Embed Size (px)
Citation preview
ANL/RP--76523
DE92 041151
GLOBAL CLIMATE CHANGE:SOCIAL AND ECONOMIC RESEARCH ISSUES
Proceedings of a ConferenceHeld at Argonne National Laboratory
on February 11-13,1992
Edited byDr. Marian RiceDr. Joel Snow
Dr. Harold Jexobson
Published May 1992
Conference Sponsored byMidwest Consortium for International Security Studies
andArgonne National Laboratory*
Members of MCISS:
Argonne National Laboratory University of ChicagoChicago Council on Foreign Relations University of IllinoisIndiana University University of IowaMichigan State University University of KansasNorthwestern University University of MichiganOhio State University University of MinnesotaPurdue University University of Notre DameSyracuse University University of Pittsburgh
University of Wisconsin
*Participation by Argonne National Laboratory supported by the University of Chicago Board ofGovernors for Argonne National Laboratory and by the U.S. Department of Energy. Opinionsexpressed in this document are those of conference participants and do not necessarily reflect theviews of sponsoring organizations. A~/
OISTIUBUTION OF THIS DOCUMENT IS
Contents
Preface v
I IntroductionJoel A. Snow 1
II Workshop in Economics 5
The Problem of Climate Change Benefit-Cost AnalysisRichard F. Kosobud 5
Comment 1A Note on Benefit-Cost Analysis and the Distribution of Benefits:The Greenhouse EffectKevin G. Quinn 26
Comment 2Issues in Benefit-Cost Analysis: Amplification Channelsand Discounting Long-Term Environmental DamagesDonald A. Hanson 27
III Workshop in Adaption and Mitigation Strategies 29
Approaching Global Warming: A Review of the Adaptationand Mitigation PerspectivesPeter M. Morrisette 29
Comment 1Why Greenhouse Warming Stays a Hot TopicRob Coppock 42
Comment 2Mitigation and AdaptationRob Coppock 44
IV Workshop in Political Institutions 45
Institutional Analysis and Global Climate Change: Design Principlesfor Robust International RegimesMichael McGinnis and Elinor Ostrom 45
Comment 1Meso-Level Regimes and Robust PlansPhilip A. Schrodt 86
Comment 2Principal-Agent RelationshipsDavid Feeny 94
Comment 3Problems of Institutional AnalysisLeonid Hurwicz 95
V General Technical Comments on Climate Change 101
Comment 1Answers to Five Questions Posed at the ConferenceJohn Firor 101
Comment 2Nurturing Multidisciplinary Research on the Global CommonsDavid Feeny 104
Comment 3The Research Agenda: Social Science Research on Global WarmingDavid Feeny 106
Comment 4Policy for Sustainable Development: Some Unresolved IssuesVernon W. Ruttan 108
Comment 5Agricultural Response to Climate ChangeVernon W. Ruttan 112
VI Conference Program 117
VII Conference Participants 119
Preface
On February 11-13, 1992, the MidwestConsortium for International Security Studies(MCISS) and Argonne National Laboratory (ANL)sponsored a conference called Global ClimateChange: Social and Economic Research Issues. Aprogram and list of participants for this conferenceare provided in Sections VI and VII of thisdocument.
The MCISS is an organization composed ofuniversities that support graduate programs ininternational security studies. It was organized in1986 to develop a program of interdisciplinaryworkshops and conferences that would encourageinteractions among participants and foster net-working that could lead to interinstitutional studiesand projects. In carrying out its mission, theMCISS has sponsored workshops and conferences onvarious topics of concern to scholars studyingsecurity issues. These have included a series ofworkshops on various aspects of national securitydecision making in both the United States and theU.S.S.R., on the ethics of low-intensity conflict, onCongress and foreign policy, and on the role ofwomen in national security. Efforts have beenmade to include scholars from fields whose work isrelated to but not traditionally seen as beingconcerned with security issues. These fields includehistory, sociology, anthropology, religious studies,and ethics. In addition, graduate students areincluded as full participants in each activity.
The MCISS is managed by the AmericanAcademy of Arts and Sciences. Since 1987, it hasbeen supported by a generous grant from theJohn D. and Catherine T. MacArthur Foundation.With renewed support from the MacArthur Founda-tion in 1990, the MCISS expanded its field ofconcern to include security in the southeast Asiacontext, the impact of military expenditures andarms acquisitions on Third World security anddevelopment, the moral role of scientists withrespect to the knowledge they generate, theimplications of the restructuring of eastern Europe,and international security and global climatechange. Additional information about MCISSactivities can be obtained from Dr. Marian Rice at
5801 South Kenwood Avenue, Chicago, Illinois60637; her telephone number is (312) 753-8162.
Argonne National Laboratory is a multipurposeU.S. Department of Energy (DOE) research labora-tory engaged in basic research in the physical, life,and environmental sciences. Argonne conductstechnology research in fission, fossil, conservation,and renewable energy. It is operated by theUniversity of Chicago for DOE, which funds about80% of the Laboratory's research and owns thesites. An important part of ANL's mission is todevelop, construct, and operate facilities for use byoutside scientists. Argonne actively encouragesjoint projects and scientist-exchange programs withuniversities and industry, including CooperativeResearch and Development Agreements (CRADAs),to speed the transfer of Laboratory developedtechnology to the public and private sectors.Argonne science and engineering educationprograms provide students at all levels withexperience in current science and technologyresearch.
The participation of ANL in the workshop wassupported primarily by the University of ChicagoBoard of Governors for Argonne National Labora-tory. The Board of Governors includes 21 seniorexecutives and scientists from universities and theprivate sector who oversee the management of ANLunder the chairmanship of the president of theUniversity of Chicago. The contract under whichthe University manages the Laboratory specifiesthat the responsibility of the Board of Governorsextends to cooperative research and educationalprograms between the scientific and technicalcommunity and ANL. To this end, the Boardsponsors various activities that bring Laboratoryand university scientists together to addressproblem areas that are of mutual interest and inwhich future cooperation may be appropriate. Thisconference explored such an area. Additionalinformation about the Board of Governors may beobtained from the University of Chicago, Office ofthe Vice President for Argonne National Laboratory,9700 South Cass Avenue, Argonne, Illinois 60439;the telephone number is (312) 252-2500.
Section I
IntroductionDr. Joel A. Snow
Associate Vice Presidentfor Argonne National Laboratory
University of Chicago
This workshop was designed to bring together agroup of scholars, primarily from the social sciences,to explore research that might help in dealing withglobal climate change. The focus for this effort waschosen on the basis of an earlier workshop thatexamined the potential economic and sociopoliticalconsequences of climate change and the correspond-ing international security implications. In thosediscussions, it became evident that social andeconomic research on climate change necessarilydepends on the evolving understanding of relatedphysical and biological phenomena and can offer atpresent only limited insight for policy. Translatingcomplex and only partially complete geophysicalpredictions into an array of policy options thatmight be useful for national and internationaldecisionmaking is a daunting task that is challeng-ing on methodological, observational, and predictivegrounds. In the interim, social science researchbased on existing data can clarify the merits andlimitations of proposed policy initiatives and providea rough guide to their likely consequences. Suchefforts will require extensive interdisciplinary dialogwithin the research community.
To further that dialog and illustrate the state ofpresent understanding, it seemed useful to focusthis workshop on three broad questions that areinvolved in coping with climate change. These are:
* How can the anticipated economic costs andbenefits of climate change be identified?
* How can the impacts of climate change beadjusted to or avoided?
* What previously studied models are availablefor institutional management of the globalenvironment?
The resulting discussions may (1) identifyworthwhile avenues for further social scienceresearch, (2) help develop feedback for naturalscientists about research information from thisdomain needed by social scientists, and (3) providepolicymakers with the sort of relevant researchinformation from the social science community thatis currently available.
Some general background is needed to provide acontext for the workshop deliberations. Thepossibility that significant changes in climate mightinadvertently result from the introduction of tracegases into the atmosphere by human activity hasbeen known for about 100 years. This possibility isby no means self-evident. Residues from industry,agriculture, transportation, and other activities ofsociety may have severe local or even regionalconsequences, but their impact is usually smallwhen compared with the vastness of the atmosphereor the oceans. Naturally-occurring phenomena,such as changes caused by the orbital motion of theplanet or the interaction of the atmosphere andoceans, involve areas and energy flows that farexceed those affected by man. For many years, thewatchwords for environmental management were to"dilute and disperse" pollutants in the planet'senormous natural reservoirs.
By 1970, systematic measurements ofatmospheric carbon dioxide (CO2) providedconvincing evidence that the concentration of thisgas was cumulatively increasing, roughly propor-tional to worldwide use of fossil energy. Theincrease was estimated to alter slightly the balancebetween the net energy coming inward from the sunand that radiated outward from the earth.Extrapolated over decades or centuries, this shiftwas then predicted to change the climate because ofthe gradual warming of the atmosphere.
Section I
This outline is simple, but the details of boththeory and measurement are complicated. Theapparent connection between fossil energy use andfuture climate change was highlighted at twoscientific workshops — Man's Impact on the GlobalEnvironment (1970) and Man's Impact on Climate(1971). These workshops served as preludes to thewidely attended United Nations Conference on theHuman Environment held in Stockholm in 1972.That conference heightened worldwide awareness ofglobal environmental problems and led to theformation of the United Nations EnvironmentProgramme (UNEP). The so-called "greenhouseeffect" has been on the international environmentalpolicy agenda ever since.
A major international scientific effort to improveman's understanding of climate change and itsorigins was subsequently mounted. Scientists havesought to refine the computational models thatdescribe the effects of CO2 and other radiativelyactive gases, to understand the global carbon cycle,and to expand the observational database. A majorscientific conference was even devoted solely to theproblem of first detection; that is, how to distin-guish the greenhouse signal from the noise causedby natural climate variability. Unfortunately,nobody has a climate meter, so climate must beestimated from weather measurements. The dataare still so obscure that no complete consensusexists on whether a distinguishable trend due togreenhouse gases has been observed, despite theconfident assertions of some scientists, which arebased on the recent series of warm years.
By the mid 1980s, a series of state-of-the-artassessments was included in the research programof the U.S. Department of Energy in an attempt tosynthesize the burgeoning research literature. Morerecently, a massive multinational effort called theIntergovernmental Panel on Climate Change (IPCC)has been initiated. The IPCC is a joint effort ofUNEP and the World Meteorological Organization(WMO) that has pulled together researchers andresearch results on climate change from across theglobe. Participants include government officials aswell as practicing scientists. Several hundredexpert contributors and reviewers drawn frominstitutions worldwide participated in the IPCCscientific assessment. The work of the IPCC hasprovided input to serious discussions about limitinggreenhouse gas emissions that have been held in
meetings leading to the 1992 United NationsConference on the Environment and Development(UNCED).
The continuing flow of new scientific informa-tion can affect the logic connecting greenhouse gasconcentrations and climate impacts. For example,consider some of the research reported on in thefirst four months of 1992:
* Data from satellite measurements of cloudcover that have become available allowestimates to be made of the effect of varioustypes of clouds at different heights on the netradiative balance.
* Data on the growth or retreat of glaciers,polar sea ice, and other ice masses have beenreported on recently, sometimes withinconsistent and conflicting results.
* Speculation continues about the mechanismthat removes from the atmosphereapproximately half of the CO2 introducedeach year by fuel combustion. Recentarguments have attributed almost half of themissing CO2 to uptake by tbs ocean.
* Changes in the salinity of the ocean arehypothesized to have cooled the northernhemisphere in recent decades, possiblymasking greenhouse warming.
The IPCC scientific assessment outlines fourmajor areas of scientific uncertainty in climateprediction: clouds, ocean processes, greenhouse gasuptake and transformation, and behavior of thepolar ice sheets. It also calls for and describes anextensive program of needed research.
In view of this extensive scientific effort and the20-year history of interest in international regimesto control greenhouse gas emissions, it is quitesurprising that more effort has not been expendedon a systematic analysis of the national andregional impacts of climate change and on the costsand benefits of various proposed control measures.Discussions of research that should be done bysocial scientists are included in recent (1991)studies published under the auspices of theInternational Social Science Council (A Frameworkfor Research on the Human Dimensions of Global
Introduction
Environmental Change) and the National Academyof Sciences (Global Environmental Change:Understanding the Human Dimension). Sources ofsignificant funding for such research have not yetemerged.
The complexity of the greenhouse problem maywell defy detailed rational analysis. A comprehen-sive model would require coupling a geophysicalmode! of oceans and atmospheres to a socioeconomicmodel describing energy, other greenhouse gassources, population, economic change, and possibly avariety of other social parameters.
As a practical matter, the geophysical modelsare very complex and require large amounts ofcomputer time, and they still do not presentlyinclude all the relevant physics. Nor do the modelsproduce results at spatial and time scales adequateto satisfy those who need a geo-based data systemthat can usefully distinguish positive and negativeimpacts. Also, geophysical predictions generallyavailable usually provide equilibrium values ofclimate variables but do not describe transientphenomena. For example, moderately reducedaverage precipitation, considered alone, might notbe very significant. If accompanied by markedseasonal asymmetry, it could lead to severe damageto the agricultural sector. Indications thatincreased average daily summer temperaturesresult primarily from higher daily night-time lowshave significance for both food production and peakelectricity demand. Furthermore, increases inturbulent phenomena, such as severe storms, couldbe of more local significance than are seasonaltemperature increases.
Underlying concerns about the validity ofclimate models remain. For example,Dr. Paul Keaney, a participant in the 1991workshop Global Climate Change and InternationalSecurity, has observed:
From 900-1000 A.D., the average globaltemperature was 1°C higher than now (andIceland's colonists found it wooded, withrelatively ice-free harbors). The reasonswhy are not well understood, but the notionthat a simulation can "back-cast" theconsequences of earth-orbit variations, evenwhen the very complicated orbital dynamicsare not included as model input parameters,
is remarkable. The notion that thetemperature rise and fall before and after1000 A.D. was due to variations in CO2concentration, and that such variation mightbe due to human activity, would be equallyremarkable.
Given the uncertain availability of data aboutthe timing and extent of climate change, theamount of confidence that can be attributed toimpact estimates is unclear. How urgent globalwarming should be on the agenda of internationalissues is also uncertain. Many believe thatpopulation growth — which drives near-term needsfor food, energy, resources, and economicdevelopment — is far more central to environmentaldegradation than global warming. Local armedconflicts may take a far higher toll in human life.
At this point, it is reasonable to ask whetherany attempt to undertake major policy initiatives tolimit greenhouse gas emissions makes sense.Nevertheless, there seems to be widespreadagreement on the following issues:
• On a relatively short time scale, animportant geophysical parameter (i.e., theearth's radiative balance) is being altered —not by nature but by man.
* The resulting changes could have profoundimpacts on humans and other components ofthe biosphere.
* These changes could be irreversible on anytime scale deemed reasonable for concertedhuman response.
• The likelihood of soon obtaining' thecapability to make truly reliable andbelievable predictions of the impacts andeffects is small.
The need to act may simply outrun the quantitativebasis for doing so.
In response to this political imperative, theremay be worthwhile strategies that are relativelyinsensitive to detailed geophysical prediction orquantitative assessments of benefits and costs.Such an approach is implied in the recent NationalAcademy of Sciences study, Polity Implications of
Section I
Greenhouse Warming, and may also serve otherdesirable environmental or social ends. Forexample, in the nearer term, growth in greenhouseemissions may be temporarily curtailed throughsubsidizing (or mandating) improved energy end-useefficiency and use of carbon-free fuels. Marketincentives that limit tropical deforestation might befound. Substitutions for energy use, such astelecommunication for transportation, might beencouraged. Each nation might be allowed tochoose its own way to meet a greenhouse gasceiling. (Recent press reports indicate thattechnical experts claim that temporary caps onemissions are feasible, even in the United States).Various international bodies or tribunals that tradeglobal warming incentives for other values might beintroduced — a policy that would be equivalent, inprinciple, to the approach established under theClean Air Act.
Part of the reason for the interest in thegreenhouse effect is that it serves as a surrogate fora whole class of problems that involve internationalcooperation, including the sharing of resources, arespect for equity and integrity, and combining theefforts of many toward a just result.
Even though an elaboration of such examples isnot provided here, it should be evident that researchis quite urgently needed, to both develop the detailsand explore the designs and consequences ofpolitical approaches that are motivated, in part, bythe broad implications of greenhouse warming.This workshop, in part, addressed that need.
The approach of the workshop was to presentcore working papers that address the themes ofeconomics, amelioration, and institutionalmanagement. First, however, an introductoryframework discussion on aspects of knowledge aboutclimate change (not included here) was led byDr. John Eddy, formerly of the UniversityCorporation for Atmospheric Research and now withthe Consortium for International Earth ScienceInformation Network. Then Dr. Richard Kosobud of
the University of Illinois at Chicago addressedeconomic issues. His application of benefit-costarguments to the Great Lakes region provided anexample of how to estimate distributed costs andbenefits. Next, Dr. Peter Morrisette reviewedadaptation and mitigation strategies, payingparticular attention to the different conceptualpostures that these approaches entail. ThenDrs. Michael McGinnis and Elinor Ostromexamined institutional issues, bringing forth asubstantial body of research on ways in which othercommon-use resources have been managed histori-cally, sometimes over centuries. Commentatorsundertook to discuss each paper as the first step ina candid and free-flowing general discussion of theconcepts that were presented. Each paper was thenrevised by the author to reflect those criticismsdeemed most cogent. Workshop participants werealso invited to contribute written comments, eitheron a particular paper or on a more general topicthat had not been otherwise addressed. The nextthree sections (II-IV) of this document contain therevised core papers, with appropriate writtencomments appended. Section V contains observa-tions and comments of a more general nature. Drs.Marian Rice, Harold Jacobson, and I served as aninformal board of editors to assemble the final text.
This workshop and these proceedings could nothave been completed without the dedicatedassistance of many individuals, includingDrs. Harvey Drucker and Terry Surles, of ArgonneNational Laboratory management, and Dr. DonHanson, whose group at Argonne served as theworkshop host and who personally oversaw theproduction of this document. Financial support wasprovided to the Midwest Consortium for Inter-national Security Studies by the MacArthurFoundation and to Argonne by the University ofChicago's Board of Governors for Argonne. Thefinal editorial and production effort was undertakenby Ms. Marita Moniger of Argonne's Informationand Publishing Division, who did a superb job underoften challenging circumstances.
Section II
Workshop in EconomicsThe Problem of Climate Change
Benefit-Cost AnalysisProfessor Richard F. Kosobud
University of Illinois at ChicagoChicago, Illinois
1 INTRODUCTION
This paper asks the question: Could benefit-costanalysis, the tool used in mainstream economics toinvestigate control of environmental externalities,play a larger role in the discussion of policies todeal with the greenhouse effect? For example, couldeconomic calculations have played a larger role inthe discussions that have centered around the 1992United Nations Conference on Environment andDevelopment (UNCED). The paper alsoinvestigates the causes of this lack of influence. Isit a result of too little research effort? Resistanceby interest groups? Methodological flaws? In thispaper, I probe selected forms of benefit-costresearch, particularly the critical discussions raisedby this type of research, in an effort to suggestwhere the chances of greater acceptance lie.
The paper begins by discussing the search foran appropriate policy: optimal, targeted, orincremental. It then describes the work being donein specifying and estimating climate change damagerelationships. A consideration of the work beingdone in specifying and estimating abatement (bothmitigation and adaptation) cost relationshipsfollows. Finally, the paper ends with an examina-tion of the search for the appropriate policyinstrument. International and methodologicalconcerns cut across these areas and are discussed ineach section. The usual introductory explanation ofthe greenhouse effect is replaced by relevantbackground information at appropriate places.
This paper concludes that there seem to be anumber of reasons that benefit-cost results playonly a limited role in policy development. A latestart and resistance to action in an uncertainproblem area are among them. There is an
unresolved tension between a focused, narrow,quantitative interpretation and a broad inter-pretation of the benefits resulting from reducing theamount of climate change damage and of the costsresulting from abating trace gas emissions. Anarrow interpretation is open to the charge ofneglecting many more qualitative aspects, and abroad interpretation is open to the charge of addingnumerous value judgments to the analysis. Thechance that more attention will be paid to economiccalculations — even if there is no new scientificevidence on the significance of global warming —should aol be underrated. There is some evidencethat the growing interest in market-basedapproaches to climate change policy and to otherenvironmental control matters is a sign of increasedacceptance. Suggestions about research directionsare made throughout this paper.
2 OPTIMAL, TARGETED, ORINCREMENTAL POLICY?
What matters for anthropogenic climate changeis the worldwide atmospheric concentration ofgreenhouse gases, not the distribution of theiremissions by country or region. (Qualification ofthis statement is required for several gases that candiffer in regional concentration; tropospheric ozoneand methane are examples. The former drifts farfrom its metropolitan-area origin, and the latter'sconcentration differs little by area, so thequalifications seem minor in a quantitative sense.)What matters for the associated damages is thedistribution of climate change — temperature,precipitation, storm frequency, and the like — bycountry and region. For countries engaged in policydeliberations, it is soon apparent that no onecountry can wisely act alone without considering
Section II
how others will act; therefore, opportunities forcoordination as well as for hard complications arisein the search for an optimal policy.
2.1 Optimal Policy
Efforts at estimating an optimal policy such asthose begun by Nordhaus (1991) provide an aptillustration of the difficulties, and light, suchresearch can provide. For a market externality,specification and estimation of the relationshipsbetween social damage and abatement cost arerequired; the point at which the marginal gains ofdamage reduction equal the marginal costs ofemission abatement determines the optimal degreeof emission reduction (or concentrations) and theoptimal tax or emission permit value. Figure 1 isadapted from the Nordhaus study and shows themarginal damage and CO2-equivalent cost functionsfor greenhouse gas (GHG) emissions.
As Figure 1 indicates, if all countries were toagree to a Pigouvian tax on trace gases of $3-$13(1980 dollars) on the basis of their CO2-equivalentcarbon content (e.g., $0.30-$1.30 per barrel ofstandard crude oil), the external costs of climatechange would be internalized.1 This tax ratewould reduce global CO2-equivalent emissions byabout 15% and specific CO2 emissions by a fewpercentage points.
Given certain conditions (to be discussedpresently) and the imposition of the optimal carbontax, there will exist at least one set of economicallymeaningful non-negative prices at which producersand consumers will make consistent allocationdecisions. Furthermore, at these equilibrium prices,no reallocation that all households will prefer willbe possible; at least one household will object.Finally, if die reader prefers a different globalPareto-efficient allocation and can alter thedistribution of goods, there is an equilibrium set ofprices to give this preferred allocation. Theunplanned or decentralized economy has led itsparticipants, in theory, to the desired optimal end,where the environmental externality has beeninternalized.
Before one discards the results obtainable fromthe fundamental theorems of welfare economics, itis worth a pause to admire their depth andappreciate their value as a framework. Providinguseful information for policy requires, however, aclose examination of the phrase "given certainconditions."
Parry (1991) addressed the question of whetherthe absence of uncertainty in the neoclassicalgrowth model used by Nordhaus made the optimalpolicy too conservative; that is, would not the risksof high damages, irreversible impacts of concentra-tions, and adjustment costs justify a more stringent
All Greenhouse Gases
Medium Damage
Low Damage
10 20 30Percentage Reduction of Total Greenhouse Gases
40
FIGURE 1 Efficient Greenhouse Gas Reduction (Source: Figure 9 in Nordhaus 1991)
Economics
policy? By introducing a subjective probabilitydistribution over damages and by introducinggrowth in trace gas concentrations, Parry usedNordhaus's model so modified to reach theconclusion that society must be risk averse towarrant more action now (a higher tax rate) than inthe calculations of Nordhaus.
Ayres and Walter (1991) examined another ofthe given conditions, the range of damage estimatesused by Nordhaus. Damage had been estimated byusing U.S. data on climate-sensitive industriesoperating under the climate change conditionsinduced by a doubling of current CO2-equivalentconcentrations and by then projecting the estimatesonto the world stage. Ayres and Walter argued thatthere would be greater impacts from, a rise in sealevel, that land in densely settled countries shouldbe more valuable, and that resettlement of apossible 100 million refugees would be expensive.They increased the tax rate from $3 to $30, whichimplies a reduction in emissions of about 23% (seeFigure 1).
It seems clear that the burden of gatheringinformation and taking measurements necessitatedby relaxing only two of the given conditions orassumptions required in the search for optimalpolicy will be staggering. Moreover, Baumol andOates (1988) have pointed out that nonconvexitiesin the production set introduced by externalities(such as the greenhouse effects by which the outputof one industry can transmit, through nonpriceeffects, adverse impacts on other industries) canmake the optimality of estimated tax rates suspect.Convexity of preferences cannot ensure that taxrates can move the economy to a preferred positionin these situations.
The search for an optimal policy encountersfurther, and perhaps graver, difficulties when it isplaced on the international agenda. For example,proposing a global carbon or CO2-equivalent taxraises issues in international tax administration,compliance, and resource transfer that aredaunting. Varying departures from competitivemarkets creates well-known problems, includingefficiency losses. The existing tax systems ofvarious countries also differ, a fact that createsstrategic opportunities for some countries to gainadvantage over others; the existing light taxation offossil fuels in some countries could enable them to
lower the gross efficiency costs of a common carbontax. Revenues from a carbon tax, if used to reducepersonal or corporate tax rates, present anothersimilar opportunity (Goulder 1991).
The deep attraction of the optimal policy searchis in its theoretical foundation: It creates incentivesfor decision makers on microeconomic issues to dowhat they do best — make cost-effective input andconsumption choices with all the information attheir disposal — and it allows decision makers onmacroeconomic issues to decide on environmentalgoals. The value of benefit-cost analysis as a bench-mark toward which policymakers may aim shouldnot be underestimated. An underlying flaw couldwell be that economists appear to be preempting thepolitical role with their calculations.
One alternative is to recommend that standardsor targets for climate change policy be establishedby the political process, while the cost-effectiveattainment of these targets be achieved throughmarket-based approaches, wherever feasible(Baumol and Oates 1988).
2.2 Targeted Policy
There are costs in giving up the pursuit: Noneof the problems are resolved by substituting targetsfor optimal goals, and the standards of scientificcritique are weakened in appraising targets. Thebasis for targets becomes more or less arbitrary inan economic or scientific sense, although possiblymore attainable in a political sense. Targets are thestuff of political careers, and a number of targets interms of GHG emission reductions have been placedon the international agenda by various govern-ments.
The International Energy Agency (IEA)publishes and updates a useful review entitledClimate Change Policy Initiatives, whichsummarizes the commitments and actions ofmembers of the Organization for Economic Coopera-tion and Development (OECD). The review revealedin November 1991 that 24 countries had made acommitment to climate change policy, all commit-ments were in the form of emission targets. Fourcountries had passed new fossil fuel taxes since thestart of 1990, but there were important exceptionsin export industries (IEA 1991, Tables 6 and 7).
Section II
Twenty of the countries established CO2emission targets alone. Most called for stabilizationat 1990 levels by 2000 or soon thereafter, one calledfor a 3-5% emission reduction, three called for a20% reduction, and two called for a 25% reduction.Four countries, including the United States, setaggregate GHG targets that were to be eitherstabilized or reduced on a CO2-equivalent basis; forthe United States, this meant a reduction in theradiatively potent chlorofluorocarbons (CFCs).Many of these commitments were made conditionalon other countries taking action.
A danger of seeking international approval bymaking targeted commitments detached from imple-mentation proposals or abatement cost calculationshas been pointed out by Schelling (1991). Thedanger is that the target is not taken seriously.This problem was brought out in the recent work ofthe Energy Modeling Forum (EMF 1991), whichasked 12 builders of energy-environment-economicmodels to run several of these targeted emissionscenarios through their models to estimate the taxrate required to achieve these reductions.
The carbon tax rate required to stabilize currentCO2 emissions by the. sar 2000 implied, on theaverage, a tax of about $50 per ton of coal and ofabout $15 per barrel of standard crude (all in 1990dollars). In other words, tax rates about equal tocurrent prices would be required for stabilization,and a doubling of these rates would be needed for a20% reduction in emissions. Note that none ofthese scenarios stabilized anthropogenic climatechange; they permitted a continuing, althoughslowed, buildup of atmospheric trace gasconcentrations.
These tax rates are high, perhaps beyond therealm of political feasibility. A full appraisal ofthese complex models and their outputs is a vastundertaking; the expected 1992 EMF report thatwill compare models when comparable futurescenarios are imposed on them will be a step in thisdirection. Among the 12 models, the optimizationmodels, which contain the most detailed economicspecifications, indicate that targets that constrainemissions in the near term will require tax rates ofthe magnitude just indicated or higher. Thisconclusion accords with one's common-sense viewthat imposing short-run adjustment costs on thefossil fuel and other industries will be expensive.
If damage estimates, uncertain as they are, donot seem to justify these short-term costs, is itpossible to establish longer-run targets thatgradually become more constraining and that callfor policy actions well beyond the next election? Orto establish sequences of policy steps that escalateover time and commit future governments? Doesthe danger of proposing targets that secure thepopular approval at the negotiating table but havelittle chance of fulfillment undermine effectiveclimate change control? Do these dangers direct ourattention to other forms for expressing policycontrol?
2.3 Incremental Policy
A more engineering type of approach to policyobjectives was proposed in a National Academy ofSciences study (1991) on mitigation options. Somedefinitions may be in order. Mitigation optionscover efforts to reduce gas emissions or enhancesinks; adaptation options cover efforts to adjust tothem; and costs incurred in either could be includedin abatement costs.
In this approach, an array is drawn up of newenergy-efficient technologies, techniques, orpractices that could rep'ace existing ways and hencereduce trace gas emissions at little or no cost.Sometimes a negative cost is claimed. The idea,therefore, is to place this array on the internationalagenda to see if agreement can be reached amongall countries to go some way down the list or to seeif pledges can be made to undertake some of theitems in the expectation that other countries willundertake their own — each effort, perhaps,requiring equal sacrifice or achieving someproportionate emission reduction.
No precise targets need be defined. Someagreement to monitor efforts and to ascertain theeffect of efforts on the buildup of concentrationswould appear to be important. A variety of policyinstruments appropriate to each instance could bechosen and used by participating governments.Illustrative examples adapted from the NAS reportare provided in Figure 2 and Table 1.
This approach makes use of the series of casestudies described as "bottom-up" in the energy con-servation literature. Mitigation Method 1 listed in
Economics
CM
8
10080604020
0.
-20
-40
-60
-80
-100
-
-
-
2 i
i i ;
•r"
3_
6 7
4j
9
1 High Cost Estimate
10
• - - - - Mid Cost Estimate
Low Cost Estimate
2 4 6Billion Tons CO2 EquiValent/Yaar lOOXAnwalUS.CO,
Equivalent Emission*
FIGURE 2 Bottom-Up Mitigation Cost Curve (Source: Based onFigure 11.1 in NAS 1991)
TABLE 1 Costs and Emission-Reduction Potentials of Various Mitigation Methods
12345678910
Mitigation Method
Resid. & comm. energy efficiency*Vehicle efficiency (no fleet change)*Industrial electric efficiency*Transportation system management11
Power plant heat rate improvements*Landfill gas collection*Halocarbonsb
Agriculture1"Reforestation11
Electric supply1"
NetImplementation
Cost (I/ton of CO2
Lo
-78-75-51-50
-20.40.9135
equivalent)
Mid
•62-40-25-22
01137
45
Hi
-47-2152135
1080
Max. PotentialEmission Red.(109ofCO2-
equiv^r)
0.90,30.50.050.050.21.40.20.21.0
PercentReduction in U.S.
CO2
(4%)
186
11115
2955
21
Emissions
CO2-Equiv.(%)
1147113
1833
13
"Category 1 includes best practice options attainable at negative or low cost.
bCategory 2 options are attainable at additional costs.
Section II
Table 1, for example, includes data from insulation,building code, heating system, and lighting studiesof residential and commercial buildings. The netimplementation costs are negative, indicating thatpresent practices in these sectors are not efficient;the market is in disequilibrium in this sector. Themitigation cost curve in Figure 2 aggregated overthese methods starts below the abscissa and rises topositive costs after reductions of 2 billion tons inCO2-equivalent emissions are achieved (25% of U.S.1988 CO2-equivalent emissions). This mitigationcost curve should be compared with the risingabatement cost curve of Figure 1, which is always inthe first quadrant (or positive) and assumes choicesare optimal before abatement occurs. Neither curvereflects the external damages of climate change.
The controversy over this approach is by nomeans assuaged by the issuance of a prestigiousacademy report. Questions have been raised abouthidden costs not accounted for in the curve andabout the incentives required to install thoseenergy-conserving technologies that are not beingundertaken at present.
At issue in these debates is the importance ofefficiency standards, education, and informationalprograms when compared with economic incentivesprovided by taxes or emission permits. To theextent that market imperfections are barriers toenergy-efficient innovations, economic incentives areless effective than command-and-control regulation.Energy conservationists argue: "Having the rightprice is helpful but not sufficient to achieve optimalefficiency improvements in the future" (Geller et al.1992).
Addressing climate change policy incrementallyappears to involve a long series of steps to reduceemissions by energy-efficient methods, many ofwhich are claimed to have negative implementationcosts. These steps would apparently require anextensive kit of policy tools to alter perceivedmarket imperfections. The following questionsarise. How can tests be constructed to determinethe actual effectiveness of particular steps? Todetermine their full implementation costs? Howshould such efforts in die international area benegotiated and monitored? Should credit be givento a regulatory measure for an innovation or to acorrelated OPEC price shock? Can ideas about
market imperfections giving rise to energyinefficiencies be formulated as testable hypotheses?
2.4 Summary Comments
The search for an optimal climate change policyby means of benefit-cost analysis is only one of theflickering candles in a very dark room. To blow itout because of its recognized limitations wouldheighten the risks of stumbling seriously on thepath to sensible policy. Yet it cannot be denied thatimportant problems remain to be resolved.
The optimal tax rate calculated by Nordhausand the emission stabilization or reductions calledfor in current proposals do not stabilize trace gasconcentrations (and hence temperature change);they slow them down. Is that what policy shouldaim at? Given the fact that there are fossil fuels inthe ground in recoverable amounts that would morethan quadruple CO2 atmospheric concentrations,should policy analysis provide guidance on a long-run goal of temperature stabilization?
The matter is complex, but the difference inconsequences between short-run and long-run policyimplementation was brought out recently in severalexperiments. Figure 3 was drawn from researchcarried out at Argonne and the University of Illinoisat Chicago. The top panel, Panel a, shows abusiness-as-usual (no greenhouse policy) globalprojection for CO2 that more than doubles concen-trations by the year 2100. On the basis of amathematical programming model that permitsconstraints to be placed on emissions andconcentrations and consequently fossil fuel use,Scenario A' was run holding concentrations constantafter 2100. A different scenario, Scenario A, wasrun holding annual emissions constant after 2000,as so many proposals advocate. Note that bothScenarios A and A' result in the same CO2concentration near 2100. In A, but not in A',concentrations continue to build up after 2100.
The implications of these concentrationscenarios for emissions, the rate of change ofconcentrations, is revealed in the middle panel,Panel b. Scenario A' emissions follow the business-as-usual emissions closely up to the year 2030,which implies relatively little impact on the fossil
10
Economics
IMS 2000 2CM 20M 2100 2130
g »•« •
, .
IMO 2000 2040 2M0 20M 2100 2120
2S0-
200-
iso-
fci
mo 20N 20M 21C0 2129
FIGURE 3 Mathematical ProgrammingProjections of CO2 Concentrations, Emissions,and Permit Prices (Source: Kosobud et al.1991)
fuel industry. Scenario A' emissions remain abovethose of Scenario A until 2060, then they diminishto the rate that stabilizes concentrations andultimately average global temperature.
The bottom panel, Panel c, is a preliminaryexploration of the policy consequences of aiming forthese two different objectives. The programmingmodel enables the researcher to interpret theshadow prices of constraints on CO2 emissions as
either tax rates in a competitive world or,equivalently, emission permit prices in a world ofcertainty. To accomplish the goal of stabilizingemissions by 2000, a carbon tax that works out tobe more than the current price of oil or coal (U.S.prices) is required. The rate declines a bit in thenext century as the economy adjusts to the shock.To accomplish the goal of stabilizing concentrationsby 2100 but permitting the economy to achieve thatgoal in an optimal way (by emitters equatingmarginal abatement costs over time), a tax rate, orpermit price, that starts low and rises with the rateof interest (5% in this experiment) is required.
Scenario A' would be preferred and welfaresuperior, providing that long-run concentrationswere the major factor in induced climate change. Ifthe rate of change of concentrations or total damageborne were the major factors, Scenario A could bepreferred. An important question for atmosphericscientists to address is the following: Which featureof trace gas buildup is most important in inducingclimate change: the level of concentrations, the rateof change, or the sum of past concentrations? Thereis little evidence that starting with modest policymeasures such as those in Scenario A' that escalateover time exacts a penalty in climate change. Thefollow-up question for social scientists is this: Whatfeatures of climate change have the largest damageimpacts?
Another key issue concerns the cooperation orconflict that has, and will, ensue in internationalnegotiations over a coordinated policy, if one isattainable. .Developing countries have argued thatthe developed countries have caused the problem, atleast to the present time, and ought to bear theheaviest burdens. If that argument does not carryenough weight, developing countries can threatenthat in the future, with economic growth, they willbe major emitters, with the upper hand inbargaining. Something like this may be in theminds of those Chinese economists of the PeoplesRepublic of China whom I have met and whoprofess a great interest in the greenhouse topic.
Yet the damage estimates that are beginning toaccumulate indicate that there will be wide regionaldifferences, resulting in winners and losers, andthat some developing countries and not thedeveloped countries will be among the potential big
O** losers. Changes in sea level are likely to threaten
11
Section II
Indonesia and Bangladesh, which have long andoccasionally low-lying coastlines. Agriculturalimpacts are likely to be more important indeveloping economies. In sum, the bargaining chipsmay alter in the future. Policy research in this areahas just begun.
3 THE SEARCH FOR CLIMATE CHANGEDAMAGE RELATIONSHIPS
The benefit-cost method has its own scissorsanalogy: The blade of the damage reduction mustfit with the blade of the abatement costs in order tocut the externality knot. The difficulty has beenthat in one environmental study after another,economists, with their measurement practices, haveestimated damages that seem to warrant little, ifany, policy action. There is a puny or no damageblade at all.2 Many environmentalists haveprotested that the estimates are too narrowlyfocused and too small; many affected industrygroups have protested that the estimates are toolarge. This issue is emerging, among others, in thepolicy analysis of the impacts of the greenhouseeffect.
Specifying the relationship seems straight-forward; estimating it is not so easy. In thissection, I examine the top-down and bottom-upapproaches to specification and the difficultiesencountered in obtaining estimates with both.Several specific estimates will be presented fordiscussion, one being the result of an effort toestimate damages for the Great Lakes region. It is,of course, self-serving for a researcher to call formore studies, but in this area, too long under-funded, I make the call in good conscience.
3.1 An Economic View of Impacts
In principle it seems simple: All one has to dois specify the micro-production functions andconsumer utility functions and include in them avector of climate variables. Increased trace gasconcentrations affect the climate vector, which thenmay or may not affect producer outputs andconsumer welfare. To the extent that climate hasan impact, one obtains the externality of a reductionin output or a reduction in the satisfaction of a
producer or consumer, independent of his or herinput or shopping decisions.
Admittedly the above framework does not tell uswhich producers or consumers are sensitive toclimate variables or how exactly to go aboutmeasuring the impacts. It is, however, suggestivein pointing toward an ordering of activities,industries, and systems by their climate sensitivity.It also suggests that decision makers on micro-economic issues may be able to find adaptivestrategies to lessen these impacts once they areaware of their coming. Finally, this frameworkplainly indicates that efforts to adapt to or mitigateclimate change will have costs, so that reducingsuch change to zero is unlikely to be optimal.
3.2 Bottom-Up or Top-Down Specification?
This distinction is becoming a battle cry amongclimate change researchers, but in this application,few blows need be struck. Top-down refers to a veryaggregate damage specification that some buildersof energy-economic-environment models have usedto study optimal tax or permit pricing under alter-native scenarios. Work in this vein is underway atthe Electric Power Research Institute (EPRI).Nordhaus has specified a linear relationship for theArgonne linear programming model, on whichresearch is now underway.
These damage equations are specified asconstraints in these models, so that activities givingrise to trace gas emissions (fossil fuel use giving riseto CO2 emissions in the two models cited) arecharged for the damages. These charges raise theprices of fossil fuels, for example, to a point in eachtime interval at which the marginal damagesgenerated by the damage equation are equated withthe marginal abatement costs of reducing, orsubstituting away from, fossil fuel use. The chargeis a shadow price, tax rate, or permit price of anoptimal policy path, providing one believes thedamage relationship.
The equations are highly aggregated and meantto be suggestive; they typically make a discounteddamage scalar depend over time on global orregional gross national product (GNP) per capita,the change in CO2 concentration (or temperature),
12
Economics
and perhaps a separate population variable. GNP,like population, is a proxy for vulnerability, andCO2 concentration change (or temperature change)is a proxy for the vector of climate variablesimpacted by the greenhouse effect.
The IPCC (1990b), adopting a wider frame ofreference for climate sensitivity than theeconomist's usual view, prepared a list that beganwith agriculture and forestry, then included naturalterrestrial ecosystems, hydrology and waterresources, human settlements, energy, transport,human health, air quality, oceans and coastal zones,and the terrestrial cryosphere. The IPCC did notattach monetary values to damages in each area,but the panel did survey the qualitative direction ofchange.
For agriculture, a key sensitive sector, thefindings of studies to date are complex and indicatethere will be winners and losers. Higher levels ofatmospheric CO2 may enhance the growth rate ofcereals such as wheat and rice but not of sorghum,millet, or maize. Use of water by crop plants maybe more efficient. How much these direct effectswill be felt in the farmer's field remains an openquestion. Negative impacts could occur at regionallevels since these CO2 impacts vary by plant typeand are decidedly nonlinear. Other weatherchanges could bring about reduced yields, stimulatenew or more virulent diseases, encourage pests, andfavor weeds.
Another key area concerns impacts on theworld's oceans and coastal zones: a rise in sea levelof 20-30 cm (about a foot expected to occur by 2050or so) will pose problems for low-lying islandcountries and coastal zones. Fresh water systemscan be adversely affected by climate changes; hotterand dryer conditions in the Great Lakes area couldreduce the levels of the freshwater lakes, affectingshipping, fisheries, and coast lines.3
3.3 Damage Estimates for theGreat Lakes Region
The most courageous (foolhardy?) way to presentthe difficulties of estimating damages is to displaysome of my own. In Table 2,1 provide estimates Irecently drew up for the Great Lakes region bydrawing on existing studies. The damage
categories list the activities and sectors I judged tobe the most sensitive in the five-state area (Illinois,Indiana, Michigan, Wisconsin, and Ohio).4
The estimation methods of Table 2 arestraightforward but based on strong assumptions.The economy of the region is projected to be biggerin 2050, but grows more slowly than in the past 50years. It has, however, the same sectoral structure,which means that agricultural and hydrologicaleconomic activities retain their relative importancein output The climate of the region is projected tochange by 2050, when global CO2-equivalent tracegas concentrations about double (assuming nocontrol policies are undertaken). The regionalclimate change is the average of the temperatureand precipitation changes reported by the EPA fromits survey of the projections of three generalcirculation models; in particular, precipitation isprojected to decline.
I have attempted to be more inclusive in Table 2than some studies and have listed a few impacts tothe health and leisure sectors in addition to thetraditional sectors that are considered climatesensitive. This attempt is meant to be suggestive,not exhaustive. I have drawn on other studies formonetary estimates that were then applied to theregion. (These other studies are the references Rl,R2,..., Rn noted in the table; they are availablefrom the author on request.)
Agriculture is a big item and corn is a bigcomponent in the region. Favorable effects ofphotosynthesis and water efficiency are more thanoffset by temperature and precipitation changes.The range, which may be too small, attempts tocapture the possible consequences of unfavorableweather during critical growing and harvestingperiods. Promising research is underway in thisarea, and new information should make theseestimates obsolete soon, if not now.
One of the interesting sensitivities of the regionis the hydrological system, including the famousjewels of the area. The reference cited, and others,project lower lake levels because thermal expansionwill be more than offset by precipitation declinesand evaporation. One study the author is currentlyworking on finds that when the cool winds of falland winter blow over the waters still warm from thesummer, the evaporation volumes are enormous.
