D. H. Maunder. 1994. UsingLandfill Gas: A UK perspective. Renewable Energy 5(5–8): 774–781.

Butler, L., and K. Neuhoff (2008). Comparison ofFeed-In Tariff, Quota and Auction Mechanisms toSupport Wind Power Development. RenewableEnergy 33 (8): 1854–1867.

BWEA (British Wind Energy Association). 2009. SmallWind Systems UK Market Report 2009.www.bwea.com/pdf/small/BWEA%20SWS%20UK%20Market%20Report%202009.pdf (accessed August 27, 2009).

———. No date. www.bwea.com/ukwed/index.asp(accessed August 20, 2009).

Cannell, M. G. R. 2003. Carbon Sequestration andBiomass Energy Offset: Theoretical, Potential andAchievable Capacities Globally, in Europe and theUK. Biomass and Bioenergy 24 (2): 97–116.

Carbon Trust. 2006. Policy Frameworks for Renewables.London: Carbon Trust.

CCC (Committee on Climate Change). 2008. Buildinga Low Carbon Economy:The UK’s Contribution toTack-ling Climate Change. London: TSO.

Chamberlain, Phil. 2008. Still Much to Be Done onPlanning. Real Power 12 (April–June): 19–24.www.bwea.com/pdf/realpower/realpower_12.pdf(accessed February 22, 2010).

Coles, R. W., and J. Taylor. 1993. Wind Power andPlanning: The Environmental Impact of Windfarmsin the UK. Land Use Policy 10 (3): 205–226.

Connor, P. M. 2003. UK Renewable Energy Policy: AReview. Renewable and Sustainable Energy Reviews 7(1): 65–82.

Conway, A. 1986. Tidal Power: A Matter of Faith,Hope and Policy. Energy Policy 14 (6): 574–577.

Cowell, R. 2007. Wind Power and“the Planning Prob-lem”: The Experience of Wales. European Environ-ment 17 (5): 291–306.

CPRE (Campaign to Protect Rural England). No date.CPRE home page. www.cpre.org.uk (accessed Feb-ruary 22, 2010).

Dale, L., D. Milborrow, R. Slark, and G. Strbac. 2004.Total Cost Estimates for Large-Scale Wind Scenariosin UK. Energy Policy 32 (17): 1949–1956.

DECC (Department of Energy and Climate Change).2008. Reform of the Renewables Obligation: GovernmentResponse to the Statutory Consultation on the RenewablesObligation Order 2009. London: DECC.

———. 2009a. Analytical Annex:The UK Low CarbonTransition Plan. London: DECC.

———. 2009b. Digest of UK Energy Statistics 2009 Edi-tion. London: DECC.

———. 2009c. Draft National Policy Statement forNuclear Power Generation (EN-6).

UK Renewable Energy Policy since Privatization 277

http://data.energynpsconsultation.decc.gov.uk/documents/npss/EN-6.pdf (accessed February 22,2010).

———. 2009d. Offshore Renewables Financial BoostKicks In. www.decc.gov.uk/en/content/cms/news/pn037/pn037.aspx (accessed August 25, 2009).

———. 2009e. The UK Low Carbon Transition Plan.London: DECC.

———. 2009f. The UK Renewable Energy Strategy. Lon-don: DECC.

———. No date, a. Energy Statistics: Prices.www.decc.gov.uk/en/content/cms/statistics/source/prices/prices.aspx (accessed August 25,2009).

———. No date, b. Planning Process in England.www.decc.gov.uk/en/content/cms/what_we_do/uk_supply/energy_mix/renewable/planning/plan_policy/england/england.aspx.

Defra (Department for Environment, Food and RuralAffairs). 2008. Greenhouse Emissions Statistics.www.defra.gov.uk/environment/statistics/globatmos/download/ghg_ns_20090326.pdf(accessed August 27, 2009).

———. 2009. UK Climate Change Sustainable Devel-opment Indicator: 2007 Greenhouse Gas Emissions,Final Figures, www.defra.gov.uk/evidence/statistics/environment/globatmos/download/ghg_ns_20090203.pdf (accessed August 27, 2009).

Drake, F. 2009. Black Gold to Green Gold: RegionalEnergy Policy and the Rehabilitation of Coal inResponse to Climate Change. Area 41 (1): 43–54.

DUKES (Digest of UK Energy Statistics). 2009.Expenditure on Energy by Final User to 2008(DUKES 1.1.6). Available at DECC n.d.a.

EEA (European Environment Agency). 2006. EUGreenhouse Gas Emissions Increase for SecondYearin a Row. www.eea.europa.eu/pressroom/newsreleases/GHG2006-en (accessed February 22,2010).

Eikeland, P. O., and I. A. Sæverud. 2007. Market Diffu-sion of New Renewable Energy in Europe: Explain-ing Front-Runner and Laggard Positions. Energy &Environment 18 (1): 13–36.

Elghali, L., R. Clift, P. Sinclair, C. Panoutsou, and A.Bauen. 2007. Developing a Sustainability Frame-work for the Assessment of Bioenergy Systems.Energy Policy 35 (12): 6075–6083.

Elliott, D. 1992. Renewables and the Privatization ofthe UK ESI: A Case Study. Energy Policy 20 (3):257–268.

———. 1994. UK Renewable Energy Strategy: TheNeed for Longer-Term Support. Energy Policy 22(12): 1067–1074.

———. 2009. Marine Renewables: A New InnovationFrontier. Technology Analysis and Strategic Management21 (2): 267–275.

Eltham, D. C., G. P. Harrison, and S. J. Allen. 2008.Change in Public Attitudes Towards a Cornish WindFarm: Implications for Planning. Energy Policy 36 (1):23–33.

European Commission. 2001. Directive 2001/77/EC ofthe European Parliament and Council of 27 September2001 on the Promotion of Electricity Produced fromRenewable Energy Sources in the Internal Electricity.Brussels: European Commission.

———. 2009. Directive 2009/28/EC of the EuropeanParliament and of the Council of 23 April 2009 onthe promotion of the use of energy from renewablesources and amending and subsequently repealingDirectives 2001/77/EC and 2003/30/EC. Brussels:European Commission.

Evans, J., and R. Green. 2003. Why Did British Electri-city Prices Fall after 1998? CMI Electricity Projectworking paper No. 26. Cambridge, UK: Universityof Cambridge.

EWEA (European Wind Energy Association). 2009.Wind Energy:The Facts. Chapter 7: Employment inthe Wind Energy Sector. www.wind-energy-the-facts.org/en/part-3-economics-of-wind-power/chapter-7-employment/ (accessed August 27, 2009).

Faiers, A., and C. Neame. 2006. Consumer AttitudesTowards Domestic Solar Power Systems. EnergyPolicy 34 (14): 1797–1806.

Foxon, T. J., R. Gross, A. Chase, J. Howes, A. Arnall,and D. Anderson. 2005. UK Innovation Systems forNew and Renewable Energy Technologies: Drivers,Barriers and Systems Failures. Energy Policy 33 (16):2123–2137.

Foxon, T. J., and P. J. G. Pearson. 2007. TowardsImproved Policy Processes for Promoting Innovationin Renewable Electricity Technologies in the UK.Energy Policy 35 (3): 1539–1550.

Grubb, M., T. Jamasb, and M. Pollitt. 2008. A Low-Carbon Electricity Sector for the UK: What Can BeDone and How Much Will It Cost? In Delivering aLow Carbon Electricity System, edited by M. Grubb,T.Jamasb, and M. Pollitt. Cambridge, UK: CambridgeUniversity Press, 462–497.

Grubb, M., and J. Wilde. 2008. Enhancing the EfficientUse of Electricity in the Business and Public Sectors.Delivering a Low Carbon Electricity System, edited by

278 Michael G. Pollitt

M. Grubb, T. Jamasb, and M. Pollitt. Cambridge,UK: Cambridge University Press, 229–256.

Hammond, G. P., S. Kallu, and M.C. McManus. 2008.Development of Biofuels for the UK AutomotiveMarket. Applied Energy 85 (6): 506–515.

Harvey, K. 1994.The Development of Combined Heatand Power in the UK. Energy Policy 22 (2): 179–181.

Hedger, M. M. 1995. Wind Power: Challenges to Plan-ning Policy in the UK. Land Use Policy 12 (1):17–28.

Helm, D. 2002. A Critique of Renewables Policy in theUK. Energy Policy 30 (3): 185–188.

Hill, A., and M. Hay. 2004. UK Renewables: Harness-ing Wind, Wave and Tide. Refocus 5 (2): 20–21.

HM Treasury. 2009. Budget 2009 Building Britain’sFuture. Economic and Fiscal Strategy Report and Finan-cial Statement and Budget Report. London: StationeryOffice.

House of Lords. 2008. Economic Affairs—FourthReport. Chapter 6: Policy on Renewable Energy.www.publications.parliament.uk/pa/ld200708/ldselect/ldeconaf/195/19509.htm#a53 (accessedFebruary 22, 2010).

Hull, A. 1995. Local Strategies for Renewable Energy:Policy Approaches in England and Wales. Land UsePolicy 12 (1): 7–16.

IEA (International Energy Agency). 2006. Energy Poli-cies of IEA Countries: The United Kingdom: 2006Review. Paris: OECD Publishing.

———. 2009. Energy Technology RD&D Database.wds.iea.org/WDS/ReportFolders/ReportFolders.aspx (accessed August 28, 2009).

Jacobsson, S., A. Bergek, D. Finon, V. Lauber, C.Mitchell, D. Toke, and A. Verbruggen. 2009. EURenewable Energy Support Policy: Faith or Facts?Energy Policy 37 (6): 2143–2146.

Jamasb, T., H. Kiamil, and R. Nepal. 2008a. Hot Issueand Burning Options in Waste Management: ASocial Cost Benefit Analysis of Waste-to-Energy inthe UK. Cambridge Working Papers in Economics:24–24. Cambridge, UK: University of Cambridge,Faculty of Economics.

Jamasb, T., W. Nuttall, M. Pollitt, and A. Maratou.2008b. Technologies for a Low Carbon ElectricitySystem: An Assessment of the UK’s Issues andOptions. Delivering a Low Carbon Electricity System,edited by M. Grubb,T. Jamasb, and M. Pollitt. Cam-bridge, UK: Cambridge University Press, 64–99.

Jarvis, I. 1986. Can a successful city-wide CHP schemebe launched in the UK?’, Energy Policy, 14 (2), 160–163.

Junginger, M., T. Bolkesjo, D. Bradley, P. Dolzan, A.Faaij, J. Heinimö, B. Hektor, Ø. Leistad, E. Ling, M.Perry, E. Piacente, F. Rosillo-Calle, Y. Ryckmans,P-P. Schouwenberg, B. Solberg, E. Trømborg, A. daSilva Walter, and M. de Wit. 2008. Developments inInternational Bioenergy Trade. Biomass and Bioenergy32 (8): 717–729.

Kannan, R. 2009. Uncertainties in Key Low CarbonPower GenerationTechnologies: Implication for UKDecarbonisation Targets. Applied Energy 86 (10):1873–1886.

Keirstead, J. 2007. The UK Domestic PhotovoltaicsIndustry and the Role of Central Government.Energy Policy 35 (4): 2268–2280.

Kelly, G. 2007. Renewable Energy Strategies in Eng-land, Australia and New Zealand. Geoforum 38 (2):326–338.

Klessmann, C., C. Nabe, and K. Burges. 2008. Pros andCons of Exposing Renewables to Electricity MarketRisks: A Comparison of the Market IntegrationApproaches in Germany, Spain, and the UK. EnergyPolicy 36 (10): 3646–3661.

Lipp, J. 2007. Lessons for Effective Renewable Electri-city Policy from Denmark, Germany and the UnitedKingdom. Energy Policy 35 (11): 5481–5495.

Lupton, M. 2008. Chasing the Dream [Micro-generation Strategy]. Engineering & Technology(17509637) 3 (12): 54–57.

MacKay, D. 2008. Sustainable Energy—without the HotAir. Cambridge: UIT.

Mander, S. 2008. The Role of Discourse Coalitions inPlanning for Renewable Energy: A Case Study ofWind-Energy Deployment. Environment and Plan-ning C: Government and Policy 26 (3): 583–600.

Marshall, E. 1993. CHP and Deregulation: The Regu-lator’s Viewpoint. Energy Policy 21 (1): 73–78.

McGowan, F. 1991. Controlling the GreenhouseEffect:The Role of Renewables. Energy Policy 19 (2):110–118.

Meyer, N. I. 2003. European Schemes for PromotingRenewables in Liberalised Markets. Energy Policy 31(7): 665–676.

Mitchell, C. 1995. Renewable Energy in the UK: Finan-cing Options for the Future. London: Campaign for theProtection of Rural England.

———. 2007. The Political Economy of SustainableEnergy. Basingstoke, UK: Palgrave.

Mitchell, C., D. Bauknecht, and P. M. Connor. 2006.Effectiveness through Risk Reduction: A Compari-son of the Renewable Obligation in England andWales and the Feed-In System in Germany. EnergyPolicy 34 (3): 297–305.

UK Renewable Energy Policy since Privatization 279

Mitchell, C., and P. Connor. 2004. Renewable EnergyPolicy in the UK 1990–2003. Energy Policy 32 (17):1935–1947.

Mueller, M., and R. Wallace. 2008. Enabling Scienceand Technology for Marine Renewable Energy.Energy Policy 36 (12): 4376–4382.

NAO (National Audit Office). 2005. Department ofTradeand Industry Renewable Energy. 11 February 2005.London: NAO.

———. 2008. Renewable Energy: Options for Scrutiny.London: NAO.

Nelson, H. T. 2008. Planning Implications from theInteractions between Renewable Energy Programsand Carbon Regulation. Journal of EnvironmentalPlanning and Management 51 (4): 581–596.

Newbery, D. M., and M. G. Pollitt. 1997.The Restruc-turing and Privatisation of Britain’s CEGB: Was ItWorth It? Journal of Industrial Economics 45 (3): 269–303.

Odenberger, M., J. Kjärstad, and F. Johnsson. 2008.Ramp-Up of CO2 Capture and Storage withinEurope. International Journal of Greenhouse Gas Control2 (4): 417–438.

Ofgem. 2002. The Review of the First Year of NETA: AReview Document, Vol. 1, July. London: Ofgem.

———. 2004. The Renewables Obligation: Ofgem’s FirstAnnual Report. London: Ofgem.

———. 2007a.The GB Queue: Problems and PossibleSolutions. www.ofgem.gov.uk/Networks/Trans/PriceControls/TPCR4/ConsultantsReports/Documents1/16962-The%20GB%20Queue%20-%20%20Reasons%20and%20remedies.ppt (accessedAugust 28, 2009).

———. 2007b. Offshore Transmission.www.ofgem.gov.uk/Networks/offtrans/Pages/Offshoretransmission.aspx (accessed August 28,2009).

———. 2007c). Reform of the Renewables Obligation2006: Ofgem’s Response. London: Ofgem.

———. 2009a. Derogations to Facilitate Earlier Con-nection of Generation—Decision on InterimApproach. www.ofgem.gov.uk/Networks/Trans/ElecTransPolicy/tar/Documents1/20090508%20derogations%20interim.pdf (accessed February22, 2010).

———. 2009b. Updated Household Energy BillsExplained. www.ofgem.gov.uk/Media/FactSheets/Documents1/updatedhouseholdbills09.pdf (accessedOctober 20, 2009).

———. Various dates. Renewables Obligation annualreports.

ONS (Office for National Statistics). 2007. House-holds: Rise in Non-family Households.www.statistics.gov.uk/cci/nugget.asp?id=1866(accessed August 25, 2009).

———. 2009. CPI series. www.ons.gov.uk (accessedAugust 27, 2009).

OPSI (Office of Public Sector Information). 2008a.Climate Change Act. www.opsi.gov.uk/acts/acts2008/ukpga_20080027_en_1 (accessed August28, 2009).

———. 2008b. Energy Act. www.opsi.gov.uk/acts/acts2008/ukpga_20080032_ en_1 (accessed August28, 2009).

———. 2008c. Planning Act. www.opsi.gov.uk/acts/acts2008/ukpga_20080029_en_1 (accessed August28, 2009).

———. 2009. 2009 No. 785 Electricity, England andWales The Renewables Obligation Order 2009.www.opsi.gov.uk/si/si2009/pdf/uksi_20090785_en.pdf (accessed March 26, 2010).

ORER (Office of the Renewable Energy Regulator).No date. ORER home page. www.orer.gov.au(accessed February 22, 2010).

Oxera. 2005. Economic Analysis of the Design, Cost andPerformance of the UK Renewables Obligation and Capi-tal Grants Scheme. Report prepared for the NationalAudit Office. London: National Audit Office.

Paish, O. 2002. Small Hydro Power: Technology andCurrent Status. Renewable and Sustainable EnergyReviews 6 (6): 537–556.

Perry, M., and F. Rosillo-Calle. 2008. Recent Trendsand Future Opportunities in UK Bioenergy: Max-imising Biomass Penetration in a Centralised EnergySystem. Biomass and Bioenergy 32 (8): 688–701.

Piterou, A., S. Shackley, and P. Upham. 2008. ProjectARBRE: Lessons for Bio-energy Developers andPolicy-makers. Energy Policy 36 (6): 2044–50.

Pollitt, M. 2002. Declining Role of the State in Infra-structure Investments in the UK. In Private Initiativesin Infrastructure: Priorities, Incentives, and Performance,edited by M. Tsuji, S. Berg, and M. Politt. Chelten-ham, UK: Edward Elgar, 67–100.

———. 2008. The Future of Electricity (and Gas)Regulation in Low-Carbon Policy World. EnergyJournal, Special Issue in Honor of David Newbery,63–94.

Pollitt, M., and J. Bialek. 2008. Electricity NetworkInvestment and Regulation for a Low CarbonFuture. In Delivering a Low Carbon Electricity System,edited by M. Grubb,T. Jamasb, and M. Pollitt. Cam-bridge, UK: Cambridge University Press.

280 Michael G. Pollitt

Porter, D., and N. Steen. 1996. Renewable Energy in aCompetitive Electricity Market. Renewable Energy 9(1–4): 1120–1123.

QEP (Quarterly Energy Prices). 2009. Average AnnualDomestic Electricity Bills by Home and Non-homeSupplier (QEP 2.2.1). Available at DECC n.d.a.

Riddell, R. 2008.TurningTides [Tidal PowerTechnol-ogy]. Engineering & Technology (17509637) 3 (16):46–49.

Roberts, S., and F. Weightman. 1994. Cleaning Up theWorld with Renewable Energy: From Possibilities toPracticalities. Renewable Energy 5 (5–8): 1314–1321.

Rodrik, D. 2008. Second Best Institutions. AmericanEconomic Review 98 (2): 100–104.

SDC (Sustainable Development Commission). 2007.Turning the Tide: Tidal Power in the UK. London:SDC.

SKM (Sinclair Knight Merz). 2008. Growth Scenarios forUK Renewables Generation and Implications for FutureDevelopments and Operation of Electricity Networks.London: BERR.

Slade, R., C. Panoutsou, and A. Bauen. 2009. Recon-ciling Bio-energy Policy and Delivery in the UK:Will UK Policy Initiatives Lead to IncreasedDeployment? Biomass and Bioenergy 33 (4): 679–688.

Snyder, B., and M. J. Kaiser. 2009. A Comparison ofOffshore Wind Power Development in Europe andthe U.S.: Patterns and Drivers of Development.Applied Energy 86 (10): 1845–1856.

Soderholm, P., and G. Klaassen. 2007. Wind Power inEurope: A Simultaneous Innovation-DiffusionModel. Environmental and Resource Economics 36 (2):163–190.

Stainforth, D., A. Cole, P. Dolley, H. Edwards, J.Wilczek, and M. Wood. 1996. An Overview of theUK Department ofTrade and Industry’s (DTI’s) Pro-gramme in Solar Energy. Solar Energy 58 (1–3): 111–119.

Stanford, A. 1998. Liberalisation of the UK EnergyMarket: An Opportunity for Green Energy. Renew-able Energy 15 (1–4): 215–217.

Stenzel, T., and A. Frenzel. 2008. Regulating Techno-logical Change: The Strategic Reactions of UtilityCompanies towards Subsidy Policies in the German,Spanish and UK Electricity Markets. Energy Policy 36(7): 2645–2657.

Strachan, P. A., and D. Lal. 2004. Wind Energy Policy,Planning and Management Practice in the UK: HotAir or a Gathering Storm? Regional Studies 38 (5):549–569.

Strbac, G., A. Shakoor, M. Black, D. Pudjianto, and T.Bopp. 2007. Impact of Wind Generation on the

Operation and Development of the UK ElectricitySystems. Electric Power Systems Research 77: 1214–1227.

Swedish Energy Agency. No date. Swedish EnergyAgency home page. www.swedishenergyagency.se(accessed February 22, 2010).

Swider, D. J., L. Beurskens, S. Davidson, J. Twidell, J.Pyrko, W. Pruggler, H. Auer, K. Vertin, and R.Skema. 2008. Conditions and Costs for RenewablesElectricity Grid Connection: Examples in Europe.Renewable Energy 8: 1832–1842.

Szarka, J. 2006. Wind Power, Policy Learning and Para-digm Change. Energy Policy 34 (17): 3041–3048.

Szarka, J., and I. Bluhdorn. 2006. Wind Power in Britainand Germany: Explaining Contrasting DevelopmentPaths. London: Anglo-German Foundation for theStudy of Industry.

Taylor, G. 2008. Bioenergy for Heat and Electricity inthe UK: A Research Atlas and Roadmap. EnergyPolicy 36 (12): 4383–4389.

Thomson, M., and D. G. Infield. 2007. Impact ofWidespread Photovoltaics Generation on Distribu-tion Systems. IET Renewable Power Generation 1 (1):33–40.

Thornley, P. 2006. Increasing Biomass Based PowerGeneration in the UK. Energy Policy 34 (15): 2087–2099.

Tickell, Oliver. 2008. Robbing Us of Renewables.Guardian, September 6.

Toke, D. 2003. Wind Power in the UK: How PlanningConditions and Financial Arrangements Affect Out-comes. International Journal of Sustainable Energy 23(4): 207–216.

———. 2005a. Are Green Electricity Certificates theWay Forward for Renewable Energy? An Evaluationof the United Kingdom’s Renewables Obligation inthe Context of International Comparisons. Environ-ment and Planning C: Government and Policy 23 (3):361–374.

———. 2005b. Explaining Wind Planning Outcomes.Some findings from a study in England and Wales.Energy Policy 33 (12): 1527–1539.

———. 2007. Renewable Financial Support Systemsand Cost-Effectiveness. Journal of Cleaner Production15 (3): 280–287.

Toke, D., and V. Lauber. 2007. Anglo-Saxon and Ger-man Approaches to Neoliberalism and Environmen-tal Policy: The Case of Financing RenewableEnergy. Geoforum 38 (4): 677–687.

Toke, D., and P. A. Strachan. 2006. Ecological Mod-ernization and Wind Power in the UK. European

UK Renewable Energy Policy since Privatization 281

Environment: The Journal of European EnvironmentalPolicy (Wiley) 16 (3): 155–166.

Upham, P., S. Shackley, and H. Waterman. 2007. Publicand Stakeholder Perceptions of 2030 Bioenergy Sce-narios for theYorkshire and Humber Region. EnergyPolicy 35 (9): 4403–4412.

Upreti, B. R., and D. van der Horst. 2004. NationalRenewable Energy Policy and Local Opposition inthe UK:The Failed Development of a Biomass Elec-tricity Plant. Biomass and Bioenergy 26 (1): 61–69.

van Dam, J., M. Junginger, A.Faaij, I. Jurgens, G. Best,and U. Fritsche. 2008. Overview of Recent Devel-opments in Sustainable Biomass Certification.Biomass and Bioenergy 32 (8): 749–780.

van der Horst, D. 2005. UK Biomass Energy since1990: The Mismatch between Project Types andPolicy Objectives. Energy Policy 33 (5): 705–716.

Warren, C. R., C. Lumsden, S. O’Dowd, and R.V.Birnie. 2005. ‘Green on Green’: Public Perceptions

of Wind Power in Scotland and Ireland. Journal ofEnvironmental Planning and Management 48 (6): 853–875.

Wind Power (The Wind Power Wind Turbines andWindfarms Database). 2009. Wind Power Produc-tion for Main Countries. www.thewindpower.net/23-countries-capacities.php (accessed February 22,2010).

Wind Prospect. 2008. Wind Prospect Group homepage. www.windprospect.com (accessed February22, 2010).

Wong, S.-F. 2005. Obliging Institutions and IndustryEvolution: A Comparative Study of the German andUK Wind Energy Industries. Industry and Innovation12 (1): 117–145.

