building resilience to climate shocks in ethiopia · 2020-03-05 · building resilience to climate...

EDITED BY JAWOO KOO, JAMES THURLOW, HAGAR ELDIDI, CLAUDIA RINGLER, AND ALESSANDRO DE PINTO

BUILDING RESILIENCE TO CLIMATE SHOCKS IN ETHIOPIA

FOOD POLICY

REPORT

BUILDING RESILIENCE TO CLIMATE SHOCKS IN ETHIOPIA

EDITED BY JAWOO KOO, JAMES THURLOW, HAGAR ELDIDI, CLAUDIA RINGLER, AND ALESSANDRO DE PINTO

A Peer-Reviewed PublicationInternational Food Policy Research InstituteWashington, DC

APRIL 2019

ABOUT IFPRIThe International Food Policy Research Institute (IFPRI), established in 1975, provides research-based policy solutions to sustainably reduce poverty and end hunger and malnutrition. IFPRI’s strategic research aims to foster a climate-resilient and sustainable food supply; promote healthy diets and nutrition for all; build inclusive and efficient markets, trade systems, and food industries; transform agricultural and rural economies; and strengthen institutions and governance. Gender is integrated in all the Institute’s work. Partnerships, communications, capacity strengthening, and data and knowledge management are essential components to translate IFPRI’s research from action to impact. The Institute’s regional and country programs play a critical role in responding to demand for food policy research and in delivering holistic support for country-led development. IFPRI collaborates with partners around the world.

This publication has been peer reviewed. Any opinions stated in this report are those of the author(s) and are not necessarily representative of or endorsed by IFPRI or UNDP. The boundaries, names, and designa-tions used in this publication do not imply official endorsement or acceptance by the authors, the UNDP, IFPRI, or its partners and donors.

Copyright © 2019 United Nations Development Programme. All rights reserved.

Photo credit: (cover) Petterik Wiggers/Panos Pictures

ISBN: 978-0-89629-359-5

DOI: https://doi.org/10.2499/9780896293595

https://doi.org/10.2499/9780896293595

CONTENTS

ACKNOWLEDGMENTS 6

LIST OF ACRONYMS, ABBREVIATIONS, AND TERMS 7

EXECUTIVE SUMMARY 9

INTRODUCTION 12

1. THE RESILIENCE LANDSCAPE 14Prapti Bhandary, Claudia Ringler, and Alessandro de Pinto

2. RESILIENCE BUILDING IN ETHIOPIA: ANALYSIS OF KEY INFORMANT INTERVIEWS 25Laia Domenech, Semhar Tesfatsion, Claudia Ringler, and Sophie Theis

3. EL NIÑO–SOUTHERN OSCILLATION IMPACTS ON AGRICULTURE AND THE NATIONAL ECONOMY 41Jawoo Koo, James Thurlow, Hua Xie, Ricky Robertson, Carlo Azzarri, Ho-Young Kwon, and Beliyou Haile

APPENDIX 1: MAIN RESILIENCE PROGRAMS IMPLEMENTED IN ETHIOPIA AND KEY FEATURES 66

APPENDIX 2: ACTIVITIES PROMOTED BY DIFFERENT RESILIENCE PROGRAMS AND ORGANIZATIONS 67

REFERENCES 69

ABOUT THE AUTHORS 76

TABLESTABLE 1 SOME DISTINGUISHING FEATURES OF ETHIOPIA’S HIGHLANDS AND LOWLANDS 26

TABLE 2 ESTIMATED NATIONAL GRAIN PRODUCTION LOSSES DURING THE 2015/16 EL NIÑO 52

TABLE 3 ESTIMATED GRAIN PRODUCTION LOSSES BY SUBNATIONAL REGION DURING THE 2015/16 EL NIÑO 52

TABLE 4 AGRICULTURE–FOOD SYSTEM SHARE OF GDP AND EMPLOYMENT, 2010/11 55

TABLE 5 ETHIOPIA’S NATIONAL ECONOMIC STRUCTURE, 2010/11 56

TABLE 6 HOUSEHOLD INCOME AND CONSUMPTION PATTERNS, ETHIOPIA, 2010/11 56

TABLE 7 GDP CHANGES DURING STRONG EL NIÑO EVENTS, AND INTERVENTION SCENARIOS 59

FIGURESFIGURE 1 BELG- AND KIREMT-DEPENDENT AGRICULTURAL AREAS IN ETHIOPIA 15

FIGURE 2 CHANGE IN MEAN DAILY MAXIMUM TEMPERATURE (OC) BETWEEN 1980 AND 2010 IN HOTTEST MONTH OF YEAR 18

FIGURE 3 DEPARTURE FROM AVERAGE PRECIPITATION IN DROUGHT AREAS, 1981–2016 19

FIGURE 4 PRECIPITATION FROM FEBRUARY 1 TO SEPTEMBER 15, 2015 (BELG AND KIREMT SEASONS) AS A SHARE OF LONG-TERM AVERAGE PRECIPITATION 20