13
Section
TABLE 2 Greenhouse Effect Damage Estimates for theGreat Lakes Region (GLR)
Damage Category
General
HealthUrban smog related (R7, Sec. E, p. 5)Heat stressCold stress
Climate variability increaseStorm frequency (R4, p. 33)
Leisure activitiesGolfing (R2, p. 3)Lawn watering (R2, p. 5)
Biological diversity in the GLR
Sectors
Annual Estimate(109 1982$)
Low High
-0.012 -0.030NANA
-0.009 -0.009
+0.010 +0.020-0.010 -0.025
NA
Agriculture (R3, p. 10-6)Corn, soybeans, wheat -2.950 -2.950
Hydrological systems (R3, pp. 5-7, 5-23)Water transportationHydroelectric powerShoreline maintenanceRecreationWater quality maintenance
Energy (R3, p. 5-53)Forestry (Rl, Table 6-4, p. 215)(R3, p. 546)ConstructionGovernment service
Totals
GLR percent of 1982 GNP
-0.010-0.002-0.012
NA-0.018
-0.297-0.150
NANA
3.460
0.6%
-0.010-0.002-0.024
-0.018
-0.297-0.350
10.825
2%
Source: Kosobud (1990).
14
Economics
Declines from a few inches up to several feet havebeen projected by 2050 or soon thereafter, withimplications for water transportation, shorelinemaintenance, fisheries, fresh water supplies, andthe like.
The energy sector will have opposing factors toconsider — the increased demands for air condition-ing offset by the decreased demands for heating.Forestry is of interest; the present furniture andvalue-added processing industries possible becauseof the quality of the trees in Wisconsin andMichigan could suffer if these trees fail to adjustand new, less valuable species replace them.
My data are intended less to persuade and moreto reveal the complexities of damage estimates andthe need for expanded work in this area (even atthe expense of reduction in certain other areas).Conclusions are in order from this effort: a range ofabout $3.5 to $11 billion in damages is obtained, or0.6% to 2% of 1982 GNP. These estimates are notall that frightening and lie between the estimates ofNordhaus for the larger U.S. economy and Ayresand Walter, who gave more emphasis to developingcountries.
My estimates, however, can be attacked as toohigh or too low, and there is merit to both argu-ments. On the high side, the point can be madethat I have failed to allow for adaptive behavior inagriculture and other sectors. A sensible farmer,seeing climate change coming (reading theproceedings of this conference for example), wouldbegin to think of new seeds, new cultivationpractices, and other innovations that would lessenthe damages.
Also on the high side, it may be argued that theeconomy of the region will look a lot different in2050 than it does today, including a larger servicesector and many more activities less sensitive toclimate change, like this conference.
On the low side, the point can be made thatthere should be more categories in the table(ecosystems, for example) and that within existingcategories, there should be many more items.Furthermore, the estimates are based on averageclimate changes for just a few variables. Whatabout storm frequency, wind velocities, strings ofhot, dry days, etc.? Perhaps even more significant,
what about the rate of change of these variables?Adjusting to a slowly changing climate (even thoughchanges over 100 years may seem slow, they areenormously more rapid than in prior known history)is one thing. What if the rate of change of climatevariables over the next 20 to 50 years turns out tobe an important damage-driving factor?
3.4 Summary Comments
The search for climate change damage relation-ships has just begun; there is a good deal of workyet to do, especially at the regional level. Althoughthe level of accuracy (hardly the strongest feature ofthe benefit-cost method) remains low, the firstresults do not point toward enormous damages, onthe average. For some regions other than the GreatLakes, the results may differ.
A source of concern is the possibility ofcatastrophic damages beyond some critical amountof climate change. Added to this is a concern aboutdamage to ecosystems and the geographical disper-sion of climate-sensitive plants and animals. Thelimitations of traditional benefit-cost analysis areall too apparent in the face of these questions.
4 THE SEARCH FOR ABATEMENTCOST RELATIONSHIPS
This section seeks to extend the understandingof efforts to estimate the costs of emission reductionor concentration stabilization. Unlike the area ofdamage estimates, which is noted for involving fewstudies with few controversies, the area of costestimates is noted for involving many studies withmany controversies.
Widespread interest in the costs of achievingenvironmental goals is relatively new on thepolitical scene in the United States. The first majorlegislative efforts in the late 1960s and especially inthe 1970 Clean Air Act Amendments (CAAA) werequite clear in indicating that improving the qualityof the air for health reasons was not to involvebenefit-cost analysis. Moreover, the target pollutionlevels — the National Ambient Air QualityStandards — were to be achieved by state andfederal rulemaking that required regulators tobecome familiar with production processes and the
15
Section II
techniques that would bring them into compliance— the command-and-control approach.
Not only does this approach place enormousinformation burdens on the regulatory agencies, butit can also create inefficient incentives, leadingsome emitters to reduce more than an amount thatis cost effective and others to reduce less. Recogni-tion of these burdens was undoubtedly of someimportance in the changes that led up to innova-tions in the 1990 CAAA that both expressed interestin cost effectiveness and permitted SO2 emissiontrading.
My own guess about policy-making motivationsis that the magnitude of expenditures on environ-mental protection (perhaps $130 billion in currentdollars) began to bear on the choice of instruments.Although GHGs were not covered in the 1990 legisla-tion, being left for later consideration, the enormityof some abatement cost estimates could have madedeep marks. That market-based approaches couldbe more cost-effective became an argument withheightened relevance.
Another aspect of thinking about climate changemay have played a role. Trace gases, especiallyCO2, are not ordinary pollutants. The carbon atomis linked to other atoms in the molecules of life.Removing all CO2 and other trace gases from theatmosphere would change the planet from green-house to icehouse. The problem of the greenhouseeffect is trying to decide what volume of emissionsor concentrations we want to live with — the kindof problem that calls for an economic calculation ofcosts.
Environmentalists in such units as the NationalResources Defense Council and the EnvironmentalDefense Fund relaxed their opposition to market-based approaches primarily, I believe, because ofthe cost effectiveness promise of these measures.The regulatory agencies themselves have becomesponsors of cost studies, and the hunt for cost-effective strategies has taken interesting turns, evencoming to a crossroad, as I shall argue.
4.1 Cost Concepts and Cost Measures
To attempt to correct a market externality likethe greenhouse effect by regulatory restriction or by
a tax rate or emission permit will affect the pricethe emitting producer receives; the price theconsumer pays; the ratio of prices of energy-intensive products to others; the terms of trade offuel exporters, fuel processors, and fuel consumers;and possibly the rate of technological progress. Notincluded in these direct and indirect effects wouldbe damages from climate changes, or budgetaryramifications, or adaptation costs incurred becauseof the permitted degree of climate change.
Welfare measures of these effects suggestthemselves some estimate of resource costs ordeadweight loss. The fully dynamic, computable,general equilibrium model that is required to dothis to the satisfaction of all is not yet in sight.Some compromises have to be made. The measuresof abatement costs included in Figure 1 discussedpreviously have been developed from a variety ofsources, including a computable general equilibriummodel, partial equilibrium mathematical program-ming models, and various recursive projectionframeworks (details from the source cited forFigure 1). These estimates are mitigation costs, inthe sense that resources are required to reduceemissions, or remove gases (reforestation), or alterthe radiative balance (e.g., painting things white).
Figure 1 estimates may be classified as "top-down," meaning they are derived from highlyaggregated energy-economic-environment modelswith econometric or macroeconomic-level estimatesof parameters and are based on assumptions,typically, of market clearing and long-run adjust-ment to price changes. They are interpretable aseither deadweight losses, or approximations, such asGDP or consumption losses after policy intervention,compared with a no-policy scenario. In program-ming models, the losses may be calculated from theshadow prices of policy constraints.
The strengths of this approach can be discernedin Figure 1. The "all greenhouse gases" curve,conceptually a horizontal aggregation overmicroeconomic-level emitters and different gases, isa convex set, the exterior points of which denoteefficient marginal mitigation efforts for theparticular degree of emission reduction. It isnonlinear, indicating that small reductions inemissions from baseline are relatively costless asmeasured in dollars per ton of carbon reduced.Larger reductions will require ever-larger resource
Economics
losses. The curve could be a basic building block forcost-effective policy to the extent it is credible:Once the environmental goal, the policy target, ischosen as a percent of emissions to be reduced, theappropriate tax rate (emission permit price) can beread off the values on the ordinate.
Figure 1 implies cross-sectional efficiency inthat cost-minimizing emitters everywhere are led toequate their marginal mitigation costs with the taxrate or permit price. In the next section, I examinethese policy instruments more critically. The issuehere is whether the top-down curve, or refinementsof it, can be the basic building block we seek.
The bottom-up or engineering approach toestimating mitigation costs is exemplified inFigure 2. Termed "technological costing" by theNational Academy of Sciences panel in its report onmitigation (1991), it contrasts with "energymodeling" described above. The NAS report goes onto note its "rough agreement," which, it is asserted,lends comfort to the validity of the Figure 2 curve.However, one curve lies below the abscissa and oneabove in the relevant policy range of reductions ofemissions.
Technological costing derives cost estimates onvarious assumptions about the technical features ofparticular mitigation options; that is, it involvesstudies at the micro-level of improvements inexisting technologies, of replacement technologies,or of altered practices that may improve energyefficiencies, and for which rough appraisals of thedirect costs of implementation may be made. Itholds out the hope of reducing trace gas emissionsby these detailed changes, often at little cost, or atcost savings.
4.2 Commentary on Cost Relationships
The debate between bottom-up and top-down(technological costing versus energy modeling)advocates may be more deep seated than appears atfirst glance. It may also have many implications. Acloser look at the issues throws light on questions ofobtaining useful estimates of the mitigation orabatement cost curve.
The debate is not only about how to estimatemitigation costs; it also "spills over" to the policydebate. I noted earlier that the bottom-up studiescould be closely tied together with the incrementalpolicy option. Instead of focusing on optimalemission reductions or eventual concentrationstabilization at some future date or on a set oftargeted goals, the negotiators could debateincremental policy moves, such as improving vehicleand building energy efficiencies, revisingtransportation systems, and others.
There are attractions to this approach: Many ofthe actions have attached to them negative costestimates, which, if true, ought to generate littleopposition. Countries could adopt those actions thatthey prefer. If each action had attached to it a tracegas emission reduction, an estimate of the aggregatereduction of emissions could be made, and eachcountry's contribution to overall reductions could beappraised.
There are questions to be asked about thisapproach. How much confidence can be placed inthe estimates of direct costs of implementing eachaction? Can a credible aggregation over vehicles,buildings, power plants, factories, farms, or the likebe carried out for each new technology, or improvedtechnology, or changed practice, so that eachcountry's contribution can, in fact, be appraised?What set of incentives or regulatory methods areenvisioned to bring about these changes in the waysthings are done presently? These questions refocusour attention on the issue of how estimates of themitigation cost function that secure widerconfidence can be obtained: Through the methods ofFigure 1 or Figure 2? Or through refinements ofone or the other? Or through a combination of thetwo?
4.2.1 Technological Costing: The Cheapand No-Regrets Way?
The reasons ordinarily given for the negativecosts or energy inefficiencies of Figure 2, which areby no means die most extreme of estimates avail-able (Geller et al. 1992, p. 5), include a lack ofinformation in the markets, high consumer discount
17
Section II
rates for energy investments, market barriers ofvarious kinds, resistance to change or irrationalbehavior, and inadequate incentives for introduc-tion, especially in the public utility sector. In otherwords, the present market solutions to energy useare not in equilibrium.
The microeconomic-level studies that underlietechnological costing vary so widely in method anddatabase that the scope of this discussion does notallow for even a sample appraisal. Severalcomments are in order on how confidence in theresults claimed can be strengthened. Highconsumer discount rates are not necessarilyirrational in terms of energy investments, especiallyamong the poor. Should the poor be urged to investin energy-efficient innovations, or in education, jobtraining, or health? A closer study of the problemsconfronting the poor is required, in my view.
Typically, only direct implementation costs areincluded for energy-efficient innovations. Secondaryeffects, such as those traced out in generalequilibrium models, are ignored. Can estimates ofthese indirect costs be included? Have alladditional information, adjustment, or managerialcosts been considered? It would be reassuring toknow that they have been. Frequently, a hiddenconstant cost assumption is "slipped in": If oneinnovation saves a certain amount, then 100 of thesame innovations save 100 times that amount.What if increasing costs characterize the extension?
There is a policy bias in each of the twoapproaches being discussed. For the advocates oftechnological costing — the bottom-up approach —the bias is toward command-and-control regulation:"Having the right price is . . . not sufficient toachieve optimal efficiency improvements . . . publicauthorities have a responsibility for promotingenergy-efficiency improvements" (Geller et al. 1992,p. 28). For the advocates of the top-down approach,the bias is toward taxes or tradeable permits orother incentive systems that get prices right. Thisis the crossroad now confronting policymakers.
4.2.2 Energy Modeling: Greenhouse GasEmission Reductions Can Be Costly
Underlying the cost curve of Figure 1 areassumptions about markets being in equilibrium:
There are a few appreciable energy inefficiencies tobe exploited. The curve does indicate that the firstfew percentages of reduction are inexpensive,primarily because of the substitution away from theCFCs. But as CO2 emissions are reduced and thefossil fuel industry is affected, marginal mitigationcosts rise.
The assumption of competitive markets inequilibrium is well known to set aside imperfectionssuch as existing subsidies, taxes, regulatorypractices, and the like. Such set-asides are highlyvisible, it should be noted, and frequently, researchis directed toward relaxing an assumption to deter-mine the consequences for the results (e.g., seeGoulder [1990] on the implications of existingtaxes).
The macroeconomic-level studies or energymodels that generate the top-down cost curve haveother problems. They vary in specification andstatistical estimation methods and are subject tospecification error, aggregation bias, and othereconometric difficulties. Mention should be made ofinadequate data and the instability of parameters inprojections over long time periods in the future.The catalog of concerns about estimates is long, butconsiderable reseat i has been conducted withrespect to these troubles. My impression is that thebottom-up approach cannot avoid many of them,and may have a few special problems of its own.
The energy modeling approach, in mostcases, features price adjustments to changes.Autonomous energy conservation, or nonpriceincreases in energy efficiency, can be introduced inmany such models by a parameter in the productionfunction. This device permits the importantdistinction to be drawn between price-inducedenergy conservation and autonomous shifts, and itprovides a measure of the significance of the former,frequently overlooked in bottom-up studies.
The bias in top-down models, and the biasamong those who attach significance to resultsobtainable from them, is toward market-basedapproaches to greenhouse policies. The introductionof emission or concentration constraints in thesemodels, or the introduction of tax rates or positivepermit prices, are assumed to lead emitterseverywhere to equate marginal mitigation costs tothe tax-rate or permit-price equivalent, thus
18
Economics
yielding interfirm or cross-sectional costeffectiveness. The decentralized market guidesconsumers and producers toward energy-efficientchoices. The Figure 1 cost curve traces outsolutions to these cost-effective decisions asemissions are increasingly reduced.
Intertemporal cost effectiveness is alsoimportant. For this condition to hold, emitters mustequate marginal mitigation costs over time, whichimplies that such costs must rise at the lon~-runinterest rate. This condition has been given lessattention than cross-sectional cost effectiveness,with the end result being that proposals to cutemissions sharply in the near term imply highinitial marginal mitigation costs that frequentlydecline over time — an indication of intertemporalcost ineffectiveness.
5 THE SEARCH FOR THE APPROPRIATEINSTRUMENT
U.S. environmental history until recently wasdominated by one policy instrument — regulation orcommand-and-control prescription as the means toobtain environmental standards. (The CAAA of1970 prohibited benefit-cost calculations on ways toobtain environmental quality levels affectinghealth.) Consideration of costs has played a largerrole recently; the IPCC (1990b) cited the costeffectiveness of market-based approaches (e.g., taxesor tradeable emission permits) to greenhouse policyas deserving further study, although there werereservations about these instruments expressed byspokespeople from developing countries.
In addition to cost effectiveness, other criterianeed to be applied to policy tools if well-informedchoices are to be made in this complex area.Implementability, flexibility, and enforcement andmonitoring suggest themselves. It would seem wiseto distinguish sharply between the policy and thepolicy tool kit. Figure 1 provides a handy illustra-tion: A policy may be defined as choosing a point orinterval emission reduction target on the abscissa,and a policy instrument(s) may be defined as theleast-cost way to move vertically to the abatementcost curve, recognizing that it is possible; and easy,to be cost ineffective and move above it.
Some policy instruments may be inseparablefrom the policy. The advocates of energy conserva-tion who see negative cost opportunities forreducing emissions frequently attribute some ofthese opportunities to badly drawn building codespecifications or inappropriate utility regulations.The policy is to reform these imperfections, and thetool is the redrafting of the regulation.
Moreover, policies directed to specific tracegases (CFCs, for example) may have a link withparticular instruments. The speed with whichinternational agreement has been reached on thephasing out of these ozone-depleting chemicals is areflection, no doubt, of the perceived risks ofmelanoma cancer. Attention is focused on reducingproduction as fast as possible and devising schemesto cut emissions from stocks of old cars andrefrigerators.
While the Montreal protocol and follow-upagreements have left implementation to participa-ting countries, provision was made for emissiontrading and other instruments. However, quotasand phase-out schedules — forms of regulatorycontrols — appear to have been the main instru-ments called upon to achieve the quick emissionreduction called for in this global problem area.The fact that the few, large production sites andenterprises involved were ready to cooperate fullymade these policy tools readily implementable andeasily monitored.
While recognizing that there are difficulties inseparating policy from instrument, this paperproceeds with the view that the distinction is clearin most cases and that it enables the researcher toappraise the benefits and costs of applying alter-native policy instruments to a policy goal. Hence, itis envisioned that policymakers have available tothem a list of various command-and-controlmeasures, market-based approaches, and govern-ment intermediate efforts to influence behavior,such as information dissemination, research, design,and development (RD&D) stimulation, and the like.The task is to match these policy instruments to thecharacteristics of the atmospheric trace gases so asto reach the environmental target in the most"satisfactory" way. Such criteria of the"satisfactory" way as implementability, cost
19
Section II
effectiveness, flexibility, and enforcement must beinterpreted in light of the greenhouse effect.
5.1 Implementability
There are two aspects of this criterion thatdeserve attention: the acceptance of an instrumentas "fair" by those affected and the feasibility of theinstrument as applied to sources and sinks of tracegases.
Tax rates (either carbon or commodity taxes)designed to reduce emissions are regressive in thatthey affect both poor families and poor countriesmore heavily than their richer counterparts, unlessoffsetting tax reductions or transfers are made. Ifcarbon taxes are the agreed upon internationalmeans to deal with global warming, then acceptanceof a rate by all nations has much to be said for it;duties on fossil fuel imports or fossil-fuel-intensivecommodities from nontaxing countries would not benecessary. Therefore, if carbon taxes are to beimplemented, the regressivity issue is on theagenda. Requiring permits results in higher energyprices, raising essentially the same issue. (If taxrates and emission permits are designed to achievethe same reduction in emissions, and if the marketoperates under conditions of certainty, the impact ofboth policy instruments on price is identical.)
Market-based approaches (either taxes ortradeable emission permits) seem more implement-able for CFCs or CO2, trace gases for which sourcesand sinks are better understood, than for the othermajor trace gases, CH4, N2O, and tropospheric O3.The sources for CH4 and N2O
a r e poorly under-stood, diverse, and, to a considerable extent,nonanthropogenic. Except for a few sources of theselatter gases, such as fertilizer use and application, itseems unlikely that market-based approaches couldbe implemented and enforced.
For the CFCs and CO2, questions arise as towhich stage of processing implementation could bestbe applied. A consumption tax or emissions permitwould force public officials to calculate the trace gascontent of numerous end-use products. Measuringand recording the carbon content of consumerproducts would generate work for an army ofinspectors and enumerators. The production orextraction stage would hold out more promise.
5.2 Cost Effectiveness
Market-based approaches would appear todominate this criterion; they do not require publicauthorities to have the detailed microeconomic dataon input and market basket options that command-and-control regulation requires. Nor do they imposerigid practices or technologies on firms andconsumers. A tax rate or permit price provides theincentive for the cost-minimizing producer orconsumer to utilize his or her information inmaking the appropriate adjustment. The publicauthorities can devote themselves to setting taxrates or determining the volume of permits. Bothcross-sectional or interfirm and intertemporal costeffectiveness can be claimed for these decentralized,price-determined instruments.
1 find the argument persuasive, but some caremust be taken in particular applications.Tropospheric ozone concentrations generated byprecursors appear to depend on many factors,including the number of hot, sunny days that falltogether. This random variability makes it difficultto devise a market-based approach that is quicklyadaptable; hence, outright bans or prohibitions maybe more cost effective. If the negative costargument of devoted energy conservationists iscorrect, regulation and reform of existing practicesmay be required, since "getting the price right" didnot work in the first instance.
As multiple trace gases are at work in thegreenhouse effect, the appropriate tax rate oremission permit price for each gas must reflect itswanning potential; otherwise, incorrect incentivesare transmitted to emitters, weakening the costeffectiveness argument. Harmonizing these rates orprices will depend on discount rates, radiativepotential, and atmospheric lifetimes. To do thiswith a reasonable chance of success requiresconstruction of a weighting index, a CO2-equivalence metric, which is not without itsdifficulties (IPCC 1990a; Nordhaus 1990).
5.3 Flexibility
Adjusting both policy and policy instruments tonew knowledge is vital in an environmental areainfused with uncertainly like climate change is.One aspect of this uncertainty is the desirability of
20
Economics
avoiding irreversible policies or instrument settingsthat impose heavy costs if they turn out to bewrong. Suppose the government had investedheavily in cold fusion, or required enterprises toprepare the way for its introduction on the groundsthat no trace gases were emitted from this energysource!
Among the market-based approaches, there isan argument favoring tradeable emissions on thiscriterion. International negotiation of tax rates, ifsuch an instrument were to be chosen, is likely tobe an arduous, time-consuming task, perhapsmaking GAIT deals look easy. Resetting tax ratesto take advantage of new information is a highlylikely and desirable event, but the renegotiationmay be more difficult than the original negotiation.
The volume of tradeable emission permits couldbe adjusted more easily. For a cumulative emissionsystem, or undated permits, initiated in the marketto accord with a targeted trace gas (CO2) concentra-tion volume, the number of outstanding permitscould be adjusted by "open market" operationscarried out by an international agency operating asa central permit bank.
Dated permits issued for a year or less wouldmake up an even more flexible system, in the sensethat next year's volume could be adjusted to nextyear's information. Such flexibility could carry aprice: Emitters carrying out long-lived investmentsin abatement technologies must have sometrajectory of tax rates or permit prices in mindwhen deciding upon the most cost-effectiveintertemporal allocation of expenditures.Unexpected changes in rates or prices over timecould upset these plans.
5.4 Enforcement
At first glance, command-and-control measuresappear to be easier to monitor and enforce thanmarket-based measures, because they are clear-cutin prescribing standards and technologies. There iscertainly more experience established with this formof regulation. However, second thoughts, alwaysthe best, cast doubt on this assertion. Checking theperformance of microeconomic decision makers mayentail expensive arrangements of personnel.
However, market-based measures, being relativelynew in the environmental field, may contain nastysurprises. The value of permits would depend inpart on compliance with the rules; cheating orcounterfeiting would dilute the value of authenticpermits.
5.5 Matching Gas Source and SinkControl with Policy Tools
Several features of GHGs bear directly on thechoice of policy tools. In the main, they aredistributed after their emission throughout theglobal atmosphere. They have both anthropogenicand natural sources and sinks that differ in theiramenability to control. A careful matching of thesesources and sinks with the policy instruments is arequirement of well-advised policy.
A detailed account of the sources and sinks andof chemical transformations would be valuable atthis point, but this is beyond the scope of this paper.What is important for our purpose is theanthropogenic share of total emissions: All CFCand related compound emissions are anthropogenic,as are almost all CO2 emissions, perhaps two-thirdsof CH4 emissions, and less than a quarter of N2Oemissions. Altering natural sources and sinks as apolicy device seems difficult and expensive, so policytools will have to be applied to human activitiesthat give rise to emissions.
The CFCs are produced at a few large sites,although emissions from old capital are scatteredand diffuse. Anthropogenic sources of CO2 arisemainly from fossil fuel use and, to a lesser extent,from deforestation; the problem here is at whatlevel or stage of processing of fossil fuels policycontrol should be instigated. Anthropogenic CH4emissions arise from processing of fossil fuels andagricultural practices. N2O arises primarily fromthe latter source.
Sink enhancement offers some promise in thecase of CO2, either through reforestation orgeoengineering plans, such as seeding ironmolecules in cold northern seas to facilitate growthof CO2-absorbing organisms, or hurling particlesthat screen out incoming ultraviolet sunlight intothe stratosphere. Sink enhancements would appear
21
Section II
to be activities deserving of study that would bebest encouraged by government research orsubsidies until they are better understood.
An overview of the matching process justdescribed is presented in Table 3 to provide aqualitative picture of three classes of policy toolsmatched with the trace gas sources. The tools aregrouped into regulatory, RD&D and subsidy(government partnership), and market-basedapproaches.
The CFCs are characterized by a few largeproduction sites with cooperative management, soboth regulation and market-based approaches haveproved effective. Reducing emissions from stocks ofold cars and refrigerators remains a difficult issue;creating incentives for proper disposal may involvea range of tools.
A more rapid phasing out of these gases ismoving up on the international agenda. Newagencies have been created in the World Bank toprovide capital at subsidized rates to developingcountries to aid in replacing the CFCs with morebenign substitutes. Implementation of a widevariety of policy instruments seems feasible and isunderway. The market-based approaches seemmost promising in terms of cost effectiveness, butregulatory measures may be the moststraightforward to enforce.
For CO2 emission control, the market-basedapproaches, when initiated at the extraction stageof processing, hold out prospects for cost effective-ness and flexibility. Enough time seems to beavailable to design systems that create the properincentives. Given die uncertainty surroundingdamage estimates, emission permits may have anadvantage over taxes in that they ensure policy-makers that the desired emission reduction will beobtained. Tax rates may yield widely varyingemission reductions.
Monitoring and enforcement may pose problemsfor new market-based approaches. If CO2 emissionsare to be constrained at intermediate or end-usestages, new match options are opened up. Energyconservationists can argue that regulatory measuresfor buildings, appliances, etc., geared to particularinefficiencies will outperform altered price patterns;tht> issue is not close to reconciliation yet.
Both CH4 and N2O, given current knowledge,offer poor prospects for market-based approaches.The chances of success seem slim, and the costs offailure are high (casting doubt on a promising newapproach). The diversity of sources suggests thatexperimentation with a variety of tools will beworthwhile if any action at all is to be undertaken.The same conclusions may be drawn fortropospheric ozone.
In general, Table 3 suggests that no one policytool will fix the problem. Applying market-basedapproaches to those areas with the greatest chanceof success — CO2 emission reduction — seems welladvised. Emission permits may be more acceptablethan taxes and may offer more control in uncertainareas; they are deserving of more intensiveresearch. Regulatory measures will continue tohave their advocates and their place. Their costeffectiveness ought to be questioned in eachinstance. Government partnership in research andin RD&D is a good buy in this uncertain area.
6 CONCLUSIONS
The argument of this paper is that economistscan contribute to the policy discussion by carryingout "plain vanilla" benefit-cost analysis, with detailson its limitations clearly marked on the package. Asecondary goal is one of clarifyihg^the implicationsof alternative targets.
Estimation of damages remains the under-developed area; perhaps because it is messy,difficult, and lacking in glitz. The lack of reliabledamage estimates in climate- sensitive areas maybe due to the fact that there are not very manydamages, but I remain unconvinced that theimpacts of extreme climatic events, stormfrequencies, soil-precipitation interactions, and thelike have been sufficiently investigated.
The controversy between bottom-up and top-down cost estimation methods is one thatemasculates the policy debate. The temptingpromise that emission reductions are achievable atnegative cost can steal center stage from morepainful choices. Unfortunately, it is going to takesome time to determine the degree of accuracy inbottom-up claims. More work on price-inducedconservation is in the mill, and this should help.
22
TABLE 3 Matching Matrix: Policy Instruments and Greenhouse Gas Control Criteria8
Regulatory Instruments
Policy Instruments
Government Demonstration Market-Based Approaches
Greenhouse Gas andCriteria
Technology Efficiency Bans andStandards Standards Quotas
Subsidies Informationand RD&D Standards
Emission Excise MarketableTaxes Taxes Permits
CFCs(a) Implementation(b) Cost Effectiveness(c) Flexibility(d) Enforcement
CO2
(a) Implementation(b) Cost Effectiveness(c) Flexibility(d) Enforcement
CH4
(a) Implementation(b) Cost Effectiveness(c) Flexibility(d) Enforcement
N2O(a) Implementation(b) Cost Effectiveness(c) Flexibility(d) Enforcement
O3 (precursors)(a) Implementation(b) Cost Effectiveness(c) Flexibility(d) Enforcement
000
00
00
00
0
0
0
0
0
0
0
0
00
•+ M good match, - « poor match, 0 « uncertain match. Ien
Section II
The choice of policy tools is complex, given themultigas, multisource, and multisink character ofthe greenhouse effect. No one tool seems able to fixthe problem to everyone's satisfaction, and there aregood reasons for this situation. Adapting varioustools to the numerous sources and sinks may provebeneficial. This policy would allow governments tochoose their preferred mix. The tool-source-sinkcombination must be appraised by economic calcula-tion for its cost effectiveness (e.g., cost of areduction in CO2-equivalent emissions per ton).Market-based approaches hold great promise on thisscore. Tradeable emission permits, as the "new kidon the block," have the attention of economistsfirmly fixed on them.
7 REFERENCES
Ayres, R.U., and J. Walter, 1991, The GreenhouseEffect: Damages, Costs and Abatement,Environmental and Resource Economics 7:237-270.
Baumol, W.I., and W.E. Oates, 1988, The Theory ofEnvironmental Policy, second edition, CambridgeUniversity Press, New York, N.Y.
Energy Modeling Forum (EMF) 12,1991, interimreports, Stanford University, Palo Alto, Calif., Aug.
Grubb, M.J., 1989, The Greenhouse Effect:Negotiating Targets, Royal Institute of InternationalAffairs, Energy and Environmental Programme,London, Nov.
Geller, H.S., et al., 1992, Getting America Back onthe Energy-Efficiency Track: No-Regrets Policies forSlowing Climate Change, American Council for anEnergy-Efficient Economy, Washington, D.C., Oct.
Goulder, L.H., et al., 1991, Effects of Carbon Taxesin an Economy with Prior Tax Distortions: AnIntertemporal General Equilibrium Analysis for theU.S., Stanford University, Palo Alto, Calif., June.
Hahn, R.W., and G.L. Hester, 1989, Where Did Allthe Markets Go? An Analysis ofEPA's EmissionsTrading Program, Yale Journal on Regulation,6:109-153.
Intergovernmental Panel on Climate Change(IPCC), 1990a, Climate Change: The IPCCScientific Assessment, report prepared for IPCCWorking Group I, J.T. Houghton et al. (editors),Cambridge University Press, New York, N.Y.
Intergovernmental Panel on Climate Change(IPCC), 1990b, Potential Impacts of Climate Change,report prepared for IPCC Working Group II,Cambridge University Press, New York, N.Y., June.
International Energy Agency, 1991, Climate ChangePolicy, Paris, France, Nov.
Kosobud, R.F., 1990, Climate Change Impacts on theGreat Lakes Region; Some Estimates, presented atthe meeting of the American Association for theAdvancement of Science, held in New Orleans, La.,on Feb. 15-20.
Kosobud, R.F., D.W. South, T.A. Daly, and KG.Quinn, 1991, Tradeable C02 Permits for Cost-Effective Control of Global Warming, in Proceedingsof the International Association of EnergyEconomists, Washington, D.C., Nov.
Manne, A.S., and R.G. Richels, 1991, Global CO2
Emissions — the Impacts of Rising Energy Costs,The Energy Journal, /2:81-107.
National Academy of Sciences (NAS), 1991, PolicyImplications of Greenhouse Warming, Report of theMitigation Panel, National Academy Press,Washington, D.C., June.
Nordhaus, W.D., 1990, A Survey of Estimates of theCost of Reduction of Greenhouse Gas Emissions,report to the U.S. Environmental Protection Agency,Washington, D.C., Feb.
Nordhaus, W.D., 1991, To Slow or Not to Slow: TheEconomics of the Greenhouse Effect, The EconomicJournal, 107:920-937, July.
Parry, I.W.H., 1991, Uncertainty, Cost-BenefitAnalysis and the Greenhouse Effect, presented atthe University of Chicago Workshop in Applicationsof Economics, held in Chicago, 111., on April 29.
24
Economics
Schelling, T.C., 1991, testimony presented beforeU.S. Congress, Department of Economics,University of Maryland, College Station, Md.,June 19.
Tietenberg, T.H., 1985, Emissions Trading,Resources for the Future, Washington, D.C.
U.S. Department of Justice (DOJ), 1991, AComprehensive Approach to Addressing PotentialClimate Change, Report of the Task Force on theComprehensive Approach to Climate Change,Washington, B.C., Feb.
U.S. Environmental Protection Agency (EPA), 1988,Potential Impacts on the U.S. of Global ClimateChange, report to the U.S. Congress, Washington,D.C.
Victor, D.G., 1990, Tradeable Permits andGreenhouse Gas Reductions: Some Issues for U.S.Negotiators, Energy and Environmental PolicyCenter, John F. Kennedy School of Government,Harvard University, Cambridge, Mass.
Weitzman, M., 1974, Prices and Quantities, TheReview of Economic Studies, 67:449-477.
8 ENDNOTES
1. A. Pigou was the English economist whoanalyzed how the imposition of a tax onpolluting commodities would internalize for theenterprise the external costs of the pollutant.The reason for the range in the tax given in theexample is a result of the range in the estimatesof damage or impacts; they ranged from low tomedium.
2. The exception to this statement, which provesthe rule, is the stratospheric ozone depletionproblem, in which the threat of skin cancer andother damages and the feasibility of substitutesfor the CFCs and other pollutants have led toremarkable international cooperation on policy.I am indebted to professor Gilbert Bassett incalling to my attention the literature on"existence value" research, which represents oneway to broaden the concept of environmentaldamages.
3. The U.S. EPA (1988) carried out a massivesurvey of potential effects of climate change onthe United States. The findings for the GreatLakes (Chapter 5) may be of special interest:Projections of climate change from three generalcirculation models under a scenario in whichCO2 concentration is doubled indicated highertemperatures (the range was from 3° to 7°C)and precipitation changes (from +0.6 mm to -0.5mm per day). Corn yields could decline up to60% (except for Minnesota), and lake levelscould decrease from a few inches to 7 feet. Theranges hint at the difficulties of making regionalclimate-change projections.
4. This is the region of responsibility of theFederal Reserve Bank of Chicago. I want tothank Dave Allardice and Bill Testa of the bankfor providing useful data on the region'seconomic and demographic characteristics.They are not responsible for the damageestimates or for errors.
5. The scales of the two abscissas are different butcan be made comparable. Figure 1 is measuredin CO2-equivalent percentage reductions, andthe optimal reduction point (where the marginalcurves cross) is a little less than 20%. Figure 2is measured in billions of tons of CO2-equivalentfor the United States. Where the mitigationcurve crosses the axis, about 2 billion tons,represents a reduction of about 25% in U.S.emissions. Note, however, that Figure 2 dataare probably average costs, not marginal costs.
25
Section II
Comment 1 on Workshop in Economics
A Note on Benefit-Cost Analysis and the Distribution of Benefits:The Greenhouse Effect
Mr. Kevin G. QuinnArgonne National Laboratory
Argonne, Illinois
The application of benefit-cost analysis toenvironmental problems in general, and to globalwarming as demonstrated by Kosobud in particular,is a very useful tool. Depending upon the limita-tions of the relevant data available (in this case, thedegree to which the marginal costs and marginalbenefits of a reduction in greenhouse gas emissionscan be quantified), benefit-cost analysis can offerinformation to society about how to improve itscondition. However, beyond the criticism of itsestimate of the Pareto optimal point (based ontheoretical considerations regarding the convexity ofthe social production frontier and of socialpreferences), benefit-cost analysis suffers from afundamental weakness: It cannot speak to thedistribution of the net benefits of implementation ofan international greenhouse policy.
Within an individual country, debate on aparticular policy intervention can effectivelyseparate the issues of achieving a potential Paretooptimum and distributing the benefits necessary toactually accomplish Pareto optimality. Thissituation occurs because (theoretically, anyway)these decisions are made in the presence of a
binding enforcement regime that can redistributebenefits as seen fit. A policy can then be introducedin the manner that achieves the best overall netbenefits, and the allocation of these benefits can betreated as a stand-alone problem.
However, in the case of international policy-making such as in the greenhouse situation, it isnot possible to divide these two issues. There is noenforcement regimen that can impose a bindingredistribution of net benefits, at least at this time.Instead, policy analysis should focus on thedetermination of what is the best attainable policy,given that each actor (country) must be better offafter the implementation of this policy, all thingsconsidered. "All things considered" would includethe possibility of side payments and the nature ofthe additional utility that a country enjoys when apolicy results in a more equitable allocation of netbenefits among countries. A game theoreticapproach may be more suitable to this problembecause it addresses this consideration directly.Benefit-cost analysis, on the other hand, maysuggest an "optimal policy" that has no chance ofbeing successfully implemented and observed.
26
Economics
Comment 2 on Workshop in Economics
Issues in Benefit-Cost Analysis:Amplification Channels and Discounting Long-Term Environmental Damages
Or. Donald A. HansonArgonne National Laboratory
Argonne, Illinois
Many environmental problems have long-termeffects. Acid rain has long-term effects on soils,forests, and exposed materials. Global climatechange has even longer-term effects. This differencein timing — between the near-term cost ofenvironmental protection and the long-termenvironmental effects — makes it difficult toconduct a cost-benefit analysis of any programdesigned to abate environmental damages. The rateat which to discount long-term environmentaldamages becomes a key question in comparisons ofbenefits and costs. This comment points out animportant facet of the discounting issue. Thediscount rate for calculating the present value offuture environmental benefits may be much lowerthan the rate of return on investment.
Cost-benefit analysis is a framework in which toevaluate policies and decisions. Because globalclimate change is a complex problem, extensions ofcost-benefit theory can be expected to add additionalinsights, particularly in the following areas:
• Distinguishing distributional effects amongnations, over time, and among generations;
• Determining the rate of discount that isappropriate for long-term environmentaldamages and separating risk aspects fromthe rate of discount; and
• Assessing amplification effects when policiesinvolve large expenditures relative to theeconomy or when affected sectors aresignificant sectors of the economy.
This comment does not add anything to thediscussion of distributional effects among nations,
but it should be pointed out that the target rate ofeconomic growth is also a distributional issue. Thepresent generation may be asked to work hard andsacrifice so that future generations can attainhigher levels of consumption.
Another comment is that for highly uncertainenvironmental problems such as global climatechange, the uncertainty component should beseparated from the time rate of discounting. Someenvironmental effects that may not be manifestedfor a very long time are reasonably predictable.However, it is true that predictions can be expectedto be more uncertain the farther in the future theyare for.
To illustrate, consider a simple aggregateeconomic growth model of consumer goods andenvironmental goods (the last of which are ameasure of environmental quality). The economypossesses the capability to transform betweenconsumer goods and environmental goods (i.e.,produce one type of good instead of another) viaexpenditures for environmental protection andabatement of environmental damages. Society canalso transform foregone current consumption intofuture consumer goods via investment. The net rateof return on investment is determined within themodel (endogenously) as it takes into account avaluation of the environmental damages that resultfrom pollution associated with production. Thecorrect formula for calculating the present value offuture consumer goods is to use the net rate ofreturn on investment as the discount factor.However, the correct method for calculating thepresent value of future environmental damages is touse a pure rate of time-preference — a rate thatmembers of society choose to discount future welfare
27
Section II
for themselves and future generations. Thevaluation of environmental damages is based on theconsumer's marginal rate of substitution betweenenvironmental quality and consumer goods. Theimplication is that any long-term environmentaldamages that are foreseen may be discounted at alow rate chosen by society. Furthermore, thevaluation of these damages may be high if consumergoods are relatively more abundant than environ-mental goods. The valuation of damages is basedon the shape of the indifference curve betweenenvironmental qualify and consumer goods.
What decisions or policies follow from thisanalysis (which does not take into account theimplications of uncertainties in both environmentaleffects and costs of policies to abate environmentaldamages)? Society faces a budget constraint in thatits total production (i.e., gross domestic product)must equal its production of current consumergoods, investment, and pollution-abatementexpenditures. Hence, when two of these variables(e.g., investment and abatement expenditures) arechosen, the third (current consumption) isdetermined from the budget constraint. Thedecision rule (or method) to calculate investment isdetermined on the basis of the net rate of return.The decision rule to calculate abatementexpenditures is determined through a cost-benefitanalysis, in which the benefits are valued on thebasis of the marginal rate of substitution and arediscounted to the present through use of the purerate of time-preference on utility. Hence, societymust balance its desire for current consumptionagainst its desire for future growth in the standardof living — a growth that depends on itsinvestment — and its desire for environmentalquality in the near term and long term. The readerseeking more information on these topics cancontact the author for additional papers.
In cost-benefit analysis, total benefits should becompared with total costs — costs that include notonly direct effects but also second- and third-roundeffects as the economy adjusts to policy changes. In
this adjustment process, there are two offsettingforces: substitution and rigidities. Bothsubstitution among consumer goods andsubstitution among factors of production allow thecost of an initial policy change to be mitigated ingenera] equilibrium. However, rigidities in themacroeconomy, such as "sticky wages" (a conditionin which labor resists a reduction in wages), causetotal losses in the economy to be amplified relativeto the original cost. (See Hanson, D.A., 1992, The1990 Clean Air Act: A Tougher RegulatoryChallenge Facing Midwest Industry, EconomicPerspectives, Federal Reserve Bank of Chicago,May/June.)
These dimensions of cost-benefit analysis makeenvironmental problems more complex to analyze,but they also open a door for policy improvements.Whenever possible, policies should exploitsubstitution possibilities and be designed tomitigate losses by allowing for smooth adjustments.
There are also interesting implications formarket-based approaches that make use ofemissions taxes or tradable permits. A firm willhave the incentive to reduce emissions up to thepoint where the marginal abatement cost (MAC)equals the emissions tax (or the tradable permitprice). Hence, if the overall reduction in emissionsis stringent, the MAC will rise and the price ofpollution (i.e., cost of the tax or permit) will be high.The resulting emissions tax revenue is in additionto total abatement costs. The emissions tax revenueand total abatement costs are either passed on tocustomers, taken out of wages, or taken out ofprofits. The emissions tax revenue, if it is large(which occurs when there is a significant carbontax), can lead to reductions in wages and profits orincreases in inflation and sectoral adjustments.Attendant macroeconomic losses are not fullyconsidered in the standard Baumol-Oates theorythat demonstrates the microeconomic efficiency ofthese instruments. Simulations on the size of themacroeconomic amplification effect are availablefrom the author.
28
Section IIIWorkshop in Adaptation and Mitigation Strategies
Approaching Global Warming:A Review of the Adaptation and Mitigation Perspectives
Dr. Peter M. MorrisetteClimate Resources Program
Resources for the FutureWashington, D.C.
1 INTRODUCTION
The debate within the scientific, policy, andenvironmental communities on what, if anything, todo about global warming appears to be focused onwhether to adapt to climate change in the future orto mitigate climate change in the present. As theissue has become increasingly politicized, the debateover these two approaches has become polarized.The two approaches, however, are not mutuallyexclusive; in fact, there is much common groundbetween them. But differences can be found in howproponents of each approach view the risks of globalclimate change and the values that underpin theseperceptions of risk. In this paper, I will brieflyoutline the progression of global warming from anobscure scientific concern into a leading interna-tional political issue. I will also review someprevious efforts by social scientists to assessattitudes and positions on global warming. I willthen examine in detail the adaptation and mitiga-tion perspectives and assess how they differ on thebasis of different conceptions of uncertainty andrisk, equity, and technology. Finally, I will examinethe adaptation and mitigation approaches from theperspective of developing countries.
2 GLOBAL WARMING'S RISE TOPROMINENCE
The global wanning issue is at the forefront ofinternational science and politics. Negotiations arecurrently underway on an international frameworkconvention on global warming, and some countrieshave even pledged to stabilize or reduce emissionsof carbon dioxide. Global warming's rise to
prominence on the international political agenda isremarkable. If we look back only a few years to themid 1980s, global warming was of interest mostly tothe scientific community, some environmentalists,and a handful of politicians. What makes globalwarming's sudden political prominence even moreremarkable is that conventional political wisdomtells us that an issue like global warming, charac-terized by great uncertainty associated with thefacts of the problem and by high potential economiccosts associated with taking action, is not likely toadvance very far in the policy process.