Wordsworth, A., and M. Grubb. 2003. Quantifying theUK’s Incentives for Low Carbon Investment. ClimatePolicy (Elsevier) 3 (1): 77.

282 Michael G. Pollitt

14

Experience with Renewable EnergyPolicy in GermanyHannes Weigt and Florian Leuthold

Germany now has more than 30 years of expe-rience in supporting renewable energy

sources (RES), of which almost 20 years includesupport for market entry. Under the ElectricityFeed-In Act (Stromeinspeisegesetz, StrEG) of1991, and continuing with the feed-in system ofthe Renewable Energy Source Act (ErneuerbareEnergien Gesetz, EEG), the share of electricitygenerated by renewables increased from about 3%in 1990 to almost 15% in 2008.

This chapter summarizes the support for RESin Germany and discusses the problems and futuredevelopments regarding market design and opera-tions. It begins with an overview, including a roadmap of applied mechanisms and quantitative mar-ket results. Next, it addresses the economic evalu-ation of the RES policy, focusing on the effi-ciency of the German feed-in tariff (FIT)approach and highlighting the impact of risk oninvestments, employment aspects of RES support,adjustment mechanisms, and interaction of RESpolicy with environmental mechanisms. Thechapter then examines the implications for marketdesign and future development given the highshare of RES. The particular focus is on wind, asthis RES technology has the highest utilizationand consequently the greatest impact on the Ger-man electricity market. Topics discussed includenetwork extension, operational issues, reserve

capacities, and related long-term aspects regardinginvestment in conventional power plants.

Renewable Energy Policyin GermanyThe German energy system largely depends onimported fossil fuels.The transport sector relies onoil imports from the North Sea, Russia, and theglobal market; the heating sector on oil and natu-ral gas, the latter imported from Russia and theNorth Sea; and the electricity sector on nuclear,coal, domestic lignite, and a growing share ofnatural gas. The overall share of fossil fuels andnuclear energy in primary energy consumption in2007 amounted to 93% (Figure 14.1). The sumtotal of energy imports in primary energy con-sumption is 60% (Bechberger and Reiche 2004).

Given Germany’s import dependence, secu-rity of supply has always been an issue in develop-ing energy policy, particularly since the 1970s oilcrisis. The issue first tended to center around coaland nuclear energy. However, with the increasedneed to address global climate change, the envi-ronmental aspect has gained importance. It isexpected that RES will play a major role in thedecades ahead: DLR (2008) forecasts a gap

between electricity demand and generation of 60to 70 GW until 2030, based on the current gen-eration portfolio. Meanwhile, the EuropeanUnion has set targets of a 20% share of RES, 20%reduction of emissions, and 20% increase inenergy efficiency, and the transformation of Ger-many’s regulated electricity and natural gas mar-kets to a competitive market framework is still inprogress.

Following is a general outline of the GermanRES policies, highlighting the different appliedmechanisms and achieved objectives in quantita-tive terms. It begins with the initial steps in the1970s and 1980s, details the first major invest-ments in wind energy in the 1990s, then discussesthe breakthrough of RES support in 2000, fol-lowed by projections of expected future develop-ments.1

1970s and 1980s: R&D and First Steps

The global oil crisis during the 1970s promptedthe move to restructure Germany’s energy sector,at first with a focus on hard coal and nuclear gen-eration, but including RES development on thefringe. Research and development (R&D) fund-ing for RES started in the mid-1970s. Govern-ment spending on R&D reached its peak duringthe first half of the 1980s, with the major share forcoal and nuclear (Figure 14.2). Most of the RES

R&D was spent on off-grid systems intended forutilization in developing countries. Government-funded programs focused on prototypes, testseries, and demonstrations, which mostly concen-trated on the development of small- and medium-scale applications.2 About 40 wind projects werefinanced during 1977 and 1989, and similar sup-port was given to solar cell research (Jacobssonand Lauber 2006).

In addition to R&D support, a first step tofoster market entry was taken when, in 1979, anRES tariff was introduced that obligated utilitiesto buy energy from RES based on avoided costs.The results were limited, however (Lauber andMez 2004).

During this period, public support for nuclearenergy decreased significantly, especially after theChernobyl accident in 1986. At the same time,domestic coal became increasingly uncompetitiveand required cross-subsidies via a special tax onelectricity prices starting in 1975. In 1980, a com-mission of the German Parliament recommendedefficiency measurements and RES as importantcornerstones of a sustainable energy policy. Thefollowing year, the recommendation wasreconfirmed by a five-year study commissionedby the Federal Ministry of Research andTechnol-ogy.

Oil34%

Gas22%

Coal14%

Lignite12%

Nuclear11%

RES7%

Hydro<1%

Wind1%

Biomass5%

Other RES

Source: BMU 2008aNote: Total 13,993 petajoules (PJ)

Figure 14.1. Primary energy consumption in Germany, 2007

284 HannesWeigt and Florian Leuthold

1990s: StrEG and Increasing Wind

As public concern about the environment grew inthe late 1980s, the German government intro-duced a number of programs to increase supportfor RES. The Ministry of Research initiated twomarket programs. Wind energy was supported bythe 100-megawatt (MW) Mass Testing Program,which was later increased to 250 MW.3 The pro-gram included a guaranteed payment of 4 eurocents (U.S. 5.4 cents) per kilowatt-hour (kWh).Solar energy was supported via the 1,000 RoofsProgram between 1991 and 1995, providinginvestment support that resulted in the installationof more than 2,000 photovoltaic systems and apeak capacity of 4 MW.

The most important support system was theElectricity Feed-In Act (Stromeinspeisegesetz,StrEG) in 1991; its feed-in tariff (FIT) systemrequired utilities to purchase and pay for RES ona fixed-tariff basis coupled to the end-user tariff.4

The StrEG, combined with the 100/250 MWProgram and soft loans provided by the state-owned Deutsche Ausgleichsbank, led to a dou-bling of wind’s installed capacities, or 4,443 MWin 1999 (Figure 14.3).

The FIT did not include a burden-sharingprovision, and utilities with large shares of windenergy faced higher costs. Therefore, a so-calledhardship clause was included in 1998 to addressthese concerns.5 The uncertainties regarding the

3,500

3,000

2,500

2,000

1,500

1,000

mn

2007

$

500

0

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

2002

2004

2006

Other Efficiency RES Fossil fuels Nuclear fission and fusion

Source: IEA 2009

Figure 14.2. Energy R&D funding in Germany, 1974–2007

2,500

2,000

1,500

1,000

Annually installed turbines Installed capacity

500

0

25,000

20,000

15,000

10,000

5,000

0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

Num

ber

of tu

rbin

es

MW

Source: BWE 2009

Figure 14.3. Wind turbine installations, 1990–2008

Experience with Renewable Energy Policy in Germany 285

FIT, as well as the end-user price decrease due tothe liberalization process and the resulting revenuelosses for wind generators, caused a decline inturbine installation after 1995 (see Figure 14.3).Besides wind energy, other RES did not obtainsignificant market share during the 1990s, as thefinancial support was insufficient even with theenactment of StrEG. The 1,000 Roofs Programhelped increase solar penetration, yet there was nodirect follow-up program.6

2000 and After: EEG and Breakthrough

In 1998, a change in the government7 led to acomprehensive market program, including areformulation of the StrEG, an ecological taxreform, further support programs for RES imple-mentation, and the decision to phase out nuclearplants.

The most important measure for RES supportwas the Renewable Energy Source Act(Erneuerbare Energien Gesetz, EEG), adopted in2000. The EEG continued the StrEG structurewith respect to dispatch priority and guaranteedpayments for RES generators. Whereas the StrEGtariffs varied with the end-user prices, the EEGtariff rates were guaranteed for 20 years once aninvestment was completed. EEG tariffs for newlyinstalled capacity decline each year by a predeter-mined percentage to account for technicalprogress and cost savings; this should eventuallyalign RES generation costs with competitive lev-els and make the support mechanism obsolete. AGermany-wide compensation mechanism wasintroduced to distribute the burden of RES sup-port equally among all consumers, and the EEGincluded provisions regarding RES grid access(Bechberger and Reiche 2004).

The EEG had a more differentiated tariffstructure than the StrEG. In addition, utilities canbenefit from the EEG tariffs. For wind energy, asite-dependent tariff allows turbines at less effi-cient locations to obtain a higher tariff. Figure14.3 shows the result, an increase in onshoreinstallations. Solar rates have increased to about50 euro cents (68 cents) per kWh, about 10 timesthe wholesale price of electricity in Germany.Biomass generation, landfill and similar gas-fired

units, and small-scale hydropower can obtain tar-iffs ranging from 7.7 to 10.1 euro cents (10.5 to13.8 cents) per kWh depending on size and fur-ther premium options (e.g., fuel cells, fuel restric-tions).

The second major policy measure to promoteRES and energy efficiency was the ecological taxreform that came into force in 1999. Additionaltaxes were introduced on motor fuels, fuel oil,natural gas, and electricity, which have graduallyincreased in subsequent years.8 Part of the taxincome is used for financing RES market support.

Other programs provide investment subsidiesand low-interest loans. A follow-up to the 1,000Roofs Program, the 100,000 Roofs Program,provides soft loans with low- and fixed-interestrates and a 20-year payback period to individualhomeowners and small-scale installation projects.In combination with the EEG, the follow-up pro-gram’s target of 350 MW was reached within 3years, and the government then had to increasethe ceiling for solar cells receiving EEG tariffs to1,000 MW.

Also in 1999, a general market support systemwas introduced with the Market Incentive Pro-gram (MAP), including direct investment subsi-dies and soft loans.Two other programs, the Envi-ronment and Energy Conservation Program andthe Environment Program, also provide soft loans,which were particularly important for the windenergy projects because they were not covered bythe MAP. The generous tariff structure of theEEG in combination with the supporting pro-grams has increased RES generation since 2000(Figure 14.4).9 The major part of this increase isattributed to wind, which quintupled its installedcapacity between 2000 and 2008 to about 24 GW(see Figure 14.3). Biomass also quintupled, from600 MW to about 3,200 MW, which led to asignificant increase in generation as a result of theplants’ high capacity factor. Photovoltaic unitsrepresent the largest relative capacity increase,from less than 100 MW to about 5,300 MW in2008. However, because of the low capacity fac-tor, PV’s total share in generation terms remainsminor.

A useful feature of the EEG is an amendmentsystem that allows the adaption of the law if mar-

286 HannesWeigt and Florian Leuthold

ket conditions change.10 After the first progressreport in 2002, the first amendment was finallypassed in July 2004. It mainly included tariffadjustments (for example, offshore wind rateswere reduced), further distinguished solar systemsby installation type (roof, facade, or stand-alone),and improved the legal position of RES genera-tors compared with grid companies.

Current and Future Developments

The EEG’s second progress report in 2004 wastransformed into an amendment that becameeffective in January 2009. The adjustments takeinto account recent developments, particularlythe lack of offshore wind and geothermal invest-ments and the cost decrease in the production ofphotovoltaic systems (Büsgen and Dürreschmidt2009). The shift of the government toward aconservative–liberal coalition with the electionsin September 2009 may lead to further adjust-ments in the EEG. Also, the abolishment of thenuclear phaseout is on the political agenda again.

Although elections almost always affectenergy policy, it is expected that Germany none-theless will continue to provide attractive invest-ment conditions for RES. The rationale underly-ing the EEG has been extended to other sectorsbeyond electricity via the new Renewable EnergyHeat Law, which should increase the share ofRES in the heating sector to 14% until 2020.

The future development of RES in the Ger-man energy market is subject to uncertainty andhas been analyzed in several studies, which differin their underlying assumptions. Wischermannand Wagner (2009) compare several future projec-tions with respect to RES development. The sce-narios can roughly be clustered in business-as-usual cases, green cases with a larger focus onenvironmental policy, and nuclear cases in whichthe phaseout of nuclear plants is withdrawn. Con-sequently, the expansion of RES differs.The RESshare of total power generation in 2030 rangesfrom 18% to 53% (Figure 14.5). Wind, particu-larly offshore, will play the largest role in futureRES developments, followed by biomass.Geothermal production will most likely play aminor role in the near future, and solar shows onlya moderate expansion in all scenarios, mainlydriven by the EEG support.

Summary of Germany’s RenewableEnergy Policy

Germany has an impressive track record inincreasing the share of RES in energy use. Given asystem with a small share of hydroelectricity gen-eration and hardly any other renewable energyutilization in the 1980s, the total share of RESincreased to 6.9% of the primary energy con-sumption and about 15% of electricity generationin 2007. Germany is the market with the second-

100 15%

12%

9%

6%

3%

0%

Per

cent

age

TW

h80

60

40

20

0

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

Hydro Wind Biomass PV Geo Share of electricity consumption

Source: BMU 2008a

Figure 14.4. Electricity generation from RES, 1990–2007

Experience with Renewable Energy Policy in Germany 287

highest amount of installed wind capacity in 2008(after the United States) and the leading marketfor installed photovoltaic (PV) capacity.

Taking the last decades as a benchmark, it issafe to assume that Germany will indeed obtain afurther increase of RES generation and mostlikely reach the European Union’s environmentaltargets until 2020. The continuous adjustment ofthe EEG allows the country to adapt to changingconditions and provide the investment securityguaranteed by the feed-in tariffs, if the tariff levelis set high enough.

Economic AnalysisAs is evident by Germany’s support for RES, theoutcome is impressive in absolute numbers. Themultitude of support mechanisms applied in thelast decades allows an economic assessment of thepolicy’s effectiveness. This section begins byexamining the economic breakdown of the FITsystem. Next, it shows the important role ofinvestment security, local support, and loan policyin providing a stable, productive environment forRES generators. It then looks at the role ofadjustment mechanisms to adapt to changingmarket conditions and analyzes the impact of

RES support on employment and general eco-nomic effects. Finally, it addresses the interactionof RES support and environmental policies,namely the emissions trading regime.

Feed-In Tariffs: Effective and Efficient?

Germany is a global leader in new wind, solar, andbiomass development (Lipp 2007). Although theFIT system, unlike a quota system, does not pro-vide a guaranteed RES share, the stated objectivesof the German government could be fulfilled andeven allow challenging future targets such as thosediscussed above to appear reasonable. However, ifthe government’s policy appears to be effective insupporting investment in RES technologies, it isalso helpful to analyze whether the policy is effi-cient from an economic point of view.

A FIT is often regarded as a non-market-basedsupport mechanism, contrary to a quota or cer-tificate system where the desired quantity is fixedand the price is determined by market partici-pants. Consequently, the latter is assumed to bemore cost-efficient in achieving specific RES tar-gets. Under a FIT market, participants can stilldecide about the output, because the quantity isfree whereas the price is fixed.Thus a FIT is simi-lar to a classical ecological tax and in the same

Hydro Wind Bio PV Geo Import Nuclear Fossil

12%

28%

25%

24%

18%

2006

BAU 1

BAU 2

BAU 3

BAU 4

Green 1

Green 2

Green 3

Nuclear 1

Nuclear 2

0 100 200 300 400 500 600 700

TWh/a

37%

51%

53%

22%

22%

Source: Wischermann and Wagner (2009)Note: Percentage values show the share of RES in total generation

Figure 14.5. Share of RES from total power generation in 2030 for different future scenarios

288 HannesWeigt and Florian Leuthold

sense faces the problem of setting the right “taxlevel.” Consequently, a FIT and a quota systemcan theoretically obtain the same outcome givenperfect information.

In the real world, however, the condition ofperfect information rarely prevails. Germany’schoice of a technology-specific FIT represents apolitically motivated market intervention that islikely to reduce economic efficiency, and it hasmany objectives, perhaps too many. The eco-nomic justification for the support of RES tech-nologies can be seen in the external effects. Onthe one hand, negative external costs fromcarbon-emitting fuels ought to be internalized.On the other, society should benefit from positiveexternalities resulting from learning curves thathelp reduce the investment costs of RES tech-nologies. In theory, learning effects (learning rate,progress ratio, or experience curve) result in areduction of unit costs or increase in performanceas experience with a product or process increases.Wright (1936) was the first to describe this con-cept when he reported that in airplane manufac-turing, labor time requirements decreased by aconstant percentage each time cumulative outputdoubled. Based on literature reviews, Abrell et al.(2009) and Wand and Leuthold (2009) find evi-dence for and estimate learning curve parameterswithin their modeling analysis for wind andphotovoltaic technologies. The fact that the EEGdefines different FITs for different technologies isthereby favored by policymakers who assert thatpolitical climate goals can be reached only viasupporting different RES technologies. Further-more, the actual law states that it aims to governthe development of energy supply in a way thatreduces external costs and increases the share ofRES energy production. The law also claims tofind an economically efficient path to reach thesegoals. Simply stated, the EEG, its implementation,and its amendments are driven by a mixture ofpolitical and economic objectives.

As the German FIT provides specific rates fordifferent technologies, a differentiated analysis isnecessary. Today Germany has a total of about 24GW installed onshore wind capacity, the second-largest share in power plants after hard coal, whichhas about 29 GW. RWI (2009) estimates the total

cost borne by all consumers for the currentlyinstalled wind capacity based on the FIT atbetween €11 billion and €20 billion ($15 billionand $27 billion).11 Given this cost estimate, thequestion is whether the policy objective toincrease the share of RES could have beenachieved at a lower cost. Butler and Neuhoff(2008) compare German and UK wind energydevelopment to evaluate the performance of theUK quota and German FIT system. They suggestthat the average lifetime costs of supporting windenergy have been higher under the quota system.Because of the on-average higher wind speeds inthe United Kingdom, the expected return oninvestment for wind will be higher given similarinvestment costs. Butler and Neuhoff (2008)adjust for the difference in wind resources andcalculate the price paid to wind generators(excluding network or balancing costs) to showthat the adjusted price for wind energy in Ger-many would have been lower than in the UnitedKingdom since the mid-1990s. Furthermore, theprice for the UK estimates accounts only for con-tracted sites, not for those actually commissioned.Particularly in the late 1990s, the average contractprices were considered not to be profitable by amajority of wind developers. Thus the price levelconsidered by Butler and Neuhoff (2008) may notreflect the full cost coverage. They also show thatthe German FIT increased competition for tur-bine production and construction, the marketsegment responsible for most of the systems’ cost.This increased competition could have had a sig-nificant impact on the price of wind energy.

Ragwitz et al. (2007) analyze the effectivenessof different RES support mechanisms in Europe.They measure the effectiveness of a supportscheme as the increase in normalized electricitygeneration compared with a reference quantity.Thus the measurement does not account for thecosts of the scheme. They observe the highesteffectiveness in countries with a FIT system, suchas Denmark, Spain, and Germany. Finon (2007)provides an efficiency estimation of wind energysupport based on the data of Ragwitz et al.(2007). Using the measure of output per capita,they show that countries that apply a FIT performbetter than countries that rely on a tradable green

Experience with Renewable Energy Policy in Germany 289

certificate (TGC) system (Figure 14.6). They alsoestimate the expected revenues of investors inonshore wind projects in the different countries.Under a FIT, the profits are lower than under aTGC system. The exception is Sweden, whichintroduced its quota in 2003 following a system oftax credits and investment subsidies.

The European Commission (2005) comparesthe prices for wind generation among severalcountries, concluding that the profit ratios ofwind generators in the United Kingdom areabout five times higher than in Germany. Thehigher price level is due to the higher riskinvolved in quota systems (see Toke 2005), thehigher administrative costs (European Commis-sion 2005), and the fact that increasing marketprices are not reflected in a FIT system. The ECalso determined that the FIT system provides bet-ter incentives for innovation and dynamic effi-ciency. Thus for wind energy, the Germanapproach has been effective in terms of quantityand is cost-efficient compared with the existingquota systems.

To our knowledge, there have been no exten-sive empirical analyses for biomass, landfill gas,and geothermal production. Similar to the sup-port for wind energy, however, the EuropeanCommission estimates a cost advantage comparedwith the United Kingdom for biogas and small-scale hydropower plants (European Commission2005). A further, not yet quantified, aspect of

biomass is its impact on the German heat market.Because of its diversity, biomass provides severaloptions for heat production, and RES areexpected to provide a significant share of the Ger-man heat demand in the future, with biomass pro-viding the largest part (BMU 2006). Bürger et al.(2008) analyze several possible support instru-ments for promoting RES usage in the heat sec-tor, providing first quantitative projections. How-ever, an economic assessment of the 2009Renewable Energy Heat Law is not yet possible.

Finally, Germany’s support for solar energyremains to be evaluated. PV generation has thehighest FIT, accounting for roughly 5 to 10 timesthe wholesale price of electricity. Although thecost of solar power decreased by 60% between1990 and 2000, reductions have been less signifi-cant since, and German customers pay more forphotovoltaic modules (retail) than in Japan or theUnited States (Wüstenhagen and Bilharz 2006).RWI (2009) estimates the costs for the 5.3 GW ofinstalled PV capacity in Germany at €35 billion($48 billion), given the 20-year feed-in guarantee.Assuming a further increase of capacity until 2010similar to recent investments levels, the costs willrise to €53 billion ($72 billion). RWI also providean assessment of the emissions abatement costsusing photovoltaic generation based on an averageemissions factor of the German power plant mixof 0.58 kilograms of CO2 per kilowatt-hour(kgCO2/kWh). They obtain abatement costs of

Austri

a

Germ

any

Spain

Italy UK

Belgium

Sweden

120

100

80

60

40

20

0

120

100

80

60

40

20

0

Out

put [

kWh/

capi

ta]

Exp

ecte

d re

venu

es [

/MW

h]

Output [FIT] Output [TGC] Revenue [FIT] Revenue [TGC]

Source: Finon 2007

Figure 14.6. Annual average wind output, 2000–2004, and expected revenues for onshore wind

290 HannesWeigt and Florian Leuthold

more than €700 ($954) per metric ton of CO2

(tCO2), which is still below the IEA (2007a) esti-mate of €1,000 ($1,362) per tCO2. The numbershighlight PV generation’s significant cost disad-vantage as a result of high investment costs andlow utilization factors. A comparison of subsidycosts under different systems and solar resources isnot yet available. Witzmann and Kerber (2007)analyze the impact of increased solar generation inGermany’s distribution network and concludethat despite the decentralized RES generation,and thus a more equalized generation distribu-tion, grid investment will be necessary.

The additional burden of RES subsidies iscollected from electricity end users via an addi-tional tariff payment. Despite the large total cost,however, the end-user price increase is moderate.In 2006, the surcharge was about 0.75 euro cent(1.02 cents) per kWh. With the expected growthof RES in coming years, this charge will steadilyincrease by about 1.5 euro cents (2 cents) perkWh and continuously decrease after 2020(Büsgen and Dürrschmidt 2009).12 Comparedwith other taxes that consumers pay on electricityprices, the EEG surcharge does not represent anexcessive fee: the concession fee amounts to 1.32to 2.39 euro cents (1.8 to 3.26 cents) per kWh,13

depending on the community size, and the energytax amounts to 2.05 euro cents (2.79 cents) perkWh. However, the overall share of governmentcharges in the German household electricity pricesums to about 40%.14

Finally, the expenses for RES can be com-pared with governmental spending for fossilenergy (see Jacobsson and Lauber 2006). Addingto the financial flow caused by the FIT, GermanRES received about €1.6 billion ($2.2 billion) inR&D funds between 1975 and 2002, plus addi-tional millions for support measurements such asthe 100,000 Roofs Program. The total R&Dspending for coal is €2.9 billion ($3.95 billion) inthe same period; subsidies for the uncompetitiveGerman hard coal are between €80 billion and€100 billion ($109 billion and $136 billion) since1975. Nuclear fission received about €14 billion($19 billion) in R&D, and the fusion program hasexpenses of about €3 billion ($4 billion). In sum-

mary, governmental support for RES is compara-ble with support for conventional fuels.

Given this fiscal record and the political goalto increase the RES share, German support hasbeen effective and cost-efficient for wind andmost likely for biomass. For solar energy, the pic-ture is less optimistic. Although in quantitativeterms the increase of solar capacity is impressive,the cost burden on society is high.This evaluationdoes not draw any conclusion whether the postu-lated objective of increasing the RES share to spe-cific targets is in itself economically reasonable.15

The economic justification for RES oftenemphasizes future expectations, however, such asfor structural cost changes through learningeffects. A study commissioned by the Ministry forthe Environment (BMU 2008b) evaluates theimpact of German RES support on the economy,estimates the occurring cost burden, and providesfuture development scenarios.16 The estimate ofthe differential costs17 resulting from the GermanRES support in the coming years finds that as aresult of the increased share of RES, costs willincrease until 2015 and steadily decrease after,because of learning-curve effects and the fossilfuel price increase. From about 2025 on, RESwill lead to a price-reducing effect (Figure 14.7).The resulting differential costs can be comparedwith the avoided external costs and further posi-tive impacts of RES generation. These includeenvironmental damages due to global warmingand pollution effects on health, crops, and materi-als.18 The BMU assumes that the avoided externalcosts are already higher than the occurring differ-ential costs.