FIGURE 5 ENABLING ENVIRONMENT KEY INTERVENTIONS 39

FIGURE 6 INTEGRATED ANALYTICAL FRAMEWORK 44

FIGURE 7 ETHIOPIA’S FIVE AGROCLIMATIC REGIONS 45

FIGURE 8 AVERAGE MONTHLY RAINFALL BY EL NIÑO–SOUTHERN OSCILLATION PHASE 46

FIGURE 9 DEVIATION IN NATIONAL RAINFALL PATTERNS DURING EL NIÑO–SOUTHERN OSCILLATION PERIODS 47

FIGURE 10 AVERAGE MONTHLY RAINFALL DURING EL NIÑO–SOUTHERN OSCILLATION PHASES ACROSS AGROCLIMATIC ZONES 48

FIGURE 11 SIMULATED YIELD DEVIATIONS FOR MAIZE AND WHEAT BY EL NIÑO–SOUTHERN OSCILLATION PHASE AND AGROCLIMATIC ZONE 49

FIGURE 12 YIELD DEVIATIONS WITHIN THREE AGROCLIMATIC ZONES DURING THE 2015/16 EL NIÑO 49

FIGURE 13 EFFECTS OF FARM TECHNOLOGIES IN RAISING YIELDS UNDER EL NIÑO AND LA NIÑA BY AGROCLIMATIC ZONE 50

FIGURE 14 ESTIMATING NATIONAL TEFF PRODUCTION LOSSES DURING THE 2015/16 EL NIÑO 51

FIGURE 15 ESTIMATED LIVESTOCK LOSSES BY SUBNATIONAL REGION DURING THE 2015/16 EL NIÑO 53

FIGURE 16 GDP LOSSES DURING THE 2015/16 EL NIÑO EVENT 59

FIGURE 17 HOUSEHOLD CONSUMPTION LOSSES DURING EL NIÑO BY POLICY SCENARIO (PERCENTAGE) 60

FIGURE 18 HOUSEHOLD CONSUMPTION LOSSES BY QUINTILE, WITH ALL INTERVENTIONS COMBINED (PERCENTAGE) 61

FIGURE 19 CHANGES IN NATIONAL POVERTY RATE AND POOR POPULATION DURING EL NIÑO BY POLICY SCENARIO 62

BOXESBOX 1 ELEMENTS OF THE GENDER, CLIMATE CHANGE, AND NUTRITION INTEGRATION INITIATIVE CLIMATE RESILIENCE FRAMEWORK 16

BOX 2 CHALLENGES AND RECOMMENDATIONS BY LIVELIHOOD AREA 38

BOX 3 IFPRI’S RURAL INVESTMENT AND POLICY ANALYSIS MODEL 55

ACKNOWLEDGMENTSThis report was developed with guidance and financial support from UNDP Ethiopia and benefited greatly from the insights of the Ethiopian Government. From UNDP, we particularly acknowledge the leadership and guidance of the UNDP Resident Representative, Ahunna Eziakonwa, the support and guidance of UNDP Country Director Louise Chamberlain, and technical guidance from UNDP’s James Wakiaga and Haile Kibret. We also would like to thank Martin Ras from the UNDP Bureau of Programme and Policy Support and Excellent Hachileka and Mansour Ndiaye from the UNDP Regional Service Center in Addis Ababa for their support. Moreover, continued support from the United Nations Country Team is acknowledged. We are also grateful to all development partners who shared their insights on strengthening climate resilience in Ethiopia. From the Government of Ethiopia, we are particularly grateful for the guidance from H.E. Ato Admasu Nebebe, State Minster, Ministry of Finance, and the continued guidance and support from Ato Fisseha Abera, Director, International Financial Institutions Cooperation Directorate, Ministry of Finance. The authors furthermore would like to acknowledge the contributions of three regional interviewers who contributed to chapter two of this report: Ms. Selamawit Zigita, who implemented the interviews in Afar Region; Mr. Getachew Kebede, who implemented the interviews in Tigray Region; and Mr. Abebe Lemessa, who implemented the interviews in Oromia Region. We are also grateful for the contributions of various IFPRI colleagues: Zhe Guo (Environment and Production Technology Division) for his contribution to the earlier analysis, and Tim Thomas (Environment and Production Technology Division) and Emily Schmidt (Development Strategy and Governance Division) for providing figures.

This work was implemented as part of the CGIAR Research Programs on Climate Change, Agriculture and Food Security (CCAFS) and Policy, Institutions, and Markets (PIM), which are carried out with support from CGIAR Fund Donors and through bilateral funding agreements. For details please visit https://ccafs.cgiar.org/donors and http://pim.cgiar.org/donors. The views expressed in this document cannot be taken to reflect the official opinions of these organizations.

LIST OF ACRONYMS, ABBREVIATIONS, AND TERMSAFS agriculture–food system

AgMIP Agricultural Model Intercomparison Project

bega dry season in Ethiopia, from October to January

belg rainy season in central Ethiopia, from February to May

BRACED Building Resilience and Adaptation to Climate Change and Disasters

CCAFS CGIAR Research Program on Climate Change, Agriculture and Food Security

CGE computable general equilibrium

CRGE Strategy Climate-Resilient Green Economy Strategy

CRS Catholic Relief Services

DFID Department for International Development (United Kingdom)

DPH drought-prone highland

DPL-P drought-prone lowland—pastoralist

DRM disaster risk management

DRR disaster risk reduction

DRSLP Drought Resilience and Sustainable Livelihoods Program

DSSAT Decision Support System for Agrotechnology Transfer

enset Ethiopian banana

ENSO El Niño–Southern Oscillation

EOC-DICAC Ethiopian Orthodox Church Development and Inter-church Aid Commission

EU European Union

FAO Food and Agriculture Organization of the United Nations

FEWSNET Famine Early Warning Systems Network

FMNR farmer-managed natural regeneration

GCAN Gender, Climate Change, and Nutrition Integration Initiative

GCM general circulation model

GDP gross domestic product

GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit (German development agency)

GRAD Graduation with Resilience to Achieve Sustainable Development (USAID program)

GRAD-II Feed the Future Livelihoods for Resilience Activity (USAID program)

GTP Growth and Transformation Plan

IFPRI International Food Policy Research Institute

ISFM integrated soil fertility management

kebele ward or neighborhood; smallest administrative unit of Ethiopia

KII key informant interview

kiremt rainy season in western Ethiopia, from June to mid-September

LRO Livelihoods for Resilience—Oromia

MAR Market Approaches to Resilience

meher main crop-growing season in Ethiopia: June, July, and August

MERET Managing Environmental Resources to Enable Transitions

MRH-C moisture-reliable highland—cereal

MRH-E moisture-reliable highland—enset

MRL moisture-reliable lowland

NASA US National Aeronautics and Space Administration

NGO nongovernmental organization

NMA National Meteorology Agency (Ethiopia)

NRM natural resource management

PCDP Pastoral Community Development Project

PSNP Productive Safety Net Programme

RESET Resilience Building and Creation of Economic Opportunities

REST Relief Society of Tigray

RIAPA Rural Investment and Policy Analysis

RPLRP Regional Pastoral Livelihoods Resilience Project

SAM social accounting matrix

SLMP Sustainable Land Management Program

SNNPR Southern Nations, Nationalities, and Peoples’ Region

SNV Netherlands Development Organisation

UAB Upper Awash Basin

USAID United States Agency for International Development

WASH water supply, sanitation, and hygiene

woreda district; third-level administrative division of Ethiopia

EXECUTIVE SUMMARYEthiopia has made consistent progress in improving development indicators, but vulnerability to extreme weather events is a continuing concern, especially for people reliant on agriculture for their livelihoods. The 2015/16 El Niño event caused both a severe drought and flooding, which highlighted the remark-able improvements in the country’s resilience and the remaining challenges in ensuring that everyone

“bounces back” relatively quickly from adverse climatic shocks.

Given the links between climate change, cyclical droughts, and poverty, and the high cost of emergency humanitarian assistance, the Government of Ethiopia and development partners decided to review the country’s resilience program-ming and identify opportunities and challenges to building greater resilience into the agricultural system. This work included three components: a review of the literature and government programs on resilience in Ethiopia; key informant interviews in several regions of the country; and quantitative crop modeling and economywide analyses to inform resilience programming.

CURRENT PROGRAMSResilience to climatic shocks depends on several key elements, including absorptive and adaptive capacity when a given climate shock or stress is experienced; the portfolio of available response options; the actions taken; and the outcomes of those responses. Using a climate resilience framework, this research examined the impacts and program responses related to the 2015/16 El Niño-Southern Oscillation (ENSO) event, recording important regional variations. Along with a literature review, key informants around the country, including government staff and practitioners, were inter-viewed to better understand the existing resilience programs and gather recommendations for future programming.

In Ethiopia’s highlands, where rainfed farming predominates, droughts are the most damaging climatic shocks, with vulnerability compounded by widespread poverty, rapid population growth, and limited access to land. In the lowlands, where pasto-ralism predominates, droughts can require much longer recovery periods, particularly if livestock sales or losses are significant. Here also, scarcity of land contributes to low resilience, and is compounded by low educational levels and poor services and infrastructure.

When failed and erratic rains in 2015/16 caused acute and widespread crop failure, asset depletion, food insecurity, and acute malnutrition, more than 10 million people needed food relief in addition to the 7.9 million people already under the country’s Productive Safety Net Programme (PSNP). The government responded quickly, along with a number of international organizations.

While funds were spent primarily on food inter-ventions, programs that could bolster resilience to climatic shocks were also in place. The big three government programs for resilience building are Ethiopia’s Program of Adaptation to Climate Change, the PSNP, and the Agricultural Growth Program. Recent irrigation investment programs also help.

Key resilience building interventions in the lowlands to support pastoralists include fodder cultivation and fodder banks; livelihood diversifica-tion activities; livestock market access improvement; water resource management; and savings and credit cooperatives. Key interventions in the highlands include watershed management; land rehabilitation and reforestation; and promotion of crop diversifica-tion. Additionally, a set of cross-cutting programs are implemented primarily in the highlands.

Each program, donor, and NGO generally uses its own approach to targeting households for inclusion in resilience interventions. Some rely on the PSNP’s screening approach. Many interviewees equated vulnerability to adverse climatic shocks

with food insecurity, suggesting programming should better target the food-insecure. Interviewees also suggested that households in more remote areas, which are generally more vulnerable, remain underserved.

The programs and projects use a wide variety of monitoring, evaluation, and learning systems. A lack of common resilience indicators and limitations on data sharing hinder both the harmonization of resil-ience activities and comparison across interventions. Interviewees reported mixed experiences with coor-dination across programs, as well as limited institu-tional capacity at lower levels, such as the kebele, for program implementation and oversight.