Understanding global warming's suddenprominence becomes even more intriguing when weconsider that scientists have long known about thegreenhouse effect. The possibility that CO2emissions from the burning of fossil fuels mightenhance the earth's natural greenhouse effect waspostulated as early as the 1890s by the Swedishchemist Svante Arrhenius (1896). Arrheniuscalculated that a doubling of atmospheric CO2concentration could lead to a global warming of fiveor six degrees celsius (a predication that is not farfrom the current projection) and that this warmingwould be greater in higher latitudes. Similarly, inthe 1930s, Callendar (1938) described the effect thatCO2 emissions from the burning of fossil fuels mighthave on global temperatures. And it was 35 yearsago when Revelle and Suess (1957) made their nowfamous remarks describing the anthropogenicrelease of CO2 as "a large-scale geophysicalexperiment," the effects of which could not bepredicted.
It has been argued that the basic scientificprinciples of the global warming problem have not
29
Sect ion HI
changed all that much in the past 50 years.Michael Glantz (1988, p. 51) of the National Centerfor Atmospheric Research (NCAR), for example,writes:
As far as the scientific aspects of the globalwarming issue are concerned (with regardonly to the burning of fossil fuels), itappears that most of the aspects discussedtoday . . . were being discussed at least fiftyyears ago.. . . While we have acquiredconsiderably more knowledge throughmodeling and paleoclimatic investigation,the scientific premise of CO2-induced globalwarming remains essentially the same as itwas in the 1930s.
Andrew Solow of Woods Hole (1991, p. 7) makesthe same point when he writes: "In many ways thebroad outline of [the global warming] debate hasremained unchanged over the past fifty years.Atmospheric science and climatology have advancedquite a lot over that period, however, and thedetails of the debate are somewhat clearer."
If the basic scientific premise of global warminghas remained essentially the same since the 1930s,what has changed, and fundamentally so, is ourknowledge and understanding of what globalwarming might mean for the planet. For example,we now have some understanding of how theclimate effects of global warming might affectagriculture, forestry, water resources, and naturalecosystems, how it might raise sea levels and floodcoastal lowlands, and how it might affect economiesand societies around the world. What has alsochanged is our perception of the problem. WhenCallendar was writing in the 1930s, the effects ofCO2-induced wanning were seen as potentiallybeneficial — expanding the margins of agricultureor postponing the onset of the next ice age(Callendar 1938). Even Revelle and Suess'swarning in the 1950s did not create much concern.It was not until the 1970s and the rise of politicaland scientific concern for the environment that thebroader implications of CO2-induced global warmingbegan to receive serious consideration. However,outside the scientific and environmental community,global warming remained a rather obscure issueuntil the mid 1980s.
Several factors have been invoked asexplanations of global warming's rapid rise on theinternational political agenda in the late 1980s.These include the North American drought of 1988and other unusual weather events around the worldin recent years that have sensitized the public andpoliticians in the United States and elsewhere tothe variability of climate; other environmentalproblems such as ozone depletion, acid rain, anddeforestation that have raised public and politicalconcern for the fate of the earth's environment; anemerging consensus in the international scientificcommunity on the climate-forcing potential ofgreenhouse gases; scientific advances that haveimproved our understanding of how the earth'svarious natural systems (climate, the oceans, andthe biosphere) interact and of how human activities,in particular, can influence these systems; and abelief that not taking action now to limit globalwarming is too great a risk to take.
There is, however, a more subtle underlyingexplanation to global warming's prominence. Theeconomist Lester Lave (1988) suggests that it is thesymbolic nature of the global warming problem thathas attracted so much attention to the issue inrecent years. Global wanning demonstrates notonly the reality of resource constraints but alsohumankind's capacity to alter environmentalsystems on a global scale. Lave argues that wehave experienced a "loss of innocence"; i.e., we canno longer ignore our environmental problems,because there are no more frontiers to exploit. In alarger sense, global warming is also symbolic ofother problems that confront modern society in thelate twentieth century: problems that are global inscale, highly complex, not fully understood, and notonly lack an obvious solution but alsofundamentally question the problem-solvingcapacity of existing institutions.
One interesting fact about global warming'sprominence is that the public has not reactedstrongly to the issue. This is evident in WillettKempton's research on lay perspectives on globalclimate change (Kempton 1991). Kempton's surveyof public attitudes shows that while the public hasheard of the global warming problem, it has verylittle understanding of it. In fact, most people inKempton's survey confused global warming with
30
Adaptation and Mitigation Strategies
stratospheric ozone depletion, and few understoodthe connection between global warming and energyconsumption (Kempton 1991). Furthermore, in thepublic's eyes, global warming lacks the immediacyand localized concern of, for example, mismanagedhazardous waste dumps or polluted drinking water;at the global scale, the issue lacks a tangibleimpact, such as the Antarctic ozone hole. Thus, atthis time, there appears to be no strong publicmandate for action driving the global warmingissue. Rather, global warming has found its mostvocal and active support among scientific, political,and environmental elites.
3 PREVIOUS ASSESSMENTS OF GLOBALWARMING
At this point, it is useful to review previousefforts at assessing attitudes toward and positionson the global wanning problem. I briefly reviewthree earlier assessments, each of which used adifferent approach. In the first of these assess-ments, Glantz (1988) described the degree to whichdifferent segments of the scientific and policycommunities believe that global warming is aproblem, and how this might influence internationalefforts to control CO2 emissions. In the secondassessment, Andrew Plantinga and I reviewed thepolitical and economic factors that would likelyinfluence the degree to which countries wouldsupport an international agreement to control theemissions of CO2 (Morrisette and Plantinga 1991).In the third assessment, O'Riordan and Rayner(1991) took a more theoretical approach towardidentifying the values and beliefs that underliedifferent "decision-making strategies" toward globalwarming.
Glantz (1988) has argued that within thescientific and policy-making communities, there arehawks, doves, and owls on the global warmingissue. The hawks, according to Glantz (p, 41), arethose who "believe that the evidence of COg/tracegases warming is very convincing and that thewarming is already under way." The doves, on theother hand, tend to dismiss the issue and "feel that[it] is yet another doomsday scenario that will mostlikely fail to materialize" (p. 41). The owls, Glantz(p. 42) explains, "have yet to make up their mindson the issue." They are waiting for moreinformation.
Glantz believes that within the scientificcommunity, the number of hawks is increasing,while the number of doves is dwindling. Thegrowing scientific evidence in support of the globalwarming theory is responsible for this shift. YetGlantz notes that there are still "persistentscientific uncertainties," and thus there are stilldoves. Within the international policymakingcommunity, Glantz believes that the owls are thelargest group, perhaps reflecting the traditionallypragmatic view of the policymaker. Glantzconcludes by noting that because there are still verydifferent perceptions of the seriousness of the globalwarming problem, particularly relative to otherproblems on the international political agenda, it isnot surprising that there is no firm consensus onthe issue. It might even be argued that it issurprising that there is as much consensus on theissue as there appears to be.
The focus of the earlier assessment of globalwarming that Andrew Plantinga and I undertookwas primarily on the policy positions of nations(Morrisette and Plantinga 1991). Our goal was tounderstand the political and economic factors thatwould condition the degree to which a countrywould be motivated to support an internationalagreement to control global emissions of CO2. Atthe risk of being overly simplistic, we divided thenations of the world into four groups on the basis oftheir stated policy positions on global warming:(1) countries advocating a cautious approach,(2) countries advocating an activist approach,(3) countries unable to act, and (4) countries unableand unwilling to act.
Among the wealthy, industrial countries, theUnited States is the principal (if not only) memberof the group advocating a cautious approach. TheUnited States maintains that not enough scientificevidence is available to justify potentially costlymeasures to mitigate emissions of C02. The activistgroup is made up of those countries in theOrganization for Economic Cooperation andDevelopment (OECD) that have endorsed CO2stabilization and reduction targets (e.g., Germany,Canada, Sweden and Japan). These countriessupport the view that the scientific evidence pointsto a significant threat from climate change and thatthe benefits of averting the threat outweigh thecosts of controlling CO2 emissions. Because of moreimmediate political, economic, and environmental
31
Section
problems, the countries of Eastern Europe and theformer Soviet Union are unable to respond to theCO2 issue with the same degree of commitmentcharacteristic of many Western European countries.Yet these countries have professed a serious concernabout the global warming problem and are willingto participate, where possible, in internationalefforts to respond to global warming.
Like the countries of Eastern Europe, manydeveloping countries are also unable to committhemselves to potentially costly efforts at controllingCO2 emissions. Furthermore, these countries havebeen less willing (if not unwilling) to supportinternational efforts to control CO2 emissions thatdo not at the same time directly address problems ofunderdevelopment. The developing countries aresuspicious of the motivations behind the interest ofthe world's rich countries in the global warmingissue and question the long-term commitments ofdeveloped countries to provide assistance todeveloping countries. The support of developingcountries will be essential to the success of anyinternational agreement on global climate change,yet their attitudes and positions toward the problemdiffer in important ways from those commonlysupported by scientific and policy elites in thedeveloped countries. I discuss the attitudes andpositions of developing countries in greater detaillater in this paper.
The assessment by O'Riordan and Rayner (1991)focuses on the underlying values and concepts ofnature that shape different decision-makingstrategies toward global warming. O'Riordan andRayner identify three decision-making strategies —prevention, adaptation, and sustainabledevelopment — and argue that each is based on afundamentally different concept of nature. Thepreventivists, they argue, believe that nature isfragile and easily susceptible to catastrophic andirreversible change. The preventivist approach,O'Riordan and Rayner (1991, p. 99) write, isunderpinned "by a conviction that it is morallywrong to distort natural processes by humanbehaviour." The adaptationist approach, on theother hand, is based on a cornucopian view ofnature in which the environment is seen as flexibleand easily capable of adapting to change. O'Riordanand Rayner (1991, p. 100) explain that "the ethical
imperative driving this position is that it is morallywrong to curtail economic development andcondemn the poor people and nations of the earth topermanent deprivation of the benefits of modernindustrial society." A synthesis of these twostrategies can be found in the sustainabledevelopment approach, which is underpinned by thebelief that global catastrophe can be avoidedthrough "careful stewardship of the limitedopportunities that nature provides for controlledgrowth" (p. 100). The sustainable development viewof nature sees the environment as generallyforgiving of many disturbances but vulnerable tocatastrophic change.
O'Riordan and Rayner believe that thesedifferent viewpoints derive not from disputes overfacts and knowledge about global change but ratherfrom different, institutionally derived ways ofapproaching or framing the problem. They explainthat preventivists tend to be egalitarian andcollectivist in their approach to problems,adaptationists are more likely to support a laissez-faire approach, and sustainable developmentaliststend to support hierarchical approaches to solvingproblems. O'Riordan and Rayner favor a "holistic"risk management strategy that could bridge thesedifferent, institutionally derived ways of framingthe problem. Such a holistic strategy, they explain,"would mean reconstructing scientific and politicalinstitutions in new ways to make decisionsinvolving global stakes at unprecedented levels ofuncertainty" (O'Riordan and Rayner 1991, p. 101).
In the remainder of this paper, I present a morefocused assessment of the adaptation and mitigationperspectives on global warming. In some importantways, this assessment is a synthesis of the threeapproaches that I have described above. Similar tothe approach described by Glantz, I focus primarilyon scientific, policy, and environmental elites,although I also examine the global warmingproblem from the perspective of the developingcountries. Like Giantz's approach and that used byMorrisette and Plantinga, I focus on the practicalpolitical and economic considerations of dealing withthe global warming problem. Finally, similar to thework of O'Riordan and Rayner, I assess some of thebasic, underlying values that characterize differentperspectives on the global warming problem.
32
Adaptation and Mitigation Strategies
4 TWO VIEWS: ADAPTATION ANDMITIGATION
For several years, natural and social scientists,policymakers, and environmentalists have beendebating over two seemingly different views abouthow society should respond to the global warmingproblem: adaptation and mitigation. Proponents ofthe adaptation perspective tend to stress thecapacity of society to adapt to changing conditionsin the environment through technological andinstitutional innovations. Proponents of themitigation perspective argue that taking steps tocurtail or eliminate the emission of greenhousegases is the only reasonable, long-term approach tothe global warming problem. Although there is notan inherent dichotomy between the two approaches(i.e., it is not an either/or decision; choosing onedoes not preclude the other), the discussion withinthe scientific and policy communities concerningthese two approaches has become polarized.
It may be, however, that the adaptation/mitigation model is not a very useful way toconceptualize the global warming problem.*Nevertheless, for better or worse, it is thepredominant conceptual model in the scientific andpolicy literature. The relatively modest purpose ofthis paper is not to argue the merits of theadaptation/mitigation model; rather, it is simply todescribe these two perspectives as they have beenoutlined in the scientific and policy literature,assessing where they differ and how they aresimilar, and offering an initial explanation on whythe debate seems to be polarized.
The release in 1991 of the National Academy ofSciences (NAS) report on the policy implications ofglobal warming, in which prospects for bothadaptation and mitigation were assessed, clearlyillustrates how polarized the discussion has become.The adaptation panel, which was one of four panels(the others were science, mitigation, and synthesis),put forward a relatively optimistic assessment ofthe capacity of the United States to adapt to amoderate, greenhouse-induced climate change. But
the reports of both the synthesis panel and theadaptation panel included dissenting opinions (inthe form of lengthy footnotes) that questioned thereasoning employed by the adaptation panel — anunusual feature for an NAS study. Also interestingis the fact that these dissenting opinions receivedmore attention in both the scientific and popularmedia than did the findings of the adaptation panel(see, for example, Roberts [1991]).
The debate over adaptation and mitigation isnot new. The broad outline of the adaptation andmitigation perspectives is clearly evident in earlyworks assessing the range of policy responses toglobal warming (for example, see Mann [1983] andSchelling [1983]), although in these earlier policyreviews, mitigation is commonly referred to asprevention. Just as Glantz and Solow have arguedthat the basic scientific foundations of the globalwarming theory have not changed all that much in50 years, I would argue that the basic structure ofthe policy debate (i.e., the concepts of adaptationand mitigation) also has not changed fundamentallysince it was first discussed seriously in the late1970s. What has changed is our knowledge aboutpotential impacts from global warming, althoughmany analysts would argue that this knowledge isstill inadequate for formulating specific policies. Inaddition, the political implications of the globalwarming issue have also changed. In the 1970s andearly 1980s, the debate over global warming was,for the most part, confined to the scientificcommunity. Today, it is at the top of politicalagendas around the world, and the stakes includenot only the fate of the earth's environment but alsothe future of the world's economy. This debateillustrates how different values can cause scientists,policymakers, and environmentalists to interpretthe same facts differently.
4.1 Adaptation Perspective
There are two themes within the adaptationperspective: first, society has a built-in capacity toadapt to changes in climate and environment;
""Discussions during the conference on Global Climate Change and International Security, in fact, reflectedsome misgivings among some conference participants about the usefulness of the adaptation/mitigationmodel.
33
Section m
second, the cost of preventing climate change maybe so prohibitive that societies will chose adaptationover mitigation. While not mutually exclusive,these two themes often appear independent of eachother in the literature. Proponents of the firsttheme tend to be agricultural scientists, engineers,resource economists, and geographers. Proponentsof the second theme tend to be economists andpolicy analysts.
Adaptationists note that according to currentscientific knowledge, even if the anthropogenicemissions of all greenhouse gases were eliminatedtoday, it is likely that the earth would stillexperience some global warming as a result of thegreenhouse gases that have already accumulated inthe atmosphere (for example, see Houghton,Jenkins, and Ephraums [1990]). If this is indeedthe case, then some adaptation to a greenhouse-warmed world will almost certainly occur in thecoming decades, regardless of what takes place inthe international political arena. Adapting toclimate change, however, is something that societiesare already accustomed to doing. For example, theadjustments and adaptations that farmers on theNorth American Great Plains have made to droughtis well documented (for example, see Riebsame[1990]).
One of the principal conclusions of the NASpanel on adaptation is that there is already a well-established capacity in society for adapting both toa wide range of climate conditions and to climatechange (NAS 1991a). The Academy panel empha-sizes the range of climates that people already livein as evidence of humanity's capacity to cope withwide variations in climate. The panel writes that"by adjusting to extremes of the current climate,society has already made investments that wouldabsorb some climate change" (NAS 1991a, p. 25).The panel further explains that "the capacity ofhumans to adapt is evident in the rapid techno-logical, economic, and political changes of the past90 years" (p. 2). Not all activities, however, will beable to adapt easily. The panel notes, for example,that water resources and natural ecosystems maybe particularly vulnerable, and that poor countrieswill almost certainly have more difficulty adaptingto climate change than rich countries. The panel'sconclusion is, however, that for a gradual andmoderate change in climate (the type of climatechange on which it based its assessment), "impacts
will be no more severe, and adapting to them will beno more difficult, than for the range of climatesalready on earth and no more difficult than forother changes humanity faces" (p. 146).
A basic premise underlying this finding is thecapacity for societies to capitalize on technologicaland institutional innovations (NAS 1991a; Ausubel1991). Technological innovations include advancesin crop varieties or farm machinery, whileinstitutional innovations involve changes in theunderlying political, economic, and social structures.The emergence of land-grant universities in theUnited States would be an example of aninstitutional innovation. With respect toagriculture, Crosson and Rosenberg (1991, p. 19)explain how innovation in the face of climate changemight take'place:
If farmers find that the alternativesavailable to them from among existingtechnologies and management practices areinadequate to compensate for the negativeimpacts of climate change, they may face theprospect of going out of farming, andperhaps leaving the region altogether. Theprospect could stimulate agriculturalresearch institutions and those charged withresponsibility for agricultural policy toinvest more research to develop a new set oftechnologies and practices better adapted tothe changed climatic regime.
Adaptationists argue that societies will continueto adapt to climate change (either natural oranthropogenic) through technological andinstitutional innovation. Indeed, Ausubel (1991)
* notes that the trend has been toward a lessening ofsocietal vulnerability to climate, and there is noreason to believe that this will not continue, even inthe face of a greenhouse-induced climate change.
The adaptation perspective is also concernedwith the potential long-term costs to society ofpursuing meaningful mitigation strategies. Someeconomists, for example, do not see the motivationfor either wealthy or poor countries to pursue costlyand uncertain efforts at controlling CO2 emissions(Lave 1988; Schelling 1991; Nordhaus 1991). Withrespect to the United States, Nordhaus estimatesthat only 3% of national income originates fromsectors that are sensitive to the effects of climate
34
Adaptation and Mitigation Strategies
change, with perhaps another 10% of nationalincome being modestly sensitive. Nordhausconcludes that the economic impact to the UnitedStates of climate change that was induced bydoubling atmospheric CO2 would be small — onlyabout one-fourth of one percent of national income.Scheliing (1991) agrees with this conclusion andsuggests that the same is likely to be true for Japanand Western Europe as well. Schelling (p. 202)argues that "the countries that can afford to takeserious measures to reduce carbon emissions maynot be able to perceive any narrow national interestin doing so."
Given limited resources and competinginterests, economists are also concerned with theopportunity costs associated with pursuingmitigation strategies, particularly in developingcountries. This concern has been expressed byNordhaus (1991, pp. 57-58):
We might. . . ask whether a majorcommitment to slowing climate changes is aworthwhile investment for developingcountries who are likely to be the regionsmost vulnerable to climate change.. . . Todevote many billions of dollars of resourcesto slowing climate change at the expense ofequivalent investments in education, energyconservation, or tangible capital indeveloping countries would probably hurtpoor countries and give little return in high-income countries.. . . Faced with thedilemma of a low social discount rate and ahigh return on capital, the efficient policywould be to invest heavily in high-returncapital now and then use the fruits of thoseinvestments to slow climate change in thefuture.
In essence, Nordhaus is arguing for a strategy ofmitigation through adaptation.1" The underlyingassumption is that increased investment indevelopment in poor countries could lead thesecountries, in the long run, to use energy moreefficiently, produce less greenhouse gas emissions,
and improve their capacity to adapt to changingclimate conditions if that becomes necessary.
Adaptationists, whether they believe in thecapacity of society to overcome problems or questionthe economic rationality of costly mitigationmeasures, do not dismiss the potential threat tosociety from climate change, although some mayargue that it is being overstated. Most adapta-tionists acknowledge that given the existinguncertainties about the climate system, there is arisk of surprises and even catastrophic change.That risk, however, supports a policy of moreresearch, monitoring, and alertness, not a policy ofpotentially costly and misguided efforts atmitigation.
4.2 Mitigation Perspective
In general terms, the mitigation perspective isbased on the proposition that there are existing,cost-effective strategies available now for reducingemissions of CO2 and other greenhouse gases.Indeed, this was the finding of the NAS synthesispanel, which indicated that "the United States couldreduce its greenhouse gas emissions by between 10and 40 percent of the 1990 level at very low cost.Some reductions may even be at a net savings if theproper policies are implemented" (NAS 1991b,p. 63). The conclusions of participants in theSecond World Climate Conference were even morespecific with respect to reductions in CO2 emissions,stating that "technically feasible and cost-effectiveopportunities exist to reduce CO2 emissions in allcountries. Such opportunities for emissions reduc-tions are sufficient to allow many industrializedcountries to stabilize CO2 emissions from the energysector and to reduce these emissions by at least 20percent by 2005" (World MeteorologicalOrganization [WMOJ1990, p. 3).
The principal belief underpinning die mitigationperspective is that according to current scientificevidence, global warming represents a potentiallyserious threat and that efforts at reducing emissions
*The idea of mitigation through adaptation is suggested by Lave (1988), who argues that it will be thepreferred approach by default because, given existing uncertainties and competing interests, the countries ofthe world will probably not agree to mitigation.
35
Section III
of greenhouse gases are a prudent insurance policyagainst this threat. Again, this was a principalconclusion of the NAS synthesis panel. "Given theconsiderable uncertainties in our knowledge of therelevant phenomena, greenhouse warming poses apotential threat sufficient to merit promptresponses. . . . Investment in mitigation measuresacts as insurance protection against the greatuncertainties and the possibility of dramaticsurprises" (NAS 1991b, p. 67). Similarly, Lave(1988) also describes efforts at greenhouse gasmitigation as an insurance policy. He writes that"most people voluntarily purchase life insurance,even though the likelihood of dying in a particularyear is very small. I suspect that people would bewilling to pay a premium for a policy that wouldprotect against an inhospitable Earth a century orso hence" (p. 464).
In addition, proponents of mitigation also tendto believe that many of the activities suggested aspossible strategies for mitigating global climatechange — improved energy efficiency, energyconservation, and fuel switching — make sense bothenvironmentally and economically, regardless ofclimate change. (This is a view shared by manyadaptationists as well.) Seductions in fossil fuel useand emissions could lead, for example, to a morecompetitive economy and less pollution. StephenSchneider (1989) calls these activities "tie-in-strategies" and argues that even if climate change isoverestimated, society will likely gain from theseactivities. Others have referred to such approachesas "win-win" L: "no regrets" strategies, althoughdefinitions of what constitutes a win-win or noregrets approach may differ greatly.
There are three other premises that oftenunderlie the mitigation perspective, although theyare not shared by all proponents of this perspective.These include a concern that economic modelingapproaches that suggest that the costs of reducingCO2 emissions would be prohibitive are narrowlybased and do not adequately account for long-termbenefits; a fear that if global warming is not abated,the future of the planet is in jeopardy; and a belief
that discussion of opportunities for adaptation ismisplaced and even morally wrong. I will discusseach of these premises.
Miller, Mintzer, and Brown (1990, p. 70) haverecently argued that "cost-benefit analyses ascurrently applied to the greenhouse problem do notprovide sufficient or appropriate grounds foropposing immediate action to mitigate globalclimate change." They base their conclusion onseveral arguments. Most troubling in their view isthat such approaches tend to emphasize short-termcosts without fully assessing long-term benefits.This is because the short-term costs are much easierfor economists to identify than the long-termbenefits. Nevertheless, they argue that economicmodels cannot simply ignore "the fact that the fullrange of consequences of changing one of the earth'smajor physical systems cannot be foreseen" (p. 71).Miller, Mintzer, and Brown are also concerned thateconomic modeling approaches (particularly cost-benefit analysis) do not adequately address thetemporal and spatial distribution of globalwarming's potential consequences, particularly withrespect to the fact that developing countries mayface a much greater burden from global warmingthan developed countries. Finally, they argue thatbecause economic models tend to emphasize short-term costs, these models do not fully appreciate thecapacity for technological innovation to reduce thecosts of CO2 mitigation in the long run.
Much of the support for strong efforts atmitigation is underpinned by a concern that therisks of unabated global warming are much greaterthan the costs associated with controlling green-house gas emissions. The formal statement fromthe 1988 Conference on the Changing Atmosphere,which called for a 20% reduction in global CO2
emissions by 2005, concluded that global warmingas well as other threats to the global atmospherethreatened international security, the world'seconomy, and the Earth's natural environment(WMO 1988).* Indeed, this threat was rankedsecond only to global nuclear war. The analogybetween global warming and nuclear war has been
*The conference, held in Toronto, was sponsored by the WMO and United Nations Environment Programmeand organized by the government of Canada. More than 300 scientists and policymakers from 46 countriesattended.
36
Adaptation and Mitigation Strategies
made more emphatically by others to underscore theargument tu^t global climate change threatens thefuture of modem civilization (for example, see Dotto[1988]). One need not turn to predictions of doom,however, to find expressions of real concern over thepotential impacts of global climate change. Forexample, the Working Group II report of theIntergovernmental Panel on Climate Change (IPCC)outlined a range of damaging impacts that climatechange might have on global agriculture andforestry, natural ecosystems, water resources,human settlements, and the world's oceans andcoastal zones (IPCC 1990).
Finally, some advocates of the mitigationperspective have questioned the reasoning thatunderlies the adaptation perspective. The criticismslevied at the NAS adaptation panel are a case inpoint. Jessica Mathews, for example, writes in theNAS synthesis report that the adaption perspectiveis based on a misguided assumption "that humaneconomic activities are largely divorced from natureand that modern technology effectively buffers us[society] from climate" (footnote to the adaptionchapter in NAS [1991], p. 45). Jane Lubchenco, amember of the NAS adaptation panel, objected tothe panel's conclusion that adaptation to climatechange in the United States may be relatively easy,calling such a conclusion "complacent." Rather,Lubchenco argues that the findings of the adaptionpanel should support a policy of mitigation.Elsewhere, Mathews (1987) has argued thatadaptation may be as costly and as politicallydifficult to achieve as mitigation, that adaptationmay widen the gap between rich countries that havethe capacity to adapt and poor countries that do not,and that mitigation will be necessary eventually,regardless of prospects for adaptation. O'Riordanand Rayner (1991) argue that for some proponentsof mitigation (preventivists, to use their label), it ismorally and ethically wrong to endorse a strategythat would allow humans to distort naturalprocesses when it is in the power of society toprevent this from happening.
4.3 Adaptation or Mitigation:A Matter of Emphasis
It would be difficult to identify any individual,organization, or nation that precisely fits either the
adaptation or mitigation approach as I haveoutlined them above. In reality, there are fewproponents of either perspective that argue for oneposition at the exclusion of the other. In otherwords, adaptationists recognize the need formitigation, and mitigationists accept the need forsome adaptation. For example, the NAS synthesispanel recommends actions that will lead both tomitigating and to adapting to climate change (NAS1991b); Mathews (1987) notes that even anaggressive strategy to prevent climate change willnot eliminate the need for some adaptation; andCrosson and Rosenberg (1991) explain that whileadaptation and mitigation strategies can be seen assubstitutes, they are also complementary. Thedifference in these perspectives is one of emphasis,guided in part by how proponents of one perspectiveor the other see the uncertainties and risksassociated with global warming, how they approachthe question of equity in determining how torespond to the problem, and how they view the roleof technology.
Both adaptationists and mitigationists recognizethat there is considerable uncertainty about theeventual magnitude of the global warming problem,its potential impacts, and the costs associated withtaking action to either mitigate or adapt to climatechange. They differ, however, on how theyapproach the risks associated with this uncertainty.Adaptationists, for example, tend to believe thatthis uncertainty supports a pragmatic approach thatdoes not lock society into costly and imprudentpolicies that, in the future, could prove misguidedand uninformed. In light of the uncertainty,adaptationists place emphasis on the risks (oropportunity costs) associated with taking action torespond to climate change. They point to thepotential loss in economic growth, lost opportunitiesfor development among poor countries, and themisallocation of limited resources to prevent anuncertain event from taking place. Mitigationists,on the other hand, see the uncertainty as one of theprimary factors supporting immediate action toforestall the problem. This uncertainty underscoresthe prospect of potentially catastrophic changes inthe world's climate that could possibly be avoided bytaking action today. Thus mitigationists emphasizethe risks associated with not taking action toaddress the global warming problem: the loss ofeconomic growth because of climate change, the loss
37
Section HI
of natural ecosystems, and the unfair burden thatthe impacts of climate change would place ondeveloping countries.
To some degree, adaptationists place moreemphasis on intragenerational equity, whilemitigationists place greater emphasis on inter-generational equity, although the adaptationists donot dismiss the issue of intergenerational equity.Proponents of adaptation are concerned with costsimposed on both developed and developingeconomies by pursuing potentially misguided effortsand controlling CO2 emissions. Resources spent oncontrolling emissions in the present may be betterspent on improving the capacity of both developedand developing countries to respond to climatechange in the future. Thus, with respect todeveloping countries, the best approach to climatechange may be improving their prospects for long-term economic development rather than forcingthem to control emissions of CO2. Mitigationists,while not ignoring intragenerational equity, areclearly concerned with the fate of future generationsand the Earth itself. Thus, imposing costs onsociety in the present is justified as necessary inorder to guarantee a liveable planet for futuregenerations. Furthermore, mitigationists see manystrategies for controlling greenhouse gas emissionsas measures that are justified in the present forother economic and environmental reasons. Theytherefore question the short-term costs of thesemeasures.
Both adaptationists and mitigationists see animportant role for technology as part of the solutionto the global warming problem. Adaptationists are,perhaps, more optimistic about the prospects oftechnological solutions than are mitigationists, butmitigationists do not discount the role thattechnology will play in developing mitigationstrategies. Most adaptationists believe that withthe exception of some catastrophic change in climate(which they see as a remote risk), the outlook forsociety to adapt to future climate change throughtechnological innovation is positive. Indeed,humanity has demonstrated a remarkable capacityto adapt to all sorts of changes in the past.Mitigationists question this rather optimisticassessment. They believe that the timing and scaleof the changes that society is likely to face fromclimate change will be unprecedented in humanhistory. Furthermore, they question the overall
success to which society has been able to bufferitself against changes in climate and the environ-ment in the past. Nevertheless, mitigationists dotend to believe that technology will reduce the costsof mitigation activities in the long run.
O'Riordan and Rayner (1991) note thatproponents of adaptation and mitigation do notdisagree over the facts of global wanning, ratherthey differ on what those facts mean. Bothadaptationists and mitigationists are concernedwith the implications of global warming for presentand future generations and for the Earth itself.Adaptationists, however, are concerned with howresponding to global warming might affect societytoday, while mitigationists are concerned with whatfailing to respond to global warming will mean forfuture generations. Adaptationists tend to see theglobal warming problem as manageable, whereasmitigationists tend to see the global warmingproblem as preventable. These differencesdemonstrate how making policy decisions aboutcomplex scientific and technical issues areinfluenced as much by values as they are by thefacts of the problem. It is likely that during thenext decade, ongoing efforts to better understandhow global warming might affect climate, and thusaffect economies and natural ecosystems around theworld, will greatly add to the knowledge base onwhich to design policies for responding to globalwarming. It is unlikely, however, that the policychoices will somehow become self evident simplybecause there is more information. The choice ofwhat, if anything, to do about global warming willultimately rest as much on values as it will on facts.
S THE PERSPECTIVE OF DEVELOPINGCOUNTRIES
It is worth looking briefly at how developingcountries view the global warming problem for tworeasons. First, the success of any internationaleffort to address the global warming problem willrequire the support of developing countries.Developing countries presently account forapproximately 28% of global CO2 emissions fromthe burning of fossil fuels (Ogawa 1991). Ifdeforestation is included — and estimates of itscontribution vary considerably — developingcountries could account for more than 40% of totalglobal carbon emissions (estimate derived from data
38
Adaptation and Mitigation Strategies
presented in Ogawa [1991] and Dannstadter[1991]). Equally important is the fact that the rateof increase in carbon emissions is much greater inthe developing countries than in the developedcountries. Because of the rapid rate of populationgrowth and industrialization, developing countrieswill provide the bulk of new carbon emissions in thecoming decades.
Second, and directly related to the theme of thispaper, many developing countries are not propo-nents of either adaptation or mitigation. Manydeveloping countries fear that they are trapped bythe global warming issue. They have neither thecapacity to mitigate CO2 emissions nor the flexibil-ity to adapt easily to changing climate conditions,yet they are likely to be among the most vulnerablecountries to the impacts of climate change (seeJodha [1989]). Developing countries are alsoconcerned that measures to control global CO2 emis-sions will have a negative impact on their efforts todevelop. Most developing countries will find it verydifficult, if not impossible, to sacrifice economicdevelopment in order to curb global CO2 emissions.It is impossible for developing countries to separatethe global warming issue from other issues ofeconomic, political, and social development. Theseinclude problems such as poverty, environmentaldegradation, political reform, and foreign debt.
Furthermore, there is an element of distrust anddisbelief on the part of developing countries towardthe developed countries. Many developing countriesbelieve that their contribution to the globalwarming problem is being overstated and aresuspicious of the international attention focused ondeforestation and environmental degradation in thedeveloping world. Some may even believe that thedeveloped countries are cynically using the globalwarming issue to limit their development.Goldemberg and Durham (1990) explain that manydeveloping countries "feel they are being trans-formed into scapegoats in a problem for whichdeveloped countries are equally, if not more,responsible." Developing countries also feel thatsince it is the developed countries that created theproblem, the developed countries must act first toresolve it.
The world's developing countries do not see thequestion of responding to global climate change as achoice between adaptation and mitigation; the issue
for them is rather how to foster development (withor without climate change). Developing countries donot dismiss global warming; on the contrary, manytake the problem very seriously, and some are veryactive in current efforts at negotiating a globalclimate change agreement. Developing countries,however, are quick to point out that the choicebetween adaptation and mitigation is a luxury fordeveloped countries.
6 CONCLUSIONS
It would be unfair to say that the generalizeddescriptions of adaptation and mitigation that Ihave provided in this paper capture the range ofviews on global warming. Bather they merelyrepresent two contrasting views along what could bedescribed as a continuum. While the adaptationand mitigation perspectives lie on different sides ofthis continuum, they are not polar opposites. Theextremes of this continuum might best be describedas the naysayers and the doomsayers. It is likelythat most perspectives about global warming (manyof which are a mix of adaptation and mitigation) liesomewhere along this continuum, between what Ihave described as the adaptation and the mitigationapproaches. Furthermore, not all perspectives onglobal warming can be neatly placed along thiscontinuum. Developing countries, for example, tendnot to see the global warming problem as a choicebetween adaptation and mitigation. Whatdifferentiates perspectives along this continuum(and off it, as well) is not whether eachacknowledges the facts of the global warmingproblem (except, perhaps, for the naysayers) butrather how each interprets those facts. Differencesbetween the adaptation and mitigation perspectivesreflect differences in the values that underpin thecurrent policy debate over how society shouldrespond to the global warming problem. Thesevalues will guide decision making about globalwarming as much as scientific and technicalinformation will.
If perspectives toward the global warmingproblem lie along a continuum in which componentsof both adaptation and mitigation mix, thenfocusing on adaptation and mitigation separatelymay not be the ideal way to conceptualize theproblem.. If this is the case, the challenge is todevise a new way of conceptualizing the global
39
Section III
warming problem that more fully captures therange of perspectives toward global warming andhow these perspectives might interact. However,simply recasting the problem is not likely to resolvethe polarizations that exist between two seeminglycontrasting views. This polarization is based moreon values that it is on the technical or conceptualmerits of various policy strategies. Nevertheless,improving our ability to conceptualize and discussvarious perspectives and policy strategies could be auseful step toward addressing the issue of values.
7 REFERENCES
Arrhenius, S., 1896, On the Influence of CarbonicAcid in the Air upon the Temperature of the Ground,Philosophical Magazine, 47:237.
Ausubel, J.H., 1991, Does Climate Still Matter?,Nature, 350:649-52.
Callendar, G.S., 1938, The Artificial Production ofCarbon Dioxide and Its Influence on Temperature,Quarterly Journal of the Royal MeteorologicalSociety, 6*4:223-240.
Crosson, P.R., and N.J. Rosenberg, 1991, Adaptingto Climate Change, Resources, 703:17-20.
Darmstadter, J., 1991, The Economic Cost ofCO2Mitigation: A Review of Estimates for SelectedWorld Regions, RFF Discussion Paper ENR91-06,Resources for (he Future, Washington, D.C.
Dotto, L., 1988, Thinking the Unthinkable:Civilization and Rapid Climate Change, WilfridLaurier University Press, Waterloo, Ontario,Canada.
Glantz, M.H., 1988, Politics and the Air Around Us:International Policy Action on Atmospheric Pollutionby Trace Gases, in Societal Responses to RegionalClimatic Change: Forecasting by Analogy, pp. 41-72,M.H. Glantz (editor), Westview Press, Boulder,Colo.
Goldemberg, J., and E.R. Durham, 1990, Amazoniaand National Sovereignty, InternationalEnvironmental Affairs, 2:22-39.
Houghton, J.T., G.J. Jenkins, and J.J. Ephraums,1990, Climate Change: The IPCC ScientificAssessment, Cambridge Univer,1 y Press,Cambridge, England.
IPCC, 1990, Potential Impacts of Climate Change,report prepared for the Intergovernmental Panel onClimate Change by Working Group II, WorldMeteorological Organization, Geneva, Switzerland.
Jodha, N.S., 1989, Potential Strategies for Adaptingto Greenhouse Warming: Perspectives from theDeveloping World, in Greenhouse Warming:Abatement and Adaptation, pp. 147-158,N.J. Rosenberg et al. (editors), Resources for theFuture, Washington, D.C.
Kempton, W., 1991, Lay Perspectives on GlobalClimate Change, Global Environmental Change,7:183-208.
Lave, L.B., 1988, The Greenhouse Effect: WhatGovernment Actions Are Needed?, Journal of PolicyAnalysis and Management, 7:460-470.
Mann, D.E., 1983, Research on Political Institutionsand Their Response to the Problem of IncreasingCO2 in the Atmosphere, in Social Science Researchand Climate Change, pp. 116-145, R.S. Chen et al.(editors), D. Reidel, Dordrecht, the Netherlands.
Mathews, J.T., 1987, Global Climate Change:Toward a Greenhouse Policy, Issues in Science andTechnology, 3(3):57-68.
Miller, A., I. Mintzer, and P.G. Brown, 1990,Rethinking the Economics of Global Warming,Issues in Science and Technology, 7(l):70-73.
Morrisette, P.M., and A.J. Plantinga, 1991, GlobalWarming: A Policy Review, Policy Studies Journal,70:163-172.
NAS, 1991a, Policy Implications of GreenhouseWarming: Report of the Adaptation Panel, NationalAcademy Press, Washington, D.C.
NAS, 1991b, Policy Implications of GreenhouseWarming — Synthesis Panel, National AcademyPress, Washington, D.C.
40
Adaptation and Mitigation Strategies
Nordhaus, W.D., 1991, Economic Approaches toGreenhouse Warming, in Global Warming:Economic and Policy Responses, pp. 33-66,R. Dombusch and J.M. Poterba (editors), MIT Press,Cambridge, Mass.
Ogawa, Y., 1991, Economic Activity and theGreenhouse Effect, The Energy Journal, 72:23-35.
O'Riordan, T., and S. Rayner, 1991, RiskManagement for Global Environmental Change,Global Environmental Change, 7:91-108.
Revelle, R., and H.E. Suess, 1957, Carbon DioxideExchange Between Atmosphere and Ocean and theQuestion of an Increase of Atmospheric C02 duringthe Past Decades, Tellus, 9:18-27.
Riebsame, W.E., 1990, The United States GreatPlains, in The Earth as Transformed by HumanAction, pp. 561-575, B.L. Turner II et al. (editors),Cambridge University Press, New York, N.Y.
Roberts, L., 1991, Academy Panel Split onGreenhouse Adaptation, Science, 253:1206.
Schilling, T.C., 1983, Climatic Change:Implications for Welfare and Policy, in ChangingClimate: Report of the Carbon Dioxide AssessmentCommittee, pp. 449-482, National Academy Press,Washington, D.C.
Schelling, T.C., 1991, Economic Responses to GlobalWarming: Prospects for Cooperative Approaches, inGlobal Warming: Economic and Policy Responses,pp. 197-221, R. Dornbusch and J.M. Poterba(editors), MIT Press, Cambridge, Mass.
Schneider, S.H., 1989, The Greenhouse Effect:Science and Policy, Science, 243:771-781.
Solow, A.R., 1991, Is There a Global WarmingProblem?, in Global Warming: Economic and PolicyResponses, pp. 7-28, R. Dornbusch and J.M. Poterba(editors), MIT Press, Cambridge, Mass.
WMO, 1988, Conference Proceedings: The ChangingAtmosphere, WMO/OMM - No. 710, WorldMeteorological Organization, Geneva, Switzerland.
WMO, 1990, Second World Climate Conference:Conference Statement, World MeteorologicalOrganization, Geneva, Switzerland.
41
Section
Comment 1 on Workshop in Adaptation and Mitigation Strategies
Why Greenhouse Warming Stays a Hot Topic
Dr. Rob CoppockPanel on Policy Implications of Greenhouse Warming
National Academy of SciencesWashington, D.C.
The rapidity with which greenhouse warmingburst onto the national and international politicalagendas is surprising. So too is the fact that it hasremained of central interest despite the lack ofunderstanding of the phenomenon exhibited by thegeneral public. Even with lackluster publicresponse, politicians and governments around theworld are advocating costly actions designed tocounter greenhouse warming.
A certain amount of attention and concern isnecessary to establish and sustain the attention ofgovernment decision makers. There are severalattributes of the greenhouse warming problem thatgenerated enough attention and concern to propel itso quickly onto the international agenda and keep itin the forefront for action.
First, it is one of a new set of global problemsthat is intimately connected to scientific analysis. Agreat deal of data has been collected and analyzedsince the early 1960s. Scientists have been care-fully laying the groundwork for decades and have asolid foundation for addressing the problem. Theywere ready in 1988 to capitalize on the NorthAmerican drought as a vehicle to bring the longer-term problem of greenhouse warming to more wide-spread attention. In short, there is a large body ofknowledge about the problem and possibleremediative actions.
Second, greenhouse warming is a vivid problemwith considerable psychological impact. Followingclose on the heels of the antarctic ozone hole andmore widespread depletion of stratospheric ozone, italso demonstrates human capacity to directly alterthe physical planet on which we depend forsurvival. Greenhouse warming is symbolic of someof our deepest fears.
The third attribute is the way worldwidescientific activity makes up for periodic reductionsin the importance accorded greenhouse wanning inany particular country. Scientists and other expertsfrom all parts of the world are working on thecauses and consequences of greenhouse warming.For example, the Intergovernmental Panel onClimate Change involved about 600 scientists fromall parts of the world. This unusually large groupof geographically dispersed experts means that thelevel of attention in a particular country can dropand international momentum still be maintained.For example, activity may develop first in theUnited States, which stimulates work in Europethat continues after interest wanes in this country.Later, that European activity may rekindle atten-tion here. The extent of international activity bothsmooths over lulls in any particular location andincreases the number of overall actions that may beundertaken.
Professional reinforcement provides the fourthelement that helps sustain attention to greenhousewanning. The problem would never have beenrecognized were it not for scientists. And it isscientists and other experts who provide the forceunderlying the present concern about greenhousewarming. Whether they work in universities,industry, or government, they derive much of theirmotivation from their professional colleagues. Oncestarted, a groundswell of concern swept throughprofessional channels and provided the basis forgovernment attention to greenhouse warming. Thepolitical base for action on greenhouse warming is ascientific constituency rather than the generalpopulation.
The fifth element also derives from the level ofscientific attention given the topic. The sheer
42
Adaptation and Mitigation Strategies
magnitude of research and the frequency withwhich new results are reported keeps attention ongreenhouse wanning. Hardly a week goes bywithout a scientific meeting somewhere on someaspect of greenhouse warming. With so muchactivity, there is bound to be a fairly regular flow ofnew ideas and findings. These, in turn, stimulateeven more analysis.
Another aspect of the global appeal ofgreenhouse warming is the fact that the greenhouseproblem involves all human activities in all partsthe world. People everywhere will feel theconsequences. More importantly, actions of eachindividual are relevant. Every human can make acontribution, however small, to slowing the onset ofgreenhouse warming.
The final element is the fact that many peoplewould like to see some of the actions that wouldslow greenhouse warming taken for other reasons.Greenhouse warming is thus an issue with naturalpolitical coalitions. This enhances the appeal ofgreenhouse wanning to many individuals andorganizations.
All of these elements combine to makegreenhouse wanning a unique issue. In essence,greenhouse warming is a political issue-based on aconstituency of scientists and experts. Because itcrosses political boundaries, however, it hasachieved a level of interest beyond other scientificand technical problems. And because of its naturalbase among these experts, it rapidly achievedprominence once access to the political agenda wasopened. And the intrinsically global character ofthe problem has provided momentum to keep itthere.
