AFTU1Africa Water and Urban 1The World Bank1818 H Street NWWashington, DC 20433www.worldbank.org

EthiopiaManaging Water Resources to Maximize Sustainable Growth

C O U N T RY WAT E R R E S O U RC E S A S S I S TA N C E S T R AT E G Y

REPORT NO. 36000-ET

CO

UN

TR

Y W

AT

ER

RE

SO

UR

CE

S A

SS

IST

AN

CE

ST

RA

TE

GY

ET

HIO

PIA

CWRAS_Tmplt_Ethiopia.qxd 5/3/06 12:39 PM Page 1

THE WORLD BANKAGRICULTURE AND RURAL DEVELOPMENT DEPARTMENT

Ethiopia:Managing WaterResources to MaximizeSustainable Growth A World Bank Water Resources Assistance Strategy for Ethiopia

© 2006 The International Bank for Reconstruction and Development / The World Bank1818 H Street, NWWashington, DC 20433Telephone 202-473-1000Internet www.worldbank.org/ruralE-mail ard@worldbank.org

All rights reserved.

This volume is a product of the staff of the International Bank for Reconstruction andDevelopment/The World Bank. The findings, interpretations, and conclusions expressedin this paper do not necessarily reflect the views of the Executive Directors of The WorldBank or the governments they represent. The World Bank does not guarantee the accu-racy of the data included in this work. The boundaries, colors, denominations, and otherinformation shown on any map in this work do not imply any judgment on the part ofThe World Bank concerning the legal status of any territory or the endorsement oracceptance of such boundaries.

The material in this publication is copyrighted. Copying and/or transmitting portions orall of this work without permission may be a violation of applicable law. TheInternational Bank for Reconstruction and Development/ The World Bank encouragesdissemination of its work and will normally grant permission to reproduce portions ofthe work promptly.

For permission to photocopy or reprint any part of this work, please send a requestwith complete information to the Copyright Clearance Center, Inc., 222 RosewoodDrive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470,http://www.copyright.com/.

All other queries on rights and licenses, including subsidiary rights, should be addressedto the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org.

A C R O N Y M S A N D A B B R E V I AT I O N S vii

E X E C U T I V E S U M M A RY xi

A C K N O W L E D G M E N T S xi

Introduction 1

Ethiopia’s Water Resources 3Resource 3Hydrological Variability: Extremes of Drought and Flood 8Interrelated Effects on Natural Resources and the

Environment 12Hydrological Variability: Endemism and Expectations 13International Waters: Tensions and Opportunities 13

Water Resources and Key Economic Sectors 15Water Supply and Sanitation 15Agriculture, Livestock, and Fisheries 19Transportaion and Market Infrastructure 23Energy and Hydropower 27Urban, Manufacturing, and Services 28Food Aid 31

Water Resources Challenges: Some Priority Responses 33Strengthening Water Resources Development and

Management 33Enhancing Economic Resilience 37

iii

Contents

1

2

4

3

iv Contents

A Hydro-Economic Model of Ethiopia: A Summary of Findings 41

Variations of the Model 41Impacts on Growth and Poverty 42Investment Scenarios 44Irrigation Investment Scenario 45Comparison of Investment Scenarios Using the Historical

Variability Model 49Baseline Scenario 49Irrigation Investment Scenario 49Multipurpose Development 52

Water Resources and Growth 53Water Security and the Minimum Platform of Water

Infrastructure and Institutions 53The Challenge of Achieving Water Security 54Economic Structure and Resilience 55The Dynamics of Institutions and Infrastructure 56

Insights for Interventions 59Lessons 60Implications for Interventions 62

A Water Resources Assistance Strategy for the World Bank 63Ethiopia’s Water Sector Strategy (2001) 63World Bank Assistance in Ethiopia 64A Strategy for World Bank Support in Water Resources 66

Appendix 1. Technical Notes on the Model 71

Appendix 2. Water-Related Programs in Ethiopia 75

Appendix 3.World Bank Projects in Ethiopia 81

Endnotes 87

References 89

5

8

7

6

Contents v

Tables2.1. Mean Annual Rainfall and Coefficients of Variation in Selected

Drought-Prone and Non-Drought-Prone Areas (1980–2001) 72.2. Trends in Rainfall Variability 1970–2001 92.3. A Chronology of Ethiopian Drought and Famines Since 1895 102.4. Drought- and Disaster-Affected Populations 113.1. Population Distribution by Settlement Size 173.2. Access to Water Supply and Sanitation 183.3. Sectoral Shares of GDP (Percentages) 193.4. Road Passability 243.5. Percentage of Reported Defects in Paved Roads 263.6. Percentage of Reported Defects in Unpaved Roads 263.7. High-Cost Scenario Road Investment Targets to 2015 273.8. Ongoing and Planned Hydropower Projects 295.1. GDP Growth Rates Under Differing Investment Scenarios 455.2. Poverty Rates Under Differing Investment Scenarios 455.3. Agricultural Production and Irrigation Returns Under the

Combined Investment Scenario, Smoothed Rainfall Variant 475.4. Baseline Scenario (No Investments) With Historical Variability 495.5. Irrigation Investment Scenario, Historical Variability Model 505.6. GDP Growth Rates, Market Infrastructure Investment Scenario,

Historical Variability Model 515.7. Poverty Rates Under the Market Infrastructure

Investment Scenario 515.8. Impacts of the Market Infrastructure Investment Scenario,

Historical Variability Model 515.9. Impacts of the Combined Market Infrastructure and Irrigation

Investment Scenario, Historical Variability Model 528.1. Ongoing International Assistance to the Water Sector

in Ethiopia 658.2. World Bank Portfolio Distribution by Major Sectors, FY2005 66

FiguresES.1. Rainfall Variability: National, Tigray, and Arsi xiiiES.2. Rainfall Variation Around the Mean and GDP Growth xviES.3. Water Storage Per Capital in Selected Countries xviiES.4. Water Infrastructure xviiiES.5. Effects of Hydrology on GDP Growth xixES.6. Effects of Hydrology on Poverty Rates xxES.8. GDP Growth With Variable Rainfall and Different

Investment Scenarios xxiES.7. Growth Rates With Smoothed Rainfall and Different

Investment Scenarios xxi2.1 Surface Runoff from Ethiopia’s Major River Basins 62.2 Rainfall Variability: National, Tigray, and Arsi 83.1 Access to Water Supply: Millennium Development Goals 173.2 Rainfall Variation Around the Mean and GDP Growth 203.3 District Mean Annual Surplus Flow Versus District Mean

Percent of Water-Related Maintenance Cost 26

vi Contents

4.1 Water Storage Per Capital in Selected Countries 355.2 Effects of Hydrology on Poverty Rates 435.1 Effects of Hydrology on GDP Growth 435.4 Projected Poverty Rates Using Smoothed Rainfall and Variable Rainfall

Means 445.3 Projected Log GDP Using Smoothed Rainfall and

Variable Rainfall Means and Variable Rainfall Ensemble 445.6 GDP Growth With Variable Rainfall and Different

Investment Scenarios 485.5 Growth Rates With Smoothed Rainfall and Different

Investment Scenarios 486.1 Balancing and Sequencing Investments in Water

Infrastructure and Management 57

Box3.1. Weak Regulation of Competing Consumptive Uses in Harar 16

Maps2.1. Ethiopia Water Resources Country Assistance Strategy 42.2. Ethiopia Water Resources Country Assistance Strategy:

Major River Basins 53.1. Ethiopia Water Resources Country Assistance Strategy:

Hydropower and Transport Infrastructure 25

vii

CAS Country Assistance Strategy

CEM Country Economic Memorandum

CWRAS Country Water Resources Assistance Strategy

ERA Ethiopian Road Authority

GDP gross domestic product

GoE Government of Ethiopia

ha hectares

IDP Irrigation Development Program

MoWR Ministry of Water Resources

NBI Nile Basin Initiative

SDPRP Sustainable Development and Poverty Reduction Program

SNNP Southern Nations, Nationalities, and Peoples’ regional states

Acronyms andAbbreviations

ix

Many people contributed to the preparation of this strategy. Thecore team was led by Claudia Sadoff (ARD) and included CatherineRevels (EWDSA); Yitbarek Tessema (AFTU1); Mekuria Tafesse (con-sultant); Mulat Demeke (Addis Ababa University, consultant); PeterMcCornick (IWMI); Mark Rosegrant, Sherman Robinson, XinshenDiao, Ken Strzepek, Paul Block, and Charles Rogers (IFPRI); MarthaSolomon (Ministry of Water Resources, Ethiopia), and TadesseKuma (Ethiopia Development Research Institute). David Grey(AFTNL) provided advice and guidance to the core team through-out this work. Aziz Bouzaher (AFTS1) and Maher Abu-Taleb(MNSRE) contributed significantly as peer reviewers.

Guidance was provided by H.E. Minster Shiferaw Jarso (Ministryof Water Resources, Ethiopia), H.E. Minister Sufian Ahmed(Ministry of Finance and Economic Development), H.E. NewayGebreab (Senior Economic Advisor to the Prime Minister), IshacDiwan (Country Director, World Bank), Jaime Biderman (SectorManager, World Bank), and Jill Armstrong (Country Manager,World Bank.)

Important contributions were also made by IJsbrand de Jong,Madhur Gautam (AFTS2), Reynold Duncan (MNSIF), E. V.Jagannathan (SASAR), Andrew Macoun, Barbara Miller, AshokSubramanian (AFTNL), Inger Andersen (MNSRE), GebreselassieOkubagzhi (AFTH3), Gemechu Ayana (Ethiopia DevelopmentResearch Institute), and Demile Yismaw (Ministry of Finance andEconomic Development, Ethiopia). Editorial services were providedby Mary Paden (consultant) and administrative support by AijazAhmad (AFTU1).

Finally, the team would like to acknowledge the input of morethan 100 stakeholders from federal and regional governments, uni-versities, research institutes, development partners, the private sec-tor, civil society, and nongovernmental organizations, whoparticipated in consultations on various drafts of this report.

The team gratefully acknowledges the financial support of theBank-Netherlands Water Partnership Program in the preparation ofthis report.

Acknowledgments

xi

Ethiopia’s primary water resource management challenges are itsextreme hydrological variability and seasonality and the interna-tional nature of its most significant surface water resources.Variability is most obviously manifest in endemic, devastatingdroughts and floods. Less apparent is the broad range of impactsvariability and seasonality have on the Ethiopian economy, even ingood rainfall years.

This report looks at, and beyond, the management hydrologicalvariability to interventions aimed at decreasing the vulnerability ofthe economy to these shocks. It helps clarify linkages between thecountry’s economic performance and its water resources endow-ment and management. It then uses this analysis to recommendboth water resource strategies and economic and sectoral policiesthat will enhance growth and insulate the Ethiopian people andeconomy from the often devastating, economy-wide effects of watershocks. The objectives of this report are to:

• Investigate and describe the role of water in the economic performance of Ethiopia from a cross-sectoral, economy-wideperspective.

• Identify broad strategies for mitigating negative and enhanc-ing positive impacts of water on the economic performance ofEthiopia.

• Define an overarching Country Water Resources AssistanceStrategy for World Bank support to water management inEthiopia, which will provide an integrated framework for devel-oping and harmonizing strategies in water-related sectors.

This report finds that unmitigated hydrological variability cur-rently costs the economy more than one-third of its growth poten-tial. The very structure of the Ethiopian economy with its heavyreliance on rainfed subsistence agriculture makes it particularly vul-nerable to hydrological variability. Its current extremely low levelsof hydraulic infrastructure and limited water resources manage-ment capacity undermine attempts to manage variability. These cir-cumstances leave Ethiopia’s economic performance virtuallyhostage to its hydrology.

To de-link economic performance from rainfall and enable sus-tained growth and development, three strategic shifts appear neces-sary. First, major investments in water resources infrastructure,

Executive Summary

institutions, and capacity to manage flows anddevelop storage at all scales must be seen as aneconomy-wide development priority. Ethiopia’sgrowth will continue to be undermined until thecountry achieves water security by acquiring aminimum platform of infrastructure, institutions,and capacity to manage its water resources.Investment on this scale may initially show lowreturns—much like a road investment, whichshows little return until it joins two cities—butgrowth will continue to be hindered until watersecurity is achieved. Particularly in the Nile RiverBasin, international cooperation on shared waterswill be crucial to achieving sustainable water secu-rity without international tensions.

Second, investments in water infrastructureshould be multipurpose in nature and should bemade in combination with the market infrastruc-ture investments and related reforms needed tofully leverage their growth potential. Particularlygiven the vast scale of investments needed inwater and transport infrastructure, it may be nec-essary to adopt a "growth pole" strategy (WorldBank 2005a) ensuring the resilience of the primaryengines of economic growth. The combination ofwater, irrigation, hydropower, roads and othermarket infrastructure investments should producedramatic synergies, and provide the incentivesand opportunity for farmers to shift out of subsis-tence agriculture into surplus/commercial agri-culture and nonagricultural activities. While it iscommon practice to spread different types ofinvestment across regions, such isolated invest-ments will not have the same growth impact ascoordinated, co-located investments.

Finally, in addition to managing the variabilityof water resources, the vulnerability of theEthiopian economy to hydrology must also bemanaged. Economy-wide, policies and investmentdecisions need to shift the pattern of developmentand the structure of the economy away from thecountry’s heavy reliance on rainfed agriculture tomake it less vulnerable to its hydrologic legacy.

THE CHALLENGE OF ETHIOPIA’S HYDRO-ECONOMY

Ethiopia’s water resources

Water has always played a central role inEthiopian society. It is an input, to a greater orlesser extent, to almost all production. It is also a

force for destruction. In Ethiopia, as in all societies,there has always been a struggle to reduce thedestructive impacts of water and increase its pro-ductive impacts. This struggle has intensified overthe past century or so as the population has growndramatically. Today, Ethiopia’s development isseriously constrained by a complex waterresources legacy and a lack of access to, and man-agement of, these water resources.

Water legacies: Natural and historical

Ethiopia’s water resources can be characterized bytwo principal features, a natural legacy and an his-torical legacy. The natural legacy is one of veryhigh, and apparently intensifying, hydrologicalvariability, coupled with marked rainfall seasonal-ity (see Figure ES.1). At first sight, Ethiopia’s waterresources endowment appears generous, with amean total surface water flow of about 122 billioncubic meters per year. The country has severalmajor rivers and lakes, and significant groundwa-ter resources and annual rainfall. However, rain-fall across much of the country is both highlyseasonal—with most of the rain falling in a single,short season—and exceptionally variable andunpredictable, both in time and space.

With poorly protected watersheds and almostno investment in water storage, a consequence ofthis hydrological variability is endemic and unpre-dictable drought and flood. The historical legacy isone of several international rivers, of which theNile is the most important. Ethiopia provides 85percent of the natural Nile River flow into Egypt.In the Nile River Basin, tensions have long beenhigh, and without the bold affirmative action forcooperation that Ethiopia is currently adopting,these tensions would likely grow as demand forwater grows with populations and economies inthe basin.

Water resources challenges: Land degradationand sedimentation

Hydrology plays a pivotal role in erosion, landdegradation, reduced soil fertility and productiv-ity, and siltation. Ethiopia’s high intensity stormscause significant erosion. The erosive effects ofrainfall are significantly augmented by Ethiopia’smountainous terrain, severe deforestation, andtraditional agricultural practices of cultivatingsteep slopes without protective measures. The lossof forest cover, in turn, is generally associated withgreater hydrological variability. Associated sedi-

xii Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Executive Summary xiii

mentation compromises productivity and short-ens the lifespan of water infrastructure for riverregulation, municipal water supplies, agriculture,and hydropower generation. The reduced regula-tion capacity also increases flood risks for down-stream communities, which poses a particular riskto the poor, who tend to live in the most vulnera-ble locations.

Water resources challenges: Droughts and floods

Droughts and floods are endemic, with significantevents every 3–5 years. Droughts destroy water-sheds, farmlands, and pastures, contributing toland degradation and causing crops to fail andlivestock to perish. During the 1984–5 drought, forexample, GDP declined by 9.7 percent, agricultureoutput declined 21 percent, and gross domesticsavings declined 58.6 percent. Drought can alsoseverely undermine hydropower generation,Ethiopia’s main source of electricity.

If rains fail, or simply come too early or too late,the entire agricultural cycle can be disrupted,because there is inadequate water storage capacityto smooth and schedule water delivery. This sensi-tivity of production to seasonal timing and the spa-tial distribution of rainfall means that hydrology

can constrain growth in years that appear “good”or “bad” according to average national data.Flooding meanwhile causes significant damage tosettlements and infrastructure, and the inundationand water-logging of productive land underminesagriculture by delaying planting, reducing yields,and compromising the quality of crops, especially ifthe rains occur around harvest time.

Water resources, water supply, and sanitation

The Government of Ethiopia’s (GoE) Comprehen-sive and Integrated Water Resources ManagementPolicy (endorsed in 1999) states that the highestpriority use of water is human and livestock con-sumption, which amounts to less than 1 percent oftotal water usage. Major efforts are being made tomeet the Millennium Development Goals(MDGs). Nevertheless, only a minority ofEthiopians, some 25 percent, have access topotable water services and 15 percent to improvedsanitation. These circumstances are particularlyonerous for women and girls who must fetchwater and then care for those who become ill fromwaterborne diseases. In addition to low coverageratios, water supply and sanitation in both urbanand rural communities is characterized by lowservice levels and lack of sustainability. The pri-

0

20

40

60

80

100

120

140

160

180

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Year

0

20

40

60

80

100

120

140

160

180

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Per

cen

tage

National rainfall variability(RFt/mean*100)

Tigray’s rainfall variability(RFt/mean*100)

Arsi’s rainfall variability(RFt/mean*100)

Figure ES.1. Rainfall Variability: National, Tigray, and Arsi

Source: Demeke (2004).

mary source of raw water for urban and ruralwater supply is groundwater. Where surfacewater is used, siltation in rivers and reservoirs hassubstantially increased treatment and mainte-nance costs. The extension and improvement ofwater supply and sanitation services is a necessaryprecondition for growth.

Water resources and the environment

The World Bank’s Country EnvironmentalAnalysis for Ethiopia concludes that the country’sprincipal environmental challenges involve com-plex cross-sectoral linkages. Two of the key envi-ronment-development linkages identified by thatstudy relate directly to the challenges posed byEthiopia’s hydrology. They are the lack of inte-grated water resources management and the landdegradation–food insecurity–energy access–liveli-hood nexus. The latter includes unsustainableagricultural land management practices andheavy reliance on biomass energy. The use of bio-mass spurs deforestation and erosion and con-tributes to a significant environmental healthproblem: exposure to smoke and indoor air pollu-tion, which causes elevated under age five mortal-ity and a high incidence of respiratory diseases,mainly in women and children. Using cropresidues and dung as fuel, rather than returningthis organic matter to the soil, causes a decline insoil fertility and deterioration in soil structure.

The degraded land is also more prone to ero-sion, leading to loss of fertility in the topsoil and toa reduction in soil depth, both of which can havean adverse effect on crop yields. The consequencesof both deforestation and degraded soil structureinclude less infiltration of rainfall, which dimin-ishes groundwater recharge; more runoff, whichcontributes to erosion and siltation; and reducedwater storage capacity in the soil, which makescrops less able to withstand drought.

Water resources and agriculture

The dominant agricultural system in Ethiopia issmall-holder production of cereals under rainfedconditions, leaving the population and economyextremely vulnerable to hydrology. Ethiopia hasthe largest livestock population in Africa.Livestock is an integral part of nearly all the mixedtype highland farming systems and the principalstore of farmer wealth. Agriculture employs 80percent of the population and accounts for almost

50 percent of the gross domestic product (GDP).Small-holder farmers, generally with less than 1hectare of land, account for about 95 percent of theagricultural output. In times of good weather,roughly 75–80 percent of the annual output is con-sumed at the household level.

It is striking that less than 5 percent of the esti-mated potential 3.7 million hectares of irrigableland in Ethiopia has been developed. Variations inrainfall and weather conditions cause consider-able volatility in agriculture growth. Hydrologicalvariability affects both high- and low-potentialareas of the country, although the impact is muchgreater in the latter. Weather variability can bothreduce yields and lead to land degradation. Yieldstypically fall if there is too little rain, but they canalso be damaged by excessive rain and flooding.Declining agricultural productivity, in turn,inhibits investments in land management andencourages the extension of agricultural landsrather than intensification of production. Thispractice can lead to further land degradation, clos-ing a vicious circle that leaves the landscapeincreasingly vulnerable to drought and rainfall-related erosion.

Because of agriculture’s dominant share of theeconomy, fluctuations in agriculture are reflectedin the GDP. These fluctuations undermined thecountry’s ability to sustain per capita growth inthe 1980s and 1990s.

Water resources and hydropower

Ethiopia has considerable hydropower potential,about 30,000 megawatts. Presently, 95 percent ofnational energy consumption is derived from fuel-wood, dung, crop residues, and human and ani-mal power. The remaining 5 percent is fromelectricity, 90 percent of which is generated byhydropower. The current dependable powercapacity in Ethiopia is 731 megawatts. Only twopercent of the country’s economically feasiblehydroelectric potential has been developed.

Power interruption is common in years ofsevere drought when water shortages disrupthydroelectric power generation. The 2002–3drought caused power interruptions that lastedfor about four months with a one-day-per-weekcomplete interruption throughout the country. Aone-day interruption was estimated to result in aloss of 10–15 percent of the GDP for the day(EEPRI 2003).

xiv Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Executive Summary xv

Because only about 15 percent of the populationhas access to electricity, there is great reliance onbiomass and fuelwood. This reliance intensifiesdeforestation and soil degradation, and increasesthe burden on women and children who tradition-ally gather and transport fuelwood and otherorganic fuels and also suffer most from indoor airpollution. While electricity will not be a feasiblesubstitute for all biomass fuel uses in the near term,the availability of hydropower will begin to shiftenergy use away from biomass and strengthenenvironmental sustainability. The GoE’s growthand poverty reduction strategy includes the aggres-sive expansion of access to affordable electricity,which will also reduce the demand for biomass andthereby protect watersheds.

Water resources, transportation, and market infrastructure

Current limited access to transportation and mar-kets undermines incentives for surplus agricul-tural production and reinforces the highlyvulnerable subsistence-oriented structure of theeconomy. Road transport accounts for roughly 95percent of the country’s passenger and freight traf-fic and provides the only form of access to mostrural communities. On average, Ethiopian farmerslive 10 kilometers from the nearest paved road and17 kilometers from the nearest commercial trans-port (Dercon and Chistiaensen 2005).

A notable characteristic of the network is thatmost all-weather roads radiate from the capital,Addis Ababa, to major towns. Direct linksbetween regions are rare, which discourages inter-regional trade. The cost of building and maintain-ing roads is high because of rugged topographyand torrential tropical rains. Because of this highcost, the systems expand slowly and many newroads are left unpaved.

Today 90 percent of Ethiopia’s roads are dry-weather roads that cannot be used effectively dur-ing the four-month-long wet season. The relianceof the economy on this small network of mostlydry-weather roads makes commerce highly vul-nerable to floods and heavy rainfall. Over theperiod 2000–2 almost 60 percent of the reporteddefects in unpaved roads and 35 percent of defectsin paved roads were directly related to water.These figures do not include rain-induced land-slides or infrastructure washouts. Moreover aclear correlation is seen between hydrological

variability and the percentage of maintenancecosts attributable to water damage. When it rainsin remote areas, farmers can produce crops butoften cannot get them to markets.

Water resources and urbanization

Ethiopia is less urbanized than most developingcountries in Africa and the rest of the world, with acurrent level of urbanization at just 12 percent. Thelargest city, Addis Ababa, is 12 times larger than thesecond largest city. Addis Ababa alone providesabout one-fifth of the overall GDP of the countryand 52 percent of the modern sector value added.Urban productivity, or value added per urbaninhabitant, is nearly four times that of rural areas,and Addis Ababa is 2.5 times more productive thanthe average of other cities and towns. These greatdisparities between urban and rural productivitywere recognized explicitly in the GoE’s AbbayRiver Basin master plan, which found that urban-ization contributes to economic diversification andprovides “the only” viable base for long-termgrowth in average incomes. By withdrawing laborfrom the agricultural sector, urbanization also con-tributes to the sustainability of agriculture and theconservation of the resource base.

As noted above, market towns are rarely inter-connected laterally, and where roads exist, pass-ability is uncertain much of the year. Thus, theirgrowth is inhibited. The dispersed settlement pat-tern leaves a greater percentage of the populationmore directly dependent on natural resource-related livelihoods and therefore more vulnerableto hydrological variability. The rural settlementpattern also compounds the challenge of mitigat-ing hydrological variability because dispersedpopulations cannot be provided drought andflood mitigation services (that is, water storageand river regulation) as cost effectively as moreconcentrated populations.

Hydrology and economy-wide performance

Ethiopia’s extreme hydrological variability isechoed in its economic performance. The vastmajority (80 percent) of Ethiopia’s population sub-sists on rainfed agriculture, and thus their welfareand economic productivity are linked to thevolatile rains. The correlation between rainfall andoverall GDP is strong, as can be seen in FigureES.2. The impact of rainfall variability, uncheckedby either physical infrastructure or strong man-

agement practices, can be felt not only on agricul-tural output but on the environment, electricity(which is 90 percent hydropower), manufacturedgoods, incomes and consumption, and prices.Poor market access and costly transportation serv-ices have created a highly fragmented economy inwhich average grain prices differ 30–70 percentacross regions. This fragmentation reflects forgoneopportunities to make both buyers and sellers bet-ter off. It also amplifies drought and flood shocksby hampering trade between surplus and deficitregions in the country.

Dynamic effects and incentives for growth

The fragility of the Ethiopian landscape and thedependence of the Ethiopian population and econ-omy on rainfed agriculture create a web of effectsthat result—at least in part—from the country’sextreme hydrological variability. This variabilityhas economy-wide impacts in good and bad years.Rainfall variability has direct impacts on the land-scape, agricultural output, transport passability,and water-intensive industry and power produc-tion. These impacts are generally unmitigatedbecause there is little hydraulic infrastructure toregulate or store water. They are often amplifiedby the fragmented nature of the Ethiopian econ-omy, which does not facilitate trade between

affected and non-affected regions of the country,and transmitted through input, price, and incomeeffects onto the broader economy.

Many of these dynamics conspire against struc-tural change in the economy. Endemic drought,fragmented and inadequate infrastructure, andexpectations that food aid shipments will disrupttransportation and tamp down food price spikesduring shortages, all promote a risk-averse subsis-tence economy and undercut incentives for farm-ers to produce surplus, marketable crops.

Classic water management responses

Ethiopia cannot fully mitigate the impact ofhydrological variability through water infrastruc-ture and management investments alone. It is esti-mated that artificial reservoir storage in Ethiopia isabout 43 cubic meters per capita in contrast to 750cubic meters per capita in South Africa (see FigureES.3). (Note that reservoir storage in NorthAmerican is 6,150 cubic meters per capita.).

If we take the storage in South Africa as a crudebenchmark of water security, it can be shown usingconventional coefficients that Ethiopia would needto invest five times its annual GDP (US$35 billion)just to achieve a similar infrastructure stock, withsignificant additional investments needed for theinstitutions and capacity required to complement

xvi Ethiopia: Managing Water Resources to Maximize Sustainable Growth

-80

-60

-40

-20

0

20

40

60

80

1982

1985

1986

1987

1988

1989

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Year

Per

cen

tRai

nfa

llV

aria

tion

arou

nd

the

Mea

n

rainfall variation around the mean

GDP growth

1997

1999 -

-

-

-

-

30

25

20

15

-10

5

0

5

10

15

20

25

Per

cen

tCh

ange

inG

DP

Gro

wth

1984

1983

1990

Figure ES.2. Rainfall Variation Around the Mean and GDP Growth

Source: Compiled from SIMA and African Rainfall and Temperature Evaluation System data.

Executive Summary xvii

those investment. This order of magnitude calcula-tion only serves to demonstrate that strategiesfocused purely on water management and infra-structure responses are not affordable.

Clearly, it will be necessary to look beyond whatis traditionally considered water resources manage-ment in order to curtail the negative effects ofhydrological variability on the performance of theEthiopian economy, by identifying both interven-tions aimed at managing hydrological variability,and interventions aimed at decreasing the vulnera-bility of the economy to these shocks.

Incentives for building resilience

Efforts to address variability start with the varietyof means traditionally associated with waterresources management, with all interventionsdesigned and implemented cognizant of environ-mental and social challenges and priorities. Asdescribed previously, these interventions includewatershed management, with programs to pro-vide sustainable livelihoods, and investments inmultipurpose hydraulic infrastructure, whichcould potentially be designed to includehydropower production, irrigation systems, andstorage adequate to mitigate both drought andfloods. These investments are envisaged to pro-

vide the security and incentives necessary toprompt private sector investment in inputs thatwill enhance productivity. This is often describedas an “input impact” on patterns of investment forsurplus production. Figure ES.4 shows the currentlow level of Ethiopia’s water infrastructure includ-ing hydro and irrigation development and accessfor potable water.

For longer-term growth, investments must alsobe considered to promote less water-dependent,more resilient, livelihoods. Current policies focuson food aid and other palliatives to ease theimpact of drought on the population and the econ-omy, but much of the problem lies in the verystructure of the economy, which leaves its peoplevulnerable to hydrological variability. The major-ity of the population depends directly upon rain-fall for subsistence agriculture, and the broadereconomy is built on domestic demand that is asunpredictable as the rains. This is compounded byinadequate infrastructure, fragmented nationalmarkets, and the current policy environment,which all prove to be disincentives for investmentsin more weather-resilient activities such as irri-gated agriculture, manufacturing, and services.

Proactive interventions are needed outside thewater sector to enhance the economy’s resilience

Water availability versus storage

0

200

400

600

800

1000

0 1,000 2,000 3,000 4,000 5,000

storage/capita (m3)

wit

hd

raw

als

per

cap

ita

(m3)

Ethiopia

S. Africa

SpainAustralia

Ethiopia

S. Africa

SpainAustralia

Figure ES.3. Water Storage Per Capital in Selected Countries

Source: Compiled from SIMA and KOID databases.

to the impacts of hydrological variability. Thedevelopment of resilient growth poles wouldappear to be a promising strategy in this regard.The scale of variability is simply too great to fullymitigate by structural or management interven-tions. Unless the structure of the economychanges, stable economic progress will continue tobe both intensely disrupted by drought and floodand quietly undermined by risk aversion.

Economywide Impacts of Ethiopia’s WaterResources: A Hydro-Economic Model

Based on these insights, a model was constructedto quantify the economy-wide impacts ofEthiopia’s water resources endowment, variabil-ity, and management. An economy-wide multi-market model was modified explicitly toincorporate hydrology into growth and povertyprojections.1 The model is dynamic, running for aperiod long enough to explore the implications ofvariability over time. The following key character-istics of the Ethiopian economy and hydrologywere drawn from the discussion above and incor-porated into the model:

• Highly variable rainfall and endemicdroughts and floods

• Fragile and degrading landscapes• Low levels of infrastructure investment to

mitigate hydrologic variability• Fragmentation of the economy, particularly

regional segmentation in agricultural mar-kets

• High transportation and marketing costs

The model was run with three variations. Allmodel variations are regional, multisectoral, mul-timarket models that incorporate hydrologywithin the agricultural production function. Theydiffer only with regard to their assumptions ofrainfall variability. Specifically:

• Smoothed rainfall: This variation assumes1995–2002 average rainfall in all years.Smoothed rainfall is the underlying assump-tion in virtually all economy-wide modeling.

• Stylized drought: This variation assumes1995–2002 average rainfall except for a styl-ized, two-year drought of average severity. Itcaptures the isolated “shock” of a singledrought.

xviii Ethiopia: Managing Water Resources to Maximize Sustainable Growth

% Developed or Served

HydroDeveloped

IrrigationDeveloped

Access toPotable Water

100

80

60

40

20

0

Figure ES.4. Water Infrastructure

Source: Authors.

Executive Summary xix

• Historical variability: This variation is a sto-chastic extension of the model that more fullyreflects Ethiopia’s historical levels of rainfallvariability, and, in addition, captures the neg-ative impacts of excessive rains on the agri-cultural and non-agricultural sectors.

Growth and Poverty Costs of Drought and Variability

The model shows that hydrological variabilitycosts the Ethiopian economy 38 percent of itspotential growth rate and causes a 25 percentincrease in poverty rates, clearly demonstratingthe extraordinary impact of drought, and particu-larly variability, on the Ethiopian economy.

A single drought event in a 12-year period—avery conservative estimate for Ethiopia—willdecrease average GDP growth rates 7–10 percent.If historical levels of variability and the partialimpacts of floods are incorporated, GDP growthrates fall 20–43 percent. Figure ES.5 shows the dif-ference in GDP growth rate projections when rain-fall is modeled as a smoothed average, when asingle drought event is incorporated, and whenhistorical levels of variability are assumed.

In terms of poverty, the impact of a singledrought will increase poverty rates 12–14 percent,causing an additional 5 million Ethiopians to be

living in poverty by the year 2015 (relative to thenondrought case), 17 million more than today.With hydrological variability, projected povertyrates rise 25–35 percent and project 51 million peo-ple living in poverty, 22 million more than today.Figure ES.6 shows the difference in the number ofpeople projected to live in poverty when rainfall ismodeled as a smoothed average, when a singledrought event is incorporated, and when histori-cal levels of variability are assumed.

It is important to note that drought shocks,which tend to be a primary focus for weather riskanalysis in Ethiopia, represent just part of thestory. The historical variability model, while stillquite conservative, better reflects the real hydro-logical impacts endured in Ethiopia. It capturesyear-on-year stochastic variations, and some of theimpacts of floods as well as droughts. These mod-els clearly suggest that variability, rather than anarrow focus on drought, must be the centralwater resources challenge for development inEthiopia.

Investment Scenarios

Figures ES.7 and ES.8 summarize the relative agri-cultural and nonagricultural GDP growth rates byinvestment scenarios, using the smoothed rainfalland the historical variability models. Overall GDP

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

GDP Growth rate Ag GDP Non-Ag GDP

An

nu

alR

ates

Smoothed DroughtVariability

Figure ES.5. Effects of Hydrology on GDP Growth

Source: Authors.

growth rates are lower in the model with variabil-ity because the enormous costs of hydrology arenot factored into the smoothed rainfall model.