QUANTITATIVE ANALYSESCrop modeling analysis was used to show the impact of ENSO-related changes in temperature and rainfall on crop productivity across the country. Yield losses were concentrated in a few subregions, with impacts greatest in the lowlands, where cereal production fell by an estimated 10.0 percent, while cattle herds were estimated to have declined by 22.7 percent in the drought-prone lowlands. This tool was also used to assess the potential of on-farm interventions. Adoption of improved agricultural technologies and intensification of management practices are shown to reduce yield losses, suggesting that investing in on-farm productivity can be highly effective in reducing vulnerability during a severe El Niño event.

A dynamic computable general equilibrium (CGE) model incorporating the estimated crop impacts allowed for economywide analysis of El Niño’s impact. A 2015/16 ENSO-type event has consider-able spillover impact beyond agriculture, leading to a decline in national GDP by 1.6 percent, as well as losses in agricultural GDP of 3.6 percent nationally and of 11.1 percent in the drought-prone lowlands. Of three potential market and social policy inter-ventions considered—food import subsidies, grain storage, and social transfers, as well as a combina-tion of the three—none are effective at limiting GDP losses. On-farm investments are needed to reduce agricultural GDP impacts and prevent negative

knock-on impacts to the rest of the economy. On the other hand, these policy interventions are effective in reducing household welfare losses. Food import subsidies reduce total consumption losses, and cash transfers and grain distribution reduce losses for poorer households. Implementing all policy options at the same time largely eliminates the welfare losses of poor households, but raises them for higher-income households.

El Niño-type events are found to increase the national poverty rate. Without interventions to mitigate impacts, a 2015/16-scale ENSO event increases the national poverty rate from 30.0 to 31.2 percent, equivalent to an additional 656,200 people living below the poverty line, during the event period. Policy interventions can be effective in helping avoid some of the increase in the incidence of poverty caused by weather shocks.

RECOMMENDATIONSBuild on the strength of Ethiopia’s current resilience programming. Efforts to improve resilience in Ethiopia’s agriculture sector should build on the existing wealth of experience, and will require a clear understanding of how interventions relate one to another, thematically and geographically, to pursue potential synergies.

Develop new monitoring tools for resilience and improve coordination. The Government of Ethiopia and partners should develop and apply a climate resilience framework for structuring development programs and monitoring outcomes. This should be supplemented with a current database of programs and organizations involved in resilience-building interventions to help identify common avenues for future collaboration and potential gaps in investments.

Improve targeting and linkages between short- and long-term programming. Programs and policies need to be customized and targeted toward commu-nities and regions based on their location, needs, and circumstances. Food security initiatives should be linked to poverty reduction, shorter-term disaster

response, and longer-term disaster risk reduction. As lowland pastoralist areas are particularly vulnerable, a strategy specifically tailored to their livelihoods should be developed.

Strengthen the enabling environment. Ethiopia has raised the productivity of cereal producers, provided social protection to poor rural households, and invested in roads and other rural infrastructure, but smallholder productivity remains low. Increasing modern input use and water management practices, developing microfinance institutions, and supporting female-headed households (and all women farmers) should be priorities.

Develop multipronged investment strategies. A portfolio of farm, market, and social policies are needed to cushion the economywide impacts of climate shocks. As the rural population grows in the medium to long term, enhancing rural resilience will require a permanent cash transfer program similar to the PSNP, accelerated infrastructure development, increased investment in agricultural R&D, and faster creation of economic livelihood options outside of agriculture.

INTRODUCTION

Ethiopia has madE consistEnt progrEss in improving its development indicators, with important declines in poverty rates and the share

of underweight children, increases in gross domestic product (GDP) growth, and

improvements in the country’s Human Development Index (World Bank 2017; UNDP

2015). Remarkably, this progress has occurred against a backdrop of an average

annual population growth rate of 2.62 percent over the last 40 years (UNDESA 2015).

Looking to the future, the government aims for Ethiopia to become a middle-income

economy by 2025 while developing a green economy (Ethiopia, Environmental

Protection Authority 2012).

Despite this progress and the government’s aims, vulnerability to extreme weather events is a long-standing and current concern, especially for people reliant on agriculture for their livelihood. The 2015/16 El Niño–Southern Oscillation (ENSO) event that caused both a severe drought and flooding in Ethiopia brought to the forefront the remarkable improvements in the country’s resilience as well as the remaining challenges in ensuring that everyone

“bounces back” relatively quickly from adverse climatic shocks. While the ENSO event in 2015/16 was stronger than that measured in 1984, its adverse impacts were much weaker, due to strong govern-ment and donor commitment and better prepared-ness. However, expenditures to address the 2015/16 crisis were significant, with humanitarian relief expenditure amounting to more than US$1.3 billion in 2016 alone, more than double the average annual humanitarian aid contribution to the country.

Achieving climate resilience, that is, the ability of an individual, a community, or the economy to absorb a climate shock and easily recover after it, is a long and dynamic process. Given the cyclical

nature of climatic shocks and their impacts on households and the economy, and cognizant of the nexus between climate change, cyclical droughts, and poverty in Ethiopia, there is a recognized need to increase the region’s resilience capacity and to reduce dependency on emergency aid to cope with climatic shocks. This need prompted the govern-ment of Ethiopia and development partners to review ongoing resilience programming, to consult stakeholders on opportunities for strengthening such programming, and to quantitatively model a series of alternative resilience strategies to help ensure that Ethiopia reaches middle-income status by 2025.

The objective of this research study is to assess the resilience of the Ethiopian economy and popu-lation to severe climate shocks, and to identify opportunities and challenges to build greater resil-ience into the country’s agricultural system. Doing so will identify elements that could contribute to addressing gaps in a major component of resilience programming. The study distinguishes between

12

highland and lowland areas, which face distinct chal-lenges regarding climatic and other natural disasters.

The first chapter provides an overview of the climatic and socioeconomic development challenges faced by Ethiopia, and discusses the concept of resilience building, including food security and other dimensions of fragility in the Ethiopian context. The second chapter draws on key informant interviews to highlight the main contributions, achievements,

and limitations of the various resilience programs in Ethiopia. The third chapter is an integrated assess-ment of ENSO impacts on Ethiopia’s economy and its population. It employs a framework that combines an analysis of ENSO impacts on crop and livestock agri-culture with an investigation of the spillover impacts from agriculture to the rest of the economy. The final chapter presents conclusions and recommendations.

INTRODUCTION 13

1. THE RESILIENCE LANDSCAPEPrapti Bhandary, Claudia Ringler, and Alessandro de Pinto

this study, whilE focusing on climatE rEsiliEncE, adopts thE broader definition of resilience from the United Nations International Strategy for

Disaster Reduction which defines resilience as: “the ability of a system, community or

society exposed to hazards to resist, absorb, accommodate, adapt to, transform and

recover from the effects of a hazard in a timely and efficient manner, including through

the preservation and restoration of its essential basic structures and functions through

risk management” (UNISDR 2017). Resilience thinking helps link and integrate sectors

such as infrastructure, social protection, health and reproductive health, and nutrition

that have traditionally been somewhat disconnected. To ensure that appropriate

connections are made, this report proposes that the government of Ethiopia

and partners develop or use a resilience framework, according to which its many

development programs can be structured and monitored for progress and outcomes.

A resilience framework considers a broad set of factors that influence resilience in a given context, including environmental conditions, the institutional environment, and the policy context. In the proposed resilience framework, also known as the Gender, Climate Change, and Nutrition Integration Initiative (GCAN) framework, resilience depends on several key elements, including the initial state of absorptive and adaptive capacity when a given climate shock or stress is experienced; the portfolio of available options; the actions taken in response to the climate signal; and the outcomes of those responses, which influence the context in which future climate shocks and stressors are experienced. Box 1 presents the elements of the climate resilience framework.