The strength of the influence of the scientificand expert communities on greenhouse warmingpolicy may lead to different policy positions invarious countries. For example, industrial, scien-tific, and other elites generally are more directlyinvolved in determining national policy in Europeancountries than in the United States. The fact thatEuropean countries currently advocate carbon taxes,emissions caps, and other policy responses togreenhouse warning much more strongly than theUnited States is consistent with a more direct policyinfluence of scientists in those countries who aredeeply concerned about greenhouse warming.
43
Section III
Comment 2 on Workshop in Adaptation and Mitigation Strategies
Mitigation and Adaptation
Dr. Rob CoppockPanel on Policy Implications of Greenhouse Warming
National Academy of SciencesWashington, D.C.
Discussions of greenhouse warming policy oftenconfuse the meaning of the terms "mitigation" and"adaptation." This confusion occurs partly becausethese terms were developed to describe physicalprocesses but have subsequently been used inappro-priately as labels for policy strategies. It alsoderives in part from misunderstanding of the basiccharacter of the two categories of intervention.
The National Academy of Sciences report PolityImplications of Greenhouse Warming develops quitespecific definitions. The NAS report refers tomitigation as actions that alter the rate of change inthe radiative balance of the planet It describesadaptation as actions that assist human or naturalsystems adjust to changed climatic conditions. Inother words, mitigation and adaptation eachdescribe actions that bring into play chains of causeand effect that lead to quite different physicalconsequences.
An important difference between mitigation andadaptation is the time at which the benefits of thesephysical processes become manifest The benefits of
mitigation actions begin comparatively soon afterthey are taken, whereas the benefits of adaptationemerge only after climatic conditions change.
Mitigation and adaptation are sometimesconfused as categories, however, because individualinterventions can lead to both types of conse-quences. For example, planting trees would bothoffset emissions of carbon dioxide — qualifying asmitigation — and contribute to an overall biospherebetter able to cope with increased atmosphericconcentrations of greenhouse gases — also qualify-ing as adaptation. Similarly, actions that improveenergy conservation and efficiency both reducegreenhouse gas emissions and alter long-termdemand for energy.
Thus the terms "mitigation" and "adaptation"are best thought of as descriptions of specificphysical phenomena that derive from given actions.It is misleading to refer to mitigation strategies oradaptation strategies because most practical actionprograms contribute to both types of consequences.
44
Section IVWorkshop in Political Institutions
Institutional Analysis and Global Climate Change:Design Principles for Robust International Regimes*
Professor Michael McGinnisPolitical Science Department
Professor Elinor Ostrom, Co-DirectorWorkshop in Political Theory and Policy Analysis
Indiana UniversityBloomington, Indiana
1 THE ROAD TO RIO
Scientific evidence suggests that humanactivities have a significant effect on the world'sclimate. Political pressures are growing to establishpolitical institutions at the global level that wouldhelp manage the social and economic consequencesof climate change. Disagreements remain about themagnitude of these effects, as well as the regionaldistribution of the detrimental consequences ofclimate change. In this paper we do not wish toenter into the complexities of these technicaldebates. Instead, we wish to challenge a seeminglywidespread consensus about the nature of thepolitical response appropriate to this globaldilemma. Specifically, we question the extent towhich the "answer" can be said to reside primarilyin the establishment of the new global institutionslikely to emerge from the first "Earth Summit" —the United Nations (UN) Conference onEnvironment and Development — scheduled forJune of 1992 in Rio de Janeiro.
In a single paragraph in a major reportprepared for this process (MacNeill et al. 1991,p. 110), the ambitious tasks awaiting the heads of
state who will attend this meeting are summarizedas follows:
• "Adopt an 'Earth Charter' setting outbroad directions for development andembodying new principles to governrelationships between governments,peoples, and the planet in the twenty-first century."
• "Adopt an agenda for action —'Agenda 2 1 ' . . . [that] will set out aninternationally agreed-upon workprogram, including targets of nationaland international performance for each ofthe critical issues . . . . The Agenda willinclude the estimated costs of achievingthose targets, and ways and means toprovide the necessary financial resources.It will also contain an operationalcommitment to transfer the needed tech-nologies to developing countries. Mostimportantly, it will designate thenational and international agencies thatwill bear responsibility for the first phaseof implementation."
•Copyright 1992 by authors. We appreciate the comments made by David Feeny, Roy Gardner, Barry Hughes,Leo Hurwicz, Richard Kosobud, Vincent Ostrom, Karen Rasler, Vern Ruttan, Phil Schrodt, and Paul Turneron an initial draft of this paper. Financial support of the National Science Foundation (Grant SES-4843901)and the Workshop in Political Theory and Policy Analysis is gratefully acknowledged by Ostrom andMcGinnis, respectively.
45
Section IV
• "Review the negotiations underway onseveral conventions, in particular those tolimit global warming, halt net deforesta-tion, and preserve the planet'sbiodiversity."
• "Initiate significant reform of our inter-national institutions to enable nations tomanage global interdependence and toimplement Agenda 21 as well as theconventions."
In MacNeill et al. (1991), the Secretary Generalfor this "summit," Maurice Strong, stronglyapplauds these "far-reaching reforms of theinternational system" and argues that the "casemade for broadening the concept of national securityto include [environmental] threats is compelling"(p. xi). Strong also suggests that the current UNTrusteeship Council should be "transformed into aforum in which the nations of the world wouldexercise their 'trusteeship' for the integrity of theplanet as a whole, including the global commons"(p. 124). MacNeill and his co-authors extol this ideaof a new "Earth Council" as "another one of theseinnovative formulations [that] reflects the quality ofimagination and the level of ambition that needs tobe applied in developing new forms of governance toguide the planet through the next turbulentdecades" (p. 124). The idea of an Earth Council istaking form in a proposal for a World Environmentand Development Forum (WEDF) that would meetannually to coordinate the implementation of theprotocols adopted as part of this process (p. 126,Figure 5.2).
It is difficult to know whether the passing dayslead toward a process that will enable individuals tobetter understand and confront the many multi-tier,complex problems that are called environmentalthreats or whether the process may itself threatenfuture capabilities to cope effectively with theseproblems.1 If most of the efforts are directedtoward reaching agreement on protocols thatrequire drastic changes in the emissions of specificchemicals and the creation of new institutionalarrangements to enforce these measures, then the"cure" could be worse than the "illness."2 Theprocess may be one of the many useful steps if mostof the efforts are directed toward (1) gaining betterunderstandings of complex, nested problems and(2) encouraging small-, medium-, and larger-scale
institutional arrangements to adopt innovativemeasures to cope more effectively with local,regional, and global environments.
The current emphasis on global solutions basedon international conventions meant to establishglobal institutions to manage environmental changemay be fundamentally misguided. Comprehensivenegotiation sessions can too easily degenerate intopublic theater. The resulting treaties may amountto nothing more than words on paper. The policiesof national governments, who are called upon totake the initiative to prevent deforestation anddesertification, have, in many instances, been amajor source of these problems. Why shouldofficials of national governments, who have obtaineda major part of their income from large-scale agro-forestry concessions, suddenly turn around and haltthese lucrative opportunities? Further, even ifprotocols move from paper tigers to regulations withteeth, no guarantee exists that the emissions beingreduced are the correct ones in a particularenvironment.
Global climate change is, by definition, a long-term process. Any effective political response mustbe based on a similarly long-range perspective. Thecurrent state of scientific knowledge stronglysuggests that changes already underway willcontinue to affect the world's climate for manydecades into the future. The political, economic,and social consequences of these climatic changesremain uncertain, but institutional innovations canbe made now that will facilitate our ability to adjustto future disruptions. Any such institutionalarrangement, however, must be long-lasting androbust with respect to uncertain and dramaticchanges. Few political institutions satisfy thesedemanding conditions.
Certain forms of global management might evenintensify problems. This could be expected if thisprocess undermines existing institutions that enableindividuals to govern and manage many local andregional commons or cause these groups to losetheir capability for innovation and discovery. Thisis certainly what has happened in many developingcountries where control over local resources wasshifted from indigenous populations to centralgovernments (Feeny 1988a; Arnold and Campbell1986; Messerschmidt 1986; Gadgil and Iyer 1989;Thomson 1977). There is a lot more to learn about
46
Political Institutions
complexly nested environments as well as complexlynested institutions before presuming that globalregulation is the way to improve global conditions.Many proposed policy changes involve issues thatcould severely exacerbate conflict between thedeveloped and developing parts of the world andeven among neighbors within a region.3 Thus,regardless of the truth or falsity of currentpredictions, issues that can affect future levels ofinternational cooperation and/or conflict are activelybeing discussed in many substantive arenas.
We presume that major efforts will be made toalter institutions so as to respond to the perceivedthreat of global wanning. We hope that thesechanges would be based on the best evolvingscientific knowledge about factors that affect globalclimate patterns. We also hope that adoptedchanges would be based on the best evolvingscientific knowledge about the factors that affecthuman behavior as this behavior in turn affectsglobal climate. Policy changes based on aninadequate understanding of how individualsrespond to institutional change can be just asdamaging to our common future as changes basedon an inadequate understanding of the generalcirculation models.
This paper should be viewed as our effort toreview recent literature on international regimes,institutional analysis, and on international securityand arrive at preliminary understandings of lessonsthat apply to designing appropriate institutionalresponses to the dilemmas of global climate change.We begin by briefly surveying the broad range ofsubstantive areas studied by researchers oninternational regimes. After summarizing the basicprinciples and implications of institutional analysesof common pool resource regimes, with a particularemphasis on the "design principles" thatcharacterize robust common-pool resource (CPR)regimes, we justify our extension of the results ofresearch on cooperation in small-scale CPRsituations to problems of global climatic change.We argue that a major difference is the diversity ofactors involved in the global arena and the need tocarefully "nest" institutional arrangements at thelocal, regional, and global levels. We offer a roughtypology of the collective actors typically engaged inmonitoring and sanctioning of internationalregimes. The diversity of possible combinations ofactor types suggests some cautionary notes about
the contribution formal models can make to ourcollective understanding of institutional responsesto global climatic change. We cone1 Je with a briefexamination of the lessons for h- jader issues ofglobal cooperations that can be derived from priorefforts at arms control.
2 AN OVERVIEW OFINTERNATIONAL REGIMES
Analysis of international cooperation hastypically examined the "international regimes" saidto exist in certain substantive areas of internationalinteractions. In this research tradition, regimes aredenned as follows:
sets of implicit or explicit principles, norms,rules, and decision-making proceduresaround which actor expectations converge ina given area of international relations.Principles are beliefs of fact, causation, andrectitude. Norms are standards of behaviordefined in terms of rights and obligations.Rules are specific prescriptions or proscrip-tions for action. Decision-making proced-ures are prevailing practices for making andimplementing collective choice (Krasner1982, p. 186).
Despite consensus on this definition, scholars donot always have the same thing in mind whendiscussing regimes. The term is used to cover broadareas such as the GATT system of internationaltrade as well as limited areas such as internationaltrade in crafts made from illegally obtainedelephant ivory. Some authors (e.g., Porter andBrown 1991, p. 20) require that a regime be basedon principles of international law established bysome treaty or convention, while others focus on theregularities of patterns of relationships rather thanformal agreements.
Haggard and Simmons (1987) argue thatregimes differ by strength, organizational form,scope, and allocational mode, but no systematicinventory of international regimes exists. Jacobson(1984) provides a systematic list of internationalgovernmental and nongovernmental organizations,but no necessary correspondence exists between theestablishment of a formal international organizationand the existence of an international regime.
47
Section IV
Classic works in the international political economyliterature focus on cases selected to demonstrate thepoints emphasized in that analysis. For example,Keohane and Nye (1989) contrast global regimes inoceans and money, as well as bilateral U.S.relations with Canada and Australia; Keohane(1980) compares the "strength" of regimes in theareas of oil, money, and trade; Krasner (1985)examines different patterns of conflict betweenNorth and South over regimes concerning civilaviation, multinational corporations, oceans,shipping, Antarctica, and official capital transfers.Proposals for new regimes often draw on lessonsfrom earlier regimes, as in the Sebenius (1991)evaluation of the prospects of a regime on globalwanning in light of the Ozone Convention and theLaw of the Sea Treaty. Of course, many otherstudies offer detailed examinations of a singleregime (see Aggarwal 1985; Benedick 1991;Sebenius 1984).
Some indication of the variety of arenas ofcollective action covered by the term "regime" iswarranted. Textbooks on international politicaleconomy (e.g., Spero 1990; Blake and Walters 1987)typically examine regimes in the following broadareas:
• International trade (commodities,agricultural products, manufactured goods,and services),
• Money (currency exchanges and financialtransactions),
• Development aid,
• Energy (especially oil), and
• Multinational corporations (includingtechnology transfer issues).
In each of the first four areas, discussion of aspecific international organization dominatesanalysis, namely, GATT, IMF, the World Bank, andOPEC, respectively. Many analyses focus ondiffering patterns within a given regime, such ascontrasts between international trade in agri-cultural products and high technology. Also, muchattention has been given to abortive efforts toestablish a New Internationa] Economic Order(NIEO) that would have provided an even more
comprehensive framework for the internationaleconomy.
Introductory world politics textbooks (e.g.,Hughes 1991, pp. 281-290) typically discuss avariety of arrangements concerning arms control,centered around the following major arms controltreaties (and associated acronyms):
• Global regimes- Partial Test Ban Treaty (PTB) and
Threshold Test Ban Treaty- Nuclear Non-Proliferation Treaty (NPT)
• U.S.-Soviet treaties- Strategic Arms Limitation Treaty (SALT) I
and II and ABM Treaty- Strategic Arms Reduction Treaty (START)- Intermediate Nuclear Forces Treaty (INF)- Hotline Agreement and other Confidence-
Building Measures (CBM)
• Regional arrangements- Treaty of Tlatelolco (Latin America)- Conventional Forces in Europe Treaty
(CFE)
Harlan Cleveland (1990/91, pp. 11-13) focuseson the following 12 examples of successfulinternational cooperation (using his terms):
• Weather forecasting,
• Eradication of infectious diseases,
• International civil aviation,
• Allocation of the frequency spectrum,
• Globalization of information flows,
• Agricultural research and development,
• UN peacekeeping and peacemaking,
• Cooperation in outer space,
• Law of the Sea Treaty,
• UN High Commissioner for Refugees(UNHCR),
48
Political Institutions
• Ozone Treaty, and
• Antarctic Treaty.
Cleveland explicitly excludes arms controlarrangements. He also briefly mentions regionalarrangements — like the European Community,NATO, and control of Mediterranean pollution(P. Haas 1989) — and global efforts of the WorldBank and UN1CEF.
Young (1989, p. 351) offers a relativelysystematic listing of international interactions onnatural resources and the environment, includingsome areas lacking any effective means ofinternational cooperation:
• Wildlife (whales, northern fur seals, polarbears);
• Marine life;
• Wild fauna and flora (trade inendangered species);
• Deep sea-bed minerals;
• Electromagnetic spectrum (broadcastfrequencies, orbital slots);
• Regional pollution;
• Radioactive fallout;
• Stratospheric ozone;
• Polar regions:- Antarctica,- The Arctic (no regime);
• Acid precipitation:- Europe,- Other areas (no regime);
• Biological diversity (no regime); and
• Global climate change (greenhouse effect)(no regime).
These lists already overlap to some extent, but afew important areas must be added, includingexamples of some unsavory business indeed:
• Human rights,
• International law of war,
• War crimes,
• Terrorism,
• Internationa] trade in illegal drugs,
• International arms trade (includingballistic missiles), and
• International trade in toxic waste.
Many other lists could be cited, but theseexamples should suffice to demonstrate thelimitations of relying on classification schemesbased on the substantive issues involved. Wepropose to use the theoretical framework ofinstitutional analysis to help suggest ways oforganizing and extending work on internationalregimes.
Debates on institutional responses to globalchange are dominated by the time-worn mind-setsof The State and The Market On the one side,economists (and other analysts) sing the praises ofmarket processes and advocate the definition ofindividual property rights as the standard solutionto most problems of resource use. On the otherside, advocates press for a single global standard ofbehavior developed in an international convention,the global stand-in for a centralized state. Afterexpressing concern for the lack of formalized rulesor institutions in a particular issue area, this secondgroup of advocates typically calls for establishmentof a new international regime, complete with aformal organization with either an I (forInternational), G (for Global) or W (for World) in itsacronym. Both markets and states are appropriatesolutions for some circumstances. However, thesetwo types of institutions do not, by any means,exhaust the range of possibilities.
3 AN OVERVIEW OF INSTITUTIONALANALYSIS
Analysis and evaluation of alternatives to stateand market has been the focus of the InstitutionalAnalysis and Development (IAD) framework that
49
Section IV
has evolved out of several decades of work bycolleagues associated with the Workshop in PoliticalTheory and Policy Analysis (Riser and Ostrom 1982;E. Ostrom 1986; Oakerson 1986; Ostrom, Feeny,and Picht 1988; V. Ostrom 1991a,b). In particular,we will draw on research on institutionalarrangements in the management of common poolresources, usually in fairly small-scale settings.4 Amajor focus of this work has been on micro-level,self-organizing, and self-enforcing capabilities.More difficult to understand and essential for thequestion at hand is how macro-political ordersenhance or detract from the capabilities of thosedirectly involved in problem solving to create newinstitutions or reform prior institutions to cope moreeffectively with their own problems. Work on multi-level institutional problems is underway but is farmore difficult than work at a single level ofanalysis.
This work is relevant to the analysis of globalchange for two fundamental reasons. First, thesubstantive natures of many local and globalproblems are similar. Second, despite the vastdifferences in scale involved in local and globalcommons, the configurations of the situations atthese levels are fundamentally similar. Thus, thetheoretical principles underlying successfulcooperation are also similar. Third, any globalregime that undermines the requisites for successfulcooperation at the local level is unlikely to besustainable in the long run.
The IAD framework helps identify the broadcomponents and relationships among thosecomponents that are considered important for workwhen doing institutional analysis. To undertakeanalysis of a problem — such as how changes ininstitutions related to global warming may affectpatterns of international cooperation and conflict —requires that the analyst draw on or developtheories that spell out more fully the specificvariables involved in an analysis and the nature ofthe mechanisms that relate variables to oneanother. For well specified problems, the analystmay be able to develop formal models that generatequite specific predictions within the range of themodel's applicability. (We evaluate the prospectsfor formal models of international cooperation onglobal climate change in a subsequent section.)
To engage in an institutional analysis, one firsthas to understand the structure of incentives facingactors in a situation, what choices are most likely tobe made, and how these choices tend to produceparticular patterns of outcomes. In other words, thefirst question is: What type of situation is this? Todo this, one identifies a focal, conceptual unit, calledan action arena, that is utilized to identify thestructure of a problem and thus explain relevantbehavior and outcomes. Action arenas include anaction situation and the types of actors in thatsituation. An action situation is composed ofparticipants holding diverse positions who makedecisions from among the alternative choices madeavailable to them in light of the information theypossess about the structure of the situation, the waytheir actions link to outcomes, and the benefits andcosts they will receive. To make a theoreticalprediction about the type of choices that varioustypes of actors will make in a situation, one needsto know: (1) the resources that an actor brings to asituation; (2) the valuation that actors assign tostates of the world and to actions; (3) the way actorsacquire, process, retain, and use knowledgecontingencies and information; and (4) the processesactors use for selecting particular courses of action.
Formal, game-theoretical models can beconstructed to analyze the expected behavior ofparticipants. Many situations can be analyzedusing these components, including competitivemarkets, principal-agent relationships, legislativeand committee decisions, the harvesting of resourceunits from common-pool resources, and thetemptation to steal resource units from a common-pool resource. For highly simplified and stylizedversions of these "universal" situations, it is evenpossible to establish the configuration ofinstitutional rules that yield optimal results(Hurwicz 1973). When the parameters of themodels capture most of the important parametersinvolved in field settings, these models can be usedto explain the types of outcomes obtained in thefield and the likely response of individuals tochanges in the incentives that the individuals face(see, for example, Plott and Meyer 1975; Shepsle1979a,b, 1989). Applying models to field settingswhose structures differ in crucial aspects from themodels results in mis-specified analysis, just as it
50
Political Institutions
does when a physical model is inappropriatelyapplied.
To address whether the models used to analyzepolicy-relevant situations (e.g., decisions related toharvesting forests or using energy sources) areappropriate requires that analysts dig deeper thanthe initial givens of a model, into the underlyingstructure of a situation that generates these givens.Here the focus of an IAD approach is on three broadclasses of underlying factors: (1) relevant aspects ofthe physical, biological, and technological world;(2) relevant aspects of the broad cultural under-standing and valuation of that world; and (3) thenested configurations of rules that affect thestrategies that actors may, must not, or must takein the situation of interest. The key problem indesigning effective institutions to control adverseeffects on the environment is matching the specificset of rules that are to be used with the attributesand dynamics of the physical, biological, andtechnological world of relevance in a particularcultural milieu. In order to simplify our analysis,we will not address the potentially important role ofcultural values on behavior toward an environmentfor our initial analysis and will focus first on theproblem of matching rule configurations to complexphysical worlds.
MATCHING RULES TO PHYSICALWORLDS
4.1 The World of Private Goods
When individuals engage in transactions relatedto private goods, matching rules to physical worldsis relatively simple. Goods are considered to be"private goods" when:
• All aspects of the goods and services thatindividuals produce, distribute, buy, andsell can be excluded from actors whowould like to use them (or, alternatively,others can avoid having externalities of aprocess dumped on them) and
• Any good consumed by one individual isunavailable for consumption by others(Ostrom and Ostrom 1977).
Rule configurations that generate an open,competitive market are optimally efficient whenused in relation to private goods (de Alessi 1988).Unfortunately, only a limited number of the goodsand services used by individuals possess theattributes of private goods. As soon as the goodsinvolved do not have these nice characteristics, itcan be shown that various aspects of "marketfailure" occur, and different types of institutionsother than those of an open, competitive marketmay be needed.
4.2 The World of Common-Pool Resources
Common-pool resources (CPRs) do not possessboth of the nice characteristics of private goods.CIRs are defined to be natural or manmaderesources in which (1) exclusion is nontrivial (butnot necessarily impossible) and (2) yield is subtract-able (Ostrom, Gardner, and Walker forthcoming).The global commons definitely share this firstattribute.6 If adverse effects to the stratosphericozone or to the factors that could lead to severechanges in weather patterns were to be averted, itwould be difficult — if not impossible — to excludeanyone who had not contributed to this effort.CPRs share the attribute of nonexclusion withanother broad class of problems that are referred toas public goods (Ostrom and Ostrom 1977).
The second attribute of CPRs that is sharedwith private goods but not with public goods is thesubtractability of the yield (Samuelson 1954).Another term used to describe this attribute isrivalness of consumption. Let us first illustrate thisattribute with examples drawn from research ongroundwater basins, irrigation systems, fisheries,forests, grazing lands, computer systems, bridges,and other natural or manmade resources thatgenerate a yield. All of these are resources thatmake available a flow of resource units over time.Examples of the resource and its resource unitsinclude: (1) a groundwater basin and acre-feet ofwater pumped, (2) a fishing grounds and tons offishharvested, and (3) a computer facility and process-ing time. The water withdrawn by a groundwaterpumper is not available to anyone else. Nor are thefish caught by one boat or the central processingunits used by one faculty member. The facilities orresources producing these units are, however, jointlyused by multiple individuals.
51
Section IV
When a few users first use renewable CPRs,what one person uses does not appear to subtractfrom what is available. Initially, the abundance ofunits relative to demand masks the subtractability.In some resources, such as fisheries, initialwithdrawals may actually increase the amountjointly available. Fish compete for food resources,and harvesting some fish allows more fish tosurvive. If there are no rules in place to limit theuse of a CPR, and if the resource unit that can bewithdrawn is valuable, then the structure of thesituation leads users to invest more and more inharvesting activities. After a threshold has beenreached, each individual's investment andharvesting activities adversely affect everyone elseusing the same resource.
Without rules to allocate resource units, userscompeting with one another for ever scarcer unitsmay engage in destructive races against oneanother, and their actions may destroy the veryresource that is generating valuable yield for them.Garrett Hardin captured a key aspect of theproblem of an open-access CPR in his now classicuse of the term "tragedy of the commons."6
Aristotle recognized a similar dilemma when hewrote: For that which is common to the greatestnumber has the least care bestowed upon it(Politics, Book II, Sec. 3). What Hardin did notcapture was the capacity of the individuals involvedin such tragedies to have sufficient insight into theproblems that they faced to restructure their ownrules and change the incentives they faced.
Where a resource produces a single, valuableresource-unit with a high level of predictabilityknown to all participants (or where storagefacilities, such as dams, enhance predictability), it ispossible to devise marketable rights or other simpleallocation rules that enable individuals to makeefficient long-term use of the resource (Hurwicz1973; Clark 1980). Fishery economists, forexample, drawing on the important work of H. ScottGordon (1954) and Milner Schaefer (1957), haverepeatedly demonstrated how a transferable quotasystem can solve the "tragedy" of a fishing commonsand lead to an efficient equilibrium (Scott 1979,1982). This has led to widespread acceptancewithin many natural resource ministries that thetransferable quota rule is the most preferred policyprescription for improving the efficient allocation offishery resources (Matthews and Phyne 1988). In
recommending transferable quotas as the favoredinstitutional solution for most fishery problems,however, Wilson and Townsend have pointed outthat the models used to generate this institutionalsolution are all based on a single-species fishery(Wilson 1982; Townsend 1986; Townsend andWilson 1987). Multi-species fisheries must becharacterized by different dynamic models thansingle-species fisheries. Complex patterns ofinteraction occur among species. Devising anefficient, effective, and equitable transferable quotasystem for a multi-species fishery is problematic(see Copes 1986; Schlager 1990).
4.3 The World of Configural Complexity
Many of the physical systems of relevance forunderstanding global environmental problems arenot adequately modeled with the simple models ofthe physical world that have been used as thefoundation for policy recommendations. Somemodels of land degradation, for example, presumethat any human intervention in a watershed thatincreases the silt load carried downstream producesa harm. Rawlins (1991) argues quite convincinglythat the movement of silt from one location toanother may enhance overall agriculturalproductivity. Whether the movement of siltimproves or degrades agricultural productivitydepends on many attributes (temperature, soildensity and depth, costs of transportation) thataffect the potential of higher versus lower sites.Thus, whether moving silt is land degradation orland improvement depends upon many other factorsthat are present in a unique configuration in anyparticular watershed.
Similarly, it appears that the effect of somechemicals on tropospheric and stratospheric ozonedepend on the amount of other chemicals present ina micro- or meso-environment. Interventions basedon incomplete understandings of these complexprocesses can inappropriately attempt to regulateall environments as if they were similar instructure. Such interventions can be ineffective(and thus inefficient because of their cost) or evencounterproductive.
A recent report released by the NationalResearch Council, for example, criticizes the U.S.Environmental Protection Agency (EPA) for focusing
52
Political Institutions
too strongly on one pollutant (Commission onGeosciences, Environment, and Resources 1991).Tropospheric ozone (smog) is generated when twoclasses of chemicals react in sunlight: (1) volatileorganic chemicals (VOCs) and (2) either nitric oxideor nitrogen dioxide (NOX) (see Stone [1992]). Insome regions of the U.S., a high ratio of VOCs toNOX exists. "In such regions," the report concludes,"it may be better to concentrate on the lessabundant pollutant to prevent the formation ofozone."
To make the situation even more difficult, onesource of VOCs has not been monitored untilrecently.7 Trees and most plants producebiogenic VOCs (such as isoprene). Heavily woodedareas would thus generate a high ratio ofunmeasured VOCs to NOX, even if most manmadeVOCs were eliminated. In such areas, applying auniform rule based on a model that smog can becontrolled by reducing VOCs would be an ineffectivestrategy. Given these recent findings, it may nowbe possible to understand why 20 years of heavyregulation has yielded little or no progress in thereduction of tropospheric ozone. Regulations havebeen based on an oversimplified, universal model ofsingle-source smog production. While all physicalprocesses follow universal laws, the particularprocesses to be controlled in any specific locationinvolve a complex, frequently nonlinear,combination of the specific values of multiplevariables.
Results from recent research on the effect ofdiverse anthropogenic aerosols on global warmingreveal spatially and temporally nonuniform pat-terns. Research reported on by R.J. Charlson andcolleagues from many of the major research labsrelated to Atmospheric Sciences in the January 24,1992, issue of Science shows a surprising effectconcerning the role of tropospheric aerosols inproducing a cooling effect. Charlson et al. (1992,p. 429) conclude that "anthropogenic sulfateproduces a mean radiative forcing of the [northernhemisphere] climate that is comparable inmagnitude but opposite in sign to the anthropogenicperturbation in forcing by greenhouse gases."Further, they stress both the spatially nonuniformdistribution of and large uncertainties about themagnitude of the presence of tropospheric aerosols.Trying to understand how these complex chemicalsinteract in different micro and meso climate systems
is essential if future regulations aimed to preventglobal warming are to be effective. Charlson andhis coauthors strongly urge a research program thatcombines data collection and modelling at smalleras well as global scales.
Another new study demonstrating nonuniformtemporal reactions has been conducted by ThomasR. Karl of the National Climatic Data Center inAsheville, North Carolina (and reported on in theFebruary issue of Scientific American; see Beardsley1992). Previous assessments of global wanninghave presumed that daytime and nighttimetemperatures would increase equally as the level ofgreenhouse gases increases. Karl's study showsthat temperatures in the Northern Hemispherehave increased primarily at night and have beenrecorded as daily minimum temperatures. "Dailymaxima, on the other hand — generally recordedduring daylight hours — have hardly increased atall in the U.S. and China and only a little in theSoviet Union" (Beardsley 1992, p. 21). Thesignificance of Karl's and Charlson's work iscurrently under debate. But if sulfate aerosols dotend to counteract warming processes — even thoseduring the day — the implications for regulationsthat focus on a reduction of pollution from fossilfuels might have counterproductive consequences.As Beardsley concludes (p. 24):
If sulfate aerosol is indeed amelioratingdaytime greenhouse warming, theimplications for public policy could beperplexing. Nighttime warming would stillbe ecologically disruptive and would stillmelt glaciers. And the provisions of the U.S.Clean Air Act and steps by other countriesto reduce pollution from fossil fuels mayhave an unintended effect. If the sulfatetheory is right... the decrease in sulfateemissions might hasten global warming.
While much more needs to be learned about thefactors affecting global warming, it does appear thatmany of the elements that enter into processes thatchange global weather patterns are not distributedin a spatially uniform manner. Since the effect ofreducing the levels of one chemical depend upon themixture of other chemicals present in theatmosphere, uniform rules are likely to producesome counterproductive results as well as someameliorative results. Successful efforts to control
53
Section IV
environmental problems may end up being thosethat are best matched to micro- or meso- levelenvironments rather than trying to devise rules at aglobal level. That suggests that there is much thatcan be learned from prior studies of smaller-scaleCPRs of relevance to global problems.
5 PRIOR RESEARCH ON ROBUSTCPR INSTITUTIONS
Prior research has included an intensiveanalysis of field settings where: (1) resource usershave devised, monitored, and followed their ownrules to control the use of a CPR and (2) theresource systems, as well as the institutions, havesurvived for long periods of time. The youngest setof institutions thus analyzed was already over 100years old. The history of the oldest institutionexceeds 1,000 years. All of the institutions studiedhave survived droughts, floods, wars, pestilence, andmajor economic and political changes. The robustCPR institutions originally included grazing andforest CPR institutions in Switzerland and Japanand irrigation systems in Spain and the PhilippineIslands (see E. Ostrom 1990). Since then, manymore cases of robust CPR institutions have beenstudied, particularly related to irrigation systems(see E. Ostrom 1992; Tang 1991,1992).
In CPR institutions that have survived for longperiods of time, operational rules-in-use have notnecessarily remained fixed since they were firstinitiated. All of these environmental settings arecomplex and variable over time. In such settings, itwould be difficult to get "the operational rules right"on the first try, or even after several tries. Theseinstitutions are "robust" or in "institutionalequilibrium" in the sense defined by Shepsle.Shepsle (1989, p. 143) regards "an institution as'essentially' in equilibrium if changes transpiredaccording to an ex ante plan (and hence part of theoriginal institution) for institutional change." Inthese cases, the participants designed basicoperational rules, created organizations toundertake the operational management of theirCPRs, and modified the rules-in-use over time inlight of past experience according to their owncollective choice and constitutional rules.
While substantial differences exist among therobust institutions, all share fundamental similari-
ties. All faced uncertain and complex environments.In contrast to the uncertainty of the environment,the populations in these settings remained stableover long periods of time. Individuals shared a pastand expected to share a future. It was important toindividuals to maintain their reputation as areliable member of a community. Extensive normshave evolved in all of these settings that narrowlydefined "proper" behavior. Many of these normsmade it feasible for individuals to live in closeinterdependence on many fronts without excessiveconflict. Further, gaining a reputation for keepingpromises, honest dealings, and reliability in onearena was a valuable asset. Prudent, long-termself-interest reinforced the acceptance of the normsof proper behavior. The most notable similarity ofall, of course, is the sheer perseverance of theseresource systems and institutions. The resourcesystems clearly meet the criteria of long-termsustainability.
The specific operational rules in each case differmarkedly from one another. Thus, they cannot bethe basis for an explanation across settings. Whilethe particular rules that are used within eachsetting cannot be the basis of an explanation for theinstitutional robustness and sustainability acrossthese CPRs, part of the explanation is based on thefact that the particular rules differ. By differing,the particular rules take into account specificattributes of the related physical systems, culturalviews of the world, and the economic and politicalrelationships that exist in the setting. Withoutdifferent rules, users could not take advantage ofthe positive features of a local CPR or avoidpotential pitfalls that could occur in one setting butnot others. Given the diversity of the settings, oneshould not expect to be able to discover a singleformulation or set of optimal mechanisms.
Instead of similar rules, there are similar designprinciples that characterize all of the robust CPRinstitutions. By "design principle" is meant anessential element or condition that helps to accountfor the success of these institutions in sustainingthe CPRs and gaining the compliance of generationafter generation of participants to the rules-in-use.8
Focusing on underlying principles rather thanspecific mechanisms may enable us to learn lessonsfrom a wide diversity of small field settings thathave relevance to the design of robust internationalregimes.9
54
Political Institutions
5.1 Design Principle 1: Clearly DefinedBoundaries
Individuals or households who have rights towithdraw resource units from the CPR must beclearly defined, as must the boundaries of the CPRitself.
Defining the boundaries of the CPR and of thoseauthorized to use it can be thought of as a "firststep" in organizing for collective action. As long asthe boundaries of the resource and/or the individ-uals who can use the resource remain uncertain, noone knows what they are managing or for whom.Without defining the boundaries of the CPR andclosing it to "outsiders," local participants face therisk that any benefits they produce by their effortswill be reaped by others who do not contribute tothese efforts. At the least, those who invest in theCPR may not receive as high a return as theyexpected. At the worst, the actions of others coulddestroy the resource itself. Thus, for any partici-pants to have a minimal interest in coordinatingpatterns of appropriation and provision, some set ofparticipants have to be able to exclude others fromaccess and appropriation rights. If there aresubstantial numbers of potential users and if thedemand for the resource units is high, the destruc-tive potential of all users freely withdrawing from aCPR could change the perception of users regardingthe likely continuance of a flow of resource-units inthe future. The more that participants fear for thefuture sustainability of the resource, the closer thesituation is to that of a one-shot dilemma, where thedominant strategy of all participants is to overusethe CPR. (This is the logic behind Aristotle'sassertion quoted above.)
5.2 Design Principle 2: Congruence betweenAppropriation and Provision Rules andLocal Conditions
Appropriation rules restricting time, place,technology, and lor quantity of resource-units arerelated to local conditions and to provision rulesrequiring labor, materials, and I or money.
Adding well-tailored appropriation and provisionrules helps to account for the perseverance of theCPRs themselves. Successful, long-enduring,irrigation institutions, for example, have all
developed methods to equate the costs of buildingand maintaining the irrigation system appropriatelyto the benefits that are achieved. Some examplesmay help the reader understand the diversity ofspecific rules that meet Design Principle 2.
• The Zanjeras of Northern Philippines. Theseself-organized systems obtain use-rights topreviously unirrigated land from a largelandowner by building a canal that irrigatesthe landowner's land and that of a zanjera.At the time that the land is allocated, eachfarmer willing to abide by the rules of thesystem receives a bundle of rights and dutiesin the form of Atars. Each Atar defines threeparcels of land located in the head, middle,and tail sections of the service area wherethe holder grows his or her crops. Responsi-bilities for construction and maintenance areallocated by Atars, as are voting rights. Inthe rainy seasons, water is allocated freely.In a dry year, water may be allocated only tothe parcels located in the head and middleportions. Thus, everyone receives water inplentiful and scarce times in rough propor-tion to the amount of Atars they possess.Atars may be sold to others with the permis-sion of the irrigation association and they areheritable (see Siy 1982; Coward 1979).
• Thulo Kulo in Nepal. When this system wasfirst constructed in 1928, 27 householdscontributed to a fund to construct the canaland received shares to the resulting systemproportionate to the amount they invested.The system has been expanded several timessince by selling additional shares. Measure-ment and diversion weirs or gates areinstalled at key locations so that water isautomatically allocated to each farmeraccording to the proportion of shares owned.Routine monitoring and maintenance areallocated to workteams so that everyoneparticipates proportionately, but emergencyrepairs require labor input from all share-holders regardless of the size of their share(see Martin and Yoder 1983; Martin 1986).
• The Huerta of Valencia in Spain. In 1435,84 irrigators served by two interrelatedcanals in Valencia gathered at the monasteryof St. Francis to draw up and approve formal
55
Section IV
regulations to specify who had rights towater from these canals, how the waterwould be shared in good and bad years, andhow responsibilities for maintenance wouldbe shared. The modern Huerta of Valencia,composed of these plus six additional canals,now serves about 16,000 hectares and 15,000farmers. The right to water inheres in theland itself and cannot be bought and soldindependently of the land. Rights to waterare approximately proportionate to theamount of land owned, as are obligations tocontribute to the cost of monitoring andmaintenance activities (see Maass andAnderson 1986; £. Ostrom 1990).
These three rule configurations differ markedlyfrom one another. The zanjeras are institutionaldevices for landless laborers to acquire use rights toland and water and might even be called communalsystems. The Thulo Kulo system comes as close toallocating private and separable property rights towater as is feasible in an irrigation system. TheValencian Huerta has maintained centuries-old landand water rights that forbid the separation of waterrights from the land being served. The Valenciansystem differs from both "communal" and "privateproperty" systems because water rights are firmlyattached to private ownership of land.
Underlying these differences, however, is thebasic design principle that the costs of constructing,operating, and maintaining these systems areroughly proportional to the benefits obtained.Slogans, such as "privatization" or "think globally,act locally" may mask important underlyingprinciples rather than providing useful guides forinstitutional reform.
5.3 Design Principle 3: Collective-ChoiceArrangements
Most individuals affected by operational rulescan participate in modifying operational rules.
CPR institutions that use this principle are ableto better tailor rules to local circumstances, sincethe individuals who directly interact with oneanother and with the physical world can modify therules over time so as to better fit them to thespecific characteristic of their setting. The presence
of good rules, however, does not account forparticipants following them. Nor is the fact thatthe participants themselves designed and initiallyagreed to the operational rules an adequateexplanation for centuries of compliance byindividuals who were not originally involved in theinitial agreement. It is not even an adequateexplanation for the continued commitment of thosewho were part of the initial agreement. Agreeing tofollow rules ex ante is an easy "commitment" tomake. Actually following rules ex post, when strongtemptations to act at variance with rules arepresent, is the significant accomplishment.
The problem of gaining compliance to rules —no matter what their origin — is frequentlyassumed away by analysts positing all knowing andall powerful external authorities that enforceagreements. In the cases described above, noexternal authority had sufficient presence to playany role in the day-to-day enforcement of the rules-in-use. Thus, external enforcement cannot be usedto explain high levels of compliance. This leads usto consider the fourth and fifth design principles.Since they are so closely related, we will discussthem both together.
5.4 Design Principles 4 and 5: Monitoringand Graduated Sanctions
Monitors, who actively audit CPR conditionsand participant behavior, are accountable to theparticipants or are the participants.
Participants who violate operational rules arelikely to be assessed graduated sanctions (dependingon the seriousness and context of the offense) fromother participants, by officials accountable to theseparticipants, or by both.
It has been surprising to find that monitoringand sanctioning are undertaken not by externalauthorities but by the participants themselves inthe robust institutions. The initial sanctions usedin these systems are also surprisingly low. Eventhough it is frequently presumed that participantswill not spend the time and effort to monitor andsanction each other's performance, substantialevidence exists from field settings as well as fromthe experimental lab (see Ostrom, Gardner, andWalker 1992).
56
Political Institutions
To explain the investment in monitoring andsanctioning activities that occurs in these robust,self-governing, CPR institutions, the term "quasi-voluntary compliance" used by Margaret Levi(1988a, chapter 3) to describe the behavior oftaxpayers in systems where most taxpayers complyis very useful. She uses the term quasi-voluntarycompliance to describe taxpayer behavior in suchtaxing regimes. Paying taxes is voluntary in thesense that individuals choose to comply in manysituations where they are not being directly coerced.On the other hand, it is "quasi-voluntary becausethe noncompliant are subject to coercion — if theyare caught" (Levi 1988a, p. 52). Taxpayers,according to Levi, will adopt a strategy of quasi-voluntary compliance when they have:
confidence that (1) rulers will keep theirbargains and (2) the other constituents willkeep theirs. Taxpayers are strategic actorswho will cooperate only when they canexpect others to cooperate as well. Thecompliance of each depends on thecompliance of the others. No one prefers tobe a "sucker" (Levi 1988a, p. 53).
Levi views coercion as an essential condition toachieve quasi-voluntary compliance as a form ofcontingent behavior. Enforcement increases theconfidence of individuals that they are not suckers.As long as they are confident that others arecooperating and the ruler provides joint benefits,they comply willingly to tax laws.
The costs of monitoring are kept relatively lowin many of the long-enduring CPUs as a result ofthe rules-in-use. Water rotation systems, forexample, usually place the two actors mostconcerned with cheating in direct contact with oneanother. The irrigator who nears the end of arotation turn would like to extend the time of histurn (and thus the amount of water obtained). Thenext irrigator in the rotation system waits nearbyfor him to finish and would even like to start early.The presence of the first irrigator deters the secondfrom an early start, and the presence of the secondirrigator deters the first from a late ending.Neither has to invest additional resources inmonitoring activities. Monitoring is a by-product oftheir own strong motivations to use their waterrotation turn to the fullest extent.
Institutional analysis that simply posits anexternal, zero-cost enforcer has not addressed thepossibility that the rules devised by participantsmay themselves have a major effect on the costs,and therefore the efficiency, of monitoring byinternal or external enforcers. Similarly, monitorswho do a good job are rewarded, and inadequatemonitors are punished.
Further, since the users tend to continuemonitoring the guards as well as each other, someredundancy is built into the monitoring andsanctioning system. A failure to deter rule breakingby one mechanism does not trigger a cascadingprocess of rule infractions, since other mechanismsare in place. Thus, the costs and benefits ofmonitoring a set of rules are not independent of theparticular set of rules adopted. Nor are theyuniform in all CPR settings. When participantsdesign at least some of their own rules (DesignPrinciple 3), they can learn from experience to craftenforceable rather than unenforceable rules. Thismeans paying attention to the costs of monitoringand enforcing as well as the benefits that those whomonitor and enforce the rules obtain.
5.5 Design Principle 6: Conflict ResolutionMechanisms
Participants and their officials have rapid accessto low-cost, local arenas to resolve conflict amongparticipants or between participants and officials.
In many models of rule-governed situations, therules that structure the strategies available toparticipants are unambiguous as well as enforced byexternal, all-knowing officials. In field settings,applying rules is never unambiguous, even whenthe participants themselves are the monitors andsanctioners. For individuals who are seeking waysto slide past or subvert rules, there are always waysthat they can "interpret" the rule so that they canargue they meet it while subverting the intent.Even individuals who intend to follow the spirit of arule can make errors.
If individuals are going to follow rules over along period of time, some mechanism for discussingand resolving what is or is not a rule infraction isnecessary to the continuance of rule conformanceitself. While the presence of conflict resolution
57
Section IV
mechanisms does not guarantee that participantsare able to maintain enduring institutions, it isdifficult to imagine how any complex system of rulescould be maintained over time without suchmechanisms. In some cases, conflict-resolutionmechanisms are quite informal, and those who areselected as leaders are also the basic resolvers ofconflict. Furthermore, networks of face-to-facecommunication are essential to the fashioning,revision, and maintenance of mutualunderstandings and social accountability.