The smoothed rainfall model shows relativelylow agricultural GDP growth rates and relativelylow returns to irrigation investments. This modelsmoothes, or averages, rainfall extremes, but it isat these extremes—during periods of drought andflood—that the returns to irrigation and drainageare greatest. Smoothed models therefore fail tocapture the extent of benefits offered by irrigationin countries such as Ethiopia that are subject togreat variability.

Under the model that incorporates historicallevels of rainfall variability, overall growth ratesare comparable for scenarios that allow eitherinvestments in irrigation or in infrastructure. Butthe two scenarios give rise to very different rela-tive growth rates between agricultural and non-agricultural GDP. The infrastructure investmentscenario gives rise to much greater non-agricul-tural benefits, which would have a markedly dif-ferent impact on the structure of the economy overtime, possibly reducing its vulnerability to hydro-logical variability and shocks.

Apparent synergies between irrigation andinfrastructure investments are much more pro-nounced under the historical variability model.Under both the smoothed rainfall model and the

stylized drought model, the lion’s share of benefitsappear to derive from market infrastructureinvestments, with irrigation investments onlyadding marginally to total growth and povertygains. Under the historical variability model, how-ever, the gains are virtually additive. This findingreflects the greater benefits of irrigation underhighly variable climatic conditions, and suggeststhat the complementarity of combined invest-ments is particularly pronounced under such con-ditions. It also suggests that investments in roadsand market infrastructure are essential for leverag-ing returns to irrigation investments, although thereverse relationship (that investments in irrigationleverage returns on market infrastructure) maynot be strong.

The model clearly calls for coordinated invest-ment in irrigation and market infrastructure tomanage the impact of hydrological variability onthe Ethiopian economy. Such coordinated invest-ment should exploit opportunities for multipur-pose infrastructure. River regulation and storageinfrastructure can be designed at all scales to meetseveral objectives: delivery of irrigation and munic-ipal water, generation of hydropower and the pub-lic good benefits of drought and flood mitigation,and protection of roads and other infrastructure.

The model does not explicitly incorporatehydropower. Generally, there are tradeoffs

xx Ethiopia: Managing Water Resources to Maximize Sustainable Growth

20032015

Smoothed

Drought

Variability

29.12

51.07

29.12

45.67

29.12 40.79

0102030405060

Figure ES.6. Effects of Hydrology on Poverty Rates

Source: Authors.

Executive Summary xxi

between the use of water for hydropower or forupland irrigation as the consumptive use of waterabove a hydropower plant means less power willbe generated. Choices may need to be made. Thereappear to be opportunities in highland Ethiopia,however, for substantial irrigation development

without requiring a significant tradeoff in terms ofhydropower generation. For example, in largeareas around Lake Tana, river regulation anddrainage will provide the incremental waterrequired for irrigation. In addition, there appear tobe many cases where small to medium

GDP Growth w/Smoothed Rainfall

0

1

2

3

4

5

6

7

Base Irrig. Roads Irri&Roads

Ann

ualG

row

th

GDP growth rate

Ag GDP growth rate

NAg GDP growth rate

Figure ES.7. Growth Rates With Smoothed Rainfall and Different Investment Scenarios

Source: Authors.

GDP Growth w/Variable Rainfall

0

0.5

1

1.5

2

2.5

3

3.5

4

Base Irrig. Roads Irri&Roads

Ann

ualG

row

th

GDP growth rate

Ag GDP growth rate

NAg GDP growth rate

Figure ES.8. GDP Growth With Variable Rainfall and Different Investment Scenarios

Source: Authors.

hydropower coupled with irrigation can be eco-nomic but where single-purpose investmentsmight not be (for example, on the rivers that feedLake Tana). Moreover, the larger storage capacityjustified by combined investments would provideeven greater public benefits. Hydropower devel-opment at all scales is potentially an integral partof both irrigation and market infrastructure devel-opment, and abundant energy is clearly needed todiversify the economy.

The GoE is currently undertaking major devel-opment programs in irrigation and market infra-structure, particularly roads and power. It isdifficult to choose a national priority between thetwo types of investment from the results of themodel. To hasten a structural shift toward non-agricultural activities, the model suggests a some-what disproportionate investment in marketinfrastructure, but this will not necessarily be thepriority in all locations. Specific investment deci-sions will need to be made cognizant of localhydrological, environmental, economic and socialconditions, and the existing stock of both irriga-tion and market infrastructure.

Water and Growth in Ethiopia: Achieving Water Security

Water security and the minimum platform ofwater infrastructure and institutions

This analysis of Ethiopia’s water resources legacyand challenges deepens our understanding of theconcept of water security (Grey and Sadoff 2005).Investments in bulk water resources and water forfood, energy, industry, and navigation and in asso-ciated institutions are initially made by all societiesin search of some implicit level of water security.2

Water security is defined here as the reliable avail-ability of an acceptable quantity and quality ofwater for production, livelihoods, and health, cou-pled with an acceptable level of risk of high socialand economic impacts of unpredictable waterevents, including the extremes of drought andflood. Water security is achieved when the overallimpact of water enhances economic growth ratherthan undermines it. Water security is a dynamicstate: different in different parts of the world,reflecting geographic, social, epidemiological, eco-nomic, and political factors, and changing overtime as these factors shift with development.

We postulate that there is a basic level of watersecurity that incorporates the idea of a minimum

platform of water infrastructure and institutions.Below this minimum platform, society and theeconomy is unacceptably impacted by water (byany mix of water shocks and/or unreliable wateravailability for production or livelihoods). The tip-ping point in achieving water security is the acqui-sition of a minimum platform of managementcapacity and infrastructure investment. The mini-mum platform has not been achieved in Ethiopia.The scale of social and related economic impacts issuch that Ethiopia’s social fabric is significantlyaffected and economic growth cannot be reliablyand predictably managed.

Implications for investment

In Ethiopia significant public investment will beneeded to provide adequate security for privateinvestment to follow and growth to ensue. Thisthinking follows the model of public investment ina road, which will have little return until it joinstwo cities, but then will attract private investmentin transportation services and economic growthfor both cities. If this is the case, it has importantimplications for how we assess the cost-effective-ness of early investments in water resources infra-structure in Ethiopia. But if the hypothesis holdstrue, growth in Ethiopia will be slowed until suchinvestments are made.

Institutions/infrastructure balance

In all countries, development of water resourcesrequires investments in both institutions andinfrastructure, but when stocks of hydraulic infra-structure are low, investments in infrastructure arelikely to receive priority. Investment in manage-ment capacity and institutions becomes increas-ingly important as larger and more sophisticatedinfrastructure stocks are built. Without the infra-structure to store and deliver water and manageflows, there is neither the need nor the incentivefor sophisticated management structures andpractices.

While developed countries are appropriatelyfocused on the implementation of integratedwater resources management, Ethiopia and otherdeveloping countries may do better to adopt aprincipled and pragmatic approach to manage-ment while putting greater emphasis on concur-rent infrastructure investments.3 Failure tounderstand the issue of balance and sequencingwithin the context of specific country circum-stances may lead to poor investment choices.

xxii Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Executive Summary xxiii

Transboundary institutions/infrastructure case

International river management is of growingimportance in a world with 260 internationalrivers and is a particular challenge for Ethiopia,whose most significant water resources are shared.Although a basic premise of water resources man-agement is that river basins are best developedand managed as an integrated whole, integratedmanagement is always legally and politically com-plex, due to the challenges of allocation betweenusers and between uses.

The management and development of interna-tional rivers is particularly challenging becausethere is no apex authority through which differ-ences can be resolved. Because the complexity ofriparian relations is an obstacle to the develop-ment of the full potential that international riversembody for growth and poverty alleviation,nations often develop river segments within theirown territories, settling for second or third bestinvestments. In extreme cases, tensions over inter-national rivers can effectively halt their manage-ment and development.

On the other hand, the need for river infra-structure is often a major driver in reachingagreements where river flow—water quantity—is the issue. This is the case beginning to bedemonstrated by Ethiopia and its neighboringriparian states on the Nile. The Nile BasinInitiative (NBI) offers considerable potential formajor cooperative development of the Nile River,including large-scale hydropower and irrigationdevelopment located in Ethiopia. In addition,opportunities for regional cooperation and inte-gration in a range of activities “beyond the river”have arisen as a consequence of strengthenedrelations built on the NBI.

Water, poverty, and wealth

In Ethiopia the centrality of water is clear. With lit-tle water resources infrastructure, relatively weakmanagement institutions and capacity, extremehydrological variability and seasonality, and ahighly vulnerable economy, Ethiopia faces anenormous challenge in building the minimumplatform of water infrastructure and managementcapacity needed to achieve water security. Majorpublic investment in water resources infrastruc-ture and institutions will be needed, but untilwater security is achieved, growth will continue tobe severely constrained.

The GoE clearly recognizes the scale and com-plexity of the water challenge and is undertakinga wide range of programs in response. Many ofthese programs can be categorized as nationalwater resource management programs. Otherspromote international cooperation (particularlyon the Nile), and still others seek to mitigate theimpact of hydrologic variability on the Ethiopianeconomy by looking beyond the traditional watersector to encourage alternative livelihoods that aremore resilient to water shocks.

Lessons

• Ethiopia’s growth will be undermined untilit attains water security, and it will requirevery large investment to achieve a nation-wide minimum platform of water infrastruc-ture and management capacity. Unmitigatedhydrological variability increases povertyrates by about 25 percent and costs theEthiopian economy about 40 percent of itsgrowth potential, leaving growth rateshostage to hydrology. Managing this vari-ability will be a prerequisite to sustainedgrowth and development and should be apriority for public investment.

• Economy-wide models that fail to incorpo-rate hydrological variability will be poor pre-dictors of growth and poverty in Ethiopia.Rainfall variability is shown to be highly rel-evant and negatively correlated to economicperformance; that is, the greater the level ofvariability modeled into the system, theworse the growth and poverty outcomes.

• Returns to irrigation and drainage invest-ments are systematically underestimated inmodels that fail to capture hydrology,because returns to irrigation and drainageare highest in times of drought and flood. InEthiopia, modeling with hydrological vari-ability doubles returns to irrigation anddrainage.

• Current incentives trap farmers in highly vulnerable rainfed subsistence agriculture.Achieving hydrological resilience willrequire structural change in the economy.Growth cannot be managed if 80 percent ofthe population and 50 percent of the econ-omy is essentially dependent upon rainfall.Bank-supported investments and interven-tions in Ethiopia should be analyzed to deter-

mine whether they provide opportunitiesand incentives for structural changes thatwould make the economy less vulnerable tohydrological variability.

• Multipurpose hydraulic/hydropower infra-structure can enhance resilience both bydirectly regulating flows and by promotingstructural shifts in the economy. Investmentscan be designed that will directly mitigatevariability by providing over-year storage fordrought and flood mitigation and irrigationto address both annual variability and unpre-dictable seasonality. Investments in powermay also help non-agricultural sectors of theeconomy by promoting a structural shift inthe economy toward greater hydrologicalresilience.

• Parallel investments in irrigation and marketinfrastructure can generate synergies forgrowth and poverty reduction, becausetogether they will provide significant incen-tives for increased farmer investments inagricultural inputs as well as incentives andopportunities outside of agriculture. Themodel suggests there are greater synergiesfrom colocated investment as hydrologicalvariability increases, and that synergies areparticularly strong for leveraging returns toirrigation investments.

• Investment is needed in water supply andsanitation to support growth and welfare.Where appropriate, urban water supplyplanning should be accompanied by invest-ments in watershed management, bulk waterinfrastructure, and river regulation infra-structure.

• Water resources management capacity andinstitutions must be strengthened. Investmentin research and education in particular will becrucial in developing the capacity needed todesign effective and appropriate waterresource management interventions, water-shed management, irrigation schemes, andwater and sanitation services. An increasedfocus on groundwater should also be seen as apriority in this regard.

• Solutions to water resources challenges needto be sought outside the “water sector,” witha greater emphasis on alternative livelihoodsto decrease the share of the population most vulnerable to hydrological shocks.Investment in market infrastructure and

hydropower development would both helpmitigate the direct impacts of drought and atthe same time coax the economy toward amore resilient structure.

• A focus on all-weather roads and regionalinterconnection is essential in creating incen-tives for the production of surplus agriculture.Road connections should emphasize linksbetween complementary regions to enablenational trade that will lessen the regionalfood price shocks of drought and floods, thusreducing dependency on food aid.

• Alternatives should be sought to the practiceof storing wealth in livestock, whichincreases rural economic vulnerability toweather shocks. “Drought proof” stores ofwealth and related financial tools such ascredit and weather insurance should beexplored to provide asset stability duringdrought episodes.

• Given the scale of investments needed inwater and transport infrastructure, it may benecessary to adopt a “growth pole” strategy,4

ensuring the resilience of the primaryengines of economic growth as a priority.This resilience is needed to ensure the sus-tained growth that will allow water securityto be achieved nationwide.

A Strategy for World Bank Support

The analysis in this report provides a roadmap fora more coherent approach to Bank support forEthiopia’s water resources challenges. To enhanceEthiopia’s resilience to drought and hydrologicalvariability, and to promote more dynamic and sus-tained growth, there are several areas that standout as priorities and comparative advantages forBank support. They include:

• Multipurpose hydraulic infrastructure develop-ment: Multipurpose dam development couldprovide an excellent platform for addressingmany of the vulnerabilities discussed in thispaper and shifting the structure of the econ-omy toward a more water-resilient path. Thefar-reaching potential benefits of such devel-opment, and the unique qualifications of theBank to support these investments, stronglyrecommends multipurpose hydraulic infra-structure development, with an emphasis onhydropower generation and regional inter-

xxiv Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Executive Summary xxv

connection, as a first priority for future Bankassistance in water resource management.

• Water supply and sanitation: Given the GoE’shigh priority for water supply and sanitation,and the Bank’s ongoing successful engage-ment in the sector, the Bank should continueits support and increase its efforts to attractand coordinate with additional financiers tohelp Ethiopia meet the water supply and san-itation MDGs.

• Watershed management: The Bank shouldundertake targeted watershed managementprojects in conjunction with any dam devel-opment and road-building projects it sup-ports. Most immediately, priority supportshould be given to the cooperative watershedmanagement project now under considera-tion within the framework of the NBI. The

Bank might also have a comparative advan-tage in supporting projects that require facil-itated cooperation in other internationalwatersheds or those that are eligible for car-bon finance.5

• Transport, market infrastructure, and private sector development: Investments in fully func-tioning market cities (growth poles) withregional infrastructure and communicationsinterconnections are potentially powerfulinvestments for managing hydrological riskin Ethiopia and could be very effective inproviding alternatives and incentives to shiftvulnerable populations toward moreresilient livelihoods. The Bank should there-fore support the development of growthpoles with adequate hydraulic infrastructure,all-weather roads, and market infrastructure.

1

Water has always played a central role in Ethiopian society. It is aninput, to a greater or lesser extent, to almost all production. It is alsoa force for destruction. In Ethiopia, as in all societies, there has alwaysbeen a struggle to reduce the destructive impacts of water and in-crease its productive impacts. This struggle has intensified over thepast century as the population has grown dramatically. Today,Ethiopia’s development is seriously constrained by a complex waterresources legacy and a lack of access to, and management of, thesewater resources.

The primary characteristics of Ethiopia’s water resources are ex-treme interannual and intra-annual rainfall variability, and the inter-national nature of its most significant water resources. Thisvariability is most obviously manifest in endemic, devastatingdroughts and floods. Less apparent is the broad range of impacts thisvariability has on the Ethiopian economy, even in good rainfall years.

This report seeks to clarify the scope and scale of the impacts of hy-drological variability on Ethiopia’s economic performance, poverty,natural resources, and socioeconomic conditions, as well as the man-ner in which water shocks are transmitted through the economy. Byunderstanding what makes the economy vulnerable to hydrologicalshocks, a strategy can be developed to make it more resilient.

The Bank’s support to Ethiopia has always included assistance towater-related sectors. It has financed projects in land and water man-agement, urban and rural water supply and sanitation, hydropower,irrigation, food security, and international waters. But this assistancehas generally been in response to specific requests from the GoErather than part of a comprehensive, integrated strategy.

In recognition of the complex economy-wide impacts of water re-sources management, and the consequent need for cross-sectoralstrategies that integrates the range of Bank-financed programs thataffect, or are affected by, water resources, the Bank’s Water ResourcesSector Strategy (2003) introduced a new vehicle, the Country WaterResources Assistance Strategy (CWRAS). The CWRAS is a Bank strat-egy document—agreed with the GoE—that lays out a comprehensiveapproach for Bank support to water-related sectors.

This CWRAS for Ethiopia will assist the Bank in defining a morestrategic approach to assisting water-related sectors and supportingimplementation of Ethiopia’s Water Sector Strategy (2001) in its ef-forts to regulate rivers, better manage watersheds, improve access tobasic services for households and industries, intensify agricultural

Introduction1

production, increase power production, and moreeffectively de-link economic performance fromrainfall and runoff.

This strategy looks beyond the management ofhydrological variability to interventions aimed atdecreasing the vulnerability of the economy tothese shocks. The report helps clarify linkages be-tween the country’s economic performance and itswater resources endowment and management. Itthen uses this analysis to recommend both waterresource strategies and economic and sectoral poli-cies that will enhance growth and insulate theEthiopian people and economy from the often dev-astating economy-wide effects of water shocks.Specifically, the objectives of this CWRAS are to:

• Investigate and describe the role of water inthe economic performance of Ethiopia from across-sectoral, economy-wide perspective.

• Identify broad strategies for mitigating nega-tive and enhancing positive impacts of wateron the economic performance of Ethiopia.

• Define an overarching strategy for Bank sup-port to water management in Ethiopia, whichwill provide an integrated framework for de-veloping and harmonizing strategies inwater-related sectors.

The focus of our analysis is the challenge ofwater resources: the management, development,allocation, and impact of the country’s overall hy-drology and water resources. It is not intended to

provide detailed investment recommendations.Nor is the report intended in anyway supercede theGoE’s current, sound water, resources strategy.Rather, this report provides analysis to strengthenunderstanding of the dynamics of water in theEthiopian economy and sets out a roadmap for amore coherent approach to Bank support forEthiopia’s efforts to manage its water resourceschallenges.

The report is structured as follows: Section 2describes the devastating impact of hydrologicalvariably on Ethiopia’s economy and its people.Section 3 provides a cross-sectoral analysis of theimpacts of water management on the Ethiopianeconomy, encompassing the environment, agri-culture, energy, water supply and sanitation, in-dustry, and transport sectors. Section 4 outlinessome of the development strategies proposed toimprove conditions. Section 5 presents a quantita-tive model, based on the qualitative description ofEthiopia’s hydro-economic dynamics, that is usedto project economy-wide growth and the effects ofhydrological variability under different policyand investment scenarios. Section 6 makes a casefor why developing countries with variable hy-drology, such as Ethiopia, need investment inwater-related infrastructure. Section 7 offers abrief summary of lessons. Section 8 outlines astrategy of Bank support for Ethiopia’s efforts tomanage its water resources to maximize sustain-able growth.

2 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

3

At first glance, Ethiopia’s water resources endowment appearsgenerous. The country has adequate average annual rainfall, severalmajor rivers and lakes, and significant groundwater resources (seemap 2.1). The total renewable surface water resources are estimatedat 122 billion cubic meters per year from 12 major river basins (seemap 2.2), as shown in figure 2.1, and 22 lakes. Renewable ground-water resources are estimated to be about 2.6 billion cubic meters(MoWR 2002a). Although less than 2 percent of these resources arediverted for use, the current estimated per capita renewable fresh-water resources of 1,900 cubic meters (MoWR 2002b) indicates anabundance of water.

However, the development and management of Ethiopia’s waterresources faces two significant challenges, a natural legacy and anhistorical legacy. The natural legacy is one of high hydrological vari-ability. Rainfall across much of the country is exceptionally variableand unpredictable, both in time (within and between years) andspace. With highly vulnerable watersheds and almost no investmentin water storage, a consequence of this hydrological variability is en-demic and unpredictable drought and flood.

The historical legacy is one of several international rivers, of whichthe Nile is the most important. The Abbay, or Blue Nile, Basin con-tributes 62 percent of the annual average flow reaching Aswan. Whencombined with the Tekezze and the Baro-Akobo Rivers, the total con-tribution from Ethiopia to the flow of the Nile at Aswan is 86 percent(BCEOM 1999). In the Nile Basin, tensions have long been high. In re-cent years, Ethiopia has placed great emphasis on developing theNile in cooperation with its neighbors. Without such bold, affirma-tive action for cooperation, tensions would grow, as demand forwater grows with populations and economies in the basin

RESOURCEEthiopia’s hydrology is greatly influenced by its highly varied topog-raphy. The heart of the country is a vast highland plateau, lying at anelevation of 1,500–3,000 meters with some peaks rising to more than4,500 meters. This central massif is divided by the deep Rift Valley,which runs from northeast to southwest. To the west, the plateauslopes gently away to the Sudan and to the wide plains of the WhiteNile and Main Nile. To the east, a steep escarpment drops to the

Ethiopia’s Water Resources2

4 Ethiopia: Managing Water Resources to Maximize Sustainable Growth M

ap2.

1

Ethiopia’s Water Resources 5

Map

2.2

plains of the lowlands; farther south, these mergeinto the great stretch of the Ogaden Desert.

The Ethiopian highlands contribute to threemajor river systems, the Nile, Awash, and Omo.The northern and central highlands drain west-ward into Ethiopia’s largest river system, theAbbay, or Blue Nile, into the Tekeze River, a tribu-tary of the main Nile, and into the Baro River, atributary of the White Nile. The eastern highlandsdrain into the Awash River, which never reachesthe sea, but is ultimately absorbed into a successionof lakes and marshes near the Djibouti border. Inthe south, the Omo River drains into Lake Turkana,and a number of streams flow into the other RiftValley lakes. In the southeast, the mountains ofArsi, Bale, and Sidamo drain toward Somalia andthe Indian Ocean, but only the Genale or Juba Riverpermanently flows into the sea. Apart from thelarger rivers, there are few perennial streams below1,000 meters (Abate 1995).

Elevations vary from the Afar Depression in thenortheast, which is 110 meters below sea level, tothe top of Ras Dashen Mountain in the north,which is 4,620 meters above sea level. In betweenthere are high mountains, high plateaus, deepgorges, incised river valleys, and low-lying plains(Gebeyehu 2002). This variation in altitude pro-duces a wide variation in climatic and microcli-matic conditions, as well as in soil and vegetationconditions. Mean annual temperatures vary from45oC in the Afar Depression to below 0oC in thehighest lands. Forty-three percent of the country,where 88 percent of the population lives, is morethan 1,500 meters above sea level.

Lake Tana, with an area of 3,600 square kilome-ters, is the largest of 15 natural lakes in Ethiopia,with a total surface area of about 6,900 square kilo-meters. Seven of the eight major natural lakes arefound in the Rift Valley. Most Ethiopian lakes exceptZeway, Tana, Langano, Abaya, and Chamo are ter-

6 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

0

10,000

20,000

30,000

40,000

50,000

60,000

Abbay

Baro-A

kobo

Omo-G

ibe

Tekez

e

Genale

–Daw

a

RiftVall

ey

Awash

Wab

e Sheb

ele

Afar-D

anak

il

Mer

eb

Ogaden

Aysha

0

10,000

20,000

30,000

40,000

50,000

60,000

Ru

nof

f(M

-m^

3/an

nu

m)

surf

ace

run

off

inM

-m

River Basins

Figure 2.1 Surface Runoff from Ethiopia’s Major River Basins

Source: MoWR

Ethiopia’s Water Resources 7

minal lakes. Lakes Shala and Abayata have concen-trations of chemicals used in production of soda ash.

Ethiopian lakes and large perennial rivers are richin fish. The sustainable yield of fish from the majorlakes is estimated to be about 35,300 tons per year.

There is limited information on Ethiopia’s re-newable groundwater resources, which are esti-mated at about 2.6 billion cubic meters, althoughthe Ethiopian Geological Survey is currently con-ducting groundwater investigations. Groundwateris an important source of raw water for urban andrural drinking supplies, and, although the coun-try’s complex geology does not lend itself to majorgroundwater developments, there does appear tobe potential to develop smaller aquifers for irriga-tion and water supply and to explore opportunitiesfor artificial recharge.

Ethiopia’s rainfall shows high spatial and tem-poral variability. The highest mean annual rainfall(more than 2,700 millimeters) occurs in the south-western highlands, and then it gradually decreasesin the north (to less than 200 millimeters), northeast(to less than 100 millimeters), and southeast (to lessthan 200 millimeters).

Based on the annual rainfall distribution pat-terns, three major rainfall regimes can be identified:

• The southwestern and western areas of thecountry are characterized by a mono-modal(single peak) rainfall pattern, with the lengthof the wet season decreasing northward.

• The central, eastern, and northeastern areas ofthe country experience a nearly bi-modal (two

peak) rainfall distribution. The two rainy sea-sons are called Belg (smaller rains fromFebruary to May) and Kiremt (big rains fromJune to September).

• The southern and southeastern areas of thecountry are dominated by a distinctly bi-modal rainfall pattern. The rainy seasons areSeptember to November and March to May,with two distinct dry periods separating them.

Ninety percent of the country’s water resourcesoccur in four river basins6 that host only 40 percentof the population. Sixty percent of the populationlives in the higher lands of the eastern and centralriver basins and depends on less than 20 percent ofthe country’s water resources (MoWR 2001).

National average rainfall figures do not capturethe large spatial variations in rainfall across dif-ferent parts of the country. Not only do arid andsemiarid or drought-prone areas have lower aver-age annual rainfall, but they also exhibit signifi-cantly larger interannual rainfall variability.Figure 2.2 shows rainfall variability, calculated as“rainfall in year t” (RFt in the figure) as a percent-age of the mean of average annual rainfall levelsduring 1980–2001. Variability is shown for thecountry as a whole, for drought-prone state ofTigray and for rainfall-adequate state of Arsi.Tigray, which has a generally lower average an-nual rainfall, exhibits substantially larger interan-nual rainfall variability.

Table 2.1 shows mean and coefficients of varia-tion of average annual rainfall levels for 1980–2001.

Table 2.1. Mean Annual Rainfall and Coefficients of Variation in Selected Drought-Prone and Non-Drought-Prone Areas (1980–2001)

Average annual rainfall Coefficient of variation Regions/zones levels (in mm) (in percent)

Drought proneTigray 710 23.2Amhara/North Wollo 832 27.3Amhara /South Wollo 901 21.3

Non-drought proneOromiya/Arsi 878 11.2Oromiya/Jimma 1,505 10.0Amhara/Gojjam (east and west) 1,180 15.4

Source: Demeke (2004).

These are extremely unfavorable in selecteddrought prone areas relative to non-drought-proneparts of the country.

Rainfall variability may actually be worsening inEthiopia. Table 2.2 shows meteorological data thatsuggest rainfall was more variable in the 1980s and1990s than in the 1970s, although the overall meanin the 1970s was lower.

HYDROLOGICAL VARIABILITY:EXTREMES OF DROUGHT ANDFLOODThe direct economic impact of hydrological ex-tremes (drought and flood) everywhere in theworld is well recognized. Ongoing work by theBank in Africa is providing quantitative analyses ofjust how serious these hydrological shocks can be.The scale of the shock on the Kenyan economy ofthe 1997–8 flood and the ensuing 1999–2000drought, for example, has been estimated atroughly 40 percent of annual gross domestic prod-uct (GDP) spread over a three-year period, withmost of this impact occurring beyond the water-re-lated sectors (that is, infrastructure damage andfalling industrial production).

Many of the poorest regions of the world havethe highest levels hydrological variability, and thepoorest of the poor and women are often thosemost sharply affected. Where this is not the case, itwill often be because sufficient investment wasmade in the hydraulic and institutional infrastruc-ture to provide a reasonable measure of water se-curity (such as in Australia and the southwesternUnited States).

The general environmental and economic im-pacts of too much or too little rain, as well as of ex-treme variability in rainfall, are described below.Section 3 discusses the impacts of this variability oneach economic sector.

Too Little Rain

Drought

In drought years, providing drinking water is par-ticularly challenging. For example, during the2002–3 drought, some urban centers went withoutnormal water supplies for weeks or even months.Water shortages in both rural and urban areas hadserious negative health and sanitation conse-quences and seriously disrupted economic pro-duction. In addition, rainfall variability affects

8 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

0

20

40

60

80

100

120

140

160

180

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Year

0

20

40

60

80

100

120

140

160

180

1980

1982

1984

1986

1988

1990

1992

1994

1996

1998

2000

Per

cen

tage

National rainfall variability(RFt/mean*100)

Tigray’s rainfall variability(RFt/mean*100)

Arsi’s rainfall variability(RFt/mean*100)

Figure 2.2 Rainfall Variability: National, Tigray, and Arsi

Source: Demeke (2004).

Ethiopia’s Water Resources 9

recharge of groundwater in some areas resulting infailure of wells. For example, in the city of Mekele,where the several-year-old water supply system in-cludes eight deep groundwater wells, it was dis-covered that the wells supplied 30–40 percent lesswater than expected. Because the capacity of thenew system did not meet the needs of the popula-tion, additional, deeper, wells will be needed.

Droughts destroy vegetation in watersheds,farmlands, and pastures. Surviving vegetation isalso at risk because during droughts livestock areconcentrated on overgrazed lands, compressingsoils and pulling out roots, which decreases thelikelihood of regeneration. Food shortages duringdroughts drive people to strip natural resourcessuch as fuelwood, crafts supplies, and wild foods.Droughts create conditions conducive to wildfiresand wind erosion. They compromise the ability ofwetlands to absorb and slow stream flows, filterand remove sediments and pollutants, and sustainhabitats for wildlife.

Drought causes crops to fail and livestock to per-ish. If rains fail, or simply come too early or too late,the entire agricultural cycle can be disrupted be-cause the sector has inadequate storage capacity tosmooth and time water delivery.

Drought can also severely undermine hydro-electric generation. During the 2002–3 drought thewater level in Koka Reservoir was an unprece-dented 3–4 meters below normal. Electric power

was rationed, with a one-day interruption perweek, in order to use available water economicallyuntil the next rainy season.

Famine

The history of drought and famine in Ethiopia is asold as the history of the land itself. Such crises havebeen traced as far back as 250 B.C. A major droughtwas recorded in the ninth century, and in thetwelfth century another famine was chronicled thatcaused the death of many cattle. The 1888–92famine was unprecedented in its scale of destruc-tion. Drought, along with heavy infestation ofpests, destroyed about 90 percent of the cattle andwiped out a third of the population. More than 19periods of widespread and severe food shortageshave been recorded in the past 100 years alone. It isclear that the problem has gradually expandedfrom the north to the rest of the country and its in-tensity has deepened (see table 2.3).

Over the past 20 years, agricultural growth hasfallen far short of the rapid population growth:Production levels have fluctuated with the rainfallpatterns and land degradation has intensified.About 10.3 percent of the population (more than 5million people) has required emergency food aideach year (see table 2.4). More importantly, the num-ber of people affected by droughts is increasing.During the broadly publicized 1984–5 drought, 8million people required assistance. In 2002–3, more

Table 2.2. Trends in Rainfall Variability 1970–2001

Average Average Average Year annual rainfall Year annual rainfall Year annual rainfall

1970 955 1981 991 1991 9711971 862 1982 989 1992 10401972 909 1983 1000 1993 10951973 912 1984 904 1994 10831974 815 1985 949 1995 10171975 944 1986 1059 1996 11361976 943 1987 1068 1997 11001977 994 1988 1166 1998 11301978 1033 1989 1111 1999 11051979 924 1990 1070 2000 10551980 935 2001 898

Standard deviation 58.63 78.39 72.43Coefficient of variation 6.31 7.60 6.85

Source: Ethiopian Meteorological Services Enterprise.

than 14 million people (22 percent of the population)required food assistance to survive. A total of 1.7million tons of food aid was provided during thatperiod, with the United States alone donating someUS$500 million worth of food aid. Even in normalyears, some 5 million people require food assistance.

Too Much Rain

Flooding

Flooding is a problem mainly along the riverineareas affecting productive agricultural land, settle-ments, and infrastructure. Flooding and sedimenta-tion also cause damage by inundating andwater-logging productive land, encouraging growthof undesirable species, and blocking road access and

trekking routes. Flooded fields often delay planting,thus reducing yields and quality of crops.

Both floods and droughts cause severe damageto roads, making many roads impassable duringthe rainy season and even during the dry season,due to poor conditions and inadequate river cross-ings. According to a recent survey, the marketroads in three of the more populous regions are im-passable between three and five months of the year(IFPRI/ILRI 2002). Many of the small-scale irriga-tion schemes developed in recent years have beenin food-insecure areas and are inaccessible by road,making it difficult for farmers to reach marketswith the products of their irrigated fields.

Floods do not significantly threaten major citiesor towns in Ethiopia. However, the river plains ofthe Abbay, Awash, Baro-Akobo, and Wabi-Shebele

10 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 2.3. A Chronology of Ethiopian Drought and Famines Since 1895

Year Affected area Triggers and severity

1895–6 Ethiopia Loss of livestock and human lives

1899–1900 Ethiopia Drought deduced from levels of the Lake Rudolf and Low Nile floods

1913–4 Northern Ethiopia Lowest Nile floods since 1695; grain price said to have risen thirtyfold

1920–2 Ethiopia Moderate drought similar to 1895–96

1932–4 Ethiopia Deduced from low level of water in Lake Rudolf in northern Kenya

1953 Tigray and Amhara (Wollo) Severity unrecorded

1957–8 Tigray and Amhara (Wollo) Rain failure in 1957; locust and epidemic in 1958

1962–3 Western Ethiopia Very severe

1964–6 Tigray and Amhara (Wollo) Undocumented; said to be severe

1969 Eritrea Estimated 1.7 million people suffered food shortage

1971–5 Ethiopia Sequence of rain failures, estimated 250,000 dead; 50 percent of livestock lost in Tigray and Wollo

1978–9 Southern Ethiopia Failure of belg (small rain season) rains

1982 Northern Ethiopia Late meher (major rain season) rains

1984–5 Ethiopia Sequential rain failure; 8 million people affected, estimated 1 million dead, and many livestock lost

1987–8 Ethiopia Drought of unrecorded severity in peripheral regions

1990–2 Northern, eastern, and Rain failure and regional conflicts; estimated 4 million people suffered southeastern Ethiopia food shortage

1995–6 Southern Oromiya and Afar Cattle death

1999–2000 Afar, Somalia, South Oromiya, Cattle death and severe water shortageTigray, and Amhara

2002–3 Ethiopia 11.3 million people required food assistance; additional 3 million needed close monitoring

Source: Webb et al. (1992) and Berkele (2002).