CLIMATE, RAINFALL SEASONALITY, AND AGRICULTURE IN ETHIOPIAThe GCAN framework emphasizes local climate conditions (signals) and the enabling environment, which includes national and local farming and food systems. Ethiopia has a complex topography and climate, with diverse rainfall patterns across the country. The availability or lack of water supplies influences agricultural production in both the crop-production zones in the highlands and the largely pastoralist areas in the lowlands (Singh et al. 2016).

The country has three distinct seasons in relation to rainfall. Belg, from February to May, is a short rainy season that supplies rainwater for crops and livestock in central Ethiopia (Degefu 1987; Gissila et al. 2004). The kiremt season, from June to

14

mid-September, delivers water for agriculture in the western part of the country (Walker 2016) and is more reliable (Singh et al. 2016) (Figure 1). The bega season, from October to January, is typically dry. These regional differences result in different growing and harvesting cycles in the country.

RainfallEthiopia has experienced a large number of droughts over the last several decades, with five major droughts since 1980 (Araya 2011; Bachewe et al. 2015). Although Conway and Schipper (2011), who studied historical rainfall patterns in Ethiopia from 1982 to 2007, recorded no marked emergent changes in rainfall over time, Funk and colleagues (2012) showed that over the previous two decades the land area receiving a level of rainfall sufficient to support crop and livestock production had decreased by 16 percent. The regions particularly affected were Oromia and the Southern Nations, Nationalities, and Peoples’ Region (SNNPR). Precipitation declines of 50–150 mm during the belg season occurred in

the south-central and eastern parts of the country, while rainfall levels in the western and southern parts of Ethiopia declined during the kiremt season. Similarly, Ethiopia’s National Meteorology Agency (NMA) found that while rainfall remained constant when averaged across the country, declines in some areas of the country have occurred since the 1990s (Ethiopia, NMA 2007).

Both seasonal and annual rainfall have exhibited high variability in many parts of the country. In addition, the frequency, magnitude, and intensity of droughts in Ethiopia have increased (Ethiopia, NMA 2007). For example, during the past 10 to 15 years, the frequency of spring droughts has increased throughout Ethiopia (Viste, Korecha, and Sorteberg 2013).

TemperatureFeyssa and Gemeda (2015) found that recent changes in temperatures are consistent with climate change projections, suggesting that the effects of climate change are already being felt in Ethiopia. The mean annual temperature in the country rose by

Figure 1 BeLg- AND KireMT-DePeNDeNT AgriCuLTurAL AreAS iN eTHiOPiA

Source: Adapted from Ethiopia Humanitarian Country Team (2015), cited in Singh et al. (2016).

SOMALI

OROMIA

DIRE DAWAHARARI

Addis Ababa

Kiremt rains area

TIGRAY

AMHARAAFAR

GAMBELA

BENISHANGUL-GUMUZ

B Areas that receive kiremt rains

SOMALI

OROMIA

DIRE DAWAHARARI

Addis Ababa

Belg rains area

TIGRAY

AMHARAAFAR

GAMBELA

BENISHANGUL-GUMUZ

A Areas that receive belg rains

1. THE RESILIENCE LANDSCAPE 15

BOx 1 eLeMeNTS OF THe geNDer, CLiMATe CHANge, AND NuTriTiON iNTegrATiON iNiTiATiVe CLiMATe reSiLieNCe FrAMeWOrK*

The Climate SignalThe climate signal represents the source of uncertainty, volatil-ity, shocks, and longer-term changes. Long-term climate changes involve shifts in average temperature and rainfall conditions, as well as in the frequency of extreme weather events, such as droughts, floods, and storms. Shorter-term climatic changes and adverse weather events also influence resilience.

The Enabling EnvironmentThe effects of climate change occur within a particular context or enabling environment, which influences the ability of individu-als and groups—across a broad scale—to absorb and respond to the impact of the changes they experience. Policies, laws, and other institutions all influence individual, household, and group responses to climate shocks and stressors. At higher levels, such factors as international commitments, international aid flows, and the degree of political stability influence the resilience of nations and regions to climate shocks and stresses.

Absorptive CapacityAbsorptive capacity is defined as the sensitivity of individuals, groups, communities, countries, or regions to shocks and stressors—that is, factors that determine the extent to which different actors are directly affected by climate shocks and stressors, and the extent of the changes they need to make to preserve or improve their well-being. For example, a smallholder farmer with a diversified livelihood that includes farm and nonfarm income sources may not experience as great a loss of income upon delayed-onset rains as a neighboring farmer whose livelihood is dependent on a single rain-fed crop. The health and nutritional status of individuals at the time of a climatic shock also affect their absorptive capacity—for example, whether or not they can withstand an increased risk of infectious disease. Other factors, such as infrastructure and the strength of the social safety net, also influence absorptive capacity at the house-hold level. Absorptive capacity at the country level would be influ-enced by such factors as the structure of the economy, the natural resource base, the level of poverty or inequality, and relations with other countries in the region.

Adaptive CapacityAdaptive capacity is defined as the ability of different actors or groups of actors to respond to climate shocks, stressors, risks, or new opportunities. This ability depends on a variety of factors that interact in different ways based on social demographics, such as gender and age. At the individual or household level, these factors include the capacity of individuals to perceive and understand cli-mate risks, access to financial capital and assets, human and social capital, access to information and technology, and time constraints. At the state or policy level, factors influencing adaptive capacity include policy makers’ perceptions and risk preferences; levels

of GDP; information systems; and the availability of technology, health systems, and access to markets.

Absorptive and adaptive capacity interact with the enabling environment to determine the range of response options avail-able to decision makers from the individual to the state level. Important gender differences often limit the range of response options available to women. For example, women tend to have less access to information about climate, less knowledge about appropriate responses to climate challenges, and less access to agricultural technologies and resources. They are also less likely to be in positions of decision-making authority in community groups, institutions, and policy-making bodies. These and other difficulties limit the potential contribution of women to increasing resilience at the household, community, and national scales, and pose the risk that adaptation will occur in ways that do not reflect women’s needs and priorities.

Response Options and the Decision-Making ContextDifferent actors—including individuals, households, groups, com-munities, and policy makers—respond differently to the climatic challenges they have experienced or anticipate. Responses can take several forms, from actions directed toward coping with the immediate impacts of a climate shock or stress, to adaptive or transformative approaches that protect or improve livelihoods and well-being outcomes over the longer run.

• Coping responses generally are strategies that utilize avail-able resources, skills, and opportunities to address, manage, and overcome adverse climate stresses and shocks in the short to medium term.

• Risk management strategies involve plans, actions, or poli-cies that aim to reduce the likelihood or impact (or both) of future negative events.

• Adaptation involves adjustments to actual or expected cli-mate stimuli to avoid harm or exploit potential benefits to return to, maintain, or achieve a desired state.

• Transformative responses aim to change the fundamen-tal attributes of a system or context to improve well-being outcomes, such as actions that address underlying social vulnerabilities.

The actions decision makers take in response to climate chal-lenges often depend on complex negotiating processes in which different actors advocate for actions that meet their own needs, preferences, and priorities. Sometimes the interests of different actors overlap, but often they diverge.

Pathways from Climate Change Responses to Well-Being OutcomesIn addition to changes in the enabling environment and in adap-tive and absorptive capacity, at least six possible pathways can

16 BUILDING RESILIENCE TO CLIMATE SHOCKS IN ETHIOPIA

be taken in response to climate shocks and stressors that can influence important well-being outcomes: (1) food production, (2) income, (3) asset dynamics, (4) labor, (5) natural resources, and (6) cooperation.