5.6 Design Principle 7: Minimal Recognitionof Rights to Organize
The rights of participants to devise their owninstitutions are not challenged by externalgovernmental authorities.
Participants frequently devise their own ruleswithout having created formal, governmentaljurisdictions for this purpose. As long as externalgovernmental officials give at least minimal recogni-tion to the legitimacy of such rules, participantsmay be able to enforce the rules themselves. But ifexternal governmental officials presume that onlythey can make authoritative rules, then it is verydifficult for local participants to sustain a rule-governed CPR over the long run. At any pointwhen participants wish to break their rules, theycan go to the external government and get localrules overturned.
5.7 Design Principle 8: Nested Enterprises
Appropriation, provision, monitoring,enforcement, conflict resolution, and governanceactivities are organized in multiple layers of nestedenterprises.
All of the more complex, enduring CPRs meetthis last design principle. This is probably one ofthe most important design principles when onebegins to think about global rather than local CPRs.If different environmental zones need to beregulated in diverse manners in order to achieve thesame outcomes, it is quite essential that there begovernance mechanisms at all levels and ways ofeffectively communicating and resolving conflictbetween units at one level and across levels.
6 THE GLOBAL RELEVANCE OFRESEARCH ON LOCAL CPR REGIMES
There are several reasons why we think thelessons learned from past research on micro CPRshave substantial applicability to the problem ofdealing with patterns of international cooperationand conflict over the potential threat of globalwarming. These include:
• The analytical structure of some of theproblems related to global warming sharessimilar features with the analytical structureof many local CPRs.
• Starting with theories and models devised forthe analysis of local CPRs may speed thework in developing theories and models at aglobal level.
• Many of the problems related to globalwarming (e.g., deforestation) are themselvesthe result of inadequate solutions at a microlevel of a complementary and interactivecommons problem.
To the extent that global warming sharesanalytical attributes with other CPRs, thetheoretical and empirical lessons learned fromstudying micro-level phenomena may "scale-up" sothat direct lessons may be learned. It is certainlythe case that the attribute of nonexdusion isshared. Is the second attribute of "subtractability"shared as well? Yes, it would appear so — only inthe reverse direction. The term "rivalness" may bean easier term to use in this connection. If onethinks of carbon emissions as "using up" a naturalcapacity to absorb carbon in the atmosphere, thenthe units of natural capacity used up by carbonemissions in one country are not available to beused up by any other country. Similar to fisheries,initial carbon emissions may actually generate jointbenefits for everyone in terms of increased plantgrowth induced by higher levels of CO2 in theatmosphere.
A second major point of similarity is the sharedconcern for finding means to cooperate despite theabsence (or disinterest) of a higher level of politicalauthority. Research on international regimes hasemphasized the ability of governments to cooperatein an environment of "anarchy," defined as the
58
Political Institutions
absence of any central authority at the global level.Political order at the international level is primarilyself enforcing, in the sense that governmentsfrequently resort to military force to achieve theirgoals or to restore the "peace." Despite its obviousshortcomings on normative grounds, war remainsthe ultimate means of conflict resolution that maynever be dispensed with until a complex, multi-tiered world government is established in someremote future time.
Since many of the micro settings whereindividuals have jointly devised their owninstitutions have existed far from the purview ofnational, regional or local government officials, thedesign of these institutions and their enforcementhas not always depended on the presence of a"State." Many of the self-organized institutionsdevised in these micro settings are not dependentupon enforcement of their rules by external officials.Further, since many of these micro institutions areneither a market nor a state, we may gain insightto alternative international institutions fromstudying these micro institutions.
Considerable caution has to be exercised inmoving from micro to macro levels, since a changein scale frequently changes the structure ofsituations dramatically. On the other hand, it isnot appropriate to dismiss the possibility of lessonsacquired from studying micro institutions applyingto macro institutions simply on the basis of adistinction between supposedly homogeneous localgroups and the heterogeneous actors involved in theglobal system. Many successful CPR regimes arecomplex structures and include individuals fromdiverse cultural groups. Homogeneity may make iteasier for groups to attain the degree of commonunderstanding essential to the operation of robustCPR regimes, but none of the design principleslisted above require actor homogeneity. Even verydiverse actors can arrive at a degree of commonunderstanding sufficient to enable them toeffectively deal with their common problems.
Analyzing complex patterns of interaction in thesimplest possible example of the structure (thebiologist's strategy) may enable the analyst tounderstand the way things work in the simplestexemplar without being overwhelmed by the sheercomplexity of the problem. To the extent that workat a micro level does not easily scale up, it still may
be the case that key variables are identified inmicro-level theories and models. Starting with thework at this level may save considerable time whencontrasted with an effort to move to the global leveldirectly. Even though it is impossible to completelydecompose global problems into entirely discretesubproblems, working upward may enable solutionsto be reached faster than starting from the globaland working downward — at least for some subsetof problems (see Ashby 1960; Polanyi 1951).
As an example of a global problem rooted ininadequate local arrangements, consider thecontribution of deforestation to global climaticchange (Cruz 1991). Deforestation may be theresult of conversion from lower-valued to higher-valued uses. It may also be the result of either(1) ill-defined property rights where the transactioncosts of gaining clear property rights (whetherstrictly private or communal) are very high, or(2) the creation a concession economy as thefoundation for resource mobilization by national-government officials (for their government as wellas for themselves). Problems like deforestation willdeteriorate still further until more effectiveinstitutional arrangements can be devised at thelocal level. Given the varieties of physicalenvironments that exist within even one country, nosingle set of rules related to the use of forests willcreate appropriate incentive systems for allindividuals whose decisions affect the regenerationof forests over time.
As discussed above, CPRs do not have the "nice"attributes that are needed for a market to workeffectively. Nor do inflexible hierarchicalarrangements work effectively in relationship tospatially nonunifonn problems. The inability toprescribe markets or hierarchies as solutions toCPR problems has profound implications for thestudy of institutions related to internationalcooperation and conflict over global change. First, itmeans that instead of seeing global change asanother "issue area" where a market or an"international regime" could be created to regulatethe problem, one should first attempt to understandthe key linked situations that affect global weatherpatterns. Some of these situations create incentivesleading actors to use various combinations of energysources in different parts of the world. Othersrelate to the structure of incentives leading actors toclear previously forested areas. Still others have to
59
Section IV
do with the structure of incentives leading actors toincrease or decrease the number of animals grazingin different parts of the world (for their effect onsurrounding terrain and their production ofmethane gas). While all of these situations jointlyaffect the level of CO2 in the world's atmosphere,the logic behind these situations differssubstantially.
Second, it means recognizing that proposedchanges in rules (or in technology) will havedifferent effects around the world. The attempt, forexample, to create a worldwide carbon emissionpermit system is likely to have drastically differenteffects in (1) open political economies that havealready established competitive markets andpolitical-legal systems that are relatively predict-able, responsive, and accountable than it would in(2) closed political economies with mercantilistmarket systems and unchecked and rampant rent-seeking among public officials. Ignoring for themoment the extraordinarily difficult task ofachieving the agreement of most national govern-ments to a particular permit system, theimplementation and monitoring of a permit systemin the United States is an entirely different problemthan the implementation and monitoring of thatsame permit system in almost any country of thedeveloping world.
Third, recognizing that the outcomes ofindividual choices that affect global change occur inmany differently structured situations that areembedded in a wide diversity of institutional andcultural settings leads to a further recognition thatno uniform set of rules imposed at the internationallevel is going to "solve" the problem of globalwanning. Whether the wide diversity of variablesthat affect global weather patterns are changed soas to reduce the risk of global warming (or,alternatively, reduce the adverse consequences thatglobal warming would have in some locations) willnot depend alone (or perhaps even strongly) onsweeping changes written in treaties made at large-scale, international meetings and ratified bynational governments. Such meetings may indeedbe crucially important if they provide anopportunity for open discussion of truly difficult andcomplex problems that need multiple solutions atmultiple scales. But the change in human under-standing of the nature of the problems involved maybe far more important in the first instance for a
change in behavior than a new internationalconvention whose pronouncements are nothingmuch more than words on paper.
On the other hand, we are cautious about casualanalogies. We fully recognize that there aresubstantial differences in the range of actorsinvolved in many global problems. In many of themicro studies, the actors involved are individualswhose own livelihood depends strongly on solutionsto use a resource system more efficiently over time.In many local commons around the world, individ-uals do devise for themselves rather ingeniousinstitutional arrangements that have enabled themto make productive use of fragile resource systemsover long periods of time. Even then, there aremany "tragedies of the commons" in these situationswhere individuals as principals interact directlywith one another. Once we move to the problem ofthe global commons, in addition to the millions ofindividuals who make choices, we add manycorporate actors who are designated as the agentsfor complex publics, and their behavior is far morecomplex. Yet many corporations and nationalgovernments can reasonably be expected to exist forlong periods of time, providing their agents withincentives to establish predictable rules of behaviorthat reduce future levels of uncertainty (Keohane1984). One issue that we emphasize below is theextent to which the nature of these collective actionorganizations affects the patterns of monitoring andsanctioning that can realistically be expected tooccur in any global regime.
7 A PRELIMINARY TYPOLOGY OFINTERNATIONAL ACTORS
Many of the prior applications of the IADframework have been to situations where the actorsare individual persons or firms. One of thechallenges of our current assignment is to apply thisframework to situations where the actors involvedare far more diverse and where the corporate natureof the actors is particularly relevant to predictionsabout activities and outcomes. A diverse array oforganizational actors participate in the global arena,including national and subnational governments,international governmental organizations (IGOs),domestic economic corporations and multinationalcorporations (MNCs), domestic nongovernmentalorganizations (NGOs), and international
60
Political Institutions
nongovernmental organizations (INGOs). Wepropose a typology of actor types based on the three-fold characterization of political, economic, andepistemic contingencies that constitute patterns oforder in human societies (V. Ostrom 1991b). Thepoint of this typology is not to locate any particularorganization in some pigeonhole but rather toidentify the combinations of interests andcapabilities that most commonly impinge oninternational cooperation.
Groups of individuals form different types ofcollective action organizations to pursue diversegoals. Individual agents in different organizationsare then authorized to undertake specific actions.In addition to their personal goals, agents ofeconomic corporations are driven by pursuit ofprofit; political agents are driven by a desire toremain in office and to achieve certain policy goals;and agents of NGOs strive to attain the purposesthat inspired formation of that type of organization.Some INGOs, such as Amnesty International, arewidely regarded as impartial sources of informationrelevant to various areas of international relations.Others, such as Greenpeace, become direct actiongroups that play a distinctive, confrontational rolein political processes.
Even though any particular association may besimultaneously engaged in economic production aswell as rule-making and knowledge-generatingactivities, different types of associations tend toemphasize one or the other type of activity. InTable 1, we use these characteristic goals and/orarenas of activity to differentiate among the type ofactors most relevant to analyses of internationalcooperation. (We differentiate among governmentsgrounded in fundamentally different forms ofpolitical order below.)
It is important to treat actors in the epistemiccategory as having distinctive interests andcapabilities rather than relegating them to aresidual "other" category left after political andeconomic actors have been identified. For example,Peter Haas (1989) emphasizes the role played bytechnical experts in establishing the politicalpreconditions for effective control of pollution in theMediterranean Sea. He argues that regimes aremost effective when based on consensus among atechnically informed "epistemic community" aboutthe nature of the problem to be solved and the
TABLE 1 Classification of Actor Types
Primary Arenaof Activity
Political
Economic
Epistemic
Scope
Domestic
Nationalgovernments;other bureau-cratic agents
Domesticcorporations
NGOs; technicalexperts
of Activities
International
IGOs
MNCs
INGOs; epistemiccommunities
merits of alternative solutions. In other words,technical experts share a "common understanding"that can provide a basis for political cooperation.The phrase "epistemic community" frequentlyappears in analyses of the prospects for increasedinternational cooperation on environmental issues(P. Haas 1992). Technical experts may play a largerole in monitoring governmental or corporatebehavior, by providing an impartial source ofinformation upon which sanctions could be based.Even if the information provided by "public interest"groups is seen as biased, competing informationprovided by more blatantly self-interested corporategroups can also be taken into account. It isimportant to realize that no one group, nor any typeof collective actor, has a monopoly on legitimateparticipation in these inherently political processes.
For our purposes, a particularly importantdistinction is the scale of the activities of collectiveaction organizations. We have in mind especiallythe standard distinction between "domestic" and"international" organizations (see Table 1). Theprominence given "sovereign states" in the inter-national relations literature and in practice requiresthis distinction be used for some analyticalpurposes, even though many important actors andnetworks transcend national boundaries. Ingeneral, agents of collective action organizationsformed by members of the same political communitymay tend to pursue different goals than agents oforganizations drawn from multiple communities.For example, agents of national governments pursuedifferent goals and fulfill different roles than do the
61
Section IV
"international civil servants" that carry out theactivities of IGOs. Similarly, domestic corporationsmay have interests that are systematically differentfrom multinational corporations, especiallyregarding such matters as restrictions on foreigntrade or currency exchanges. Although domesticand international NGOs tend to resemble each othermore closely, in some circumstances the distinctionbetween NGOs and INGOs may be important.
Some organizations are difficult to locate in anyone of these categories. Should cartels of drugdealers be treated as multinational corporations?Or should drug cartels and terrorist groups beincluded in the same NGO (or INGO) category asreligious and scientific organizations? Do somenational governments engaged in the internationalarms trade tend to take on the interests andcharacteristics of purely economic agents? Despitesome remaining ambiguities, these basic distinctionsamong governments, IGOs, domestic corporations,MNCs, and domestic and international NGOsremain valid for many analytical purposes.
8 NEGOTIATIONS AND THE DESIGNOF NESTED REGIMES
Collective-action organizations with a widevariety of interests will contend in upcomingnegotiations to establish a global convention onlimiting emissions of greenhouse gases. Thegovernments of advanced industrial democraciesmust respond to pressures from domestic environ-mental groups as well as the economic corporationsthat produce the vast array of consumer goodsdeemed essential to our modern society. Govern-ments of developing countries must strike a balancebetween external inducements to minimize environ-mental degradation (especially deforestation anddesertification) and domestic pressures for economicgrowth and development. Multinational anddomestic corporations seek stable and predictablelaws and regulations that will facilitate theircontinuing pursuit of profits. Private voluntaryorganizations will press the claims of scientists andenvironmentalists, no doubt too vigorously for thecomfort of the leaders of established governmentsand corporations. Finally, international civilservants will press for establishing additional levelsof bureaucracy at the international level as well asinsuring access to more regularized sources of
income. The task of reconciling such a widediversity of interests is truly formidable.
Although we cannot provide a comprehensivesurvey of the interests of the organizations that willbe involved in establishment of a global climateregime, we can discuss some general designprinciples that may facilitate establishment of astable and robust regime. The "design principles"discussed above are relevant to the development ofany regimes intended to be robust and long lasting.We realize that difficulties exist in scaling upprinciples of local commons regimes to the globalscale, but it is not necessary to apply theseprinciples to global institutional arrangement as asingle regime. It may indeed be counterproductiveto visualize establishment of a single, integrated,global management scheme covering all issues ofglobal warming, deforestation, desertification, andother forms of environmental degradation.
A more appropriate conceptualization is basedon the idea that sustainable global regimes mustmake sense at all levels of aggregation — local,regional, national, transnational, and global.Institutional arrangements at these multiple levelsmust be nested in such a way that the institutionsat each level are robust with respect to the type ofchallenges that are likely to arise at that level. Forif the local manifestations of a global regime fail tosatisfy the design principles identified above, thenthat regime is subject to fragility and eventualdissolution.
This nesting cannot be expected to happenautomatically. Quite the contrary is the case, sincesuccessful management of resources by some localgroup often imposes negative externalities on othergroups. In such circumstances, some means mustbe found by which representatives of affected groupscan come together to establish mutually beneficialinstitutional arrangements. Despite the currentemphasis on state-to-state negotiations, this type ofinteraction should be expected to take place at alllevels of aggregation.
This process itself raises the standard problemsof principal-agent relations, that is, the difficulty ofinsuring that an agent selected to act in the nameof a principal will indeed pursue the interests of theprincipal rather than acting solely out of selfinterest. Ideally, arrangements at local and
62
Political Institutions
intermediate levels are incentive compatible, in thesense that the interests of agents are somehowmade to coincide with the interests of the principal.The design principles identified above, when appliedto institutional arrangements at the local level, maybe said to constitute necessary but not sufficientconditions for establishment of a robust globalregime. In order to establish a robust, multilevelregime, additional design principles can be used toguide efforts to make institutional arrangements atthe local, regional, and global levels mutuallyreinforcing.
One important design principle relevant to alllevels of aggregation is that institutional arrange-ments must remain tied to the physical realities ofthe environmental situation. Thus, if the bound-aries of a river basin do not correspond to theofficial boundaries among sovereign states, groupswithin all affected states must be included (whetheror not all national governments are directlyrepresented). An example of relatively successfulcooperation among a large group of heterogenousgovernments is the control of pollution in the Medi-terranean Sea (P. Haas 1989,1990). Intergovern-mental negotiations should not be considered theonly, nor even the best, forum for the cooperationneeded to address the consequences of environ-mental change. Sometimes international agree-ments play an important role, but they rarely (ifever) tell the whole story. Although it might seemobvious that a global problem like greenhousewarming necessarily requires a solution that issimilarly global in scope, no global regime canremain robust if it neglects to take account of localcircumstances or the conflicting interests of smaller-scale collective action organizations.
Perhaps the most important nesting principle isthat the interests of all relevant groups must beincorporated in the ultimate agreement, or else theregimes's sustainability will be undermined by thosegroups whose interests are excluded. It is notenough to impose an arrangement by majority vote(of, for example, the UN General Assembly), forthose groups whose interests are not representedwill dismiss the result as unfair or illegitimate.Instead, we should expect complex linkages betweendifferent issues, as occurred in negotiations over theLaw of the Sea Treaty (Sebenius 1984). Negotiatorsshould also be attentive to possibilities for side
payments of various sorts, in order to complete amutually advantageous package deal.
In addition to the standard problems ofcollective action, negotiations over global climatechange are complicated by an unusual degree ofinherent uncertainty. Although some regions maybenefit from global warming while others suffer, theidentity of potential winners and losers remainsvery uncertain. Also, regions that benefit fromearly aspects of climatic change might later sufferdire consequences as changes accumulate oraccelerate. In effect, all governments and otherorganizations may consider themselves to bepotential losers, simply because even successfuladjustments to changing environments are likely toprove costly. But these costs remain in the future,whereas the policy changes being proposed mayimpose very real short-term costs on certaingovernments or corporations.
This is an example of a classic dilemma ofcollective action: a large group of potentialbeneficiaries facing diffuse and uncertain gains ismuch harder to organize for collective action thanclearly defined groups who are being asked to suffereasily understandable costs. To overcome theresistence of "blocking coalitions" most adverselyaffected by any agreement (Sebenius 1991), one oftwo types of leadership is crucial. First, smallprivileged groups may find it in their interests todirectly provide the public good (Olson 1965).Second, some political (or other) entrepreneurs maytake the lead in educating members of the relevantgroups as to the nature of the common problem theyface. The record of robust CPR regimes reviewedabove clearly demonstrates the ability of individualsto solve difficult dilemmas of collective action evenin the absence of a dominant leader.
In the international political economy literature,the role of the United States in establishing thepost-World War II economic order is emphasized, asare the difficulties of achieving policy coordinationin the absence of a hegemon (Keohane 1984).Advocates of a more active environmental policy areoften disappointed about the lack of U.S. leadershipin the second sense, but recent work has empha-sized the leadership role played by "epistemiccommunities" of technical experts in the establish-ment of a wide range of regional or international
63
Section IV
regimes, including the Montreal Protocol on ozonedepletion (P. Haas 1992). Indeed, scientific groupshave played a major role in bringing ihe issues ofglobal climate change to the attention of policy-makers and the mass public. But the crucial stepwill still require joint action on the part of a largenumber of diverse groups. Our basic point is thatthese discussions should be informed by the lessonsdrawn from analyses of the principles underlyingsuccessful patterns of cooperation at the local level.
Reliance on hegemonic leadership to establishcooperation is short sighted, because that hegemonicposition is unstable in the long run (since free-ridersgain relative to a hegemon paying for public goodsproduction). Although technical experts can helpeducate the public about the nature of the problemof climate change, an even more effective techniqueis available to those seeking to design a sustainableclimate regime. As long as individuals and groupspursue their own self interest, a regime is mostlikely to be sustainable if it so configures individualand group interests as to support the achievementof collective goals.
Designers of a global regime to deal withclimatic change can draw important lessons fromthe designers of the U.S. Constitution, whoconcentrated on finding ways to balance poweragainst power, to use contestation among individ-uals to provide a firmer foundation for achievingpublic order (see V. Ostrom 1987). Rather thanseeing the diversity of actors and interests as animpediment to agreement, it is important to put thisdiversity to use in order to increase the robustnessof any resulting regime. In particular, the diverseinterests of groups can be used to monitor variousaspects of the agreement. If the monitoringactivities of different groups overlap to some extent,then each can serve as a check on the other,providing increased confidence that sufficientcompliance will be forthcoming.
Statistics provided by official governmentsources can play some role in helping to monitorgreenhouse emissions, but they cannot remain thesole source of information. If so, then governmentofficials might have strong incentives to misreportlevels in order to protect their own government frompunishments, and they might also be subject tobribery by corporate officials eager to cover up theirown excessive use of carbon fuels. Insuring regular
alternative sources of information is essential.Scientists can play an essential role in this process,for the same sort of activities needed to improvebasic scientific understanding of global climatechange will also contribute to any potential abilityto monitor any future agreement.
Where should sanctioning of violations ofgreenhouse gas restrictions be focused? Typicalproposals require that exports from states whoseaggregate carbon use exceeds the allotted permitsbe subject to tariffs or other trade restrictions. Butthis would seem to be the wrong target, since it isthe behavior of corporations that must be changed ifthese emissions are to be reduced. Governmentsmight try to identify the guilty parties throughregular police procedures, but this assumes a levelof consistency between international treaties anddomestic law that may be rather difficult tomaintain in all circumstances. Thus, some non-governmental monitoring of individual violatorswould seem to be required as a supplement andcheck on official governmental activities.
In addition to assigning responsibility forviolations of existing agreements, other types oflegal claims may also become part of the ongoingapplication of a global climate regime. It isgenerally assumed that the detrimental conse-quences of climate change will be intense, at leastfor some regions and some time periods. To whatextent can these "victims" of global climate changebe said to have a valid legal claim for compensationfrom those groups less severely affected by, orperhaps even benefitting from, climatic changes?Settlement of such claims would be greatlyfacilitated by establishment of clear principles oflegal liability, but it is difficult to imagine how thevalidity of such claims could be ascertained.
Implicit normative principles have already beeninvolved in negotiations over global warming, inparticular, the widespread perception that someexceptions should be provided for developingcountries in order to not unduly hamper theirefforts to achieve better standards of living througheconomic growth. Some advocates propose provid-ing a disproportionate amount of marketableemissions permits to developing countries, as ameans of indirect income transfer. But do we reallywant to institutionalize a process by which anygroup disadvantaged by climate changes can press
Political Institutions
claims for compensation in national and interna-tional courts?
Environmental change may sow the seeds fornew issues of conflict over access to resources, but apoorly designed global regime may itself exacerbateconflict by merely providing governments somethingnew to fight over. For example, richer oil-exportingstates (especially Saudi Arabia) allowed other OPECmembers to produce disproportionate levels in orderto provide a more equitable distribution of the gainsthis cartel provided. However, this arrangementitself exacerbated distributional issues, includingcontributing to tensions that led to Iraq's invasionof Kuwait and the subsequent Gulf War. It isimportant to insure that global institutions, onceestablished, do not themselves become catalysts forfuture wars.
The distributional consequences of globalclimate change may well be significant, but they arealso likely to become clear only slowly. Thus,sufficient time may exist to develop some sharedsense of fairness. Even though conflict over climaticchange cannot be prevented, institutional arrange-ments can be established that facilitate theresolution and management of conflict that doesarise.
Any political bargain on the scale of theproposed global regime limiting emissions ofgreenhouse gases should be grounded on some clearfundamental principles, both normative andpractical (see Dorfman 1991). The principles cannotbe imposed but must instead emerge from a processof contestation in which a wide range of feasiblealternatives are given careful consideration. In thenext section, we return to the literature oninternational regimes to propose some alternativeforms of institutional arrangements that aregenerally overlooked because of the tendency ofmost analysts of international relations to focusexclusively on the behavior of national governments.
9 MONITORING AND SANCTIONINGIN INTERNATIONAL REGIMES
Having earlier proposed a classification schemebased on the dominant objectives of organizationsand their scope of activities, we would now like tofocus on two of the dimensions that are closely
related to Design Principles 4 and 5 describedabove. That is, in this section, we examine how thetypes of actors involved in monitoring andsanctioning may affect the types of incentives andlikely outcomes produced. We realize that the typesof actors involved in constitutional decisions affectwho participates in a regime and which type ofauthority and scope rules are used, and we also areconcerned about the types of dispute resolution thatcan be adopted depending on the types of actors.But a preliminary focus on monitoring andsanctioning activities is warranted by theirfundamental importance in any effort to achievemeaningful and sustainable changes in the level ofcooperation over global change issues.
We distinguish categories of regimes accordingto (1) what type of actor is primarily responsible formonitoring compliance with the rules of that regimeand (2) what type of actor is sanctioned for ruleviolation. This latter type of actor is generally thesame as the type of actor whose behavior isrestricted, although this relationship is notnecessarily the only option. For example, nationalgovernments could be sanctioned for the activities ofdomestic corporations that violate rules laid down ina given regime.
Since the extent to which an IGO can beconsidered to be a separate actor remains in doubt,we do not consider the possible existence of regimesin which IGOs are sanctioned. (Efforts by theReagan administration to withhold payment of U.S.dues to some UN organizations can be interpretedas sanctions against some of their previousactivities.) However, IGOs can fulfill importantinformation-gathering functions, includingmonitoring the behavior of governments, corpora-tions, and other organizations (Jacobson 1984).Also, for these purposes, there seems to be littlereason to distinguish between domestic corporationsand MNCs, or between NGOs and INGOs. Thisleaves us with four types of actors engaged inmonitoring activities (IGOs, governments, corpora-tions, private groups) and three actor types subjectto sanctions (governments, corporations, privategroups or individuals).
Not all of the resulting 12 categories arerelevant to the existing set of international regimes.Illustrative examples of relevant categories aregiven in Table 2. These categories are rough, and
Section IV
TABLE 2 An Actor-Based Classification of International Regimes
Entity thatMonitors
Government
IGO
Corporation
Private groupor individual
1
2
3
6
Government
Arms controllaws of war
Nuclearnonproliferation
Trade restrictionsfrequency spectrum
Human rightsenvironmental
Entity that Gets Sanctioned
Corporation
4 LDC-MNC relations
5 Sea-bed mining(Law of the Sea)
9 Cartels
7 Future environmental?
Private Groupor Individual
8 War crimesNHO
many real world cases will fail to fall clearly withina single category, especially since many differenttypes of groups may be engaged in monitoringactivities. Nonetheless, we believe this categoriza-tion has some utility in preparing the way for amore precise analysis of international regimes. InTable 2, numbers are assigned to each of thecategories discussed below.
9.1 Category 1: Government Monitored-Government Sanctioned
Realist theorists of international politics focusalmost exclusively on situations in which govern-ments play both roles, that is, regimes that fallwithin Category 1. In his classic study of inter-national order, Bull (1977, chapter 3) argues that"sovereign states" (which he interprets as individualmembers of "international society") are primarilyresponsible for performing all functions of rulemaking, monitoring, interpreting, legitimizing, andenforcement at the international level. Forexample, compliance with U.S.-Soviet arms controltreaties has been monitored by each government'sintelligence agencies.
The international law of war also fits into thiscategory. Many observers are reluctant to describewar as an orderly activity, yet war has long served
as a prominent institution of "international society"(Bull 1977), and certain restraints on the use offorce are widely (but not universally) observed. Theapplication of regime analysis to war or securityrivalries remains controversial, but we take theposition that any regime "/ill include some mixtureof cooperation and conflict, and thus that someregimes will be rather more conflictual than others.
Even relatively clear examples of the impositionof sanctions on national governments illustrate theextent to which international law falls short of thestandards of domestic legal systems. Since Iraqiintervention in Kuwait was widely seen as aviolation of international agreements againstaggression, it was relatively easy to arrange globalagreement on the imposition of economic sanctionson Iraq. Yet, this example reinforces concerns aboutthe fairness of applying sanctions to collectivitiesrather than to individual actors raised in TheFederalist (see V. Ostrom 1987). After all, theindividual policymakers responsible for Iraqi actionswere able to circumvent the deleterious effects ofeconomic sanctions, which fell hardest on the poorand powerless elements of Iraq's population.Although economic sanctions have been used severaltimes in recent decades, the use of physical force toenforce UN resolutions remains very unusual.Despite President Bush's claims that these eventssignal the birth of a new world order, it is more
€6
Political Institutions
widely believed that the ability of the U.S.government to garner such widespread support formilitary enforcement must attributed to specialcircumstances that are unlikely to be repeated.
9.2 Category 2: IGO Monitored-GovernmentSanctioned
In Category 2, an IGO is established to monitorgovernments' compliance with an internationalagreement. For example, the Nuclear Non-Proliferation Treaty established an internationalagency (the IAEA) intended to monitor possibleviolations. (We discuss some lessons of arms controlregimes below.) As for any other bureaucracy, caremust be taken to somehow make it in the selfinterest of individual bureaucrats to reportaccurately and reliably, and the data reported byIGOs cannot be honestly evaluated without someconsideration of the purposes that willfulmisrepresentation might serve. However, theactions of most IGOs are generally seen asrelatively impartial, perhaps because IGOs have toosmall a resource base to sustain extravagant graftor corruption, at least in comparison to manynational governments.
9.3 Category 3: Corporation Monitored-Government Sanctioned
Category 3 encompasses the standard set ofinternational economic regimes. Although typicallyestablished by negotiations among nationalgovernments, economic corporations are the actorsmost directly involved in monitoring compliance,with rules against imposing trade barriers or unfairtrade practices. Corporations keep careful track ofany advantage gained by their international com-petitors. Once violations are detected, corporationsand economic interest groups typically press theirown governments to seek redress of the violationsthrough negotiations with the other government.For example, accusations of Japanese dumping ofautomobiles on the American market were resolvedby a bilateral agreement on voluntary importrestraints. Other types of policy changes includeimposing penalties on corporations engaged inunfair trade practices (see Jackson 1990). Thus, thepolicies of national governments lie at the center ofeconomic regimes, even though the monitoring is
usually handled by those corporations most directlyaffected by rule violations. This central location ofgovernments opens up several opportunities forrent-seeking and other predatory activities,especially by governments relying on concessionsfrom MNCs as their primary source of revenue.(We discuss this problem below.)
9.4 Category 4: Government Monitored-Corporation Sanctioned
Category 4 includes examples of agreementsbetween a multinational corporation (MNC) and aless developed country (LDC) that serves as "hostcountry," in the sense that some of the economicactivities of that MNC occur within that country'sterritory. LDC governments can use sanctioningmechanisms ranging from increased taxation tooutright expropriation (or "nationalization") of MNCproperty. MNCs have access to national courts, butthese are unlikely to impose real punishment ontheir own governments. MNCs, on the other hand,can frequently threaten moving their businessactivities elsewhere. Indeed, it is often assertedthat, in most circumstances, MNCs have unfairbargaining advantages over their LDC hosts. Amajor unachieved goal of negotiations concerningestablishment of a New Internationa] EconomicOrder was to establish some internationalregulation of the activities of MNCs.
9.5 Category 5: IGO Monitored-CorporationSanctioned
Instances of Category 5 regimes, in which IGOsmonitor compliance of economic corporations tointernational agreements, point the way towardestablishment of a fundamentally new internationallegal framework. A potential example is the Law ofthe Sea Treaty, which calls for establishment of aninternational agency that would regulate miningactivities on the deep sea-bed outside the 200-mileoffshore exclusive economic zones also established inthat treaty. The purpose of this InternationalSeabed Authority is to insure that profits from thesemining activities are distributed to all peoples of theworld. Concern with the precedent set by thisexample of international control of a still-emergingarea of economic activity led the Reagan adminis-tration to refuse to sign the LOS (Law of the Sea)
67
Section IV
treaty. The extent to which similar internationalagencies will be established to redistribute profitsmade in other areas considered to be the "commonheritage of mankind" remains to be seen.
9.6 Category 6: Private Monitored-Government Sanctioned
In Category 6 regimes, NGOs monitor govern-ment activities that may be subject to sanctioningunder emerging regimes, as in die area of theprotection of individual human rights. Govern-ments have long monitored the ways in which othergovernments treat ethnic or religious minorities ofconcern to them. More recently, private groups,such as Amnesty International and the variousgroups established in Eastern Europe after thesigning of the Helsinki Accords, have achievedcredibility as relatively impartial monitors ofhuman rights violations. In this case, application ofsanctions to a government is clearly appropriate,since it is the oppressive activities by the govern-ment itself, or by some of its agents, that is ofconcern. Thus, the link between the identity of therule violator and the target of sanctioning is moredirect than in economic regimes.
Scientists and environmental action groups arelikely to emerge as the most credible monitors offuture international environmental regimes. Inmany proposed schemes, governments would bearthe brunt of punishment for allowing domesticcorporations to violate internationally establishedrestrictions. For example, in the Ozone Convention,national governments agreed to establish guidelinesfor certain emissions for all industrial activitieswithin their borders. Each government is delegatedto incorporate these into its own legal system. Thenit remains up to the governments involved to decidehow to allocate these limits to domesticcorporations.10
9.7 Category 7: Private Monitored-Corporation Sanctioned
Innovative proposals to establish a market inpollution permits (e.g., Victor 1991) might retain amajor governmental influence in setting limits onthese permits, but the underlying logic is funda-mentally different. Since these permits would be
purchased by corporations, the onus for exceedinglimits would fall on individual corporations. Thatis, the overall emissions within a given territorymight exceed the limit that would have been setunder some international convention, but as long asthe corporations polluting that territory hadobtained sufficient permits, there would be r oviolations of that regime. A regime of that typewould fall under a new category of private monitor-ing and sanctioning of corporations, marked byCategory 7 in Table 2.
9.8 Category 8: Government Monitored-Private Sanctioned
In Category 8 regimes, private groups orindividuals are subject to sanctions imposed byother governments. Proposals for establishment ofa New International Information Order (NIIO), inwhich international news media would have beensubject to explicit regulation by governmentalauthorities, would point toward such a regime, sincedisgruntled governments would supposedly havebeen able to punish news organizations forunflattering news reports, but this regime was notestablished. Examples of Category 8 regimes areextremely rare, with perhaps the best examplebeing the war crimes trials after World War II, inwhich the victorious powers established tribunals toestablish individual guilt for crimes consideredunacceptable within the standard tenets of the lawsof war. More recent examples include abduction ofreputed terrorists to stand trial in the country inwhich their victims lived, or the practice of usingU.S. courts to try deposed dictators for corruption(e.g., Marcos and Noriega). Complete realization ofthis type of regime awaits establishment of a newconstitutional order at the global level.
9.9 Other Combinations
The remaining categories are less directlyrelevant to international regimes, at least as theyare normally conceptualized. An example of thetype of issue that would arise in a regime in whichcorporations monitor and sanction each other wouldbe typical cartel arrangements, which are oftentechnically illegal. In other situations, corporationsmay monitor each other but receive sanctions fromsome external source, as in cases where one
68
Political Institutions
corporation sues another for patent infringement.Our understanding is that such disputes canusually be handled under the confines of existing"private international law," which differs from"public international law" in that it deals with theactivities of groups and individuals rather than theactivities of public authorities (usually in terms ofnational governments). Further development of theinternational legal system toward an increasedsimilarity to domestic legal systems might wellprove a key step in transforming our current state-centered system into an entirely new form of worldorder, but for now, the remaining cells in Table 2are left blank.
10 ADDITIONAL CONSIDERATIONS
In order to expand this monitoring andsanctioning analysis to include the remainingdesign principles discussed above, a systematicsurvey of international regimes could be organizedaround the following additional questions:
• How many actors are involved? Two, a few,or many?
• Do actors have common or complementaryinterests? Are the actors relativelysymmetric in resources and perceptions? Ordo gross inequities exist in these areas? Doactors share a common understanding of theissues?
• Was the regime established by a singlehegemonic power? Or by a small group ofleaders in that field of endeavor, after theyhad attained their position of leadership? Orwas it established before oligopolistic position(or its equivalent) could be attained?
• What are the procedures for change inmembership? Do new members qualify byattaining recognition as a sovereign state?Or must they engage in some qualifyingactivities? Must current members vote onadmission of new members?
• How are rules established and altered? Canrules be changed during regular meetings ofrepresentatives of participating parties? Ormust special global conventions be held?
* What arenas exist for the resolution ofdisputes? How are decisions enforced on theparties? How likely is it that disputingparties will resort to military force?
In lieu of a comprehensive survey, discussion ofa few examples may be useful. Bilateral regimesinclude relations between adversaries (U.S.-Sovietarms control) and allies (demilitarization of theU.S.-Canadian border). Antarctica is a classicexample of a regime established by a relativelysymmetric group of leading states that took controlbefore the rest of the world could get involved; onlythose governments actively engaged in Antarcticresearch can quality for participation in Antarcticgovernance. The Law of the Sea involves allcountries of die world, and it is a rare example ofefforts to impose regulations on an emergingtechnology (deep sea-bed mining) before it canestablish its own practices. The Law of the Seaemerged from some 15 years of complex globalnegotiations, but the Ozone Convention requiredconsiderably less time to complete and to revisetowards stricter standards.
In summary, the basic ideas underlying theanalytical framework presented here are that thoseinternational regimes that are most effective arelikely to have been established by interested partieswith common or complementary interests, sharedcognitive understandings, and relativelysymmetrical capabilities. But probably the mostimportant point is that a successful regime mustinspire those actors most directly affected by thatregime to engage in regular monitoring activities.Actors should also have access to a graduated scaleof sanctions, flexible procedures for revising rules,and a well-established arena for resolution of thedisputes that will inevitably arise. In theinternational arena, the diversity of actors greatlycomplicates analysis, and considerable researchmust be completed before the utility of thistheoretical framework can be adequ iely assessed.
11 FURTHER REFLECTION ON THENATURE OF GOVERNMENT ACTORS
Another important consideration is the nature ofthe political order within which any "nationalgovernment" or "state" acts (V. Ostrom 1991a).11
Open political orders in which contestation is the
69
Section IV
norm can have dramatically different consequencesfor possible forms of international cooperation thanpolitical orders more closely resembling theunaccountable sovereign portrayed in Hobbes'sLeviathan. In open political systems, for example,groups of individuals are allowed to form varioustypes of associations that would be deemedillegitimate or unlawful in a closed political system.Thus, it would be difficult to rely on NGOs oreconomic corporations as monitors in anyinternational regimes among Hobbesian sovereigns.Such associations would be encouraged in regimesformed by governments of open political orders.Furthermore, the state-centered monitoringactivities characteristic of a Hobbesian system mayremain insufficient to sustain global cooperation onclimate change, since there is no reason to expectthat sovereign states would always provide accurateinformation on their own activities. (We addressthe limitations of state-centered monitoring in ourdiscussion of arms control regimes below.)
Any international regime that includesgovernments of very different political orders suffersfrom additional complications. This difficulty ismost apparent in regimes covering North-Southrelations. Many governments in the Third Worldare based on nontax sources of revenues (nationalrevenues obtained largely from rents, concessions,and/or donors), hidden taxes, and/or brute force.Government officials, who do not need to rely on theconsent of taxpayers and obtain substantial revenuefrom mineral sales and from selling agriculturalproduce at world market prices (while buying atlower controlled prices), face fewer incentives toinvest resources efficiently to generate real economicgrowth than do governments that are answerable tocitizen-taxpayers. When large quantities of donorfunds are added to the fiscal resources of a centralgovernment, centralized opportunities to seek rentsincrease still further (Bates 1983). Amos Sawyer,reflecting on the emergence of autocracy in Liberia,stresses the perverse effect of overreliance onobtaining revenues from forest concessions:
The impact of these arrangements on thesociety was profound. First, they increasedthe capability of the government in amanner that further strengthened institu-tional capacities at the center. Theproprietary role of the government — thepresident, in other words — was enormously
increased. Reliance on rents, royalties, andprofits gave the presidency an independentexistence, with the capability to operatewithout any of the pretensions of account-ability that would have been required hadthe president been dependent on income orother taxes raised directly from the people(Sawyer 1992, p. 261).
There are two problems at the internationallevel when some of the national participants relyheavily on rents for revenue — particularly whenthese rents are based on forest products. It is notjust that these national governments have a delicatepolitical issue to face in limiting the extraction offorest resources. Such an action might eliminatethe major source of revenue to the governmentitself. Furthermore, if a government has not yetestablished accountability to its own citizens, it ishard to know how to develop trust andaccountability at an international level.
12 THE LIMITATIONS OF COMPUTERAND GAME MODELS
We now discuss some implications of thismultilevel theoretical perspective for thedevelopment of formal models of internationalenvironmental regimes, specifically simple gametheoretic models and large-scale computer models.Much of the concern with global warming is basedon the conclusions of large-scale computer models ofatmospheric circulation patterns and relatedphysical processes. These models provide a firmerscientific basis for similar concerns about thesustainability of current trends popularized in aseries of modeling exercises initiated by theinfluential book Limits to Growth (Meadows et al.1972). (For extensive reviews of computer models ofdynamic feedback relationships among variousdemographic, technological, resource, economic, andpolitical factors, see Hughes 1985; Onuf 1983.)
The political economic micro foundations of bothtypes of global models tend to be overly restrictiveor poorly articulated. Such models typicallypresume the existence of readily manipulable policy"levers" that can immediately change the behaviorof all relevant actors. In order to evaluate theconsequences of alternative policy proposals,modelers must presume that the underlying
70
Political Institutions
patterns of behavior will remain the same, whetherthat behavior occurs by habit or as the aggregateresult of market processes. But we have empha-sized the importance of devising institutionalarrangements that enhance capabilities forartisanship in which individuals creatively designnew responses to emerging problems. In addition,the vast array of private actors affecting globalenvironmental change will adapt to any policychange implemented by an international convention.A considerable tension exists in relying on models ofconstant processes while trying to change theunderlying processes themselves.
One adaptation that can be expected istechnological innovation. Substantial evidenceexists that when relative factor prices have changedsubstantially, technological changes have occurredwith a distinct factor-saving bias (Binswanger 1974,1977; Binswanger and Ruttan 1985). When regula-tions specify technology rather than changing thoseincentives that help to induce technological innova-tion, less positive adaptations are stimulated. Nomodel that fails to consider these responses canadequately represent the implications of anyglobally mandated policy change. Previous critiquesalong these lines have focused on the potentiallybeneficial consequences of technological innovationinspired by increasing resource costs, but it is alsoimportant to consider the implications of inducedinnovation in political and economic institutions(see, for example, North and Thomas 1973; Davisand North 1971; North 1981, 1990, 1991).
Simple 2x2 game models stand in stark contrastto massive computer models of global dynamics. Byeschewing any pretensions to completeness, simplegame models, especially of Prisoner's Dilemma,Chicken, and various coordination games, haveproven to be very useful devices for highlightingsome fundamental dilemmas of collective action.One can draw on many of the recent developmentsin game theory, public choice theory, and micro-economic theory to go beneath the surface of an"issue area" in an attempt to isolate structures ofsituations that occur in many different guises andforms across different issue areas.
A particularly influential body of researchfocuses on lessons drawn from simple game modelsin which self-interested or egoistic actors cooperate
in the absence of any centralized authority.Influential works by Axelrod (1984) and Oye (1986)emphasize that cooperation (in repeated Prisoner'sDilemma games) is facilitated in situationsinvolving a small number of actors with common orcomplementary interests and who perceive a longshadow of the future (or low discount rate).Cooperation can also be facilitated by increasing thetransparency of a regime (meaning the extent towhich individual behavior can be monitored andsanctions focused on rule violators). Other modelersexamine the implications of various factors on therange of equilibria solutions for these simplemodels. For example, a wider range of cooperativebehavior becomes feasible if players are allowed tolink their behavior in different issue areas (E. Haas1980; McGinnis 1986) or have access to some meansof informally correlating their behavior (McGinnisand Williams 1991).
Although the basic insights of these modelsremain relevant to global cooperation onenvironmental change, these lessons should beapplied with care. Nearly all game models supposethat the 2 (or n) players are fundamentallysymmetric in the sense that they play similar roles(see Gardner and Ostrom 1991). Although gameplayers differ in their exact preferences and in theirrelative resource endowments, they typically havethe same or similar set of possible actions. Moregenerally, regimes may include a diverse set ofactors, with very different interests and capabilities.