Although the neutrality of the BMU con-cerning RES can be doubted, the general trend ofits study is similar to other studies conducted.Wischermann and Wagner (2009), who providean overview and comparison, find that the pro-jected differential costs all show a peak around2015, ranging from €2.5 billion to €6.2 billion($3.4 billion to $8.4 billion), and negative differ-ential costs after 2040 at the latest. The studies donot answer whether the expected cost advantageof RES in the future justifies the current supportexpenses or whether the efficiency gains are partlydestroyed by the effects on the conventional gen-eration fleet.

Experience with Renewable Energy Policy in Germany 291

Risk, Supporting Mechanisms,and Local Support

The German support policy aims at a maximumof RES expansion. Hence the schemes weredesigned such that potential investors areincentivized to invest as much and as early as pos-sible. One important way to entice investors is acombination of direct financial support (the FITschemes) and an indirect benefit, the reduction ofuncertainty and exposure to market risks. Theweakness of this approach is that particularly thelatter benefit for investors causes costs within thesystem that cannot be quantified easily. The costsare basically shifted from the RES investors andborne by the customers. A study commissionedby the Ministry for the Environment (BMU2008b) shows that most of the entrepreneurs inthe RES segment are satisfied with the GermanFIT and view it as the major contributor to suc-cessful development.

Contrary to a quota system, the German FITschemes provide investors with a hedge againstthree different types of risks: price, volume, andbalancing risk (Mitchell et al. 2006). First, theguaranteed feed-in tariffs provide RES investorswith price security regarding the sales price.Theyalso provide RES generators with a perfect andcost-free hedge against price volatility, which isrelatively high in liberalized electricity markets.

The volume risk is eliminated by the priority ofthe RES feed-in. Because the system operator isobliged to accept the supplied renewable genera-tion, sales will largely equal the potential produc-tion. Finally, balancing risk is eliminated by theFIT as well as by the exemption from balancingrequirements. Other market participants in a lib-eralized electricity market are typically penalizedif the projected and actual load profiles diverge. InGermany, the generator is not penalized for fluc-tuating generation, but the costs are passedthrough to end users. This protection from bal-ancing risk on the generation level does not pro-vide incentives to improve the reliability of RESgeneration and thus may induce higher socialcosts in the long run.

As a consequence, the low-risk situationincreases the ability of investors to finance theirprojects via the financial markets, reduces thecosts of capital, and provides long-term stability,incentives, and resources for innovations (Mitchellet al. 2006). Ragwitz et al. (2007) compare theeffectiveness of different support mechanismswith the expected profit for RES investors. Theyshow that for Spain and Germany, despite lowproducer profits, the effectiveness in quantityterms is best. This is due to the low market riskfaced by investors.

Another important risk- and cost-reductionfactor of the German RES scheme is the shallow

14

12

15

12

9

6

3

0

10

8

6

4

2

0

2000

2002

2004

2006

2008

2010

2012

2014

2016

2018

2020

2022

2024

2026

2028

2030

Avoided external costs Differential costs Differential costs w/o PV

Ext

erna

l cos

ts b

n /

a

Diff

eren

tial c

osts

bn

/a

Source: BMU 2008b

Figure 14.7. Estimated differential costs and avoided external costs as a result of German RES production

292 HannesWeigt and Florian Leuthold

grid connection charges. RES investors have topay only for the connection to the nearest gridconnection point. The system operator must real-ize necessary grid reinforcement investments, andthe occurring costs are redistributed to end usersvia the network charges. After taking into accountthe opposing social effects of shallow grid con-nection charges, this procedure tends to be a sim-ple redistribution of costs within the system.According to Klessmann et al. (2008), in case ofshallow grid connection charges, social cost couldbe lower than when the RES investor would needto raise capital for the investments because oflower capital costs. However, the issue of regu-lated network expansion must then be revisited,which raises questions about regulatory ineffi-ciencies and the realization that shallow connec-tion charges do not provide a decent investmentsignal for plant siting.

Although, the FIT of the StrEG and EEGprovided the basis of the German RES support,the guaranteed tariffs alone would not have led tothe observed capacities in the market withoutadditional measures.The different supporting pro-grams mainly provided investors with access tolow-cost loans. For solar energy, the 1,000 and100,000 Roofs Programs were crucial in provid-ing private investors with the required capital.Theprograms aimed at individuals, associations, foun-dations, and small-scale enterprises and guaran-teed a long-term loan with very low- and fixed-interest rates (Bechberger and Reiche 2004). Thehigh remuneration of the EEG alone provedinsufficient as the investment process significantlyslowed in 2003 with the phaseout of the 100,000Roofs Program.

For other RES projects, the MAP was par-ticularly important. Financial support, in the formof direct investment subsidies and soft loans, wasavailable for solar thermal, biomass and biogas,small-scale hydro, and geothermal plants(Bechberger and Reiche 2004).The success of theprogram led the Ministry of Environment to con-tinue it after the planned expiration in 2003. Forwind energy, the Environment and Energy Con-servation Program and the Environment Programwere crucial elements because MAP did not sup-port wind. The programs particularly targeted

small enterprises, local companies, and public–private partnerships. The importance for windenergy is highlighted by the fact that between1990 and 2002, 95% of the granted €10 billion($13.6 billion) went to wind projects (BMU2003).

Local and public support for RES should notbe underestimated. From the beginning, thedevelopment of the RES sector was closely con-nected to associated companies and the localpopulation (Bechberger and Reiche 2004). In theearly 1990s, municipal utilities and market intro-duction programs at the federal level formed thebasis of the sector. They bridged the gap betweenthe first government programs in the 1980s andmarket support.The large number of local feed-inlaws also highlights the public’s interest in RES(Jacobsson and Lauber 2006), and Germans havewidely supported the increase in RES capacityexpansion (Zoellner et al. 2008).

About 50% of the installed wind capacity inthe country is locally owned. Local initiativesreduce possible “not-in-my-backyard” (NIMBY)resistance to the erection of new facilities, arecheaper than corporate projects with a higheroverhead, and consequently lead to a higherinstalled overall capacity (Toke 2007). Further,German municipalities have to identify possibleareas for wind farms in their spatial planning pro-cesses; this in turn speeds up the investment proc-ess and additionally reduces local resistancebecause municipalities are free to choose where tobuild (Reiche 2002).

From the above, it can be argued that compo-nents exist in the German support scheme thatmore or less do not affect welfare but do shift costsfrom RES producers to other marketers, such asin the case of shallow grid connection charges.Other components, such as the merit-order effect(see the section titled Operational Aspects below)and the indirect benefits from reduced risk expo-sure, are likely to have welfare effects, particularlyin the long term. The economic problem of theindirect benefits for RES investors becomesapparent when examining the extent to whichRES can replace conventional capacities. Thiscapacity replacement potential, or capacity credit,is defined as the fraction of RES energy by which

Experience with Renewable Energy Policy in Germany 293

conventional capacity can be reduced without aloss in security of supply. The capacity factor ofwind energy is estimated to be rather low (com-pare the discussion in the section titled StrategicAspects below). Consequently, the reduction inrisk exposure incentivizes investments in REScapacities that are basically added to the existinggeneration fleet but cannot replace existing units.

Adjustment Mechanisms andPolitical Conditions

RES are assumed to have significant learning-curve effects that will decrease future investmentcosts. An inflexible feed-in scheme with fixed tar-iffs would lead to an excessive burden for con-sumers, because the benefit of reduced generationcosts would only increase the producers’ rent.Furthermore, a fixed feed-in payment could ham-per innovation and lead to inefficiencies duringchanging market conditions.

To avoid these effects, the German FIT incor-porates two mechanisms. First, the guaranteedtariff is reduced each year by a predefined per-centage level for newly installed capacity.19 Thusthe generation costs of RES have to decrease toremain profitable. In the long term, the supportcan be stopped, as it will become more profitablefor RES generators to participate in the openelectricity market. As an example, wind genera-tion currently receives an average feed-in tariff ofabout 9 euro cents (12 cents) per kWh (based on2007 values). The price for a yearly electricitypeak-load future in 2008 topped €100 ($136) permegawatt-hour (MWh) as a result of increasedfuel prices (VIK 2009). Consequently, a windfarm owner could possibly earn a higher profit inthe wholesale market than under the FIT, but at ahigher risk because of wind’s intermittent genera-tion. High electricity prices due to eitherincreased fuel costs or tighter emissions restric-tions will most likely render RES generationcompetitive in the future.

The second mechanism is the periodic revi-sion process of the EEG, which evaluates progressand allows the government to adapt to changes inthe market environment.This amendment processtheoretically ensures that political expectations

and actual market forces correlate; for example,the reduction of the feed-in tariff for solar genera-tion in the last amendment accounts for the costreduction in recent years, and similarly, theincrease for offshore wind generation accounts forthe lack of project progress. However, the experi-ence with the two amendments so far also showsthat the process is lengthy, influenced heavily bylobbyists, and produces a high degree of uncer-tainty (Agnolucci 2006). This last point is a prob-lem for investors and could possibly result in stop-and-go behavior.

Institutional arrangements assume an impor-tant role in the stable development of RES sup-port, particularly regarding who holds responsi-bility for RES. With the shift in 2002 from theMinistry of Economics, which traditionally ismore aligned with industry, to the Ministry ofEnvironment, the German RES policy hasstronger political support, which was lackingwhen the FIT was introduced (Lipp 2007). In1999–2000, the Ministry of Economics delayedthe FIT draft bill several times until the parlia-mentary parties presented their own bill to speedup the process (Bechberger and Reiche 2004).Other important administrative support for REScan be seen in the decision to phase out nuclearenergy in Germany, which will lead to a higherdemand for RES investment in the middle andlong run. However, this decision has reappearedon the political agenda after the election in Sep-tember 2009.

Employment Effects andEconomic Impacts

Support for RES has led to expanded investmentsin production and construction of RES genera-tion technologies that would not have been prof-itable without the support. Such investments haveled to positive gross employment effects in acountry where unemployment hovers at up to10%. As a result, politicians have increasinglystressed the employment effects of RES support.Particularly the Green Party claims a positiveimpact of supporting RES and a sustainableeconomy on both the employment situation in

294 HannesWeigt and Florian Leuthold

Germany and the competitiveness of Germanindustry in the global economy (Bündnis 90/DieGrünen 2009).

There are also negative effects, however, suchas increased energy costs for consumers because ofthe refinancing of RES support and adjustmentsin production and transport (Hillebrand et al.2006). Second, investments in RES generationwill reduce investments in fossil generation, and aportion of the newly formed jobs will be offset byadditional unemployment in the conventionalenergy sector. Also, the lower public budget as aresult of RES support will have a negative impacton government investment and spending (Lehr etal. 2008).

The gross effect of RES support is easy toassess. A study commissioned by the Ministry ofEnvironment (BMU 2007) quantifies the overallemployment numbers in the RES sector to about250,000 people employed in 2007 (Table 14.1).The ministry expects the subsequent trendingincrease to result in about 400,000 employees in2020, although not all sectors will have a similardevelopment; for example, the photovoltaic andbiofuels sectors are assumed to have reached theirpeaks (BMU 2007).

The net effect of RES support is less clear anddifficult to estimate, however. BMU (2007)assumes that the net employment effect will be inthe range of 50,000 to 120,000 additional jobs,depending on assumptions about the export capa-bility of the German RES industry. Pfaffenbergeret al. (2003) estimate a cumulative net loss ofabout 20,000 jobs within a 20-year period.Hillebrand et al. (2006) analyze the net effects of

German RES support with econometric meth-ods.They observe that the positive overall effect ofthe first years as a result of investments will evenout and eventually become negative. Lehr et al.(2008) develop an input–output vector for RESand several policy scenarios. They show that thenet effect is positive throughout the scenarios andthat exports in particular are an important driverof future development, whereas domestic invest-ments lead to a rather small positive labor effect.

Thus the net employment effect of RES poli-cies appears slightly positive at best, contrary topolitical perception. Furthermore, the argumentthat labor-intensive energy generation is favorableis part of the discussion (Michaels and Murphy2009). Note that local employment impactsregardless of the overall economic effect are alwaysimportant for elected officials. A large proportionof RES investments in generation and facility pro-duction occurs in areas with high unemploymentrates, particularly eastern Germany (Handelsblatt2007), and thus contributes to the industrialdevelopment of those regions. Finally, the Ger-man RES policy provides other countries thatsupply the German market with a positive netemployment effect.

Interaction of Environment Policies

RES support in Germany embraces several instru-ments to support the market entry of RES suppli-ers. Beside pure support systems, accompanyingenvironmental aspects have gained increasedattention, finally culminating in the establishmentof the European Union Emission Trading System(EU ETS). Environmental policies that targetenergy markets have overlapping objectives andinteract to a specific degree with the RES supportmechanism.20 This interaction produces bothsynergies and conflicts (Gonzáles 2007).

The main interaction of the EU ETS andRES support instruments is the impact on theCO2 emissions level. The ETS introduces a priceon emissions that in turn increases the electricityprice. In contrast, RES support desires to increasethe share of renewable generation and thus indi-rectly leads to a reduced emissions level. A largershare of RES reduces the demand for emissions

Table 14.1. Gross employment figures of the RESsector in Germany

1998 2002 2004 2006 2007

Wind 16,600 53,200 63,900 82,100 84,300

Solar 5,400 12,700 25,100 40,200 50,700

Hydro 8,600 8,400 9,500 9,400 9,400

Geothermal 1,600 2,400 1,800 4,200 4,500

Biomass 25,400 29,000 34,200 45,200 44,200

Totala 66,600 118,700 160,490 235,640 249,300

Sources: BMU 2007, 2008caIncluding biofuels, services, and publicly financed positions

Experience with Renewable Energy Policy in Germany 295

allowances and thus has a decreasing effect onallowance prices and in turn on electricity prices.However, the allowance price reduction isobtained by using abatement technologies (RES)that typically would not be cost-efficient in a pureETS scheme.

Rathmann (2007) analyzes the impact of windenergy on emissions allowance prices and its feed-back on market prices, concluding that in thepresence of an ETS, additional wind generationcan reduce allowance prices (and thus electricityprices). For the first trading period between 2005and 2007, he estimates a reduction of electricityprices by €6.4 ($8.7) per MWh for Germanysolely by reduced emissions prices, not taking intoaccount price impacts as a result of a changeddispatch. The savings are higher than the increasein electricity prices of €3.8 ($5.2) per MWh fromthe feed-in tariffs.

Traber and Kemfert (2009) analyze the com-bined impact of the German FIT and the ETS onthe electricity market using a bottom-up modelapproach with strategic company behavior. Theyshow that the FIT has a total additional burden ofabout €5 ($6.8) per MWh, of which the producerprice decrease is the largest share. The decreasecan be divided into a substitution effect and apermit price effect of roughly the same size, illus-trating the feedback effect of the ETS. They alsodemonstrate that the emissions reduction in Ger-many as a result of the EEG has no impact on theoverall European emissions level; the additionalreduction in Germany reduces the burden onlyfor other member states.

Abrell and Weigt (2008) analyze how the EUETS and renewable support mechanisms influ-ence one another by applying a static openeconomy computable general equilibrium (CGE)model of Germany incorporating different con-ventional and renewable generation technologies.They test the impact of a pure ETS scheme, agreen certificate scheme, and the combination ofboth schemes on generation investment and mar-ket prices. They observe that the restrictionsimposed by the ETS and a quota are achieved byreducing the output of electricity and shiftinggeneration to green or emissions-free technolo-gies, or both. If both environmental restrictions

are regarded simultaneously, a lower CO2 pricecan be observed. The lower carbon price in turnsupports “dirty” generation from fossil fuels andreduces the investments in CCS technologies,which aligns with the findings of other studies(e.g., Böhringer and Rosendahl 2009; Pethig andWittlich 2009). The three environmental policycases analyzed in Abrell and Weigt (2008) lead to awelfare decrease of about 0.6‰ compared with anunrestricted benchmark (Table 14.2).

A study commissioned by the Ministry ofEnvironment (BMU 2008b) evaluates the interac-tion of German RES support with other climateand environmental instruments. Regarding emis-sions trading, the study highlights the importanceof considering RES generation when setting thetarget cap. In contrast, Germany’s ecological taxreform shows little interaction with RES support.Particularly in the electricity sector, the impact oftaxes is negligible, because energy from RES istaxed similarly to fossil energies concerning endusage. Also, differentiated taxes on fuels for elec-tricity generation have been annulled by Euro-pean guidelines. Thus the possible regulatingeffects in the fuel choice in favor of RES are notachievable. Finally, an introduction of specificnon-RES taxes would not induce any furtherinvestments in RES, as these are defined by theunaffected FIT. Support for combined heat andpower production (CHP) shows no direct interac-tion, because plants can profit from only one sup-port scheme. Conflicts can arise in the priorityfeed-in of both RES and CHP electricity in casesof network congestion.

Table 14.2. Impact of emissions trading and RESsupport on the German economy

BAU ETS QuotaETS +quota

Electricity price(%)

100 107.58 103.48 107.42

Carbon price(€/tCO2)

— 10.04 — 7.29

Green certificateprice (€/MWh)

— — 26.1 12.1

Welfare (‰) 0 –0.598 –0.602 –0.64

Source: Abrell and Weigt 2008Note: €1 = $1.36 as of this writing

296 HannesWeigt and Florian Leuthold

Last but not least, all measures show an indi-rect impact, because they all represent an addi-tional burden for consumers and lead to priceincreases, which induce consumption effects.

Summary of the Economic Analysis

The German RES policy based on fuel-type-specific FITs and investment programs provideinvestors with a strong framework that has led to alarge increase in total numbers. The approach wasclearly effective but lacks efficiency. For windenergy, Germany’s support payments are lessexpensive than in other countries, whereas theyare extremely expensive for solar. The implemen-tation of a technology-specific FIT has beendriven not only by economic factors, but also by amix of economic and political reasons, includinglearning-curve expectations. The political argu-ment for the implementation of more than oneFIT is that learning for more than one technologycan be achieved, which hedges uncertain out-comes and addresses the issue of security of sup-ply.

Market Implications of RESSupport: The Case ofWind EnergyBecause of its development, wind energy repre-sents the renewable source with the largest impacton market operations. This section focuses onGermany’s experiences with wind energy. Theexperience from wind integration can be largelytransferred to solar generation and is characterizedby its stochastic nature. They differ only in thestochastic production profile. In contrast, biomass,hydro, and geothermal production can be oper-ated just as conventional generation plants; hencetheir impact on electricity market operations willnot be stressed here.

This discussion begins by briefly highlightingthe characteristics of the German wind energysector.This is followed by an analysis of the effectsof wind energy on operational aspects in the shortterm and strategic impacts in the long term. It is

argued that the integration of wind generationnecessitates transparent market structures for allenergy market segments: spot, forward, and bal-ancing. Furthermore, efficient locational pricesignals are required in order to support efficientcongestion management and extension planningfor both network and RES generation locating.

Wind Energy in Germany

Wind has been the fastest growing RES source inGermany. This strong growth was made possibleby the support scheme of the German govern-ment, discussed earlier in this chapter. The majorbreakthrough for RES technologies, particularlyfor wind, came with the EEG in 2000. As a resultof the EEG feed-in tariffs, installed wind capacitygrew from about 2 GW in 1997 to about 24 GWin 2008 (BWE 2009). The tariffs were readjustedin 2004 and 2008. Table 14.3 displays the feed-intariffs as stipulated in the EEG. The tariffs areguaranteed for a period of 20 years, during whichthey change from higher initial tariffs to lower-end tariffs according to the yield of the unit.These numbers are known by the investor inadvance. The tariffs decrease beginning with theyear in which the law became effective, such thatturbines that go online earlier receive higher sup-port. Hence the EEG fosters the growth of windin two ways: the investor does not have uncer-tainty about the electricity price, because it alwaysreceives the guaranteed tariff; and the EEG stipu-lates a priority feed-in for RES-generated elec-tricity, which means that the Transmission SystemOperator (TSO) always accepts and dispatcheselectricity from wind turbines first.

Germany has both onshore and offshore windpotential. For the time being, increased windgeneration has been achieved by onshore technol-ogy, but for future developments, it is estimatedthat offshore technology will be of greater impor-tance (DEWI et al. 2005). Germany treats windenergy as intermittent base load generation(Klessmann et al. 2008). Even with larger turbines(compare the discussion of StrEG earlier in thischapter), however, a single turbine cannot replacean entire fossil plant. Therefore, wind generationtakes place by building so-called wind farms,

Experience with Renewable Energy Policy in Germany 297

which are an agglomeration of turbines.The gen-erated energy is collected locally at a bus, whichthen connects the facility to the distribution ortransmission grid. Nevertheless, onshore windusage leads to a rather decentralized generationstructure, because the farms are widespread overnorthern and central Germany. The exploitationof feasible onshore wind farm locations is alreadyhigh in Germany. Hence associated operationalmarket issues can be easily observed andaddressed.

Operational Aspects

Integrating wind in the existing German marketstructure is challenging, because wind energy hasto be treated differently in the market than the“classical” generation technologies. Major fieldsof the energy market organization that have to beconsidered in order to assess the market impact ofwind are the applied support scheme, grid code,balancing regulations, and collateral market regu-lations (Klessmann et al. 2008). These fields ofinterest are the distinctive criteria when compar-ing country examples.

The German market design distinguishes aforward market and a balancing market. A real-time market has only recently come into force.The term “spot market” is also used, but it refersto the day-ahead market.Thus most market trans-actions concerning energy delivery within oneday (intraday) have to take place via balancing,which is managed by the TSOs. Wind energy

operators are exempt from the balancing require-ment. Consequently, three elements are particu-larly important for the integrated operations ofelectricity markets with extensive wind energyproduction: economic integration in forwardmarkets (ahead of delivery); operational integra-tion in forward markets; and operational integra-tion in balancing markets (compare Klessmann etal. 2008).

The forward market in Germany is dividedinto bilateral over-the-counter (OTC) and stand-ardized future trades. The future trades are man-aged by the European Energy Exchange (EEX).The market is dominated by OTC transactions,whereas the EEX made up only about 19% of themarket in 2007 (Schöne 2009). The balancingmarket is required because electricity is highlyinelastic in a short-term perspective. Hence thedemand for electrical energy must be balancedwith the supply (generation) at each point in time,or otherwise the system will collapse. Althoughconcepts such as demand-side management andinterruptible load programs exist, for the timebeing balancing has to be achieved to a largeextent by supply-side effects. The balancing serv-ice comprises primary reserve, secondary reserve,and tertiary reserve. The system operator providesthese balancing services to compensate physicallyfor flows that deviate from the schedules.21

Klessmann et al. (2008) state that balancing windnormally requires the cheaper tertiary reserve.

This is partly due to the way wind is inte-grated in the German market. Regular marketplayers such as traders and fossil generators have toannounce their schedules a day ahead and sendthem to the system operators.These marketers areobliged to pay if their actual flow deviates fromthe schedule that is specified in the imbalance set-tlement rules. Wind, however, is exempt from thisrule. Additionally, wind generators do not have toprovide primary reserve. The TSO assumes manyobligations of the wind operators that a regularmarket participant would have to fulfill individu-ally. The TSO integrates the stochastic generationprofile of the wind producers in the merit order ofthe entire plant fleet. For this purpose, the TSOdefines monthly standard profiles of wind produc-tion based on its wind predictions. These profiles

Table 14.3. Feed-in tariffs in Germany to EEG 2008and EEG 2004 (in €/MWh)

EEG 2008 EEG 2004

Windonshore

Annual decrease:1%

Annual decrease:2%

Initial tariff 92 78.7

End tariff 50.2 49.7

Wind off-shore

Annual decrease(from 2015): 5%

Annual decrease(from 2008): 2%

Initial tariffa 130 87.4

End tariffa 35 59.5

aCommissioning until 2015: bonus of 20 €/MWhNote: €1 = $1.36 as of this writing

298 HannesWeigt and Florian Leuthold

are sold to all players that deliver electrical energyto final customers. For them, the profiles are con-stant. The cost associated with converting inter-mittent wind generation into a constant profileare borne by the TSO and passed through to allnetwork users via a system usage charge. Conse-quently, distribution and transmission systemoperators pay the feed-in tariffs to the wind gen-erators. These costs are then in turn socializedover all network users. Hence German consumerssupport the extension of wind capacity by directlybearing the costs resulting from the feed-in tariffand assuming the costs resulting from some of theuncertainties associated with wind operation, par-ticularly forecasting, scheduling, and balancing.The advantage of this approach is that the TSOmanages a large number of wind farms and canhedge stochastic wind input geographically. Itcould also realize certain economies of scale inforecasting and minimizing forecast errors. Thusthe TSO manages to integrate the stochastic gen-eration profiles in the fossil dispatch partly by day-ahead settlement and partly by balancing. How-ever, only for wind-caused balancing, the Ger-man TSOs have introduced an own tender inaddition to the regular reserve tender, whichshows one of the system’s disadvantages. In addi-tion, in October 2006, the EEX introduced anintraday market. Nevertheless, the traded volumesare still low, and empirical data on the impact ofthis development are absent. The late establish-ment and illiquidity of the intraday market meansthat wind forecast errors are relatively high,because most of the forecast must be conducted24 hours ahead.