Ethiopia’s Water Resources 11

basins are prone to flooding. Urbanization con-tributes to flooding risk because large paved andimpervious areas increase runoff. Furthermore, theconstruction of stormwater sewers and the realign-ment and culverting of the natural stream channelsallows water to be transmitted to the drainage net-work more quickly. Urbanization also increases thepotential costs associated with floods as invest-ments are made in productive assets that are vul-nerable to flood damage.

Massive Soil Loss and Land Degradation

Seventy-nine percent of Ethiopia’s land has a slopein excess of 16 percent, and at least one-third of thisarea has a slope of 30 percent or more (Campbell1991). Torrential or tropical rains, combined withthe absence of adequate cover, result in a heavyflush of water that washes away the soil. Crops thathave not yet adequately rooted during the earlypart of the rainy season may also be washed awayalong with the soil. Farmers cultivating the ruggedterrains in the regions of Tigray and Amhara(Wollo, North Gondar, and North Shoa) frequentlyabandon their farmland after the topsoil is washed

away.7 The soil depth is reported to be 10 centime-ters or less in many areas (Hurni 1988.).

Studies indicate that in many areas erosion ex-ceeds soil formation. Soil formation is estimated atabout 15 tons per hectare per year. In the highlandregions of the Abbay Basin, soil loss in areas culti-vated through traditional practices amount to122–128 tons per hectare per year. This figure morethan doubles in the absence of vegetation. In theBaro subcatchment, the upper plateau with 5–60percent slope produced soil loss of 46–425 tons perhectare per year. In the Tekeze Basin, Quiha sub-catchment soil loss amounted to 33 tons per hectareper year (NEDECO 1997).

The erosive effects of rainfall are significantlyaugmented by Ethiopia’s severe deforestation, itsmountainous terrain, and traditional agriculturalpractices of cultivating steep slopes without pro-tective measures. The loss of forest cover, in turn, isgenerally associated with greater hydrologicalvariability. Ethiopia’s high-intensity storms causesignificant erosion, especially at the beginning ofthe season when the soil is dry. Landslides, gullyerosion, and stream bank erosion commonly result

Table 2.4. Drought- and Disaster-Affected Populations

Drought/disaster- affected Proportion affected Year population (million) (percent)

1980–1 2.82 7.71981–2 3.70 9.81982–3 3.30 8.51983–4 4.21 10.51984–5 6.99 17.01985–6 6.14 14.51986–7 2.53 5.81987–8 4.16 9.31988–9 5.35 11.61989–90 3.21 6.81990–1 7.22 14.81991–2 7.85 15.61992–3 4.97 9.61993–4 6.70 12.61994–5 3.99 7.31995–6 2.78 4.91996–7 3.36 5.81997–8 4.10 6.81998–9 7.19 11.7

1999–2000 10.56 16.62000–1 6.24 9.6Average 5.37 10.32002–3 14.30 22.0

Source: Demeke (2004).

from a combination of hydrology and human dis-turbance.

Deforestation, watershed degradation, roadbuilding, and quarrying all contribute to the fre-quency of landslides that destroy farm lands andinfrastructure. Gullies are created mainly as a re-sult of concentrated runoff in natural drainagelines where there is little or no vegetation coverand along footpaths or tracks, such as those nearwatering points, settlement areas, and frequentlyused grazing fields. Stream bank erosion causesdamage to grazing and croplands adjacent torivers, destroys the approaches to bridges and cul-verts, and causes bridge failure by underminingfootings.

Where soils fail to drain properly followingheavy rains and floods, the water table rises anddeposits salt on the soil surface. The salt inhibitsplant growth by disturbing the osmotic relations inthe root zone. Secondary salinity is also induced byinadequate drainage in irrigation schemes and fre-quently flooded areas.

Power Generation and Irrigation Reduced by Sedimentation

High sediment loads in rivers can reduce powergeneration capacity, interfere with irrigation, con-tribute to flood risk, and affect clean water sup-plies. Ethiopia’s annual average sediment yieldranges between 10 tons per square kilometer (in thesouthwestern parts of the country) to about 1,500tons per square kilometer (in the northern and east-ern parts of the country).

When sediments settle in reservoirs, the capac-ity for power generation is reduced in proportionto the sediment ingress into the reservoir. In addi-tion, concentration of sediment at the power inletshas hampered operation of dam bottom outlets aswell as power intakes.

While the impact of flooding on the energy sec-tor has been relatively limited, flooding has causeddamage to generation equipment at the MelkaWakana and Tis Abbay power plants. The KokaDam—which provides the only control on theAwash River for irrigation and hydropower gener-ation and regulates water supplies to more than 70percent of the nation’s large-scale irrigation—haslost its flood control capacity because of siltation.Available water for downstream development hassharply declined. The reduced regulation capacityhas caused flood damage on farmlands in the

upper and middle valleys (MoWR 2002b).Irrigated agriculture is also threatened by sedi-

mentation in the storage reservoirs and conveyancestructures, which adds to the overall operating andmaintenance costs. Sedimentation affects watersupply to towns that rely on surface water sourcesas well.

In the city of Gonder, for example, during con-struction alone, 20–30 percent of the projected vol-ume of the new water system’s capacity was lostdue to siltation caused by deforestation upstreamin the Angereb River. An additional five wells anda pipeline had to be constructed at a cost of ap-proximately Br5 million to augment the town’swater supply during the rainy season.

INTERRELATED EFFECTS ONNATURAL RESOURCES AND THEENVIRONMENTThe Bank’s recently completed CountryEnvironmental Analysis for Ethiopia concludesthat the country’s principal environmental chal-lenges involve complex cross-sectoral linkages.Two of the key environment-development linkagesidentified by that study relate directly to the chal-lenges posed by Ethiopia’s hydrology. They are thelack of integrated water resources managementand the “land degradation–food insecurity–energyaccess–livelihood” nexus. The latter includes un-sustainable agricultural land management prac-tices and heavy reliance on biomass energy.

The use of biomass spurs deforestation and ero-sion and contributes to a significant environmentalhealth problem: exposure to smoke and indoor airpollution, which causes elevated under age fivemortality and a high incidence of respiratory dis-eases, mainly in women and children. Using cropresidues and dung as fuel, rather than returningthis organic matter to the soil, causes a decline insoil fertility and deterioration in soil structure. Thedegraded land is also more prone to erosion, lead-ing to loss of fertility in the topsoil and to a reduc-tion in soil depth, both of which can have anadverse effect on crop yields.

The consequences of both deforestation and de-graded soil structure include less infiltration of rain-fall, which diminishes groundwater recharge; morerunoff, which contributes to erosion and siltation;and reduced water storage capacity in the soil,which makes crops less able to withstand drought.

12 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Ethiopia’s Water Resources 13

HYDROLOGICAL VARIABILITY:ENDEMISM AND EXPECTATIONSWhile the catastrophic effects of drought and floodextremes are apparent, it is less well recognizedthat even in years of average precipitation, expec-tations of high variability and endemic droughtsand floods affect economic performance and po-tentially the structure of the economy. In countrieswhere hydrological variability is high and invest-ments to achieve water security are inadequate,variability is a constant economic risk to small in-vestors (such as farm families) and large ones (suchas industries), as well as to the nation.

The expectation of variability and the unpre-dictability of rainfall and runoff are likely to con-strain growth and diversification by encouragingrisk-averse behavior at all levels of the economy inall years, as economic actors, particularly the poor,focus on minimizing their downside risks ratherthan maximizing their potential gains. Because theyknow they could lose everything in a single flood ordrought, farm families quite rationally will not in-vest in land improvements, advanced technologies,or agricultural inputs, and thus constraining agri-cultural output and productivity gains. Lack of suchinvestments can lead to land degradation and de-sertification, which will result in a vicious circle ofreducing production and deteriorating assets.

Controlling these real and perceived risks canproduce an “input impact” whereby farmers investin inputs that will increase yields and sustain pro-ductivity. Recent research carried out by Derconand Christaensen (2005) suggest that when down-side risks to farmers are reduced by one standarddeviation, the use of fertilizer may increase byabout 7 percentage points and fertilizer applicationrates by 43 percent.

Similarly, interruptions in water supply andtransportation services caused by hydrologicalvariability are a significant disincentive for invest-ments in industry and services. This disincentivewill slow the diversification of economic activitiesand maintain an economic structure that is basedlargely on low-input, low-technology, agriculturalproduction.

Understanding and mitigating the full impact ofhydrological variability on economic performancewill require a better understanding of the role of ex-

pectations and the incentives created by entirely ra-tional risk aversion.

INTERNATIONAL WATERS:TENSIONS AND OPPORTUNITIESA serious challenge for Ethiopia is the fact that itshares so many international rivers. Ethiopia sits atthe headwaters of three significant transboundaryrivers: the Nile, the Gash, and the Juba-Shebele.8

These waters are shared with numerous riparianstates: Burundi, Democratic Republic of Congo,Egypt, Eritrea, Kenya, Rwanda, Somalia, Sudan,Tanzania, and Uganda.

There are tensions, to a greater or lesser extent,between riparian nations on all internationalrivers. In some cases these tensions underminebroader relations among riparian states and in-hibit potential economic growth by blocking re-gional integration, both in river-relatedproduction and in trade and infrastructure inter-connection “beyond the river” (Grey and Sadoff2002). Tensions over shared rivers encourage theadoption of less economically efficient polices thatfocus on self-sufficiency (for example, in agricul-ture and power) rather than on trade and integra-tion. There is a real risk that these tensions couldresult in the diversion of strategic human re-sources and policy focus from economic develop-ment to security concerns related to water andeven a diversion of financial resources to militarypreparedness.

Recognizing this, the GoE is engaged in a seriouseffort to promote cooperative development andmanagement of the largest of its shared rivers, theNile, through the Nile Basin Initiative (NBI).Ethiopia is playing a leadership role in the NBI,which is building trust and capacity across the NileBasin and provides the framework for preparingand implementing a major program of investment.

Within the Eastern Nile Subsidiary ActionProgram, Egypt, Ethiopia, and Sudan are currentlypreparing investments in irrigation, hydropower,watershed management, and flood mitigation andhave begun to identify a longer-term program ofmajor, cooperative, multipurpose development ofthe river. This riparian cooperation brings huge op-portunities and even the potential for transforma-tional change in Ethiopia.

15

This section looks at the specific impacts of Ethiopia’s extreme hy-drological variability on its key economic sectors and services: waterand sanitation, agriculture (including livestock and fisheries), trans-portation and market infrastructure, energy, urban development,and industries such as manufacturing, banking, and tourism.

WATER SUPPLY AND SANITATIONThe GoE’s National Water Resources Policy states that the highestpriority use of water is human and livestock consumption, whichamounts to less than 1 percent of total water usage. Still, according tothe GoE, only a minority of Ethiopians—42 percent—has access topotable water services and some 11 percent have access to improvedsanitation. Coverage is highest in urban areas, where about 83 per-cent of the population has access to improved water supply and 55percent to improved sanitation facilities.

In most areas, women and girls spend considerable time fetchingwater from wells or rivers, and then care for those who become illfrom waterborne diseases, tasks which keep them from engaging inincome-generating activities and attending school.

The primary source of raw water for urban and rural water supplyis groundwater, though some towns rely on surface water.Groundwater contamination is not generally a problem, but in someareas there are high concentrations of substances such as iron and flu-oride. However, rainfall variability can affect recharge of groundwa-ter, resulting in failure of wells, as in Mekele.

Where surface water is the source of raw water, such as in Gonder,Ambo (80 kilometers west of Addis Ababa) and Nazret, siltation hasdriven up the cost of drinking water. Surveys of Legedadi Reservoir(upper Awash River basin), which provides the water supply toAddis Ababa, indicate that between 1986 and 1996 raw water turbid-ity increased from an average of 123 formazine turbidity units to 520formazine turbidity units, with consequent increases in treatmentcosts as well as other operation and maintenance costs (Addis AbabaWater and Sewerage Authority 2000).

Water Resources and Key Economic Sectors3

Competing Uses

Although the National Water Resources Policygives highest priority to drinking water for thepopulation and livestock, regulations regarding al-location of water resources between consumptiveuses are not always enforced. In Harar, for exam-ple, a new water supply system must now be con-structed at a very high cost because Lake Alemaya,the source of the city’s water supply, was practi-cally pumped dry by farmers growing chat for ex-port (see Box 3.1).

Coverage Targets and Required Investment

Ethiopia’s population is expected to grow by al-most 27 million by 2015. Most of the growth willcontinue to be in rural areas, but 17 percent will bein medium-sized towns and 5 percent in urbanareas, putting stress on already inadequate waterand sanitation systems (see table 3.1).

The GoE’s Sustainable Development andPoverty Reduction Program (SDPRP) identifieswater and sanitation among the key issues to be ad-dressed in order to reduce poverty by enhancingrapid economic growth and improving service de-livery. The SDPRP targets for the medium and longterm, consistent with the Millennium DevelopmentGoals (MDGs), are to halve by 2015 the proportionof people without sustainable access to safe drink-ing water and sanitation. To achieve this about 35million more people in rural areas will need im-proved water and sanitation services and 8 millionmore people in urban areas will need sanitation fa-

cilities (see Table 3.2 and Figure 3.1). The estimatedinvestment required to meet the MDGs is over $170million.

The GoE, with international support, is makingreal progress in the sector. Between 2002 and 2005the percentage of people with access to potablewater increased from 30 to 42 percent, and the per-centage of those with access to sanitation servicesincreased from 7 to 11 percent. Current aid alloca-tions for water and sanitation from the Bank andthe African Development Bank amount to US$164million, with another US$60–70 million per year al-located by the GoE.

Water supply and sanitation in both urban andrural communities are characterized by low levelsof service and lack of sustainability. User fees areoften so low that they do not even provide for ade-quate maintenance of existing facilities. Urbanpiped water supply systems often operate wellbelow capacity due to maintenance problems.Rationing and service interruptions are frequent.Water utilities tend to be understaffed and staff areoften underpaid, inadequately trained, and lackingin the resources required to do their jobs.

Although water supply systems in the largercities have recently been improved, they need to beexpanded to meet the demands of populationgrowth and the planned industrial zones. In ruralcommunities, water systems have too often beeninstalled without adequately training the commu-nities to manage and maintain them. When systemsbreak down, communities often do not have thewherewithal to obtain spare parts and maintenancesupport.

16 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Box 3.1. Weak Regulation of Competing Consumptive Uses in Harar

For about 50 years, the people of the city of Harar re-lied on Lake Alemaya for water supply. As recently2000, the depth of the lake was approximately 8 me-ters. But a combination of low rainfall and increasedreliance on the lake for irrigation caused the lakelevel to drop to 20 centimeters. The lake has virtuallydisappeared.

Farmers in the area had been growing chat for localconsumption. Since about 2000, the demand for chatincreased and is now Ethiopia’s second leading export

crop. To meet demand, farmers brought in high volumepumps and drew water from the lake around the clockfor irrigation. Ethiopia has regulations for allocation ofwater resources between consumptive uses, but owingto lax enforcement, Lake Alemaya is now almost gone.

The water supply situation in the town is desperate.Families are paying a high price to have water deliv-ered by vendors. The town has stopped giving licensesfor new commercial and industrial businesses until thewater supply situation is solved.

Source: Authors.

Water Resources and Key Economic Sectors 17

Limited Financing

Budget resources combined with donor and bilat-eral assistance have been insufficient to significantlyimprove coverage and will be insufficient to meetPoverty Reduction Strategy Paper and MDG targetsfor water supply and sanitation unless funds can bechanneled more efficiently and effectively. Donorsand bilaterals have funded a myriad of projects withdifferent policies and implementation arrange-ments, some through the federal GoE and some di-rectly to regions, towns, or rural communities. As aresult, transaction costs have been high.

For urban water supply, full cost recovery re-quires that user fees are set at a level sufficient tocover operating and maintenance costs as well as

investment costs for systems improvements andexpansion. Ethiopia has not achieved full cost re-covery in the past, and as a consequence utilitieshave not been in a position to expand or improvetheir systems, nor to attract private sector financingto do so. The GoE is actively exploring ways inwhich to align the immediate imperatives of pro-viding affordable services to growing populations(which may require time-bound, targeted subsi-dies), with the financing needs and options of util-ities whose long-term goal is full cost recovery.

Decentralization and Capacity Constraints

Until recently, the federal GoE has been responsi-ble for identifying, planning, and implementing

100.0

80.0

60.0

40.0

20.0

0.0

Pop

ula

tion

(m)

urban rural total urban rural total

Water Supply Access

access (2000) added for MDGs (2000-2015) no access

Figure 3.1 Access to Water Supply: Millennium Development Goals

Source: Adapted from World Health Organization Joint Monitoring Program.

Table 3.1. Population Distribution by Settlement Size

Rural Small towns Medium towns Urban Total

Settlement size (by population) <2,000 2–10,000 10–75,000 >75,000

Population 2000 51.5 m (82%) 1.4 m (2%) 7.0 m (11%) 2.6 m (4%) 62.5 m

Population 2015 67.8 m (76%) 1.7 m (2%) 14.8 m (17%) 4.9 m (5%) 89.2 m

Number of settlements 125,000 580 65 1

Source: MoWR (2002).Note: “m” = millions

water supply and sanitation improvements. Manyof these responsibilities are now being shifted tothe regional and local governments, consistentwith the GoE’s policy on decentralization.

The GoE’s overall policy on decentralization isreflected in the National Water Resources Policy,which states that ownership and management au-tonomy should be devolved to the lowest possiblelocal level. The GoE has provided block grants toworedas (districts), which can be used to improvewater supply, among other things, at the discretionof the local governments. But regional and localgovernments have limited capacity to carry outthese new responsibilities. In the future, theMinistry of Water Resources (MoWR) will focus onfacilitation and regulation of the sector rather thanon implementation.

There is wide disparity in capacity between themore developed regions (Amhara, Oromiya,Tigray, Harari, and outhern Nations, Nationalitiesand Peoples [SNNP]) and the emerging regions(Afar, Somali, Benshangul, and Gambela). Thestandard approaches to improving water supply inthe highlands simply do not work in the pastoral-ist regions. Emerging regions will likely need sig-nificant capacity building and continued supportfrom the MoWR until they develop capacity toplan, prioritize, implement, and monitor their owninvestment programs.

Sanitation

With the exception of Addis Ababa, which has asewerage system, households are expected to fi-nance, install, and maintain their own sanitation fa-cilities—latrines and septic tanks—though somesubsidies are provided for demonstration projects.The GoE’s role has typically been to promote sani-tation and hygiene and to provide an enabling en-vironment for private sector service providers. Thelow coverage level—15 percent of the population—is partly attributed to the fragmentation of respon-sibilities across different government institutionsand lack of effective coordination. In addition,towns and cities have not integrated planning forsanitation and hygiene with water supply im-provements, as called for in the National WaterResources Policy.

Most towns do not have adequate means of sep-tic removal (vacuum trucks) and many do not havedesignated areas for waste disposal or do not ade-quately enforce regulations. This situation con-tributes to environmental degradation of the citiesand, in some cases, to contamination of the watersupply. The SDPRP recognizes the need for sewer-age systems in a number of growing cities and forexpansion of the Addis Ababa sewerage systemover the coming 10 years. The investment needs forurban sanitation will be considerable.

18 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 3.2. Access to Water Supply and Sanitation

2000 2015 MDG GAP (2000–15)

Population Access Access Pop Access Access Added Investment(million) (million) (percent) (million) (million) (percent) (million) (US$/year)

WaterRural 51.5 6.7 13 73.8 41.7 57 35.0 97Urban 11.0 8.0 73 17.5 15.8 90 7.8 34Total 62.5 14.7 24 91.3 57.4 63 42.8 131

SanitationRural 51.5 3.1 6 73.8 39.1 53 36.0 12Urban 11.0 6.1 62 17.5 14.2 81 8.1 28Total 62.5 9.2 15 91.3 53.3 58 44.1 40

Source: Adapted from World Health Organization Joint Monitoring Program.

Water Resources and Key Economic Sectors 19

Urban Wastewater

Urbanization, industry, and services all create ad-ditional water demand and generate wastewater,potentially diminishing both water quantity andquality. Abstraction for these purposes will lowerflow levels and require tradeoffs among users.Flows from urban watersheds are likely to containsignificant quantities of pollutants.

A reduction in the natural river flow togetherwith a discharge of poor quality drainage water canhave severe negative impacts on downstream usersand the environment. The health of downstreamsettlements is at risk. Habitats both within andalongside rivers are particularly rich, often sup-porting a high diversity of species. A large reduc-tion in riverflow combined with a heavy pollutionload will alter micro-habitats, of which wetlandsare a special case. It is particularly important toidentify any endangered species whose survival isendangered because of their restrictive ecologicalrequirements. In addition, fish and other animalsmay become contaminated by untreated waste-water, which has significant implications for prod-uct quality and public health.

AGRICULTURE, LIVESTOCK, AND FISHERIESAgriculture accounts for the lion’s share of theEthiopian economy (almost 50 percent of GDP), fol-lowed by services (38 percent) and industry (11percent), and employs the majority of the popula-tion (80 percent). The economy has shown littlesign of transformation over the past two decades:the share of agriculture remains high, with only aslight decline, from 52.9 percent in the 1980s to 48.2percent in the 1990s (see Table 3.3.)

Because of its dominant share in the economy,fluctuations in agriculture are reflected in the GDP,

as shown in Figure 3.2. These fluctuations have un-dermined the country’s ability to sustain per capitagrowth rate in the 1980s and 1990s.

Until the late 1950s, Ethiopia was not only self-sufficient in staple food but was actually a net ex-porter of food grain. In the 1960s the average percapita food production was about 280 kilogramsper year. However, since the early 1970s, domesticfood supply has often failed to meet the food re-quirements of the people. Even though sufficientfood has been produced in most good years, aver-age food production during the last decade re-mained almost stagnant. Over the past 30 years, percapita food production has declined to about 160kilograms per year.

At present, more than half of the population isfood insecure, mainly because of limitations ofrural land holdings, where more than one-third ofthe households farm less than 0.5 hectares of landunder rainfed agriculture with minimal agricul-tural inputs. Depending on rainfed agricultureputs Ethiopians at the mercy of the country’s ex-treme hydrological variability. Drought is by farthe most important constraint on agriculture at na-tional level, but flooding and frost can be equallyimportant in specific areas.

In addition to crop loss or damage in a givenyear, flood and drought can take a heavy toll onlong-term productivity. The loss of topsoil becauseof heavy rains diminishes organic matter and avail-able nutrients, compromising soil fertility. Loss ofvegetation cover in the highlands and degradationof soil structure reduces water infiltration rates,which lowers base flows, and hence water avail-ability for crops. Lower infiltration rates, in turn,can increase runoff and thus the likelihood and in-tensity of floods.

Declining agricultural productivity, in turn, in-hibits investments in land management and en-courages the extension of agricultural lands (rather

Table 3.3. Sectoral Shares of GDP (Percentages)

Sector/year 1980–1 to 1991–2 1992–93 to 2000–1

Agriculture 52.9 48.2Industry 11.9 10.8Services 35.2 44.0

Source: MoFED.

than intensification of production) leading to fur-ther land degradation, closing a vicious circle thatleaves the landscape increasingly fragile and in-creasingly vulnerable to drought and rainfall-re-lated erosion.

Given the dominance of agriculture in theEthiopian economy, the weak and variable perfor-mance of this sector will significantly impact thebroader economy. Impacts will be transmittedthrough price and income effects, loss of assets, dis-incentives to investment, and public health.

Rainfed Agriculture

This section describes in detail the impacts of hy-drological variability on the country’s major rain-fed crops: cereals, pulses, oilseeds, and coffee. Itdiscusses the limited amount of irrigated agricul-ture. It also describes the current situation andtrends regarding livestock and fisheries.

The dominant agricultural system in Ethiopia issmall-holder production of cereals under rainfedconditions, with a total area of approximately 10million hectares.

Cereals

Fluctuations in cereal yield levels are extremelyhigh and closely follow the pattern of rains. Thereis a significant correlation between national cerealyield and national average rainfall.0 National yieldlevels declined by as much 25 percent in 1984, 17percent in 1994, and 10 percent in 1997. Because ofthese fluctuations, the mean cereal yield showedlittle change over the years and remained at lessthan 12 quintals per hectare. With less than 1 per-cent growth rate in yield, nearly all increase in pro-duction over the years was due to expansion in thearea of cropland. Expanding cropland area, ratherthan enriching the soil of current cropland, can con-tribute to deforestation and land degradation.

Farmers do not plant if there is no rain at plant-ing time. Planting during the small rainy season(belg) is often interrupted if there is no rain early inthe season (February–March). Little or no rain inJune could also mean that most cereals (with the ex-ception of teff, which can be planted in July) cannotbe planted. If the rains decline or disappear afterplanting, near complete or complete crop failure is

20 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

-80

-60

-40

-20

0

20

40

60

80

1982

1985

1986

1987

1988

1989

1991

1992

1993

1994

1995

1996

1997

1998

1999

2000

Year

Per

cen

tRai

nfa

llV

aria

tion

arou

nd

the

Mea

n

rainfall variation around the mean

GDP growth

1997

1999 -

-

-

-

-

30

25

20

15

-10

5

0

5

10

15

20

25

Per

cen

tCh

ange

inG

DP

Gro

wth

1984

1983

1990

Figure 3.2 Rainfall Variation Around the Mean and GDP Growth

Source: Compiled from SIMA and African Rainfall and Temperature Evaluation System data.

Water Resources and Key Economic Sectors 21

possible. Seriously affected areas are generally ex-cluded from reporting, and, hence, the area undercultivation declines sharply in years of unfavorableweather.

Pulses and Oilseeds

Pulses and oilseeds yields are even more sensitive toweather and pest conditions than cereals. The an-nual yield of pulses declined by 37 percent in 1998,24 percent in 1984, and more than 10 percent in 1983,1985, 1987, and 1993. Fluctuations in the yield ofoilseeds were of an even greater magnitude, declin-ing by 10 percent or more in 8 of the 21 years. Thereis a similar extreme swing in the area under cultiva-tion of pulses and oilseeds, as virtually no plantingoccurs under adverse weather conditions. Becausethe risk of crop failure is so high, farmers may be un-willing to spend money on fertilizers and other in-puts. Fertilizer is applied to only 14 percent of pulsesand 7 percent of oilseed crops.

Coffee

Coffee is Ethiopia’s most important export earner,accounting for 40–60 percent of exports in value.Various taxes on the crop constitute an importantsource of government revenue. About 15 millionpeople in Ethiopia are believed to depend on coffeefor their livelihoods. However, Ethiopia’s coffee pro-duction is very sensitive to rainfall conditions:Drought and untimely rains have frequently causedsevere fluctuations in the level of the coffee harvest.

Total coffee output has showed no sustained in-crease over the last two or three decades. Coffeeproduction has not yet rebounded to the level at-tained just before the 1984–5 drought. After thedrought, production dropped 40 percent, muchgreater than the drop in food crop production. In2002–3, the coffee harvest declined by 30 percentdue to drought in coffee-producing areas of thewestern, southwestern, and eastern parts of thecountry, according to the U.S. Famine EarlyWarning System Network. Unlike the annual cropsthat may rebound the next year, coffee productiontakes up to four years to recover to its original level.

Because of drought and low coffee prices, farm-ers in many parts of the country are reportedly up-rooting their coffee trees and replacing them withchat (a mild, leafy stimulant) or with maize. An es-timated 75 percent of coffee farmers in theHararghe highlands (home to the famous, highlyaromatic Harar coffee beans) have uprooted theirtrees and planted chat, according to the Oromiya

Coffee Farmers Cooperative Union. Removal oftree crops, and the forest cover often associatedwith coffee cultivation, contributes to deforestationand soil erosion.

Irrigated Agriculture

Less than 5 percent (about 200,000 hectares) of theestimated potential 3.7 million hectares of irrigableland in Ethiopia is under irrigation. Four types ofirrigation are practiced: traditional (38 percent),modern communal (20 percent), modern private (4percent), and public (38 percent).

• Traditional irrigation: Has been practiced inEthiopia for centuries. The traditionalschemes are generally based on community-constructed diversions, which are frequentlywashed away with rainy season floods, re-quiring significant investments in labor by thecommunity during the growing season.

• Modern communal: Small-scale irrigationwas promoted after the 1974 land reforms andactively encouraged following the 1983drought. These schemes were developed bythe GoE or nongovernmental organizationswith beneficiary farmers usually operatingand maintaining them through users’ associ-ations.

• Modern private: Were initially developed inthe 1950s and 1960s by foreign investors forsugar and cotton farms. This category ceasedto exist when these farms were nationalized inthe 1970s, but they have begun to re-emergein the past 10 years through denationalizationand the development of new enterprises.

• Public systems: Replaced the private systemsafter nationalization and are now being aban-doned or privatized.

Irrigation Development Program

The GoE’s Irrigation Development Program (IDP)calls for an additional 274,000 hectares to comeunder irrigation within its 15-year plan period of2002–16, an increase of 135 percent over currentlevels. Total investment requirements for the planare estimated at US$1,683 million.

The IDP has set average annual growth rate tar-gets for developing irrigated areas. These include:4.5 percent during the short term, 5.5 percent dur-ing the medium term, and 6.5 percent during the

long term. For the short term, these targets translateinto irrigated area of 147,000 hectares of large- andmedium-scale irrigation schemes and 127,000hectares of small-scale irrigation schemes.According to these targets, new projects will add274,000 hectares to the 197,000 hectares alreadyunder irrigation, resulting in a countrywide total of471,000 hectares of irrigated farmland by 2016.

Since the IDP was launched, more recent gov-ernment plans aim to develop about 750,000hectares of irrigated land by the end of the plan-ning period. Currently three projects (Kesem,Tendaho, and Koga) are under construction, total-ing 97,000 hectares. Another 12 projects are sched-uled for completion by 2010, adding about another259,000 hectares (Tafesse 2004).

The IDP projects are expected to reduce the na-tional cereals deficit by 11 percent and deficits inseed cotton and sugar crops by 24 percent each.These estimates are based on the assumption that50 percent of newly irrigated areas will be devotedto cereals and 50 percent to cotton and sugar crops.

Livestock

Ethiopia has the largest livestock population inAfrica, about 35 million tropical livestock units. Atthe national level, livestock accounts for 11 percentof export earnings, second only to coffee, and about15 percent of GDP (30 percent of the agriculturalGDP). It is estimated that 40–90 percent of the cashincome of rural households is generated throughthe livestock sector. In addition, cattle provide draftpower and rural transport. Cattle dung is a sourceof soil fertility and fuel for household energy(Tafesse 2004).

Livestock are an integral part of nearly allmixed-type highland farming systems. Livestock isalso the major component of the livelihoods of thepastoralists in the arid and semiarid lowlands ofthe country. Pastoralism is practiced extensively,particularly in the Somali and Afar regions, in theBorana zone of the Oromiya region, in the SouthOmo zone of the SNNP region, in the Benshangulregion, and in the Gambela region. Ethiopia’s low-land pastoral areas encompass more than 60 per-cent of the country (500,000 square kilometers) andcontain more than 11 million animals and 7 millionpeople.

Increasing human and livestock populationpressure has caused serious natural resource

degradation, and droughts are becoming more fre-quent. The pastoralists’ capacity to cope withdrought has declined to the point where survival ofviable pastoral production systems is threatened.

Drought in a pastoral system can degrade bothland and water resources, making recovery moredifficult with each cycle. When water is scarce, live-stock populations become concentrated around theremaining water sources, removing vegetation,compacting soil, polluting water, and acceleratingland degradation. The high concentration of animalwaste often pollutes the water sources, which canspread parasitic infestation to both animals and hu-mans, posing major risks for public health.

Historically, pastoralists coped with drought bymigrating to greener pastures. However, migrationis now constrained by the expansion of settlers dueto fast population growth. Some pastoralists stilltry to cope with drought by moving beyond theirnormal range. For instance, in 1999–2000, Somalipastoralists from the zones of East Gashamo andWarder traveled hundreds of kilometers towardGode looking for pasture and water. Such move-ment beyond traditional boundaries may result inconflict and loss of lives.

Drought reduces the availability of livestockfodder. In the highland areas the availability ofcrop residues limits the impact of drought, but inpastoral areas grazing land is the only source offeed. A rough estimate of the feed balance in goodand bad rainfall years shows that feed productioncovers all the requirements only in exceptionallygood years. There is a 35 percent deficit in normalyears, and a 70 percent deficit in drought years.

As supplies of grazing and water diminish,households reduce herd size by selling first smallruminants and later cattle such as older males andyoung animals unlikely to survive stress. But masssales of this nature saturate the market, leading toa decline in the value of livestock and, as happenedin 2003, a reduction in the price of meat. The declinein livestock prices is in sharp contrast to the rise incereal prices during drought. The eroded livestock-cereal terms of trade during drought often putsherders and mixed farmers in marginal areas at aconsiderable disadvantage (Webb et al. 1992).

Many livestock owners do not sell their animalsin time, hoping that the rains will come before it istoo late. But the animals lose weight and value andeventually die under prolonged drought.Losses of

22 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources and Key Economic Sectors 23

animals and shrinkage of grazing resources due torecurrent drought have reduced the number of an-imals per household in the pastoral areas.

Mesfin (2002) estimated that during the1999–2000 drought, herd size declined by as muchas 80 percent in Oromiya and Somali, and between45 and 50 percent in Afar and the SNNP region. InBorana, South Oromiya, the number of heads ofcattle per household reportedly fell 37 percent from1980–1 to 1996. The impact of such decline in thenumber of animals owned is obviously devastatingsince the pastoralists have no other store of wealth.