Well-Being OutcomesWell-being outcomes include (1) food and nutritional security, (2) environmental security, (3) gender equality, (4) health, and (5) quality reproductive health as five final outcomes affected by the interaction of climate shocks and stresses with various responses to these challenges at different scales. Five interre-lated “environments” mediate these out-comes: (1) the food environment, (2) the social/work environment, (3) the health environment, (4) quality reproductive health services, and (5) the living environ-ment. The food environment encompasses the availability of food, the quality of diets, access to food (including both market access and affordability), and the stability of the food supply over time. The social/work environment involves shifts in the live-lihood roles and responsibilities of men, women, and children. The health environ-ment includes health stresses, control and prevention of epidemics and infectious dis-ease outbreaks, and healthcare practices and infrastructure. Quality reproductive health services focus on ensuring effec-tive emergency obstetric care, address-ing unmet needs of family planning, and implementing the Minimum Initial Service Package for reproductive health in cri-ses and beyond. The living environment includes changes in the availability and quality of natural resources as well as the physical infrastructure, such as health cen-ters, schools, shelters from disasters, and sanitation systems.

Importantly, considerable linkages, trade-offs, and synergies occur among environments, development outcomes, time frames, and groups of people. For example, practices that improve food avail-ability and access in the food environment through the increased use of chemical

fertilizers or pesticides may have negative implications for envi-ronmental outcomes, such as water quality. In terms of trade-offs over time, some responses to climate shocks and stressors may yield short-term benefits that have negative implications for future resilience. For example, selling assets to meet con-sumptive demands following a climate shock may improve nutritional status in the short term but have negative implica-tions for long-term availability of and access to food. Moreover, differences occur in terms of how the costs and benefits of the chosen response options are distributed among different groups of people.

*Under GCAN, the International Food Policy Research Institute, with input from USAID and its implementing partners, developed a conceptual framework (Bryan et al. 2017a) that integrates climate resilience, gender, and nutrition. The purpose of the framework is to identify and describe key elements of climate resilience while highlighting its linkages with gender and nutrition.

Source: Bryan et al. (2017b).

CLIMATE RESILIENCE FRAMEWORK

RANGE OF RESPONSE OPTIONS

DISTURBANCES

TRANSFORMATIVE CAPACITY

DECISION-MAKING CONTEXT

RESPONSE

PATHWAYS

Coping | Risk Management | Adaptation | Transformation

INCOMEFOODPRODUCTION

TRADE-OFFS AND SYNERGIES BETWEENPEOPLE, SCALES, AND OUTCOMES

ASSETS

ABSORPTIVE CAPACITY ADAPTIVE CAPACITY

Food environmentSocial/work environment

Health environmentLiving environment

Food and nutritional securityGender equality

Health Environmental security

OUTCOMES

RES

ILIE

NC

E/V

ULN

ER

AB

ILIT

Y F

EED

BA

CK

LO

OP

EM

ISS

ION

S/S

EQ

UES

TRA

TION

COOPERATIONNRMLABOR HUMAN CAPITAL

Note: NRM = natural resource management.


about 1.3°C between 1960 and 2006, and minimum temperatures increased more than maximum temperatures over the past decade. The rise in temperature is more pronounced in May and June. Conway and Schipper (2011) showed year-round warming in all regions of Ethiopia, with an annual warming of 1.2oC (range of 0.7oC–2.3oC) by the 2020s and of 2.2oC by the 2050s (range of 1.4oC–2.9oC), resulting in increased frequency of heat waves and higher rates of evaporation. The expectation is that rising temperatures will lead to more frequent extreme weather events. In fact, parts of Ethiopia have already experienced substantial warming of temperatures (Figure 2).

EL NIÑO–SOUTHERN OSCILLATION EVENTS SINCE THE 1980SSeveral studies have illustrated the relationship between annual rainfall and El Niño–Southern Oscillation (ENSO) events in Ethiopia (Haile 1988; Attia and Abulhoda 1992; Nicholls 1994; Eltahir 1996). The anomalies of sea surface temperature in the South Atlantic and Indian Oceans, exacerbated by anthropogenic activities, impacted the rainfall supply in the country in the 1980s (Haile 1988; Wolde-Georgis 1997). Haile (1988) discussed the occurrence of drought in northern Ethiopia associated with deviations from rainfall every 3 to 5 years, whereas it was every 8 to 10 years for the whole country. Wolde-Georgis (1997) described ENSO years as having heavy belg rains and declines in kiremt rains. Furthermore, Attia and Abulhoda (1992) and Eltahir (1996) showed that ENSO reduced the flow of the Nile River, whose tributaries supply 85 percent of water in the highlands of Ethiopia, resulting in drought condi-tions in both 1986/87 and 1987/88. Nicholls (1994), Webb and von Braun (1994), and Ayalew (cited in Wolde-Georgis 1997) described famines associated with these drought periods. A 1984 drought hit an extensive area ranging from northern and central to eastern Ethiopia, thus impacting a large number of people (Singh et al. 2016). The 2015/16 drought covered similar areas (Figure 3) and showed an even larger precipitation decline than the 1984 drought.

In 2015, the NMA of Ethiopia announced the failure of belg rains, followed by a severe delay and erratic pattern of kiremt rains (Singh et al. 2016), which resulted in the worst drought in 50 years (Davison 2015). Figure 4 maps this precipitation decline—only 50–75 percent of historical rainfall was received in the northern, central, and eastern parts of Ethiopia. Failed belg and delayed/erratic kiremt rains caused acute and widespread crop failure, asset depletion, and food insecurity, as well as acute malnutrition. More than 10 million people needed food relief in addition to the 7.9 million people already under the country’s Productive Safety Net Programme (PSNP) (USAID 2016). It was also reported that approximately 6 million children were at risk from hunger, disease, and lack of water because of the El Niño–related drought (UNICEF 2016). Approximately 450,000 children were expected to be treated for

Figure 2 CHANge iN MeAN DAiLY MAxiMuM TeMPerATure (OC) BeTWeeN 1980 AND 2010 iN HOTTeST MONTH OF YeAr

Source: Map generated from data provided by NASA (2019).

Note: Cells that show no significant trend at 10% significance level are masked out.

< -5

-5 – -4

-4 – -3

-3 – -2

-2 – -1

-1 – -0.25

-0.25 – 0.25

0.25 – 1

1 – 2

2 – 3

3 – 4

4 – 5

> 5


severe acute malnutrition, and a further 2.2 million children and pregnant and lactating women for moderate acute malnutrition. The failure of the 2015 belg rains particularly affected smallholder farmers and pastoralists in the northeastern Afar and northern Somali Regions (OPM and HESPI 2018). In addition to the impacts of water shortages for agri-culture and livestock activities, the El Niño drought affected domestic water supplies used for drinking, cooking, washing, and personal hygiene.

In response to the failure of rains, the government began to scale up food assistance and responded

relatively quickly to the drought beginning as early as July 2015 in some areas (USAID 2016; OPM and HESPI 2018). Response funds were sourced from several development and resilience budgets. According to the country’s Humanitarian Requirements Document (UNOCHA Ethiopia 2016), the National Disaster Risk Management Commission and the World Food Programme would mobilize funds to cover 7.6 million people, while the Catholic Relief Services–led Joint Emergency Operation would cover the needs of 2.6 million people. Several other groups increased or redirected investment.

Figure 3 DePArTure FrOM AVerAge PreCiPiTATiON iN DrOugHT AreAS, 1981–2016

Source: Singh et al. (2016), using Climate Hazards Group InfraRed Precipitation with Station data (http://chg.geog.ucsb.edu/data/chirps/).

Note: Affected drought area: Continuous northern, eastern, and central Ethiopia with precipitation deficit of at least 15 percent during historical period, 1981-2016, February 1 to September 1.