The major problem in Germany, however, isthat there are misleading incentives for several rea-sons. First, the market structure for balancingservices, consisting of basically four different ten-dering systems, causes low liquidity and highprice levels (Klessmann et al. 2008). In addition,the costs of converting the stochastic wind inputinto synthetic constant profiles appear too high.For example, LBD (2007) estimates that the costsrange between €3.4 and €5.4 ($4.6 and $7.4) permegawatt-hour of renewable energy (MWhRES),whereas the average cost charged by networkoperators is about €8.3 ($11.3) per MWhRES.

Another reason is the congestion managementmechanism. Currently, Germany has a singleelectricity price at the wholesale market (EEX),which is determined independently of the loca-tion of demand and supply, thus ignoring conges-tion in the network. Leuthold et al. (2008) modelthe German electricity market and show that thisuniform pricing scheme is less efficient from awelfare perspective than a full nodal pricingapproach, which assigns a price to each node ofthe network taking into account all relevant mar-ket and capacity information. Because of the sin-gle price in Germany, the schedules that areannounced to the TSOs often are not feasiblefrom a technical load flow perspective. In thiscase, theTSO has the right to change the dispatchof the power plant, deviating from the marketequilibrium at the EEX (“redispatch”). Thisredispatch is cost-based in Germany, which meansthat the costs that occur can be socialized by theTSO via the network charges. An alternative doesexist, however, in the form of market-basedredispatch or countertrading. In the latter case, anown merit order is determined for the redispatchby an independent coordination office orexchange, which is a more transparent approachbecause the order books can be made public afterthe market clearing, whereas in the currentscheme the TSOs individually decide what is tobe redispatched (Wawer 2007). In the case of con-gestion, wind energy has the priority right,meaning that it is dispatched first or curtailed lastcompared with fossil units. Regarding the politi-cal wish to expand wind, locational price signalscan have opposing effects. Klessmann et al. (2008)argue that the problem of not having locationalprice signals provides no incentives for wind tolocate efficiently, which leads to efficiency lossesas a result of reduced market response and lesssystem-optimized behavior. On the other hand,locational price signals impose market risk onwind operators that they do not bear now, whichbrings efficiency gains as a result of lower riskpremiums and the lower required support pay-ments. This argument ignores the efficiencyeffect, however: a more efficient congestion man-agement scheme would affect the entire marketoutcome, including fossil production. From an

Experience with Renewable Energy Policy in Germany 299

economic point of view, the introduction oflocational price signals should be mandated,because they improve incentive structure in thenetwork. Higher FITs could compensate for thedifferent risk situations for wind operators.

Klessmann et al. (2008) find that in 2005, theloss through curtailment of wind turbines causedby network congestion accounted for 5% of theannual yield. Jarass and Obermair (2005) showthat there is an increasing tendency to wind pro-duction curtailment. This fact leads to the lastpoint of the operational aspects: grid connectionof wind facilities. In Germany, a wind farm inves-tor now pays for the connection to the nearestgrid access point. If other grid reinforcements arerequired to transport the generated electricity tothe demand locations, they are carried out by thesystem operator. This is a so-called “shallow” costallocation method (e.g., Rious et al. 2008),because the generator does not take into accountthe effect of its investment on the overall gridcongestion situation. Other than building a windfarm, the planning and construction of new high-voltage lines are a long-term issue and will beaddressed in the next section.

Strategic Aspects

The strategic or long-term aspects of integratingwind production into the German electricitymarket mainly affect investment. One issue con-cerns grid investments. Looking at the continen-tal European region managed by the Union forthe Co-ordination of Transmission of Electricity(UCTE), including Germany, Leuthold et al.(2009) analyze the impact of wind capacity exten-sions on the grid. The wind extension scenariosfor Europe are ambitious, with 114.5 GW (IEA2007b) and 181 GW (Greenpeace International/EWEA 2005) of installed capacity. Based on awelfare maximization approach and using thenodal pricing methodology, the optimal invest-ment strategy is determined. Leuthold et al.(2009) show that the remarkable extension ofwind capacity leads to an increase in total Euro-pean welfare compared with a current fuel mixscenario. They conclude that developing the net-work at existing bottlenecks—mainly cross-

border connections—should be encouraged nowby regulatory authorities. With a more moderatewind expansion of 114.5 GW, the optimal gridinvestments are smaller. If the additional windcapacity becomes too great (181 GW), theneeded grid extensions will increase comparedwith the actual situation. Although the gridextensions determined are most likely feasiblefrom economic and political viewpoints, theirstudy rests on two strong assumptions: the windcapacity extension is supposed to occur through-out Europe and not be concentrated within a sin-gle or only a few countries, such that additionalcross-border flow issues due to wind are minor;and the stochastic nature of wind production isneglected, which may lead to an underestimationof balancing costs.The first constraint is importantfor Germany and its neighbors, as Germany plansto develop large-scale offshore wind. DEWI et al.(2005) estimate that Germany will have 30 GWinstalled offshore capacity in 2030. Based on thisassumption, Leuthold et al. (2008) and Weigt et al.(2010) model the economic impact of offshorewind expansion on Germany using nodal pricingas an analysis tool for the German market.Leuthold et al. (2008) assume that the entire off-shore wind capacity will be erected in the NorthSea and connect to the grid at the best possiblenode on the German coast. They find that theoffshore capacity has a price-decreasing andwelfare-increasing effect. However, the presentgrid cannot transmit the 30 GW input to the loadcenters; absent extensive grid expansion, only 8 to13 GW appears feasible. Additionally, in times ofhigh wind production, serious congestion mightoccur to and within the Benelux countries. Incontrast, Weigt et al. (2010) assume that the off-shore wind can be connected to the well-meshedparts of the grid in central Germany via high-voltage direct current (HVDC) technology. Intheir scenario, the price-decreasing and welfare-increasing effect of wind can be realized at theGerman demand centers without additional seri-ous congestion problems.

The two studies assume, however, that thewind energy can be sufficiently backed up byother generation such that there will be no bal-ancing or network stability problems. This

300 HannesWeigt and Florian Leuthold

assumption is valid as long as the considered timeperiods are within the lifetime of the existing fos-sil plant fleet. In this case, increasing balancingrequirements might lead to increasing balancingcosts, but contingency issues appear to be minor.Hence there are decreasing effects on price andCO2 emissions. In the short-term, these effectshave been quantified by some studies. Sensfuß etal. (2008) find that the price effect (“merit ordereffect”) of wind input can range from €0 perMWh during low load periods up to €36 ($49)per MWh during peak load periods. Weigt (2009)finds that the average electricity price in Germanyis about €10 ($13.6) per MWh lower in 2006–2007 because of wind input. When subtractingthe average support expenses of €5.4 ($7.4) perMWh in 2006 and €7 ($9.5) per MWh in 2007,he concludes that Germany still realizes a net ben-efit of wind energy, even when possible additionalgains of reduced emissions allowance prices arenot accounted for. Based on a literature review,Klessmann et al. (2008) determine that the mini-mum average price advantages are €3 ($4.1) perMWh in 2005 and €6 ($8.2) per MWh in 2006.Thus a short-term advantage of increasing windcapacities appears to be proven as a result of theextremely low marginal cost (almost €0 perMWh) of wind production.

The long-term aspect is less clear, however. Aswind production has marginal costs near zero andis exempted from the balancing requirements buthas an intermittent character, the structure of theconventional plant fleet is likely to change. Inves-tors in conventional technology face higheruncertainty regarding the full load hours andprobability to be dispatched in the future. Theexisting base load technologies in Germany,nuclear and lignite, are characterized by highcapital and low fuel costs, which require high fullload hours to recover their fixed costs. Therefore,these types of technology will be associated withhigher investment risk in a long-term perspectivewith constantly high or even increasing windenergy production. Hence a shift in fossil genera-tion technologies to plants with higher variablecosts but lower capital costs could be a conse-quence. The effect of this development could belower electricity prices on average, but with very

high price spikes during peak load times account-ing for the higher marginal cost of the conven-tional generation fleet.22 In addition to thedescribed price and emissions effects, anotherissue is the capacity credit of wind energy. Weigt(2009) estimates that the present installed windcapacity does not allow a significant reduction ofinstalled conventional capacity, because the lowerboundary of Germany’s wind capacity credit isonly about 1%, which means that for the timebeing, basically the entire installed wind capacitymust be backed up by conventional technologies.

Summary of the Market Implications

Regarding the operational aspects of integratinglarge amounts of intermittent renewable genera-tion such as wind energy into electricity markets,it can be concluded that wind producers enjoylow market risk and reasonable financial supportvia FITs. This appears to be a major driver for thecountry’s extensive wind capacity expansion. Thereduction of uncertainty specifically appears to bea crucial aspect in the German feed-in system. It issometimes argued that a quota system is prefer-able, because the market can then determine theoptimal technology investment strategy. Somestudies show, however, that in the presence of cer-tain framework factors, the optimal policy choicecan deviate from the theoretically favorableinstrument. One of these factors appears to be thepresence of uncertainty (compare Fischer et al.2003). However, shortcomings in the Germansystem derive from the opacity and inefficiency ofassociated regulatory issues:

• The balancing market is an administrativeentity and is thus illiquid and inefficient; thesame is true for the process of integratingwind into the electricity market, which isdone by converting intermediate wind inputinto synthetic constant profiles.23

• The long absence and present illiquidity ofthe intraday market means that the windforecast errors are relatively high, as most ofthe forecast has to be conducted 24 hoursahead; shorter gate closure or handling timewould help decrease forecast error.24

Experience with Renewable Energy Policy in Germany 301

• The German congestion managementscheme does not provide locational price sig-nals and therefore does not offer incentives toTSOs or wind farm operators to optimizetheir locations.

Regarding the strategic aspects of integratingwind energy production into electricity markets,the question arises how to account for the lowcapacity credit of wind capacities in an economicassessment. This could be accomplished byincluding the required incremental capital cost ofconventional backup in the investment calculationof additional wind capacity expansion. It is ques-tionable, however, whether society will acceptadditional large-scale expansions of both windand conventional capacities even if it shouldappear to be economic. Technological innovationcould help overcome this problem. A major driverto increase the capacity credit of wind is decen-tralized storage facilities. Using these facilities,large wind farms could handle the intermittencyon the spot and act as a constant generator. Thiscould be particularly interesting in the context ofoffshore expansion, where large amounts of windare transmitted from only a few power bus barsoffshore. However, storage facilities are not yetavailable at reasonable cost. Another long-termaspect is grid investment issues. To address theseissues, the aspect of locational price signals isimportant. Based on these signals, sociallyfavorable grid enforcements could be identifiedmore easily and transparently.The introduction offinancial transmission rights or similar financialproducts can be used to support this process.These rights could also help determine the realnetwork cost of certain grid connection projects.Furthermore, the integration of today’s nationaland regional electricity markets in Europe canhelp increase the capacity credit of wind energy asa result of greater geographic coverage.

ConclusionsWhen reviewing the case of Germany, it can beseen that the record of renewable energy pro-grams is remarkable. In 2007, the total share of

RES in Germany accounted for 6.9% of the pri-mary energy consumption and about 15% of elec-tricity generation. In 2008, Germany had thesecond-highest amount of installed wind capacity,after the United States, and was the world leaderfor installed photovoltaic capacity. One reason forthat is that Germany provides a strong frameworkfor investors which means that society bears mostof the market risks, as the FITs are guaranteed for20 years and RES units are exempted from bal-ancing requirements. Another reason appears toconsist in the technology-specific German FITscheme.

However, the evaluation of the system’s eco-nomic efficiency is ambiguous. Germany’s sup-port payments for wind are less expensive than inother countries, but they are very expensive forsolar, presumably because of environmental fac-tors. Also, the implementation of a technology-specific FIT is driven by a mix of economic andpolitical reasons. The economic argument in sup-port of RES technologies is that learning effectscan be achieved that produce positive externaleffects. The political argument for the implemen-tation of more than one FIT is that learning formore than one technology can be achieved, hedg-ing uncertain outcomes and addressing security ofsupply.

Another conclusion from the German exam-ple is that nonintermittent production (e.g.,biomass, geothermal) can be easily integrated intothe existing market structure as they can betreated the same as other conventional units.Intermittent production (e.g., wind, solar), how-ever, must be backed up by conventional units;hence the capacity credit of these RES technolo-gies is currently low. Nonetheless, there are short-term and long-term effects. In the short term, theRES production from sources with low marginalcosts had a price-decreasing effect in recent years,whereas in the long term, the conventional plantportfolio is likely to gravitate to more flexibleunits with lower fixed and higher variable costcomponents, which would then lead to increasingprices on average.

In addition to that effect, evidence for othermarket inefficiencies could be found. However,the causes of inefficiencies within the system inte-

302 HannesWeigt and Florian Leuthold

gration process appear to be manifold, and costscannot be allocated clearly. The balancing marketis important in the case of intermittent produc-tion. However, Germany’s balancing market is nota real market, but rather an administrative entity,and as such it is illiquid and inefficient. Addition-ally, the long absence and present illiquidity of theintraday market means that wind forecast errorsare relatively high, as the majority of forecastsmust be conducted 24 hours ahead; shorter gate

closure or handling time will reduce forecasterrors. Lastly, Germany’s congestion managementscheme does not provide locational price signals;therefore, it does not offer incentives to eitherTSOs or wind farm operators to optimize theirlocations. Hence, it is not always possible to dis-tinguish between costs and inefficiencies causedby renewable energy integration and failures inthe structure of the different sub-markets for elec-tricity.

AppendixHydro Wind Biomass PV Geo

1990

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

40

35

30

25

20

15

10

5

0

GW

Source: BMU 2008a

Figure 14.8. Installed RES generation capacities, 1990–2007

Wind Solar (thermal) Solar (PV)0.5

0.4

0.3

0.1

0

1985

1987

1989

1991

1993

1995

1997

1999

2001

2003

2005

0.2

2

1.6

1.2

0.4

0

0.8

Win

d, S

olar

/k

Wh

PV

/k

Wh

Source: BMU 2008b

Figure 14.9. Cost reduction for RES generation

Experience with Renewable Energy Policy in Germany 303

Notes

1. This case study is based on Bechberger andReiche (2004), Büsgen and Dürrschmidt (2009),Gan et al. (2007), Jacobsson and Lauber (2006),Lipp (2007), and Toke and Lauber (2007).

2. Other programs, such as the GROWIAN projectto develop a wind turbine with several megawatts,were only demonstration projects.

3. With the start of the program in 1989, 20 MW ofwind energy were installed.

4. Wind and solar power obtained 90% of the end-user tariff; all other small-scale plants (< 500kW)obtained 80%, and large-scale plants (500 kW to 5MW) obtained 65%.

5. The clause provided a twofold 5% cap. The firstcap was at the utility level and allowed a cost pass-through to the Transmission System Operator(TSO) if the amount of purchase obligationsexceeded 5% of the utilities’ total deliveries. Thesecond cap was at theTSO level and withdrew thedispatch obligation for theTSO (Agnolucci 2006).

6. Further support for RES was initiated at themunicipal level, as the 1989 tariff regulationallowed utilities to sign contracts with RES inexcess of their long-term avoided costs. SeveralGerman cities adopted this model. Additional sup-port came from state governments that adoptedspecific support programs highlighting the stillhigh public support for RES (Jacobsson andLauber, 2006).

7. The conservative–liberal coalition was replaced bya social democratic–green coalition.

8. To reduce the burden for the domestic companies,energy-intensive industries and transportationpaid reduced taxes. Because biofuels were exemptfrom the tax, they obtained an advantage of 14.2euro cents (19.3 cents) per liter, which led to anincrease in biodiesel production (Bechberger andReiche 2004).

9. For installed capacities, see Figure 14.8 in theAppendix at the end of this chapter.

10. This amendment process became moreadministration-led with the shift of responsibilityfor RES from the Ministry of Economic Affairs tothe Ministry of the Environment after the share ofthe Green Party in the government coalitionincreased in the 2002 elections (Wüstenhagen andBilharz 2006). This shift of responsibility alsostrengthened the general position of RES, as theMinistry of Economic Affairs tended to supportthe incumbent energy companies.

11. Wind turbines receive a site-dependent premium,so the actual FIT varies for each turbine.The valueof €11 billion ($15 billion) represents the lowerboundary, assuming no extra premiums, and theupper value represents the extreme case that allturbines receive the full premium.

12. These numbers are derived using the EEG amend-ment of 2004. They will likely increase as a resultof the higher remuneration for offshore wind withthe amendment of 2009, and further amendmentsmay provide additional burdens or offsets.

13. The concession fee is a levy by the municipalityfor the right to build a network in the municipal-ity.

14. End-user prices in Germany range from 16 to 20euro cents (22 to 27 cents) per kWh.

15. See also Finon (2007) and Fouquet and Johansson(2008) for a general assessment of different supportmechanisms.

16. An estimate of the cost reduction in RES invest-ments is presented in Figure 14.9 in the Appendixat the end of this chapter.

17. The differential costs estimate the cost burden as aresult of the RES fee compared with the supplycosts without the feed-in requirement.

18. The estimates are based on cost projections byKrewitt and Schlomann (2006): external costs in€/t [$/t] for CO2 (70 [95]), SO2 (3,280 [4,468]),NOx (3,320 [4,523]), PM10 (12,000 [16,348]), andNMVOC (870 [1,185]).

19. A RES plant constructed in 2007 receives the2007 FIT for 20 years, whereas a plant constructedin 2008 receives the lower 2008 FIT for 20 years,and so on.

20. For effects regarding the interaction with energy-efficiency targets, see Meran and Wittmann (2008)and Sorrel and Sijm (2005).

21. In Germany, each running unit with an installedcapacity of 100 MW or more must be capable ofproviding primary reserve. The primary reservemust be available within 30 seconds and for at least15 minutes. The primary reserve market is struc-tured as a monthly tender. The market clearing isbased on a primary reserve merit order, which inturn is based on a capacity price, and the pricingfollows the pay-as-bid rule. The secondary reserveunits must be available within 5 minutes and haveat least a gradient of 2% per second of their cap-acity, which means they must reach full load after50 seconds at maximum. Depending on the typeof provider, the secondary reserve must be madeavailable for several hours, e.g., 4 hours for

304 HannesWeigt and Florian Leuthold

pumped storage plants. The secondary reservemarket is also structured as a monthly tender. Themarket clearing is based on a secondary reservemerit order, which in turn is based on a capacityprice, and the pricing follows the pay-as-bid rule.The tertiary reserve (also called minute reserve)must be available within 15 minutes and must alsobe able to be switched off within 15 minutes. Thecontracted capacity must be available for the entiretime slice. The time slices are 4 hours, resulting in6 predefined periods over the day. The tertiaryreserve market is structured as a daily tender. Themarket clearing is based on a tertiary reserve meritorder, which in turn is based on a capacity price,and the pricing follows the pay-as-bid rule. Inaddition to the capacity payment, an energy pay-ment is provided in the event of usage. Klessmannet al. (2008) state that the economic value of thedifferent balancing services decreases in the sameorder, from primary to tertiary.

22. The same effect is likely to occur in times of lowwind energy production.

23. See Vandezande et al. (2010) for a recent discus-sion on balancing market designs.

24. See Weber (2010) for a recent discussion onintraday market issues.

ReferencesAbrell, J., J. Herold, and F. Leuthold. 2009. Modeling

the Diffusion of Carbon Capture and Storage underCarbon Emission Control and Learning Effects.Innovations for a Low-Carbon Energy SystemWorking Papers, WP-RD-04. Dresden, Germany:Dresden University of Technology.

Abrell, J., and H. Weigt. 2008.The Interaction of Emis-sions Trading and Renewable Energy Promotion.Economics of Global Warming Working Papers,WP-EGW-05. Dresden, Germany: Dresden Uni-versity of Technology.

Agnolucci, P. 2006. Use of Economic Instruments inthe German Renewable Electricity Policy. EnergyPolicy 34 (18): 3538–48.

Bechberger, M., and D. Reiche. 2004. RenewableEnergy Policy in Germany: Pioneering and Exem-plary Regulations. Energy for Sustainable Development8 (1): 47–57.

BMU (Federal Ministry for the Environment, NatureConservation and Nuclear Safety of Germany).2003. Entwicklung der erneuerbaren Energien—Stand:

August 2003. (Development of Renewable Energies– As of: August 2003.) Berlin: BMU.

———. 2006. Eckpunkte für die Entwicklung undEinführung budgetunabhängiger Instrumente zurMarktdurchdringung erneuerbarer Energien imWärmemarkt. (Key Points for the Development andImplementation of Budget Independent Instrumentsfor the Market Penetration of Renewable Energieson the Heat Market.) Berlin: BMU.

———. 2007. Erneuerbare Energien: Arbeitsplatzeffekte2006. (Renewable Energies: Employment Effects2006.) Berlin: BMU.

———. 2008a. Development of Renewable Energy Sourcesin Germany in 2007: Graphics and Tables. Berlin:BMU.

———. 2008b. Analyse und Bewertung derWirkungen desErneuerbare-Energien-Gesetzes (EEG) aus gesamt-wirtschaftlicher Sicht. (Analysis and Evaluation of theEffects of the Renewable Energy Law (EEG) in Wel-fare Terms.) Berlin: BMU.

———. 2008c. Bruttobeschäftigung 2007 -eine ersteAbschätzung-. (Gross Employment 2007 – A FirstEstimation.) Berlin: BMU.

Böhringer, C., and K. E. Rosendahl. 2009. GreenServes the Dirtiest: On the Interaction betweenBlack and Green Quotas. Discussion paper no. 581.Oslo: Statistics Norway.

Bündnis 90/Die Grünen. 2009. Der grüne neueGesellschaftsvertrag. (The Green New Social Con-tract.) Berlin: Bundestagswahlprogramm.

Bürger, V., S. Klinski, U. Lehr, U. Leprich, M. Nast,and M. Ragwitz. 2008. Policies to Support Renew-able Energies in the Heat Market. Energy Policy 36(8): 3150–59.

Büsgen, U., and W. Dürrschmidt. 2009.The Expansionof Electricity Generation from Renewable Energiesin Germany: A Review Based on the RenewableEnergy Sources Act Progress Report 2007 and theNew German Feed-In Legislation Energy Policy 37(7): 2536–45.

Butler, L., and K. Neuhoff. 2008. Comparison ofFeed-In Tariff, Quota and Auction Mechanisms toSupport Wind Power Development. RenewableEnergy 33 (8): 1854–67.

BWE (Wind Energy Association). 2009. Anzahl derWindenergieanlagen in Deutschland. (Number of WindTurbines in Germany.) www.wind-energie.de/de/statistiken/ (accessed June 2, 2009).

DEWI (Wind Energy Institute), E.ON Netz, EWI(Institute of Energy Economics), RWETransportnetz Strom, and VE Transmission. 2005.Energiewirtschaftliche Planung für die Netzintegration von

Experience with Renewable Energy Policy in Germany 305

Windenergie in Deutschland an Land und Offshore biszum Jahr 2020. (Energy-Economic Planning of theNetwork Integration of On- and Offshore WindEnergy in Germany until 2020.) Berlin: GermanEnergy Agency (DENA). www.dena.de/de/themen/thema-kraftwerke/projekte/projekt/netzstudie-i/, accessed January 31, 2009, (in Ger-man).

DLR (German Aerospace Center). 2008. Leitstudie2008:Weiterentwicklung der Ausbaustrategie ErneuerbareEnergien vor dem Hintergrund der aktuellenKlimaschutzziele Deutschlands und Europas. (PilotStudy 2008: Improvement of the Extension Strategyof Renewable Energies in the Context of the Cur-rent Climate Policy Objectives of Germany andEurope.) Berlin: BMU.

European Commission. 2005. The Support of Electricityfrom Renewable Energy Sources. Com(2005) 627.

Finon, D. 2007. Pros and Cons of Alternative PoliciesAimed at Promoting Renewables. EIB Papers.Kirchberg, Luxembourg: European InvestmentBank, Economic and Financial Studies.