Full recovery from a given drought year has be-come increasingly difficult (Webb et al. 1992) be-cause of the increase in the frequency of thedrought and shortage of feed even in normal years.In 1999–2000, a rural household on average owned4.1 head of cattle (including oxen) and about 21percent were found to have no cattle at all (MoFED2002). This loss of assets undermines consumer ex-penditures and future investments in the economy.

Fisheries

Although fish consumption is not commonplace inEthiopia, it has increased in popularity in AddisAbaba and other urban centers in and around theRift Valley lakes area over the last 20 years.Catching fish in the Rift Valley lakes has so far keptpace with the increased demand. Elsewhere thefishing potential has hardly been touched.

About 120,000 square kilometers of Ethiopia’sarea is covered by water and water courses, whichmay be exploited for fisheries directly or throughenhancement. However, knowledge of stock sizesand the means to exploit them are not yet knownfor each fishery waters. Empirical estimates put thecountry’s total fishery potential at 40,000–50,000tons per year. Currently only 20–30 percent of thisamount is exploited, leaving considerable room forexpansion.

TRANSPORTAION AND MARKETINFRASTRUCTUREIn many years, areas of Ethiopia with high agricul-tural potential can produce enough food to meetthe needs of the people in the food deficit areas.However, because of poor communication andtransport infrastructure and the lack of storage andmarketing facilities, people living in food-deficit

areas continue to face famine and food insecuritywhile producers in surplus regions endure unat-tractively low prices.

In most areas, farmers sell their surplus in smallopen-air local markets because high transport costsprevent them from accessing market towns andurban areas that have a wider range of participants,including wholesalers, retailers, part-time farmer-traders, brokers, agents, assemblers, processors,and consumers.

Anywhere outside the small all-weather roadnetwork, farmers and potential investors must as-sume there will be disruptions and unpredictablecosts involved in moving their goods to market.The risk associated with these uncertain costs, andthe additional uncertainties associated with inade-quate market infrastructure once goods are deliv-ered, is a strong disincentive for investment inmarketable agricultural surpluses and non-agricul-tural production, which helps keep the economytrapped in a low-level subsistence equilibrium.

Improved market functioning would benefit bothproducers and consumers by reducing marketingmargins, thereby raising and stabilizing farm in-comes and reducing the cost of food to consumers.Well-functioning agricultural markets promotefarmers’ incentives to use productivity-enhancinginputs, invest in land management, and reduce de-pendence on food aid for survival.

Road Network

Ethiopia’s rugged topography and torrential trop-ical rains make the cost of building and maintain-ing roads extremely high. Inadequate roadmaintenance adds so greatly to the costs of trans-port that many operators are reluctant to provideservices during the rainy seasons. Human portagesand pack animals are still the main means of trans-port in the country.

Only 25 percent of Ethiopia’s area is served by amodern road transport system, and only a rela-tively small percentage of those roads are pavedand generally passable year round (see Table 3.4and Map 3.1)The road network is made up of an es-timated 33,000 kilometers of roads, 3,800 kilome-ters (12 percent) of which are paved all-weatherroads and the remaining 29,000 kilometers aregravel or earth surfaced. More than half of the roadnetwork needs to be rehabilitated or reconstructed,

and most of the remaining network is either in, orfalling into, a state in which routine and periodicmaintenance may be insufficient.

Road density in Ethiopia is one of the lowest inAfrica: At 27 kilometers of roads per 1,000 squarekilometers of land it is well below the African av-erage of 50 kilometers per 1,000 square kilometers.Some 70 percent of farms were reported to be morethan half a day’s walk from an all-weather road in2002 (FDRE 2002) and 17 kilometers to the nearestcommercial transport (World Bank 2005b). Even so,road transport accounts for roughly 95 percent ofthe country’s passenger and freight traffic and pro-vides the only form of access to most rural com-munities.

Most of the all-weather roads radiate from thecapital, Addis Ababa, to major towns. Direct linksbetween regions are rare, with few roads directlylinking regional towns to one another. As a conse-quence, the road distance between two towns isoften twice the air distance.13 This significantly di-minishes the potential returns to interregionaltrade, and little such trade is seen. If it were prof-itable, interregional trade could help to integrateand strengthen the national economy by creatingtrade between surplus and deficit regions and re-gions producing complementary outputs.

Rain Damages Roads

Unpaved roads are extremely vulnerable to floods,landslides, and gully erosion. Unless rain waterdrains properly from unpaved roads, the road sur-facing construction material is washed away, re-sulting in a bumpy washboard surface. Such roadsrequire frequent expensive maintenance. In addi-tion, they create significant inconvenience for trav-

elers and costs for vehicle owners. According toPavement Management Branch of the EthiopianRoad Authority (ERA), water-related damage ac-counted for over 35 percent of the defects in federalpaved roads and 60 percent of the defects in un-paved roads reported from 2000 through 2002 (seeTables 3.5 and 3.6).

Tafesse (2004) demonstrates that water/hydrol-ogy related defects are markedly more pronouncedwhere the mean annual water surplus (that is,runoff) is higher (see Figure 3.3). The actual costsfor water-related road maintenance vary accordingto the annual rainfall, from a low of 4 percent inDire Dawa, where annual rainfall is less than 100millimeters, to 28 percent in Sodo, where the aver-age annual rainfall is 700 millimeters.

The ERA also reports that floods following tor-rential rains cause frequent landslides andwashouts of drainage structures. Few records exist,however, as to the extent, location, cause, duration,and cost of traffic disruptions. ERA identifies wa-tershed degradation as the root cause of increasedflood intensities.

Road Sector Development

In 1997, the GoE launched a 10-year Road SectorDevelopment Program (RSDP) (1997–2007). Underthe first five years, completed in 2002, the focus wason rehabilitation of the core road networks.Substantial progress was achieved in reopeningnearly all of the classified roads,15 and accessibilityimproved and the percentage of roads in good con-dition was increased. The total length of pavedroads was increased to 1,760 kilometers from 840kilometers in 1994, and total length of unpavedroad was increased from 2,877 kilometers to 3,585

24 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 3.4. Road Passability

Number of months the Number of months the dry dry season road is impassible Percent of woreda cut off

Region season road is impassible for four-wheel vehicles during the impassable period

Tigray 5.4 6.0 (probably 4.0) 22.0Amhara 4.6 3.3 27.9Oromiya 3.2 2.3 28.8Total 4.3 3.5 26.4

Source: Amha and Gabre-Madhin (2003).

Water Resources and Key Economic Sectors 25

Map

3.1

kilometers. At the same time, substantial progresswas made in institutional capacity building and insecuring the required levels of funding for mainte-nance and capital works. Road condition surveysindicate significant improvement in the perfor-mance of both paved and unpaved roads.

Based on the lessons learned from the first phaseof program, the second phase has been plannedthrough 2015 under low- and high-cost scenarios.

The low-cost scenario envisions the rehabilitationof 1,014 kilometers, the upgrading of 1,864 kilome-ters, the construction of 5,693 kilometers includinglow-class roads, and the implementation of com-prehensive policy and institutional reforms thatcommenced during RSDP I. In addition it includesheavy and routine maintenance in several types ofroads, bridge works, and studies for follow-on op-eration. The estimated cost is Br19.5 billion (US$2.2

26 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

30

25

20

15

10

5

00 500 1000 1500

District mean annual surplus flow

Dis

tric

tmea

n%

ofW

ater

rela

ted

mia

nten

ance

cost

Figure 3.3 District Mean Annual Surplus Flow Versus District Mean Percent of Water-RelatedMaintenance Cost

Source: Tafesse (2004).

Table 3.5. Percentage of Reported Defects in Paved Roads

Year Water related Deformation related Other

2000–1 56.45 11.40 8.12001–2 21.74 17.84 4.5

Notes: Water-related defects include ditch clearing, gully formation, and culvert/bridge maintenance; deformation-related defectsinclude rutting, potholes, high traffic, and heavy axle load; other defects include pavement cracking and so on.

Table 3.6. Percentage of Reported Defects in Unpaved Roads

Year Water related Deformation related

2000–1 67.76 26.172001–2 51.60 29.54

Source: Tafesse (2004).

Water Resources and Key Economic Sectors 27

billion). The high-cost scenario involves the reha-bilitation of 1,168 kilometers, the upgrading of2,045 kilometers, and the construction of 8,383 kilo-meters of roads at an estimated cost of Br62.7 bil-lion (US$7.2billion) (see Table 3.7).

ENERGY AND HYDROPOWEREthiopia has an immense renewable energy poten-tial in the forms of hydropower, solar, and wind, aswell as significant untapped geothermal resources.However, the per capita electricity consumption isamong the lowest in the world. According to theGoE, only 15 percent of the population has accessto electricity, and the bulk of consumption (82 per-cent) is in the household sector.

Presently, 95 percent of national energy con-sumption is in the form of fuelwood, dung, cropresidues, and human and animal power. The re-maining 5 percent is electricity, 90 percent of whichis generated by hydropower.

This dependence on biomass energy causes in-creased deforestation, shortages of fuelwood, anddegradation of rural ecosystems. The country’sforests have diminished to less than 7 percent oftheir estimated natural extent. The demand forwood products, especially fuelwood, is expected toincrease at around 3 million tons annually.

Deforestation in watersheds has contributed tohigh rates of siltation in the country’s reservoirs,which limits the useful life of hydropower struc-tures (and those used for municipal water sup-plies and irrigation as well), which furtherincreases the cost of modern energy. For example,the Bati and Kombolcha reservoirs in Wello,

Amhara,have completely silted up, night storagereservoirs in the sugar estates of Wonji andMetahara near Addis Ababa suffer from sedimen-tation, and the storage capacity of the Koka Damin the Awash Valley was reduced by 21 percentbetween 1959 and 1981 (HALCROW 1997). Thisloss of reservoir capacity has caused reduced hy-dropower generation, reduced water availabilityfor irrigation, and reduced capacity to regulate theriver, thereby exposing the river basin to generalwater shortages and to flooding.

Ethiopia currently has 731 megawatts of de-pendable power, most of it hydropower. TheEthiopian interconnected system provides electric-ity primarily to the cities of Addis Ababa, Nazret,Dire Dawa, and Harar, and their immediate sur-roundings and presently has 600,000 customers.The self-contained system provides electricity tothe remaining load centers in Ethiopia from iso-lated mini-hydro and diesel installations.

The present network is severely overloaded andsubject to low and fluctuating voltages, which lim-its its ability to provide connections for large-scalecommercial and industrial users. The main issuesfacing the power sector are the low level of accessto electricity provided, the inability to connect largenew commercial and industrial customers, and thehigh cost of future generation investments.

Hydropower electrical systems are especiallyprone to disruption by drought. During the 2003drought, for example, power disruptions (caused byload shedding)16 were common. Each one-day inter-ruption in power during this period was estimatedto result in a loss of 10–15 percent of that day’s GDP(EEPRI 2003.) Energy-intensive and water-intensive

Table 3.7. High-Cost Scenario Road Investment Targets to 2015

Indicators 2003–4 2014–5

Proportion of asphalt roads in good condition 49% 86%Proportion of gravel roads in good condition 34% 88%Proportion of rural roads in good condition 36% 52%Road density/1,000 km2 33 km 91 kmRoad density/1,000 population 0.51 km 1.07 kmProportion of area more than 5 km from a road 63% 28%Average distance to a road 11 km 4 km

Source: Government of Ethiopia (1996a).

farms were among the hardest hit. This vulnerabil-ity can be mitigated by developing alternativesources, such as geothermal, or by expanding gen-eration and the grid so that temporal and spatialvariations in hydrology can be balanced.

Inexpensive, reliable electricity could promotesignificant investment in the more power-intensiveindustry and service sectors. In agriculture, it couldincrease the profitability of fertilizer productionand pumping for irrigation, both of which wouldhelp improve crop reliability and yield.

The GoE’s strategy is to aggressively improve thesupply of affordable electricity in hopes of reducingthe demand for biomass and, thereby, protecting thewatersheds. Other key facets of the strategy includedevelopment of the country’s substantial hy-dropower resources by both the private and publicsector; liberalization of generation transmission, dis-tribution, and supply in isolated areas; commercial-ization and decentralization of operations; andstrengthening the regulatory environment.

Ethiopia’s economically feasible hydroelectricpotential is estimated to be 100 times more thancurrent production.17 Recognizing the potential ofthis resource and the country’s need for reliable en-ergy, MoWR has designed an ambitious five-yeardevelopment program to meet the country’s pro-jected demand. MoWR has two projects currentlyunder construction (Tekeze and Gilgel Gibe II), fivemore in development, and an additional sevenunder study. While these developments will fur-ther increase the dependency of the system on hy-dropower, future energy developments areanticipated to be an alternating mix of thermal andhydropower (EEPCO 2003). See Table 3.8.

URBAN, MANUFACTURING, AND SERVICES

Urban Centers

Ethiopia is less urbanized than most developingcountries in Africa and the rest of the world. Only12 percent of the population lives in cities. In 2003there were roughly 300 urban centers with popula-tions of more than 5,000. Only 23 of these had apopulation of more than 50,000, and only one hadmore than 250,000 inhabitants. The largest city,Addis Ababa, is 12 times larger than the secondcity, Dire Dawa. Addis Ababa is growing at a rateof 3.3 percent, quite low by world standards and

only slightly above the natural growth rate. The av-erage growth rate of Ethiopia’s urban centers hasdeclined from 5.5 percent in 1960–75 to 4.5 percentin 1975–89 and to only 3.8 percent in 1989–99, whilethe growth of the population as a whole has re-mained almost constant.

Urban areas contributed 34 percent of Ethiopia’sGDP in 1997, compared with 30 percent in 1987.Addis Ababa alone provides about one-fifth of theoverall GDP and 52 percent of the modern sectorvalue added. Urban productivity, or value addedper urban inhabitant, is nearly four times that ofrural areas, and Addis Ababa is 2.5 times more pro-ductive than the average of other cities and towns.The GoE’s Abbay River Basin master plan statesthat urbanization contributes to economic diversi-fication and provides “the only” viable base forlong-term growth in average incomes.

Ethiopia suffers from the relative weakness of itsnetwork of second rank cities, or regional capitals.These market towns are rarely interconnected lat-erally, or at all, to major road arteries. In manyparts of the country the topography and frequentdissection by rivers make both road constructionand maintenance extremely expensive. High andintense rainfall and uncontrolled runoff from de-graded lands cause floods and road washouts. Thedispersed population pattern leaves most of thepopulation more directly dependent on natural re-source–related livelihoods and therefore more vul-nerable to hydrological variability. The ruralsettlement pattern also compounds the challenge ofmitigating hydrological variability because a dis-persed population cannot be provided droughtand flood mitigation services (that is, water storageand river regulation) as cost effectively as moreconcentrated populations.

With 80 percent of the population in agriculture,weak performance in that sector slows demandthroughout the economy by diminishing anddestabilizing incomes and thus demand for non-agricultural production. Much of Ethiopia’s non-agricultural production relies on raw materialsfrom the agricultural sector, which can be in shortsupply during droughts. The unreliability of rawmaterials, particularly in combination with unreli-able water supplies, discourages prospective in-vestors.

Weak and unpredictable growth in non-agricul-tural opportunities keeps labor in subsistence agri-culture on fragile lands. This contributes to landdegradation as poor farmers extend cultivated

28 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources and Key Economic Sectors 29

areas without the means and incentives to invest inland stewardship and modern inputs, further low-ering agricultural productivity.

Manufacturing

Manufacturing is at a very early stage in Ethiopia,with less than a 7 percent share of the GDP.Processing of agricultural raw materials is the dom-inant manufacturing activity, with food and bever-ages, tobacco, textiles, clothing, leather, and woodaccounting for 65 percent of the manufacturingoutput in the 1990s. The share of food and beverageproduction in the manufacturing value added in-creased from 31 percent during the prereform pe-riod of 1983–4 to 1991–2 to 39 percent in the reformperiod of 1992–3 to 1999–2000.

The manufacturing and services sector is charac-terized by a weak enabling environment, low levelsof domestic and foreign investment, and low growthand productivity. Operations generally run well

below capacity due to the high cost and difficultiesassociated with transportation and marketing, andthe unreliable supply of raw materials and power.This uncertainty regarding supply and marketing,plus an inconsistent local demand for non-agricul-tural products, undermines investments.

A survey of public and private industries con-ducted by the Central Statistical Authority esti-mated that capacity utilization averaged 60 percentin public and 39 percent in private manufacturingenterprises in 1998–9. About 40 percent of the man-ufacturing enterprises claimed market-relatedproblems as the most serious constraint and 29 per-cent reported that a shortage of raw materials wasthe most serious constraint (CSA 2000). Shortagesof raw materials derive directly from the weaknessand unreliability of production in the agriculturalsector, which in turn is driven in large part by thevariability of hydrology.

Every major drought leaves hundreds of thou-sands of livestock dead and many more weak and

Table 3.8. Ongoing and Planned Hydropower Projects

Installed capacity List of projects Status as of 2005 (megawatts)

Ongoing construction projects (2005) Year of completionTekeze/TK_5 2008 300Gilgel Gibe-II 2008 420

Ongoing study projects Level of of studyBaro Hydropower Project Feasibility study 800Karadobi Multipurpose Project Prefeasibility study 1,700Genale GD-3 Multipurpose Project Feasibility study 245Genale GD-6 Hydropower Project Prefeasibility study 244Wabe Shebele WS-18 Multipurpose Project Feasibility study 87

Upcoming study/design projects (2005–10) Level of study/designBaro Hydropower Project Design 800Genale GD-3 Multipurpose Project Design 245Wabe Shebele WS-18 Multipurpose Project Design 87Karadobi Multipurpose Project Feasibility/Design 1700Genale GD-6 Hydropower Project Feasibility/Design 244Mendaia Hydropower Project Prefeasibility/ Feasibility 1,700Border Hydropower Project Prefeasibility/ Feasibility 1,780Tams Multipurpose Project Prefeasibility/ Feasibility 1,020Mabil Hydropower Project Prefeasibility/ Feasibility 1,440Birbir Hydropower Project Prefeasibility/ Feasibility 467Dobus Multipurpose Project Prefeasibility/ Feasibility 741Didesa Hydropower Project Prefeasibility/ Feasibility 308

Source: MoWR.

prone to disease. As a result, the cycle of droughtcontinuously diminishes potential growth in thelivestock population and compromises the qualityof the national herd. Poor quality of hides and skinshas contributed to low capacity utilization in theleather sector, which in turn has affected labor pro-ductivity. Labor productivity in the leather sectoris very low and declining (Ministry of Trade andIndustry 2003).

Large- and medium-scale edible oil enterpriseswere reported to exploit only about 20 percent oftheir production potential, mainly because edibleoil is imported through food aid. One nongovern-mental organization was reported to be selling ed-ible oil, which it received as developmentassistance from U.S. Agency for InternationalDevelopment, at 21 percent below the market price(Befekadu et al. 2000/1).

The performance of the non-agricultural sectoris also affected by power interruption in years of se-vere drought when water shortages cause inter-ruptions in hydroelectric power generation. Forinstance, during the drought of 2002–3 power wasunavailable throughout the country on one day aweek for about four months. Each interruption wasestimated to result in a loss of 10–15 percent of theGDP for that day (EEPRI 2003). According to theEthiopian Economic Policy Research Institute, thecountry appeared to be moving from rainfed agri-culture to “rainfed industry.”

Banking

Since the early 1990s, the banking sector has beenrestructured and liberalized to lay the groundworkfor private entry into the financial sector and toprovide for licensing and supervision of banks andinsurance banks. The number of banks increased tonine (six of which are private) and the number ofbranches expanded rapidly. Total deposits held bythe banks increased by 130 percent (from Br10.1 bil-lion in 1994–5 to Br21.1 billion in 2000–1). Loansdisbursed to the nongovernment sector increasedfrom Br5.4 billion in 1992–3 to Br17.4 billion in2000–1 (FDRE 2002).

However, a high rate of default has severely con-strained the activities of the banking sector. Theratio of nonperforming to total loans was estimatedat around 36 percent in September 2002 (WorldBank 2003). Poor lending practices, the war withEritrea, and fluctuations in agricultural prices (due

to variations in weather, leading to changes in sup-ply) have contributed to the deterioration of thebanks’ portfolios. Default rates on fertilizer loansare also high in rural areas, but the CommercialBank of Ethiopia (the only supplier of loans in re-cent years) is protected against such bad loans bythe regional governments.18

The banking system in Ethiopia is character-ized by excess liquidity, reflecting low return tocapital, high risk, and inadequate lending prac-tices. Loans to the agricultural sector account forless than 15 percent of total lending and arelargely composed of short-term fertilizer loans.The high risk of crop failure and fluctuations ofagricultural prices, together with a land policythat disallows the use of land as collateral, havemade agriculture unattractive for the financialsector. The industrial sector is also too weak(partly to the weakness in agriculture) to attractloans. Agricultural and industrial sector loans to-gether account for less than 30 percent of the totallending, while the service sector accounts for over60 percent (World Bank 2003).

Drought-induced default has affected the per-formance of microfinance institutions and othercredit programs in the rural areas.19 Agriculturalloans tend to default all at once during drought,making them unattractive (undiversified) portfo-lios for banks as well as for microfinance institu-tions. Borrowers have also been affected by the riskof low prices and depressed economic activitieseven in normal years.

The challenge is to lay an institutional and infra-structural foundation to protect agriculture and in-dustry from weather risks and thereby attractinvestment loans. Foreign investors cannot be at-tracted when the domestic banking sector is sick orweak and foreign banks are absent.

Tourism

Ethiopia has enormous potential as a tourism des-tination. Apart from the African game and culturalexperiences available to visitors, it has a rich arrayof historic and natural sites that are truly worldclass. A number of problems, including the highcost and poor quality of tourism services, have con-strained the development of the tourism industry.Ethiopian tourism promotion also needs to over-come the images of drought, famine, and war thatmany foreigners associate with the country.

30 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources and Key Economic Sectors 31

FOOD AIDFood aid has saved many lives in Ethiopia over theyears, and it cuts across other sectors, includingagriculture, transport, and manufacturing.However, food aid may actually be creating disin-centives for the types of agricultural investmentsthat could result in long-term food security andmore sustainable livelihoods. Until Ethiopia devel-ops the infrastructure and market incentives togrow and market its own agricultural produce, itwill remain highly vulnerable to the vagaries of theweather.

Safeguarding the lives and health of vulnerablegroups is the rational, overriding concern of na-tional drought relief programs. From a politicalperspective, it is often more palatable for donorcountries to provide food aid than financial aid, be-cause the donated crops are produced and pur-chased in their own economies. However, a systemof donated food aid that is relatively easily accessi-ble and not fungible creates incentives to rely onfood aid as a rational response to drought ratherthan investing fungible development funds inproactive drought risk management.

In practice, emergency aid or relief assistanceusually supplies specified goods, such as seeds,medicines, and trucks or specific emergency mea-sures such as repair of the water supply and sink-ing bore holes. Food aid, while greatly valued, isinflexible and usually cannot be redirected to re-construction or more general development supportat the end of the drought if a more rapid than an-ticipated agricultural recovery is achieved.

Food aid can also create perverse incentives. Theexpectation of food aid can discourage domesticprivate sector investment in marketable agricul-tural production. In years of shortage, when pricesare potentially highest, it is reasonable to assumethat food aid will arrive and both lower prices (bysupplying deficit regions and often leaking into the

broader economy) and drive up costs (by compet-ing for transport and distribution services).Continuous food aid also enables people to main-tain subsistence agriculture on marginal lands,thereby sustaining settlements and livelihoods thatare extremely vulnerable to water shocks and ad-vanced environmental degradation.

Food aid has a significant influence on the trans-port sector and on incentives for surplus agricul-tural production. When food aid arrives at the port,transport prices soar and trucks become unavail-able to national farmers marketing their product.During the 1984–5 drought, while overall GDP fell9.7 percent, transport and communication distrib-utive services rose by 8.7 percent, probably to an in-crease in the long-haul transport of food aid.

Since there is no coordination in the arrival ofships and domestic transport, and to the emer-gency nature of the food import, all available trucksare diverted to the Port of Djibouti to meet theships, leaving few to transport food from any re-maining productive agricultural areas to the fooddeficit areas.

A recurring problem has been the stop-go atti-tude toward drought planning: efforts to mitigatethe effects of drought often fall by the wayside oncethe rains return. National policies favor food aidover proactive contingency planning, ignoring theneed to build effective mitigation and responsestrategies at subnational levels.

Without the active involvement of affected localpopulations in planning responses, strategies tomitigate the effects of drought often ignore whatlocal people do to help themselves in times ofdrought, with the unintended effect of undermin-ing local coping strategies and weakening local so-cial capacity. As a consequence, planning fordrought is frequently centralized at the nationallevel, and early warning systems are largely de-signed to serve as tools for food aid planning as op-posed to active contingency planning.

33

The previous two sections have detailed how Ethiopia’s dependenceon rainfed agriculture creates a web of effects on agricultural output,industry, and power production, which are exacerbated by a lack ofboth hydraulic and market infrastructure. These impacts are trans-mitted through input, price, and income effects onto the broadereconomy. By creating continual uncertainty, the impacts of hydrol-ogy can be seen in investments and patterns of growth. Many of thesedynamics conspire against structural change in the economy.Endemic drought, fragmented and inadequate and infrastructure,and expectations that food aid shipments will disrupt transportationand tamp down food price spikes during shortages all promote a risk-averse subsistence economy and undercut incentives for farmers toproduce surplus, marketable crops.

This section looks at management options to address the complexsituation described above. Obviously, it is crucial to develop waterstorage infrastructure at all scales. But Ethiopia cannot fully mitigateagainst the impact of hydrological variability through water infra-structure investments alone. Clearly, it will be necessary to look be-yond traditional water resources management in order to curtail thenegative effects of hydrological variability on the Ethiopian economyby also identifying interventions aimed at decreasing the vulnerabil-ity of the economy to these shocks. Unless the structure of the econ-omy changes, stable economic progress will continue to be bothintensely disrupted by drought and flood and quietly undermined byrisk aversion.

STRENGTHENING WATER RESOURCESDEVELOPMENT AND MANAGEMENTOpportunities exist to significantly strengthen water resources de-velopment and management practices. Attention must be given to in-terventions that moderate hydrological variability, mitigate thephysical impacts of hydrologic shocks on the environment and wa-tersheds, and arrest land degradation. Such interventions will in-clude infrastructure investments to provide storage and regulateriver flows and runoff and land and soil management practices to en-hance the natural capacity of watersheds to moderate hydrology.

Water ResourcesChallenges: Some

Priority Responses

4

Strengthening Institutions and Capacity

Water resources management is a core issue for de-velopment in Ethiopia. In all countries, informationand human and institutional capacity are essentialfor effective water resources management. GivenEthiopia’s challenging hydrology, the need for suchcapacity is great, but capacity in the country is low.Efforts to strengthen capacity are ongoing inEthiopia and should be seen as a continued priority.

Investment in research and education in partic-ular will be crucial to develop the capacity neededto design effective and appropriate water resourcemanagement interventions, watershed manage-ment, irrigation schemes, and water and sanitationservices. An increased focus on groundwatershould also be seen as a priority in this regard.

Increased Water Storage

To mitigate against the economic impacts of watershocks in Ethiopia, greater water storage capacity,both natural and manmade, large scale and smallscale, will be needed. Given both seasonal and in-terannual variability, significant over-year storagewill be particularly important. Storage design musttake explicit account of the country’s extreme hy-drologic variability and high sedimentation rates. Awide range of infrastructure investments is needed.

Public investments are needed to provide thewater security and incentives necessary to promoteprivate sector investment in inputs that will en-hance productivity. This is often described as an“input impact” on patterns of investment for sur-plus production.

Large-Scale Water Storage

Artificial reservoir storage in Ethiopia is currentlyestimated to be about 43 cubic meters per capita incontrast to 750 cubic meters per capita in SouthAfrica and 6,150 cubic meters per capita in NorthAmerica (with most developed countries having fig-ures in the thousands) (see Figure 4.1). Ethiopiawould need to invest five times its annual GDP(US$35 billion) to achieve the water storage capacityof South Africa, taking such capacity as an arbitrarybenchmark of water security. This order-of-magni-tude calculation demonstrates that strategies fo-cused purely on water infrastructure responses arenot affordable, so a number of other approaches arealso outlined in this section.

Any hydrologic infrastructure developed forEthiopia must be designed to meet multiple needs.Reservoirs can regulate the variability of surfacewater in streams or rivers. They can retain excesswater from periods of high flows for use during pe-riods of drought and store floodwaters to reduceflood damages downstream. Reservoirs can deliverirrigation water; generate hydropower; slow sedi-mentation downstream; create fisheries; providetourism opportunities; help manage watersheds;control flooding; mitigate drought; and providedrinking water, local transport, and recreation.

In Ethiopia, many natural sites are well suited tocreate storage reservoirs at relatively low costs andwith relatively little environmental and social dis-ruption. In the design and economic analysis of po-tential hydropower development, the benefitsassociated with drought and flood protection, withsedimentation control, and even with navigationshould be—but rarely are—explicitly considered infiguring the economic rates of return.

The Koka reservoir is an good example of a mul-tipurpose reservoir whose construction made it pos-sible to develop about 50 medium- and large-scaleirrigated farms totaling 65,000 hectares distributedalong the valley from Wonji (80 kilometers southeastof Addis Ababa) to Asayita, and to generate 110megawatts of hydropower at three run-of-the-riverpower stations. However, severe soil erosion is lead-ing to rapid loss of Koka’s storage capacity.

When reservoirs serve multipurpose uses, com-plex operational tradeoffs arise and are most acutewhere hydrology is variable and droughts andfloods endemic. For example, reservoir levels mustbe kept low enough to provide flood retention ca-pacity while at the same time full enough to con-serve water in the event of a drought. Similarly,water releases for power generation must beweighed against requirements for irrigation, all inline with system requirements such as environ-mental flows. Issues such as these are commonlyresolved through the adoption of reservoir operat-ing rules agreed upon by the various stakeholders.

Given the shared nature of most of Ethiopia’srivers, this broad range of potential benefits andpotential positive and negative externalities needto be explored in a transboundary context.

Small-Scale Water Storage

Water harvesting at the household level is an im-portant strategy to bridge water deficits and assistin addressing food insecurity. However, harvest-

34 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources Challenges: Some Priority Responses 35

ing structures must be designed in accordance withintended usage and local circumstances, if rainwa-ter harvesting is to be cost effective. The amount ofwater needed to sustain household agriculture indrought years is substantial.

There is an ongoing debate about the currentwater harvesting program, which involves con-struction of a household water storage structure(commonly a concrete-lined pit with a capacity of40 cubic meters of water, which will irrigate about100 square meters of land) intended for life-savingirrigation for field crops as well as for vegetablegardening. One writer has claimed (Bezabih 2003)that the structures cost an individual farmer a min-imum of Br1,500 (a little less than US$200).However, the return is only Br40–50 per annum.Nevertheless, farmers have reported using thesmall amount of water in their structures to growvegetables for home consumption and in somecases for market sales. The amount of water col-lected can also be increased in some cases.

Water harvesting in the form of runoff manage-ment for supplemental irrigation may have signif-icant potential to enable double cropping. Forexample, Vertisol20 areas near major cities (for ex-

ample, Ada and Becho areas around Addis Ababa),which currently grow mainly teff,21 could use waterharvesting to add a crop of vegetables during thelong dry season. Water harvesting could be partic-ularly successful in these areas since the tempera-ture is cooler and seepage is lower.

Water Supply and Sanitation Services

The first priority for water is for human and live-stock consumption. The benefits of improvingclean water supply and sanitation services are wellknown. The potential economic impacts includedecreased incidence of waterborne diseases, whichpromotes better health and greater productivityand incomes; savings on medication; and time sav-ings, particularly for women and girls, which couldallow for more study or opportunities for incomegenerating activities.

Since the setting of wastewater goals and stan-dards is the responsibility of river basin authoritiesand other public water resources managementagencies, planning will require close collaborationbetween utilities and these bodies. The clear needfor investments in infrastructure will require

Water availability versus storage

0

200

400

600

800

1000

0 1,000 2,000 3,000 4,000 5,000

storage/capita (m3)

wit

hd

raw

als

per

cap

ita

(m3)

Ethiopia

S. Africa

SpainAustralia

Ethiopia

S. Africa

SpainAustralia

Figure 4.1 Water Storage Per Capital in Selected Countries

Source: Compiled from SIMA and KOID databases.

accompanying reform, much of which is under-way, to focus on the financial sustainability of util-ities, to make them more accountable for theirperformance, and to integrate sanitation planningand hygiene promotion with improvement ofwater supply.

In some cases, effective urban water supplyplanning may require concomitant investments inwatershed management, bulk water infrastructure,or river regulation infrastructure. In these cases, se-quencing onsite and offsite investments will be im-portant. These decisions should be made in tandemwith local demand and the capacity of local insti-tutions. Benefits of water and sanitation services,such as cleaner downstream water, should be con-sidered explicitly in planning.

Irrigation and Drainage

Appropriately designed irrigation investments canprovide a secure supply of agricultural water tobuffer production during all but the most extremeclimatic events and protect the subsector from thegreater part of hydrological variability. It can en-hance food security and the reliable delivery ofmarketable and exportable agricultural products,helping to stabilize prices and encourage invest-ment in both the agricultural and non-agriculturalsectors.

Ethiopia has significant potential for irrigatedagriculture. Critical to its development is the activeparticipation of the private sector in small-,medium- and large-scale enterprises. Attractingprivate investment requires, in addition to access torelatively reliable water resources at reasonablecosts, the appropriate policy environment for in-vestors, security of land tenure, and the necessaryaccess to financial resources.

Investment in irrigation should also be accom-panied by attention to water conservation, soil con-servation, soil conditioning, and soil productivity.Crops to be grown under irrigation should be care-fully selected to maximize the return from invest-ment. Most cereals are not as profitable as cashcrops (cotton, sugarcane, and so on), horticultureproducts (fruits and vegetables) and floricultureproducts. Developing fisheries in lakes and reser-voirs could provide additional income and food forthe population.

Potential for Irrigation From Groundwater

Ethiopia’s groundwater resource potential is esti-mated at 2.6 billion cubic meters. One distinct ad-

vantage of groundwater is its relative reliability be-cause it is buffered from rainfall variability. TheEthiopian Geological Survey plans to conduct anassessment of groundwater for sustainable ex-ploitation and determine its location and extent.