While the n-person, iterated Prisoner's Dilemmagame has frequently been equated to the commonsproblem (as if there were only one commonsproblem), a much richer array of specific n-persongames are found in situations that have the twoinitial attributes of CPRs. Depending upon thespecific rules-in-use, CPR problems can berepresented as rent dissipation problems, assign-ment problems, or technological externalityproblems. Some of these "typical" structures foundin CPR situations lend themselves both tomathematical modeling (see, for example, Gardnerand Ostrom 1991; Weissing and Ostrom 1991a,b)and to empirical research conducted in anexperimental laboratory (see Ostrom, Gardner, andWalker 1992, forthcoming; Walker, Gardner, andOstrom 1991; Walker and Gardner forthcoming;Ostrom and Walker 1991).
71
Section IV
More elaborate, extensive form game modelshave recently been used to model the variety ofmonitoring activities that can be used to maintainstable common pool resources regimes in situationsin which rule violations cannot be entirelyeliminated (Weissing and Ostrom 1991a,b). Theselatter models demonstrate that it is often effectiveto combine different types of monitoring activities,with individuals playing specialized monitoringroles being supplemented by self monitoring by theindividuals most directly involved. Since there is noway to eliminate all temptation for shirking on thepart of monitoring officials, it is useful for theindividuals involved to also monitor the level ofeffort exerted by officials selected to do themonitoring. Even then, rule violations are nottotally eliminated, but their effects are sufficientlyconstrained to allow the overall regime to remain inoperation. In short, equilibrium solutions do notrequire perfect cooperation.
The monitoring models of Ostrom and Weissingsuggest important lessons for the regime literature,much of which is based on much simpler modelscontaining only a single type of actor. Even whennational governments establish some intergovern-mental organization intended to monitor compliancewith some specific agreement, they also engage in asubstantial amount of self monitoring. (We discussthis issue in the context of arms control regimesbelow.) But in practice, the most effective monitorsof compliance with international trade agreements,for example, are not national governmentsthemselves but rather those economic corporationsseeking to prevent their competitors from gainingan unfair advantage by circumventing domesticlaws or international treaties.
Our purpose in presenting this preliminarytypology of international regimes is to suggest thata much wider range of possibilities exists than hasgenerally been realized in the existing literature oninternational regimes. Although the activities ofcorporations and interest groups receive consider-able attention in case studies, formal models ofinternational cooperation continue to focus on asingle, undifferentiated set of actors. Richlydetailed case studies of specific regimes thatincorporate the actions of a wide range of actorsgenerally lack a coherent, well-articulatedtheoretical framework that makes sense of
systematic differences in the interests andcapabilities of different actor types.
But it is the modeling literature that is mostguilty of ignoring the diversity of actor types. Fartoo sweeping conclusions about the limitedpossibilities of cooperation have been based onmodels including only a single undifferentiated setof actors. In our opinion, to fairly model the varietyof configurations available for use in designinginternational regimes, the goals and behaviorpossibilities characteristic of each of the followingactor types would have to be specified:
* Governments: Open political orders, coercivepolitical orders, and concession economy-based;
* Corporations: Domestic and MNCs;
• NGOs and INGOs; and
• IGOs.
Unfortunately, we cannot point to any examplesof multiple-actor-type models of internationalregimes, but we strongly encourage theirdevelopment. Consider, for example, the problem ofmonitoring compliance with emissions standardsthat might be established as part of a global climateregime. If corporations find that compliance withsome of these controls would lower their profits,they are likely to seek some way of circumventingthem. Yet other aspects of the regulations mighthave relatively little effect on profit margins and yeta more direct impact on environmental degradation.A corporation's competitors would be particularlyconcerned about violations that would unfairlylower costs of production, whereas nongovernmentalenvironmental groups would be particularlyconcerned with violations that adversely affect theenvironment, whether or not they affect prices.
A model of a game with multiple monitors couldbe set up, in which corporate competitors wouldreceive benefits from uncovering the first type ofviolations, and NGOs could receive benefits(through increased exposure to potential contribu-tors or providing satisfaction to their currentmembers) by detecting the second type of violations.An additional step might be added to represent the
72
Political Institutions
ability of conflict resolution mechanisms to correctlydetermine whether or not violations had actuallyoccurred. Officials of the conflict resolution arenas(whether representatives of governments, IGOs, orother institutions) would have an incentive to keepthe proportion of correct judgements sufficientlyhigh to insure a continued perception of the entirescheme as fair and equitable, but they may also besubject to the temptations of bribes to overlookcertain violations. A formal model of this monitor-ing scheme might help clarify the extent to whichmultiple types of monitoring activities can becombined to achieve lower equilibrium rates ofcheating or other properties.
This informal sketch suggests that considerableinsights might be derived from very simple modelsof international regimes in which these differentactor types interact on a regular basis, even if onlya few types are included in any given model. Suchmodels might help specify circumstances underwhich different arrangements most effectivelyenhance the prospects for stable patterns ofcooperation. Empirical analyses can suggest therange of possibilities that have already beenrealized in various circumstances, but formalmodels complement empirically based studies bygiving them a broader and more rigorousfoundation.
Tn summary, we advise that computer models ofthe consequences on global environment of alterna-tive policy changes be used with great caution.Simple game models can play an important role inhelping clarify the underlying dilemmas of collectiveaction, and computer models can help strengthenour intuition about the implications of feedbackprocesses in complex systems. Response to globalchange is a massive, ongoing process of evaluationand re-evaluation. Yet if the modeling effort isattempted too early in this process, one may developa perfectly good model, but a model of a differentsituation than the one of interest. Our focus shouldinstead lie on enhancing the ability of individualsand collective groups to design institutionalarrangements that facilitate creative, locally basedadaptation to changing circumstances.
13 LESSONS FROM ARMS CONTROLREGIMES
One important means to understand the rangeof possible solutions to collective action problems atthe international level is to examine some examplesin some detail. In this section we attempt to drawsome lessons from examples of arms controlregimes. (For an excellent survey of the lessonsthat can be drawn from the Law of the Sea Treatyand the Ozone Convention, see Sebenius 1991.)
Some scholars are concerned that globalenvironmental change will be a new source ofconflict, since unaccustomed deprivations willexacerbate contention over access to variousresources (Homer-Dixon 1991; Jodha 1988).However, the novelty of environmental conflict caneasily be exaggerated, since resource issues havelong been a source of international conflict. Sinceefforts to eliminate international conflict areunrealistic, the basic problem has always beenmanaging conflict and preventing disputes fromescalating into war. Similarly, some global climatechange will occur no matter what. A robust globalregime must find some means of dealing with theconflict that arises from these changes. Thus, abrief examination of earlier efforts to manageconflict would seem to be in order.
It is also the case that security regimes aremuch simpler to examine, since analysis of securityregimes can reasonably focus on the behavior ofnational governments. Although any analysis ofeconomic and environmental regimes must alsoinclude the behavior of other types of actors, manyof the same principles that have proven successfulin dealing with security problems may prove equallyuseful in more complex configurations of actors.
The historical record of arms control effortsprovides several lessons for the types of monitoringactivities and sanctioning options needed toestablish and maintain international cooperation.In evaluating these examples, it is important tokeep in mind that the connection between thoseactors engaged in monitoring and those being
73
Section IV
sanctioned is more direct and clear than in otherareas of international relations. Regimes in otherareas require more complicated relationships amonga wider range of actors. Since national security isone of the primary concerns of all nationalgovernments, these governments are exactly theright actors to monitor activities related to theirown security, but other types of actors more directlyaffected by economic or environmental regimes arelikely to prove more effective monitors in theseissue areas.
Analysis of international cooperation on securityissues has long been plagued by high-blown rhetoricand unrealistic expectations (neither of which is inshort supply in the literature on globalenvironmental change). In classic "idealist"accounts of the interwar period, emphasis istypically placed on the establishment of formalinstitutional structures meant to supersede existingnational governments or new legal systems intendedto restrain governments from undertaking certainactivities. Yet the record of the League of Nations(which supposedly established a global collectivesecurity system), the Kellogg-Briand Pact (whichsupposedly outlawed war as an instrument ofnational policy), or the Washington NavalConference (which set limits on naval forces) isbleak. Most efforts undertaken since World War IIhave had less grandiose goals, and some of theseefforts have proven remarkably successful.
Bilateral arms control treaties between theUnited States and the Soviet Union have been selfmonitored, with both governments relying on whatare euphemistically called "national technicalmeans" (spy satellites and other forms of espionage)to observe the other superpower's force deploymentsand weapons tests. Indeed, meaningful armscontrol treaties became feasible only afterreconnaissance satellites were developed thatenabled each superpower to monitor the otherstate's military activities with a relatively highdegree of confidence (Gaddis 1987). Scientificadvances have played a similarly crucial role inidentifying the problem of global climate change aswell as providing means for monitoring itsdevelopment.
In the SALT treaties, accusations of violationswere raised in a bilateral, ad hoc institution of theStanding Consultative Committee, which served
primarily as a means of conveying concerns to theother side. Each government had other means to letthe other know of its ability to detect violations,although neither would tell the other all that itknew for fear of revealing the limits of their ownintelligence-gathering capability. Throughout theperiod of these treaties, there was considerablecontroversy about the extent of Soviet cheating,which occasionally became an important issue inU.S. domestic politics. In retrospect, it is clear thatthe Soviets repeatedly pushed at the edges of theagreement. It is hard to determine, however,whether this means they were (1) trying to get awaywith whatever they could, (2) probing to find thelimits of U.S. detection capabilities, or (3) simplytaking advantage of the latitude provided byambiguities of the treaties. In any event, thisuncertainty greatly complicated later negotiations,since critics could continue to insist that betterdetection capabilities were needed.
These controversies necessitated more elaborateinspection schemes in INF and START treaties,including relatively permanent inspection stationsto be established outside certain productionfacilities and provisions for a limited number ofsurprise inspections elsewhere in each country.These provisions can be seen as a successfulachievement of terms long advocated by the UnitedStates, even though these same provisions led someU.S. military experts to express concern that theSoviets would gain too much information from theseactivities! It now appears that many of theseremarkable procedures may never actually beimplemented, given the recent dissolution of theSoviet Union and ongoing massive cuts in strategicarsenals.
One disturbing lesson of U.S.-Soviet strategicarms control is how politicized monitoring effortscan be, even when undertaken by supposedlyobjective experts. Even in areas in which technicalconsensus is available, expert advise may have littleimpact on policy. One of the most importantmultilateral arms control agreements is the 1963Partial Test Ban Treaty, which prohibited nuclearweapons tests in outer space, the oceans, or theatmosphere. This treaty was to some extent aresponse to a massive public outcry against world-wide radioactive fallout from atmospheric tests,specifically the detection of radioactive isotopes inmilk products. In retrospect, limitations of the
74
Political Institutions
treaty weve relatively easy to obtain, since any testsin these areas could be easily detected from remotelocations (i.e., from satellites) and sinceunderground tests were allowed to continue.
Progress toward a comprehensive test ban hasbeen virtually nil, despite a widespread consensusamong seismologists that any meaningful levels ofunderground tests could be distinguished fromnaturally occurring earthquakes by a relativelyminor expansion of the worldwide system ofearthquake monitors, many of which could beremotely controlled. Despite this clear realization ofthe technical feasibility of a comprehensive test ban,both sides insisted on being able to continue to testin order to develop new weapons systems (such ascomponents of SDI) as well as to ensure the potencyof existing warheads. This example demonstratesthat developing a consensus s?raong technical"epistemic communities" is not generally sufficientto insure international cooperation, even thoughsuch a consensus may often be a necessaryprecondition for establishing a sustainable regime(see P. Haas 1992). But technical agreement isuseful only up to a point. Eventually some politicalcompromises must be made. Since compromises arelikely to attract continuing controversy, the need forinformation dissemination and political bargainingnever goes away.
Some types of global regimes for control of CO2
emissions would involve extensive inspectionprocedures that corporations are likely to find verytroublesome, given their natural reluctance toreveal industrial secrets to their competitors.Indeed, exactly these concerns have slowed progresson verifying international treaties supposedlycontrolling the spread of chemical and bacterio-logical weapons. In this area, the basic problem isthat the same factories can produce chemicalweapons or commercially valuable products.Intrusive inspection schemes would be required todetermine whether chemical weapons are beingproduced.
The nuclear nonproliferation regime demons-trates the problems entailed in comprehensiveagreements among actors with diverse interests.The basic logic of this agreement was the followingdeal: states capable of building nuclear weaponspledged to help nonweapons states build nuclearpower plants in exchange for their agreement not to
try to obtain nuclear weapons for themselves. Theproblem is that civilian and military uses of nucleartechnology cannot be neatly demarcated, since manyof the same procedures useful for power supply arealso relevant to the development of nuclearweapons.
An international agency (the IAEA) wasestablished to monitor the nonproliferation regime,and it was staffed by technical experts in the field.However, the inadequacy of its monitoring effortswas dramatized by inspections of Iraqi facilitiesafter the Gulf War, which revealed a very extensiveprogram in nuclear weapons development (see Pilat1992). This example suggests a deeper problemwith the IAEA. Many of its clues concerning whereto look for evidence of nuclear weapons facilitiescame from U.S. intelligence sources, and thisreliance on the CIA exposed the IAEA to charges ofbeing an American "lackey." On the other hand, isnot clear how independent sources of intelligencecould be arranged for use by an internationalagency. Nor is it obvious that an internationalintelligence service should be welcomed.
In environmental regimes, some way must befound to monitor compliance without forcingcorporations to forfeit trade secrets to theircompetitors. This suggests the importance ofremotely positioned sensors, especially on satellites.Limits on emissions of greenhouse gases may provemuch easier to monitor than weapons production,since it has been suggested that so-called "carbontaxes" could be levied on energy inputs. If this werefeasible, direct monitoring of the output of individ-ual factories might not be required. However, thisprocedure would require that monitors have accessto a large amount of information concerningeconomic transactions, which might easily befalsified. Also, under many proposed schemes,sanctions would be applied to all exports of anoffending country, even if only a few corporationswere responsible for violations of emissions limits.In the long run, more intrusive technical monitoringstations would need to be established in order toattribute responsibility to individual corporations.After all, it is the behavior of individuals andcorporations that must be changed if the problemsof global climate change are to be addressed —simply changing governmental policies solvesnothing unless these policies successfully elicit thedesired response by other actors.
75
Section IV
To return to the simpler confines of the U.S.-Soviet case, strategic arms control negotiations werecomplicated by their differing (and changing) viewstoward the goal they should be pursuing. In theearly stages of the SALT talks, the United States(under the Nixon administration) emphasized thedesirability of maintaining a situation of mutualassured destruction (MAD) that the Sovietscountered was less ethical than relying on defense.Later, during the Reagan's administration's pursuitof a space-based strategic defense system, theUnited States emphasized the ethical merits ofdefence over deterrence, and Soviet leaders andReagan's domestic critics sang the praises of theMAD doctrine. Also, throughout these negotiations,the United States stressed its concern about thedestabilizing nature of the land-based multiple-warhead intercontinental ballistic missiles (ICBMs)that play such a dominant role in the Soviet forcestructure. U.S. concern with Soviet ICBMs led toinsistence on complicated limitations on variouscategories of weapons systems in hopes of indirectlyshaping the Soviet force structure in ways theSoviets were unwilling to accept. In resisting theseefforts, the Soviets forced U.S. negotiators to acceptoccasionally ambiguous terms. Later innovations inSoviet weapons (which they may well have beenplanning at the time the treaties were signed) couldbe interpreted as violating these ambiguous terms.This perception of Soviet cheating in turn increasedpressures for obtaining even more specificrestrictions on Soviet forces, which the Sovietsresisted even more forcefully. In short, a lack ofcommon understanding of the nature of the problemcontinued to complicate strategic arms negotiations.
On the other hand, leaders of both superpowersshared a clear and common understanding of thedangers of losing control of their own forces duringa crisis. Perhaps the most successful area of armscontrol does not concern weapons per se, but ratherthe establishment of direct communication linksbetween U.S. and Soviet leaders that were painfullylacking during the most serious confrontation of theCold War era, namely, the 1962 Cuban MissileCrisis. Establishment of a teletype "Hot Line" andits subsequent upgrades in technology, along withother confidence-building measures (Hughes 1991,p. 286), are included under the rubric of "quiet armscontrol" (Carnesale and Haass 1987). Thesecommunication links definitely served their purpose.The record of quiet arms control suggests important
lessons for advocates of comprehensive conventionson greenhouse warming and other aspects of globalclimatic change. Crucially important small stepstoward a better future were most easily taken whenleaders eschewed trying to solve all the world'sproblems at once.
In this brief examination of arms controlregimes, we have drawn a series of lessons forenvironmental issues. The most fundamental lessonis that, despite the highly technical nature of bothnuclear and environmental issues, the heart of thematter remains contention among collective actionorganizations (whether they be states, corporations,or private interest groups). Internationalcooperation is an inherently political matter thatcannot be reduced to simple technical formulas.
14 CONCLUSIONS
In this paper, we have drawn lessons from awide range of research traditions. This survey ofinstitutional research on local commons and oninternational security regimes should lead one to becautious about insisting on global steps towardglobal solutions. The most important institutionalchanges for coping more effectively with the conflictlikely to grow out of global change issues may bethe development of open information andcommunication regimes to enhance the capabilitiesof actors at many levels to come to a better commonunderstanding of both the physical as well as theeconomic and social aspects of global climatepatterns. Global steps may well be accomplishedmost effectively if many small steps are taken first,providing these steps can be sustained andcomplement wider efforts.
We conclude with a brief recapitulation of themajor conclusions of our analysis.
• Analyses of responses to common-poolresource dilemmas have demonstrated that awide variety of monitoring and sanctioningschemes can be established and maintainedby the individuals and groups directlyinvolved in collective action dilemmas.Continued reliance on choosing between thestandard policy options of market-basedprivatization or state-based management isfar too constraining.
76
Political Institutions
• The current emphasis on global solutionsbased on international conventions onmanaging environmental change reflects anincomplete understanding of the inherentlymulti-level nature of this problem. A globalregime is more likely to be robust withrespect to significant changes if it impliesinstitutional arrangements at the local levelthat satisfy the design principles identified inanalyses of common-pool resource regimes.
• The existing literature on internationalregimes exhibits excessive concentration onthe behavior of national governments andintergovernmental organizations. Much moreattention should be paid to the incentivesand activities of economic and other types ofcollective actors. Policy proposals emergingfrom these analyses should consider ways inwhich these diverse interests can becombined to establish robust regimes.
• Emphasis should be placed on understandingthe ways in which actors and outcomes atone level of interaction (local, national,regional, global) affect other levels. Institu-tional arrangements at different levelsshould be "nested" in order to insure thattheir effects reinforce each other rather thanacting at cross purposes.
• Scientific groups have important contribu-tions to make in furthering our collectiveunderstanding of global processes and inmonitoring compliance with any futureagreement. However, if comprehensivemodels of physical processes are to be used toevaluate the implications of alternative policyregimes, some way must be found to incor-porate the implications of the technologicaland institutional innovations that are likelyto be induced by global climate change.
• Examples from "quiet arms control" andregional pollution control suggest that earlyconcentration on smaller and moremanageable portions of inordinately complexproblems can establish a sound foundationfor future cooperation.
* Changes in the global environment will havedifferential effects. New forms of conflict canbe expected to arise. However, we areunlikely to be able to predict the forms suchconflict will take. Emphasis should be placedon developing effective means of managingconflict whenever it occurs.
• Contestation should be encouraged at alllevels. Information should be solicited fromindividuals and collective groups with vary-ing interests and perspectives. Arenas forconflict resolution at the local, national,regional, and global levels should beestablished to help contesting groups arriveat common solutions.
15 BIBLIOGRAPHY
Aggarwal, Vinod K, 1985, Liberal Protectionism:The International Politics of Organized TextileTrade, University of California Press, Berkeley,Calif.
Arnold, J.E.M., and J.G. Campbell, 1986, CollectiveManagement of Hill Forests in Nepal: TheCommunity Forestry Development Project, pp. 425-454, in Proceedings of the Conference on CommonProperty Resource Management, sponsored byNational Research Council, published by NationalAcademy Press, Washington, D.C.
Ashby, W. Ross, 1960, Design for a Brain: TheOrigin of Adaptive Behavior, 2nd edition, JohnWiley & Sons, Inc., New York, N.Y.
Axelrod, Robert, 1984, The Evolution of Cooperation,Basic Books, New York, N.Y.
Bates, Robert, 1983, Essays on the PoliticalEconomy of Rural Africa, University of CaliforniaPress, Berkeley, Calif.
Beardsley, Tim, 1992, Night Heat, ScientificAmerican, 266:21-24, Feb.
Benedick, Richard E., 1991, Ozone Diplomacy,Harvard University Press, Cambridge, Mass.
77
Section IV
Berkes, Fikret, 1989, Common Property Resources:Ecology and Community-Based SustainableDevelopment, Belhaven Press, London, England.
Binswanger, Hans P., 1974, The Measurement ofTechnical Change Biases with Many Factors ofProduction, American Economic Review, 64:964-976,Dec.
Binswanger, Hans P., 1977, Measuring the Impactof Economic Factors on the Direction of TechnicalChange, pp. 526-550, in Resource Allocation andProductivity in National and InternationalResearch, Thomas M. Aradt, Dana G. Dalyrumple,and Vernon W. Ruttan (editors), University ofMinnesota Press, Minneapolis, Minn.
Binswanger, Hans P., and Vernon W. Ruttan(editors), 1985, Induced Innovation: Technology,Institutions and Development, John HopkinsUniversity Press, Baltimore, Md.
Blake, David H., and Robert S. Walters, 1987, ThePolitics of Global Economic Relations, 3rd edition,Prentice-Hall, Inc., Englewood Cliffs, N.J.
Blomquist, William, 1992, Dividing the Waters:Governing Groundwater in Southern California, ICSPress, San Francisco, Calif., forthcoming.
Bromley, Daniel W., 1991, Sustaining the GlobalCommons: Environmental Policies throughInternational Cooperation, paper prepared byDepartment of Agricultural Economics, Madison,Wise, for U.S. Agency for InternationalDevelopment, Bureau for Science and Technology,Office of Rural Development, Washington, D.C.
Bull, Hedley, 1977, The Anarchical Society,Columbia University Press, New York, N.Y.
Carnesale, Albert, and Richard N. Haass (editors),1987, Superpower Arms Control, Ballinger,Cambridge, Mass.
Charlson, R.J., et al., 1992, Climate Forcing byAnthropogenic Aerosols, Science, 255:423-430,Jan. 24.
Clark, Colin W., 1980, Restricted Access to Common-Property Fishery Resources: A Game-TheoreticAnalysis, pp. 117-132, in Dynamic Optimization andMathematical Economics, P.T. Liu (editor), PlenumPress, New York, N.Y.
Cleveland, Harlan, 1990/91, The Management ofPeace, The GAO Journal, pp. 4-23, winter.
Commission on Geosciences, Environment, andResources, 1991, Rethinking the Ozone Problem,National Research Council, Washington, D.C.
Copes, Parzival, 1986, A Critical Review of theIndividual Quota as a Device in FisheriesManagement, Land Economics, 62(3):278-291.
Coward, E. Walter, Jr., 1979, Principles of SocialOrganization in an Indigenous Irrigation System,Human Organization, 3S(l):28-36.
Cruz, Maria Concepcion, 1991, Population Pressure,Deforestation, and Common Property Institutions:An Overview, presented at the Second AnnualMeeting of the International Association for theStudy of Common Property (LVSCP), held inWinnipeg, Canada, on Sept. 26-29.
Dahlman, Carl J., 1980, The Open Field System andBeyond: A Property Rights Analysis of an EconomicInstitution, Cambridge University Press,Cambridge, United Kingdom.
Davis, Lance E., and Douglass North, 1971,Institutional Change and American EconomicGrowth, Cambridge University Press, Cambridge,United Kingdom.
de Alessi, Louis, 1988, How Markets AlleviateScarcity, in Rethinking Institutional Analysis andDevelopment: Some Issues, Alternatives, andChoices, Vincent Ostrom, David Feeny, andHartmut Picht (editors), ICS Press, San Francisco,Calif.
Dorfman, Robert, 1991, Protecting the GlobalEnvironment: An Immodest Proposal, WorldDevelopment, /9(l):103-110.
78
Political Institutions
Feeny, David, 1986, The Exploration of EconomicChange: The Contribution of Economic History toDevelopment Economics, in The Future of EconomicHistory, Alexander J. Field (editor), Kluwer-NyhoffPublishing, Dordrecht, the Netherlands.
Feeny, David, 1988a, Agricultural Expansion andForest Depletion in Thailand, 1900-1975, pp. 112-143, in World Deforestation in the TwentiethCentury, J.F. Richards and R.P. Tucker (editors),Duke University Press, Durham, N.C.
Feeny, David, 1988b, The Development of PropertyRights in Land; A Comparative Study, pp. 272-299,in Toward a Political Economy of Development: ARational Choice Perspective, Robert H. Bates(editor), University of California Press, Berkeley,Calif.
Feeny, David, Fikret Berkes, Bonnie J. McCay, andJames M. Acheson, 1990, The Tragedy of theCommons: Twenty-Two Years Later, HumanEcology, ;S(1):1-19.
Gaddis, John Lewis, 1987, Learning to Live withTransparency: The Evolution of a ReconnaissanceSatellite Regime, pp. 195-214, in The Long Peace,Oxford University Press, New York, N.Y.
Gadgil, M., and P. Iyer, 1989, On the Diversificationof Common-Property Resource Use by Indian Society,pp. 240-272, in Common Property Resources, FikretBerkes (editor), Belhaven Press, London, England.
Gardner, Roy, and Elinor Ostrom, 1991, Rules andGames, Public Choice, 54:171-185.
Gardner, Roy, Elinor Ostrom, and James Walker,1991, The Nature of Common-Pool ResourceProblems, Rationality and Society, 2(3):335-358,July.
Gleick, Peter, 1991, Water and Conflict, presentedat the Conference on Environmental Change andAcute Conflict, sponsored by the American Academyof Arts and Sciences and the Peace and ConflictStudies Program of the University of Toronto, Sept.
Gordon, H. Scott, 1954, The Economic Theory of aCommon Property Resource: The Fishery, Journal ofPolitical Economy, 62:124-142, April.
Grubb, M., 1989, The Greenhouse Effect: Negoti-ating Targets, Royal Institute of InternationalAffairs, London, England.
Haas, Ernst B., 1980, Why Collaborate? Issue-Linkage and International Regimes, World Politics,32:357-405.
Haas, Ernst B., 1990, When Knowledge Is Power:Three Models of Change in International Organiza-tions, University of California Press, Berkeley, Calif.
Haas, Peter M., 1989, Do Regimes Matter?Epistemic Communities and Mediterranean Pollu-tion Control, International Organization, 43:378-403.
Haas, Peter M., 1990, Saving the Mediterranean:The Politics of International Environmental Coopera-tion, Columbia University Press, New York, N.Y.
Haas, Peter M. (editor), 1992, Knowledge, Power,and International Policy Coordination, InternationalOrganization, 46:1, special issue, winter.
Haggard, Stephan, and Beth A. Simmons, 1987,Theories of International Regimes, InternationalOrganization, 42:491-517.
Hardin, Garrett, 1968, The Tragedy of theCommons, Science, 762(l):243-248.
Hardin, Garrett, 1978, Political Requirements forPreserving Our Common Heritage, pp. 310-317, inWildlife and America, H.P. Bokaw (editor), Councilon Environmental Quality, Washington, D.C.
Hayami, Yujiro, and Vernon W. Ruttan, 1985,Agricultural Development: An InternationalPerspective, revised edition, Johns HopkinsUniversity Press, Baltimore, Md.
Homer-Dixon, Thomas F., 1991, On the Threshold:Environmental Changes as Causes of Acute Conflict,International Security, 76(2):76-116.
Hughes, Barry, 1985, World Models: The Bases ofDifference, International Studies Quarterly,25:77-102.
79
Section IV
Hughes, Bariy, 1991, Continuity and Change inWorld Politics, Prentice-Hall, Inc., Englewood Cliffs,N.J.
Hurwicz, Leonid, 1973, The Design of Mechanismsfor Resource Allocation, American Economic Review,63(2):l-30.
Intergovernmental Panel on Climate Change, 1990,Global Change, Report 12, in The InternationalGeosphere-Biosphere Programme: A Study ofGlobal Change, Stockholm, Sweden.
Jackson, John H., 1990, Restructuring the GATTSystem, The Royal Institute of International Affairs,London, published by Council of Foreign RelationsPress, New York, N.Y.
Jacobson, Harold K, 1984, Networks of Inter-dependence, 2nd edition, Alfred A. Knopf, Inc., NewYork, N.Y.
Jodha, N.S., 1988, Potential Strategies for Adaptingto Greenhouse Warming: Perspectives from theDeveloping World, pp. 147-158, in GreenhouseWarming: Abatement and Adaptation,Norman Rosenberg et al. (editors), Resources for theFuture, Washington, D.C.
Keohane, Robert 0., 1980, The Theory of HegemonicStability and Changes in International EconomicRegimes, 1967-1977, pp. 131-162, in Change in theInternational System, Ole Holsti, RandolphSiverson, and Alexander George (editors), WestviewPress, Boulder.Colo.
Keohane, Robert O., 1984, After Hegemony,Princeton University Press, Princeton, N.J.
Keohane, Robert 0., and Joseph S. Nye, Jr., 1989,Power and Interdependence: World Politics inTransition, 2nd edition, Little, Brown & Co., Boston,Mass.
Riser, Larry, and Elinor Ostrom, 1982, The ThreeWorlds of Action. A Metatkeoreticul Synthesis ofInstitutional Approaches, pp. 179-222, in Strategiesof Political Inquiry, Elinor Ostrom (editor), SagePublications, Beverly Hills, Calif.
Krasner, Stephen D., 1982, Structural Causes andRegime Consequences: Regimes as InterveningVariables, International Organization, 36:185-205.
Krasner, Stephen D., 1985, Structural Conflict: TheThird World against Global Liberalism, Universityof California Press, Berkeley, Calif.
Latin American and Caribbean Commission onDevelopment and Environment, 1990, Our OwnAgenda, Inter-American Development Bank,Washington, D.C.
Levi, Margaret, 1988, Of Rule and Revenue,University of California Press, Berkeley, Calif.
Maass, Arthur, and R.L. Anderson, 1986, ...and theDesert Shall Rejoice. Conflict, Growth, and Justicein Arid Environments, Krieger, Malabar, Fla.
MacNeill, Jim, Pieter Winsemius, and TaizoYakushiji, 1991, Beyond Interdependence. TheMeshing of the World's Economy and the Earth'sEcology, Trilateral Commission book, OxfordUniversity Press, Oxford, England.
Marine, Alan S., and Richard G. Richels, 1991,International Trade in Carbon Emission Rights: ADecomposition Procedure, American EconomicReview, 8J(2):135-139, May.
Manne, Alan S., and Richard G. Richels,forthcoming, Global C02 Emission Reductions —The Impacts of Rising Energy Costs, The EnergyJournal.
March, James G., and Johan P. Olsen, 1989,Rediscovering Institutions: The OrganizationalBasis of Politics, Free Press, New York, N.Y.
Martin, Edward G., 1986, Resource Mobilization,Water Allocation, and Farmer Organization in HillIrrigation Systems in Nepal, Ph.D. dissertation,Cornell University, Ithaca, N.Y.
Martin, Edward G., and Robert Yoder, 1983, TheChherlung Thulo Kulo: A Case Study of a Farmer-Managed Irrigation System, pp. 203-217, inAppendix I of Water Management in Nepal:Proceedings of the Seminar on Water ManagementIssues, Ministry of Agriculture, Agricultural ProjectsServices Center, Kathmandu, Nepal.
80
Political Institutions
Matthews, R., and J. Phyne, 1988, Regulating theNewfoundland Inshore Fishery: Traditional Valuesversus State Control in the Regulation of a CommonProperty Resource, Journal of Canadian Studies,23:158-176.
McCay, Bonnie J., and James M. Acheson, 1987,The Question of the Commons: The Culture andEcology of Communal Resources, University ofArizona Press, Tucson, Ariz.
McGinnis, Michael D., 1986, Issue Linkage and theEvolution of International Cooperation, Journal ofConflict Resolution, 30:141-170.
McGinnis, Michael D., and John T. Williams, 1991,Configurations of Cooperation: CorrelatedEquilibria in Coordination and Iterated Prisoner'sDilemma Games, presented at the 25th NorthAmerican Meeting of the Peace Science Society(International), held at the University of Michigan,Ann Arbor, Michigan, on Nov. 15-17.
Meadows, Donnella H., Dennis L. Meadows, JorgenRanders, and William W. Behrens III, 1972, TheLimits to Growth, Universe Books, New York, N.Y.
Messerschmidt, Donald A., 1986, People andResources in Nepal: Customary ResourceManagement Systems of the Upper Kali Gandaki,in Proceedings of the Conference on CommonProperty Resource Management, pp. 455-480,sponsored by National Research Council, publishedby National Academy Press, Washington, D.C.
Morgenstern, Richard D., 1991, Towards aComprehensive Approach to Global Climate ChangeMigration, American Economic Review, S/(2):140-145, May.
National Research Council, 1986, Proceedings of theConference on Common Property ResourceManagement, National Academy Press, Washington,D.C.
Neitschmann, Bernard, 1991, EnvironmentalConflicts and Indigenous Nations in CentralAmerica, presented at the Conference onEnvironmental Change and Acute Conflict,sponsored by the American Academy of Arts andSciences and the Peace and Conflict StudiesProgram of the University of Toronto, Sept.
Nincic, Miroslav, 1986, Can the U.S. Trust theU.S.S.R. ?, Scientific American, 254(4):33-41.
Nordhaus, William D., 1991, A Sketch of theEconomics of the Greenhouse Effect, AmericanEconomic Review, 5/:146-150.
North, Douglass C, 1981, Structure and Change inEconomic History, W.W. Norton & Co., Inc., NewYork, N.Y.
North, Douglass C, 1990, Institutions, InstitutionalChange and Economic Performance, CambridgeUniversity Press, New York, N.Y.
North, Douglass C, 1991, Institutional Innovationfor Agricultural Development: Constraints,Problems and Promise, prepared for Consultation onInstitutional Innovations for SustainableAgricultural Development into the 21st Century,held at Rockefeller Foundation Conference Center,in Bellagio, Italy, on Oct. 14-18.
North, Douglass C, and Robert Paul Thomas, 1973,The Rise of the Western World: A New EconomicHistory, Cambridge University Press, New York,N.Y.
Oakerson, Ronald J., 1986, A Model for theAnalysis of Common Property Problems, inProceedings of the Conference on Common PropertyResource Management, pp. 13-30, sponsored byNational Research Council, published by NationalAcademy Press, Washington, D.C.
Okidi, Charles, 1991, Environmental Stress andConflicts in Africa: Case Studies of AfricanInternational Drainage Basins, presented at theConference on Environmental Change and AcuteConflict, sponsored by the American Academy ofArts and Sciences and the Peace and ConflictStudies Program of the University of Toronto, Sept.
Olson, Mancur, 1965, The Logic of Collective Action,Harvard University Press, Cambridge, Mass.
Onuf, Nicholas, 1983, Reports to the Club of Rome,World Politics 36:121-146.
Ostrom, Elinor, 1986, An Agenda for the Study ofInstitutions, Public Choice, 48:3-25.
81
Section IV
Ostrom, Elinor, 1990, Governing the Commons: TheEvolution of Institutions for Collective Action,Cambridge University Press, New York, N.Y.
Ostrom, Elinor, 1992, Crafting Irrigations forIrrigation Systems, ICS Press, San Francisco, Calif.
Ostrom, Elinor, and Vincent Ostrom, 1977, PublicGoods and Public Choices, pp. 7-49, in Alternativesfor Delivering Public Services, E.S. Savas (editor),Westview Press, Boulder, Colo.
Ostrom, Elinor, and James Walker, 1991,Communications in a Commons: Cooperationwithout External Enforcement, pp. 287-322, inLaboratory Research in Political Economy,Thomas R. Palfrey (editor), University of MichiganPress, Ann Arbor, Mich.
Ostrom, Elinor, Roy Gardner, and James Walker,forthcoming, Rules, Games and Common-PoolResources, University of Michigan Press, Ann Arbor,Mich.
Ostrom, Elinor, Roy Gardner, and James Walker,1992, Covenants with and without a Sword: Self-Governance Is Possible, American Political ScienceReview, June.
Ostrom, Vincent, 1987, The Political Theory of aCompound Republic: Designing the AmericanExperiment, revised edition, University of NebraskaPress, Lincoln, Neb.
Ostrom, Vincent, 1991a, The Meaning of AmericanFederalism,, ICS Press, San Francisco, Calif.
Ostrom, Vincent, 1991b, Some Ontological andEpistemological Puzzles in Policy Analysis,presented at the Annual Meeting of the AmericanPolitical Science Association, held in Washington,D.C., on Aug. 30.
Ostrom, Vincent, David Feeny, and Hartmut Picht(editors), 1988, Rethinking Institutional Analysisand Development: Issues, Alternatives, and Choices,ICS Press, San Francisco, Calif.
Oye, Kenneth A. (editor), 1986, Cooperation underAnarchy, Princeton University Press, Princeton,N.J.
Palco, Joseph, 1992, Poles Apart, Science Thrives onThin Ice, Science, 255:276-278, Jan. 17.
Pilat, Joseph F., 1992, Iraq and the Future ofNuclear Proliferation: The Roles of Inspections andTreaties, Science, 255:1224-1229, March 6.
Pinkerton, Evelyn (editor), 1989, Co-operativeManagement of Local Fisheries: New Directions forImproved Management and Community Develop-ment, University of British Columbia Press,Vancouver, British Columbia, Canada.
Plott, Charles R., and Robert A. Meyer, 1975, TheTechnology of Public Goods, Externalities, and theExclusion Principle, in Economic Analysis ofEnvironmental Problems, Edwin S. Mills (editor),National Bureau of Economic Research, New York,N.Y.
Polanyi, Michael, 1951, The Logic of Liberty:Reflections and Rejoinders, University of ChicagoPress, Chicago, 111.
Porter, Gareth, and Janet W. Brown, 1991, GlobalEnvironmental Politics, Westview Press, Boulder,Colo.
Rawlins, Stephen L., 1991, Status of OurInstitutional Capacity to Monitor the AgriculturalProduction Consequences of Environmental Changes,prepared for Consultation on Institutional Innova-tions for Sustainable Agricultural Development intothe 21st Century, held at Rockefeller FoundationConference Center, in Bellagio, Italy, on Oct. 14-18.
Roberts, Leslie, 1991, Greenhouse Role in ReefStress Unproven, Science, 253:258-259.
Robinson, David, 1991, Common Pool Resources,Heritage and the Global Commons: Observationsand Management Suggestions, M.A. thesis, SydneyUniversity Law School, Sydney, Australia.
Rosenberg, Norman, William E. Easterling, PierreR. Crosson, and Joel Darmstadter (editors), 1989,Greenhouse Warming: Abatement and Adaptation,Resources for the Future, Washington, D.C.
82
Political Institutions
Ruttan, Vernon W., 1990, Resource andEnvironmental Constraints on Sustainable Growthin Agricultural Production: Report on a Dialog,staff paper P90-33, University of Minnesota,Institute of Agriculture, Forestry and HomeEconomics, St. Paul, Minn.
Samuelson, Paul, 1954, The Pure Theory of PublicExpenditure, Review of Economics and Statistics,36:387-389.
Sand, Peter H., 1990, Lessons Learned in GlobalEnvironmental Governance, World ResourcesInstitute, Washington, D.C.
Sawyer, Amos, 1992, The Emergence of Autocracy inLiberia: Tragedy and Challenge, ICS Press, SanFrancisco, Calif.
Schaefer, Milner, 1957, Some Considerations ofPopulation Dynamics and Economics in Relation tothe Management of the Commercial MarineFisheries, Journal of the Fisheries Research Boardof Canada, 74:669-681.
Schlager, Edella, 1990, Model Specification andPolicy Analysis: The Governance of CoastalFisheries, Ph.D. dissertation, Indiana University,Bloomington, Ind.
Scott, Anthony, 1962, The Economics of RegulationFisheries, in Economic Effects of Fishery Regulation,pp. 22-61, R. Hamlisch (editor), Food andAgriculture Organization, Rome, Italy.
Scott, Anthony, 1979, Development of EconomicTheory of Fisheries Regulation, Journal of theFisheries Board of Canada, 36:725-740.
Scott, Anthony, 1982, Regulation and the Locationof Jurisdictional Powers: The Fishery, Osgoode HallLaw Journal, 20:780-805.
Sebenius, James K, 1984, Negotiating the Law ofthe Sea: Lessons in the Art and Science of ReachingAgreement, Harvard University Press, Cambridge,Mass.
Sebenius, James K., 1991, Designing Negotiationstoward a New Regime: The Case of GlobalWarming, International Security, 75(4):110-148.
Shepsle, Kenneth A., 1979a, InstitutionalArrangements and Equilibrium in Multi-dimensional Voting Models, American Journal ofPolitical Science, 23(l):27-59.
Shepsle, Kenneth A., 1979b, The Role ofInstitutional Structure in the Creation of PolicyEquilibrium, in Public Policy and Public Choice,pp. 249-283, Douglas W. Rae and T.J. Eismeier(editors), Sage Yearbooks in Politics and PublicPolicy 6, Beverly Hills, Calif.
Shepsle, Kenneth A., 1989, Studying Institutions:Some Lessons from the Rational Choice Approach,Journal of Theoretical Politics, /(2):131-149, April.
Siy, R.Y., Jr., 1982, Community ResourceManagement: Lessons from the Zanjera, Universityof the Philippines Press, Quezon City, Philippines.
Southgate, Douglas, and C. Ford Runge, 1990, TheInstitutional Origins of Deforestation in LatinAmerica, staff paper P90-5, University of Minnesota,Institute of Agriculture, Forestry and HomeEconomics, St. Paul, Minn.
Spero, Joan E., 1990, The Politics of InternationalEconomic Relations, 4th edition, St. Martin's Press,Inc., New York, N.Y.
Stone, Richard, 1992, NRC Faults Science behindOzone Regs, Science, 255:26, Jan. 3.
Suhrke, Astri, 1991, Environmental Change,Population Displacement, and Acute Conflict,presented at the Conference on EnvironmentalChange and Acute Conflict, sponsored by theAmerican Academy of Arts and Sciences and thePeace and Conflict Studies Program of theUniversity of Toronto, Sept.
Tang, Shui Yan, 1991, Institutional Arrangementsand the Management of Common-Pool Resources,Public Administration Review, 5J(1):42-51.
Tang, S. Yan, 1992, Institutions and CollectiveAction: Self-Governance in Irrigation, ICS Press,San Francisco, Calif.
83
Section IV
Task Force on International Development andEnvironmental Security, 1991, Partnership forSustainable Development: A New U.S. Agenda forInternational Development and EnvironmentalSecurity, Environmental and Energy StudyInstitute, Washington, D.C.
Thatcher, Peter S., 1990, Alternative Legal andInstitutional Approaches to Global Change,Colorado, Journal of International EnvironmentalLaw and Policy, i.101-126.
Thomson, James T., 1977, Ecological Deterioration:Local-Level Rule Making and Enforcement Problemsin Niger, in Desertification: EnvironmentalDegradation in and around Arid Lands, pp. 57-79,M.H. Glantz (editor), Westview Press, Boulder, Colo.
Townsend, Ralph E., 1986, A Critique of Models ofthe American Lobster Fishery, Journal of Environ-mental Economics and Management, i3:277-291.
Townsend, Ralph, and James Wilson, 1987, AnEconomic View of the "Tragedy of the Commons", inThe Question of the Commons: The Culture andEcology of Communal Resources, pp. 311-326,Bonnie J. McCay and James M. Acheson (editors),University of Arizona Press, Tucson, Ariz.
Victor, David G., 1991, Limits of Market-BasedStrategies for Slowing Global Warming: The CaseofTradeable Permits, Policy Sciences, 24:199-222.
Walker, James, and Roy Gardner, forthcoming, RentDissipation and Probabilistic Destruction ofCommon Pool Resources: Experimental Evidence,Economic Journal, in press.
Walker, James M., Roy Gardner, and ElinorOstrom, 1991, Rent Dissipation and BalancedDeviation Disequilibrium in Common PoolResources: Experimental Evidence, in GameEquilibrium Models II: Methods, Morals, andMarkets, pp. 337-367, Reinhard Selten (editor),Springer-Verlag, Berlin, Germany.
Weissing, Franz, and Elinor Ostrom, 1991a,Irrigation Institutions and the Games IrrigatorsPlay: Rule Enforcement without Guards, in GameEquilibrium Models II: Methods, Morals, andMarkets, Reinhard Seltan (editor), Springer-Verlag,Berlin, Germany.