2008 20122004200019961988 19921984

450

500

550

600

650

700

750

800

850

Tota

l pre

cip

itat

ion

(mm

)

Longterm average: 656.9292mm

20161981


http://chg.geog.ucsb.edu/data/chirps/

In July 2015, US$143,761 of funds from the Office of Foreign Disaster Assistance of the United States Agency for International Development (USAID) were provided to GRAD (the former Graduation with Resilience to Achieve Sustainable Development program, now titled Feed the Future Livelihoods for Resilience Activity but still known by the GRAD acronym as GRAD-II) for seed and livestock activi-ties in Guraghe Zone of SNNPR. CARE provided US$249,715 for livestock support to implementing partner Relief Society of Tigray (REST) in the Southern Zone of Tigray Region in December 2015. Furthermore, a livestock feed activity was initiated by Vétérinaires sans Frontières Germany in mid-August 2015 in three woredas (districts) in Afar Region using a crisis modifier (a funding mechanism designed to provide timely responses to suspected or apparent crises by development partners already

operating on the ground) in a project funded by European Civil Protection and Humanitarian Aid Operations. Save the Children similarly introduced a crisis modifier into Peace for Development, a Somali Region project funded by the UK Department for International Development (DFID). In July 2015, DFID reallocated approximately $185 million toward emergency response, focusing on funding to the United Nations Humanitarian Response Fund, WASH (water supply, sanitation, and hygiene) projects under UNICEF, and the PSNP. The European Union (EU) provided approximately $68 million to the PSNP, with around $44 million released in December 2015 for use in 2016. The EU also provided $11.2 million toward therapeutic feeding programs through its nutrition program with UNICEF. By February 2016, USAID had reprogrammed $10 million of Feed the Future and water resources funds in response to the drought (USAID 2016). By March 2016, the govern-ment of Ethiopia had committed $700 million to the emergency response.

However, the full-scale humanitarian response started only several months after the failed belg and poor kiremt rainfall seasons. As a result, short-term coping mechanisms, such as large-scale sale of livestock by pastoralist communities and reduc-tions in food intake, led to longer-term challenges, including increased malnutrition in children (OPM and HESPI 2018). The meher (main crop-growing season) assessment concluded that the expected harvest was far below expectations, with some regions experiencing crop losses of between 50 and 90 percent (UNOCHA Ethiopia 2016). Widespread decline in meher crop yields were reported from field assessments in Amhara, Tigray, and Oromia Regions. Pastoralists in Afar and Somali Regions were some of the first and hardest hit (Seaward 2016). The comprehensive response included, for the first time, child protection as a specific operational response area in the Humanitarian Requirements Document. Moreover, at woreda and kebele (district and ward) levels, specific institutional arrangements on the drought response were implemented, such as “daily follow-up,” ensuring that means were put

Figure 4 PreCiPiTATiON FrOM FeBruArY 1 TO SePTeMBer 15, 2015 (BeLg AND KireMT SeASONS) AS A SHAre OF LONg-TerM AVerAge PreCiPiTATiON

Source: Singh et al. (2016), using Climate Hazards Group InfraRed Precipitation with Station 4.8-km (1/20-degree) precipitation dataset (http://chg.geog.ucsb.edu/data/chirps/).

Note: Long-term average is 1981-2014.

500 65 75 85 95 105 115 125 135 145 200

Precipitation as Percent of Average (%)




in place for distribution of food and other materials for managing the response (OPM and HESPI 2018). Interventions such as the Targeted Supplementary Feeding Programme aimed at providing nutritious food specifically to children under five and pregnant and nursing women.

THE ENABLING ENVIRONMENT Climatic shocks do not occur in a vacuum. As described in Box 1, national policies, laws, and insti-tutions—but also international flows of aid, foreign direct investment, trade relations, and overall stability—affect the response options to climatic shocks and stressors. One important element of the national institutional environment is land tenure. Land is important in all economies, including in largely agrarian communities where land is the largest productive asset, and particularly in those countries that undergo rapid economic and population growth. Tenure security increases incentives to invest in land, including in soil and water conservation technologies and in trees, which have longer pay-back periods than annual crops. In 1975, land in Ethiopia became public property supported by long-term use rights. In the late 1990s, the government of Ethiopia imple-mented its first land certification program, which was followed by a second-level land certification program in the early years of the following decade. Deininger, Ali, and Alemu (2011) found that the land rights reform increased tenure security, land-related invest-ment, and rental market participation. Plots were registered in the names of both husband and wife, but there continued to be a gap in women’s partici-pation in land rights–related activities, suggesting that women’s rights lagged behind those of men. Quisumbing and Kumar (2014) studied gender gaps in land rights in detail and found that gaps in knowledge about the land rights vested in the land certificates reduced the adoption of soil conservation practices as well as the planting of tree crops and legumes by women farmers. The authors suggested that closing the gendered knowledge gap about legal rights is an important step toward improving

adoption of soil conservation technologies and other climate-smart agricultural practices.

Another element of the enabling environment worth mentioning is climatic forecasts that can help improve planning for droughts and long-term adverse climatic events, such as ENSO. Such forecasting capability has improved in recent years, allowing governments to prepare better by requesting or reserving resources for food aid, by increasing payouts and other resources as part of food safety net programs and operations, and by readying disaster relief operations. At the same time, many donors (and their taxpaying constituencies) continue to respond retroactively to climatic events, once local impacts have become known, reducing the full potential benefit of improved forecasting of adverse climatic events in low-income countries. One such forecasting group, FEWSNET (Famine Early Warning Systems Network), with more than 20 country offices, including an office in Ethiopia, issues monthly reports and 6- to 12-month forecasts of risks to food security, developed based on climate, agricultural production, food prices, trade, and nutrition information, to help decision makers and relief agencies plan and prepare for potential disasters. However, research on the adoption of climate services has shown that women are often excluded from such information. Providing informa-tion on climate change and climate-smart agricul-ture practices to the husband does not mean that this information will necessarily be passed on to the wife (Tall et al. 2014; Twyman et al. 2014). Thus, building resilience to climatic shock requires a gendered approach.

IMPACTS ON AGRICULTURE AND THE ECONOMY: ABSORPTIVE AND ADAPTIVE CAPACITIESAgriculture is essential to the Ethiopian economy because it is a major contributor to economic activity and because the majority of the popula-tion is employed in this sector (Taffesse, Dorosh, and Gemessa 2012). According to data from the World Bank’s World Development Indicators (World Bank 2017, based on the International Labour


Organization’s ILOSTAT database as of March 2017), relative employment in agriculture increased from around 50 percent of the population in the 1990s to 60 percent in the following decade, and to 70 percent by 2015. Increasing the resilience of vulner-able people to climate stressors in agriculture is thus essential to support Ethiopia’s development goals. Evidence indicates that both long-term climate change and shorter-term climate shocks such as ENSO events have negative impacts on crop and livestock production.

Crop productionSignificant uncertainties persist regarding the effects of historical climate variability on agricultural productivity in Ethiopia. Although rainfall is a main determinant of crop yields, analysis of subnational rainfall and crop yield data shows a weak relationship between local rainfall and staple crop yields (Conway and Schipper 2011), suggesting that factors such as management are also important in determining crop yields. This finding is similar to that of Lobell and others (2008), who analyzed relationships between observed harvests and monthly temperature and precipitation for five regions in Africa. Relationships were significant in only 17 out of 41 cases. However, as shown later in this report, there is a clearer rela-tionship between specific ENSO events and agricul-tural production, particularly during the 2015/16 El Niño (see Chapter 3).