Groundwater availability estimations indicatethat a significant proportion of this resource couldbe economically used for irrigation and that tube-well irrigation is likely to be a cost-effective option.These wells, ranging from small diameter on-farmsingle farmer–managed to large community-owned wells, may also be used for domestic waterby rural households. Tubewell irrigation can sup-plement canal irrigation as well as irrigate areasthat do not have canal irrigation facilities.

Affordable water-lifting devices, such as smalldiesel or electric pumps or muscle-powered treadlepumps, have dramatically improved poor people’saccess to groundwater in Bangladesh, easternIndia, and Nepal. For smallholders, the capital re-quirements to develop groundwater irrigation aregenerally low and its productivity is higher com-pared to surface irrigation (Sally 2002).

Drainage Improvement to Protect SoilProductivity

Where drainage is inadequate, the surface runofffrom irrigation may accumulate in vertisol fieldsand cause waterlogging, which can lead to soildegradation. Management practices should be de-veloped to minimize water accumulation on thesoil surface in vertisol areas. Subsurface drainagetechniques could be costly but effective and longlasting. Other options such as camber beds,broadbed, and furrows and open ditches may beless costly but are only short-term solutions.

Watershed Management

Improvements in watershed management will be acrucial element in managing water resources and,more broadly, riverine ecosystems. Improved wa-tershed management can slow watershed erosion,moderate the hydrological cycle, regulate runoffand groundwater, improve infiltration (hencewater retention and base flows), and reduce poten-tial flood damage. Good watershed managementwill help minimize the impacts of drought andflood on agriculture and sustain agricultural pro-ductivity. It can also reduce soil erosion and sedi-mentation rates, thereby extending the live-storage(useable) capacity of reservoirs, reduce water treat-

36 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources Challenges: Some Priority Responses 37

ment plant costs, and diminish the costs and in-crease the life of irrigation infrastructure.

Increasingly, reservoir operators recognize thatsubstantial amounts of storage are lost each yeardue to sedimentation, and that sediment manage-ment in both watersheds and dams is vital formaintaining asset value. Programs for integratedwatershed management will improve the sustain-ability of hydropower production, particularly inwatersheds upstream of the existing and plannedreservoirs. Hydropower providers are fully awareof the impact of sedimentation on reducing the ef-fectiveness of the assets and are increasingly advo-cating sound watershed management practices toreduce sedimentation.

An important lesson from experience in Ethiopiais the need to include stakeholders effectively in theplanning, design, and management of watershedinterventions. This includes extensive consultationwith stakeholders, ensuring that resettlement isdone well, investing in community management ofwatersheds, and ensuring that local people becomebeneficiaries. A strong positive feedback loop inthis arrangement is that the development of hy-dropower to provide electricity will meet the en-ergy needs of more of the population, therebyreducing the demand for biomass and reducing thepressures on the watersheds.

Drought Preparedness

Adjustments to minimize agricultural losses dur-ing a drought include the development of localwarning systems, grain storage arrangements, se-lection of climate-appropriate crops, developmentof a flexible mixed crop and livestock economy,and a system for supplementing income throughsecondary occupations and seasonal migration foremployment. The importance of each of these re-sponses and their degree of success will depend onthe present system of land use in any area, the typeof soil, the local level of technology, the economiclinkages, and the social milieu. However, the re-sponse must also depend upon what communitiesand individuals perceive as the options open tothem (Sally 2002).

Early Warning Systems

Effective early warning systems are a critical com-ponent of any drought preparedness program, withmajor benefits for agricultural and hydropower pro-

duction. However, there is currently a deficiency ofreliable meteorological data and observation net-works, as well as an inadequate information deliv-ery system. The country must further develop anumber of drought preparedness tools by strength-ening early warning systems, market informationsystems, and permanent diagnosis and research cen-ters (agriculture, hydrology, meteorology). Thesetools can be used to collect and analyze agriculturalproduction data in order to identify and target po-tential risk areas, thus allowing agencies to initiateappropriate urgent operations.

Livestock Survival

To improve drought preparedness in the livestocksector, investments are needed in small feederroads, improved water management, fodderbanks, and range improvement. Funds should alsobe set aside to provide rapid responses such as de-stocking, water tankering, and the control ofhuman and animal disease outbreaks.

ENHANCING ECONOMICRESILIENCEThe following sections offer options for proactiveinterventions outside the water sector to enhancethe economy’s resilience to the impacts of hydro-logical variability.

Encourage Alternative Livelihoods

Most Ethiopians depend directly on rainfall forsubsistence agriculture. Thus, Ethiopia’s broadereconomy is built on domestic demand that is as un-predictable as the rains. Investments and policiesmust be considered to create opportunities for lesswater-dependent, more resilient livelihoods.

In a large part of the highlands, the rapidly grow-ing population is outstripping the carrying capacity.As more and more people collect fuelwood and ex-pand their agricultural holdings, the critical uplandareas of watersheds are becoming deforested andprone to erosion. Ultimately, population planning,community action, and alternative livelihoods willbecome the keys to the restoration of highland wa-tersheds. Interventions to provide alternative en-ergy sources, sustainable fuelwood production, andalternative livelihoods (either outside of traditionalextensive agriculture or outside of sensitive water-sheds) would not only improve livelihoods but

would also help to preserve forest cover and soilsand moderate the hydrological cycle.

The most vulnerable livelihoods are those re-liant on rainfed agriculture in fragile and degrad-ing watersheds. It is always the poorest farmerswho expand production on the most vulnerableland—with the steepest slopes and thinnest soils—and who have the least capacity to invest in landmanagement and soil conservation. Investments inirrigated agriculture, resettlement outside degrad-ing watersheds, and off-farm employment wouldboth protect vulnerable populations and slow landdegradation, which in turn could moderate the im-pacts of hydrological shocks.

Seek Price and Income Stabilization

With 85 percent of Ethiopia’s population depen-dent on agricultural incomes, shocks to farm pro-ductivity must be reduced. Better road and marketinfrastructure for agricultural production coulddeepen the domestic market and help to smoothfood supplies and prices. Expanded opportunitiesfor off-farm employment would also help lessenthe economic shocks arising from fluctuations inthe agriculture sector by providing supplementalincome to a greater share of the population.

Furthermore, the prevalent assumption thatemergency aid should be food for work or food-based income-supplementing programs should bechallenged. In increasingly better integrated andmore complex systems, cash for work, foodcoupons, and other cash transfer mechanisms maybe more efficient than food-based interventions,and help to smooth incomes in rural drought-prone areas.

Enable Private Investment in Irrigation andin Non-Agricultural Activities

The historical development focus on rural rainfedagriculture has served to sustain a highly vulnera-ble agrarian economy. Investment policy in irri-gated agriculture and energy sectors need to befurther liberalized to attract both domestic and for-eign private investors. The investment climate andproperty rights regimes should ensure long-termsecurity and adequate return on investments.Investors must be assured that land tenure andwater use rights are flexible enough to promotecomparative advantage in food staples and cashcrops. Ownership and transfer rights in land

should guarantee sufficient incentives to inducelong-term investment in Ethiopia.

These rights must be matched by access to ruralcredit and finance and dissemination of technologyand good practices in water use. Strategies areneeded that combine formal irrigation systemswith informal pro-poor systems and affordabletechnologies, such as small-scale irrigation, waterharvesting, and tubewell irrigation. Agriculturalextension and credit policies need to be geared to-ward creating easy access to small-scale, affordabletechnologies. In addition, policies need to promoteaccess to affordable power (likely hydropower)once the necessary infrastructure is put in place.

Invest in All-Weather Roads and Market Infrastructure to ShiftExpectations and Incentives

To mitigate against the economic impact ofEthiopia’s endemic flooding and high erosion rates,all-weather roads should be a priority interventionin affected areas. Improving the coverage and qual-ity of Ethiopia’s road network will make the econ-omy less vulnerable to disruptions caused by watershocks, ensuring that producers can move theirproduct to market. This would diminish current dis-incentives to investment in all sectors that arise fromthe high costs and uncertainty of transport services.

It is particularly important for the agriculturesector that roads are dependably passable and ad-equate storage and marketing facilities are avail-able in order to limit the fluctuations in marketedoutputs, reducing risks, and enhancing expecta-tions. Better functioning markets will provide in-centives for investment both within and beyondthe agriculture sector.

Invest Strategically in Interregional Roads

Interregional roads could promote interregionaltrade, and offer marketing and transport alterna-tives that would mitigate the economic impacts ofwater-related transport disruptions. The existingpattern of roads radiates from the capital, AddisAbaba, to major towns, while adjacent regionsoften lack links between them. Such interregionalroads, if they existed, would facilitate the growth oftrade between regions, bring surplus production todeficit areas, help smooth weather related priceshocks, and promote a more integrated nationaleconomy and greater national unity.

38 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources Challenges: Some Priority Responses 39

Ensure That Other Infrastructure Investments are Designed to WithstandErosion and Flooding

Where high rates of sedimentation are seen or pre-dicted, hydraulic infrastructure such as reservoirsand canal systems should be designed with ade-quate sediment capture capacity. Treatment sys-tems should be budgeted adequate funding foroperation and maintenance.

Expand Hydropower Generation Capacity

Ethiopia has enormous potential for hydropowerdevelopment; that is, national water resources areestimated to have the potential to generate as muchas 30,000 megawatts of power. Hydropower pro-duction, and hence the economy, can be better in-sulated from the effects of water shocks byincreasing installed generation capacity, which is amedium-term goal of the energy strategy and of theWater Sector Development Plan (MoWR 2002b).

The hydropower development plan aims to pro-mote the achievement of national socioeconomicgoals through efficient and sustainable develop-ment of water resources to produce hydro-electric-ity. Reliable and sufficient supplies of electricitywill spur investment. A grid of spatially disbursedgeneration facilities would improve the reliabilityof power supplies for the economy.

Access to electricity as an alternative to the useof biomass fuels will have a significant impact onthe land and forest resources of the country, boost-ing environmental sustainability and agriculturalproductivity. To the extent that hydropower can besubstituted for biomass fuels, women in particularwith benefit from such a shift. In addition, hy-dropower infrastructure can be designed and op-erated to mitigate against flood, drought, andsediment transport, all of which have major eco-nomic impacts in Ethiopia.

Promote Regional Power Trade

Ethiopia’s vast hydropower potential offers oppor-tunities for mutually beneficial power trade in theregion. While current national demand may be in-adequate to justify rapid expansion of hydropowercapacity, regional power trade offers a significantopportunity to make large-scale projects economic.The availability of cheap power would help makethe Ethiopian economy more competitive. In addi-

tion, the revenues earned from regional powertrade could help the economy develop.

By developing a network of linked energy pro-ducing facilities, Ethiopia can increase economic re-silience to water shocks. For example, two or moreinterconnected hydropower-based systems mayprove complementary in reducing the effects ofseasonal river flows, interannual flows (includingdroughts and floods), and differences in reservoircapacity. Whereas thermal generation has highvariable operating costs relative to hydropower,savings can be achieved by using hydropowerwhen it is available.

Likewise, hydropower can generate power dur-ing excess flows for thermal plants and receivepower from thermal plants when reservoirs are de-pleted. Power plants can benefit from mutual as-sistance during maintenance or forced outages.Reservoirs on an interconnected system could beoperated in such a way that each power plant (orcountry) might not need large reserve capacity.Economies of scale might be attained in new gen-erating capacity as markets improve.

Regional cooperation in power productioncould strengthen regional relations and promoteincreased economic cooperation and integration.Ethiopia is currently exploring opportunities for re-gional power trade with Sudan and Egypt throughthe NBI and with Djibouti. In particular, coopera-tive power investments with Egypt and Sudan,producing sufficient energy to meet Ethiopia’s do-mestic demand and to justify transmission toSudan and Egypt, could potentially ease long-standing historical concerns arising from the threecountries’ dependence on the Nile. The economicrate of return of supplying power to Sudan andEgypt has been estimated to be 32 percent (MoWR2002a). The EIRR for providing power to Djiboutiis lower at around 20 percent (Demeke 2004), butstill financially very attractive.

Ethiopia could potentially transmit powerthrough Egypt’s grid to the Middle East and evento Europe. As a clean, renewable source of power,hydroelectricity may be a very attractive alterna-tive to diesel- or coal-fired and nuclear-poweredsources of energy for Europe. In addition, develop-ing this renewable energy might make Ethiopia el-igible for carbon credits.

In addition to hydropower, cooperatively con-structed dams could have other downstream ben-efits. They would regulate sediment transport thatis washed from Ethiopia’s denuded hillsides

through irrigation infrastructure in Sudan to LakeNasser in Egypt. Unless the runoff is controlled,silting will severely damage downstream reser-voirs (Thurow 2003). In addition, should the coun-tries work cooperatively on operating rules, itmay be possible to store Nile waters in the cool cli-mate of the Ethiopian highlands and provideevaporative savings in the river system, therebyincreasing the total volume of water available tothe riparians.

Diversify and Export Manufactured Goods

Expansion of reliable electricity coverage willbroaden opportunities for other economic invest-ments and likely promote non-agricultural activi-ties such as industry and services that are lessvulnerable to water shocks. Similarly, improvingthe cost and reliability of transport will enable thesuccess and expansion of manufacturing and ser-vices. Increased exports of manufactured goodscould help insulate urban manufacturing from thenegative effects of diminished local demand duringtimes of drought. Increasing the resilience of theagriculture, energy, and transport sectors willtherefore also improve the resilience of urban pro-duction.

Promote Financial Risk Mitigation and Credit

Current credit and banking systems do not providerobust risk mitigation products. Whatever the ex-tent of investment made in hydrologic infrastruc-ture and water resources management, Ethiopia’shydrologic variability is so extreme that it willlikely always pose significant risks for agriculturaland non-agricultural investors. Japan, for example,suffers flood damage each year despite multibilliondollar expenditures on structural flood protection.It is therefore important to explore opportunities toprovide insurance, credit, or banking vehicles tohelp producers manage this risk. Ethiopia has justlaunched a highly innovative pilot program inweather risk insurance.

Shift Assets From Livestock to Rural Banks

Storing wealth in livestock increases rural eco-nomic vulnerability to weather shocks. Duringdroughts, when incomes collapse, livestock assetvalues also decline. The correlation of these inci-dents increases the vulnerability of the population,and it is clear that in a drought cycle the loss of as-sets marks a turning point toward poverty.“Drought proof” stores of wealth and related toolssuch as credit and weather insurance should be ex-plored to provide asset stability during droughtepisodes. Furthermore, smaller livestock herdswould require less drinking water, diminish thepotential for animal waste-related waterborne dis-eases, and slow degradation of the landscape.

40 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

41

To help clarify and quantify the impacts of Ethiopia’s current waterresources variability, and to explore potential impacts of mitigatinginterventions including investments in irrigation and market infra-structure described in the previous section, we developed a computermodel that examines three proposed interventions under three rain-fall patterns.

The results of the model should provide a better understanding ofthese dynamics and inform efforts to formulate and develop inter-ventions that will enhance growth and insulate the Ethiopian peopleand economy from the often-devastating economy-wide effects ofwater shocks.

A complex series of model runs provides the following insights forpolicymakers:

• Standard economic models, which assume average rainfallsover a long period, are inappropriate for Ethiopia because theyfail to capture the significant economic impact of the country’shydrological variability. A model that accounts for this vari-ability projects a much bleaker—but more realistic—picture ofthe country’s economic future.

• Investments in irrigation appear much more profitable whenhydrological variability is taken into consideration.

• Investments in irrigation and market infrastructure (roads, stor-age facilities, and so on) appear to spur similar growth impacts,and there appear to be synergies between these types of invest-ments, particularly under conditions of hydrological variability.

• Investments in market infrastructure will also help non-agri-cultural sectors of the economy, thus promoting a structuralshift in the economy toward greater diversification and hydro-logical resilience.

VARIATIONS OF THE MODELThe first task in developing an effective quantitative economic modelfor Ethiopia was to find a way to incorporate the country’s extremelyvariable hydrology, which has such a great influence on its economy.Most economic models project average growth rates in agriculture,which implicitly assumes average, steady, and constant rainfall over

A Hydro-Economic Model of Ethiopia:

A Summary of Findings

5

time. This type of model misses the hydrologicalshocks and uncertainties that plague Ethiopianfarmers. We tested three variations of our model asdescribed below. All variations are regional, multi-sectoral, multimarket models that incorporate hy-drology within the agricultural productionfunction. They differ only with regard to their as-sumptions of rainfall variability.

The model does not incorporate groundwater,hydropower or the dynamics of food aid, and thatdata are always imperfect. For the assumptionsused in the model, as well as more detail on thevariations, see appendix 1.

• Variation 1: Smoothed rainfall.22 This variationassumes 1995–2002 average rainfall in allyears. This type of model is used in virtuallyall economic modeling. Rainfall is implicitlyincorporated into models through the as-sumption of average agricultural growthtrends, because the assumption of averageproduction implies average annual rains (orirrigated agriculture.) However, this assump-tion is unrealistic virtually everywhere. In thecase of Ethiopia it proves to be drastically mis-leading.

• Variation 2: Stylized drought. This determinis-tic variation of the model seeks to capture theimplications of drought by incorporating astylized, two-year drought (Diao et al. 2004).

• Variation 3: Historical variability. The final vari-ation is a stochastic extension of the basemodel that attempts to more fully reflectEthiopia’s historical levels of rainfall variabil-ity. Rather than using a smoothed hydrologywith a single shock event, this model capturesannual fluctuations whose associated cumu-lative losses affect growth performance andpotential over time. Where the smoothedrainfall and the stylized drought models as-sume that the yield function plateaus at a cer-tain level of rainfall, this model showsdeclining yields when rainfall is greater thana 50-year flood level, so that the negative im-pact of excessive rains (waterlogging andfloods) is also captured, both for agriculturaland nonagricultural sectors. This variationprovides a finer, less stylized—and truer—picture built on Ethiopia’s hydrologicalrecord, yet is still quite conservative (Strzepeket al. 2004).

IMPACTS ON GROWTH AND POVERTYIn the absence of increased investment, the modelindicates that the national economy and the agri-cultural sector will grow at a lower rate than pro-jected population growth through the year 2015.Assuming historically average trends in agricul-tural production and without incorporating anyhydrological variability, per capita incomes willfall and 11 million more Ethiopians will be living inpoverty by the year 2015.

However, the situation is probably much worse.The two models that incorporate the country’s vari-able hydrology show a bleaker—and probablymore realistic—future. The impact of a single, two-year drought within the model’s 12-year time hori-zon will lower the projected average rate of GDPgrowth by 10 percent and will mean another 5 mil-lion Ethiopians will live in poverty, 16 million morethan today.

When historical levels of rainfall variability, andthe impacts of both drought and flood are conserv-atively captured in the model, projected GDPgrowth rates decline by 38 percent and the numberof people living in poverty to increase by 25 per-cent, for a total of over 51 million.

Figure 5.1 clearly demonstrates the extraordi-nary impact of drought and, particularly, hydro-logical variability, on the Ethiopian economy. Asingle drought in a 12-year period will decrease av-erage GDP growth rates 7–10 percent. If historicallevels of variability and the partial impacts offloods are incorporated, GDP growth rates fall20–43 percent. Figure 5.2 shows that the impact ofa single drought will increase poverty rates 12–14percent, and that hydrological variability of themagnitude Ethiopia has historically experienced,raises poverty rates 25–35 percent.

The historical variability variant of the modelbest reflects the real hydrological impacts enduredin Ethiopia. It captures year-on-year stochasticvariations, and impacts of floods as well asdroughts.

The historical variability model provides a rangeof results rather than a single result for each simula-tion. Figure 5.3 shows this range, or ensemble, of re-sults in green. While this ensemble does notencompass all possibilities, the wide range of annualresults for the 12-year series suggests potential cata-strophic scenarios. Figure 5.3 compares GDP growth

42 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

A Hydro-Economic Model of Ethiopia: A Summary of Findings 43

rate projections using the smoothed rainfall model(the red line), and the mean of the historical vari-ability model ensemble (blue). From this graph it isclear that as greater hydrological variability is cap-tured significantly lower growth rates are projected.

When poverty rates are compared (figure 5.4),the results are even more dramatic. The historical

variability model (mean in blue) shows signifi-cantly higher levels of poverty. Again, the widerange of green line scenarios suggests potentiallydevastating poverty rates.

These models clearly demonstrate the criticalimportance of explicitly incorporating hydrologicalvariability in modeling the Ethiopian economy.

0.00

0.50

1.00

1.50

2.00

2.50

3.00

3.50

GDP Growth rate Ag GDP Non-Ag GDP

An

nu

alR

ates

Smoothed DroughtVariability

Figure 5.1 Effects of Hydrology on GDP Growth

Source: Authors.

20032015

Smoothed

Drought

Variability

29.12

51.07

29.12

45.67

29.12 40.79

0102030405060

Figure 5.2 Effects of Hydrology on Poverty Rates

Source: Authors.

Assumptions regarding hydrological variabilityhave a considerable impact on growth and povertyprojections, and the level of variability is negativelycorrelated to economic performance; that is, thegreater the level of variability modeled into the sys-tem, the worse the growth and poverty outcomes.Simply put, models of the Ethiopian economy thatfail to incorporate hydrological variability will bepoor predictors of growth and poverty.

INVESTMENT SCENARIOSNext we modeled three investment options: in-vestment in irrigation, in market infrastructure,and in both irrigation and market infrastructure.We investigated each investment scenario undereach of the three hydrological variations de-scribed above, producing nine scenarios. This sec-tion compares the results of the three investment

44 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Yearly Mean and Variable Climate GDP Base Simulation

3.85

3.90

3.95

4.00

4.05

4.10

2003 2006 2009 2012 2015

Simulation Year

GD

P(l

ogm

illi

onU

S$)

Variable Climate Sets Means Variable Climate mean

Figure 5.3 Projected Log GDP Using Smoothed Rainfall and Variable Rainfall Means andVariable Rainfall Ensemble

Source: Authors.

Yearly Mean and Variable Climate Poverty Rates Base Simulation

35

40

45

50

55

60

65

70

2003 2006 2009 2012 2015Simulation Year

Pov

erty

Rat

e(%

)

Variable Climate Sets Means Variable Climate mean

Figure 5.4 Projected Poverty Rates Using Smoothed Rainfall and Variable Rainfall Means

Source: Authors.

A Hydro-Economic Model of Ethiopia: A Summary of Findings 45

scenarios to the baseline scenario described ear-lier, and describes their differing results undereach of the three hydrological variations, all interms of their impacts on GDP growth andpoverty rate reduction. Tables 5.1 and 5.2 give acomparison of the overall results. The followingsection examines the results of each investmentscenario using only the historical variably rainfallmodel, deemed the most realistic in making plan-ning projections for Ethiopia. Additional technicaldetails regarding the investment scenarios can befound in Appendix 1.

IRRIGATION INVESTMENTSCENARIOThe effect of irrigation is modeled as the yield dif-ferential between irrigated and rainfed crop pro-duction. The irrigation investment scenario usesthe GoE’s irrigation plan, which calls for moderateirrigation expansion. According to the 2002Irrigation Development Program of the WaterSector Development Plan, the GoE plans to doublethe current irrigated area of grain from the current137,000 hectares to 274,000 hectares and triple theirrigated area of cash crops from the current 63,000

Table 5.1. GDP Growth Rates Under Differing Investment Scenarios

GDP Growth rates (2003–2015)

Percent change Percent Percent from baseline Stylized change from Historical change from

Smoothed with smoothed single baseline variability baseline with Model assumptions rainfall rainfall drought with drought (stochastic) variability

ScenariosBaseline 2.87 2.57 1.78Irrigation Investment 3.43 19.51 3.09 20.23 2.73 53.37Infrastructure Investment 4.42 54.01 2.53 42.13Irrigation and

Infrastructure Invesment 4.56 58.89 4.25 65.37 3.43 92.70

Source: Authors.

Table 5.2. Poverty Rates Under Differing Investment Scenarios

Poverty (percent of population living in poverty 2015)

Percent change Percent Percent from baseline Stylized change from change from

Smoothed with smoothed single baseline Historical baseline with Model assumptions rainfall rainfall drought with drought variability variability

ScenariosBaseline 45.89 51.38 57.45Irrigation Investment 42.52 -7.34 47.9 -6.77 53.18 -7.43Infrastructure Investment 39.39 -14.16 54.39 -5.33Irrigation and

Infrastructure Invesment 37.92 -17.37 43.43 -15.47 50.42 -12.24

Source: Authors.

hectares to 200,000 hectares by the end of the 15-year planning period. This would increase the totalirrigated area to approximately 470,000 hectares.Of the increased irrigated area, half (about 130,000hectares) are to be small-scale projects, equally di-vided into short-term (2002–6), medium-term(2007–11), and long-term (2012–6) projects. Theother half (about 147,000 hectares) are large- andmedium-scale projects, most of which will be com-pleted by 2015–7. Note that more recent govern-ment targets are more ambitious, aiming tocomplete 100,000 hectares by 2006 and develop750,000 hectares by the end of the planning period.

The program estimates the average cost of large-and medium-scale irrigation development at aboutUS$7,000 per hectare. It assumes that small-scale ir-rigation development costs are approximatelyUS$4,800 per hectare. The costs of other inputs,such as the fertilizers and improved seeds that areoften needed to increase yields on the irrigatedland, were not included in the model estimatesowing to a lack of available information. Thus, thereturns to irrigation investments are potentiallyoverstated. On the other hand, the 12-year timehorizon does not capture the full benefits of later-year investments. All irrigation assets can be as-sumed to have a functional life considerably longerthan 12 years.

Under the smoothed rainfall model, significantincreases in both GDP (6.7 percent) and agricul-tural GDP (3.5 percent) are projected. The totalpopulation living in poverty is projected to beabout 37.8 million by 2015, 3 million less than thebaseline scenario projections, but still 9 millionmore than the current level (see Tables 5.1 and 5.2).

In the stylized drought scenario, the benefits ofirrigation investments increase. Investment in irri-gation saves 3 million people from falling intopoverty. Although projected returns on irrigationinvestments in cash crops do not change signifi-cantly, returns to irrigation investments in cerealproduction more than double so that investmentsin small-scale irrigation could be recovered overthe 12-year period.

Under conditions of historical variability, theprojected benefits of irrigation are even greater. A53 percent increase in the GDP growth rate is pre-dicted under this model, compared with a roughly20 percent increase under both the smoothed rain-fall and stylized drought models. This result isfairly intuitive because it is at the greatest extremes

of drought and flood that the returns to irrigationinvestment are greatest. Not only does the exclu-sion of rainfall variability bias all growth projec-tions, but it especially underestimates returns toinvestments in irrigation. Poverty results under thethree variants of the model were similar.

Market Infrastructure Investment Scenario

This scenario evaluates the potential gains frompublic investment in market infrastructure, includ-ing transportation and storage facilities, to reducethe farmer’s marketing costs. Transportation in-vestments can enable interregional trade, whichcould strengthen food security during droughtsand floods, lessen price shocks, sustain servicesyear round, and promote structural changes in theeconomy toward less water-vulnerable activities.

We assumed that investments in roads and othermarket infrastructure could improve the servicesector’s productivity and reduce marketing mar-gins between producers and consumers. We esti-mated this as a 1.7 percent annual increase in theservice sector’s productivity, which lowers themarketing margins steadily over time. In total, pro-ductivity in the service sector is projected to rise by20 percent in 12 years, with marketing marginsfalling about 1 percent per year.

Under the smoothed rainfall model, total GDPincreases 19.6 percent above the no-investment sce-nario. The total poverty ratio falls to 39.4 percent,compared with 46 percent in the baseline scenario(5.8 million fewer people) and 42.5 percent (2.8 mil-lion fewer people) in the irrigation investment sce-nario (see Table 5.2).

Under the smoothed rainfall model, improve-ments in transport and market infrastructure de-livered more than twice the growth rate ofinvestments in irrigation. But under the historicalvariability model, the impacts of the two invest-ment scenarios were similar.23 This is explained bythe fact that the historical variability model on onehand captures far greater returns to investments inirrigation but, on the other hand, captures dimin-ished returns to roads and non-agricultural activi-ties because it takes into account the damage doneby flooding. Under the variability model, invest-ments in market infrastructure will increase therate of GDP growth by 42 percent and lower thepoverty rate by 5 percent (see Tables 5.1 and 5.2).

46 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

A Hydro-Economic Model of Ethiopia: A Summary of Findings 47

Combined Infrastructure and IrrigationInvestment Scenario

Parallel investments in irrigation and market infra-structure produce the greatest GDP growth andpoverty reduction and, of course, require the great-est investment. When the model with smoothedrainfall is run with combined investments in agri-culture and infrastructure, the overall GDP growthrate increases 58.9 percent over the no-investmentbaseline scenario (rising from 2.9 to 4.6 percent).This is the highest increase in overall GDP growthrates in any scenario. The irrigation-alone scenariogives a growth rate of 19.5 percent and the infra-structure-alone scenario give a growth rate of 54percent (see Table 5.1).

Strong synergies can be seen in the impact of thiscombined investment on the cumulative returns toirrigation investment. The per hectare return to ir-rigation under this scenario increases to US$10,350(from US$4,400), raising both returns to grain pro-duction (up to US$9,600 from US$2,800) and non-grain production (up to US$11,100 from US$6,100)(see table 5.3).

Combined investment also produced significantgains in the other rainfall models. The stylizeddrought model projects a 65 percent increase inGDP growth rates, and the historical variabilitymodel shows a 93 percent increase, the largest in-crease of all.

Poverty rates also decline significantly withcombined investments, 15 percent under a stylizeddrought assumption and 12 percent assuming his-torical variability (see Table 5.2). Each rainfall vari-

ation model achieves its greatest reduction inpoverty under the combined investment strategy.

Under both the smoothed rainfall model and thestylized drought model, the lion’s share of benefitsappear to derive from market infrastructure in-vestments, with the irrigation investments onlyadding marginally to total growth and povertygains. Under the historical variability model, how-ever, the gains are virtually additive.

These findings suggest three conclusions. First,irrigation has greater returns under highly variableclimatic conditions. Second, the complementarityof combined investments will be particularly pro-nounced in Ethiopia. Third, investments in roadsand market infrastructure are essential for leverag-ing returns to irrigation investments, although thereverse relationship (that investments in irrigationleverage returns on market infrastructure) may notbe strong.

Agricultural and Non-Agricultural GDP Growth: Changing the Structure of the Economy

Figures 5.5 and 5.6 show the results of the differentinvestment scenarios according to the smoothedrainfall model and the rainfall variability model.These figures separate the GDP growth intoAgricultural GDP and non-agricultural GDPgrowth. Agricultural GDP growth is essential in theshort run, but growth in non-agricultural GDP—manufacturing and services—should be encouragedbecause, over time, it will shift the structure of the

Table 5.3. Agricultural Production and Irrigation Returns Under the Combined Investment Scenario,Smoothed Rainfall Variant

Grain production Value of grain Value of Returns by 2015 production by 2015 other crops to irrigation

(million tons) (million $US) (million $US) (2003–15)

Baseline 11.0 3,236 1,335 —

Irrigation 11.3 3,259 1,451 US$4,400/hectareUS$2,800 (grains)

US$6,100 (other crops)

Combined irrigation 11.5 3,398 1,552 US$10,350/hectareand infrastructure US$9,600 (grains)

US$11,100 (other crops)

Source: Authors.

economy away from agriculture and toward moredrought-resilient activities.

The smoothed rainfall model (Figure 5.5) showsrelatively low agricultural GDP growth rates andrelatively low returns to irrigation investments.Since returns to irrigation are highest when cli-mates are variable, smoothed models fail to capture

the extent of benefits offered by irrigation in coun-tries such as Ethiopia.

The historical variability model (Figure 5.6)shows that investments in irrigation have a muchgreater impact when swings between drought andflood are assumed to be the norm. Overall GDPgrowth rates are still lower in the historical vari-

48 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

GDP Growth w/Smoothed Rainfall

0

1

2

3

4

5

6

7

Base Irrig. Roads Irri&Roads

Ann

ualG

row

th

GDP growth rate

Ag GDP growth rate

NAg GDP growth rate

Figure 5.5 Growth Rates With Smoothed Rainfall and Different Investment Scenarios

Source: Authors.

GDP Growth w/Variable Rainfall

0

0.5

1

1.5

2

2.5

3

3.5

4

Base Irrig. Roads Irri&Roads

Ann

ualG

row

th

GDP growth rate

Ag GDP growth rate

NAg GDP growth rate

Figure 5.6 GDP Growth With Variable Rainfall and Different Investment Scenarios

Source: Authors.

A Hydro-Economic Model of Ethiopia: A Summary of Findings 49

ability model because the enormous costs of hy-drology are not factored into the smoothed rainfallmodel. Combined investment in both irrigationand market infrastructure provides real gains inthe agriculture GDP growth rate over the single-sector investment scenarios.

For investments in irrigation or infrastructuresingly, overall growth rates (blue) are comparablebut the two scenarios give rise to very different rel-ative rates of growth between agricultural (red) andnon-agricultural GDP (white). This, in turn, willhave markedly different impacts on the structure ofthe economy over time and its likely vulnerabilityto future hydrological variability and shocks. Tohasten a structural shift toward non-agricultural ac-tivities, the model suggests a disproportionate in-vestment in market infrastructure. This type ofinvestment will help shift the structure of the econ-omy toward a more water-resilient growth paththat will also have a lesser impact on Ethiopia’s vul-nerable and highly degraded landscape.

COMPARISON OF INVESTMENTSCENARIOS USING THEHISTORICAL VARIABILITY MODELSince the historical variability rainfall model is the most accurate for Ethiopia, this section ex-plores in detail the three investment scenariosunder this model.

BASELINE SCENARIOThe model that most fully captures Ethiopia’s an-nual rainfall variability projects a striking 38 per-cent decline in the GDP growth rate and a 25percent increase in the number of people living inpoverty (11 million people) by 2015. Table 5.4

shows the no-investment baseline for the GDPgrowth rate, the agricultural GDP growth rate, thenon-agricultural growth rate, and the poverty ratesin 2003 with projections for 2015 under the histori-cal variability model.