Weissing, Franz, and Elinor Ostrom, 1991b,Irrigation Institutions and the Gares IrrigatorsPlay: Internal versus External K ue Enforcement,presented at the Workshop on Games in Hierarchiesand Networks, held at the Max-Planck-Institut furGesellschaftsforschung, in Koln, Germany,Sept. 5-7.
Wilson, James A., 1982, The Economical Manage-ment of Multispecies Fisheries, Land Economics,5S(4):417-434.
World Commission of Environment and Develop-ment, 1987, Our Common Future, Oxford UniversityPress, Oxford, England.
World Resources Institute, 1991, Compact for a NewWorld, World Resources Institute, Washington, D.C.
Young, Oran R., 1989, The Politics of InternationalRegime Formation: Managing Natural Resourcesand the Environment, International Organization,45:349-375.
16 END NOTES
1. It is somewhat paradoxical that environmentalregulations also substantially increase the costof monitoring environmental conditions. Acurrent example is the strict measures adoptedby the signatories to the Antarctic Treaty thatmake conducting research more difficult. "Notonly are there environmental impact state-ments to file, but there are also rules ontreatment of indigenous wildlife, prohibitionson the importation of alien species to thecontinent (no more sled dogs, for example), andrules on waste disposal" (Palco 1992, p. 277).
2. See Manne and Richels (1991, forthcoming)and Nordhaus (1991) for estimates of the costsinvolved in meeting proposals to limit totalcarbon emissions to their 1990 level. See alsoGrubb (1989) and Morgenstern (1991).
3. See papers by Gleick (1991), Okidi (1991), andSuhrke (1991) presented at the Conference onEnvironmental Change and Acute Conflict inToronto in the fall of 1991 and also those byJodha (1989), Latin American and CaribbeanCommission on Development and Environment
84
Political Institutions
(1990), Task Force on International Develop-ment and Environmental Security (1991), andWorld Resources Institute (1991).
4. See E. Ostrom (1990); Gardner, Ostrom, andWalker (1991); Gardner and Ostrom (1991);Tang (1992); Blomquist (1992); NationalResearch Council (1986); McCay and Acheson(1987); Berkes (1989); Pinkerton (1989);Ostrom, Feeny, and Picht (1988).
5. See Robinson (1991) for an imaginativeapplication of the results of our micro-CPRenvironments to problems of the globalcommons.
6. Hardin also contributed to the contemporarypresumption that there were only two solutionsto this kind of problem. The "only" alternativesthat he saw to the commons was what hecalled "a private enterprise system," on the onehand, or "socialism" on the other (Hardin 1978,p. 310). He proposed that change would needto be instituted with whatever force wasneeded. In other words, "If ruin is to beavoided in a crowded world, people must beresponsive to a coercive force outside theirindividual psyches, a 'Leviathan' to useHobbes's term" (p. 314). Contrary to manycasual, contemporary descriptions, the Englishcommons was not an open-access CPR (seeDahlman 1980).
7. Monitoring problems plague many efforts tounderstand patterns of global warming. Thehypothesis that global warming was respon-sible for the "bleaching" of coral reefs in theCaribbean cannot be established due to the
lack of adequate data. Researchers arecalling for a major monitoring effort butpoint out that the number of differentvariables required and the number of sitesrequired would lead to a monitoring effortthat would cost millions of dollars a year(see Roberts 1991).
8. We do not think it is possible to elucidatenecessary and sufficient principles for enduringinstitutions, because it takes a fundamentalwillingness of the individuals involved to makeany institution work. No set of logical condi-tions are sufficient to insure that all sets ofindividuals are willing and able to make aninstitution characterized by such conditionswork.
9. These principles are drawn from E. Ostrom(1990) and are explained in more detailtherein.
10. This logic is similar to procedures used inallocation of the electromagnetic frequencyspectrum: The ITU allocates frequency rangesto national governments, which then allocatespecific frequencies to local broadcasters,according to their own rules and procedures.In the United States, this allocation is carriedout by the FCC, and broadcasters closelymonitor each other's behavior.
11. See Neitschmann (1991), which examines thehistorical struggle of the Miskito Indians toretain control over what they consider to betheir own territory and resource base by a"foreign" state, even though it is this "state"that is recognized in the international realm.
Section IV
Comment 1 on Workshop in Political Institutions
Meso-Level Regimes and Robust Plans
Professor Philip A. SchrodtDepartment of Political Science
University of KansasLawrence, Kansas
The paper by McGinnis and Ostrom isimportant for at least three reasons. First, itprovides a number of useful guidelines for theconstruction of political regimes that can controlcommons problems (global warming being oneexample) for generations yet that do not require anexternal enforcing authority. Second, the paperintroduces some important concepts and theoreticaltools from political science, history, and economicsthat can be used to further study such regimes.Finally, and I believe most importantly, the paper isan excellent example of the type of study that needsto be done if the social control of global warming isto be taken seriously. McGinnis and Ostrom remindus that it is not enough to impose rules; these rulesmust be backed by stable self-monitoring and self-correcting institutions.
Since I am in agreement with most of thepoints of the paper, my remarks are largelysuggested extensions rather than criticisms. I focuson three issues: (1) I extend their discussion of localsystems and international systems to the nationaland regional level; (2) I discuss a typology for howsystems might fail, and (3) following John Eddy'sopening comments at the conference, I make someobservations on communication between the naturaland social sciences.
1 LEVELS OF ANALYSIS AND SELFORGANIZATION
The empirical studies in the McGinnis andOstrom paper, which are based on Ostrom's book,Governing the Commons, are generally local. In aterms of the basic levels of analysis shown inTable 1, most of their analysis is either at the
micro level (CPR problems) or the macro level(international regimes).
As McGinnis and Ostrom point out, whendealing with the international systems as a whole,there is a good empirical reason for focusing onrelatively autonomous local systems. Since thesesystems have been stable in the absence of meso-level (i.e., state) intervention, they are applicable tothe quasi-anarchic macro level. Ironically, stableregulatory regimes at the international level mayhave more in common with those at the local levelthan they do with those at intermediate levels, sincethe intermediate levels usually presuppose theexistence of political institutions with monitoringand enforcement capabilities. One should note,however, that the concept of "the state" is verymuch an illusion derived from long experience; infact, national governments are nothing more thanunusually stable and long-lived regimes that solvean assortment of collective goods problems.
This recognition by Ostrom and McGinnis —that the study of CPR regimes must commonlybegin with an understanding that there is nothingspecial about the state — is the political scienceequivalent of Isaac Newton's recognition that themoon was falling. To the common wisdom, it isclear that the moon is not falling; to the common
TABLE 1 Levels of Analysis
Micro Local and individualMeso National and regionalMacro International system
86
Political Institutions
wisdom, it is clear that systems with sovereignenforcement are very different than those that donot have this characteristic (i.e., the internationalsystem or local CPR systems). Only by recognizingthat the common wisdom is wrong can one begin todevelop the theoretical tools to understand thefundamental workings of the system.
McGinnis and Ostrom characterize the robustCPR systems as "self organizing." This character-ization is correct in the sense that the design andimplementation of these systems is not imposed inthe short term by a sovereign; contrast, for example,the design of an aircraft or a computer. However,the utility of the concept of short-term "selforganization" in its pure form is somewhat limited,for the simple reason that humans plan, and eventhe most simple human organizations are theproduct, to some extent, of foresight and design. Inthis respect, the CPR regimes are quite differentthan physical self-organizing systems such as theconvection cell in a thunderstorm or primordialFNA.
Rather than self organization, I find thefollowing two characteristics most important forrobust CPR regimes. First, they are self monitoring:These systems solve the problem of quis custodietipsos custodies (i.e., who guards the guards, or whomonitors the monitors). As McGinnis and Ostrompoint out, solutions other than a sovereign exist forthis problem, and some of them are surprisinglysimple. Second, they are self correcting: They cantransmit a working set of rules across generations;they can correct deviations from those rules; andthey can adapt the rules to fit new circumstances ina fashion that preserves the system.
In the long term, the self-correcting andadaptive character of robust systems does provide aform of self organization. For example, theConstitution of the United States was originallydesigned to handle the collective goods problem of asmall, isolated, agrarian set of newly independentcolonies; 200 years later that same documentgoverned an industrial superpower with territoryspanning a continent and global political commit-ments. While the original document was not selforganized in the short term (i.e., its authors hadvery explicit and carefully argued reasons forconstructing it in the fashion they did), it hasbecome more of a self-organized system as it has
adapted to systems vastly different than those forwhich it was originally conceived.
2 STABLE REGIMES AT THE MESO LEVEL
Since the examples provided by McGinnis andOstrom are either at the micro or macro level, I willsuggest some at the meso level, since much of thework on solving the global warming commonsproblem is likely to involve regimes at either thenational or regional level, even though the problemitself is global. (This is not to imply that thoseauthors are unaware of these examples; I am simplyusing this opportunity to elaborate on them.) Ingeneral, I think many of the characteristics of thesesystems will reinforce the lessons from theMcGinnis and Ostrom studies of local systems. Inchoosing these examples, I looked for cases that hadlifetimes in excess of two centuries and survivedconsiderable turmoil in their external environment.
2.1 National Systems: Federal
The U.S. Constitution (1789 to the present) andthe Swiss confederational system (15th century tothe present) are examples of two federal systemsthat have the characteristics of self correction andself monitoring. Both survived the IndustrialRevolution; in the process, both the United Statesand Switzerland became among the wealthiestnations on earth and survived a variety of changesin their external environment.
The success of these two systems has led manypolitical scientists to see federal systems as theobvious answer to stable meso-level systems. Thepast 50 years have indicated otherwise; the federalscheme has failed or at least appears to be failing incountries such as Pakistan, Nigeria, Canada,Yugoslavia, and Belgium. Federalism alone is notthe answer. Additional research is needed todetermine what rules, implicit or explicit, operate tostabilize federal structures that, on their surface,appear potentially very tenuous.
2.2 National Systems: Hierarchical
The two long-lived hierarchical systems of noteare the Tokagawa Shogunate (1600-1868) and the
87
Section IV
British system of a parliamentary monarchy (1688to the present). (My thanks to Barry Hughes forpointing out this example.) These systems are inmany ways quite different (e.g., the Tokagawasystem was considerably more static than theBritish), but they are both characterized bydecentralized control under the guidance of verygeneral guidelines, a relatively rigid class structure,the constant accountability of the nobility to thecentral system (through the sankin kotai system inJapan* and the Parliament in Britain), and theadvantage of the physical isolation of island nations.Interestingly, both systems provided a "soft landing"into the industrial age, absorbing revolutionariesfrom the hinterlands (industrialists from Kyushuand Scotland respectively) rather than breakingunder revolutionary movements. Both systems alsoevolved highly autonomous bureaucratic servicesthat insulate much of policy from democraticinfluences.
2.3 Regional Systems: The WestphalianState System
The Westphalian nation-state system (1648 tothe present), which evolved in Europe following thecentury and a half of chaos occasioned by the end offeudalism and the Protestant Reformation, can beusefully considered a stable regional system, eventhough it is typically considered a macro-level"international" system. The regional system,composed of a relatively small number of terri-torially contiguous, legally equal, sovereign states,was unprecedented in European history and yetsurvived democratization, industrialization, and themajor wars of the first half of the 20th century.
The Westphalian system underwent a numberof self modifications. The system adapted to at^nsition from aristocratic to democratic rule. Iteasily expanded to encompass non-Europeanindustrialized states such as the United States andJapan in the late 19th century, then went on to beadopted throughout the rest of the planet. Thesystem developed, and in a sense co-opted, a varietyof international organizations. It has changed many
of its own norms. For example, the "state monopolyon violence," now considered fundamental to thesystem, did not operate fully until the 19th centurywith the control of dueling, mercenaries, and state-sanctioned piracy
2.4 Religion
The empirical — as opposed to theological —phenomenon of organized religion provides a finalexample of an extraordinarily stable system at themeso level; in this case, the institutions aretypically transnational rather than strictly regional.
Religious belief provides psychological benefits;moreover, virtually all religious communities alsoprovide an unusually stable environment thatprovides some insurance against random misfortune(illness, bereavement, etc.) through the pooling ofresources. The collective goods problem is solvedbecause there is a very strong norm, which, ifbelieved, provides a long-term rationale againstdefecting from the coalition. In the Judeo-Christian-Islamic tradition, this is usually a veryexplicit promise of a pleasant and eternal afterlife(i.e., by engaging in short-term "irrational" altruisticbehavior, one maximizes long-term benefits).Buddhism and Hinduism provide much the samenorm, albeit through a diametrically oppositemechanism: an escape from eternal life rather thanthe promise of it.
These promises are social constructs, notempirical facts, and they are self perpetuating onlyto the extent that they succeed. When they do, themyth is reinforced in economic terms: Individualstrading within religious groups in which the socialnorms are enforced by the community obviouslyhave much lower information costs than do individ-uals trading in groups in which members must beconstantly alert to the possibility of defection. All ofthe following — Quakers and other dissentingindustrialists in 18th century England and Scot-land; the diamond trade among orthodox Jews inAmsterdam, New York, and Tel Aviv; the agrariancommunities of the Amish in Pennsylvania, Indiana,
This was the system of "alternate attendance" established by Iemitsu Tokugawa in 1635, which requiredmost of the nobility (daimyo) to spend every other year in the shogun's court in Edo.
88
Political Institutions
and Iowa; and the Mormons in 19th century Illinoisand Utah — provide conspicuous examples of theeconomic advantages of self-enforcing communitieswith strict norms, which, in many instances, survivedespite the efforts of sovereign authorities.Virtually any reasonably stable religious groupprovides at least a modest example of the practicalbenefits of membership in a community withstrongly held norms. By doing good, one can dovery well. This prosperity, in turn, appears toprovide empirical evidence of the validity of thegroup's beliefs, which reinforces the believability ofthe unobservable system of sanctions and rewards.
This mechanism provides the potential for anexceedingly stable social unit. This is preciselywhat we observe empirically: Religious institutionsare unquestionably the most long-lived and stable ofall large-scale human institutions, surviving andoften thriving in chaotic times that destroy all otherforms of organization. Empires and ideologies arefragile and passing wisps of fashion compared to thetwo-and-a-half millennia record of Judaism,Confucianism, and Buddhism. The Christianmonastic orders preserved Roman law through thedisorder and social upheavals of the Dark Ages;Islam spread a creed from Morocco to Malaysia inan age when goods and people moved by oar and ox-cart. The assessment of modern revolutionariessuch as Jefferson, Paine, Robespierre, Lenin, andMao — that a civil religion of political ideology (beit democratic liberalism or communism) should becreated to replace the function of deistic religion —was correct with respect to the social function ofreligion and, quite curiously, consistently andcompletely incorrect with respect to the ability ofthose secular formulations to actually provide asubstitute for religion.
These meso-level systems are suggested aspoints of departure for the further study of robustregimes. They seem to share four of the charac-teristics of local systems discussed by McGinnis andOstrom; for example, they are all relativelydecentralized;* all of the systems except religion are
relatively isolated; and the British, Tokagawa, andWestphalian systems all provided outwardly rigidstructures in response to periods of considerablecivil war and instability. Beyond this, however,there is probably ample room for studying thesuccesses and failures of robust meso-level systems.
3 HOW SYSTEMS FAIL
McGinnis and Ostrom discuss a number ofcharacteristics of successful systems. To extendthis, I think it is also useful to consider how theymight fail. On the basis of their studies and otherwork, I would suggest the following typology:
3.1 Design
A system may fail because it never worked inthe first place (i.e., the design was flawed). Forexample, most people would now concede that thecentrally planned economic systems devised in theearly 20th century (most notably the model of theSoviet Union that was either copied or imposed in avariety of places) didn't work for some fairlystraightforward reasons involving information andincentives.
Lest this seem unduly obvious, it should benoted that such systems persisted for the betterpart of a century despite their problems. Thissituation occurred because political systems,particularly those that have substantial coercivepower and are relatively isolated, can be very slowto respond to feedback, when compared with marketsystems. Sovereign states are, by definition,monopolies, and the cost of competitive entry is veryhigh. Once a political coalition is established that issustainable (a criterion quite different fromefficiency), it is likely to be maintained through thelifetimes of those who established it, as occurredwith the Soviet system. This phenomenon is notunique to leftist regimes; the fascist regimes inSpain and Portugal showed a similar behavior.
The day-to-day decision making is decentralized, although the imposition of the rules of the system may becentralized. For example, this distinguishes the social rigidity of Tokagawa Japan from the administrativerigidity of the France of Louis XIV, the Spain of Philip II and Franco, or the Soviet Union of Brezhnev. TheBritish system could survive 60 years of George III; the French could not survive 20 years of Louis XTV.
89
Section IV
Political systems generally have slow feedback, andit is difficult, in the short term, to distinguishbetween those that are truly robust and those thatare merely surviving in a relatively staticenvironment.
3.2 Rent Seeking
Hugh Nixon, in the March 15, 1991, issue ofScience, observed "All successful systems accumu-late parasites." Rent-seeking behavior — a phenom-enon well-studied by development economists butoddly ignored by political scientists — is a secondreason systems may fail. Corruption, to use the lesspolite word, is a failure of the self-monitoringfunction of the system.
The experience of the United States withProhibition provides an excellent example of failuredue to rent seeking. (The prohibition on theconsumption or manufacture of alcoholic beveragesbegan with the ratification of the 18th Amendmentto the Constitution in 1919; the 21st Amendmentreversed this in 1933. The 18th Amendment hasbeen the only change to the U.S. Constitution thathas been subsequently repealed.) Contrary to apopularly held belief in the United States, Prohibi-tion was quite successful; the rate of alcoholism andhealth problems associated with alcohol droppeddramatically during the 1920s compared with rateof the previous half-century. The system wasabandoned, however, with the rise of gangsters suchas Chicago's Al Capone and the corruption of thejudiciary and law enforcement agencies. Prohibitionof alcohol provided too many opportunities for rentseeking by law enforcement officials and under-mined a legal system that had previously functionedwith a relatively low level of corruption. In the end,the public health benefits from Prohibition werejudged less important than the costs of corruption inthe legal system. In solving one problem, rentseeking may create another. The United Statesstarted with an alcoholism problem and ended witha judicial corruption problem.
3.3 Information Cost
A system can fail because the information costsare too high. "Information" here is meant broadlyto include all of the costs of monitoring and
administering a program, including the cost ofallocation penalties rather than simply ofmonitoring. As McGinnis and Ostrom point out,robust CPR systems tend to have low informationcosts.
If the information load of a system is higu, oneof two things can happen. First, the system can bogdown into spending most of its finite resources inhandling information rather than in actually chang-ing the environment. For example, this criticismhas been leveled at the U.S. Superfund program forthe cleanup of toxic waste sites, which spends thebulk of its efforts on litigation rather than cleanup.
Alternatively, if the system attempts to bypassthe information costs of fairly administering asystem, enforcement becomes essentially random,and probabilities of getting away with cheating orbeing unjustly punished become roughly equal.Since we are dealing with a collective goods situa-tions here (i.e., in the absence of punishment, anindividual has more to gain from cheating thanobeying the law), when the likelihood of beingpunished is roughly identical whether one iscooperating or defecting, it becomes rational todefect, and the CPR system fails.
Recognition of this fact may account, in part,for McGinnis's and Ostrom's observation thatsuccessful systems tend to use graduated ratherthan severe sanctions. In a system where mistakesof judgement are possible (i.e., all systems),individuals are likely to demand more information(proof) for severe sanctions than for modestsanctions. The information load increases as theseverity of the punishment increases, andeventually the system becomes unenforceable in aself-monitoring system. This is counterintuitive ifone looks only at the simple benefit-cost calculations(i.e., for a fixed probability of detection, thelikelihood of compliance should increase with theseverity of punishment), but it becomes clear whenone does the appropriate calculations, taking intoaccount the collective characteristics of thesituation.
3.4 Transmission Failure
The final source of system failure occurs whena system is changed in the process of being
90
Political Institutions
transmitted from generation to generation; it canalso occur if the organization fails to adaptappropriately to new circumstances.
In the biological self-organization literature,Eigen and Schuster (1978, p. 555) consider theproblem they term the "error catastrophe":
Genetic reproduction is a continuouslyself-repeating process, and as such differsfrom a simple transfer of a messagethrough a noisy channel [i.e., the classicalproblem of information theory]. For eachsingle transfer it requires more than justrecovery of the meaning of the message,which, given some redundancy, wouldalways allow a fraction of the symbols tobe reproduced incorrectly. It is alsonecessary to prevent any further accu-mulation of mistakes in successivereproduction rounds. In other words, afraction of precisely correct [optimized]copies must be able to compete favorablywith the total of their error copies. Onlyin this way can the [optimal copy] bemaintained in a stable distribution.
[These results] can be summarized asfollows: Any mechanism of selectiveaccumulation of information involves anupper limit for the amount of digits to beassembled in a particular order. If thislimit is surpassed, the [information] willfade away during successive reproduc-tions.
This problem is particularly strong when anorganization attempts to maintain rules andbehaviors that are effective during crisis situationsbut costly and inconvenient in noncrisis situations.Any "mutations" (i.e., modifications of theprocedure) that are easier in the noncrisis environ-ment will be selected for and compete against the"correct" rules that were originally developed to copewith the crisis.
Wohlstetter's classic study of the V.S.intelligence failures prior to the attack on PearlHarbor (Wohlstetter 1962; also see Janis 1982)provides an excellent example of this phenomenonin action. The military knew what it should havebeen doing in a state of heightened tensions with
Japan. Rules more effective than those actuallyemployed in Hawaii were amply available in theformal institutional memory of the military; themeans of improving performance were no mystery.However, two decades of peace had allowed a greatdeal of "genetic drift" to occur in the actualorganizational heuristics. For example, command-ers recognized the necessity of maintaining anti-aircraft gun crews on alert, but they mutated thisrule to provide an exception for Sundays. Theaccumulated errors provided opportunities that theJapanese were able to exploit through carefulplanning.
Ironically, erroneous transmission ofinformation, normally considered a liability, canalso provide a source of inadvertent innovation. Anorganization that is totally efficient and accurate ininformation transmission will give up some advan-tage in innovation. Total transmission efficiency ismost useful when it occurs in a very stable environ-ment; the Confucian bureaucracy of imperial China,which spent an inordinate amount of time in ensur-ing transmission fidelity, comes to mind. Flexibleorganizations with transmission errors, in contrast,should be better able to generate the innovationsrequired in a rapidly changing situation. The U.S.response to Pearl Harbor, which was basically tojunk the existing military structure and build acompletely new one, was completely appropriate inthis regard. This conclusion is also consistent withthe observations about the importance of experi-mentation and failure — as opposed to rationalplanning — found in bureaucratic maxims aboutinnovation, such as "It's easier to apologize than toask permission."
4 CRITERIA FOR DEVELOPING ROBUSTSYSTEMS
McGinnis and Ostrom provide a number ofsuggestions on how robust systems might bedeveloped and, in particular, caution against tryingto do this with overly simple modeling techniques. Iwould add two additional points to their list.
First, efficiency may be less important thansustainability in a robust system. The motto of TomWest, the computer development manager in TracyKidder's novel, Soul of a New Machine, is, "Noteverything worth doing is worth doing well."
SI
Section IV
Biological systems, which evolved rather than weredesigned, tend to be grossly inefficient in manyregards but are all very robust.
A legal system that is only partially successfulmay be more robust than one that attempts to becompletely successful. For example, by mostaccounts, the complex U.S. income tax system failsin the sense that most people under-report income.However, the system is successful in the sense thatit is largely self enforcing and generates substantialamounts of income by using relatively smallamounts of coercion. Consistent with McGinnis andOstrom's observations, that system's use of carefullygraduated sanctions reduces the rent-seekingbehavior that is characteristic of many internalrevenue systems.
A system for the control of global warmingmust be sufficiently thorough to solve, or at leastcontrol, the basic physical problems of globalwarming. Interest-group politics being what theyare, the temptation will probably exist to solveadditional sociai and environmental problems at thesame time, to make the effort at global cooperationmore efficient. Experience would indicate that suchefforts are likely to be counterproductive.
The other point to note is that complex systems(for example, the transportation or telecommunica-tion systems of industrialized countries or anymature legal system, in addition to biologicalorganisms or ecosystems) are not designed, they aregrown; they arise as a response to prior selectivepressures. The more frequently an organization isconfronted with a particular problem, the moreefficient is its response. However, the first time theorganization is so confronted, its response willprobably be inefficient and possibly unpredictable,since the interaction effects of all of the rules of asystem are unknown. One can rarely take theformal organization chart and rule books of ahuman system and deduce from those the outcomeof a specific change. Nor are such systems neces-sarily logically consistent; were that the case, therewould be no need for lawyers. These observationsargue against trying to overdesign a global warmingregime, particularly in view of the uncertaintyconcerning even the physics of the problem.
5 CONCLUSION
As I indicated at the beginning of thesecomments, robust CPR regimes are an importanttopic that are very relevant to the global warmingproblem. The McGinnis and Ostrom paper, andmore generally Ostrom's work in Governing theCommons, provide examples of how to proceed withsystematic studies of these problems.
In his remarks opening this conference, JohnEddy suggested the need for greater cooperationbetween die natural and social sciences. Let meextend this suggestion with two comments, the firsta reminder as to why there has been so little fundedcooperation in recent years, the second to indicatethat communication may be easier than it wouldfirst appear.
Until the onset of the Reagan administration in1981, cooperation between the social and naturalsciences had been encouraged in policy research.Under David Stockman's aegis, however, it wasnoticed that social scientists had an awful tendencyto say things the Reagan ideologues did not want tohear. Economists said that tax cuts would lead to amassive expansion of the federal deficit and thatderegulation of the savings and loan industry wouldlead to corruption and bankruptcy; politicalscientists pointed out that the Soviet Union waseconomically stagnant and riven with ethnic divi-sions; and sociologists pointed out that cutbacks insocial programs might lead to increases in home-lessness, drug abuse, and poverty. Uncomfortablewith the message, the Reagan administration dideverything possible to cut funding for themessengers. While funding was restored in someareas, such as the National Science Foundation, it isstill absent from others. Without being too much ofa disciplinary chauvinist, I would suggest that theabsence of social scientists from the policy processhas been due at least as much to the accuracy,rather than the failures, of their predictions.
The other important point is that the socialsciences and the natural sciences have much morein common now in terms of their methodologiesthan they had 50 years ago. Robert Pool, reportingin the August 18,1989, issue of Science on the
92
Political Institutions
Santa Fe Institute, a collaborative effort ofphysicists and economists to share ideas, observes:
The physical scientists were flabbergastedto discover how mathematically rigoroustheoretical economists are. Physics isgenerally considered to be the mostmathematical of all the sciences, butmodern economics has it beat... . Theflip side of the physicists' surprise at therigor of the economists was the econo-mists' astonishment at the physicists' lackthereof.
Social scientists are now quite conversant inmathematical modeling, statistical methodologies,experimental design, and computer simulation.C.P. Snow's gap between the "two cultures" maystill apply to the humanities and the naturalsciences, but there has been a great deal ofconvergence between the methodologies of the socialand natural sciences. Some of the concepts mostrelevant to the problem of understanding globalwarming (for example, general systems, feedback,and adaptation) have been part of the study ofsocial behavior for decades. General systems theorywas at the forefront of the "behavioral revolution" inpolitical science in the 1950s, through individualssuch as David Easton in American politics,Karl Deutsch in comparative politics, and Morton
Kaplan in international relations; economics wasput into a rigorous mathematical form in the late19th century by conscious emulation of themathematical models of physics.
This being said, the problem of understandingthe social and political implications associated withthe control of global warming is not going to be easyto solve. The social system is at least as complex asthe physical; both have long lag periods and poorlyunderstood feedbacks. In the social as well as thephysical and biological realms, intuition may notalways be a good guide, and purely deductivemodels, divorced from the study of the operation ofactual systems, are unlikely to be very useful. TheMcGinnis and Ostrom paper shows that these socialand political issues can be studied and provides aworthwhile agenda for future research.
6 BIBLIOGRAPHY
Eigen, Manfred, and Peter Schuster, 1978, TheHypercycle, Springer-Verlag, New York, N.Y.
Janis, Irving, 1982, Groupthink, Houghton-Mifflin,Boston, Mass.
Wohlstetter, Roberta, 1962, Pearl Harbor: Warningand Decision, Stanford University Press, Stanford,Calif.
Section IV
Comment 2 on Workshop in Political Institutions
Principal-Agent Relationships
Professor David FeenyDepartment of Economics
McMaster UniversityHamilton, Ontario, Canada
In recent decades, economic analysis has beenextended to situations in which there are importantasymmetries of information between actors. Oneimportant class of these situations are principal-agent relationships, in which one party, the agent,acts on behalf of the other, the principal. In suchsituations of incomplete and asymmetric informa-tion (in which the agent may be better informedthan the principal), is it possible to devisemechanisms to ensure that the agent acts in thebest interests of the principal? How can weconstruct relationships so that physicians act in thebest interests of patients, lawyer in the bestinterests of their clients, or public officials in thebest interests of their constituents?
When complex, multilayered problems such asthose associated with global warming are beingconsidered, the concept of principal-agent relation-ships has two important applications. First theconcept of the ideal principal-agent relationshipscan be used as a normative standard. Under idealcircumstances, leaders (agents) of groups (such asnation states) will negotiate to maximize net socialbenefits on behalf of their constituents (theprincipals). Thus one could examine the set offeasible agreements that could occur if there wasperfect agency (i.e., situation in which agent acts in
the best interests of the principal) among all ofthose who took part in a given round of inter-national (or national or subnational) negotiations.
Second, recognizing that perfect agency doesnot accurately characterize most regimes, one couldexamine the likely set of feasible agreements inactual practice. Even in democratic systems,imperfectly competitive political arenas mean thatnarrow, well-organized special interests are oftendisproportionately represented relative to thegeneral social interest. In less accountable regimes,the scope for the agents to pursue their owninterests at the expense of the principals is oftenquite substantial.
Principal-agent relationships are also relevantin a situation of decomposable, nested hierarchies.Such hierarchies may be a mechanism throughwhich coordination on a global scale may beachieved in a relatively decentralized fashion. Insuch hierarchies, the leader of a small-scale groupwould act as agent on behalf of the members whennegotiating with higher levels in the hierarchy.Such systems of nested hierarchies would appear tobe of great relevance in the context of issues of theglobal commons.
94
Political Institutions
Comment 3 on Workshop in Political Institutions
Problems of Institutional Analysis
Professor Leonid HurwiczDepartment of EconomicsUniversity of MinnesotaMinneapolis, Minnesota
The following is mostly an elaboration of thecomments I made during the discussion stimulatedby the McGinnis-Ostrom paper (McG-0 hereinafter).It will be seen that I have added references to otherpapers by these two authors, kindly made availableto me after the conference. I had been ignorant ofthis literature, and I am grateful both to theauthors and to the organizers of the conference forso broadening my horizons.
MacNeil et al. (1991, p. 220, as cited on p. 1 ofMcG-O) states that the design of appropriateinstitutions is on the Rio "Earth Summit" agenda:"the heads of state . . . will be asked to . . . (4)initiate major reforms of'international institutionsto enable nations to manage global interdependenceand to implement Agenda 21 as well as theconventions.'"
The use of the term "institutions" in the abovequotation illustrates the multiplicity of meaningsgiven to this term and the frequently resultingconfusions.1 The relevant dichotomy in the presentcontext is between (1) using "institution" asreferring to an organization, agency, or othermission-oriented entity on the one fesad and(2) interpreting "institution" as referring (in one ofseveral possible senses) to rules governing behavior.The above quotation seems to use the term in itsfirst (1) meaning. But the focus of the McGinnis-Ostrom analysis concerning problems due to theprospect of global warming, and in particularproblems being faced by the Rio conference, is —and I believe should be — on the issue of existing orneeded international regimes relevant to the ecologyof the globe. The concept of such regimes is broaderthan but includes institutions in the second (2)sense. From now, unless stated to the contrary, thesecond (2) meaning is intended. This is not to
minimize the possible need for (l)-sense institutionswhen it comes to the implementation of rulesinherent in (2)-sense institutions.
McG-O raises a host of important issues, bothsubstantive and methodological, but it also takesstrong positions on what institutional solutions aredesirable and are likely to be successful. I amrather sympathetic with many of these positions,especially insofar as they favor decentralization.However, the extent to which one can hope to avoidexternal (third-party) enforcement and the extent towhich external enforcement can be made effectiveare complex and do not lend themselves, in myopinion, to simple generalizations. To analyze thisproblem, I find it convenient to formulate somewhatmore precisely the meaning I attribute to(2) institutions.
In McG-O, and even more so in the twoauthors' other writings, game-theoretic models areused at times to formalize various institutionalarrangements. But there is, I believe, a distinctionto be made between their formulation and the one Iam about to state. The difference between the twoformulations is, in turn, related to the distinctionbetween what is technically called a game and agame form (or mechanism). What the latter twoconcepts have in common is the specification of theplayers' permissible strategy domains (i.e.,permissible sequences of moves that are, in general,conditioned on the preceding moves of the variousplayers).2 Where they differ is in functionsdefining the consequences of strategic choices made.In a game, associated with each player is his/herpayoff function, specifying the utility level (anumber) that the player would derive from eachpermissible configuration of strategy choices.3 It isthis payoff that the player seeks to maximize. But
95
Section IV
the payoff functions can be viewed as beingcomposed of two elements: the physical4
consequences of the strategic choices made (more orless global warming, degree of destruction of forests,state of health of different groups, etc.) and thepreferences or values that different players attachto these consequences. The physical consequencesof strategic choices are specified by what we call anoutcome function. Here, as for the payoff function,the independent variable (what the mathematiciancalls the domain) is the configuration of strategieschosen. But while for each payoff function, thedependent variable is a (utility) number, thedependent variable for the outcome function is aphysical5 event. Thus a player's value of thepayoff function for a given configuration of strate-gies is given by his/her utility of the outcomeassociated (by the outcome function) with thatconfiguration.6 Thus a game form is defined by thestrategy domains and the outcome function, while agame is defined by the strategy domains and thepayoff functions.7
The rules of a game, as ordinarily conceived,are contained in the game form, because the rulesspecify which strategies are permitted or possibleand the consequences of strategy choices. Notethat if the players' preferences changed while therules of the game remained the same, the payofffunctions would change, but the outcome functionwould not and hence the game form would not.
Given that game rules are specified by thegame form and that institutions (2 sense) are ofteninterpreted as rules, it is tempting to think of aninstitution (again, in 2 sense) as the component ofthe game form that is subject to change by humanaction.9 But when we examine specific examples,we see that typically it is a whole class or categoryof game forms,10 rather than a single game form,that is thought of as a specific institution.11 Onemight say that an institution is an attribute of gamerules, or even that it is a rule about rules."12
Following this digression about the basicconcepts, let us return to the problem of enforce-ment of rules, or — more broadly — what one mightcall effective implementation, such a crucial issue ininstitutional design for problems of global warming.
Superficial reading of certain comments in theliterature of game theory might make one believethat there is no problem of enforcement, since it isoften stated that Nash equilibrium solutions (innoncooperative games) are self enforcing. But thisclaim is only valid when it is assumed that theplayers will not go outside their strategy domains(i.e., will not use "illegal" strategies) and that thepostulated outcome function can be made effective.But why should these assumptions be satisfied?
Consider first the problem of "legal" versus"illegal" strategies. Typically, it is physicallypossible for players to use strategies outside thepermissible domain. Unless players have inter-nalized the prohibitions, the relevant (to be callednatural) game form contains all strategies, legal andillegal; that is the natural domain. As for theconsequences, the corresponding natural game formcontains the actual consequences, whether they aredetermined by the laws of physics, the reactions ofothers, an enforcement system, or other factors. Ifthese various factors act so as to discourage the useof illegal strategies, the original ("desired") gameform will prevail. Otherwise, some or all playerswill use illegal strategies, and the results will bedifferent from what would have happened at a Nashequilibrium of the "desired" game form.
Typically, when new institutions are designed,the game is augmented in still another way:players are added. These additional players may bereal persons ("third parties") or they may be"artificial persons" (organizations, agencies,committees, etc.). For example, when a democraticsystem is introduced, a parliament (an artificialperson) may be brought into existence. On theother hand, an enforcement system may use thirdparties such as police, judges, etc., whose (collective)mission is to discourage the use of illegal strategies.The game form that takes into account the naturaldomains, the natural outcome function, and theextra players is called the augmented game form.
But, assuming that only legal strategies will beused, a mechanism (game form) cannot be madeeffective unless its outcome function operates asprescribed. That may, in particular, involve taxes,subsidies, fines, and hence transfers of goods and
96
Political Institutions
money. A machinery is needed to specify variousprocedural details and also to see that transfers areactually made (e.g., to collect taxes or make socialsecurity payments to those who are entitled). If allof this is successfully done, I speak of effectiveimplementation.13
To effectively implement a goal function meansthat the augmented game form yields equilibriumoutcomes consonant with that function. Sincetypically there is cost in operating the system(whether for information processing or enforcement),the ideal "first best" solution is usually notattainable. But even the "second best" solutions arenot always effectively implementable. Therecommendations contained in the McG-0 papersuggest procedural prescriptions (meta-rules)designed to result in communities adopting rulesthat are enforceable.
McGinnis and Ostrom point to situations inwhich, given a transgression on the part ofindividual A, there is an individual B who has boththe motivation and the ability to note that atransgression has occurred and who can eitherdiscourage (or even stop) the transgression or reportit to someone who can do so. Individual B is what Ihave called an intervenor; a player who has themotivation and ability (power, assets, even powersof persuasion or charisma) to intervene in such amanner as to minimize the probability or eventotally prevent illegal behavior and to make thedesired outcome function operational. But anintervenor (in the sense of this definition) may notbe a person directly affected by the transgression.Consider a policeman stopping a speeder, or aprosecutor determined to root out corruption.Effective implementation may be a result of the factthat there are participants (individuals or groups)who believe in the system as designed, and who,without having direct (i.e., "material") interest,14
are determined to see it work, and who, again, aresomehow endowed with the powers to accomplishthis. They qualify as intervenors. On the otherhand, there may be a chain consisting of individualA (the object of enforcement activities) and theenforcers E1,..., E11"1, and En. Individual En isspecial in that it is in his/her ("material") interestthat A obey the rules. The behavior of person A ismonitored by enforcer E1, who in turn is monitoredby E2, etc. It is further assumed that each E1
among the enforcers E2,..., En has the power to
enforce proper (supervisory or enforcement)behavior on the part of the underling E'"1, whosebehavior E' monitors. This configuration may becalled "closing the circle." Here En is an intervenor,having both the ("material") interest in seeing thatA behaves and the powers (indirectly) to bring thisabout.15
Without studying the nature of the goal to beimplemented and the nature of the facts bearing onprobable ("natural") behavior, it is difficult for oneto generalize as to the need for external (third-party) enforcement. Perhaps the GATT experience,the Montreal Agreement, or disarmamentnegotiations provide clues as to what can beachieved without third-party intervenors on a globalscale.
Similarly, one cannot be too dogmatic aboutgradual versus "wholesale" reforms. When manycomponents are interdependent, the gradualapproach may not be feasible. On the other hand,such radical changes across the board also havetheir dangers and difficulties. The experiences ofCentral and Eastern European countries emergingfrom communism and command economies illustratethis point,
A few additional comments are in order. FirstI will mention a technical point concerning whatMcGinnis16 calls two-level games in which, forinstance, elected representatives of organizations(lower-level bodies) are engaged in bargaining. Thetraditional approach of treating theserepresentatives as being motivated by a utilityfunction17 — when in fact the representatives areunder instruction generated by (for example)majority votes in the lower bodies — is clearlyinappropriate. But there are ways of dealing withsuch situations without postulating a utility-maximizing behavior. For example, in a paper18
dealing with the implementability (in Maskin'ssense) of goal functions,191 replaced Maskin'sassumption that each player2 is maximizing hisor her utility or preference by the assumption thateach player has a choice function. This choicefunction could reflect the majority votes of the lowerbody that the player represents. The importantpoint is that such a choice function need not berationalizable in terms of a transitive, complete,ordering relation. Nevertheless, one still obtainscounterparts of Maskin's implementability
97
Section IV
theorems, both for necessary and sufficientconditions. It is true that when the lower bodies'choice functions are "irrational," the meaning ofimplementability is weakened, but this seemsunavoidable. In any case, models of this type aremore representative of international negotiationsand deserve further study.
The second issue is about design versusspontaneous development or self-organizingbehavior. If the world were satisfied withspontaneous development of institutions dealingwith global ecological issues, we would not (I think)be witnessing Montreal or Rio. History provideswhat to me are clear cases of conscious design, theU.S. Constitution being perhaps a prime example.But looking at the matter from the point of view ofwhat we have called the natural game form, oneencounters a philosophical question. One can arguethat the natural game form cannot be altered by thedesigners. But then, what is the role of design?Perhaps it is only a matter of discovering that thenatural game form contains more advantageoussolutions than those previously arrived at and ofpersuading others that this is so. In that case, therole of design is purely at the level of cognition. Buthaving a new design adopted may involve playing agame based on the natural game form. The two-stage Reiter-Hughes model points in that direction.
It should be noted that the outcome of such agame need not result in a resource allocation that isPareto-superior to the previous one. In other words,as is often the case when decisions are made bymajority vote, there may be losers as well aswinners. It is in this context that the (theoretical)role of benefit-cost analysis can be understood.Subject to a technical point concerning thecustumary underlying assumptions (to be discussedbelow), if the benefit-cost analysis shows that acertain change would yield positive social (i.e.,consumer plus producer) surplus, we conclude thatit is, in principle, possible to find a compensationscheme so that there would be no losers while someor all would be gaining. But, as has often beenpointed out in the literature, it does not follow thatthe proposed is desirable even if such compensationsare not made.
Further, there is the technical problemmentioned above: Customarily, the calculations
underlying benefit-cost analysis are made as ifpreferences were quasi-linear (i.e., characterized byconstant marginal utility of money). In that case,one cait separate the issue of efficiency from that ofdistribution. But there is usually no justificationfor that assumption, and then distributional issuesaffect conclusions as to efficiency. This fact doesnot, in my opinion, invalidate the benefit-costapproach as such, but it does call for estimationprocedures that are more sophisticated. Further-more, there is then no longer a unique, efficient,social-resource allocation but rather a wholespectrum of different efficient arrangementscorresponding to different distributive policies.Together with the problem of compensations, thisissue, far from being an abstruse point of theory, is,I believe, at the heart of the conflicts between themore developed "North" and the less developed"South." Similarly, the debates concerningappropriate discount rates may, to some extent, beinterpreted as distributional conflicts between thepresent and future generations.
Finally, I will mention the suggestion,discussed in some detail in Kosobud's paper,concerning the use of market-type methods (inparticular, tradeable pollution permits) in dealingwith pollution and other environmental hazards.The basic and rather natural idea is to make thepollution activity into a negative commodity (i.e.,commanding a negative price) and then takeadvantage of the known "invisible hand" advantagesof competitive markets. However, it has beenshown by D- Starrett21 that when such negativecommodities are adjoined to the conventional ones,the phenomena of detrimental externalities such aspollution result in nonconvex production possibilitysets. Consequently, it is likely that no fixed pricevector will balance supply and demand. This doesnot mean that a market equilibrium is impossible,but it will not be a "perfectly competitiveequilibrium." For a market process to yield efficientresource allocation — which is the objective — itbecomes necessary to use "nonlinear prices" (i.e.,prices that vary as functions of the volume traded).But to find the appropriate nonlinear formulas forthese prices, some informational centralization maybe unavoidable, and so some of the customaryadvantage of the market process may be lost.
98
Political Institutions
END NOTES
1. See E. Ostrom, 1986, An Agenda for the Studyof Institutions, Public Choice, 48:3-25. As willbe seen, I tend to side with Ostrom's inter-pretation of the term "institution" (in the senserelated to rules and labeled as 2 in whatfollows) rather than the equilibrium behaviorsense given to it by A. Schotter, 1981, TheEconomic Theory of Social Institutions,Cambridge University Press. However,because of my somewhat broader interpretationof what is meant by rules, the contrast is notas strong as might appear. See L. Hurwicz,forthcoming, Towards a Framework forAnalyzing Institutions and InstitutionalChange, in Democracy and Markets: Participa-tion, Accountability and Efficiency, S. Bowles,H. Gintis, and B. Gustafsson (editors),Cambridge University Press, and L. Hurwicz,forthcoming, Implementation and Enforcementin Institutional Modeling, in Political Economy:Institutions, Information, Competition, andRepresentation, W. Barmnett, M. Hinich, andN. Schofield (editors), Cambridge UniversityPress.