Future changes in climatic conditions create further uncertainty. Jones and Thornton (2003) predicted that in the Ethiopian highlands surrounding Addis Ababa, maize yields might increase substan-tially with a changing climate, but overall yields for all main crops will essentially be unchanged. Thornton and others (2010) showed that yields of maize and beans in the tropical highland mixed systems of Africa are projected to increase, sometimes substantially, under some climate change scenarios. Kassie and colleagues (2015) showed that by 2050, maize yields in the Central Rift Valley of Ethiopia might decrease by 20 percent compared with the 1980–2009 period due to climate change. Araya

and others (2015) showed that the median maize yield could increase by 1.7 percent, to 2.9 percent, during the period 2010–2039, but the range then widens from a 6.3 percent decrease to a 4.0 percent increase during 2040–2069, conditional on the model used. The Agricultural Model Intercomparison Project (AgMIP) of the US National Aeronautics and Space Administration (NASA) uses four different crop models to generate changes in yield projec-tions under five different general circulation models (GCMs) for the period 2000–2050. Countrywide results indicate general agreement across GCMs on the types of effects expected from climate change for some crops (that is, negative for maize, wheat, sorghum, and groundnuts, and positive for rice) but some disagreement between models for other crops, such as barley, millet, and cassava (mostly negative), and for soybeans (NASA 2019).

Despite considerable climate variability, agri-cultural production trends in Ethiopia have been strongly positive. Overall, production of cereals has nearly doubled since 2006, and yields have grown by about 22 percent. However, national averages of staple crop yields are still at least 50 percent below on-station yields (Global Yield Gap and Water Productivity Atlas 2019), suggesting large yield gaps and potential for farmers to lift yields through changes in on-farm management, including addi-tional use of agricultural inputs such as fertilizers and irrigation.

Nevertheless, climate variability and extremes, such as ENSO events, as well as long-term climate change, will likely put pressure on future crop yields. As a result, production capacity will be under increased stress given the projected effects of climate change, with added impacts from both biotic and abiotic stresses. The net effect is that crop yields are projected to increase but remain below a non-climate-change trajectory of yield improvements (see, for example, IFPRI 2017), with yield declines in times of climatic extremes. Moreover, climatic shocks will remain a major challenge for climate resilience and response options. Alemu, Korecha, and Mohamod (2017) analyzed the effects of


historical ENSO phases from 1980 to 2014 on cereal productivity in the Upper Awash Basin (UAB), where belg rains account for about 183–310 mm of rain per year while kiremt rains contribute 465–906 mm. The authors found average cereal yield reductions of 16.0 percent and 5.3 percent due to El Niño and La Niña events, respectively, in the southern, south-eastern, and central regions of UAB, and 10.1 percent during El Niño and 9.1 percent during La Niña in the western highlands, with even higher reductions for some individual crops.

Livestock Rain is an important source of water for herds and for growth of forage. Studies assessing linkages between livestock population and rainfall distribution (Desta and Coppock 2002; Angassa and Oba 2007; Tache and Sjaastad 2010) suggest a strong relationship—the drought years of 1984, 2000, 2002, 2009, and 2011 show comparatively high mortality rates of livestock because of feed deficiency and water scarcity. These same studies show low mortality rates of cattle, sheep, goats, and camels during the wet years of 1996, 2003, 2006, 2010, and 2014.

ENSO events (El Niño and La Niña) have become regular occurrences for pastoral communities of Ethiopia, leading to rainfall uncertainty that affects pasture/rangeland, thereby influencing livestock populations (Hassen et al. 2017; Kandji, Verchot, and Mackensen 2006; Korecha and Sorteberg 2013). Hassen and colleagues (2017) studied the effect of reduced rainfall on livestock populations in eastern Ethiopia. Results illustrate the declining cattle and sheep population during El Niño (diminishing rainfall) in Shinile Zone, whereas goats and camels, which are more drought tolerant, are better adapted to low amounts of rain. Further, mortality rates of cattle rise with declines in the mean annual rainfall (compared with normal distribution) during ENSO (Hassen et al. 2017). The long travel distance to find better grazing areas and water resources could contribute to the high mortality rates of livestock during ENSO (Hassen et al. 2017). Megersa and others (2014) described comparable outcomes in Borana pasture

areas, where scarcity of grazing lands and changes in climate conditions have triggered decreasing popula-tions and productivity of livestock. Given uncertainty in the occurrence of rainfall, pastoralists tend to settle near water sources and to use seasonal mobility and destocking as adaptation options to alleviate the detrimental effects of ENSO.

Additionally, the change in the composition of rangeland plant species from grasses to woody plants because of climate change may have impacted cattle and sheep populations more than those of goats and camels (Abebe et al. 2012). Camels and goats can consume more bushy plant species than can sheep and cattle (Abate, Ebro, and Nigatu 2010; Megersa et al. 2014). Moreover, camels are preferred by herders due to their ability to produce relatively high amounts of milk despite environmental stress, thus providing income and nutrition to herders’ families (Tilahun et al. 2016; Hassen et al. 2017).

The Agriculture Knowledge, Learning, Documentation and Policy (AKLDP) Project examined the early impacts of the 2015/16 drought in North and South Wollo and Wag Himra of Amhara National Regional State (ANRS). Farming of livestock, composed of cattle, oxen, sheep, and goats, is a major source of economic livelihood in these areas. In a climate-normal year (that is, without ENSO), appro-priate meat quality is achieved in the belg period given sufficient forage and water, commanding better prices during trading (AKLDP Project 2016). However, the lack of rain in the 2015 belg period reduced meat quality and production, resulting in low meat prices in North and South Wollo and Wag Himra. Livestock prices in November 2015 were more than 50 percent below those in December 2014 in the study areas of ANRS. Inferior meat quality due to poor pasture, over-supply of livestock in the markets, and the “forced” nature of sales (inability of smallholders to negotiate for better prices because traders are aware of their urgent need for cash) were some of the reasons for the low prices, as were herders’ needs to align livestock numbers to lower water and feed availability (AKLDP Project 2016).


Economic growth and povertyThe impacts of climate variability go beyond crop and livestock production to the broader economy. Block and others (2008) utilized a multimarket model with detailed analysis of climate variability to assess its impact on economic growth and poverty. The results showed that climate variability significantly reduces agricultural GDP growth, and because of the importance of agriculture in the economy, this also reduces nonagricultural and total GDP growth. Importantly, poverty also increases due to climate variability, adding urgency to policies, programs, and investment to improve resilience and response to weather shocks. Analysis presented later in this report provides new estimates of ENSO’s impacts on food security and the Ethiopian national economy, including losses in GDP and worsening household poverty (see Chapter 3).

A focus on enhanced resilience is particularly important in Ethiopia’s dryland areas, where according to Cervigni and Morris (2016), the agricul-ture-dependent population is expected to grow by about 60 percent between 2010 and 2030. A major shortcoming in the evidence base or literature on ENSO and other climate shocks in Ethiopia is a lack of differentiation between lowland (largely dryland areas) and highland (more productive areas) condi-tions and policy needs. This report evaluates policy options in light of their effects on highland versus lowland economies and populations.

NEED FOR ACTION: GROWING VULNERABILITY IN RURAL AREASExpenditures on humanitarian responses in Ethiopia average US$700 million per year and spiked during 2016 in response to the extreme ENSO event. Most of the funds were spent on food security interven-tions. Development programs have increased their flexibility to quickly expand the reach and intensity of

support during disasters, for example by expanding the quantity and coverage of cash transfers or food-for-work programs under the government’s PSNP. Knippenberg and Hoddinott (2017) showed that households with access to PSNP experience fewer months of food insecurity than do households without such access. Following a reported drought shock, households with PSNP access returned faster—after around two years—to their predrought food security level, whereas households not receiving any PSNP payments did not return to their (slightly higher) predrought food security levels until four years after the drought, on average.

Rural resilience can be achieved only if infrastruc-ture development accelerates and if investments in agricultural research and development are increased. Even today only one-half of all Ethiopian farmers use chemical fertilizers, and growth in irrigated area, though commendable, remains too low to dramatically affect resilience outcomes in the next two decades. Crop insurance is also key to building resilience against climatic shocks. However, demand for insurance among poor farmers is low due to the associated high risks and costs, as well as low trust and insufficient understanding of insurance (Dercon et al. 2014).