IRRIGATION INVESTMENTSCENARIOIncreasing irrigation has long been seen as the mostdirect strategy to alleviate the impact of droughtand ensure food security. It is generally assumedthat increased irrigation will reduce the risk farm-ers face from climate variability, giving them theincentive to purchase fertilizers, improved seeds,and other inputs that can increase agricultural pro-ductivity. Without increased irrigation, the greatunpredictability of rains in Ethiopia is an over-whelming disincentive to investments in agricul-tural improvements.

Currently only 2 percent (200,000 hectares) ofEthiopian cropland is irrigated; that is, 60 percentis used for cereal production and the other 40 per-cent for other crops. Cereal production occupiesabout 70 percent of all cultivated land. Irrigationaccounts for just 1.5 percent of total cereal produc-tion compared with more than 2 percent of othercrop production.

The effect of irrigation is modeled as the yielddifferential between irrigated and rainfed crop pro-duction. According to the Agricultural SampleSurvey data of 1997 and 2000, this differential forgrain is 40 percent (that is, on average, irrigationcan increase grain yield up to 40 percent). A simpleback-of-the-envelope calculation shows that if in-creased irrigation areas can help close this yieldgap in the next 12 years, it would increase annualgrowth in grain production 3 percent. The model

Table 5.4. Baseline Scenario (No Investments) With Historical Variability

2003–15 Baseline

GDP growth rate 1.78Agriculture GDP growth rate 1.47Nonagriculture GDP growth rate 2.17Poverty rate 2003 44.23%Poverty rate 2015 57.45%

Source:Authors.

projects that if all of Ethiopian grain productioncould achieve this growth rate, the nationalpoverty rate would fall to 33 percent by 2015 andthe absolute number of people living in extremepoverty would fall by 1 million.

However, given that irrigation currently ac-counts for only 1.5 percent of total cereal produc-tion, it is unrealistic to expect that investments inirrigation alone can significantly reduce the yieldgap nationwide. Many researchers have shownstrong diminishing returns to the large-scale irri-gation investment, and there is increasing concernthat irrigation cannot be the primary focus for ad-dressing food security. Caution is therefore neededin promoting large irrigation investment projects aspanaceas for food security.

Returns from irrigation investments are ex-tremely sensitive to price changes, which derivefrom many sources. For example, prices may be af-fected by food aid. In drought years when demandexceeds supply, grain prices rise. If food aid grainsfill this gap, or are sold in domestic markets at con-trolled prices, price spikes will be tamped. The flowof cheap grain will undercut the profitability of in-vestments in grain irrigation and serve as a broaderdisincentive for surplus production by farmerswho could otherwise reap a profit through interre-gional trade in drought years.

For exportable crops, such as coffee, returns toinvestment will largely be driven by world marketprices. External price shocks can significantly alterthe calculation of potential returns from irrigationinvestments. When irrigation is used to increase theproduction of nonexport cash crops, the increasedsupply could drive down prices and reduce net re-turns from the investment.

The irrigation investment scenario with histori-cal variability shows a 53 percent increase in theprojected GDP growth rate, far greater returns than

either the smoothed rainfall or stylized droughtmodel, because when variability is modeled in allyears, irrigation becomes much more profitable(see Table 5.5).

Market Infrastructure Investment Scenario

An important contrast between the scenarios istheir potential equity implications. Irrigation bene-fits can easily be captured by a small set of farmers,while improved roads and lower marketing costsin general benefit the majority of farmers. The ben-efits derived from improved transport and marketinfrastructure24 are nonexcludable public goodsthat can be shared by all agricultural and non-agri-cultural traders, generating social welfare benefitsby reducing marketing margins and enabling lesswater-vulnerable activities.

A final, important point is that this scenario,which raises incomes without significant increasesin production, would have a lesser impact onEthiopia’s vulnerable and highly degraded land-scape. Because soil degradation is an extremely se-rious issue in Ethiopia, this difference is not trivial.

Investments in market infrastructure will in-crease the rate of GDP growth by 42 percent andlower the poverty rate by 5 percent under the his-torical rainfall variability model (see Table 5.6).Irrigation and market infrastructure investmentswould have roughly equal impacts.

Table 5.7 shows that the historical variabilitymodel suggests both types of investments wouldreduce poverty by a similar amount. The modelpoints to a relatively high rate of growth in non-agricultural (less water vulnerable) production inresponse to investments in market infrastructure,suggesting this type of investment will help shiftthe structure of the economy toward a more water-resilient growth path (see Table 5.8).

50 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 5.5. Irrigation Investment Scenario, Historical Variability Model

Baseline Irrigation investment Percent change

GDP growth rate 1.78 2.73 53Agriculture GDP growth rate 1.47 3.13 113Nonagriculture GDP growth rate 2.17 2.14 -1Poverty rate 2003 44.23 44.23 0Poverty rate 2015 57.45 53.18 -7

A Hydro-Economic Model of Ethiopia: A Summary of Findings 51

Combined Market Infrastructure andIrrigation Investment Scenario

This scenario examines the potential synergy ofcombining investments25 in irrigation and marketinfrastructure. It finds that parallel investments inboth would significantly increase growth, reducepoverty, and double returns to irrigation both forgrain and nongrain crops regardless of hydrologi-cal assumptions.

This model captures only the synergies gener-ated through price and income effects of these com-bined investments. The qualitative analysisprovided in this report, however, strongly suggeststhat significant additional synergies will derive

from colocating investments in irrigation and inroads and market infrastructure because togetherthey will provide strong incentives for increasedfarmer investments in agricultural inputs, and in-centives and opportunities outside of agriculture. Itis interesting to note that apparent synergies be-tween irrigation and infrastructure investments aremuch more pronounced under the historical vari-ability model. Under both the smoothed rainfallmodel and the stylized drought model, the lion’sshare of benefits appear to derive from market in-frastructure investments, with irrigation invest-ments only adding marginally to total growth andpoverty gains. Under the historical variabilitymodel, however, the gains are virtually additive.

Table 5.6. GDP Growth Rates, Market Infrastructure Investment Scenario, Historical Variability Model

2003–15 Historical variability Percent change over baseline

No investment (baseline) 1.78Irrigation investment 2.73 53Infrastructure investment 2.53 42

Source: Authors.

Table 5.7. Poverty Rates Under the Market Infrastructure Investment Scenario

2003–15 Percent of population living in poverty in 2015Percent change from baseline

No investment (baseline) 57.45Irrigation investment 53.18 -7Infrastructure investment 54.39 -5

Source: Authors.

Table 5.8. Impacts of the Market Infrastructure Investment Scenario, Historical Variability Model

Market Baseline: infrastructure

2003–15 No investment investment Percent change

GDP growth rate 1.78 2.53 42Agriculture GDP growth rate 1.47 1.64 12Nonagriculture GDP growth rate 2.17 3.61 66Poverty rate 2003 44.23% 44.23% 0%Poverty rate 2015 57.45% 54.39% -5%

Source: Authors.

This reflects the greater benefits of irrigation underhighly variable climatic conditions, and suggeststhat the complementarity of combined investmentsis particularly pronounced under such conditions.It also suggests that investments in roads and mar-ket infrastructure are essential for leveraging re-turns to irrigation investments, although thereverse relationship (that investments in irrigationleverage returns on market infrastructure) may notbe strong.

The combined investment scenario generatesmore growth in the economy overall and, like theinfrastructure investment scenario, gives rise to dif-fering agricultural and non-agricultural rates ofgrowth that will shift the structure of the economyaway from agriculture and toward more water-re-silient non-agricultural activities, though at a moremeasured pace than the infrastructure-alone sce-nario (see Table 5.9).

MULTIPURPOSE DEVELOPMENTAlthough the model described above does not ex-plicitly incorporate hydropower, a coordinated in-vestment strategy should clearly exploitopportunities for multipurpose infrastructure tomanage the impact of hydrological variability onthe Ethiopian economy. River regulation and stor-age infrastructure can be designed at all scales tomeet several objectives: delivery of irrigation andmunicipal water, generation of hydropower, thepublic benefits of drought and flood mitigation,and protection of roads and other infrastructure.Hydropower development at all scales is poten-tially an integral part of both irrigation and market

infrastructure development, and abundant energyis clearly needed to diversify the economy.

Generally, there are tradeoffs between the use ofwater for hydropower or for upland irrigation; thatis, the consumptive use of water above a hy-dropower plant means less power will be generated.Choices may need to be made, but in highlandEthiopia—for example, the large areas around LakeTana—there appear to be opportunities for substan-tial irrigation development without diminishing thepotential for hydropower generation. In addition,there appears to be many cases where small tomedium hydropower developments coupled withirrigation can be economic, but where single-pur-pose investments might not be (for example, on therivers that feed Lake Tana). Moreover, the largerstorage capacity justified by combined investmentswould provide even greater public benefits.Hydropower development at all scales is potentiallyan integral part of both irrigation and market infra-structure development, and abundant energy isclearly needed to diversify the economy.

The GoE is currently undertaking major develop-ment programs in irrigation and market infrastruc-ture, particularly roads and power. It is difficult togeneralize a priority between the two types of in-vestment from the results of the model. To hasten astructural shift toward non-agricultural activities,the model suggests a somewhat disproportionate in-vestment in market infrastructure, but this will notnecessarily be the priority in all locations. Specific in-vestment decisions will need to be made cognizantof local hydrological, environmental, economic, andsocial conditions, and the existing stock of both irri-gation and market infrastructure.

52 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 5.9. Impacts of the Combined Market Infrastructure and Irrigation Investment Scenario,Historical Variability Model

2003–15 Baseline Combined investments Percent change

GDP growth rate 1.78 3.43 93Agriculture GDP growth rate 1.47 3.29 124Nonagriculture GDP growth rate 2.17 3.59 65Poverty rate 2003 44.23% 44.23% 0%Poverty rate 2015 57.45% 50.42% -12%

Source: Authors.

53

Since the origins of human society, there has been a constant strug-gle to reduce the destructive impacts of water and increase its pro-ductive impacts.26 Many of the earliest civilizations, and particularlythose on the floodplains of the world’s great rivers, succeeded by har-nessing and managing water, thereby increasing production and re-ducing the risk of destruction.

Today, water resources development and management remain atthe heart of the struggle for sustainable development, growth, andpoverty reduction. This has been the case in all industrial countries,most of which invested early and heavily in both water infrastructureand institutions,27 and it is the case in all developing countries, mostof which have not invested sufficiently in water infrastructure and in-stitutions. In some developing countries, such as Ethiopia, the unmetchallenge of managing their water legacy is almost without prece-dent. Unless they are able to do so, sustainable growth and povertyeradication cannot be achieved.

WATER SECURITY AND THE MINIMUMPLATFORM OF WATER INFRASTRUCTURE AND INSTITUTIONS

Water Security

Investments in water for food, energy, industry, and navigation, andin associated institutions, are initially made by the state, the city, thefirm, the farm, or the family. Their goal is to achieve some implicitlevel of water security. We can define water security as the reliableavailability of an acceptable quantity and quality of water for pro-duction, livelihoods, and health, coupled with an acceptable level ofrisk of unpredictable water events (including the extremes of droughtand flood) that have high social and economic impacts.

Water security is achieved when water underpins economicgrowth rather than undermining it, or, in other words, when the netimpact of water on growth is positive. Water security is demonstratedwhen investments in water focus on enhancing growth rather thanmeeting basic needs and mitigating risk. It is the point at which vul-nerability to drought, flood, and disease, or a lack of access to water-related services, no longer creates an overwhelming obstacle to

Water Resources andGrowth6

growth. Water security is a dynamic state: differentin different parts of the world reflecting geo-graphic, social, epidemiological, economic, and po-litical factors and changing over time as thesefactors shift with development.

The Minimum Platform of WaterInfrastructure and Institutions

We can postulate that there is a basic level of watersecurity, which incorporates the idea of a minimumplatform of water infrastructure and institutions.Below this minimum platform, society and theeconomy are unacceptably impacted by watershocks and/or unreliable water for production orlivelihoods. The tipping point in achieving watersecurity will be the acquisition of a minimum plat-form of management capacity and infrastructureinvestment. Below the minimum platform, waterobstructs growth; above the minimum platformwater enhances growth. Until the minimum plat-form is achieved, the scale of social impacts (for ex-ample, morbidity, mortality, and resource conflict)and related economic impacts (for example, frominstitutional failure, production inefficiencies, anddisaster shocks) can be such that social fabric is sig-nificantly affected and economic growth cannot bereliably and predictably managed.

THE CHALLENGE OF ACHIEVINGWATER SECURITYThe level of institutions and investment requiredfor water security will differ across countries andacross economic actors as a consequence of the nat-ural hydrology, the structure of the economy, its re-silience to water shocks, and risk aversion. Anation’s hydrology—the absolute levels of waterresource availability, its variability, and its spatialdistribution—will largely determine the institu-tions and the types and scale of infrastructureneeded to manage, store, and deliver the resource.

An Easy Hydrologic Legacy

Relatively low rainfall variability, with rain dis-tributed through the year resulting perennial riverflows sustained by baseflows, results in hydrologythat is “easy” to manage. In temperate parts of theworld, much of which are now industrialized,achieving the minimum platform for water secu-rity was straightforward and required compara-

tively low levels of skill and investment, primarilybecause water was sufficient, widespread, and rel-atively reliable.

Once the minimum platform was achieved,growth was able to proceed without water as a sig-nificant constraint. Growth allowed further water-related investments, which increased watersecurity beyond the minimum platform. As returnsfrom new water investments gradually dimin-ished, the infrastructure and institutional platformbecame mature, with water a reliable input to pro-duction and risks (that is, insurance costs) thatwere fully acceptable. At maturity, the need and in-centive for developing new investments are low,while the returns from, and the incentives for, man-aging existing assets are high.

A Difficult Hydrologic Legacy

“Difficult” hydrology, on the other hand, includesareas of absolute water scarcity and, at the other ex-treme, low-lying land where there is severe floodrisk. Even more difficult hydrology is where rain-fall is markedly seasonal as in the tropics where oneor two rainy seasons per year is typical, because ashort season of torrential rain followed by a longdry season requires the storage of water. More dif-ficult still is high interannual climate variability,where extremes of flood and drought create unpre-dictable risks to individuals, nations, and regions.Perhaps most difficult of all is a combination of ex-treme seasonality (intra-annual) and variability (in-terannual), which is characteristic of many of theworld’s poorest countries. Ethiopia clearly is acountry with extremely difficult hydrology.

With increasingly difficult hydrology, the level ofinstitutional refinement and infrastructure invest-ment needed to achieve water security becomes sig-nificantly greater. Thus in these countries water is amuch higher constraint to growth than in the tem-perate nations with their easy hydrology. In addi-tion, achieving the minimum platform is also muchhigher, a quantum leap in some cases. This leap hasnot been achieved in many poor countries, so basicwater security is lacking and water remains a keyconstraint on growth, an unreliable input to pro-duction and the cause of major economic shocks. Incountries with difficult hydrology, there are few in-centives to manage what little infrastructure there is.While the returns to society from investing in watersecurity could be very high, there may simply be in-sufficient wealth to invest.

54 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources and Growth 55

Taking this argument further, we postulate thatsocieties in areas of water scarcity and/or high cli-mate variability have remained poor and in a low-level equilibrium trap at least in part because it hasbeen impossible for them to make the compara-tively large investments needed to reach the mini-mum water infrastructure platform.

A Transboundary Hydrologic Legacy

The management and development of rivers whosebasins fall within the borders of more than one stateis exceptionally complex owing to the anarchy of in-ternational relations. The UN Convention on theNon-Navigational Use of International Water-courses, which seeks to address many of these diffi-cult issues, was under preparation for some 20 yearsand adopted by the UN General Assembly in 1997(Salman and Boisson de Chazournes 1998). Evenafter such prolonged negotiations, however, the con-vention has not yet entered into force because an in-sufficient number of states have ratified it.

Over generations, the border between Franceand Spain evolved so that today it mostly followsthe watershed, a line of least dispute. In developingcountries, colonial rule resulted in geographic di-visions that usually have no geographic logic andseriously compound the challenge of achievingwater security. Britain, France, Portugal, Italy, andGermany carved up Africa in a way that has led toa complex legacy of international river basins.Guinea has 14 international river basins,Mozambique has eight. Similarly, river basins areshared by a large number of countries: the NileBasin has 10 countries, the Niger Basin has nine.Similar types of colonial boundaries challengeother areas, such as South Asia.

Cooperating with one state on one river is a chal-lenge for poor nations. With more than one state ona river, or if on more than one river, or both, thechallenge can be extremely high. Planning for opti-mal development is also complicated by the factthat relevant historical and current data—where itexists at all—is likely to be fragmented, nonuni-form, and owned by different riparian states. Infact, there may be few if any studies of the river sys-tem as a whole.

Poor river basin management can result in manylost opportunities and increased costs: environ-mental costs to the river, economic costs in subop-timal development of the river, costs from politicaltensions over the river, and costs of all the other op-

portunities foregone through non-cooperation(Grey and Sadoff 2002). The legacy of internationalrivers can significantly affect the potential for man-aging and developing water for growth andpoverty alleviation.

ECONOMIC STRUCTURE AND RESILIENCEThe structure of the economy will affect the mini-mum level of institutions and infrastructure neces-sary for water security. How much does theeconomy rely on water resources for income gen-eration and employment? How vulnerable is it towater shocks? For example, a country like Ethiopia,whose rainfall is highly seasonal and/or variableand relies heavily on rainfed agriculture, is highlyvulnerable. A country whose most productive as-sets or areas lie in flood plains is vulnerable.Vulnerable countries may require more extensiveinvestments to achieve water security.

Water vulnerability will affect geographic re-gions and sectors differently, creating incentivesand disincentives for specific economic activities.These activities will influence both the structure of the economy and spatial patterns of growth, thusaffecting overall growth and equity outcomes. Notonly will vulnerable economies regularly suffergreater setbacks from water shocks, but their vul-nerability will likely prove a strong disincentive fordomestic or foreign entrepreneurial investmentsthat could shift the economy toward a more diver-sified, water-resilient structure. More diversifiedeconomies that are less water dependent or wealth-ier economies that can more easily compensatethose harmed by drought or flood, might accepthigher levels of hydrological uncertainty withoutslowing growth-focused investment.

This analysis suggests that efforts to guide struc-tural change in the economy toward achievinggreater economic resilience to water shocks canhelp a country attain and improve water security.

Risk Aversion

In the poorest countries, where survival is a realconcern for much of the population and there arefew functional social safety nets, economic actorstend to be extremely risk averse, investing onlyafter there is significant demonstration of returns.Levels of risk aversion will influence the thresholdat which water security can be achieved.

While the catastrophic effects of drought and flood extremes are apparent, it is less well recognized that even in years of average precipi-tation, expectations of high variability and en-demic droughts and floods will affect economicperformance and, potentially, the structure of the economy.

The expectation of variability constrains growthand diversification by encouraging risk-averse be-havior at all levels of the economy in all years, aseconomic actors, particularly the poor, focus on minimizing their downside risks rather thanmaximizing their potential gains. Farm familieswill quite rationally not invest in land improve-ments, advanced technologies, or agricultural in-puts, thus constraining agricultural output andproductivity gains. Insecure land tenure arrange-ments compound these disincentives to invest inland management.

Lack of such investments can lead to land degra-dation and desertification, which will result in a vi-cious circle of reducing production anddeteriorating assets. Similar disincentives for in-vestments in industry and services will slow the di-versification of economic activities and maintain aneconomic structure that is based largely on low-input, low-technology agricultural production.

The poorest countries may well face the highestrisks, yet have the most risk-averse populations,the lowest infrastructure investment, and theweakest institutions. This could well be a very seri-ous low-level equilibrium trap: These countriesbegin at a very low level, but must ascend to quitea high level of institutions and investment toachieve water security.

Financing Water Security

Generally, early investments in water security weremade by governments. In Wealth of Nations,Adam Smith wrote that government should con-struct public works when these works are “of sucha nature, that the profit could never repay the ex-pense of any individual or small number of indi-viduals, and which it therefore cannot be expectedthat any individual or small number of individualsshould erect or maintain.” When basic water secu-rity is achieved and additional investments inwater (for example, in irrigation or power genera-tion) can be highly profitable. At that point privateinvestors will respond. All rich countries haveachieved water security with a publicly funded

minimum platform in place. Most poor countriesnot have achieved this level.

THE DYNAMICS OF INSTITUTIONSAND INFRASTRUCTURE

The Institutions/InfrastructureInterrelationship

All countries require investments in both water in-stitutions and infrastructure. But what is the bal-ance and sequencing between the two types ofinvestments?

Human society has always managed water.Village wells and ponds, canal systems, andearthen embankments in the flood plains of greatrivers have all been developed for millennia andeach has been accompanied by institutional sys-tems to plan, develop, manage, and maintain it. Insome cases, such institutions were an early form ofgovernment, an emerging public sector to managepublic goods. A private sector emerged to provideservices to meet public demand. Water institutionsreflect the culture and political economy withinwhich they fit: public to private, centralized and hi-erarchical to decentralized and participatory, andrules-based to market-based.

When we look across the world at the develop-ment of water resources, particularly during theperiod of industrialization, we see a pattern (seeFigure 6.1). At all times, concomitant investmentsmust be made in infrastructure and institutions,but when stocks of hydraulic infrastructure arelow, investment in them will be a priority.Investment in management capacity and institu-tions becomes increasingly important as larger andmore sophisticated infrastructure stocks are built.

Most developed countries fall into the Type 3category, where significant infrastructure invest-ments have been made (in some cases arguably ex-cessive investments) and where efforts are bestdirected toward strengthening water resourcesmanagement. In Type 1 countries, most of theworld’s poorest countries, infrastructure stocks areso low that investments in management do nothave the same high returns. Middle-incomeeconomies are often in the Type 2 category.

Without the infrastructure to store and deliverwater and manage flows, there is neither the neednor the incentive for sophisticated managementpractices. While developed countries are appropri-

56 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Water Resources and Growth 57

ately focused on implementing integrated water re-sources management, developing countries maydo better to adopt a principled and pragmatic ap-proach to management while putting greater em-phasis on concurrent infrastructure investments.28

The Institutions/Infrastructure Balance

Failure to understand the issue of balance and se-quencing within the context of specific country cir-cumstances may lead to poor investment choices.One potential danger is that most donor nations areType 3 countries strongly focused on water re-sources management, but the priority of Type 1client countries may be investment in water re-sources infrastructure. Type 3 donors must not lettheir perspectives misinform the development pri-orities of Type 1 countries. Of course, there must bemanagement and capacity investments at allstages, but the emphasis in investment may varyaccording to the “type” of the country.

The Transboundary Institutions/Infrastructure Case

Shared management of international rivers is ofgrowing importance in a world of 260 internationalrivers, shared by about 90 percent of the world’s

population. A basic premise of water resourcesmanagement is that river basins are best managedand developed as an integrated whole. In practice,however, management is always legally and polit-ically complex, due to the challenges of allocationbetween users and uses. There is no final authoritythrough which differences can be resolved.

Although criteria for allocating water and itsbenefits can be drawn from a growing body of cus-tomary international water law, there is no consen-sus on the criteria for equitable allocation (Sadoff etal. 2003). Developed (Type 3) economies have usu-ally achieved a relative equilibrium in establishingfit-for-purpose institutional arrangements, includ-ing treaty regimes, dealing with issues of river in-frastructure, and the quantity and quality of waterflows. In many cases, the need for river infrastruc-ture, such as navigation locks, weirs and dykes forflood management (for example, the Rhine), or hy-dropower facilities (for example, the ColumbiaRiver), drove the neighboring countries to adoptinstitutional solutions. Recently there has been agrowing emphasis on improving water qualitythrough joint institutional solutions to restoreriverine and lacustrine ecosystems (for example,the Danube). The success of such arrangements isenabled by broader international relations andagreements.

Returns on Investment

ManagementInvestments

Developing Developed

Type 1 Type 2 Type 3

InfrastructureInvestments

Figure 6.1 Balancing and Sequencing Investments in Water Infrastructure and Management

Source: World Bank (2002).

Both middle-income (Type 2) and developing(Type 1) economies continue to face great chal-lenges. Because the complexity of riparian relationsis such an obstacle, some nations seek to developriver segments within their own territories indepen-dently, often settling for the second or third best op-tion from a basinwide perspective. In extreme cases,tensions over international rivers can effectively halttheir management and development.

The need for river infrastructure often drivescountries to reach agreements about river manage-ment. For example, the 1960 Indus Basin Treaty be-tween India and Pakistan was a prerequisite for theIndus Basin development program that divided therivers. The 1959 Nile Agreement between Egyptand Sudan was for the Aswan High Dam thatstores water for Egypt. A focus on the benefits ofcooperative management—say, for water quality—

and of agreed (even cooperative) development—say, for irrigation and power—can lead to viabletransboundary institutions.

Ethiopia and Water Security

Ethiopia is an economy with little water resourcesinfrastructure and relatively weak management in-stitutions and capacity. It also suffers from extremevariability and seasonality. It is a highly vulnerableeconomy dependent on rainfed subsistence agri-culture and pervasive risk-averse behavior, all ofwhich put it into a particularly “deep hole” interms of achieving a minimum infrastructure plat-form and water security. Major public investmentin water resources infrastructure and institutionswill be needed to enable Ethiopia to achieve itsgrowth potential.

58 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

59

This section briefly summarizes the lessons of the past several sec-tions and examines their implications for interventions. In the nexttwo sections we examine what the GoE and the Bank have done andare planning to do to deal with Ethiopia’s water resources problems.Finally, we make specific recommendations for Bank investments.Although the Bank can address many issues, there are some recom-mended interventions that it will probably not address. Thus, thissection serves as a short summary of findings and recommendations.

We have documented Ethiopia’s extreme hydrological variabilityand shown how it has kept the country in a poverty trap. The riskcaused by hydrological variability discourages individual investmentin irrigation, market infrastructure, and agricultural inputs.

The appearance of food aid during times of famine saves lives, butcan also undermine the development of a national market economyin agriculture. An economy based on rainfed agriculture cannot pros-per in the face of such extremes of drought and flood, but the GoE,even with the help of donors, has not been able to invest in sufficientwater infrastructure projects to provide a base of water security.Meanwhile, the country’s environmental resource base suffers con-tinuing degradation as the poor strip watersheds for fuelwood, silta-tion plugs up hydropower and irrigation reservoirs, and erosion andpoor drainage destroy cropland.

The way out of the trap seems to be through combined invest-ments in water infrastructure projects and management, irrigation,and market infrastructure. Investments in water infrastructure andirrigation will help smooth out the spikes of variability that plaguethe country and help it achieve a base level of water security.Investment in market infrastructure will help it move away from anoverdependence on agriculture and toward development of its man-ufacturing and service sectors. As these investments help stabilize theeconomy, changes need to be made in banking and credit manage-ment so that funds are available for individual investment andhealthy market processes can develop.

Insights for Interventions7

LESSONS

• Water resources development and management isa core issue for development in Ethiopia. Ethiopia,which experiences extreme hydrological vari-ability and is highly dependent on rainfedagriculture, lacks the minimum infrastruc-ture, institutions, and capacity to achievewater security. Without serious investment inwater security, Ethiopia will be held hostageby its hydrology and be unable to break out ofits cycles of famine and food aid.

• Droughts are only part of the story, which includesvariability, seasonality, and floods. The modelsdemonstrate that while measurable droughtssignificantly shock the economy, “normal”year-on-year variability, unpredictable sea-sonality, and the impacts of waterlogging andflood (which did not even capture the fullrange of infrastructure damage) consistentlyundermine growth and poverty gains.

• Failure to incorporate hydrological variability intoeconomywide modeling will result in significantand systematic overestimations of growth and un-derestimations of poverty. Most models makeprojections based on average rainfall and agri-cultural production. Because such modelssmooth over the production and price spikescaused by significant hydrological variations,they mask the extreme incidents that thrustvulnerable poor farmers into poverty. This isparticularly significant in Ethiopia, where rain-fall is highly variable, droughts and floods areendemic, and a significant portion of the pop-ulation subsists on rainfed agriculture.

• Failure to incorporate hydrology into economywidemodeling will disproportionately undervalue re-turns to irrigation and drainage investments. Oursuite of models clearly demonstrated that asvariability increases, returns to irrigation anddrainage increase. This is fairly intuitive be-cause it is at the extremes—during periods ofdrought and flood—that the returns to irriga-tion and drainage are greatest. Economywidemodels without explicit hydrology will sys-tematically undervalue the contribution of ir-rigation and drainage, in particular, in theEthiopian economy.

• Increased multipurpose water infrastructure is apowerful investment. To mitigate the economicimpacts of variability, greater water storage ca-pacity, both natural and manmade, small scale

and large scale will be needed. Future schemes,wherever possible, should be designed as mul-tipurpose projects, potentially providing bene-fits such as watershed management, floodcontrol, drought mitigation, drinking water, ir-rigation, fisheries, local transport, sedimentmanagement, and systemwide reduction ofevaporation losses and recreation. Providinghydropower as an alternative to biomass fueland integrating targeted watershed manage-ment protection can also help slow deforesta-tion and soil degradation, which will mitigatesome hydrological impacts.

• Investments in irrigation and drainage can helpensure agricultural production and incomes.Appropriately designed irrigation anddrainage investments can provide a securesupply of agricultural water to buffer produc-tion during all but the most extreme climaticevents and protect the subsector from thegreater part of hydrological variability. It canenhance food security and the reliable deliv-ery of marketable and exportable agriculturalproducts, helping to stabilize prices and en-courage investment in both the agriculturaland nonagricultural sectors.

• Water resources management capacity and institu-tions must be strengthened. Given Ethiopia’schallenging hydrology, the need for such ca-pacity is great—but capacity in the country islow. Efforts to strengthen capacity are ongoingand should be seen as a continued priority.Investment in research and education in par-ticular will be crucial to develop the capacityneeded to design effective and appropriatewater resource management interventions,watershed management, irrigation schemes,and water and sanitation services.

• Groundwater resources should be more fully ex-plored. Ethiopia’s groundwater resources areestimated to amount to 2.6 billion cubic me-ters, but significant work remains to be donein determining how to develop these re-sources sustainably. One distinct advantageof groundwater is its relative reliability, dueto buffering from rainfall variability.

• Dependable clean water supplies are needed forhuman health and productivity and as an incen-tive for investing in manufacturing and services.The clear need for investments in infrastruc-ture will require accompanying reform, muchof which is underway. In some cases, effective

60 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Insights for Interventions 61

urban water supply planning may requireconcomitant investments in watershed man-agement, bulk water infrastructure, or riverregulation infrastructure.

• Watershed management is crucial for water re-sources, livelihoods, and the environment.Improvements in watershed management,land use planning, and forest managementwill be crucial in managing water resourcesand riverine ecosystems more broadly.Improved watershed management can slowwatershed erosion; moderate the hydrologi-cal cycle; regulate runoff and groundwater;improve infiltration, water retention, andbase flows; and reduce potential flood dam-age. An important lesson from experience inEthiopia is the need to effectively includestakeholders in the planning, design, andmanagement of watershed interventions.

• Limited and costly power supplies force reliance onbiomass, the collection of which is women’s burdenand degrades watersheds. Ethiopia’s current en-ergy balance relies heavily on the use of fuel-wood, crop residues, and dung. In ruralEthiopia women traditionally gather andtransport fuelwood and other organic fuels,and women suffer from indoor air pollutionfrom the smoke of cooking with biomassfuels. At the same time, removing these or-ganic materials from the soil contributes sig-nificantly to land degradation.

• Parallel investments in irrigation and market in-frastructure generate synergies for growth andpoverty reduction. Synergies can be expected toderive from co-locating investments in irriga-tion with investments in roads and market in-frastructure, because together they willprovide significant incentives for increasedfarmer investments in agricultural inputs andincentives and opportunities outside of agri-culture. The model suggests there are greatersynergies from co-located investment as hy-drological variability increases and that syn-ergies are particularly strong for leveragingreturns to irrigation investments.

• Proactive interventions are needed outside of thewater sector to enhance the economy’s resilience tothe impacts of hydrological variability. They mightinclude policies that promote alternative liveli-hoods to decrease the share of the populationmost vulnerable to hydrological shocks, pro-tection of agricultural incomes, and infrastruc-

ture investments explicitly designed to with-stand hydrological shocks to diminish costlyinfrastructure damage and associated cost in-creases in other infrastructure using sectors.

• Market infrastructure investments promote non-agricultural growth and provide immediate growthand poverty gains. In the medium term, they shiftthe structure of the economy away from agriculture,making it less vulnerable to hydrology. In the in-frastructure investment scenario, GDP growthcomes primarily from nonagricultural sources.Immediately, and more markedly over time,the structure of the economy will shift awayfrom agriculture and toward more drought-re-silient, nonagricultural activities.

• Investments in market infrastructure have less im-pact on the landscape. The infrastructure in-vestment scenario raises incomes withoutsignificantly increasing agricultural produc-tion by lowering transactions and marketingcosts and encouraging nonagricultural activi-ties. This suggests that welfare gains could becaptured under this scenario without the useof additional land and water resources foragriculture and might therefore have a lesserimpact on Ethiopia’s vulnerable and highlydegraded landscape.

• Infrastructure investments are public goods at lessrisk of capture than irrigation investments.Irrigation benefits can easily be captured by asmall set of farmers, while improved roadsand lower marketing costs benefit the major-ity of farmers. The benefits derived from im-proved transport and market infrastructureare nonexcludable public goods that can beshared by all agricultural and nonagriculturaltraders, generating social welfare benefits byreducing marketing margins and enabling arange of less water-vulnerable activities. Roadinvestments are arguably more socially pro-ductive because their gains are derived froma decrease in transactions costs to both pro-ducers and consumers (a real public-good ef-fect) rather than a transfer from consumers toirrigated farmers.

• Drought preparedness and agricultural resilienceneeds further strengthening. Greater emphasiscould be placed on the development of localwarning systems, selection of climate-appro-priate crops, and a range of cultivation practices and animal-rearing systems andtechnologies.

Alternatives should be sought to the practice of stor-ing wealth in livestock, which increases rural economicvulnerability to weather shocks. Drought-proof storesof wealth and related tools such as credit andweather insurance should be explored to provideasset stability during drought episodes.