Fischer, C., I. W. H. Parry, and W. A. Pizer. 2003.Instrument Choice for Environmental ProtectionwhenTechnological Innovation Is Endogenous. Jour-nal of Environmental Economics and Management 45 (3):523–45.

Fouquet, D., and T. B. Johansson. 2008. EuropeanRenewable Energy Policy at Crossroads – Focus onElectricity Support Mechanisms. Energy Policy,36(11): 4079–4092.

Gan, L., G. S. Eskeland, and H. H. Kolshus. 2007.Green Electricity Market Development: Lessonsfrom Europe and the US. Energy Policy 35 (1): 144–55.

González, P. D. R. 2007. The Interaction betweenEmissions Trading and Renewable Electricity Sup-port Schemes. An Overview of the Literature. Miti-gation andAdaptation Strategies for Global Change 12: 8.

Greenpeace International/EWEA. 2005. Wind Force12:A Blueprint to Achieve 12% of theWorld’s Electricityfrom Wind Power by 2020. Brussels: Global WindEnergy Council.www.gwec.net/fileadmin/documents/Publications/wf12–2005.pdf (accessedFebruary 15, 2009).

Handelsblatt. 2007. Solarindustrie bringt Jobs fürOstdeutschland. (Solar Industry provides Jobs forEastern Germany.) www.handelsblatt.com/unternehmen/industrie/solarindustrie-bringt-jobs-fuer-ostdeutschland;1370712 (accessed October 15,2009).

Hillebrand, B., H. G. Buttermann, J. M. Behringer,and M. Bleuel. 2006. The Expansion of RenewableEnergies and Employment Effects in Germany.Energy Policy 34 (18): 3484–94.

IEA (International Energy Agency). 2007a. Energy Poli-cies of IEA Countries: Germany, 2007 Review. Paris:OECD Publishing.

———. 2007b. World Energy Outlook 2006. Paris:OECD Publishing.

———. 2009. Energy Technology R&D Statistics.www.iea.org/textbase/stats/rd.asp (accessed June 2,2009).

Jacobsson, S., and V. Lauber. 2006. The Politics andPolicy of Energy SystemTransformation: Explainingthe German Diffusion of Renewable Energy Tech-nology. Energy Policy 34 (3): 256–76.

Jarass, L., and G. M. Obermair. 2005. WirtschaftlicheZumutbarkeit des Netzausbaus für ErneuerbareEnergien. (Economic Reasonability of NetworkExtensions for Renewable Energies.) Zeitschrift fürEnergiewirtschaft 29 (1): 3–10.

Klessmann, C., C. Nabe, and K. Burges. 2008. Pros andCons of Exposing Renewables to Electricity MarketRisks: A Comparison of the Market IntegrationApproaches in Germany, Spain, and the UK. EnergyPolicy 36: 3646–61.

Krewitt, W., and B. Schlomann. 2006. Externe Kostender Stromerzeugung aus erneuerbaren Energien imVergleich zur Stromerzeugung aus fossilen Energieträgern.(External Costs of Electricity Generation withRenewable Energies in Comparison to Generationwith Fossil Fuels.) Stuttgart: BMU.

Lauber, V., and L. Mez. 2004. Three Decades ofRenewable Energy Politics in Germany. Energy &Environment 15 (4): 599–623.

LBD. 2007. EEG-Gutachten: Gutachten zurAngemessenheit der Aufwendungen für die Veredelung desEEG-Stromaufkommens durch die Übertragungs-netzbetreiber. www.lbd.de/cms/pdf-gutachten-und-studien/0710-LBD-Gutachten-EEG.pdf (accessedAugust 24, 2009).

Lehr, U., J. Nitsch, M. Kratzat, C. Lutz, and D. Edler.2008. Renewable Energy and Employment in Ger-many. Energy Policy 36 (1): 108–17.

Leuthold, F., T. Jeske, H. Weigt, and C. vonHirschhausen. 2009. When the Wind Blows overEurope: A Simulation Analysis and the Impact ofGrid Extensions. Electricity Markets WorkingPapers WP-EM-31. Dresden, Germany: DresdenUniversity of Technology.

Leuthold, F., H. Weigt, and C. von Hirschhausen.2008. Efficient Pricing for European Electricity

306 HannesWeigt and Florian Leuthold

Networks: The Theory of Nodal Pricing Applied toFeeding-In Wind in Germany. Utilities Policy 16 (4):284–91.

Lipp, J. 2007. Lessons for Effective Renewable Electri-city Policy from Denmark, Germany and the UnitedKingdom. Energy Policy 35 (11): 5481–95.

Meran, G., and N. Wittmann. 2008. Green, Brown,and Now White Certificates: Are Three One TooMany? A Micromodel of Market Interaction. Dis-cussion paper 809. Berlin: German Institute for Eco-nomic Research.

Michaels R., and R. Murphy. 2009. Green Jobs: Fact orFiction? Washington, DC: Institute for EnergyResearch.

Mitchell, C., D. Bauknecht, and P. M. Connor. 2006.Effectiveness through Risk Reduction: A Compari-son of the Renewable Obligation in England andWales and the Feed-In System in Germany. EnergyPolicy 34 (3): 297–305.

Pethig, R., and C. Wittlich. 2009. Interaction of Car-bon Reduction and Green Energy Promotion in aSmall Fossil-Fuel Importing Economy. Workingpaper no. 2749. Munich: CESifo.

Pfaffenberger, W., K. Nguyen, and J. Gabriel. 2003.Ermittlung der Arbeitsplätze und Beschätigungswirkungenim Bereich erneuerbarer Energien. (Estimation ofEmployment Effects in the Renewable EnergiesSector). Düsseldorf: Hans-Böckler-Stiftung, BremerEnergie Instituts.

Ragwitz, M., A. Held, G. Resch,T. Faber, R. Haas, C.Huber, P. E. Morthorst, S. G. Jensen, R. Coenraads,M. Voogt, G. Reece, I. Konstantinaviciute, and B.Heyder. 2007. OPTRES.Assessment and Optimisationof Renewable Energy Support Schemes in the EuropeanElectricity Market. Final Report. Karlsruhe, Germany.www.optres.fhg.de/ (accessed August 24, 2009).

Rathmann, M. 2007. Do Support Systems for RES-EReduce EU-ETS-Driven Electricity Prices? EnergyPolicy 35 (1): 342–49.

Reiche, D. 2002. Renewable Energies in the EUMember States in Comparison. In Handbook ofRenewable Energies in the European Union, edited by D.Reiche. Frankfurt/Main: Verlag Peter Lang, 13–24.

Rious, V., J.-M. Glachant, Y. Perez, and P. Dessante.2008.The Diversity of Design ofTSOs. Energy Policy36 (9): 3323–32.

RWI (Rhine-Westphalia Institute for EconomicResearch). 2009. Die ökonomische Wirkung derFörderung Erneuerbarer Energien: Erfahrungen ausDeutschland. (The Economic Impact of RenewableEnergy Support: Experiences from Germany).Essen, Germany: RWI.

Schöne, S. 2009. Auctions in the Electricity Market. Berlin:Springer-Verlag.

Sensfuß, F., M. Ragwitz, and M. Genoese. 2008. TheMerit-Order Effect: A Detailed Analysis of the PriceEffect of Renewable Electricity Generation on SpotMarket Prices in Germany. Energy Policy 36 (8):3086–94.

Sorrel, S., and J. Sijm. 2005. Carbon Trading in thePolicy Mix. Oxford Review of Economic Policy 19 (3):420–37.

Toke, D. 2005. Are Green Electricity Certificates theWay Forward for Renewable Energy? An Evaluationof the United Kingdom’s Renewables Obligation inthe Context of International Comparisons. Environ-ment and Planning C: Government and Policy 23 (3):361–74.

———. 2007. Renewable Financial Support Systemsand Cost-Effectiveness. Journal of Cleaner Production15: 280–87.

Toke, D., and V. Lauber. 2007. Anglo-Saxon and Ger-man Approaches to Neoliberalism and Environmen-tal Policy: The Case of Financing RenewableEnergy. Geoforum 38 (4): 677–87.

Traber, T., and C. Kemfert. 2009. Impacts of the Ger-man Support for Renewable Energy on ElectricityPrices, Emissions and Firms. Energy Journal 30 (3):155–78.

Vandezande, L., L. Meeus, R. Belmans, M. Saguan,and J.-M. Glachant. 2010. Well-Functioning Balan-cing Markets: A Prerequisite for Wind Power Inte-gration. Energy Policy, 38 (7): 3146–3154.

VIK (German Association of Industrial Energy Usersand Self-Generators). 2009. Die Entwicklung amEEX-Terminmarkt. www.vik-online.de/fileadmin/vik/EEX/Terminmarkt.pdf (accessed June 2, 2009).

Wand, R., and F. Leuthold. 2009. Feed-In Tariffs forPhotovoltaics: Learning by Doing in Germany?Innovations for a Low-Carbon Energy SystemWorking Papers WP-RD-03. Dresden, Germany:Dresden University of Technology.

Wawer, T. 2007. Konzepte für ein nationalesEngpassmanagement im deutschen Übertragungs-netz. (Concepts for a National Congestion Manage-ment Scheme in the German Transmission System).Zeitschrift für Energiewirtschaft 31 (2): 109–16.

Weber, Christoph. 2010. Adequate Intraday MarketDesign to Enable the Integration of Wind Energyinto the European Power Systems. Energy Policy38 (7): 3155–3163.

Weigt, H. 2009. Germany’s Wind Energy: The Poten-tial for Fossil Capacity Replacement and Cost Sav-ing. Applied Energy 86 (10): 1857–63.

Experience with Renewable Energy Policy in Germany 307

Weigt, H., T. Jeske, F. Leuthold, and C. vonHirschhausen. 2010. Take the Long Way Down:Integration of Large-Scale North Sea Wind UsingHVDC Transmission. Energy Policy 38 (7): 3164–3173.

Wischermann, S., and H.-J. Wagner. 2009. RenewableEnergy in Energy Scenarios for Germany. Paper pre-sented at 10th IAEE European Conference, Energy,Policies and Technologies for Sustainable Econo-mies. September 2009, Vienna, Austria.

Witzmann, R., and G. Kerber. 2007. Capacity of Dis-tribution Networks due to Power from Photovoltaic.EW: Das Magazin für die Energiewirtschaft 419 (4):50–54.

Wright, T. P. 1936. Factors Affecting the Cost ofAirplanes. Journal of the Aeronautical Sciences 3: 122–28.

Wüstenhagen, R., and M. Bilharz. 2006. Green EnergyMarket Development in Germany: Effective PublicPolicy and Emerging Customer Demand. EnergyPolicy 34 (13): 1681–96.

Zoellner, Jan, P. Schweizer-Ries, and C. Wemheuer.2008. Public Acceptance of Renewable Energies:Results from Case Studies in Germany. Energy Policy36 (11): 4136–41.

308 HannesWeigt and Florian Leuthold

15

Renewable Electricity Support:The Spanish ExperienceLuis Agosti and Jorge Padilla

As in many other countries, the policy of theSpanish government toward the electricity

sector has had three sometimes conflicting objec-tives: low prices, security of supply, and environ-mental sustainability. Over the last few years, thegovernment has promoted investment in renew-able energy, both wind and solar, with thoseobjectives in mind. The expansion in low-marginal-cost renewable capacity was expected topush wholesale prices (excluding subsidies)downward; reduce the traditional dependency ofthe Spanish economy on imported primaryenergy sources; and meet the government’s emis-sions reduction targets.

Investment in renewable generation has beensupported by means of a dual feed-in tariff (FIT)scheme.1 Renewable generators can choosewhether to sell their energy at a regulated fixedprice or to sell at the market price and receive anadditional premium. Investors in renewables alsobenefit from low administrative and regulatorybarriers and relatively favorable grid access condi-tions in Spain in comparison with other coun-tries.

The policy of the Spanish government hassome aspects to commend but can also be criti-cized on a number of fronts.The European Com-mission’s assessment of renewable energy supportschemes relies on two indicators relative to their

effectiveness and efficiency.The effectiveness indi-cator shows for each technology the increase inrenewable generation compared with its estimatedmidterm realizable potential.2The efficiency indi-cator compares the support received with thegeneration cost of the generation units covered bythe scheme. The scheme is considered more effi-cient when that difference is small.

The government’s support scheme has provedto be effective. In 2008, renewable capacityaccounted for 42% of all generation capacity and22% of total electricity production.Today Spain isthe second-largest wind power generator in theEuropean Union (EU). In November 2009, windgeneration reached record output levels, repre-senting 50% of total demand for several hours.Solar photovoltaic (PV) generation has experi-enced an extraordinary increase in recent years,from 146 megawatts (MW) installed capacity in2006 to 3,342 MW in 2008. Indeed, Spaininstalled more than 40% of the world’s total solarunits in 2008.

Regarding wind generation, the policy hasbeen both effective and efficient. Figure 15.1shows the effectiveness and efficiency—expectedprofit, in terms of the cost of renewable supportmeasured in euro cents per kilowatt-hour(kWh)—of the promotion schemes used by vari-ous EU member states in connection with wind

generation. The support scheme in Spain is themost effective, while the cost of the scheme is nogreater than the average of the member states.Thepicture for solar generation is somewhat different.The rapid growth in solar deployment describedabove is in part the result of an extremely gener-ous FIT. As the number of solar units in the mar-ket increased well above what was planned, theSpanish government found itself committed topay billions of euros over a 20-year period. Inevi-tably, this led to a sudden reduction of the feed-intariffs for this technology in 2009.

The current debate in Spain concerns theimpact of the current policy toward renewablegeneration on the future performance of thecountry’s electricity sector, in particular, onwholesale and retail electricity prices, systemoperation, and security of supply. Has the Spanishgovernment gone too far in its support of renew-able energy? Has it sacrificed the goals of eco-nomic efficiency and security of supply to theenvironmental objectives?

This chapter seeks to answer these questionsby describing the current status and future chal-lenges of the policy of support to renewable gen-eration in Spain. It also explores potential solu-tions to the problems identified. The discussion

begins with an examination of the current regula-tory framework for renewable energy, as well as adetailed description of existing renewable capacityand its expected development. The chapter thenlooks at the challenges affecting the future expan-sion of renewable capacity in Spain—in particular,the effects that large volumes of renewable energyhave had, and are expected to have, on the opera-tion of the Spanish electricity market, investmentincentives for different technologies, and retailprices. Finally, it gives possible solutions to theproblems posed by large-scale deployment ofrenewable generation.

Renewable Energyin Spain TodayRenewable energy production is part of theRégimen Especial (Special Regime, or SR),which includes all electricity produced fromrenewable sources, combined heat and power(CHP), and solid waste from units below 50 MW.Conventional units—nuclear, large hydro, coal,and combined cycle gas turbines (CCGTs)—operate under the Régimen Ordinario (Ordinary

–1 0 10%

5%

10% AT

DE

IE

ES-FP ES-MO

CZ-MO UK

ITBE-Wallonia

FR

CZ-FP

FI

SE

LTHU

15%

20%

25%

2 3 4 5Flanders

Expected profit [euro cents/kWh]

Effe

ctiv

enes

s in

dica

tor

Feed-in tariffs Quota/TGC Tax incentives/Investment grants

BE-PL

6 7 8

Source: European Commission 2008Notes: MO = market option; FP = fixed price option

Figure 15.1. Effectiveness and efficiency of support schemes for onshore wind generation in Europe in 2006

310 Luis Agosti and Jorge Padilla

Regime, or OR). In 2008, the SR was responsi-ble for 23% of electricity generated and repre-sented 32% of total installed capacity, as describedin Table 15.1.

Table 15.1. Generation mix in Spain, December 2008

Electricitygenerated

Installed capacity

Technology GWh % MW %

Hydro (> 50 MW) 21,428 7.5 16,657 18.3

Nuclear 58,973 20.6 7,716 8.5

Coal 46,275 16.1 11,359 12.5

Fuel/gas 2,378 0.8 4,418 4.9

CCGT 91,286 31.8 21,675 23.9

Ordinary Regime 220,340 76.9 61,825 68.0

Hydro (< 50 MW) 4,416 1.5 1,979 2.2

Wind 31,393 11.0 15,874 17.5

Other renewables 7,183 2.5 4,069 4.5

Nonrenewablesa 23,308 8.1 7,132 7.8

Special Regime 66,300 23.1 29,054 32.0

Total 286,640 100.0 90,879 100.0

Source: REE 2008aInclude CHP and solid waste

Figure 15.2 shows the volumes of renewablecapacity (excluding large hydro units) in severalEU countries, based on 2007 data. In absolute

terms, Spain ranked second in renewable capacity,after only Germany. In relative terms, renewablesrepresented 21% of total capacity in Spain, thesecond-highest proportion after Portugal.

Among the different technologies included inthe SR, wind generation has a predominant role.As shown in Table 15.2, onshore wind capacityhas grown from 886 MW in 1998 to 15,709 MWin 2008. Wind generation, together with CCGT,has dominated the evolution of the Spanish gen-eration mix in the last decade. These two tech-nologies together accounted for 94% of all cap-acity installed between 2003 and 2008. Solar PVrepresents only 3% of total capacity, although thepace of investment in the last two years has beenextraordinary.

As a result of the growth in installed renew-able capacity, output from the SR has experienceda substantial increase in recent years. As Figure15.3 shows, the combination of increased outputand a recession-induced drop in demand meantthat during 2009, SR generation representedmore than 30% of electricity demand—a figurethat rises to above 40% if the production fromlarge hydro units that operate in the OR wereadded in.

0

Portu

gal

Spain

Germ

any

Luxe

mbo

urg

Austri

a

Czech

Rep

ublic

Irelan

d

Hunga

ry

Greec

e

Belgium

Norway

Estonia

United

King

dom

Italy

Poland

Franc

e

5,000 24%

19%

21%

16% 13% 12% 11% 11% 7% 7% 6% 5%5% 5%

4% 2%

10,000

15,000MW

20,000

25,000

30,000

Onshore wind Small hydro

Percentage of total capacity

Photovoltaic Other

Source: CEER 2008Note: Figures exclude large hydro capacity

Figure 15.2. Renewables mix in EU countries, 2007

Renewable Electricity Support: The Spanish Experience 311

Policy Goals

The key goal of the renewables policy of theSpanish government is the reduction in CO2

emissions. In 2008, the production of electricityfrom renewable sources, excluding large hydrounits, saved more than 15 million tons (Mt) ofCO2,

3 which represents approximately 4% ofyearly emissions (MARM 2009). Current annualCO2 emissions represent a 52% increase withrespect to base-year emissions, or almost 37%

above the Spanish target under the Kyoto Proto-col. The policy of support to investment inrenewables is a natural response to the significantincrease in CO2 emissions.

Another fundamental goal is the reduction ofprimary energy dependence. Domestic primaryenergy production (including nuclear power) cur-rently accounts for just 19% of total primaryenergy supply in Spain. With hardly any naturalgas or oil production, the country relies onnuclear and coal production for the bulk of

0

1996

1997

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007

2008

2009

50,000

100,000

GW

h

150,000

200,000

250,000

300,000

0%

10%

5%

15%

20%

25%

30%

35%

Demand Share of SR

Source: CNE, 2009b

Figure 15.3. Evolution of demand and share of the SR

Table 15.2. Evolution of installed renewable and other SR capacity (MW)

Year Biomass Solar WindHydro(< 50MW)

Hydro(> 50MW)

CHPSolidwaste

Solid wastetreatment

1998 81 1 886 1,300 16,454 3,673 334 N/A

1999 88 2 1,686 1,439 16,526 4,206 351 29

2000 144 2 2,296 1,469 16,526 4,929 339 81

2001 230 4 3,508 1,562 16,588 5,352 449 157

2002 352 8 5,066 1,594 16,587 5,567 461 324

2003 454 12 6,324 1,666 16,658 5,628 468 420

2004 469 23 8,522 1,708 16,657 5,694 585 469

2005 499 48 10,097 1,765 16,657 5,706 585 538

2006 540 146 11,891 1,893 16,657 5,836 579 624

2007 557 695 14,423 1,913 16,657 6,060 569 527

2008 582 3,342 15,709 1,965 16,657 6,168 579 554

Source: CNE 2009bNote: SR includes renewable, CHP, and solid waste and excludes large hydro generation

312 Luis Agosti and Jorge Padilla

domestic energy sources, both under pressure tobe reduced in coming years.4The current govern-ment has clearly stated its intention to abandonnuclear power in favor of renewable energy. In thecase of coal, mining activity is for the most partheavily subsidized and likely to disappear in themedium term.5 Renewables will then be the solesources of indigenous energy. Only if renewablegeneration grows at above 8% per year from 2008onward will Spain be able to maintain its energydependence below the current level of 80%.Without the contribution of renewables, finalenergy dependence would be above 90% by2020.6

A third goal of the renewables policy in Spainconcerns electricity prices. The expansion inrenewable generation, which is characterized bylow variable costs and intermittency, was expectedto exert downward pressure on wholesale elec-tricity prices, at least in the short and mediumterms. Provided that additional capacity wasowned by nonincumbent generators, it wouldcontribute to address the government’s repeatedlystated concerns about the use and abuse of marketpower in the electricity generation market. Theeffect of this expansion in capacity on marketprices was also seen as a way to partly offset thecost of the subsidies to renewable generation.

Renewable Regulation

The legal framework of the SR has evolved sub-stantially over the last 15 years. The first majorpush for the promotion of renewable energy camewith the enactment of the Spanish Electricity Lawin 1997 and Royal Decree (RD) 2818/1998,which established a reward scheme for each tech-nology, consisting of a fixed premium over themonthly average market price.

In 1999, the government approved the Plande Fomento de las Energías Renovables (Plan toPromote Renewable Energy, or PFER), settinginvestment objectives for each technology withthe aim of meeting 12% of primary energy con-sumption in 2010 from renewable sources(including large-scale hydro). It also approved sev-eral measures to encourage market participationby renewables. Renewables units were allowed tosign bilateral contracts with retail companies andalso became eligible to receive a capacity paymentfor participating in the market. This was removedin 2007.

In 2004, this regime was modified by RD436/2004, which allowed renewable generatorsto choose either to sell their output to a distribu-tion company at a fixed feed-in tariff or to go tothe market and receive the market price plus apremium, and in some cases a further incentive.7

0

2004 2005 2006 2007 2008

10,000

81%

48%26% 27% 27%

19%52%

74%73%

73%

20,000

GW

h

30,000

40,000

50,000

60,000

70,000

80,000

Sales at fixed price Sales at market option

Source: CNE 2009b

Figure 15.4. Distribution of SR production between fixed price and market options

Renewable Electricity Support: The Spanish Experience 313

The possibility of choosing between these twooptions was in itself a significant advantage com-pared with RD 2818/1998. In addition, RD436/2004 implied an increase in terms of absolutecompensation, quite large in some cases. As aresult, after its enactment, most of the SR produc-tion switched from the fixed feed-in tariff to themarket option, as shown in Figure 15.4.

A 2005 review of the PFER showed that theobserved growth in renewable energy was insuffi-cient to meet the goals defined under the plan. Asa consequence, the PFER was updated in 2005with the Plan de Energías Renovables (Renew-able Energy Plan, or PER) 2005–2010. Accordingto that plan, by 2010 at least 12% of primaryenergy consumption and 29.4% of the total gen-eration of electricity should come from renewablesources. The PER 2005–2010 argued the need toreinforce the government’s support of renewableenergies through various measures, includingstronger economic incentives.

This gave rise to RD 661/2007, which pro-vides the current framework for the promotion ofrenewables. The economic regime is similar tothat under RD 436/2004. Each unit may choosebetween the fixed tariff and the market option,

and a generation plant operating within the SRcan change from one alternative to the other if itstays with the chosen alternative for at least oneyear. The level of subsidies is somewhat higherthan provided by RD 436/2004.

Table 15.3 shows the evolution of the premi-ums and fixed prices under the different royaldecrees. As can be seen, RD 436/2004 alreadyrepresented a large increase in the value of theincentives, especially in the case of solar PV andsolar thermal generation, which were subse-quently increased with RD 661/2007. In order tolimit the potential cost of the support schemes,RD 661/2007 introduced a cap-and-floormechanism for the market option premium ofsome technologies, with the intention of bothreducing the risk for investors and limiting themaximum cost of subsidizing renewables.RD 661/2007 also established that once 85% ofthe 2010 capacity targets for each technology hadbeen achieved, the remuneration scheme foradditional capacity would have to be redefined. Inthe case of solar PV, the 2010 goal was exceededin 2008. Installed capacity in 2008 equaled 3,342MW, compared with a 2010 target of 371 MW.