2. The concept of a player's strategy domaincovers the three categories in Ostrom (1986):permitted, required, and prohibited. Every-thing outside the strategy domain is prohibited;everything within it is permitted. If a move isrequired, it constitutes an element of everystrategy hi the domain.
3. There are, of course, as many payoff functionsas there are players.
4. The term "physical" should be interpreted quitebroadly, possibly to include even emotionalstates and behaviors. In economic models, theoutcome is usually the resulting resourceallocation. In voting models, it may be thecandidate elected or the issue prevailing in areferendum.
5. Typically multidimensional (see the precedingfootnote).
6. Some notation may help avoid ambiguity. Letthere be n players, and let Sj be the strategychosen by the i-th player. If h represents) theoutcome function, and the given strategyconfiguration (n-tuple) is S = (S l t.... Sn), thenthe outcome produced by that configuration ish(s). Now let the i-th utility function bedenoted by u\ and the i-th payoff function bedenoted by p1. We then have the relationp'(s) = u'(h(s)); that is, the i-th payoff functionis the composition of the i-th utility functionwith the outcome function.
7. Using the notation of the preceding footnote,and denoting the strategy domain (i.e., the setof permissible strategies) of the i-th player byS{> we defined a game form by the n+1 entities(Sp ..., Sn, h), while a game is defined by the2n entities (Sx,..., Sn, p
1 pn).
8. However, strategy domains are limited by lawsof physics, and the outcome function is also, atleast in part, determined by these laws as wellas other factors. That is why we say that gamerules are contained in (but are not coextensivewith) the game form.
9. As distinct from components due to physicallaws, etc. The distinction made here is closeto, if not identical with, that between physicaland "deontic" statements in R. Gardner andE. Ostrom, 1991, Rules and Games, PublicChoice, 70:121-149.
10. More precisely, with a class of thosecomponents of game forms that are influence-able by humans. We may note here that sincethe rules, in addition to governing behavior,are generated by human actions, they may inparticular be viewed as solutions to a priorgame. See S. Reiter and J. Hughes, 1981, APreface on Modeling the Regulated U.S.Economy, Hofstra Law Review, 9(5):1381-1421.This aspect of the situation bears a relation-ship to Schotter's view of institutions asbehavior patterns that are equilibrium solu-tions of games.
99
Section IV
12. Consider, for instance, the institution ofmarriage. We may think of a particular type ofmarriage (say monogamous) as being definedby one game form, while another (say polyga-mous) is defined by another game form. But itis the common features of these two gameforms (and other game forms defining othertypes of marriage) that correspond, in myinterpretation, to the notion of the institutionof marriage as such.
12. Thus, I interpret Ostrom's rules as being such"rules about rules." Typically, we are subjectto many institutions, or what Ostrom (1986)calls a configuration of rules. Therefore, aconfiguration of rules in Ostrom's sense is theintersection of sets of game forms, with eachset defining one of the institutions of theconfiguration. Clearly, as pointed out byOstrom, the class of game forms admissibleunder the configuration is narrower than theindividual institutions would permit.
13. This goes beyond the concept of implementa-tion as defined by Maskin, 1977, NashEquilibrium and Welfare Optimality, M.I.T.mimeo, where enforcement and the properworkings of the outcome Junction are taken forgranted. Hence the adjective "effective" usedhere.
14. There may be an indirect interest in, forexample, maintaining a reputation forimpartiality and incorruptibility, which may beprofessionally valuable. But one should notexclude the possibility of idealism, or ethics, orideology.
15. This is different from the McG-0 irrigationrotation example, where everyone has a directinterest in his/her predecessor observing therules and where everyone is an intervenor withboth motivation and power.
16. From a paper done jointly with John T.Williams, which the authors were kind enoughto make available to me but which is not yet infinal form available for quotation.
17. Or, more generally, by a preference relation,whether or not it can be represented by autility function. An example of a preferencerelation that cannot be so represented is theso-called lexicographic ordering, in which, forexample, one good (such as one's child'sprobability of survival) holds absolute priorityover another good (such as one's chance to save$10). It may be noted parenthetically thatcertain criticisms of the benefit-cost analysis(voiced in connection with the Kosobud paper)can be interpreted as saying that when suchlexicographic ordering is present, no finite rateof substitution is appropriate, and theconventional calculus techniques do not apply.
18. See L. Hurwicz, 1986, On the Implementationof Social Choice Rules in Irrational Societies, inSocial Choice and Public Decision Making,Essays in Honor of KJ. Arrow, W.P. Heller,R.M. Starr, and D.A. Starrett (editors),Volume 1, Chapter 4, pp. 75-96.
19. Here we permit functions to be multivalued(i.e., the term function covers correspondencesas well as single-valued relations).
20. In the above example, each representative of alower body is a player.
21. In D. Starrett, 1972, Fundamental Non-Convexities in the Theory of Externalities,Journal of Economic Theory, 4:180-199.
100
Section V
General Technical Comments on Climate Change
Comment 1:Answers to Five Questions Posed at the Conference
Dr. John Firor*Advanced Study Programs
National Center for Atmospheric ResearchBoulder, Colorado
/. What are the projections of global averagetemperature increase?
The Intergovernmental Panel on ClimateChange (IPCC), assuming a "business as usual"scenario of emissions, projected a 0.3°C temperatureincrease per decade, with an uncertainty range of0.2°C to 0.5°C per decade. The authors of therecent update of the IPCC study looked at the manypapers that have been published since the originalreport and decided that there was no scientific needto adjust this earlier projection.
Climatologists have long recognized that thenatural variability of the climate system issufficiently large that the warming generated byhuman-induced emissions will not be clearly visiblein the global data until the end of this century orlater. Even so, the impatience of everyone to seeprogress in specifying the global rate of change hasled groups to make comparisons of model calcula-tions with the heating deduced from global datasets.
A number of groups have attempted suchcomparisons. They conclude that the global averagesurface temperature has increased approximately a
half degree in the last 100 years, whereas climatemodels retrospectively forecast an increase over thesame period of between a half degree and onedegree. Taken at face value, these comparisonsfavor those models that project the smallest futureheating, but when they are viewed in the context ofthe large natural variability, the comparisonsmerely remind us that it will be some years ordecades before the question of the rate of heatingcan be answered observaticnally. It should also beremembered that the same natural variability willprevent the 0.3°C-per-decade warming from being asteady increase.
Sea level is projected to : ise 6 cm per decade,with an uncertainty range of 3 to 10 cm per decade.
2. What are the prospects for projecting regionalclimate change?
The regional effects projected by IPCC are asfollows: Land surfaces will warm more rapidly thanthe ocean. In winter, high northern latitudes willwarm more than the global mean. Temperatureincreases in southern Europe and central NorthAmerica will be larger than the global mean and be
*Dr. John Firor of the National Center for Atmospheric Research is a distinguished climate scientist who hasdevoted much of his career to the description and analysis of the physics of the earth's atmosphere. He hasparticipated in and advised many national and international organizations concerned with climate changeand its impacts. At this workshop, he played a role as a resource expert on the underlying factual knowledgebase of climate science. In this capacity, he was asked to address several questions relevant to social scienceresearch. His responses follow. — Editors
101
Section V
accompanied by reduced summer precipitation andsoil moisture. There will be apparent regionalvariations in the rise in sea level because of therising or subsiding motions of islands andcontinental margins.
These projections are both coarse and uncertainand of little help to planners interested in preparingfor climate change at a specific locality. Therefore,there is a vigorous effort underway in variouslaboratories to improve the ability of climate modelsto simulate the sensitivity of regional climate toincreases in infrared-trapping gases in theatmosphere.
Much of the difficulty in foreseeing regionalchanges arises from the fact that global climatemodels are themselves coarse grained (i.e., a singlepoint used for calculation in the model represents alarge area on the surface of the earth, andimportant features on the surface such asmountains can be represented only crudely).Limitations in computer speed and memory preventmodelers from quickly increasing the spatialresolution of global climate simulations.
The approach that may be most effective forpromptly producing improvements consists ofimbedding a fine-mesh regional model of an area ofparticular interest into a global climate model. Tothe extent that features in the local regions (such asmountains, shorelines, and areas of vegetation) canbe thought of as processing large-scale flows of theatmosphere to produce the detailed local climates,this approach can be expected to produce progressin our ability to project regional climate changes.
3. Do the climate models produce outlooks for agroup of climate variables closely related tosocietal impacts?
Many variables are involved in generalcirculation model (GCM) calculations and could, inprinciple, be retained and displayed as time series.Those directly produced at each level above eachmodel grid point include pressure, density,temperature, wind, humidity, downflowing andupwelling radiation (both solar and long wave), andany cloud at that point. At the surface, bothprecipitation and evaporation are computed and soilmoisture is derived. For the ocean portion of the
model, variables such as salinity and the area of seaice are also calculated. Additional variables couldbe derived (e.g., amount of cloud cover or ocean-wave height). Newer models are beginning toinclude the concentrations of a few chemicals asvariables.
Some of these variables — in particular soilmoisture — are now routinely computed because ofrequests from scientists interested in responses ofagriculture to climate changes. Variables not nowretained in the data sets produced by modelcalculations could, if needed, be stored for later use,and new variables derivable from the ones alreadycalculated could be added to the calculation, butalways at a sacrifice of the length of time that canbe simulated with a fixed computer resource.
4. What is the feasibility and what are the costs ofproposed geoengineering options for respondingto climate change?
For the most part, proposed geoengineeringplans have not been studied in any detail and noanswer to the question is possible. The reason forthe lack cf study arises from the fact that most suchproposals involve complex systems, so a great dealof work needs to be done to convert the idea intoone that can be quantified in a climate or relatedmodel.
For example, the notion suggested by alaboratory experiment — that iron fertilization ofcertain ocean areas might enhance biologicalactivity and, through this activity, the ocean storageof atmospheric carbon — will require extensivework on ocean biology and integration of thisbiological knowledge into calculations of carbonfluxes to and from the ocean and transfers withinthe oceans. The ocean biological system issufficiently complicated that a single oceanexperiment now would not provide final answers. Ifchanges in carbon fluxes were observed in such anexperiment, one would not necessarily know thechain of events that produced the changes; thus, theexperiment could not readily be scaled up. Further,without additional knowledge of the oceanic carboncycle, one would not know whether the carbon takenup would stay stored for the centuries required orwhether unacceptable shifts in the ocean food-chainmight be an accidental result of the plan. Thus,
102
General Technical Comments on Climate Change
decades of work, large teams of scientists, and alarge amount of funding for field observations andmode! experiments would be required to begin toassess the feasibility of the proposal.
Some of these steps are underway. Scientistsfamiliar with ocean biology in the suggested regionshave studied the proposal and presented argumentsthat the net long-term carbon storage would besmall, thus beginning the usual scientific process oftesting a new idea. Other groups have also begunto apply what they already know to the question. Ifwork over several years by many groups shows afavorable possibility for carbon storage and a lack oftroublesome side effects, we should then expectplans to be formulated for some sort of field test ofthe proposal.
The iron fertilization suggestion envisions amethod of removing one of the causes of human-induced climate change from the atmosphere. Inthat respect, it is similar to the proposal that weplant trees to remove carbon from the air. Mostother geoengineering proposals, however, aredesigned to counteract one feature of the resultingclimate change and, as such, have a defect thatslows their consideration. A major climate-changeevent, such as the one that we are now inducing, iscomplex, involving all aspects of weather andclimate — cloudiness, wind, temperature, andprecipitation. The defect of most geoengineeringproposals is they attempt to slow one kind ofclimate change (temperature increase) by thecreation of another kind of climate change (e.g.,change in cloudiness or atmospheric transparency).So far, none of the proposals has included a carefulstudy of the effects of the different kinds of climate
change that would be produced by the proposedaction.
Finally, all geoengineering proposals share theproblem that their true impacts are as difficult toproject as the impact of additional carbon dioxide inthe atmosphere. We see evidence in the prepara-tions for the meeting to be held in Rio that a roughestimate of the global average rate of climatechange may be sufficient to trigger actions to slowthat rate of change. But a much more precisecalculation would likely be required before peoplewould intentionally set about to modify the climate.
5. What climate variables co-vary with globalaverage temperature?
Both average absolute humidity and globalrainfall should increase as temperature increases.However, the rainfall would not be distributedevenly or even proportionally to current rainfall.The estimate from IPCC is that precipitation willincrease in winter in mid latitudes and at allseasons in high latitudes and in die tropics.
Scientists have speculated about whetheratmospheric variability will change as temperatureincreases, or whether the strength of hurricanes willincrease with higher surface temperatures, but noconclusions are yet possible.
It is, of course, very likely that the frequency ofhot spells will increase and die frequency of coldspells will decrease in a hotter climate, but, again,giving details of such a change will requireimproved regional calculations.
703
Section V
Comment 2:Nurturing Multidisciplinary Research on the Global Commons
David FeenyDepartment of Economics
McMaster UniversityHamilton, Ontario, Canada
Both an improved understanding of the causesand consequences of global warming as well as theexploration of responses to global warming requirethe integration of knowledge from a wide variety ofdisciplines in the natural sciences, social sciences,and humanities. The need for high-quality researchdrawing on a number of disciplines has a number ofimplications.
There are a variety of examples of successfulmultidisciplinary enterprises that have conducteduseful research over an extended period of time. Itis possible, however, that the challenges involved inthe creation and maintenance of multidisciplinaryresearch teams and institutes to date have been lessformidable than the challenges facing the creationof appropriate institutional arrangements forresearch on global warming. Global wanningappears to require both greater integration among anumber of disciplines in the natural sciences andgreater integration with social science disciplines.Multidisciplinary success in the health sciences andagricultural sciences has probably been lessdemanding.
Successful multidisciplinary settings wouldappear to share several characteristics. First, it isimportant that there be a critical mass ofinvestigators from each key discipline to provideongoing personal development and intellectualsupport within each discipline. It is also importantthat each member of such a research enterprise becommitted to the conduct of multidisciplinary teamresearch, at least over the medium term. In such asituation, there is an incentive to bear the heavyinitial costs of learning how to communicate amongthe disciplines. It is also important to nurture aculture of mutual respect for the contributions ofeach member of the team. The involvement ofsenior, well-respected investigators from a numberof disciplines is important to legitimize the
intellectual enterprise and provide useful rolemodels for more junior investigators.
Standard academic programs in graduateschools are often not well suited for the training ofinvestigators for such settings. It remainsimportant, however, that such investigators obtainsolid training in their own discipline. Suchtraining, when combined with practical experienceworking in teams (perhaps as a research associateof a faculty supervisor or in a postdoctoral position),might provide a viable training path for disciplinaryscholars capable of working effectively inmultidisciplinary teams.
The traditional departmental/disciplinary-basedstructure of academic settings is also often not wellsuited to the conduct of multidisciplinary research.The reality is that publications in journals in one'sdiscipline are weighted more heavily than publica-tions elsewhere. Successful ongoing multidisci-plmary programs of research will require that theincentive structure be altered to recognizeachievements in other fields. Disciplinary-baseddepartmental committees and chairs will have toengage in communication with departments on therelative merits of various outlets. Explicit ex anteagreements on the relative weights to be accorded toaccomplishments in various activities might beuseful (e.g., 25% for teaching, 35% for researchwithin one's discipline, 25% for multidisciplinaryresearch, and 15% for service).
Institutional mechanisms to support thetraining of investigators capable of contributing tomultidisciplinary research and to support theconduct of such research will need to be furtherelaborated to enhance the capacity to providerigorous and timely basic and applied research onglobal warming. Medium-term commitments will berequired to provide the incentives to overcome the
104
General Technical Comments on Climate Change
initial fixed costs of establishing interdisciplinary global wanning research will have importantcommunication and collaborative working styles. spillover benefits in other areas.Accomplishments in creating such arrangements for
105
Section V
Comment 3:The Easearch Agenda:
Social Science Research on Global Warming
Professor David FeenyDepartment of Economics
McMaster UniversityHamilton, Ontario, Canada
The institutional analysis group discussed anumber of topics and methodological approachesthat would constitute important components of thesocial science contribution. I have elaborated uponthose discussions to record a somewhat personallyidiosyncratic' set of items for the social scienceresearch agenda.
A number of social science and historicaldisciplines contribute relevant case studies andcomparative case studies. Economic and tech-nological history studies that address the factorsthat affect the speed and direction of bias intechnological change are of obvious relevance andwill be useful in informing presumptions about thedegree of substitutability within and among naturaland man-made inputs to production. Case studiesof existing and historical nested hierarchies are alsovery important.
Norman Rosenberg reviewed the Resources forthe Future MINK study (Missouri, Iowa, Nebraska,and Kansas) for the group. The study included acareful descriptive analysis of the economic activi-ties of sectors in the region that are sensitive tochanges in climate (agriculture, forestry, waterresource management, energy). Carefully calibratedmodels describing inputs and outputs were devel-oped. Simulations were then conducted to assessthe independent effects of autonomous economicchange, autonomous climate change, and their inter-action. The results represent one meaningful andattractive approach for assessing economic-change/climate-change interactions. Additional studiessuch as the MINK study and alternative approachesfor the assessment of economic-change/climate-change interactions appear to be warranted.
A number of disciplines in the social scienceshave developed measures that reflect nonpecuniary
objectives and aspects of human quality of life otherthan the material standard of living. The socialindicators approach provides a broad measure ofquality of life. Within health-services research,there are a variety of broadly and narrowly focusedmeasures of health-related quality of life. Prefer-ence measurement approaches developed to evaluatethe health-related, quality-of-life value of changes inhealth status may be an especially useful model fordeveloping methods for eliciting preferences overnonpecuniary environmental outcomes.
In recent decades, the social sciences haveexploited small-scale laboratory experiments tosharpen the ability to test theory or the efficacy ofproposed new institutional mechanisms. Therehave been a number of informative studies onsmall-group collective-action problems, committeebehavior, voting rules, and other topics relevant tomanaging the global commons.
The social sciences have also conducted large-scale field experiments to inform assessments ofmajor social policy issues. Examples includenegative-income tax, alternative pricing policies forpublic utilities, and health-care expenditureinsurance experiments. Although field experimentsare difficult and expensive and although suchexperiments involve ethical issues, the conduct ofsuch research has much to contribute to theformulation of well-informed, effective policyinterventions.
Studies on the design of incentive-compatiblemechanisms are crucial for improved policyformulation for global climate change problems.Under existing arrangements, private actors takeinto account private costs and benefits that poorlyreflect the underling social costs and benefits oftheir actions. New institutional arrangements that
706
General Technical Comments on Climate Change
would more closely align private and social costsand benefits would help to ameliorate global changeproblems. Our understanding of the design,creation, and maintenance of such incentive-compatible mechanisms needs to be furtherdeveloped. Better understanding is needed at thelevel of theoretical/analytical knowledge. It is also
needed at the level of practical implementation andthe transition path from one set of institutionalarrangements to another. A variety of analyticalapproaches, including noncooperative and coopera-tive game theory, are of relevance. These analyticaltechniques need to be complemented by experi-mental methods and case studies.
707
Section V
Comment 4:Policy for Sustainable Development:
Some Unresolved Issues
Professor Vernon W. RuttanDepartment of Agriculture and Applied Economics
University of MinnesotaSt. Paul, Minnesota
1 INTRODUCTION
We are now in the midst of the third wave ofsocial concern since World War II about theimplications of natural resource availability andenvironmental change for the sustainability ofimprovements in human well-being.
The first wave of concern in the late 1940s andearly 1950s focused primarily on the quantitativerelationship between resource availability andeconomic growth. The response to this first wave ofconcern was technical change in raw materialproduction. The second wave of concern, in the late1960s and early 1970s, focused on the risingdemand for assimilation by the environment ofresiduals derived from growth in commodityproduction and consumption — primarily at thelocal and regional level. Since the mid 1980s, thesetwo early concerns have been supplemented by athird. This more recent concern centers around theimplications for environmental quality, foodproduction, and human health of a series ofenvironmental changes such as global warming,ozone depletion, acid rain, and others that areoccurring on a transnational scale. Before actionsto resolve these more recent challenges tosustainable economic growth are taken, a number ofunresolved issues must be confronted to allow acommitment to this resolution to be translated intoan internally coherent reform agenda. In this note,I discuss three of these issues.
2 ISSUE OF SUBSTITUTABILITY
One area where our knowledge is inadequate iswith respect to the role of technology in wideningthe substitutability among natural resources andbetween natural resources and reproducible capital.Economists and technologists have traditionally
viewed technical change as widening the possibilityof substitution among resources (of fertilizer forland, for example) (Solow 1974; Goeller andWeinberg 1976). The sustainability communityrejects the "age of substitutability" argument. Theloss of plant genetic resources is viewed as apermanent loss of capacity. The elasticity ofsubstitution among natural factors and betweennatural and man-made factors is viewed asexceedingly low (James, Nijkamp, and Opschoor1989; Daly 1991). This is an argument, ineconomists' language, over the form of theproduction function. If a combination of capitalinvestment and technical change widens theopportunity for substitution, imposing constraintson present resource use could leave subsequentgenerations less well off. On the other hand, if realoutput per unit of natural resource input isnarrowly bounded (i.e., cannot exceed some upperlimit that is not too far from where we are now),catastrophe is unavoidable. While the argumentbetween adherents to these two views is often castin philosophical terms, empirical research shouldlead toward a convergence. But since the dataneeded to resolve the issue will never be adequate,continued disagreement can be expected.
3 ISSUE OF OBLIGATIONS TOWARDTHE FUTURE
The second issue is one that has dividedtraditional resource economists and thesustainability community. That is the issue of howto deal analytically with the obligations of thepresent generation toward future generations. Theissues of intergenerational equity is at the center ofthe sustainability debate (Pearce, Barbier, andMakrandaya 1990; Solow 1991). Environmentalistshave been particularly critical of the approach usedby resource and other economists in valuing future
108
General Technical Comments on Climate Change
benefit and cost streams. The conventional approachinvolves the calculation of the "present value" of aresource development or protection project bydiscounting the cost and benefit stream by some"real" rate of interest (i.e., an interest rate adjustedto reflect the costs of inflation).
The critics insist that this approach results in a"dictatorship of the present" over the future. Atconventional rates of interest, the present value of adollar of benefits 50 years into the futureapproaches zero. "Discounting can make molehillsout of even the biggest mountain" (Batie 1989,p. 1092). Solow has made the same point in moreformal terms. He notes that if the marginal profit(marginal revenue less marginal cost) to resourceowners rises more slowly than the rate of interest,production is pushed nearer in time, and theresource would be exhausted quickly (Solow 1974,p. 3; Lipton 1991).
A question that has not been adequatelyanswered is this: As a result of the adoption of awidely held sustainability "ethic," would themarket-determined discount rates decline towardthe rate preferred by those advancing thesustainability agenda?1 Or would it be necessary toimpose sumptuary regulations in an effort to inducesociety to shift the income distribution morestrongly toward future generations? It is clear, atleast to me, that in most countries, efforts toachieve sustainable growth must involve somecombination of (1) higher contemporary rates ofsaving (i.e., deferring present consumption in favorof future consumption) and (2) more rapid technicalchange, particularly technical changes that willenhance resource productivity and widen the rangeof substitutability among resources.
4 ISSUE OF INCENTIVE-COMPATIBLEINSTITUTIONAL DESIGN
A third area in which knowledge needs to beadvanced is the design and implementation ofinstitutions that are capable of internalizing (withinindividual households, private firms, and publicorganizations) the costs of actions that generate thenegative spillover effects (i.e., the residuals) thatare the source of environmental stress. Underpresent institutional arrangements, importantelements of the physical and social environment
continue to be undervalued for purposes of bothmarket and nonmarket transactions. Traditionalproduction theory implies that if the price to a userof an important resource is undervalued, it will beoverused. If the price of a factor (the capacity of theatmosphere to absorb greenhouse gases, forexample) is zero, it will be used until the value ofits marginal product to the user approaches zero.This condition will be true, even though it may beimposing large social costs on society.
The dynamic consequences of failure tointernalize these costs are even more severe. In anenvironment characterized by rapid economicgrowth and changing relative factor prices, failureto internalize resource costs will bias the directionof technical change (Ruttan 1971). The demand fora resource that is priced below its social cost willgrow more rapidly in this environment than in asituation where substitution possibilities areconstrained by existing technology. As a result,"open access" resources will undergo stress ordepletion more rapidly than they would in a worldcharacterized by a static technology or even byneutral (unbiased) technical change.
The design and implementation of incentive-compatible institutions (institutions capable ofachieving compatibility between individual,organizational, and social objectives) remain, at thisstage, an art rather than a science. The incentivecompatibility problem has not been solved even atthe most abstract theoretical level.2 This deficiencyin institutional design and implementation capacityis evident in our failure to institute policies capableof achieving contemporary distributional equity,either within countries or among rich and poorcountries. It impinges with even greater force onour capacity to design institutions capable ofachieving intergenerational equity.
5 AN UNCERTAIN FUTURE
In closing, I would like to emphasize how far weare from being able to design either an adequatetechnological or an adequate institutional responseto the issue of how to achieve sustainable growth.At present, there is no package of technology thatcan ensure the sustainability of growth inproduction at a rate to meet the demands that arebeing placed on either developed or developing
70.9
Section V
countries. Sustainability is appropriately viewed asa research agenda rather than as a guide to practice(Ruttan 1988; Graham-Tomasi 1991).
At present, the sustainability community hasnot been able to advance a program of institutionalinnovation or reform that can provide a credibleguide to the organization of sustainable societies.We have yet to design the institutions that canensure intergenerational equity. Few wouldchallenge the assertion that future generations haverights to levels of sustenance and amenities that areat least equal to those enjoyed (or suffered) by thepresent generation. They also should expect toinherit improvements In institutional capital(including scientific and cultural knowledge) neededto design more productive and healthyenvironments.
My conclusion with respect to institutionaldesign is similar to that which I have advanced inthe case of technology. Economists and other socialscientists have made a good deal of progress incontributing the analysis needed for "coursecorrection." But capacity to contribute toinstitutional design and implementation remainslimited. The fact that the problem of designing andimplementing incentive-compatible institutions(institutions capable of achieving compatibilitybetween individual, organizational, and socialobjectives) has not been solved at even the mostabstract theoretical level means that institutionaldesign proceeds in an ad hoc trial-and-error basisand that the errors continue to be expensive.Institutional innovation and reform shouldrepresent a high-priority research agenda.
6 END NOTES
1. The question of the impact of the use of apositive discount (or interest) rate on resourceexploitation decisions is somewhat more complexthan often implied in the sustainabilityliterature. Simply lowering the discount rate tofavor the natural resource sector will not ensureslower exploitation of natural resources if themarket rate of interest remains high.Recipients of the lower interest rates maytransfer the revenue from resource exploitationto investments that have higher rates of return
rather than reinvesting to sustain the flow ofresource benefits. Furthermore, high rates ofresource exploitation can be consistent with eitherhigh or low interest rates. In the case of forestexploitation, for example, a low discount rate favorsletting trees grow longer and planting trees thattake longer to grow. On the other hand, a lowdiscount rate will make it profitable to invest inmineral exploitation, land and water development,or other investment projects that might otherwisebe unprofitable. That is why, in the past, resourceeconomists and environmentalists have argued infavor of higher interest rates on public waterresource projects (Norgaard 1991; Price 1991;Graham-Tomasi 1991). As an alternative to lowerdiscount rates, Mikesell (1991) suggests takingresource depletion into account in project benefit-cost analysis.
2. The concept of incentive compatibility wasintroduced in a 1972 paper by Hurwicz (1972).In that paper, he showed that it was notpossible to specify an infonnationallydecentralized mechanism for resouice allocationthat simultaneously generates efficient resourceallocation and incentives for consumers tohonestly reveal their true preferences. For thecurrent state of knowledge in this area, seeGroves, Radner, and Reiter (1987).
7 REFERENCES
Batie, Sandra, 1989, Sustainable Development:Challenges to the Profession of AgriculturalEconomics, American Journal of AgriculturalEconomics, pp. 1085-1101, Dec.
Daly, Herman E., 1991, From Empty WorldEconomics to Full World Economics: Recognizingan Historical Turning Point in EconomicDevelopment, Environmental Working Paper No. 46,pp. 18-26, in Environmentally Sustainable EconomicDevelopment: Building on Bruntland, RobertGoodland, Herman Daly, and Salah El Seraly(editors), World Bank, Washington, D.C., July.
Goeller, H.E., and Alvin M. Weinberg, 1976, TheAge ofSubstitutability, Science, 797:683-689,Feb. 20.
110
General Technical Comments on Climate Change
Graham-Tomasi, Theodore, 1991, Sustainability:Concepts and Implications for Agricultural ResearchPolicy, pp. 81-102, in Agricultural Research Policy:International Quantitative Perspectives, Philip G.Pardey, Johannes Roseboom, and Jack R. Anderson,Cambridge University Press.
Groves, Theodore, Roy Radner, and Stanley Reiter(editors), 1987, Information, Incentives, andEconomic Mechanisms, University of MinnesotaPress, Minneapolis, Minn.
Hurwicz, Leonid, 1972, On InformationallyDecentralized Systems, pp. 297-333, in Decision andOrganization, C.B. McGuire and Roy Radner(editors), North Holland Publishing Co.,Amsterdam, the Netherlands.
James, D.E., P. Nijkamp, and J.B. Opschoor, 1989,Ecological Sustainability in Economic Development,pp. 27-48 in Economy and Ecology: TowardSustainable Development, F. Archibugi and P.Nijkamp (editors), Dordrecht, the Netherlands.
Lipton, Michael, 1991, Accelerated ResourceDegradation by Third World Agriculture: Createdin the Commons, in the West, or in Bed, paperpresented at Seminar on Agricultural Sustainabilityand Poverty Alleviation: Issues and Polices, held inFeldafing, Germany, on Sept. 23-27.
Mikesell, Raymond F., 1991, Project Evaluation andSustainable Development, Environmental WorkingPaper No. 46, pp. 54-60, in EnvironmentallySustainable Economic Development: Building onBruntland, Robert Goodland, Herman Daly, andSalah El Seraty (editors), World Bank, Washington,D.C., July.
Norgaard, Richard B., 1991, Sustainability asIntergenerational Equity: The Challenge toEconomic Thought and Practice, World Bank, Officeof the Chief Economist, Asia Region, Washington,D.C., May 22.
Pearce, David W., Edward Barbier, and AnilMakrandaya, 1990, Sustainable Development-Economics and Environment in the Third World,Gower Publishing Company, Brookfield, Vt, pp. 23-56.
Price, Colin, 1991, Do High Discount Sates DestroyTropical Forests?, American Journal of AgriculturalEconomics, 42:71-85, Jan.
Ruttan, Vernon W., 1971, Technology and theEnvironment, American Journal of AgriculturalEconomics, 53:707-717, Dec.
Ruttan, Vernon W., 1988, Sustainability Is NotEnough, American Journal of AlternativeAgriculture, 3:128-130, spring/summer.
Solow, Robert M., 1974, The Economics of Resourcesor the Resources of Economics, American EconomicReview, 64:1-14, May.
Solow, Robert M., 1991, Sustainability: EconomistsPerspective, J. Seeward Johnson Lecture, WoodsHole Oceanographic Institution, Marine PolicyCenter, Woods Hole, Mass., June 14.
Ill
Section V
Comment 5:Agricultural Response to Climate Change
Professor Vernon W. RuttanDepartment of Agriculture and Applied Economics
University of MinnesotaSt. Paul, Minnesota
1 DISCUSSION
The complex interrelationship between globalclimate change and agricultural production willbecome one of the most significant policy issues, inboth developed and developing countries, in the firstdecades of the 21st century. Global and regionalclimate change will modify both agriculturalproduction capacity and its location. And theintensity of agricultural production will contributeto environmental change at both the regional andglobal levels.
The best single source currently available forevaluating the impact of climate change onagriculture is the recent book by Martin L. Parry,Climate Change and World Agriculture.1 Parrynotes that two major types of adjustments may beanticipated at the farm and the regional level.These are changes in land use and changes in farmmanagement.2 But there is almost no discussion byParry, or in the climate change literature moregenerally, of the capacity of the agriculturalresearch system to "invent around" the newtemperature and rainfall regimes associated withglobal warming. Yet the historical record suggeststhat modest investments in agricultural researchhave enabled societies characterized by vastlydifferent resource endowments to achieve relativelyrapid growth in agricultural production. In ourbook Agricultural Development and in otherwritings, Yujiro Hayami and I have developed andtested, against the experience of both developed anddeveloping countries, an "induced technical change"model, in which the direction of technical change isinduced by changes in relative resourceendowments.3 Countries with comparable levels ofland per worker have been able to achieve widelydifferent levels of output per hectare and per worker(Figure 1). It is clear that changes in relativeresource endowments have, in the past, had a
pervasive impact on the adoption of technicalinnovations by farmers, on the supply of theindustrial inputs in which technology is embodied,and on the direction of research efforts byagricultural scientists and technologists in bothpublic- and private-sector institutions. This inducedbehavior must be more explicitly built into theinteractive models if we are to build more crediblemodels to estimate the impacts of climate change.
The fact that the agricultural research systemhas in the past been reasonably successful, at leastin the presently developed countries, in respondingto growth in demand and to changes in resourceendowments cannot, of course, be taken asassurance that it will be able to do so in the future.4
Its capacity to do so will depend both on the rate ofclimate change and on the level of investment inagricultural research. We now have in place aneffective system of international agriculturalresearch centers (IARCs) funded by a donorsconsortium, the Consultative Group forInternational Agricultural Research (CGIAR). Inspite of the high rates of return generated by theCGIAR institute, there were, beginning in the mid1980s, growing signs of "aid-weariness" on the partof the donor countries. A more serious concern,however, is the limited national agriculturalresearch capacity in the poorest developingcountries.6 Even Brazil, which was remarkablysuccessful in strengthening its national agriculturalresearch capacity in the 1970s, has experiencedsubstantial erosion of that capacity since the mid1980s. And biological technology, needed to makethe location- specific technical changes needed toadapt to climate change, will not be available.
The capacity of agricultural research systems,even in the most highly developed countries, torespond to the effects of global warming will dependon the rate at which climate change occurs. We are
112
General Technical Comments on Climate Change
tY/A)
20
CJ
do
Q
! •u.OUJ
oc
IS
: * f
ECT
a.
o.
RAL
o
o<
0.1
Ta(72|-''
Eg (48)
••\v
Ba(13>
'*\v
Japan 1(211/
1n(44)/ NZC88
USA
France 1880(51>'' ' UK. 1880184)
.Br(55)
SAI72) /
Par(40)
5 10 50 100
AGRICULTURAL OUTPUT PER MALE WORKER tLOG. SCALE)
250
FIGURE 1 Agricultural Output
on the edge of a transition from an older Mendelian-based biological technology to a new molecular-based biological technology. At a conference held atthe University of Minnesota in July of 1989, a groupof leading agricultural scientists attempted to assessrelative contributions of the "old" and "new"biotechnology to agricultural production over thenext several decades.
Several conclusions are particularly relevant tothe issue of global warming.6
• Advances in conventional technology willremain the primury source of growth in cropand animal production over the next quartercentury. Almost all increases in agriculturalproduction in the future must come from
113
Section V
further intensification of agriculturalproduction on land that is presently devotedto crop and livestock production. Until wellinto the second decade of the next century,the necessary gains in crop and animalproductivity will continue to be generated byimprovements resulting from conventionalplant and animal breeding and from moreintensive and efficient use of technical inputsincluding chemical fertilizers, pest controlchemicals, and higher-quality animal feeds.
Advances in conventional technology will,however, be inadequate to sustain thedemands that will be placed on agriculture aswe move into the second decade of the nextcentury and beyond. Advances in crop yields,particularly in wheat and rice, have largelyaccrued as a result of efforts by plantbreeders to (1) change plant architecture tofacilitate increases in plant populations perhectare and (2) increase the ratio of grain tostraw rather than to increase total drymatter production. Advances in animal feedefficiency have come about throughdecreasing the proportion of feed consumedthat is devoted to animal maintenance andincreasing the proportion used to produceusable animal products. There are severephysiological constraints to continuedimprovement along these conventional paths.These constraints are most severe in thoseareas that have already achieved the highestlevels of productivity — areas in westernEurope, North America, and parts of easternAsia.
Advances in basic science, particularly inmolecular biology and biochemistry, areopening new possibilities for supplementingtraditional sources of plant and animalproductivity growth. The possibilities rangefrom the transfer of growth hormones intofish to conversion of lignocellulose into edibleplant and animal products. The realizationof these possibilities will require areorganization in the performance of agricul-tural research. An increasing share of thenew knowledge generated by research willreach producers in the form of proprietaryproducts or services. This means that theincentives exist to draw substantially more
private-sector resources into agriculturalresearch. Within the public sector, researchorganizations will have to move increasinglyfrom a "little science" to a "big science" modeof organization. The capacity of national andinternational agricultural research systemsto successfully negotiate the transition fromthe old to the new biological technology wiilbe of critical importance in determining theimpact of climate change on agriculturalproduction as we move beyond the firstquarter of the next century.
2 END NOTES
1. Parry, M.L., 1991, Climate Change and WorldAgriculture, Earthscan Publications, London,United Kingdom, in association with theInternational Institute for Applied SystemsAnalysis and the United Nations EnvironmentalProgram, 1990. See also the review byV.W. Ruttan in Environment, 33:25-29,July/Aug., 1991.
2. Parry, M.L., Climate Change, p. 119.
3. Yujiro Hayami and V.W. Ruttan, AgriculturalDevelopment: An International Perspective, TheJohns Hopkins University Press, Baltimore,Md., 1st ed. 1971, 2nd ed. 1985; also V.W.Ruttan, 1971, Technology and the Environment,American Journal of Agricultural Economics,53:707-717, Dec.
4. There is now a large body of researchdocumenting the underinvestment in agricul-tural research in both developed and developingcountries. The high annual rates of return thatare documented in these studies, which typicallyrange from 30% upward to 100%, are an indica-tion of underinvestment. See V.W. Ruttan, 1982,Agricultural Research Policy, University ofMinnesota Press, Minneapolis, Minn., pp. 242-248. For a more recent compilation, see RubenG. Echeverria, 1990, Assessing the Impact ofAgricultural Research, Vol. II in Methods forDiagnosing Research System Constraints andAssessing the Impact of Agricultural Research,R.G. Echeverria (editor), International Servicefor National Agricultural Research (ISNAR), theHague, the Netherlands.
114
General Technical Comments on Climate Change
5. See Philip G. Pardey and Johannes Roseboom, 6. V.W. Ruttan (editor), 1989, Biological and1989, ISNAR Agricultural Research Indicator Technical Const: aints on Crop and AnimalSeries: A Global Data Base on National Productivity: Report on a Dialogue, Staff PaperAgricultural Research Systems, Cambridge P89-45, University of Minnesota, Department ofUniversity Press; Philip G. Pardey, Johannes Agricultural and Applied Economics, St. Paul,Roseboom, and Jock R. Anderson (editors), 1991, Minn, Nov.Agricultural Research Policy: InternationalQuantitative Perspectives, Cambridge UniversityPress, London, United Kingdom, forthcoming;Johannes Roseboom and Philip G. Pardey, 1990,Measuring the Development of NationalAgricultural Research Systems, Staff Notes9095, International Service for NationalAgricultural Research (ISNAR), the Hague, theNetherlands, Nov.
775
Section V
116
Section VI
Conference Program
FEBRUARY 11,1992
6:00 - 8:00 Reception and Dinner
8:00 - 10:30 WelcomeJoel Snow, Associate Vice President for Argonne National Laboratory, University of ChicagoHarvey Drucker, Associate Laboratory Director, Argonne National Laboratory
Keynote Address and Discussion
Where Are We Now? Where Are We Going?John Eddy, University of Michigan
FEBRUARY 12,1992
9:00 -12:15 PLENARY SESSION I: DISCUSSION OF BACKGROUND PAPERS
9:00 - 9:10 Opening RemarksHarold Jacobson, University of Michigan (Chair)
9:10 -10:05 Workshop in EconomicsThe Problem of Climate Change Benefit-Cost Analysis
Richard Kosobud, University of Illinois at ChicagoDiscussant: Khairy Tourk, Illinois Institute of Technology
10:05 -10:15 Break
10:15 -11:10 Workshop in Adaptation and Mitigation Strategies
Approaching Global Warming: A Review of the Adaptation and Mitigation PerspectivesPeter Morrisette, Resources for the FutureDiscussant: Rob Coppock, National Academy of Sciences
11:10 -11:20 Break
11:20 -12:15 Workshop in Political Institutions
Institutional Analysis and Global Climate Change: Design Principlesfor Robust International Regimes
Michael McGinnis and Elinor Ostrom, Indiana UniversityDiscussant: Philip Schrodt, University of Kansas
12:30 - 2:00 Lunch
2:00 - 5:15 Small Workshop Groups
5:30 - 7:30 Dinner
117
Section VI
FEBRUARY 13,1992
9:00 - 10:20 PLENARY SESSION II: REVIEW OF WORKSHOP DISCUSSIONS,WITH EMPHASIS ON CROSS-CUTTING ISSUES
Joel Snow, Associate Vice President forArgonne National Laboratory, University of Chicago
10:20 - 10:40 Coffee Break
10:40 - 12:00 Plenary Session II (continued)
12:15 -1:45 Lunch
1:45 - 3:15 Small Workshop Groups
3:15 - 3:30 Coffee Break
3:30 - 3:45 Closing Remarks and AdjournmentHarold Jacobson, University of Michigan (Chair)
118
Section VII
Conference Participants
Mr. David AdmiraalDepartment of Political ScienceUniversity of Illinois
Dr. Gilbert W. Bassett, Jr.Department of EconomicsUniversity of Illinois at Chicago
Dr. Gale BoydEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Dr. Bob Coppock, Staff DirectorPanel on Policy Implications of
Greenhouse WarmingNational Academy of Sciences
Ms. Kirsten DowCenter for Technology,
Environment and DevelopmentClark University
Dr. John EddyConsortium for International
Earth Science InformationNetwork
Professor David FeenyDepartment of EconomicsMcMaster University
Dr. John W. Firor, DirectorAdvanced Study ProgramsNational Center for Atmospheric
Research
Professor Roy GardnerDepartment of EconomicsIndiana University
Dr. David GoetzDepartment of Political ScienceUtah State University
Dr. Donald HansonEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Dr. Madeleine HostiInstitute for Social ResearchUniversity of Michigan
Professor Barry B. HughesGraduate School of International
StudiesUniversity of Denver
Professor Leonid HurwiczDepartment of EconomicsUniversity of Minnesota
Professor Harold Jacobson,Director
Center for Political StudiesUniversity of Michigan
Professor Roger KaspersonDepartment of Government and
GeographyClark University
Dr. David KeeRegional Air & Radiation DivisionU.S. Environmental Protection
Agency
Dr. Elizabeth KirkSenior Program OfficerAmerican Association for the
Advancement of Science
Professor Richard KosobudDepartment of EconomicsUniversity of Illinois at Chicago
119
Section VII
Professor Douglas MacLeanDepartment of PhilosophyUniversity of Maryland,
Baltimore County
Professor Howard MargolisGraduate School of Public Policy
StudiesUniversity of Chicago
Ms. Annette MathenyDepartment of Political ScienceUniversity of Illinois
Professor Michael McGinnisPolitical Science DepartmentIndiana University
Mr. Daniel McGrathEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Dr. Peter MorrisetteClimate Resources ProgramResources for the Future
Professor Elinor Ostrom,Co-Director
Workshop in Political Theory andPolicy Analysis
Indiana University
Mr. Kevin PorterSchool of EngineeringUniversity of Illinois
Mr. Kevin G. QuinnEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Dr. Kilaparti RamakrishnaInternational and Environmental
LawThe Woods Hole Research Center
Mr. Geoffrey ReynoldsWorkshop in Political Theory and
Policy AnalysisIndiana University
Dr. Marian RiceExecutive AssociateAmerican Academy of Arts and
Sciences
Dr. Norman RosenbergEnergy and Natural Resources
DivisionResources for the Future
Professor Vernon RuttanDepartment of Agriculture and
Applied EconomicsUniversity of Minnesota
Professor Philip A. SchrodtDepartment of Political ScienceUniversity of Kansas
Dr. Joel SnowAssociate Vice President for
Argonne National laboratoryUniversity of Chicago
Mr. David SouthEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Dr. David StreetsEnvironmental Assessment and
Information Sciences DivisionArgonne National Laboratory
Professor Khairy TourkStuart School of BusinessIllinois Institute of Technology
Mr. Paul TurnerWorkshop in Political Theory and
Policy AnalysisIndiana University
Mr. J. Michael WalshEconomic Analysis and Planning
DepartmentChicago Board of Trade
Ms. Judith L. WeddleDepartment of Political ScienceUniversity of Kansas
720
Conference Participants
Dr. Dee Wemette Dr. Richard WinterEnvironmental Assessment and Environmental Assessment and
Information Sciences Division Information Sciences DivisionArgonne National Laboratory Argonne National Laboratory
121