IMPLICATIONS FOR INTERVENTIONS

• Bank-supported investments and interven-tions in Ethiopia should be analyzed to deter-mine whether they provide opportunities andincentives that will make the economy lessvulnerable to hydrological variability.

• Any economic modeling done for Ethiopiashould incorporate the country’s hydrologicalvariability. Returns to irrigation and drainageinvestments are systematically underesti-mated in models that fail to capture hydrol-ogy because these returns are highest in timesof drought and flood. In Ethiopia, modelingwith hydrological variability doubles returnsto irrigation and drainage.

• Water infrastructure and management capac-ity should be priorities for public investment.Ethiopia’s growth will be undermined until itattains water security, and it will require sig-nificant investment to achieve a minimumplatform of water infrastructure and manage-ment capacity.

• Road development should focus on qualityand regional interconnection. Investments inhigh-quality all-weather roads leading to re-gional centers will create incentives for theproduction and marketing of surplus agricul-ture. Roads should also be built between com-plementary regions to enable national tradeduring droughts and floods.

• Interventions should be analyzed to deter-mine whether their spatial location leveragesthe potential for synergies across investmentsin hydraulic infrastructure, irrigation, androads and market infrastructure.

• Infrastructure and road building projects,wherever appropriate, should include a wa-tershed management component to ensurethat projects are environmentally and eco-nomically sustainable and not undermined bysiltation, inundation, or erosion.

• Infrastructure investments require comple-mentary institutional strengthening and ca-pacity building, and potentially reforms toenhance effectiveness, efficiency, sustainabil-ity, and accountability.

• Opportunities should also be explored to com-plement water investments with insurance,credit, or “drought proof” stores of wealth tohelp producers manage weather risk. Ethiopiahas just launched a highly innovative pilot pro-gram in weather risk insurance.

• Solutions to water resources challenges shouldbe sought outside the water sector. Investmentin market infrastructure and hydropower de-velopment would both help to mitigate the di-rect impacts of drought and at the same timecoax the economy to a more resilient structure.Investments in market infrastructure will pro-mote a structural shift toward growth in ser-vices and manufacturing, which are lessvulnerable to rainfall variations.

• Regional cooperation in power productioncould strengthen regional relations and pro-mote increased economic cooperation and in-tegration.

• Food aid programs should be designed mind-ful of their impacts on longer term develop-ment. Development projects should bedesigned keeping in mind the complexitiesintroduced by emergency food aid programs.

62 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

63

The previous sections lay out the real challenges faced by Ethiopia indeveloping and managing its water resources. This section examinesongoing government and donor initiatives and seeks to define a strat-egy for Bank assistance that focuses on high impact interventions forwhich the Bank has some comparative advantage.

ETHIOPIA’S WATER SECTOR STRATEGY (2001)The GoE has recently developed a Comprehensive and IntegratedWater Resources Management Policy (endorsed in 1999), followed bya Strategy and a Sector Development Program, aimed at taking an in-tegrated approach to water resources development. The policy es-tablishes the overall goal of promoting efforts to utilize availablewater resources efficiently and equitably, thereby contributing to thecountry’s socioeconomic development on a sustainable basis.

The GoE clearly recognizes the scale and complexity of the waterchallenge and is undertaking a wide range of programs to meet thischallenge. Many of these programs can be categorized as nationalwater resource management programs. Others promote internationalcooperation (particularly on the Nile), or seek to mitigate the impactof hydrologic variability on the Ethiopian economy by looking be-yond what is traditionally considered the water sector to encouragealternative livelihoods that are more resilient to water shocks. The ob-jectives of some of the programs are described below.

National water management programs include activities that:

• Seek to better secure the timely availability and effective use ofwater, such as investments in watershed restoration and con-servation, water storage (including ponds, surface and subsur-face dams, and rainwater harvesting at household, community,local, and regional levels), irrigation infrastructure, and urbanand rural water supply (and sanitation).

• Reduce water-related risks faced by the farmers and familiesmost vulnerable to water shocks through agricultural coopera-tives, rural credit programs, expansion of agricultural extensionprograms, incentives for crop diversification, and community-driven development assistance targeted to food-insecure andpastoralist areas.

A Water ResourcesAssistance Strategy for the World Bank

8

• Incorporate multipurpose investments (forexample, addressing power, irrigation, floodcontrol, and watershed protection in an inte-grated way).

Programs that address international watershedcooperation include fostering cooperative develop-ment of transboundary waters (particularly on theNile River, where major progress is beingachieved) to increase the benefits that can be de-rived from Ethiopia’s water resources and to stabi-lize regional relations.

Programs to encourage changes in the structureof the Ethiopian economy to make it less vulnera-ble to hydrological variability include:

• Voluntary resettlement schemes designed tomove populations into lower-risk and higher-productivity areas

• Private sector reform and investment incen-tives to encourage more diversified invest-ment and the adoption of more advanced andless vulnerable production methods and tech-nologies

• Transport infrastructure investments that willimprove market access and the function ofmarkets within the country that will allowinter alia the transfer of food from surplus todeficit areas

• Investments in hydropower production,transmission and rural electrification, reduc-ing dependence on fuelwood, creating newproductive opportunities, and potentiallyearning foreign exchange

Most of Ethiopia’s international assistance forwater-related projects currently goes to irrigationand water supply and sanitation efforts. (See table8.1 and appendix 2.)

WORLD BANK ASSISTANCE IN ETHIOPIAThe Bank’s broad engagement in Ethiopia isguided by a set of strategic documents includingthe Poverty Reduction Strategy Paper, the recentCountry Economic Memorandum (CEM), and theforthcoming Country Assistance Strategy (CAS).The recommendations arising from this CountryWater Resources Assistance Strategy are consistentwith, and reinforce, those of CEM and will help toinform the forthcoming CAS.

More specifically, the Bank’s recent CEM high-lights Ethiopia’s extreme vulnerability to droughtthat is sustained by a heavy reliance on rainfed tra-ditional agriculture. It recommends that the GoEadopt a more balanced growth strategy that willfoster deeper inter-sectoral linkages. Key recom-mendations includes:

• Rehabilitate and/or complete existing damsand related irrigation schemes in order to at-tract private investment.

• Speed up privatization of state farms and pro-vide the appropriate complementary publicgoods to attract private investors.

• Develop multipurpose dams in strategic loca-tions for irrigation and hydropower, withphysical linkages—such as roads and tele-com—to urban areas.

• Develop complimentary economic infrastruc-ture to reap the potential for large-scale mech-anized rainfed agriculture.

• Support livestock production in particularareas.

The Bank currently supports a portfolio of on-going projects that deal directly or indirectly withwater resources. They include the Water Supplyand Sanitation Project, the Ethiopian SocialRehabilitation Development Fund, the FoodSecurity Project, the Energy II Project, the EnergyAccess Project, the Agricultural Research andTraining Project, and the Pastoralist CommunityDevelopment Project. Ongoing research in small-scale irrigation and rainwater harvesting is alsobeing supported. In addition, the Bank is providingsubstantive support to the NBI, in which Ethiopiais the leading donor and facilitator. Appendix 3 de-tails existing projects.

However, to date, there has been no attempt tolook across these projects, rationalizing and/orleveraging activities to support the country’s newintegrated water resources management policy andstrategy and to specifically address the need tomaximize resilience to hydrological variability andassociated economic shock. The current Bank-sup-ported portfolio breaks down as follows:

Alignment of the Portfolio

The sectoral balance of the portfolio aligns reason-ably well with the insights provided in this study.The bulk of resources are committed for transport

64 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

A Water Resources Assistance Strategy for the World Bank 65

and roads investments, with similarly large invest-ments designated for agriculture and energy. It isnotable, and consistent with this analysis, that alarge proportion of resources are committed toemergency activities. A clear goal over time will beto diminish emergency needs and focus more re-sources on transforming the Ethiopian economy.

While watershed management is addressedthrough components in several Bank projects,29 it isstriking that there is no dedicated watershed sup-port in the Bank portfolio. There are, however,some activities under preparation. Other donorsare active in these areas. Private sector develop-ment is yet another important area where activitiesare under preparation.

This analysis found that the current weak incen-tives for commercial agricultural production and

non-agricultural activities contribute to the contin-uation of an extremely vulnerable, subsistence-based agrarian economy. To better align itsportfolio with opportunities for increasing the re-silience of the Ethiopian economy to drought, theBank should make priority investments in multi-purpose water infrastructure, water supply andsanitation, irrigation and drainage, market infra-structure and private sector development, and tar-geted watershed management.

Proposed Modifications to the CurrentWorld Bank Portfolio

Several relevant infrastructure projects that are inthe planning stage should be reviewed with therecommendations of this report in mind.

Table 8.1. Ongoing International Assistance to the Water Sector in Ethiopia

Amount

US$ In thousands, Project/program type Project/program name Financier (thousands) other currency

Capacity building Water Resources Development and Utilization Program UNDP 5,000Institutional Setup Studies of the Abbay Basin France 1,540 EuroGround Water Development and Water Supply Japan 4,453

Training CenterWide area and local area networking World Bank 300

Hydropower Baro and Karadobi Hydropower Feasibility Study Norway NKr40,000

Irrigation Small-Scale Irrigation Project5,300 ha IFAD 22,6002,400 ha France 10,050600 ha Ireland 1,340Agricultural Support Program 7,500 ha AfDB 33,992Koga Irrigation and Watershed 6,000 ha AfDB 42,460

Water resources Genale Dawa Master Plan AfDB 4,950Awash Flood Mitigation AfDB 2,488

Water Supply Rural Water Supply UNICEF 12,120and Sanitation Water Supply and Sanitation World Bank 100,000

Kibremengist Town BADEA 6,000Debrebirhan Town Water Supply and Sanitation AFD/France 6,399Three Towns Water Supply and Sanitation Study Germany (KfW) ¤553Harar Water Supply ADB 28,263

Notes: ha, hectares; UNDP, United Nations Development Programme; IFAD, International Fund for Agricultural Development; AfDB, AfricanDevelopment Bank; UNICEF, United Nations Children Fund; BADEA, Arab Bank for Economic Development in Africa; AFD, Agence Française de Devéloppement.Source: MoWR Policy and External Relations Department.

For example the successors to the Energy IIProject are under preparation, and planning is un-derway on projects addressing roads and irrigationand drainage. In particular, the Bank should exam-ine the spatial distribution of these projects to seewhether synergies might be captured by adoptinga growth pole approach.

Planned roads projects should give special con-sideration the balance to between all-weather versusdry weather roads and to strategic interconnectionsbetween regions with potential trade opportunities.

Targeted watershed management projectsshould also be considered to complement ongoingor planned investments (by the Bank or others) inroads or dams.

Food-based emergency aid, while a traditionaland politically expedient response, undermines in-centives for surplus production and domestictrade. All emergency aid should be designed withimpacts on longer-term national development keptin mind.

Support for ongoing weather insurance pilotprograms should be continued, and rural bankingand credit schemes should be explored. From theperspective of water vulnerability, farmers andpastoralists should increasingly be encouraged touse banks for their savings. In direct contrast tolivestock, the traditional store of wealth, savingsaccounts can sustain asset values during droughts.Furthermore, managing herd sizes will diminishdemands on water during droughts and help con-

trol pollution loads where drinking water sourcesand livestock are in close proximity.

The Bank is in discussions with the GoE regard-ing potential support for ongoing resettlement pro-grams. Before lending support, the Bank shouldanalyze these programs to determine whether theproposed settlement patterns and livelihoods willbe adequately resilient to hydrological variability.Since the objective is to move farmers away fromvulnerable food-insecure areas, adequate incen-tives and investments should be put in place to en-sure that their livelihoods are more diversified andless dependent on rainfall.

A STRATEGY FOR WORLD BANKSUPPORT IN WATER RESOURCESThe needs in Ethiopia are great, and it is difficult todraw comparisons regarding the urgency of needsacross different sectors. From the perspective of en-hancing Ethiopia’s resilience to drought and hy-drological variability, however, there are severalareas that stand out as priority investments:

• Multipurpose hydraulic infrastructure de-velopment, including both hydropower andirrigation

• Water supply and sanitation• Watershed management• Transport, market infrastructure, and private

sector development

66 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Table 8.2. World Bank Portfolio Distribution by Major Sectors, FY2005

Sector US$ (million) Percent

Agriculture/rural development 309.5 13.05Capacity building 151.2 6.38Energy 332.7 14.03Education 40.0 1.69ESRDF 136.8 5.77Health 100.0 4.22HIV-AIDS 59.7 2.52Transport/road 596.9 25.17PRSC 130.0 5.48Water supply and sanitation 100.0 4.22Others (emergency, cultural) 414.4 17.48Total 2371.2 100

Notes: ESRDF, Ethiopian Social Rehabilitation and Development Fund; PRSC, Poverty Reduction Strategy Credit.Source: Government of Ethiopia (1996b).

A Water Resources Assistance Strategy for the World Bank 67

Multipurpose Storage Development

Multipurpose storage development addresses manyof the vulnerabilities discussed in this report andwould help shift the structure of the economy to-ward a more resilient path. Multipurpose dams canbe large or small. They can be designed to createstorage capacity, regulate flows during droughtsand floods, generate hydropower, provide irrigationwater, and develop fisheries opportunities. In lightof our finding that returns to irrigation should begreater than expected in Ethiopia, irrigation shouldbe given particular consideration.

Any project of this nature includes both invest-ment in the infrastructure itself and in the institu-tions and capacity required to operate itappropriately and sustainably. Properly designedmultipurpose dams necessarily include concomi-tant investments in watershed management.Investments should be targeted and co-located tocapture synergies. Because Ethiopia’s most promis-ing development sites are on transboundary rivers,continued international cooperation will be essen-tial. This cooperation could strengthen regional re-lations and economic integration, bringingadditional benefits.

The potential costs and complications that mayarise as a consequence of safeguard requirementsand potential opposition to dam development mustbe accepted by both parties before major effortsbegin. If and when the parties are satisfied that thedevelopment benefits outweigh the costs of engage-ment, the partnership will hold great promise.

Drought and Flood Control

Water storage capacity will immediately providethe country’s water resources managers with apowerful and direct means to mitigate the devas-tating impacts of drought and floods. To fully uti-lize this capacity, concomitant investments in earlywarning systems and capacity building are needed.In addition, the river regulation provided by thisinfrastructure could be used to develop productiveuses such as hydropower, irrigation, and fisheriesand to safeguard ecosystem needs by ensuring ad-equate environmental flows.

Hydropower Development

Ethiopia’s untapped hydropower potential is enor-mous, and many of the potential development sitespose moderate environmental and social impacts.

The development of hydropower is a noncon-sumptive use of water that would provide a clean,renewable source of energy to help fuel theEthiopian economy. The availability of electricitycould reduce fuelwood reliance, with all of the as-sociated benefits to the environment, women, andchildren previously discussed. Reliable energy isan essential factor in providing the means and in-centives to transform the economy.

While the current expressed demand for energyis small, historically most countries have success-fully developed energy infrastructure ahead of de-mand. In a country where less than 15 percent ofthe population has access to modern energy, it isappropriate to assume that demand will grow.Given this extraordinarily low level of energy ac-cess, it might be appropriate to consider the provi-sion of energy as a basic need analogous to roadsand water supply, a crucial component of the min-imum infrastructure platform needed to transformthe country’s subsistence economy into a more vi-brant and diversified one.

There also appears to be real potential for energyexport to Djibouti, Sudan, and Egypt, which couldhelp justify significant investments in hydropowerdevelopment beyond Ethiopia’s immediately pro-jected needs. Having trade relations could also helpstabilize regional relationships and integrate re-gional economies. Under the NBI, Ethiopia is ex-ploring opportunities to develop its hydropowerpotential cooperatively with its downstream neigh-bors, Sudan and Egypt. Downstream neighborscould also benefit from Ethiopian hydropower de-velopments because hydropower is a noncon-sumptive use of water and operation of the plantcould regulate downstream flows and diminishsediment loads.

Irrigation and Drainage Development

The model developed in this study suggests that re-turns to irrigation and drainage are routinely un-derestimated when hydrological variability is notexplicitly incorporated into economic modeling.Irrigation opportunities exist in many parts of thecountry. Some indications suggest that the poten-tial for cost-effective irrigated cereals is limited,and that irrigation investments will be most attrac-tive for the production of higher value cash crops.However, the potential for irrigation and drainageinvestments is clear, and these benefits will growdramatically when concurrent investments aremade in market infrastructure and private sec-

tor–friendly policies are strengthened. If power be-comes available locally and domestic fertilizer pro-duction grows, agriculture may becomeincreasingly profitable. Similarly, if agri-processingindustries expand in response to both energy andinfrastructure availability, demand for raw materi-als should grow.

Synergies: Greater Than the Sum of the Parts

The development of multipurpose dams, by defin-ition, will capture many of the synergies discussedhere. Co-located investments in market infrastruc-ture and targeted watershed activities, discussedbelow, will also greatly strengthen the impact ofmultipurpose investments. Such development canput in place the opportunities and the incentivesneeded to provide more resilient and more diversi-fied economic growth.

Caveats

The Bank’s rigorous safeguard policies are an im-portant assurance of the quality of its work. At thesame time these requirements are sometimesviewed as onerous by our clients and may be per-ceived to slow project preparation and add costs tothe bottom line. It must also be anticipated that de-velopment of dams in Ethiopia will meet with sig-nificant scrutiny and potential opposition locallyand internationally, particularly if it is found thatlarge dams are the best option for certain sites.These risks must be clear before the Bank engages,and both parties must be willing to accept them be-fore major efforts begin.

World Bank’s Comparative Advantage

The Bank has a strong comparative advantage insupporting Ethiopian development of multipur-pose dams, demonstrated most recently by thesuccess of Gilgel Gibe. Multipurpose dams,whether large or small, can be complex and costly,requiring considerable financing and expertise.Many of the most promising development sitesare in the Nile River Basin, which, historically,made them difficult to develop. Even now withthe impressive successes of the NBI, developmentof the sites will require significant diplomacy, par-ticularly if regional partners are to be deeply in-volved. The Bank, as the lead donor and facilitatorof the NBI, is uniquely well placed to helpEthiopia develop these sites within the frameworkof the NBI.

The far-reaching potential benefits of multipur-pose dam development, and the unique qualifica-tions of the Bank to support these investments,strongly recommends support of multipurposedam development with an emphasis on hy-dropower generation and regional interconnectionas a first priority for future Bank assistance in waterresource management.

Water Supply and Sanitation

Investments in water supply and sanitation shouldaim to help Ethiopia achieve the MDG of reducingits unserved population by half by 2015. This willrequire more financing as well as a concerted effortto strengthen institutions and national capacity forimplementation by public and private sector actors.

To ensure that funds can be channeled more ef-ficiently and effectively, the GoE has developed astrong policy and strategy for reform. Its invest-ment program now has “rules of engagement,”streamlined implementation arrangements, andprinciples of monitoring and evaluation to improvetransparency and accountability. In addition, it isessential that water resource concerns, such asgroundwater management, siltation, and water-shed management, be integrated into water supplyand sanitation projects.

Caveats

Past interventions by the GoE and various donorshave not always been well coordinated, resulting ininefficiency and low sustainability. The Bank re-cently approved the $100 million Water Supplyand Sanitation Project to help finance implementa-tion of the GoE’s water supply and sanitation in-vestment program. This project represents onlyone-fifth of the level of investment needed over thecoming five years for Ethiopia to be on target forachieving the MDGs for water supply and sanita-tion. It primarily supports capacity building in thepublic and private sector through training and on-the-job support.

World Bank’s Comparative Advantage

The Bank’s ongoing involvement in the WaterSupply and Sanitation Project points to the need tocontinue its support to the sector as well as increaseefforts to attract additional financiers and help co-ordinate support to Ethiopia in meeting the watersupply and sanitation MDGs.

68 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

A Water Resources Assistance Strategy for the World Bank 69

Watershed Management

Good watershed management can slow soil ero-sion in the watersheds, moderate runoff and po-tential flood damage, and improve infiltration ofwater into soils and groundwater acquifers. Thesechanges help moderate the hydrological cycle andsustain agricultural productivity. In addition theycould reduce sedimentation rates, thus extendingthe useable capacity of reservoirs and reducingwater treatment plant costs.

Moderate Hydrology

Effective watershed management can help regulaterunoff and augment groundwater recharge, limit-ing the impacts of drought and flood on agricultureand reducing costly flood damage to roads andother transport infrastructure.

Mitigate Land Degradation and SlowSedimentation in Rivers and Reservoirs

Appropriate watershed management can also re-duce soil erosion, help to sustain agricultural liveli-hoods, and diminish the costs and increase the life ofdownstream reservoirs and irrigation infrastructure.

Caveats

There have been many efforts to strengthen water-shed management in Ethiopia, undertaken by awide range of actors. These experiences have high-lighted the need to effectively include stakeholdersin the planning, design, and management of wa-tershed interventions. An important lesson fromthe analysis in this study is that watershed man-agement may significantly promote economicgrowth if the current pattern of settlement and eco-nomic activities is maintained. Watershed manage-ment efforts should be designed to address the coreissues—sustainable livelihoods and ecosystems—and may require real innovations in methodology,financing, alternative livelihoods, or migration.

World Bank’s Comparative Advantage

Many groups support the GoE in watershed man-agement. For some projects, however, the Bank hasa comparative advantage. First, in any watershedwhere the Bank is supporting a multipurpose damproject, it should implement a watershed manage-ment plan. The Bank would have a clear advantagein designing and implementing the two, either astwo components of a single project or two stand-

alone projects if, for example, the watershed man-agement project began implementation prior to thefinalization of the dam project. Analogously, Bank-supported road investments might also benefit fromassociated watershed management investments.

The Bank might also have a comparative advan-tage in projects that confer significant benefits todownstream countries. Given the Bank’s involve-ment in the NBI, this advantage could be particu-larly pronounced with regard to sedimentationbenefits on the Nile. Finally, the Bank may have acomparative advantage in watershed managementinvestments that are eligible for carbon offsets, anew and promising area of potential finance.

Most immediately, a cooperative watershedmanagement project is under consideration withinthe framework of the NBI. This should be given pri-ority support in the Bank’s portfolio.

Transport, Market Infrastructure, andPrivate Sector Development: Growth Poles

Although investments in transport, market infra-structure, and private sector development are gen-erally considered to fall outside the realm of waterresources management, they are powerful hydro-logical risk-management investments. Investmentin these areas will help diminish the internal frag-mentation of the Ethiopian economy, which com-pounds the impacts of regional water shocks andhelps shift the economy toward a more water-re-silient structure. Their scale, scope, and priorityshould therefore be enhanced.

Transport disruptions and interregional trade

Improving the coverage and quality of Ethiopia’sroad network will make the economy less vulnera-ble to transportation disruptions caused by floods,ensuring that producers can move their product tomarket. Interregional roads would facilitate thegrowth of trade between regions, bringing surplusproduction to deficit areas, helping to smoothweather-related price shocks, and promoting amore integrated national economy.

Economic diversification

In the medium term, this study suggests that in-vestments in fully functioning market cities withregional infrastructure and communications inter-connections could be particularly effective in pro-viding alternatives and incentives to shift

vulnerable populations toward more resilientlivelihoods. It also suggests that the quality ofroads is essential, and thus investments in all-weather roads should take increasing priority.

Investments in fully functioning market cities(growth poles) with regional infrastructure and

communications interconnections are potentiallypowerful investments for managing hydrologicalrisk in Ethiopia. The Bank should therefore supportthe development of growth poles with adequatehydraulic infrastructure, all-weather roads, andmarket infrastructure.

70 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

71

The model is explicitly dynamic, running for a pe-riod long enough to explore the implications of hy-drological variability and drought or flood eventsover time. In constructing the model considerationwas given to the fact that Ethiopia’s economy andhydrology are characterized by:

• Highly variable rainfall and endemicdroughts and floods

• Fragile and degrading landscapes• Low levels of infrastructure investment to

mitigate against hydrologic variability• Fragmentation of the economy, particularly

regional segmentation in agricultural markets• High transportation and marketing costs

A national multisectoral and multiregionalmodel was modified to capture these characteris-tics. It builds on a spatial multimarket model that isenhanced to capture cross-sectoral linkages, partic-ularly between agricultural and non-agriculturalsectors. It also captures the important links be-tween water demand and supply and economic ac-tivity, focusing largely on agriculture. The modeldisaggregates agricultural commodities by regionof origin, where regions are defined to correspondas closely as possible to watersheds.

The model captures key features of theEthiopian context, such as weak or nonexistentlinkages between prices on world markets andprices in Ethiopian regions that are geographi-cally and economically “distant” from accesspoints to world markets. In such an environment,regional hydrological variability will have poten-tially more serious impacts, since the local impactcannot be buffered or ameliorated through marketlinks to other regions. There also may be signifi-cant threshold effects whereby prices must riseabove critical values before inducing trade with

other regions or fall below a certain level to gainaccess to world markets.

Regional differentiation was modeled by classi-fying nine development domains; that is, spatialrepresentations of three factors considered impor-tant for rural development: agricultural potential,market access, and land altitude.

The structure of the economy is modeled as in-cluding 34 disaggregated agricultural commoditiesand two aggregate non-agricultural commodities.Production of all the 36 commodities is further dis-aggregated into 56 zones. The supply function forthe 28 crop products at the zonal level is a combi-nation of a yield function (a function of own priceand productivity parameters) and an area function(a function of the prices for all the 36 commodities,including the prices for the two aggregated non-agricultural commodities and a shift parameter).The supply functions for the livestock and non-agricultural sectors depend on the prices for all thecommodities and productivity parameters. Whileconstraints on total agricultural land or other re-source factors are not modeled explicitly, themodel does impose constraints on the elasticitiesemployed in the supply functions in order to avoida simultaneous increase in the output of all prod-ucts in a given year.

At this stage of the model’s development it can-not explicitly address:

• The dynamics of food aid• The direct impact of hydrology on power• The tradeoff between hydropower and agri-

culture

Hydrology is incorporated into the model pri-marily as an input in agricultural productionthrough the agricultural yield function. Multi-market models typically specify only commodity

Appendix 1. Technical Notes

on the Model

supply functions (that is, output as a function ofprices rather than inputs) and, hence, do not ex-plicitly specify production technology. In thismodel, technology differences are captured by incorporating two types of supply (or yield) func-tions for the same crop: one production functionfor irrigated and a second for rainfed yields. Thedifference in technologies is reflected by the yieldgap between the two supply functions for thesame crop.

DATA FOR HYDROLOGICALVARIATIONSThe smoothed rainfall variation assumes1995–2002 average rainfall in all years. The stylizeddrought variation models the variability of rainfalland its effect on the economy as exogenous shocksto the yield and area only for rainfed crops.

Rainfall shortages are modeled as deviationsfrom a 100-year mean rainfall level calculated foreach of the 56 zones. We assume that the droughtaffects both the yield and area cultivated, and thenegative effect at the zonal level is assumed to beproportional to the value of the rainfall deviation ineach zone during the benchmark drought of1997–8. The difference between the results derivedfrom the smoothed rainfall scenario and the styl-ized drought scenario were interpreted to reflectthe impact of the drought on economic perfor-mance and poverty projections and, also impor-tantly, the significant impact of explicithydrological modeling on the predictive power ofthe model itself.

The historical variability variation captures an-nual rainfall fluctuations whose associated cumu-lative losses affect growth performance andpotential over time. An ensemble of historical hy-drographs was used to derive yield functions re-flecting the impact both of drought and of flood ina given year. Nine 12-year combinations weretaken from the 100 years of recorded data available:1900–12, 1913–24, 1925–36, 1937–48, 1949–60,1961–72, 1973–84, 1985–96, and 1989–2000. All ofthe combinations were modeled to start from thesame base year, 2003, for ease of comparison. Thisstochastic model provides a range of results ratherthan a single result for each simulation.

DATA FOR POVERTY RATES AND GDPFor all scenarios:

• To model rural incomes and poverty, percapita demand for rural and urban areas wasspecified for each commodity at the zonallevel as functions of all commodity prices andzonal-level rural and urban per capita in-comes. It was constrained by zonal-level ruraland urban budget constraints such that aver-age rural and urban per capita expenditureson all commodities were equal to the averagerural and urban per capita income.

• Zonal level rural and urban total incomes areendogenously determined in the model andequal to the sum of production revenues fromzonal level agricultural and nonagriculturalactivities. The average rural and urban percapita income at the zonal level is obtained bydividing the zonal level total income by itsrural and urban population. The populationgrowth rate is assumed the same across zones.Since the model does not take into account ofintermediate input costs, prices for the agri-cultural goods are adjusted such that totalagricultural revenue is close to the value ofagricultural GDP. Together with the twononagricultural sectors, which represent in-dustrial and service GDP, total income equalsthe country’s total GDP.

• The poverty rate in the initial year (2003) istaken from the GoE’s official poverty rate basedon 1999/2000 Household Income, Consump-tion, and Expenditure Survey. According to thesurvey, the national absolute poverty rate is44.2 percent, with a rate of 45.5 percent in ruraland 36.9 percent in urban areas.

DATA FOR AGRICULTURE AND TRADEWe assume an integrated national market withdifferent marketing margins across zones. Thusfor each individual commodity, total supply at thenational level, net trade, equals total demand atthe national level. Marketing margins reflect both

72 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Appendix 1. Technical Notes on the Model 73

the distance from Addis Ababa, a proxy for mar-ket access, and food deficit or surplus situationsfor each region. For a commodity produced in aregion with a food surplus, producer prices willbe lower than that in the Addis Ababa market,and the farther from the capital city, the lower theprice. For the same commodity produced in a fooddeficit region, consumer prices will be higher thanthat in the capital city, and the farther from thecity, the higher the price.

The model also considers the price gap betweendomestic and international markets. The importparity prices are defined as border prices (CIFprices) plus transportation and other marketingcosts from the port to Addis Ababa. Export parityprices are FOB prices minus transportation andmarketing cost. For a specific commodity, if tradedata show the commodity is imported, we assumethat the domestic price faced by consumers ishigher or equal to the import parity price.Similarly, if trade data show the commodity is ex-ported, domestic producer prices are assumed tobe lower or equal to the export parity price. Forthese traded commodities, domestic prices are

linked to world prices, and hence are exogenouslydetermined in the model.

For most agricultural products, such as pulses,roots, oil crops, and livestock products, the tradedata show no significant imports or exports. Wetherefore assume that they are not initially traded,and domestic prices are modeled as endogenouslydetermined in the domestic markets at a levelbelow the import parity price but above the exportparity price. If growth in domestic supply cannotmatch growth in domestic demand, which ismainly driven by population growth, prices forsome of these products will rise in the domesticmarket and imports may occur over time.

Productivity increases over the period are basedon 1995–2002 average growth rates in individualcrop yields, area expansion, livestock production,and the industrial and service sector GDP. Duringthis period, about 80 percent of the increase in totalcrop production was due to area expansion, whichoccurred at an annual growth rate of about 1.5 per-cent. Growth in crop productivity (yield) has beenextremely slow, reaching only about 0.5 percentper year.

75

Appendix 2. Water-Related Programs

in Ethiopia

Ongoing and Upcoming Irrigation Projects

Ongoing study/design projects Status (2005) Hectares

Kesem Design 30,000Tendaho Design 60,000Gumera Feas/Design 13,976Arjo Dedessa Feas/Design 14,280

Upcoming study/design projects (2005–10) Level of study/design Hectares

Megech Feasibility/Design 31,821Ribb Feasibility /Design 19,925Anger Feasibility /Design 14,450Negesso Feasibility /Design 22,815Humera Feasibility /Design 42,965Errer and Gololch Feasibility /Design 14,000Angereb Feasibility/Design 16,535Upper Beles irrigation project Feasibility/Design 53,700Bilate (Rift Valley) irrigation project Feasibility/Design 7,800Gilgel Abbay irrigation project Feasibility/Design 23,577

Ongoing construction projects Status (2005) Hectares

Kesem Construction 30,000Tendaho Construction 60,000Koga Construction 7,000

Upcoming construction projects (2005-10) Status (2005) Hectares

Part of Gode West Construction 6,600Gumera Construction 13,976Arjo Dedessa Construction 14,280Megech Construction 31,821Ribb Construction 19,925Anger Construction 14,450Negesso Construction 22,815Humera Construction 42,965Errer and Gololch Construction 14,000Angereb Construction 16,535Upper Beles Construction 53,700Bilate (Rift Valley) Construction 7,800

Source: MoWR.

76 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Ongoing and Upcoming Hydropower Projects

Ongoing study projects Level of of study Installed capacity (megawatts)

1 Baro Hydropower Project Feasibility study 8002 Karadobi Multipurpose Project Prefeasibility study 1,7003 Genale GD-3 Multipurpose Project Feasibility study 2454 Genale GD-6 Hydropower project Prefeasibility study 2445 Wabe Shebele WS-18 Multipurpose Project Feasibility study 87

Upcoming study/design projects (2005–10) Level of study/design

1 Baro Hydropower Project Design 8002 Genale GD-3 Multipurpose Project Design 2453 Wabe Shebele WS-18 Multipurpose Project Design 874 Karadobi Multipurpose Project Feasibility/design 17005 Genale GD-6 Hydropower Project Feasibility/design 2446 Mendaia Hydropower Project Prefeasibility/ feasibility 1,7007 Border Hydropower Project Prefeasibility/ feasibility 1,7808 Tams Multipurpose Project Prefeasibility/ feasibility 1,0209 Mabil Hydropower Project Prefeasibility/ feasibility 1,44010 Birbir-R Hydropower Project Prefeasibility/ feasibility 46711 Dobus Multipurpose project Prefeasibility/ feasibility 74112 Didesa Hydropower project Prefeasibility/ feasibility 308

Ongoing construction projects (2005) Year of completion

1 Tekeze/TK_5 2008 3002 Gilgel Gibe-II 2008 420

Source: MoWR.