Table 15.3. Evolution of incentives for renewable generation

RES technology RD 2818/1998 RD 436/2004 RD 661/2007

1998 2004 2007

Premium FIT Premium FIT Premium

(€/MWh) (€/MWh) (€/MWh) (€/MWh) (€/MWh)

PV < 100 kWp 60.0 414.4 N/A 440.0 N/A

PV > 100 kWp 30.0 216.2 187.4 [229–417] N/A

Solar thermal electricity 20.0 216.2 187.4 269.0 254.0

Wind < 5 MW 31.6 64.9 36.0 73.0 29.0

Wind > 5 MW 31.6 64.9 36.0 73.0 29.0

Geothermal < 50 MW 32.8 64.9 36.0 69.0 38.0

Mini hydro < 10 MW 32.8 64.9 36.0 78.0 25.0

Hydro 10–25 MW [0–35.8] 64.9 36.0 [66–78] 13.0

Hydro 25–50 MW [0–35.8] 57.7 28.8 [66–78] 13.0

Biomass (biocrops, biogas) 28.0 64.9 36.0 [107–158] [61–115]

Agriculture residues 28.0 57.7 28.8 [96–130] [30–97]

Municipal solid waste 22.0 50.5 21.6 [38–53] [30–97]

Notes: For some technologies, the value of the subsidies varies with the size of the installation; brackets indicate the minimum andmaximum value of the subsidy; €1 = $1.36 as of this writing

314 Luis Agosti and Jorge Padilla

The government therefore established a newremuneration scheme for solar PV units in RD1578/2008.

In 2009, the renewable sector reduced itsexpansion significantly as a result of various fac-tors, including new administrative barriers,uncertainty about the economic framework thatwill apply once the targets for 2010 are accom-plished, and the economic recession. Investmentin wind and solar developments has almost ceased.

There is little doubt, however, that the sectorwill resume its expansion in the future. Spain isobliged to implement European Directive2009/28/EC on the promotion of the use ofenergy from renewable sources,8 which requires itto produce 20% of final energy consumption fromrenewable sources by 2020. Moreover, the Span-ish government has made the promotion ofrenewables a national priority. It is currently pre-paring a new Law on Energy Efficiency andRenewable Energies, which will translate intoSpanish law the EU regulations on energy andclimate change, and a new 2011–2020 Plan onRenewable Energies, which will establish indi-vidual targets for each type of technology and the

corresponding measures that will have to beadopted in order to attain those targets.

Data for 2008 show that Spain is well posi-tioned to meet the 2020 challenges resulting fromthe European directive in electricity, but not so inheating and transport. In 2008, the share of finalenergy produced by renewables was approxi-mately 9%, with 5% corresponding to the elec-tricity sector and 4% to the heating and transportsectors (MITYC 2009a). According to the Span-ish government, the 2020 scenario would require40% of electricity to be produced from renewablesources, which corresponds to 9% of total energyconsumption.9 The remaining 11% necessary tomeet the 20% target would have to be achieved bythe heating and transport sectors. These estimatesinvolve rapid annual growth rates of 6% for elec-tricity and 11% for heating and transport, asshown in Figure 15.5.

Assessing Current PolicyHas the support scheme implemented by theSpanish government successfully contributed to

Conventional Renewable electricity

Renewable heating and transport

50,0002008 2020

60,000

70,000

80,00091%

5%4%

90,000kteo

100,000

CAGR 11%

CAGR 6%

CAGR 0.6%

80%

9%

11%

110,000

120,000

130,000

140,000

Source: Author’s estimates, assuming 40% of electricity coming from renewables in 2020Notes: Final energy consumption growth 1.8% per annum as estimated for Spain during the period 2007–2016 in IEA 2009; ktoe = kilotons of oil equivalent; CAGR = compound annual growth rate

Figure 15.5. Renewable energy growth till 2020 and contribution to final energy consumption

Renewable Electricity Support: The Spanish Experience 315

its three main policy goals of environmentalsustainability, lower electricity costs, and securityof supply? As explained in detail in the followingdiscussion, the current support scheme has provedeffective but quite expensive, is likely to havecaused an increase in electricity prices, and raisesseveral concerns in relation to the operation of thesystem and its reliability, especially in the longterm.

Environmental Sustainability:Effectiveness and Efficiency

As noted above, the Spanish renewable electricitysupport scheme has proved to be highly effective.Its overall 2010 target is within reach, even ifsome of the subtargets for individual technologiesdefined in PER 2005–2010 and RD 661/2007will likely be missed, as shown in Table 15.4. Theincrease in renewable capacity has significantlycontributed to reducing the high level of CO2

emissions.The effectiveness of the scheme has come at

considerable cost, however. As shown in Table15.5, in 2008 the average subsidy received by SRproduction units per MWh produced was 75% ofthe average market price. Subsidies to the SR areexpected to reach €4.7 billion ($6.4 billion) by theend of 2009 (CNE 2009c). Notably, 40% of thatamount will go to solar PV units, which representonly 3.7% of total SR generation.

Table 15.4. Capacity targets and currentdeployment status, September 2009

Septem-ber 2009(MW)

Target for2010(MW)

% accom-plished

CHP 6,464 9,215 70%

Solar PV 4,824 371 1,300%

Solar thermal-electric 82 500 16%

Wind 17,269 20,155 86%

Wind repoweringa 0 2 0%

Hydro (< 10 MW) 1,414 2.4 59%

Biomass 674 1,567 43%

Residuals 269 350 77%

Total 30,996 36,558 85%

Source: CNE 2009baRepowering involves replacing first-generation wind turbineswith modern multimegawatt wind turbines

Lower Electricity Costs

In 2009, a combination of factors—increasedrenewable capacity, reduced demand, and highlevels of available hydro production—led to arapid increase in the share of energy produced bythe SR, which represented 28% of total genera-tion in January–October 2009. The increased roleof renewable energy has had major implicationsfor the performance of the Spanish electricity sys-tem. In particular, the surge in renewable genera-

Table 15.5. Implicit subsidies for the SR, 2008

Capacity(MW)

Energy(GWh)

Total pricereceived(€/MWh)

Market price(€/MWh)

Implicit pre-mium(€/MWh)

Premium overmarket price

CHP 6,136 21,088 99.2 64.5 34.7 54%

Solar 3,454 2,541 453.0 64.5 388.6 603%

Wind 15,982 31,869 100.4 64.5 36.0 56%

Small hydro 1,945 4,632 96.2 64.5 31.8 49%

Biomass 583 2,487 116.5 64.5 52.0 81%

Solid waste 569 2,732 87.8 64.5 23.3 36%

Solid waste treatment 573 3,156 111.4 64.5 47.0 73%

Total SR 29,242 68,504 113.4 64.5 49.0 76%

Source: CNE 2009bNote: €1 = $1.36 as of this writing

316 Luis Agosti and Jorge Padilla

tion has had an impact on the electricity pricesfaced by end users. Measuring that impact is acomplex exercise, however, because the increasein renewable generation produces several effectsthat may operate in opposite directions.

First, consumers of electricity pay the highsubsidies received by renewable generators viahigher regulated network access tariffs. These tar-iffs are set by the government periodically and arepaid by all consumers that have access to the elec-tricity grid. They are set taking into account thecosts of the transmission and distribution grids,other costs of the system, and the subsidies paid tothe SR.10 The weight of these subsidies in thecalculation of the access tariff is significant. Notsurprisingly, the policy of subsidizing investmentin renewable generation has led to increases in theaccess tariffs paid by the average consumer. Thosetariffs have increased from €15.70 ($21.29) perMWh in 2008 to €33.30 ($45.16) per MWh in2009—an increase of 112%, half of which iscaused by the renewable electricity supportscheme.11 As these charges account for around50% of final prices paid by end consumers, theincrease in the access tariffs has had a significantimpact on the overall cost of electricity and henceon consumer welfare (CNE 2009a).

Second, the increase in renewable generationcan reduce wholesale market prices in the shortterm. In Spain, most of the production of all ORand SR units is sold in a day-ahead market, whichis a pool-type market in which all generatorsreceive the marginal prices of hourly auctions,provided they submit supply offers at prices belowthe clearing price of the market. Conventionalunits are expected to submit bids based on theirvariable costs. Renewable SR units, whether theyhave opted for the FIT or the market price pluspremium option, bid in all their production atzero, as in most cases this is close to their marginalcost, and because they need to ensure that all theirproduction is sold in the market. As a result, alloutput from the SR, including the output ofrenewable units, is cleared, reducing the residualdemand faced by the OR units. An increase in theoutput supplied by the SR reduces the demandfaced by the only price-setting units in the day-ahead market—those within the OR. As a result,

the increase in SR supply results in lower whole-sale prices (Sáenz de Miera et al. 2008).

Note, however, that any reduction in spotprices would reduce the remuneration received byexisting production units. The adoption of apolicy aimed at promoting renewable generationcauses an adverse shock to the actual and futureprofitability of those units which may render theinvestment made on them unjustified. To theextent that that policy, or at least some of its cen-tral elements, could not have been anticipated bymarket participants at the time their investmentdecisions were made, the adverse impact of thepolicy on the profitability of incumbent plantsmay be seen as a form of regulatory taking whichshould be compensated. Note, in addition, thatthe exit from the market of inadequately remu-nerated plants would bring prices up and under-mine the price goal of renewable promotion. So itappears that an aggressive renewables policy mayhave to go hand in hand with some form of com-pensatory measures aimed at compensatingincumbent players for unamortised investmentsmade before the new regulations were adopted.Those compensations may be defended on bothfairness and efficiency grounds.

The net effect of an increase in renewablecapacity on end-user prices is therefore ambigu-ous. As illustrated in Figure 15.6, the net effectcould be higher or lower end-user prices,depending on the slope of the supply curve in theday-ahead market and the magnitude of theabove-market premium paid to incentivize invest-ment in renewable capacity.

Figure 15.6 depicts two styled examples. Inboth Case 1 and Case 2, the production of theSR, QSR, reduces the demand faced by OR gen-erators from QT to QOR, which causes a reduc-tion in the wholesale market-clearing price fromPT to POR. As OR units are remunerated at themarket-clearing price, the reduction in that priceresults in a reduction in the wholesale cost of elec-tricity given by area A. On the other hand, the SRunits are remunerated at a price equal to PRE,irrespective of the market-clearing price. A lowermarket-clearing price simply translates into agreater premium for the units operating under theSR. Hence an increase in the production of the

Renewable Electricity Support: The Spanish Experience 317

SR results in extra payments to the SR equal toarea B. The net result of the increased productionof the SR will be a reduction (increase) in thewholesale cost of electricity when area A is larger(smaller) than area B. In Case 1 the cost of therenewables subsidy is greater than the savingsderived from the reduction in the wholesalemarket-clearing price. In Case 2, on the contrary,the cost of the subsidy is more than offset by thereduction in the wholesale market-clearing price.Case 1 is characterized by a relatively flat supplycurve, whereas Case 2 involves a relatively steepercurve.

The net effect of the policy of supportinginvestment in renewables via FITs in Spain islikely to have increased the cost of electricity, andhence the price paid by end users. This is becauseat present, the supply curve in the Spanish day-ahead market is quite flat at the margin, as a resultof the large number of CCGTs competing in thesame range of costs. Thus the impact of theincrease in renewable capacity on the day-aheadmarket price can be expected to be fairlysmall—in any event, smaller than the overall costof the subsidies paid to the owners of renewableelectricity assets. Note in this respect that duringthe first half of 2009, the day-ahead market priceranged between €35 and €40 ($47 and $54) perMWh, which, given that the floor value of theFIT for wind production is €71.30 ($96.70) perMWh, implies a very large premium. Moreover,the intermittent nature of wind and solar technol-ogy imposes additional pressure on electricityprices. The production of wind plants is highly

volatile and requires thermal capacity to be avail-able for backup, which increases requirements forspinning reserves,12 and these have to be appro-priately remunerated.

The Spanish renewable electricity supportpolicy is likely to lead to additional increases inelectricity prices in the medium term and evenmore so in the long term. In the medium term,unless a significant reform of the FITs occurs, theoverall cost of the subsidies to the SR will increasein direct response to the increase in renewablecapacity. Moreover, the impact of further addi-tions of renewable capacity on day-ahead prices islikely to decrease as installed capacity goes up andmarket prices approach the competitive bench-mark.

In the long term, the increase in renewablecapacity will likely crowd out some thermalcapacity, causing prices to increase. Note in thisrespect that the expansion of renewable capacityhas already resulted in a significant reduction inthe production of conventional thermal units.From November 2008 to October 2009, outputfrom CCGTs and coal plants was 17% lower thanin the previous 12 months. Some of this reductionis the result of a fall in electricity demand, butsome of it (no less than 7%) is the direct conse-quence of the increase in SR production.13 Thereduction in the production of conventional ther-mal units has led to lower utilization rates, whichhave fallen from 53% in 2008 to around 44% in2009.

DT

Case 1

PT

Net subsidy

Reduction in themarket price POR

PRE

QSR P

DOR

QOR QT

A

DT

Case 2

PT

Net subsidy

Reduction in themarket price POR

PRE

QSR P

DOR

QOR QT

BB

A

Figure 15.6. Effect of renewable energy promotion in the wholesale generation market

318 Luis Agosti and Jorge Padilla

Security of Supply

The significant increase in renewable capacity hasimplications for the operation of the system in theshort term as well as for its performance in themedium and long term.

As is well known, the integration of a largevolume of intermittent generation into an elec-tricity system poses several operational challenges.Red Eléctrica de España (REE), the Spanish sys-tem operator, has made considerable efforts tointegrate the production of wind and solar unitsefficiently and reliably. For instance, in 2006,REE created the Centre for the Control of theSpecial Regime (CECRE), with the specific mis-sion of maximizing renewable generation, espe-cially wind generation, while ensuring systemsecurity.

CECRE’s mission is made difficult by thevolatile nature of the production of wind (and alsosolar) plants. According to REE, forecasting theproduction of wind plants 24 hours ahead typi-cally involves average forecasting errors of about20% (Rodríguez 2007). These errors require con-siderable additional reserves. When forecast pro-duction is above actual production, the systemrequires upward reserves. When, instead, forecastproduction is below actual production, the properoperation of the system demands downwardreserves.

Managing wind plants in Spain is particularlycumbersome, as their production is negatively

correlated with electricity demand: production islow during the summer and winter peaks, but itresults in energy “spills” during off-peak hours.14

For example, as shown in Figure 15.7, onNovember 2, 2008, wind production was 3,200MW above forecast. With a low demand, 20,000MW, the system ran out of downward reservesvery rapidly. To balance the system, the REE hadto reduce wind production by 2,400 MWbetween 7:00 and 9:00 a.m.

Another common problem occurs when largenumbers of wind turbines need to be shut downbecause of high wind speeds. For example, onJanuary 23–24, wind turbines in the north ofSpain experienced wind speeds above 200kilometers per hour and had to be disconnected.As a result, the difference between forecast andactual production was above 5,000 MW, as shownin Figure 15.8. The system operator had to coverthis difference by quickly switching production toCCGTs.

In summary, wind, and to some extent alsosolar, generation has introduced significant uncer-tainty in the operation of the system. As a result,REE has had to increase the reserve margin so asto be able to respond in a timely and efficientmanner to potential variations in renewables pro-duction. In particular, REE has relied on a com-bination of gas-fired and hydroelectric generationto provide the necessary reserves.

The reliability of the system on a forward-looking basis will hinge on a number of factors.

Wind generation Forecast

0

0:00

1:00

2:00

3:00

4:00

5:00

6:00

7:00

8:00

9:00

10:0

011

:0012

:0013

:0014

:0015

:0016

:0017

:0018

:0019

:0020

:0021

:0022

:0023

:00

1,000

MW

h

2,0003,0004,0005,0006,0007,0008,000

Source: Data obtained from REE database.

Figure 15.7. Actual versus forecast generation from wind, November 2, 2008

Renewable Electricity Support: The Spanish Experience 319

The size and availability of the CCGT fleet isperhaps the most important of all. This is a reasonfor concern, however, because the profitability ofCCGT plants fell dramatically in 2009, and theirprospects in the near future are not particularlybright either. As explained above, the increase inrenewable capacity has led to a reduction in theutilization rates of the CCGTs, as well as a drop inthe prices received when they operate.The reduc-tion in the number of hours of operation has alsocaused unit variable costs to rise as a result of,among other things, the increased maintenancecosts resulting from more variable operating pat-terns and the “take or pay” obligations that forceCCGT operators to pay for gas volumes they pur-chased under long-term contracts regardless ofwhether those volumes are effectively consumed.The forecast evolution of renewable capacity,which could be large enough to cover almost100% of the demand sometimes, suggests that theutilization rates and profitability of the CCGTscurrently in operation will remain low in the nearfuture.

The implication of this is that the availabilityof backup thermal capacity may be at risk.Incumbent CCGT operators will be tempted toclose down or “mothball” some of their existingcapacity. Moreover, they (as well as other potentialentrants) will no longer have the incentive tointroduce additional thermal capacity into the sys-tem. The evidence shows that this is indeed the

case: several investment plans have been discon-tinued, which means that CCGT capacity in thefuture will fall short of what was expected notlong ago. Tighter supply conditions may lead tohigher spot prices at times when wind output islow, and that may allow the conventional unitsthat remain in the market to operate profitably.Nevertheless, what constitutes a rational responsefrom a private point of view could have disastrouseffects from the viewpoint of the system’s securityof supply.

Other factors that may also contribute to inte-grating renewable generation efficiently includethe ability of the system to manage demand fluc-tuations, the availability of hydro pump unitswhose operation can be optimized to provideadditional energy storage, the development ofinterconnections with neighboring countries, andthe existence of a transportation grid that can suc-cessfully accommodate the deployment of renew-able units in areas that are not necessarily locatedwhere the demand is. Unfortunately, Spain is notterribly well placed with respect to any of thesefactors, which limits the ability of REE to avoidenergy spills and minimize reserve requirements.For example, the interconnections with theneighboring countries, especially with France,represent a negligible proportion of peak demand,and the transmission grid is designed to reflect thegeneration fleet in place prior to the deploymentof wind generation units.

0

10:0

011

:00

12:0

013

:00

14:0

015

:00

16:0

017

:00

18:0

019

:00

20:0

021

:00

22:0

023

:00

0:00

1:00

2:00

3:00

4:00

5:00

6:00

7:00

8:00

9:00

2,000

MW

h

4,000

6,000

8,000

10,000

12,000

14,000

23 24

+5,000 MW

Wind production Forecast

Source: Data obtained from REE data base.

Figure 15.8. Actual versus forecast generation from wind, January 23–24, 2008

320 Luis Agosti and Jorge Padilla

Policy RecommendationsThe renewable electricity support schemeadopted by the Spanish government no doubt hascontributed positively to achieve its environmen-tal objectives, but it has done so at a high cost interms of both the efficiency and reliability of theSpanish electricity system. From an efficiencystandpoint, as explained above, the drastic expan-sion of renewable capacity in Spain is likely tohave raised the cost of electricity in the shortterm, and its medium- and long-term effects onprices are bound to be adverse from the viewpointof end users. Regarding security of supply, thepolicy of the Spanish government has contributedto reducing the dependency of the system onimported energy sources. On the other hand, anelectricity system where a significant proportionof electricity demand is covered by wind and solargeneration units can work reliably only if there issufficient thermal capacity in reserve. This in turnrequires remunerating that backup capacityappropriately in a market context where, preciselybecause of the expansion of renewable generation,the output of those thermal units and the pricereceived by them are bound to fall.

The Spanish government has no room torenegotiate its environmental commitments.What it can do, however, is rethink its currentpolicy to ensure that those commitments aremade compatible with the efficient and reliableoperation of the system in the short, medium, andlong term. Otherwise, the environmental policyof the government may come into question.

What should be done, then? First, and mostimportant, the Spanish government should care-fully consider the extent and cost of its supportscheme. It should investigate whether its environ-mental objectives can be more cheaply achievedby means of policies aimed at promoting energyefficiency. It should revisit its forecasts regardingfuture electricity demand and generation capacityin order to rigorously assess how much support tooffer the deployment of renewable generationcapacity, taking into account its potentiallyadverse effects on economic efficiency and secu-rity of supply, as discussed above.

The available evidence suggests that this isprecisely what the government is starting to do.The Spanish government is taking steps to limitthe costs of its support policy, even when that mayimply curtailing the expansion of renewable gen-eration. Thus, for example, Royal Decree-Law(RDL) 6/2009 imposes additional administrativeburdens on the installation of renewable capacity.Only those production units with the necessaryauthorizations and credible construction planswill be eligible to register into the SR. In prom-ulgating this RDL, the government explicitlyacknowledged that if the expansion of renewablecapacity were to follow the observed trend, theeconomic and technical viability of the wholeelectricity system could be seriously jeopardized.The government has now imposed yearly caps onthe expansion of certain technologies (Secretaríade Estado de Energía 2009). Wind generationcapacity will not be able to increase by more than1,855 MW in 2010 and 1,700 MW in 2011 and2012. Solar PV capacity expansions will becapped at 500 MW until 2012.The premium paidto solar PV units has also been significantlyreduced, from €442 to €320 ($599 to $434) perMWh.

These measures may nonetheless prove insuf-ficient. It thus makes sense for the Spanish gov-ernment to look for alternative ways to achieve itsenvironmental objectives at lower cost. One obvi-ous recommendation is to proceed further inreducing support for solar PV, which despite thereduction noted above remains extraordinarilyhigh and is hard to justify on environmental orindustrial policy grounds. Some of the savingscould be invested in research and development(R&D) into solar energy. Given the current stateof knowledge regarding solar energy, investing inR&D is likely to prove more effective and effi-cient than investing in installations that can soonprove obsolete.

Second, the Spanish government should alsoassess whether the overall cost of meeting its 2020targets can be lowered by placing increasedemphasis on renewable energy in the heating andtransport sectors, such as through the promotionof biofuels, and also on the promotion of energyefficiency. The current situation of renewables in

Renewable Electricity Support: The Spanish Experience 321

these sectors suggests great potential for improve-ment, especially in the transport sector, whererenewable energy represents less than 2% of totalconsumption (as compared with the 10% targetset for 2020).

Some estimates indicate that biomass and solarthermal heating are much more economic meansof reducing emissions than other technologiessuch as wind or hydro generation (IEA 2009, 96).Policies aimed at increasing energy efficiencyshould also be a priority (European Commission2009). The current strategy of achieving signifi-cant emissions reductions by focusing almostexclusively on the supply side of the electricitysystem is too costly. A multipronged approachinvolving demand-side interventions in the elec-tricity markets and actions in heating and trans-port may prove more effective and efficient.

Third, once the government has decided howmuch renewable generation to promote, it shouldconsider reforming the generation market so as toensure that the increase in renewable capacitydoes not distort the operation of the system orcreate security of supply concerns. In this respect,the government should assess how to reward theavailability of backup capacity in the form ofCCGTs and open cycle gas turbines.The govern-ment should perhaps increase the capacity pay-ments made to CCGTs and other flexible backupunits. Currently, the value of the capacity pay-ments is fixed at €20,000 ($27,126) per MW peryear for the existing conventional units and lim-ited to 10 years. In the case of capacity paymentsfor future investments, the value of the capacitypayment is calculated according to the value of anindex that measures the excess of capacity of thesystem. This capacity payment has a cap of€28,000 ($37,976) per MW per year and isawarded for only 10 years as well. These figuresmay be insufficient to remunerate generationunits that will run only at peak or as backupcapacity. A more definitive conclusion on theappropriate level of payments would require fur-ther research.

Also, the government could introduce a mar-ket for short-term backup capacity, which cansend the appropriate signals to investors consider-ing whether to enter the market to cover the flex-

ibility needs of a system with significant amountsof intermittent capacity.15 Or if none of thisworks, it could consider raising or removing thecap on the electricity price of the day-ahead mar-ket, currently set at €180 ($244) per MWh. Thisvery low value, which is below any estimates ofthe value of lost load, clearly risks limiting theability of peaking units to earn their cost of capi-tal. If it fails to undertake any of these measures,the Spanish electricity system will face a consider-able risk of underinvestment in thermal units,particularly combined cycle and peak gas tur-bines, which could eventually undermine its abil-ity to reliably meet electricity demand.

Fourth, the government should promoteinvestment in the development of new grid infra-structures, subject to appropriate cost–benefitanalyses, so that the generation of wind and solarunits can be transported and distributed more effi-ciently to those areas of the country wheredemand is located. According to REE, its invest-ment plans would make it technically possible forthe Spanish grid to connect and serve more than40,000 MW of wind power and more than 8,000MW of solar thermoelectric capacity by 2016, incomparison with 15,800 MW of wind power and82 MW of solar power in 2008.

Fifth, the government should also revisit thequestion of how to pay for the support offered tothe SR. Who should contribute to fund the gov-ernment’s environmental policy? Should the costof that policy be borne exclusively by end users ofelectricity in proportion to their usage?This is nota for-gone conclusion. For one thing, the benefitsof that policy are spread over the entire popula-tion. For another, by distributing the burden in away that is unrelated to usage, the governmentwould be minimizing the impact of its environ-mental policy on the competiveness of energy-intensive manufacturing and services industries.

ConclusionsThe Spanish government has made investment inrenewable electricity generation a national prior-ity and believes it will be one of the key drivers ofeconomic development in the coming years.16 Its

322 Luis Agosti and Jorge Padilla

policy has no doubt been effective. Spain is one ofthe countries in the world with more wind andsolar PV electricity generation units in absoluteand relative terms. Expanding renewable capacityhas also been fairly costly and has raised concernsregarding the efficiency and reliability of theSpanish electricity system. This chapter has iden-tified and explained the related problems and pro-posed how some of them might be addressed suc-cessfully. The Spanish government shouldproactively tackle the problems generated by itsrenewable electricity support power as a matter ofurgency. The country’s economy could be seri-ously damaged if its electricity power system wereunable to meet its needs efficiently and reliably.The environmental objectives of the governmentcannot and should not be achieved at the expenseof the competitiveness of the Spanish industry.

Notes

1. FIT schemes are widely regarded as particularlyeffective. See, e.g., CEER (2008); IEA (2008,2009).

2. The realizable potential represents the maximumachievable potential, assuming that all existing bar-riers can be overcome. The assessment of the real-izable midterm potential of renewable technolo-gies for European countries up to 2020 was car-ried out using the Green-X model. See Ragwitzet al. (2005).

3. Relative to a counterfactual scenario where thesame levels of generation were provided byCCGTs.

4. According to the IEA methodology, the primaryenergy form for nuclear energy is not the heatingvalue of the nuclear fuel used, as this is difficult toestablish unambiguously. Instead, the heat contentof the steam leaving the reactor for the turbine isused as the primary energy form, and it is there-fore considered as a domestic source of energy.

5. The plan for the support of domestic coal requiresthe reduction of coal production to 9.2 Mt by2012. See European Commission (2006, 2009).

6. This scenario assumes that primary energy growsat an annual rate of 1.3%, as assumed in the IEAenergy balances for Spain, and that outputs fromnuclear and domestic coal remain at 2008 levels.

7. Under the fixed feed-in tariff option, generatorssell their output at a fixed price regardless of theprice in the wholesale electricity market; underthe premium option, they receive a price equal tothe market price plus a premium.The values of thefixed feed-in tariffs, premiums, and incentiveswere fixed in relation to the tarifa media dereferencia (TMR; average reference tariff), whichmeasures the average cost of electricity in a givenyear.

8. Directive 2009/28/EC of the European Parlia-ment and of the Council of 23 April 2009 on thepromotion of the use of energy from renewablesources and amending and subsequently repealingDirectives 2001/77/EC and 2003/30/EC.

9. The Ministry of Industry, Tourism and Trade hassuggested on several occasions that the 2020objectives require 40% of electricity to haverenewable origin by 2020. See, e.g., MITYC(2009b).

10. These include inter alia the costs of the systemoperator (REE), the market operator (OMEL),and the electricity regulator (CNE).

11. Average access tariffs, excluding the ex ante defi-cit, as described in CNE (2009a).

12. On average, wind generation in Spain requiresover 600 MW of additional and expensivereserves.

13. This assumes that 100% of the demand reductionis compensated by a reduction in CCGT produc-tion.This is a conservative assumption, as the ordi-nary regime production, excluding coal andCCGTs, has also decreased in comparison with2008.

14. According to REE, if the current trend continues,energy spills will become a constant in the Spanishsystem. The report estimates that from 2014onward, energy spills will reach between 0.4 and2.3 TWh for 3% to 10% of hours in a year, orbetween 0.6 and 3.6 TWh from 5% to 12% ofhours, depending on the scenario. See Secretaríade Estado de Energía (2009).

15. Currently, no such market exists. The tertiaryreserves market comprises the reserves required toincrease or decrease production within a timeframe of 15 minutes for a period above 2 hours.Additionally, this service rewards only productionand not capacity. With the modification of thecapacity payments in 2007 (see Orden [Order]ITC/2794/2007, Annex III), the system operatorwas expected to develop a short-term capacitymarket or availability service, which was intended

Renewable Electricity Support: The Spanish Experience 323

to incentivize thermal capacity availability duringpeak hours by having the the system operator andgenerators sign bilateral contracts. However, thismarket has not been developed yet, nor is itexpected to be developed in the short term.

16. A number of agents and institutions have analyzedthis belief and reached different conclusions. Forexample, the Spanish renewable energy associa-tion, APPA, concludes that the contribution ofrenewables to economic growth is positive interms of employment and economic growth(APPA 2009). On the contrary, the Instituto Juande Mariana finds that the renewable policy inSpain destroys 2.2 jobs for each green job created(Calzada Álvarez 2009). However, it does notappear that any rigorous economic analysis hasbeen done of the issues.

ReferencesAPPA (Association of Producers of Renewable

Energy). 2009. Estudio del ImpactoMacroeconómico de Las Energías Renovables en laEconomía Española. (Macroeconomic Impact ofRenewable Energy in the Spanish Economy).www.appa.es/descargas/NP_APPA_Estudio_Impacto_ER_Espana.pdf(accessed February 22, 2010).

Calzada Álvarez, Gabriel. 2009. Study of the Effects onEmployment of Public Aid to Renewable EnergySources. www.juandemariana.org/pdf/090327-employment-public-aid-renewable.pdf (accessedFebruary 22, 2010).

CEER (Council of European Energy Regulators).2008. Status Review of Renewable and Energy Effi-ciency Support Schemes in the EU. Brussels:CEER.

CNE (National Energy Commission). 2009a. BoletínMensual de Indicadores Eléctricos y Económicos, Octubre2009. (Monthly Bulletin on Economic and ElectricIndicators, October 2009) www.cne.es/cne/doc/publicaciones/iap_indicadores-oct09.pdf (accessedFebruary 22, 2010).

———. 2009b. Información Estadística sobre las Ventas deEnergía del Régimen Especial. (Statistical Informationon Special Regime Energy Sales.) www.cne.es/cne/Publicaciones?id_nodo=143&accion=1&soloUltimo=si&sIdCat=10&keyword=&auditoria=F (accessedFebruary 22, 2010).

———. 2009c. Informe 19/2009 de la CNE sobre laPropuesta de Orden Ministerial por la que se Revisan las

Tarifas de Acceso Eléctricas a Partir del Día 1 de Julio de2009. (Report 19/2009 from the CNE on the Pro-posal about the Ministerial Order to Review Elec-tricity Access Tariffs from July 2009 Onwards.)www.cne.es/cne/doc/publicaciones/cne86_09.pdf(accessed February 2, 2010).

European Commission. 2006. National Plan of Strate-gic Reserve of Coal 2006–2012. NN 81/2006.http://ec.europa.eu/competition/state_aid/register/ii/by_case_nr_nn2006_0060.html (accessedMarch 2, 2010).

———. 2008. The Support of Electricity fromRenewable Energy Sources. http://ec.europa.eu/energy/climate_actions/doc/2008_res_working_document_en.pdf (accessedMarch 2, 2010).

———. 2009. Study on the Energy Savings Potentialsin EU Member States, Candidate Countries andEEA Countries. Final Report. http://ec.europa.eu/energy/efficiency/studies/doc/2009_03_15_esd_efficiency_potentials_short_report.pdf(accessed February 22, 2010).

IEA (International Energy Agency). 2008. DeployingRenewables: Principles for Effective Policies. Paris:OECD Publishing.

———. 2009. Energy Policies of IEA Countries: Spain2009. Paris: OECD Publishing.

MARM (Ministry of the Environment and Rural andMarine Affairs). 2009. Inventario de Emisiones a laAtmósfera de España, Edición 2009, Serie 1990–2007,Sumario de Resultados de GEI. (Inventory of Emis-sions to the Spanish Atmosphere, 1990–2007 Series.Summary of GEI Results.) www.mma.es/secciones/calidad_contaminacion/pdf/Sumario_Inventario_de_Emisiones_GEI-_serie1990–2007.pdf (accessed March 2, 2010).

MITYC (Ministry of Industry, Tourism and Trade).2009a. La Energía en España 2008. (The Energy inSpain, 2008) www.mityc.es/energia/balances/Balances/LibrosEnergia/ENERGIA_2008.pdf(accessed March 2, 2010).

———. 2009b. Sebastián: “España tiene uncompromiso firme y ambicioso con las energíasrenovables”. June 5, 2009. www.mityc.es/es-ES/GabinetePrensa/NotasPrensa/Paginas/Rusia050609.aspx (accessed March 2, 2010).

Ragwitz, Mario, Joachim Schleich, Claus Huber,Gustav Resch, Thomas Faber, Monique Voogt,Rogier Coenraads, Hans Cleijne, and Peter Bodo.2005. FORRES 2020: Analysis of the RenewableEnergy Sources’ Evolution up to 2020. Stuttgart:Fraunhofer IRB Verlag.

324 Luis Agosti and Jorge Padilla

REE (Red Eléctrica de Espana) 2008. Informe delSistema Eléctrico.

Rodríguez, Juan M. 2007. REE (Red Eléctrica deEspaña), Integración en el Sistema y Operación.(Integration and Operation in the System). Pre-sented at CNE’s Jornadas sobre la Perspectiva Actualy Evolución de las Energías Renovables en España(Conference on Current Perspective and Evolutionof Renewable Energy in Spain). December 2007,Madrid.

Sáenz de Miera, Gonzalo, Pablo del Rio González, andIgnacio Vizcaino. 2008. Analysing the Impact ofRenewable Electricity Support Schemes on PowerPrices:The Case of Wind Electricity in Spain. EnergyPolicy 36 (9): 3345–59.

Secretaría de Estado de Energía (Secretary of State forEnergy). 2009. Resolución de 19 de noviembre de2009 (Resolution of November 19, 2009). http://noticias.juridicas.com/base_datos/Admin/res191109-itc.html (accessed March 2, 2010).

Renewable Electricity Support: The Spanish Experience 325

Conclusions: Whither RenewableGeneration?Boaz Moselle, Jorge Padilla, and Richard Schmalensee

The authors of this book have assessed the cur-rent use and future potential of renewable

energy technologies for electricity generation.These are technologies that, as Godfrey Boyleputs it in Chapter 2, “enable constantly replen-ished renewable energy flows to be harnessed toproduce power in forms useful to humanity on asustainable basis.” In particular, the authors haveaddressed the following two questions: What isthe case for promoting renewable energy? andWhat are the implications for power markets andsystems of the widespread adoption of renewablegeneration? In this concluding chapter, we sumup the answers provided in preceding chaptersand discuss some possible policy implications.

The Case for PromotingRenewable PowerKenneth Gillingham and James Sweeney argue inChapter 5 that the transition from carbon-basedtechnologies to renewable energy is inevitable inthe very long run. As noted by Erin Mansur inChapter 3, we are likely to see a significantincrease in the role of renewables in meeting elec-tricity demand in most developed countries incoming decades. Recent experience surveyed in

this book, notably Germany (by Hannes Weigtand Florian Leuthold in Chapter 14) and Spain(by Luis Agosti and Jorge Padilla in Chapter 15), isconsistent with this trend. The goal of the Euro-pean Union is to achieve significant reduction incarbon intensity by 2020, with renewable genera-tion playing an important role. As explained byChristopher Jones in Chapter 12, about 60% ofthe world’s wind capacity was installed in Europeat the end of 2007, and the EU is committed to a20% share for renewable energy in its total energymix by 2020, compared with about 8.5% in 2005.

The growth in renewable generation is not—and, for the foreseeable future, is not likely tobe—driven by market forces, but is largely theresult of government intervention. As describedby Richard Schmalensee in Chapter 11, there arefour basic kinds of support schemes targeted spe-cifically at renewable generation: feed-in tariffs,output subsidies, investment subsidies, and outputquotas. Feed-in tariffs guarantee a predetermined,above-market price for power over a period ofyears.They have been used widely within the EU,notably in Germany and Spain, and have provedto be a powerful mechanism to promote invest-ment, minimize costs to consumers (for a givenlevel of output and mix of renewable technolo-gies), and maximize production. Output subsidiespaid on top of market price, though not as widely

employed, can be shown to be as effective as, andless distortionary than, feed-in tariffs. Investmentsubsidies are less efficient than feed-in tariffs oroutput subsidies yet are widely used all over theworld to promote the deployment of renewablepower plants. Finally, output quotas typicallyrequire agents operating in power markets to gen-erate or procure a minimum fraction of energyfrom renewable sources.

Renewable generation support schemes,whether feed-in tariffs, output subsidies and quo-tas, or investment subsidies, are commonly justi-fied as a means of curbing CO2 emissions andtherefore as part of the overall response to globalwarming. Indeed, U.S. and especially EU energypolicies place renewables at the forefront of thefight against climate change (see Chapters 11 and12). Climate change is expected to adverselyaffect many economic sectors and natural systemsand increase human mortality and morbidity. Theneed to address this threat is beyond dispute. Withrespect to renewable generation, the relevant anddifficult question from an economic perspective iswhether current and planned support schemesrepresent an efficient response to the challengesposed by climate change.

From an economic viewpoint, the startingpoint for designing an efficient climate policy is amarket-based mechanism to internalize the exter-nalities associated with emissions of greenhousegases, such as a carbon tax or emissions trading(cap-and-trade) system. A properly designed car-bon tax or cap-and-trade system can in theoryreduce emissions at least cost.

Those market-based instruments are by defi-nition technology-neutral.1 That is, they affectdifferent technologies in terms of their green-house emissions only, and therefore do not favorone technology (e.g., renewable technologies)over other low-carbon forms of generation (e.g.,nuclear power or energy efficiency) on a priorigrounds. Under those mechanisms, technologiescompete on their merits and electricity is gener-ated efficiently, using those technologies that aremore economical on the basis of both their actualcosts of production and the market cost of theiremissions. Investment in renewable energy may

be an efficient response to those market-basedinterventions, but there is no guarantee that this isthe case.

Market-based solutions would also affect allCO2-emitting sectors in a similar way and wouldnot place the entire decarbonization burden onthe electricity sector. Most of the jurisdictionsanalyzed in this book have committed to ambi-tious renewable generation targets even thoughthey perhaps could have more easily and eco-nomically achieved the decarbonization of theireconomies through a properly designed carbontax.

Governments almost universally have adoptedspecific policies to promote renewable generationrather than relying on market-based solutions,however. This is so even though alternative tech-nologies may offer the same environmental ben-efits at lower cost. Of course, policy choices arethe outcome of political processes that are oftencomplex, and it would be naive to expect thatthose outcomes generally coincide with the rec-ommendations of economics textbooks.

Existing renewables policies and future policyproposals have been defended using both eco-nomic and noneconomic arguments. The variouschapter authors have a range of views regardingthose justifications and the necessity or desirabilityof specific support mechanisms for renewablegeneration. The view of the book editors isderived from three observations.

First, it is plausible that the true size of thenegative externality associated with CO2

emissions—the social cost of carbon—is higherthan the current price of CO2 emissions (e.g., inthe EU Emission Trading System, or ETS), andthat, as Boyle argues in Chapter 2, the differenceexceeds the incremental cost of renewable genera-tion relative to conventional generation. This is anecessary condition for supporting renewablegeneration, though by no means sufficient (forexample, one might be able simply to raise theprice of CO2 emissions).

Second, it is also plausible that other forms ofemissions reduction—such as nuclear, carboncapture and storage (CCS), or energy efficiencypolicies2—are either more costly (at the relevantmargin) than renewables or constrained by

328 Boaz Moselle, Jorge Padilla, and Richard Schmalensee

noneconomic factors, such as the public accept-ability issues around nuclear power.

The third observation is an empirical one, forwhich Part IV of this book provides ample sup-port: political processes do not easily deliver out-comes that resemble the first-best outcome oftechnology-neutral, market-based mechanisms toreduce greenhouse gas emissions. Why this is soraises many interesting questions of politicaleconomy, but it is clear that governments facegreat difficulty in credibly committing to a long-term carbon price that is close to the social cost ofcarbon or high enough to support significantinvestments in low-carbon technologies. Even theEU ETS, which is an impressive and uniqueachievement, has to date given a level of carbonprices that induces switching from coal to gas-fired generation but is not high enough to supportsuch investments.

Drawing these three observations together,the editors therefore recognize that the need torespond to the challenges posed by global warm-ing, combined with real-world constraints onpolicymaking processes, may provide a valid justi-fication for specific support mechanisms forrenewables, especially for less costly forms. How-ever, no plausible combination of assumptionsjustifies a “pay whatever it takes” approach. Theshadow value of investing in a given technology isfinite, limited, if by nothing else, at least by theopportunity cost of investing more in lower-costtechnologies. In particular, it is hard to justifyextensive investment in deploying extremelyexpensive forms of renewables, such as solarphotovoltaic in Spain, Germany, and parts of theUnited States and offshore wind in the UnitedKingdom.

This logic for supporting some specific meas-ures to promote renewable generation does notdepend on the kinds of market failure argumentsthat are often cited in favor of such measures.Although it is possible in theory to identify manysuch market failures (see Chapter 5), in most casesthey appear unlikely to be material enough tojustify the kind of large-scale interventions theyare used to support. In general, the claimed “non-climate-change externality” arguments forrenewables support appear rather weak.

For example, investing in renewables to createjobs seems like poor economic policy. This isbecause electricity generation is generally highlycapital-intensive, the time frame for investment istoo long to allow for a temporary boost inemployment in the short run (during a recession),and in the long run the level of employmentdepends on structural macroeconomic factors andis unlikely to be affected by policies of this nature.As an illustration, in Spain, the rapid growth ofrenewable generation before and after the GreatRecession has had, at most, a de minimis impact onits labor markets and has not offset in any mean-ingful way the growth of unemployment duringthe current crisis.

With regard to industrial policy, almost everymajor government in the developed world seemsto believe that it will become a world leader innew energy technologies, and that such leadershipwill generate positive externalities on other com-panies within the energy sector as well as on otherindustrial and services sectors. As yet, however,evidence is slim in support of that claim for anynation. In particular, little evidence exists thatextensive deployment of renewables createslearning-by-doing (or knowledge) spillovers thatcannot be captured within firms (see, e.g., Chap-ter 5). Moreover, as in other industries, it is likelythat these positive externalities are much strongerat the R&D level. Subsidies for wide-scaledeployment of expensive renewable technologiesare therefore unlikely to be justified on the basiseither that the cost is overestimated (because theywill bring down future costs) or that they willstimulate an infant industry.

An alternative justification for existing andproposed support schemes for renewables con-cerns their impact on security of supply. In mostmarkets, individual participants deal effectivelyand efficiently with uncertainty regarding demandand supply through price incentives, bilateralinsurance contracts, and related devices.The con-clusion that supply insecurity justifies governmentaction here must rest on the identification of oneor more specific externalities that can be cor-rected by nonmarket intervention.

From a U.S. perspective, Gillingham andSweeney point out in Chapter 5 that an important

Conclusions: Whither Renewable Generation? 329

externality may relate to the costs associated withnational security, but this concerns oil and haslittle relevance to renewables other than biofuels.Looking at the EU, where security of supply ishigh on the political agenda and is routinely citedas a justification for renewables subsidies, BoazMoselle in Chapter 4 identifies significant exter-nalities in the context of the key relevant security-of-supply issue: dependence on Russia and Alge-ria for natural gas. Although EU policymakersoften refer to generalized import dependence, fewsecurity issues appear to exist around the use ofimported coal, given that many diverse and his-torically friendly countries have enormous coalreserves. Moselle discusses the extent to whichrenewable generation in the EU, as opposed toother energy sources, may replace Russian andAlgerian gas. He concludes that dependence onimported gas gives rise to real security-of-supplyissues in the EU, especially in eastern Europe, andthat market outcomes may not provide an effi-cient response to those risks because of “moralhazard” types of concerns that give rise to exter-nalities. He argues, however, that these externali-ties do not justify the specific promotion ofrenewables, because it is not clear how muchRussian or Algerian gas would be displaced byrenewable generation, and because other forms oflow-carbon generation could have similar orgreater impact on security of supply.

In summary, then, although the environmen-tal benefits of renewable generation are clear, theeditors are skeptical of economic and geostrategicjustifications for policies specifically aimed at pro-moting renewables. They do, however, recognizea political rationale for existing renewable supportschemes: subsidizing renewables may be politi-cally more acceptable than taxing firms and con-sumers for polluting. As in many other instances,first-best economic policy choices may not bepolitically feasible. In such cases, we need to lookfor second-best policy responses. Existing andplanned support schemes can be rationalized onlyas second-best responses to the challenges posedby climate change. Yet the authors widely agreethat current and anticipated policies need consid-erable fine-tuning before they can be properlyregarded as the right policy instruments even in a

second-best world.The editors would add that theconstraints that lead to second-best policy out-comes can change, and as the true cost of notusing technology-neutral, market-based mecha-nisms becomes clearer over time, the opportunitymay arise to move closer to first-best.

Renewable Generation andPower Systems in PracticeCurrent policies toward greenhouse gas abate-ment in the electricity generation sector in theEU, in many EU member states, in some U.S.states and regions, and under serious considera-tion at the federal level in the United States com-bine a technology-neutral market-based mecha-nism (e.g., the EU ETS and similar cap-and-tradeschemes in California and the U.S. Northeast)with technology-specific interventions forrenewables, nuclear, and CCS. This mix of poli-cies has profound implications for wholesalepower markets.

Existing and planned policies may raise ques-tions about the current model of liberalizedpower markets. A large part of the benefits fromliberalization have come about through improvedinvestment decisions. In fact, dynamic efficiencyconsiderations underlay the move from regulatedto liberalized generation. However, current poli-cies mean that market forces will have limitedimpact on future investment decisions, as thosewill be driven, to a large extent, by the environ-mental and security-of-supply concerns of gov-ernments. In the United Kingdom, for example,choices for new investment are affected by a verylarge number of different environmental programs(as discussed by Michael Pollitt in Chapter 13),whose cumulative effect is likely to be determina-tive.3

At a more concrete level, existing and plannedpolicies are bound to have a significant impact onthe economics of conventional generation plants.As shown by James Bushnell in Chapter 9, theexploitation of wind resources will have a signifi-cant impact on market prices. It will also affect thenet load shape, the difference between load and

330 Boaz Moselle, Jorge Padilla, and Richard Schmalensee

supply from intermittent sources. This is likely toexhibit relatively higher spreads between peak andaverage demand for thermal production. As a con-sequence, the equilibrium investment mix ofnonwind resources will shift toward less base loadand more peaking capacity. The experiences inSpain and Germany show that the widespreaddeployment of intermittent generation (wind)leads to increased price volatility, with very low(on occasions even zero or negative) prices whenwind patterns make for high levels of output attimes of low load (see Chapters 14 and 15).

This combination of spot price volatility andthe need for large numbers of peaking plants withvery low utilization is likely to significantly ex-acerbate the “missing money” problem that canalready arise in wholesale power markets, where acombination of the lack of real-time pricing formany consumers, transmission system operator(TSO) behavior, explicit regulatory interventions(price caps), and implicit regulation means thatmarkets do not provide the right signals to getsufficient peaking capacity.

Essentially three approaches can be taken tothis problem. The first is to observe that marketsare quite capable of dealing with problems of thissort via the price mechanism, including for goodsthat are not storable. For example, demand foraccommodation on or near a ski slope is highlyweather related. High prices for that accommoda-tion at times of peak demand (e.g., around Christ-mas and during school holidays in winter) ensurea high level of investment and the absence of linesoutside ski chalets. Similarly, if there are no pricecaps and investors do not expect future govern-ment intervention, then they should be expectedto build sufficient capacity to cope with demand,investing in backup plants that would tend to runinfrequently, on the basis that the spot priceswhen they did run would be high enough thatthey could recover their fixed costs over time. Atthe same time, demand should become moreresponsive to short-term price signals as theprevalence of price spikes makes it worthwhile todevote time and make necessary capital invest-ments so as to be able more easily to reduce orshift demand at times of peak load.

In many (probably most, and perhaps all)jurisdictions, however, governments lack the abil-ity to credibly commit not to intervene in energymarkets when prices are high. Indeed many liber-alized energy markets, including several in theUnited States and in a number of EU memberstates, already have explicit price caps in place.These caps in part reflect political sensitivities butare also a reaction to well-founded concerns thatwholesale power markets are more susceptiblethan many others to market abuse.