Appendix 2. Water-Related Programs in Ethiopia 77

Project Project starting ending

Project Donation Loan Donor date date Project location Collaborators Remarks

Aquifer studies for 4.45 - JICA 1998 2005 To train MoWR A number of staffpotable water supply regional staff in the regions

trained

Awash River Flood - AfDB 2002 2005 On the MoWRControl studies Awash River

Biodiversity strategy 5.00 - UNDP/GEF 1999 2005 At federal level IBCR Conservation and action plan use of biodiversity

resources for sus-tainable agriculture

Dynamic farmers 2.50 - UNDP/GEF 1995 2004 ORS, SNNP IBCR with Conservation of land region, TRS, the regional races in drought

AfRS offices prone areas and use of the materials

Emergency drought 15.00 - World Bank 2003 2006 35 drought MoARD, Community-driven recovery project affected woredas Woreda development

in AfRS, TRS, Bureau of SNNP regional Agriculture

state, ORS, AfRS, and and SRS Communities

Ethiopia Nile 3.00 - AfDB 2003 - AfRS, ORS, MoWR, EMABasin Conservation and CIDA TRS, SNNP Study Project regional states,

BRS, GRS

Ethiopian Nile 1.56 - French 2002 2005 Nile Basin MoWR, CSA, Based on the results, Basin Development government EGS, and holistic develop-Authority Project MoARD mental approach to

be followed

Food security 2.18 - BMZ 2002 2005 AfRS,ORS, TRS, MoFED Giving support to capacity building and SNNP food security

regional states coordination offices

Food security - 85.0 WB 2002 2009 AfRS, ORS, TRS, MoARD with Community-driven project and SNNP the four development

regional states regional agricultural

offices

Forest management 7.50 - BMZ 1995 2004 ORS GTZ and Develop feasible in Adaba-Dodola ORS Bureau approach for conser-

of Agriculture vation of forests

Genalle-Dawa 0.52 - AfDB 2001 2005 Southern east of MoWR, Basin master plan Ethiopia (ORS, EMA, studies SNNP regional EGS, and

states, and SRS) NSL

Budget in US$(million)

Donor-Supported Watershed-Related Projects in Ethiopia

78 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

Project Project starting ending

Project Donation Loan Donor date date Project location Collaborators Remarks

Household energy 3.30 - BMZ 1998 2005 AfRS, ORS, TRS, GTZ , GTZ assists MoARD and protection of and SNNP MoARD research and devel-natural resource regional states opment and imple-

menting agencies in four regions to increase biomass energy efficiency

Hydropower site 1.44 - Norwegian 1995 AfRS and ORS MoWR, studies government regions

Koga Irrigation 1.35 39.69 AfDB 2002 2007 Wechecha MoWR, Good progress has Development Woreda, EMA, and been achievedProject Wastern Gojam AfRS

Land use planning 8.23 - BMZ 1997 2006 ORS GTZ, ORS Support and dis-and resource land and seminate integrated management in natural watershed manageORS resource ment

office

Microenterprise 17.00 - USAID 2002 2007 AfRS Virginia Integrated food development and Technical security and environ-agricultural University mental protection in extension and AfRS the Amhara region

offices

Pastoral Community 60.00 - World Bank 2003 2008 AfRS, SRS, ORS, MoFA and Community-driven Development and SNNP the four developmentProject regional states regional

agricultural offices

Promotion social 10.50 - BMZ 1997 2005 TRS Bureau of Extend services with forestry in Tigray Agriculture agro-forestry actions

and Natural to farmersResources of Tigray

Small irrigation 15.61 31.53 Irish and 1999 2005 AfRS, ORS, TRS, MoWR, Small-scale irrigation projects French and SNNP regions projects started in

governments regional states four regions

Small-scale - 23.5 World Bank 1996 2004 In all the regions ESRDF with Demand-driven com-irrigation projects except SRS communities munity investment;

communities have to be prepared to cover 10% cost

Support to the 1.70 - BMZ 1998 2004 At federal level IBCR Preserve forest Biodiversity Institute genetic resources as

a requirement for sustainable utilization

Budget in US$(million)

Donor-Supported Watershed-Related Projects in Ethiopia (continued)

Appendix 2. Water-Related Programs in Ethiopia 79

Project Project starting ending

Project Donation Loan Donor date date Project location Collaborators Remarks

Sustainable 4.50 - UNDP 2003 2006 Regional and EAP Policy formulation environmental federal EAP and implementation management offices of the policies

Technology 15.00 - CIDA 2004 2008 All regional MoARD, Launched in transfer and market states ILRI, July 2004accessibility and regions

The rehabilitation 11.00 - Belgian 2004 2009 TRS Katholieke Improved agricultural of dryland agri- government Unviersity, management and culture and forests Leuven; environmental pro-in Ethiopia: ecology Mekele tectionand management University;

RES

Water resources 5.00 - UNDP 2002 2007 In 10 regions MoWR with Rehabilitation of development and the regional irrigation schemes, utilization offices water supply, and

sanitation

Notes: JICA, Japan International Cooperation Agency; AfDP, African Development Bank; UNDP/GEF, United Nations DevelopmentProgramme/Global Environment Facility; CIDA, Canadian International Development Agency; BMZ, German Federal Ministry for EconomicCooperation and Development; USAID, U.S. Agency for International Development. ORS, Oromia national regional state; TRS, TigrayRegional State; AfRS, Afar Regional State; SRS, Somali Regional State; BRS, Benishangual Gamuz National Regional State; GRS, GambellaNational Regional State; EAP, East Asia and the Pacific.IBCR, Institute of Biodiversity Conservation and Research; MoARD, Ministry of Agriculture and Rural Development; EMA, Ethiopian MappingAuthority; CSA, Central Statistical Authority; EGS, Ethiopian Geological Survey; MoFED, Ministry of Finance and Economic Development;GTZ, German Technical Cooperation; NSL, National Soils Laboratory; ESRDF, Ethiopian Social Rehabilitation and Development Fund; ILRI,International Livestock Research Institute; REST, Relief Society of Tigray.Source: A desk study by Abiy Astatke for the ENSAP watershed management project background paper.

Budget in US$(million)

Donor-Supported Watershed-Related Projects in Ethiopia (continued)

81

Appendix 3.World Bank Projects in

Ethiopia

World Bank–Supported Projects in Ethiopia

Current portfolio

Lending:

P050383 Food Security Project Total costs: Other 2002The development objectives are to build the resource base US$110.16 million; agricultureof poorer rural households, increase their employment and IDA US$85.0 millionincomes, lower the real costs of food, and improve nutrition levels especially for children under five years of age and pregnaant and lactating mothers (by e.g., better management of rainfed agriculture; investment in small-scale irrigation; better natural resource management through “zero tillage” techniques on farms; and watershed-level activities to conserve soil, reverse soil degradation, improve water harvesting, and so on)

P081773 Emergency Drought Recovery Project Total costs: Food 2003The objective is to enable the GoE to help affected families US$61.7 million; Security, survive the crisis, retain productive assets, and develop IDA US$60.0 million Natural sustainable livelihoods, as well as contribute to stabilizing Resource the macroeconomy. The main outcomes of the project will Managementbe to contribute to: (1) defend critical human development expenditure by the GoE during the next year from the pressures caused by the current crisis; (2) maintain the avail-ability of critical economic inputs (particularly those needed in rural areas) and support the economic recovery process; (3) rehabilitate and/or create communal assets using commu-nity labor of the affected population; and (4) provide income supplements and protect private assets of the affected popu-lation. The project uses a community-driven development approach, and, based on community priority investment needs, small-scale water development operations could be financed. The Food Security Office under the newly restruc-tured Ministry of Agriculture and Rural Development is managing the project.

P000736 Energy II Total costs: Power and 1998The objectives are to increase the efficiency and sustainability US$295.0 million Energyof Ethiopia’s power sector and to increase electricity use for IDA US$200.0 millioneconomic growth and improved quality of life and to improve utilization efficiency of rural renewable energy. The project comprises the following components: (1) decentralization of

continued

82 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

P000736 the Ethiopian Electric Power Corporation’s (EEPCO) account-(continued) ing system; (2) decentralize part of EEPCO’s billing system as

a pilot, which requires branch office renovation, computer, software, and accessory installation, as well as other training and consulting services; (3) manpower development to support EEPCO’s restructuring to give more control to regional man-agement; (4) construction of the Gilgel Gibe hydroelectric plant to increase generation capacity; (5) establishment of a dam safety, maintenance management system, and optimiza-tion of water use; (6) strengthen rural energy capability by adapting technology and training research personnel overseas in various energy fields; (7) complete the Woody Biomass inventory data collection and remote sensing survey indica-ting the extent and productivity of grazing lands and under-taking a livestock carrying capacity analysis; (8) studies including sector restructuring study, a generation expansion study, a tariff study, and a valuation of fixed assets; and (9) assistance converting the state-owned utility (EEPCO) into a commercial operation, as well as other restructuring measures.

P049395 Energy Access Project — PE Total costs: Power and 2003The objectives are to: (1) establish a sustainable program for US$199.12 million; Energyexpanding the population’s access to electricity and improving IDA US$132.7 millionthe quality and adequacy of electricity supply, thus supporting broad-based economic development and helping to alleviate poverty; (2) reduce environmental degradation and improve energy end-use efficiency; (3) reduce the barriers to the wide-spread adoption of renewable energy technologies, in particu-lar solar photovoltaic and micro hydropower generation in rural areas, thereby contributing to the reduction in green-house gas emissions via displacement of kerosene and diesel that would otherwise be used for lighting and electricity generation; and (4) provide technical for institutional and capacity building of key sector agencies, including for regulatory, fiscal, and institutional reforms in the mining sector.

P000733 Agricultural Research and Training ProjectThe project will make the Ethiopian agricultural research system more efficient and sustainable by: (1) building institu-tional capacity to manage the entire agricultural research system; (2) rationalizing the way the programs are identified, designed, and prioritized according to an acceptable eco-nomic criteria; (3) improving their relevance through decen-tralization; (4) drawing on the relevant pool of knowledge available internationally; (5) consolidating and strengthening the existing research centers; (6) expanding research facilities into less favorable agro-ecological settings; and (7) building human resource capacity in collaboration with the Alemaya University of Agriculture. There are three components in this project. First, agricultural research management will assist in operationalizing the Ethiopian Agricultural Research Organization, designing and implementing a decentralized research system, implementing new systems and procedures for efficient agricultural research management, and develop-ing a strong program of research-extension-farmer linkages.

World Bank–Supported Projects in Ethiopia (continued)

continued

Appendix 3. World Bank Projects in Ethiopia 83

P000733 Second, strengthening of the agricultural research system Total costs: Research 1998(continued) will rehabilitate and expand the existing research system to US$60.0 million;

make its operation more efficient and its work environment IDA US$60.0 millionmore conducive to research work. Third, human resource development will provide technical assistance and training for the researchers in the research system and those at the university.

P000764 Water Supply Development and Rehabilitation Project Total costs: Urban water 1996The main objectives are to ensure the long-term viability of US$65.5 million; supplywater supply and sanitation operations in line with the GoE’s IDA US$35.7 millionregionalization policies and, in the long run, improve the health and productivity of the population. The objectives will be met by providing assistance for: (1) capacity building of the regional governments and water supply and agencies for management of urban water supply and sanitation schemes and rural water supply operations; (2) formulation of policies to ensure long-term financial and managerial viability of water supply operations, establishment of regulatory arrange-ments, and sound investment planning by the regional governments and water supply agencies; and (3) short- and medium-term physical rehabilitation, augmentation, and establishment of urban water supply and sanitation schemes in war-affected and the most deficient regions. The project will consist of the following components: (1) institutional capacity building, which includes water supply sector capacity building and rural water supply pilot scheme (this component will provide studies, technical assistance, training, equipment, service rigs, mobile workshops, vehicles, and the construction of pilot schemes); (2) feasibility and engineering, which will consist of water and sanitation scheme feasibility, engineering designs, and building works designs; and (3) physical works comprising water and sanitation scheme rehabilitation as well as building works.

P000771 Social Rehabilitation Project IDA US$120.0 million Social 1996The project will provide to poor, mainly rural communities protectionthe assets and services needed to improve their economic and social standards and will provide financial and technical sup-port to communities and community groups to construct or rehabilitate and maintain basic economic and social infra-structure and services that they have prioritized and in which they are willing to invest. In addition, it will support commu-nity actions aimed at environmental conservation and reha-bilitation. Efforts will be focused on poor communities and, especially, poor women. To achieve these goals, the project will strengthen community capacity in project identification, implementation, and maintenance. The project will also support the establishment of a Welfare Monitoring System to strengthen capacity to collect and analyze information related to causes, patterns, and incidence of poverty and to improve the basis for developing poverty reduction strategies, assessing the impact on the poor of economic policy reform, and evalu-ating the benefits of other poverty reductions interventions. The components of this project will be as follows: (1) construc-tion and rehabilitation of social and economic infrastructure

World Bank–Supported Projects in Ethiopia (continued)

continued

84 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

P000771 and improvement of related services, including basic health (continued) and education, small-scale irrigation, urban sanitation, and

rural water supply; (2) welfare monitoring system; and (3) capacity building, training, and research.

P075915 Pastoral Community Development Project Total costs: Community 2003The proposed project would consist of (1) the development US$60.0 million; action of an institutional framework that would empower pastoral IDA US$30.0 million programcommunities and facilitate demand-driven investments at community and zonal levels, (2) a social and economic investment fund with components for demand-driven micro-investments and for larger zonal infrastructure investments, and (3) a component for research and extension on a variety of social and economic issues

P073925 Capacity Building for Decentralized Service Delivery Program Economic and Sector Decentrali- 2003The proposed project aims to enhance decentralized service Work zationdelivery performance by initiating long-term public sector capacity building at the federal, regional, and local levels. Building public sector capacity through institutional reforms, systems development, and training would promote effective policy formulation and implementation, particularly in key sectors such as infrastructure, health, and education. The proposed project, envisaged as a first phase of long-term Bank lending and nonlending support, seeks to enhance service delivery performance through a coordinated program of: (1) federal civil service reforms, (2) regional capacity building, and (3) local government restructuring and empowerment. Over the life of the proposed project, key reform and capacity building activities will be piloted in selected federal, regional and local government entities.

AAA:

Standard:

P059215 Small-Scale Irrigation and Conservation Project Total costs: Irrigation 1999The project, supported by Credit l765-ET for US$7.0 million US$22.0 million; and equivalent, was approved in FY87. The project was cofinanced IDA US$7.0 million drainageby the International Fund for Agricultural Development (US$11.0 million equivalent), and the Organization of Petro-leum Exporting Countries Fund (US$4.0 million equivalent). The credit was closed in FY97. The project’s principal objec-tives were to increase food production in drought prone areas through the development of small-scale irrigation for 25,000 farm families and control land degradation in the highlands to benefit 300,000 families. There were five components: (1) institutional capacity building through technical assistance, staff training, and provision of equipment; (2) construction of small-scale irrigation infrastructure; (3) provision of agricultural support services; (4) soil conservation through biological bund stabilization; and (5) support for women’s activities. The project was implemented during a period of intense political, administrative, and social turmoil that dramatically slowed

World Bank–Supported Projects in Ethiopia (continued)

continued

Appendix 3. World Bank Projects in Ethiopia 85

implementation. The project, the institutional capacity-building components, were very successful in training a cadre of competent irrigation, agricultural extension, and soil conservation professionals in the line agencies. Farmers have formed themselves into water users’ associations that have taken responsibility for scheme operation and maintenance and achieved a substantial level of cost recovery.

Future portfolio

Lending:

P076735 Water Supply and Sanitation Project Total costs: Water and 2004By shifting from the centralized approach to identifying, finan- US$120.00 million; sanitationcing, and implementing water supply and sanitation projects IDA US$100.00 to a demand-driven approach, whereby local governments millionand communities would identify projects and appraisal and fund management would be carried out at the regional and/or woreda level, the project would support the GoE’s policies of decentralization of resources management to the lowest pos-sible local level and promote the involvement of all stake-holders so as to improve efficiency in service delivery. By requiring demonstration of financial and economic feasibility of proposed investment and financing, operations and main-tenance plans and tariffs phasing in full-cost recovery for urban water supply and sanitation project proposals, the project would support the GoE’s policy of recognizing water as an economic good and improve the level of cost recovery in both urban and rural water supply to ensure sustainability.

P074020 Public Sector Capacity Building Project

The objective is to improve the scale, efficiency, and respon- Total costs: Civil service 2004siveness of public service delivery at the federal, regional, and US$397.80 million; reform, local level; to empower citizens to participate more effectively IDA US$100.00 decentraliza-in shaping their own development; and to promote good gov- million tion, public ernance and accountability. This objective will be achieved financialby scaling up Ethiopia’s ongoing capacity building and institu- manage-tional transformation efforts in six priority areas: civil service ment, insti-reform, district-level decentralization, urban management tutional de-capacity building, tax systems reform, justice system reform, velopment, and information and communications technology. judicial

reform

AAA:

P076064 Irrigation note Economic and sector Irrigation 2002work and drainage

Standard:

P049250 SPEC.CTRY PROG. II Unidentified (standard Irrigation 2004International Fund for and Agricultural Develop- drainagement project)

World Bank–Supported Projects in Ethiopia (continued)

continued

86 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

P075405 Ethiopia: Poverty Reduction Strategy Paper Poverty Reduction Unidentified 2003For 2002–5; finished and circulated in August 2002. Prepared by the Ethiopian authorities and contains a declaration of the GoE’s commitment to poverty reduction, a description of the existing extent and patterns of poverty, the main elements of its existing poverty-reduction strategy, and a three-year macroeconomic framework as well as a three-year policy matrix. It presents a broad outline for reducing poverty that is grounded in the development strategy laid out in the National Development Program. The poverty reduction strategy is centered around promoting economic growth and increasing the income-earning capacity of the poor. It comprises four elements: (1) an agriculture development-led-industrialization strategy, and food security; (2) civil service and judicial reforms; (3) governance, decentralization, and empowerment; and (4) capacity building. It identifies preliminary quantitative targets for various dimensions of poverty, as well as targets for increases in average income.

Notes: IDA, International Development Association (of the World Bank Group).Source: Authors.

World Bank–Supported Projects in Ethiopia (continued)

87

Endnotes

1. The model includes 56 climatic zones, 34 agricultural com-modities, two nonagricultural commodities, and distin-guishes between irrigated and nonirrigated areas.

2. Here the term "institutions" incorporates capacity, organiza-tions, policies, rules, and agreements, and integrated waterresources management includes investment in water insti-tutions and infrastructure.

3. This is the message of the World Bank’s 2003 Water ResourcesSector Strategy.

4. See the World Bank’s recent (2005) Country EconomicMemorandum, “Ethiopia: A Strategy to Balance andStimulate Growth.”.

5. Carbon finance provide resources for projects that reducegreenhouse gas or carbon emissions (for example, throughclean energy or reafforestation).

6. These include the Abay (Blue Nile), Tekeze, Baro Akobo, andOmo Gibe basins, largely occupying the western and south-western parts of the country.

7. For instance, a farmer cultivating steep slope in DelantaDawunt reported to have lost 1 hectare of his 3 hectare landover a period of 10 years. The bedrock is all that remainsonce the top soil is washed away (Demeke et al. 2003).

8. There is debate whether the Awash River should be consid-ered a transboundary river shared with Djibouti.

9. Bojo and Cassells (1995) estimated a gross annual financialloss of US$117 million due to nutrient losses and a furtherUS$6 million due to soil erosion.

10. The correlation coefficient between national cereal yield andnational average rainfall is 0.338 (significant at slightly lessthan 10 percent level of significance).

11. Chat has been banned in the United States, Canada, and theUnited Kingdom.

12. Tropical livestock units = 250 kilograms live weight.13. For instance, the existing road distance between Bure (a

town in west Gojam, close to Wollega) and Dese (Wollo) is618 kilometers. The air distance between the two is only 298kilometers and the proposed direct road link (betweenGojam and Wollo) is estimated at 381 kilometers.

14. This calculation might well underestimate costs. The actualfinancial expenditures for the two consecutive years wereanalyzed to determine what percentage of the total mainte-nance cost went to water- related defects. Only those activ-ities that are purely related to water-related measures suchas ditch clearing, culvert maintenance, and so on are con-sidered, while on the other hand activities such as regravel-ing and resurfacing activities that may have been caused bywashing away of surface material or weakening of the sub-grade by capillary action and a combination of other factorswere not considered due to difficulties of refining the data.

15. Those designed and maintained for long-term vehicularuse.

16. When the power supply is too low to meet the demand ofall customers, utilities are forced to temporarily cut offpower to some of their consumers—or shed customerloads—in order to protect and stabilize the power system.This is termed “load shedding.”

17. To establish the economic potential of energy in Ethiopia,water regulation options were considered that requiredam/diversion weir lengths below 750 meters and damsheights below 120 meters. Installed capacity correspondingto output of 145,000 gigawatts per hour per year, at an av-erage plant utilization factor of 0.6, would be about 87,000.

18. Under the new arrangement (introduced in 2000) regionalgovernments pay from their budget for defaulting farmers.

19. More than 80 percent of the fertilizer sales are financedthrough government credit in Ethiopia. See, for instance,Demeke (2004).

20. Vertisols are dark black soils rich in expandable clay miner-als which swell and shrink substantially with moisture.

21. Teff is a staple crop used to produce traditional injera bread.22. Results for this model are taken from Diao et al. (2004).23. A stylized drought model was not run for this scenario be-

cause that variant of the model would show no impact. 24. Market infrastructure includes roads, transport and storage

facilities, communications, and other infrastructure invest-ments that will decrease the costs of transporting and mar-keting goods.

25. One must be cautious in comparing the results of these sce-narios. It must be recognized that the impacts of these twoinvestment tracks are somewhat interdependent. The twoscenarios are also drawing on different types of data andmaking different, heroic, assumptions. While informationabout approximate costs of irrigation investment perhectare is available, there is no similar information currentlyavailable for the costs of investment in roads or other mar-ket infrastructure. Direct comparisons cannot be drawn be-tween similar investments targeting irrigation versus roadnetworks. Moreover, it may be unrealistic to believe thatroads and market infrastructure can be improved simulta-neously nationwide over the next 12 years. Some regionsand farmers will neither be able to access improved roadsand markets or capture their benefits. However in the roadscenario there is less concern about potential negative priceeffects from increased supplies of grain and other cropproducts, because improving roads and market conditionscan make prices more favorable to both consumers andfarmers without affecting those farmers who cannot accessmarkets through price effects.

26. This section is drawn from Grey and Sadoff (2006).27. Here the term “institutions” incorporates capacity, organi-

zations, policies, rules, and agreements; integrated water re-sources management includes investment in waterinstitutions and infrastructure.

28. This is the message of the World Bank’s 2003 WaterResources Sector Strategy.

29. Watershed interventions are potentially supported throughcommunity driven development components under theSmall-Scale Irrigation Project, the Pastoral CommunityDevelopment Project, the Food Security Project, and theEmergency Drought Recovery Project.

88 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

89

Abate, Zewdie. 1995. Water Resources DevelopmentIn Ethiopia: An Evaluation Of Present ExperienceAnd Future Planning Concepts. Reading, UK:Ithaca Press.

Addis Ababa Water and Sewerage Authority. 2000.“Water Demand Management.” Report, AddisAbaba.

Amha, Wolday, and Eleni Gabre-Madhin. July2003. “An Analysis of the Structure, Conduct,and Performance of the Ethiopian GrainMarket.” Report, AEMFI, Ethiopia andInternational Food Policy Research Institute,Washington, DC.

Awulachew, S. B., D. J. Merrey, A. B. Kamara, B.Van Koppen, F. Penning de Vries, and E. Boelee.2004. “Experiences and Opportunities forPromoting Small-Scale Micro Irrigation andRainwater Harvesting for Food Security inEthiopia.” Working Paper 98. InternationalWater Resources Management Institute.August. http://www.iwmi.cgiar.org/pubs/working/WOR98.pdf.

BCEOM (in Association with BRGM and ISL). 1998.Abay River Basin Integrated DevelopmentMaster Plan Project.

BCEOM. 1999. “Abbay River Basin IntegratedDevelopment Master Plan Project: MainReport.” Ministry of Water Resources, FederalDemocratic Republic of Ethiopia, Addis Ababa.

Befekadu Degefe, Berhanu Nega, and GetahunTafesse. 2000/1. “Second Annual Report on theEthiopian Economy.” Section II, Volume II.Ethiopian Economic Association: Addis Ababa.

Berhane, Gulilat. 2003. “Present and Future WaterResources Development in Ethiopia Related toResearch and Capacity Building.” Paper pre-sented at the MoWR/EARO/IWMI/ILRIInternational Workshop, Addis Ababa,December 2–4.

Berkele, Yonis. 2002. “Magnitude of Famine inPastoral Areas of Ethiopia.” Paper presented atthe Roundtable on Drought and Famine in thePastoral Regions of Ethiopia, Pastoralist ForumEthiopia, Addis Ababa, December 23–4.

Bezabih, Mulugeta. 2003. “Is There ReliableStrategy to Benefit From Irrigation?” Nation(weekly Amharic newspaper) 2 (75).

Bojo, Jan, and D. Cassells. 1995. “Land Degradationand Rehabilitation in Ethiopia: AReassessment.” AFTES and ENVLW WorkingPaper 17, World Bank, Washington, DC.

Burgess, S. 1992. “Water Harvesting for CropProduction.” Paper presented at the SecondNational Conference on Rainwater CatchmentSystems, Kenya, Nairobi, August 30 toSeptember 4.

Campbell, J. 1991. “Land or Peasants? TheDilemma Confronting Ethiopian ResourceConservation.” African Affairs 90 (358): 5–21.

Cour, J. M. 2002. “Sustainable Development andPoverty Reduction Strategy Revisited: A Demo-Economic Conceptual Framework and itsApplication to Ethiopia.” Paper presented at theEthiopian Economic Association 2003International Conference on the EthiopianEconomy, Addis Ababa, December 19.

CSA (Central Statistical Agency). 2000. “Report onLarge and Medium Scale Manufacturing andElectricity Industry Survey.” Addis Ababa.

Dadi, Legesse, Asfaw Negassa, and S. Franzel. 1992.“Marketing Maize and Teff in Western Ethiopia:Implications for Policies Following MarketLiberalization.” Food Policy 17 (3): 201-213.

Demeke, M. 2004. “Hydrological Variability inEthiopia: Its Impact and Some StrategicConsiderations.” Background Paper forEthiopia Water Resources Assistance Strategy ofthe World Bank, University of Addis Ababa,Addis Ababa.

References

Demeke, M., B. Kassahun, and A. Berhanselassie.2003. “Socioeconomic Baselines Survey ofLivelihoods in Delanta Dawunt, North Wollo,Amhara Region.” Report, Oxfam GB, AddisAbaba.

Dercon, Stefan, and Luc Christiaensen. 2005.“Consumption Risk, Technology Adoption andPoverty Traps: Evidence From Ethiopia.” Draft,World Ban, Washington, DC.

Diao, Xinshen, Mark W. Rosegrant, PeterMcCornick, Charles Rodgers, and JordanChamberlin. 2004. “Irrigation vs. Roads:Investment Options in Ethiopian AgriculturalGrowth—A Spatial Multi Market Model Analysisfor 2004–2015.” Draft, International Food PolicyResearch Institute, Washington, DC.

EEPCO (Ethiopian Electric Power Corporation).2003. “Ethiopian Power System ExpansionMaster Plan.” Addis Ababa.

EEPRI (Ethiopian Economic Policy ResearchInstitute). 2003. “Quarterly Report on theMacroeconomic Performance of the EthiopianEconomy.” Addis Ababa.

Estifanos, Eyob. 2003. “Public Policy and PrivateInvestment in the Air Transport Sector.” Paperpresented at the Ethiopian EconomicAssociation 2003 International Conference onthe Ethiopian Economy, Addis Ababa,December 19.

Ethiopian Road Authority. 2000/2002. “RoadCondition Survey Report.” Pavement Manage-ment Branch, Addis Ababa, Ethiopia.

FDRE. 2002. “Ethiopia: Sustainable Developmentand Poverty Reduction Program.” Report,Ministry of Finance and Economic Develop-ment, Addis Ababa.

Fisseha, Messele. 2000. “Watershed Deteriorationin the Ethiopian Nile Sub-Basin and its Effectson the Water Resources.”

Gayessa, Abera. 1999. “Agricultural Marketing inEthiopia: A Review of Constraints and ProposedMechanisms for Improvement.” Paper pre-sented at the Workshop on Operation,Maintenance and Organization of IrrigationSchemes in Amhara National Regional State,Bahir Dar, October 18–20.

Gayessa, Abera. 2002. “Marketing Policy ForAgricultural Products In Ethiopia.” Paper pre-sented at the Regional Workshop, “Markets andMarket Information for Agricultural Products,”Nairobi, Kenya, January 21–5.

Gebeyehu, Admasu. 2002: “Research andDevelopment in Land and Water Resources.”Paper presented at the MoWR/EARO/IWMI/ILRI International Workshop, Addis Ababa,December 2–4.

Gebremedhin, Berhanu, and D. Peden. 2002.“Policies and Institutions to Enhance the Impactof Irrigation Development in Mixed Crop-Livestock Systems.” Paper presented at theMoWR/EARO/IWMI/ILRI International Work-shop, Addis Ababa, December 2–4.

Government of Ethiopia.1996a. Road SectorDevelopment Program. Addis Ababa.

Government of Ethiopia. 1996b. Social Rehabili-tation and Development Fund. Addis Ababa.

Grey, David, and Claudia Sadoff. 2002. “WaterResources and Poverty in Africa: EssentialEconomic and Political Responses.” Paper pre-sented at the African Ministerial Conference onWater, Abuja, Nigeria, April 29–30.

Grey, David, and Claudia Sadoff. 2006. “Water forGrowth and Development.” In ThematicDocuments of the IV World Water Forum, ed.David Grey and Claudia W. Sadoff, Mexico City:Comision Nacional del Agua.

HALCROW. 1990. “Master Plan for theDevelopment of Surface Water Resources in theAwash Basin.” Addis Ababa.

Hurni, H. “Degradation and Conservation of theResources in the Ethiopian Highlands.”Mountain Research and Development 8 (2/3):123–30.

Kebede, G. 1992. The State and Development inEthiopia. Highlands, NJ: Humanities Press.

Mebratu, Desta, and Mulugeta Tamire, eds.. 2002.Energy in Ethiopia: Status, Challenges, andProspects: Proceedings of the Energy Conference.Addis Ababa: Publisher.

MCE (Metaferia Consulting Engineers). 2001.“Assessment of Experiences and Opportunitieson Medium and Large-Scale Irrigation inEthiopia.” Background paper, World Bank,Washington, DC.

MCE (Metaferia Consulting Engineers). 2002.“Study on Research and Development Activitiesin The Water Sector.” Addis Ababa, Ethiopia.

Mesfin, Tafesse. 2002. “Immediate Causes for theFamine in Pastoral Areas: Immediate Causes forthe Famine—The Drought Dimension.” Paperpresented at the Roundtable on Drought andFamine in the Pastoral Regions of Ethiopia,Addis Ababa, December 23–4.

90 Ethiopia: Managing Water Resources to Maximize Sustainable Growth

References 91

Ministry of Trade and Industry. 2003. “Ethiopia:Trade and Transformation Challenges, Annex10.” Manufacturing Sector, Diagnostic TradeIntegration Study, Addis Ababa.

MoFED (Ministry of Finance and EconomicDevelopment). 2002. “Development andPoverty Profile in Ethiopia.” Report, WelfareMonitoring Unit, Addis Ababa.

MoWR (Ministry of Water Resources). 1996. “Omo-Gibe Basin Integrated Development Master PlanStudy,” Final Report, Executive Summary.Addis Ababa.

MoWR (Ministry of Water Resources). 2002a.“National Water Supply and Sanitation MasterPlan.” Interim Master Plan, DHV ConsultantsBV and DHV Ethiopia. Addis Ababa.

MoWR (Ministry of Water Resources). 2002b.“Water Sector Development Program.” Report,Addis Ababa.

MoWR (Ministry of Water Resources). “AbbayRiver Basin Integrated Development MasterPlan Project, Phase III.” Report, Addis Ababa.

NEDECO-Netherlands Engineering Consultants.1997. “Tekeze River Basin Integrated Develop-ment Master Plan Project.” Report, Location.

Sadoff, Claudia W., Dale Whittington, and DavidGrey. 2003 Africa’s International Rivers: AnEconomic Perspective. Washington, DC: WorldBank.

Sally, H. 2002. “Advances in Integrated WaterResources Management Research in Agri-culture.” Paper presented at the MoWR/EARO/IWMI/ILRI International Workshop,Addis Ababa, December 2–4.

Salman, Salman, and Laurence Bosisson deChazournes. 1998. “International Watercourses:Enhancing Cooperation and ManagingConflict.” World Bank Technical Paper 414,World Bank, Washington, DC.

Strezepek, Kenneth, Paul Block, and MarkRosegrant. 2004. “Analysis of InvestmentStrategies for Ethiopia: Effects of StochasticVersus Deterministic Modeling, PreliminaryResults.” Background paper, International FoodPolicy Research Institute, Washington, DC.

Tafesse, Mekuria. 2004. “Background Paper:Ethiopia Country Water Resources AssistanceStrategy of the World Bank.” World Bank,Washington, DC.

Thurow, Roger. 2003. “Changing Course: Ravagedby Famine Ethiopia Finally Gets Help From theNile/For Generations, Politics Kept TributariesFlowing By, Bringing Bounty to Egypt/ ParchedCorn, Rushing Water.” The Wall Street Journal,November 26.

Webb, Patick, Joachim von Braun, and YisehacYohannes. 1992. “Famine in Ehtiopia: PolicyImplications of Coping Failure at National andHousehold Levels.” Research Report 92,International Food Policy Research Institute,Washington, DC.

World Bank. 2002. “China Country Water ResourcesAssistance Strategy.” Washington, DC.

World Bank. 2003. “Water Resources SectorStrategy.” Washington, DC.

World Bank. 2005a. “A Strategy to Stimulate andBalance Growth in Ethiopia.” A CountryEconomic Memorandum Report 29383-ET,World Bank, Washington, DC. http://web.worldbank.org/WBSITE/EXTERNAL/COUN-TRIES/AFRICAEXT/ETHIOPIAEXTN/0,,contentMDK:20216510~hlPK:574361~menuPK:625036~pagePK:141137~piPK:217854~theSitePK:295930,00.html

World Bank. 2005b. “Well-Being and Poverty inEthiopia: The Role of Agriculture and Agency.”Washington, DC.