In those circumstances, the second approachto dealing with intermittency, price spikes, andthe need for backup is to provide a subsidy toinvestment in generation (and on efficiencygrounds, perhaps also to flexible load) in the formof “capacity payments”. A number of designissues around the capacity markets determinethese payments, which have been widely explored(in theory and in practice) in the United States,much less so in the EU, where few jurisdictionscurrently have any form of capacity payment inplace. Getting those payments at the right levelappears to be key to ensuring that the movetoward renewable energy does not compromisethe reliability of electricity systems.

The third approach would be to conclude thatmeasures of this kind are ineffective, and thatcompetitive markets with a significant proportionof intermittent (and low-marginal-cost) renew-able energy are not able to deliver the necessaryrents to induce adequate investment in peakingcapacity. From the discussion above, it should beclear that the editors do not share that conclusion.The prospect of re-regulation of electricity aimedat ensuring generation adequacy, while at thesame time giving governments more direct con-trol over technology choices, is nonetheless a realprospect,4 though in the editors’ view, it is prob-lematic and potentially worrisome.

That the shift to renewable energy may lead tonew market distortions and regulatory challengesdoes not justify giving up on markets. No doubtregulators will have to adapt their tools to accom-modate the environmental and security-of-supplyconcerns of governments, but they should do sowithout throwing off the discipline that a com-petitive market imposes not only on firms and

Conclusions: Whither Renewable Generation? 331

consumers, but also on policymakers. Bad poli-cies, like poorly managed firms, tend to fail themarket test quickly.

Power markets that integrate large amounts ofwind face several other policy issues. First, apartfrom the investment issue discussed above,increasing penetration of intermittent generationgives rise to operational concerns because of thedifficulty of predicting output ahead of time.However, significant improvements have takenplace in forecasting wind generation output overtime (see Chapter 2). Moselle argues in Chapter 4that as a result, the operational issue can in princi-ple be made relatively straightforward and, subjectto solving the problem of providing investmentincentives to ensure sufficient availability of peak-ing plants, the problem becomes simply a furthercost element, albeit potentially a very materialone, rather than a security-of-supply concern.

Second, a subsidy per megawatt-hour ofrenewable generation distorts price signals andcan lead to inefficiencies. For example, the mas-sive introduction of wind generation can lead tonegative prices, reflecting an inefficient outcomewhere wind generators, which could stop run-ning at zero social cost but at the expense of losingtheir feed-in tariff payments, pay others (e.g.,nuclear generators) to incur real costs to produceless electricity (Schmalensee reports on negativespot prices in west Texas in Chapter 11). Otherinefficiencies may arise when the level of paymentis unrelated to the market price (e.g., the absenceof an incentive to produce at peak hours or atpeak times of the year).

A third problem concerns the potential needfor new investment in transmission, as new pat-terns of generation, reflecting different patterns oflocation for new installed capacity, lead to chan-ging flow patterns in transmission systems. Chris-tian von Hirschhausen explains in Chapter 10 thatthe complexities of harnessing renewable energyto generate electricity are relatively simple incomparison to transporting that electricity overlong distances to large demand centers. He alsodescribes the many market and institutional prob-lems, including regulatory and technological risksand rent-sharing issues, which make it difficult toovercome the transmission bottlenecks that limit

the effective deployment of renewables. Ifrenewables are to play a more important role,policymakers must address these difficult prob-lems.

The location of large-scale renewable genera-tion will depend on a number of factors, includ-ing the location of appropriate resources, such aswater, wind, sunshine, and sources of biomass; therelevant transportation costs for some of theseresources; issues around siting; and the costs ofdeveloping the necessary distribution and trans-mission systems. Efficient locational pricing, asdiscussed by William Hogan in Chapter 7, willhelp provide incentives for efficient decisionsconcerning the location of generation and trans-mission facilities, such as by ensuring higherprices in areas where transmission bottlenecksexist. However, it would be daunting even intheory to attempt to fully decentralize transmis-sion investment decisions via locational pricing, ifonly because transmission investments are lumpyand commonly eliminate the nodal price differ-ences that justified the investments. In practice,therefore, more extensive work is needed ondeveloping and implementing appropriate meth-odologies for transmission planning.

Moreover, one effect of the need to locategeneration close to renewable resources, and farfrom NIMBY-minded citizens, is likely to be amuch greater mismatch between the locations ofgeneration and load. New arrangements for trans-mission planning thus need to include arrange-ments to coordinate planning among multipletransmission owners. U.S. experience suggeststhat electricity systems operated by regional trans-mission operators (RTOs) and independent sys-tem operators (ISOs) are more effective in inte-grating wind generation. Those systems leveragetheir operational and transmission integration tofacilitate deployment of renewable generation (seeChapter 11). A new approach to infrastructureplanning is foreseen in Europe. Under the thirdinternal market package, two new Europeanorganizations, the Agency for the Cooperation ofEnergy Regulators (ACER) and the EuropeanNetwork of Transmission System Operators forElectricity (ENTSO-E), will have to cooperateand coordinate to put in place new infrastructure

332 Boaz Moselle, Jorge Padilla, and Richard Schmalensee

and grid access rules to ensure that the growth inrenewable energy is sustainable and the availablecapacity is used efficiently (see Chapter 12).

Finally, large-scale deployment of renewablegeneration raises questions as to the nature ofregulation. Richard Green maintains in Chapter 8that energy regulation will need to adjust, forexample by recognizing the need for new regu-lated investments (e.g., in “smart grids” to com-plement distributed generation), but that thereshould be no step change in the fundamentalscope of energy regulation. He also suggests thatregulators should continue to focus on the pro-motion of competition, and indeed that competi-tive models will have extra advantages in the con-text of a transition to renewables because of theirsuperior performance in stimulating innovation.

ConclusionExcept possibly for nuclear terrorism, globalwarming is widely regarded as the most importantchallenge to our globalized society. Policymakersare considering ways of confronting this challengeand are also putting a substantial amount ofmoney on the table. A significant share of thatmoney is being allocated to renewable generationsupport schemes. For a variety of reasons, a lot ofthat money will be spent inefficiently. Because thesums involved are very large, it is imperative thatwe discuss how to move toward a decarbonizedelectricity system at the least possible cost. Thisrequires moving away from generic and arcanedebates about policy goals and toward analyzingthe design and implementation of renewable gen-eration support schemes, as well as the relativemerits of renewable generation policies in com-parison with other decarbonization policies. It ishoped that this book represents a nontrivial stepforward in that direction.

Notes1. A “market-based instrument” is defined in this

context to refer to an instrument that puts a price

on greenhouse gas emissions (or at least on CO2

emissions), either directly via a tax or indirectly viacap-and-trade. Under this definition, therefore, aninstrument like tradable renewables certificates (or,hypothetically, a tax on all nonrenewable genera-tion) does not qualify as a market-based instru-ment.

2. In industrialized countries, as explained by JoséGoldemberg in Chapter 6), energy efficiency willbe an attractive, though limited, alternative torenewable generation. In developing countries,energy efficiency programs are less likely to substi-tute for the promotion of carbon-free or low-carbon technologies. In those latter countries,energy demand is bound to grow and should bemet by deploying clean and efficient technologiesearly in the process of development.

3. These programs include the EU ETS; the nationalUK Emissions Trading Scheme; the RenewablesObligation Certificates, which fund renewablesdeployment differentially by technology; invest-ment subsidies for renewables, known as capitalgrants; feed-in tariffs for new microrenewables;support for cogeneration; new support measurescurrently being developed for nuclear and CCS;obligations on suppliers to install energy efficiencymeasures to meet the Carbon Emissions Reduc-tion Target (CERT); as well as a number of othermeasures.

4. Recent proposals from the British energy regula-tor (formerly a leading advocate of liberalizedenergy markets) include, at one extreme, the crea-tion of a single central buyer responsible for pro-curing wholesale power and contracting to pro-cure new generation investment, which it wouldsecure against long-term contracts (see Ofgem2010).

ReferenceOfgem. 2010. Project Discovery: Options for Delivering

Secure and Sustainable Energy Supplies, February 3.London: Ofgem.

Conclusions: Whither Renewable Generation? 333

Index

Note: Page numbers in italics indicate figures and tables. Page numbers followed by an ‘n’ indicate notes.

“20-20-20” proposal, 235–2368.5 GW Severn Barrage, 264

Agency for the Cooperation of EnergyRegulators (ACER), 246–247

Airtricity, 187–188, 188American Recovery and Reinvestment Act

(2009), 42, 43amorphous silicon (a-Si), 12–13Annual Energy Outlook (EIA), 42–43, 43ARRA (American Recovery and Reinvestment

Act) (2009), 42, 43atemporal market failures, 72–87, 76, 80, 81,

87

backup generation, investment in, 61, 61–64balance-of-system (BoS) costs, 14–15Biofuels Directive (EU), 234biomass, United Kingdom, 268–269business-as-usual (BAU) scenario, EEA-MENA

2050 case study, 192–193Business Energy Investment Tax Credit

(BEITC), 215–216

CAISO-TEAM, 200California, NHR technologies, 221, 221–223,

222cap-and-trade programs for CO2, 161capacity market

electricity market, 123equilibrium model analysis, 171, 171–172,

172, 173carbon dioxide, policy instruments for, 85–86,

161carbon emission caps, 129–130carbon market impact

equilibrium model analysis, 174–175, 175carbon trading, 145CCGTs (combined cycle gas turbines),

310–311, 319–320Centre for the Control of the Special Regime

(CECRE), 319

Clean Development Mechanism (CDM),105–107, 106

climate policy, recent investments in renewables,36–38

coal supply, 54, 54combined cycle gas turbines (CCGTs),

310–311, 319–320concentrating systems for solar energy, 9–11, 13countertrade markets, 116, 119–121crystalline silicon, 12current energy sources, 34–36, 35, 36

decarbonization policies, 254–257, 255, 256,257, 265–268, 274–276

Department of Energy (DOE). see U.S.Department of Energy (DOE)

Desertec Project, 26–27, 189–190, 190dish Stirling solar concentrators, 11dispatch protocol for low-carbon technologies,

132–133

economic dispatch, selection of low-carbontechnologies, 132–133

economies of scaleatemporal market failures, 76policy instruments for, 85

EEA-MENA 2050 case study, 191–198, 192,194, 195, 196

EEG (Germany), 285, 286–287, 287efficiency potential, 98–101, 100, 101, 102EIA (Energy Information Administration),

Annual Energy Outlook, 42–43, 43Electricity Feed-In Act (Stromeinspeisegesetz,

StrEG) (1991), 285, 285–286electricity generation targets, European Union,

143electricity markets

alternative designs, 116–121capacity market, 123connection standards, 121–122countertrade markets, 116, 119–120, 121design for a low-carbon future, 120–121

effect of renewable energy, 138–142, 140financial transmission right (FTR), 118hybrid infrastructure, 124independent system operators (ISOs), 116integrated locational electricity market design,

117–119, 121investment in, 124–125, 161–163locational marginal pricing (LMP), 118reform, 115–116regional transmission organizations (RTOs),

116–119smart systems, 122vertically integrated electric utilities, 122–124

electricity systemsfundamental characteristics, 113–115integration of renewable energy sources,

24–27, 26ELMOD model, 191Emission Trading System (ETS), 236emissions trading schemes, 101–107, 104, 105,

106, 107energy

consumption, 94defined, 7efficiency potential, 98–101, 100, 101, 102sources of, 34–36, 35, 36, 93storage of, 25–26

energy availability profiles, equilibrium modelanalysis, 175, 175–176, 176

Energy Information Administration (EIA),Annual Energy Outlook, 42–43, 43

energy-only market, equilibrium model analysis,168–171, 169, 170

“An Energy Policy for Europe”, 235, 248–251energy utilities, and a low-carbon system,

142–145Energy Watch Group, renewables forecasts, 45environmental externalities

atemporal market failures, 73policy instruments for, 82–83

Environmental Protection Agency (EPA),43–44, 44

equilibrium model of electricity investmentoverview, 163–165analysis and results, 168–176demand data, 165, 165–169, 166, 167thermal generation cost data, 168, 168WECC subregions, 164, 164wind generation data, 165–168, 166, 167

equilibrium model of generation investment,177–178

ETS system, 2050 zero-carbon target, 245–246European Commission, World Energy Technology

Outlook—2050 (WETO), 40–42, 41European Network of Transmission System

Operators for Electricity (ENTSO-E), 247European supergrid projects, 26, 187–188, 188,

191–198, 192, 194, 195, 196European Union (EU)

“20-20-20” proposal, 235–236Biofuels Directive, 234electricity generation targets, 143Emission Trading System (ETS), 236feed-in tariffs (FIT), 237, 242First Strategic EU Energy Review, 235,

248–251future energy targets, 243–244green certificates, 237imports, 243infrastructure issues, 246–247Renewable Electricity Directive, 234,

238–242renewable energy policy, 233–238, 242–246tendering schemes, 236–237trade in renewable certificates, 243zero-carbon system, 244–246

European–North African Supergrid(EEA-MENA 2050) case study, 191–198,192, 194, 195, 196

exceedance measure, 178–179 n6

feed-in tariffs (FIT)effectiveness of, 269–271European Union (EU), 237, 242Germany, 288–291, 290, 292green energy policy mandates, 133United Kingdom, 265, 269–271

financial transmission right (FTR), 118First Strategic EU Energy Review

key excerpts, 248–251objectives, 235

flow-based deviations, 77–79, 85–88forecasts of renewables, 25, 26, 39–45, 40, 41,

43, 44, 46FORRES 2020 Project, 44–45

gallium arsenide (GaAs), 12gas industry

CCGTs (combined cycle gas turbines),310–311, 319–320

EU dependence on Russian gas, 53–59implications for regulation, 153Ukraine, 53

336 Index

gas supply, 54–55, 55generation investment, equilibrium model,

177–178generation revenues, regulation of, 151–152geothermal energy, 8, 21Germany

economic analysis of RES support, 292–297,295, 296

Electricity Feed-In Act (Stromeinspeisegesetz,StrEG) (1991), 285, 285–286

feed-in tariffs (FIT), 288–291, 290, 292market implications for RES support,

297–302R&D, 284, 285renewable energy policy, 283–288, 284, 285,

287, 288Renewable Energy Source Act (Erneuerbare

Energien Gesetz, EEG), 285, 286–287,287

RES generation, 303green certificates, 237, 241–243green energy policy mandates, 129–134green tags, markets for, 37grid systems, 13–14, 122

see also supergrids

heat pumps, 157 n29Howard, Ron, 134 n5HVDC grid expansion

EEA-MENA 2050 case study, 193–194, 194,195

hybrid infrastructure, 124hydroelectric technologies, 20

IEA (International Energy Agency). seeInternational Energy Agency (IEA)

imperfect foresightintertemporal market failures, 78–79policy instruments for, 84–87

import dependence, 52–59independent system operators (ISOs), 116information market failures

atemporal market failures, 74–75policy instruments for, 83–84

infrastructure investmentelectricity market, 124–125, 128–129and planning issues, 246–247supergrids, 198–202

integrated firms model of regulation, 141integrated locational electricity market design,

117–119, 121Integrated Planning Model, 43–44, 44

Intergovernmental Panel on Climate Change(IPCC 2007), renewables forecasts, 45

intermittencychallenges to EU renewable energy policy,

246green energy resources, 127–128investment in backup generation, 61, 61–64problems for renewable generators, 143–144renewable energy sources, 36system balancing, 59–61

International Energy Agency (IEA)nuclear energy, 323 n4World Energy Model (WEM), 39–40, 40World Energy Outlook, 39

International Energy Outlook (DOE), 39–40intertemporal market failures, 77–79, 85–88investment in backup generation, 61, 61–64

Joint Coordinated System Plan 2008 (JCSP’08),185–187, 186

joint implementation (JI), 104–105

labor market supply–demand imbalancesatemporal market failures, 72–73policy instruments for, 82

learning-by-doing (LBD)green policy mandates, 130–131intertemporal market failures, 78–79policy instruments for, 86–87, 87

locational marginal pricing (LMP), 118low-carbon technologies

design for, 120–121selection in energy dispatch, 132–133utilities in, 142–145

low-temperature solar thermal-electrictechnologies, 9–10

market failuresatemporal market failures, 72–77, 76intertemporal market failures, 77–79policy instruments for, 69–89, 80, 81, 87and resource use, 71

market poweratemporal market failures, 76–77policy instruments for, 85

medium- and high-temperature solar electrictechnologies, 10–11

national security externalitiesatemporal market failures, 73–74policy instruments for, 83

Index 337

natural gas industryCCGTs (combined cycle gas turbines),

310–311, 319–320EU dependence on Russian gas, 53–59implications for regulation, 153Ukraine, 53

network externalitiesintertemporal market failures, 79policy instruments for, 87

New York Independent System Operator(NYISO) model, 124

Non-Fossil Fuel Obligation (NFFO)costs and delivery, 262, 262–264, 263, 264United Kingdom, 258–260

nonhydro renewable (NHR) technologiesoverview, 210, 210–213, 211, 212Business Energy Investment Tax Credit

(BEITC), 215–216in California, 221, 221–223, 222deployment, 214–217, 216, 217Public Utilities Regulatory Policies Act

(PURPA) (1978), 214–215R&D, 213–214Renewable Electricity Production Tax Credit

(REPTC), 215–216renewable portfolio standards (RPSs),

217–220, 218, 219Residential Renewable Energy Tax Credit

(RRETC), 215state policies, 217–221, 218, 219, 220support of, 213–217in Texas, 223, 223–224wind power technologies, 224–227, 225

North African supergrid projects, 189, 189–190,190

North Sea Wind Energy Super Ring project,189, 189

nuclear energy, 312–313NYISO model, 124ocean thermal energy conversion (OTEC),

9–10offshore wind, 19, 19, 273–274onshore wind, 266–268OR (Spain), 310–311OTEC (ocean thermal energy conversion),

9–10

parabolic trough concentrators (PTCs), 10–11peak load pricing, 149

see also price regulationperfect markets, 71

see also market failures

photovoltaic (PV) electricity, 11–16, 15, 16Piebalgs, Andris, 236policy intervention, and import dependence,

56–59political consequences for supergrid projects,

201–202power, defined, 7power towers, 11price regulation, 147–151

see also peak load pricingPTCs (parabolic trough concentrators), 10–11Public Utilities Regulatory Policies Act

(PURPA) (1978), 214–215PV electricity, 11–16, 15, 16

R&Dgreen policy mandates, 130intertemporal market failures, 78NHR technologies, 213–214policy instruments for, 86

RECs (renewable energy credits), 37redispatch, selection of low-carbon technologies,

132–133Régimen Especial (Special Regime, or SR),

310–311Régimen Ordinario (Ordinary Regime, or

OR), 310–311regional transmission organizations (RTOs),

116–119regulation

electricity industry, 138–142, 140, 145–153gas industry, 153generation revenues, 151–152

regulatory failuresatemporal market failures, 75policy instruments for, 84

remote power, PV systems for, 13Renewable Electricity Directive (EU)

feed-in tariffs (FIT), 242national targets, 239–241, 240results of, 234trade in renewable certificates, 241–242

Renewable Electricity Production Tax Credit(REPTC), 215–216

renewable energyfundamental issues, 70policy implications, 330–333promoting, 327–330regulation of electricity industry, 138–142,

140in restructured markets, 160–161

renewable energy credits (RECs), 37

338 Index

renewable energy policyEuropean Union (EU), 233–238, 242–246Germany, 283–288Spain, 310–322United Kingdom, 265–268United States, 210–217

Renewable Energy Source Act (ErneuerbareEnergien Gesetz, EEG), 285, 286–287, 287

renewable energy sourcesoverview, 95–98, 97, 98, 99costs and intermittencies, 36integration into electricity systems, 24–27, 26

renewable energy technologiescosts, 127, 161economics, 21–24, 22, 23, 24future direction of, 38–39investment in, 161–163policy mandates, 129–134recent investments, 36–38resources, 125uncertainties, 125–126

renewable generators, 142–145Renewable Heat Incentive (RHI), 265renewable portfolio standards (RPSs)

green policy mandates, 131–132implementation of, 37NHR technologies, 217–220, 218, 219

renewables forecasts, 25, 26, 39–45, 40, 41, 43,44, 46

Renewables Obligation (RO) scheme, 260,260–262, 261

Residential Renewable Energy Tax Credit(RRETC), 215

resource use, and perfect markets, 71restructured markets, 160–163retail competition model of regulation, 140retail prices, regulation of, 152–153RFF Haiku model, 45RHI (Renewable Heat Incentive), 265RO scheme, 260, 260–262, 261ROCs (Renewables Obligation Certificates). see

RO schemeRPSs (renewable portfolio standards), 37

green policy mandates, 131–132implementation of, 37NHR technologies, 217–220, 218, 219

RTOs (regional transmission organizations),116–119

Russian gas, EU dependence, 53–59

Scandinavian electricity market, 194–196, 196Severn project, 264

single buyer model of regulation, 141smart grids, 122solar chimney, 10solar energy, 7–11Solar Grand Plan, 183–185, 184solar photovoltaic (PV) electricity, 11–16, 15,

16solar ponds, 9solar radiation, 184Spain

Centre for the Control of the Special Regime(CECRE), 319

combined cycle gas turbines (CCGTs),310–311, 319–320

current policy assessment, 315–320, 316, 318,319, 320

current policy status, 310–315, 311, 312,313, 314, 315

electricity costs, 316–318, 318nuclear energy, 312–313policy goals, 312–313policy recommendations, 321–322regulatory framework, 310–311, 311, 312renewable electricity support scheme, 316,

316renewable regulation, 313, 313–315, 314,

315supply security, 319, 319–320, 320

SR (Spain), 310–311stimulus packages

American Recovery and Reinvestment Act(2009), 42, 43

investment in green activities, 107–108, 108stock-based deviations, 72–87, 76, 80, 81, 87stock-based environmental externalities

intertemporal market failures, 77policy instruments for, 85–86

storage of energy, 25–26StrEG (Germany), 285, 285–286submarine wind energy superhighways,

187–188, 188supergrids

development challenges, 198–202EEA-MENA 2050 case study, 191–198, 192,

194, 195, 196European projects, 26, 187–188, 188evaluating, 190–191infrastructure investment, 198–202political consequences, 201–202typology, 182–183, 183, 184U.S. projects, 183–187, 184, 186, 188see also grid systems

Index 339

system balancing, 59–61

technologies for renewable energy, economicsof, 21–24, 22, 23, 24

tendering schemes, 236–237Texas, NHR technologies, 223, 223–224thin-film PV technologies, 12–13tidal power technologies, 8, 20–21, 264too-high discount rates

atemporal market failures, 75policy instruments for, 84

tradable green certificate (TGC), 269–270transmission investment, 133–134

Ukraine, role in gas industry, 53uneconomic dispatch, selection of low-carbon

technologies, 132–133Union of Concerned Scientists, renewables

forecasts of, 45United Kingdom

biomass, 268–269decarbonization policies, 254–257, 255, 256,

257, 274–276feed-in tariffs (FIT), 265, 269–271future policy, 272–274Non-Fossil Fuel Obligation (NFFO),

258–260offshore wind, 273–274policies since 1990, 257–265policy compared to other countries, 269–272,

270renewable energy policy, 253–254renewable energy policy assessment, 265–268Renewable Heat Incentive (RHI), 265Renewables Obligation (RO) scheme, 260,

260–262, 261tidal projects, 264–265tradable green certificate (TGC), 269–270wind power technologies, 266–268

uranium supply, 54, 54U.S. Department of Energy (DOE)

Annual Energy Outlook (EIA), 42–43, 43

Energy Information Administration (EIA),42–43, 43

International Energy Outlook, 39–40U.S. Environmental Protection Agency (EPA),

43–44, 44U.S. supergrid projects

integration of wind, solar and biomassgeneration, 187, 188

Joint Coordinated System Plan 2008(JCSP’08), 185–187, 186

Solar Grand Plan, 183–185, 184utilities, and a low-carbon system, 142–145

vertically integrated electric utilities, 122–124

wave power technologies, 8, 20–21, 264Waxman-Markey bill, 43–44, 44WEM (World Energy Model), 39–40, 40Western Electricity Coordinating Council

(WECC) subregions, 164, 164WETO (World Energy Technology Outlook –

2050), 40–42, 41wholesale competition model of regulation, 140wind power

annual growth rate, 35forecasting, 25, 26worldwide capacity, 35

wind power technologiesequilibrium model analysis, 172, 172–174,

174Germany, 297–302, 298offshore wind, 19, 19onshore wind farms, 266–268planning problems, 266–268potential of, 17–19, 18, 224–227, 225principles of, 16–17United Kingdom, 272–274

World Energy Model (WEM), 39–40, 40World Energy Outlook (IEA), 39World Energy Technology Outlook—2050

(WETO), 40–42, 41

zero-carbon system, 244–246

340 Index