Achieving climate resilience in rural areas will require the generation of much faster economic live-lihood options outside agriculture, which in turn will necessitate a much faster opening to and support of foreign direct investment. Clearly, the share of people employed in the agriculture sector must decrease substantially for the country to achieve middle-income status.

The question remains of how emergency responses can be better linked to ongoing develop-ment programming. The next chapter discusses this question in detail, drawing on a literature synthesis and the experiences of key informants in Ethiopia.


2. RESILIENCE BUILDING IN ETHIOPIA: ANALYSIS OF KEY INFORMANT INTERVIEWSLaia Domenech, Semhar Tesfatsion, Claudia Ringler, and Sophie Theis

a kEy ElEmEnt of thE climatE rEsiliEncE gEndEr, climatE Change, and Nutrition Integration Initiative (GCAN) framework is understanding

Ethiopia’s response options and the decision-making context. To reduce dependency

on emergency aid and the vulnerability of communities and ecosystems to climatic

shocks, a series of programs and projects were initiated by the Ethiopian government,

multilateral agencies, and bilateral donors. These aimed to build resilience and

simultaneously boost economic growth, create jobs and livelihoods, and strengthen

access to health and education for all communities.

This chapter sheds light on the main contribu-tions, achievements, and limitations of existing resilience programs in Ethiopia to provide insights to strengthen future climate resilience programming in the region. The findings are drawn from interviews with key informants, at both national and regional levels, who are in charge of a specific resilience projects or programs in the country or are familiar with resilience programming in Ethiopia.

METHODOLOGYKey informant interviews (KIIs) were implemented with government officials, donors, and international and local nongovernmental organizations (NGOs) in the capital city of Addis Ababa and in three main geographic regions of Ethiopia: Tigray, Oromia, and Afar. Interviewees were selected from agencies implementing key resilience programs; in turn, they suggested additional interviewees, resulting in a snowball sampling approach.

As mentioned above, Ethiopia has two distinct geographic areas: highlands and lowlands (Table1), with highlands concentrated in the northwestern parts of the country. Climatic shocks and livelihoods differ markedly in these two areas. Highlanders are affected by land degradation, and the population is mostly agrarian—that is, characterized by their use of crop systems or mixed crop and livestock systems. Lowlanders are mostly affected by rainfall variability, droughts, and high temperatures; given the low rainfall levels in this region, the population consists mostly of pastoralists and agropastoralists.

In total, 47 stakeholders were interviewed between January and March 2018: 18 from govern-ment organizations, 6 from donor organizations, 15 from international NGOs, 5 from local NGOs, and 3 from research institutions. To obtain a general overview of resilience programs, 21 stakeholders were interviewed in Addis Ababa. In addition, 9 KIIs were implemented in Oromia Region, 10 in

25

Tigray Region, and 7 in Afar Region. Depending on interviewees’ responsibilities, they were asked questions about a specific program or, alternatively, more general questions about different resilience programs implemented by their organization.

The KIIs included questions on what needs the resilience program aims to address, which popula-tion is targeted and through what interventions, how programs are monitored, the implementation capacity, and the extent to which programs are adaptable to changing situations. Interviewees were asked to comment on key successes, failures, and entry points for strengthening resilience program-ming in Ethiopia. The information is treated confi-dentially and therefore no names or organizations

are linked with any specific interview or statement made in this report.

DO NATIONAL AND REGIONAL POLICIES SUPPORT RESILIENCE GOALS?The three big government programs for resilience building are Ethiopia’s Program of Adaptation to Climate Change, the Productive Safety Net Programme (PSNP), and the Agricultural Growth Program. All fall under the government’s Climate-Resilient Green Economy (CRGE) Strategy and its Growth and Transformation Plans (GTPs). In addition, substantial programs are targeted at pastoralists in lowland areas. Key among these are the Intergovernmental Authority on Development’s Drought Resilience and Sustainable Livelihoods Program (DRSLP), which covers pastoralist areas across the Horn of Africa, and the Pastoral Community Development Project (PCDP).

Most stakeholders interviewed said they think that national policies support resilience goals. “These policies are good and well written,” explained one interviewee. However, the greatest challenge is to ensure proper implementation, understanding, and follow-up as well as practical alignment between policies. Some stakeholders interviewed requested the development of a well-planned, stand-alone resilience policy or strategy. “This policy should be nationally endorsed and include national, regional, and woreda-level provisions,” said one interviewee.

One of the most relevant policies for resilience building in Ethiopia is the CRGE Strategy, which was launched in 2011 and identifies green economy opportunities that can help Ethiopia reach middle-income status by 2025 while keeping greenhouse gas emissions low. However, the policy has been criticized for (1) being confined to the national level, because it has not been decentralized to the regions, for (2) not emphasizing climate resilience, and (3) for lacking specific climate resilience indicators. The government is currently developing such indicators.

The first Growth and Transformation Plan (GTP I), from 2010-2015, aimed to increase productivity and

TABLe 1 SOMe DiSTiNguiSHiNg FeATureS OF eTHiOPiA’S HigHLANDS AND LOWLANDS

Highlands Lowlands

regions Tigray and parts of Oromia

Afar Region, and Guji and Borana Zones of Oromia Region

Primary livelihoods Agrarian: teff, wheat, other crops

Pastoralism: goats, cattle, camels, or agropastoralism: livestock with some crop production

Primary climatic shocks

Excess rainfall and floods

Rainfall variability, droughts, and high temperatures

Sources of vulner-ability

Deforestation and land degradation, small landholding sizes

Lack of infrastructure and basic services, limited market access, conflict, land degra-dation, insufficient watering points and grazing land in the dry season

Sources of resilience Some existing infra-structure and services, such as health, educa-tion, and financial services

Social capital

Main impacts of climatic shocks

Crop loss Water shortage, lack of grazing land, undernutrition, human and livestock disease

Source: Authors.


growth of the country, but did not initially consider climate. GTP II, lasting from 2015/16 to 2019/20 considers climate and disaster management. All sectors will have to evaluate the achievements of GTP II, which will help to determine if vulnerability to climate change has been reduced in the country. An important challenge identified is that one-half of the country is disconnected from the other half in terms of growth. “A growth strategy for the lowlands, arid [lands], and semi-arid lands is very much needed,” according to one interviewee.

The disaster risk management (DRM) policy approved in 2013 and the social protection policy from 2014 were also identified as relevant policies. The social protection policy should support the DRM policy and help prevent shocks and disasters. However, “there is lack of coordination and no synergy between both policies,” explained one interviewee. Another interviewee added that the paradigm shift from a reactive response to a proactive risk management approach like the one promoted by the DRM policy has not yet been realized in Ethiopia.

Several stakeholders pointed at specific policies they would like to see amended or created to improve resilience building:

• The land use policy should be refined and reformulated to consider the reality of resilience programs and the communal land tenure system.

• Implementation of the environmental protection policy needs to be improved because communal land where NGOs work are often encroached upon.

• The livestock management policy needs to be integrated with land rehabilitation work. Due to a lack of coordination, rehabilitated land is sometimes destroyed by poor livestock systems management a few years after being rehabilitated.

• The national policy on water use, irrigation, and water pricing is malfunctioning and should be

revised. A water resource master plan is urgently needed.

• A policy to promote payments for ecosystem services could improve resilience: If downstream communities would pay and help upstream communities to work on water harvesting and terracing, the former would be less affected by floods.

• Policy or procedure to support the sustainability of implemented programs: Programs would be more sustainable and resilience improved if a government institution led the work started after an NGO leaves at the end of the project period.

• Resilience-building policies should include the private sector.

• The policy of the Charity Societies Agency, which uses the 70:30 budget criterion for direct and indirect costs, should be amended so more funds can be allocated to training and evalua-tion activities. Training support interventions and monitoring and evaluation costs are considered indirect costs, with a 30 percent limit, which can limit the capacity of the interventions.

MAIN VULNERABILITIES TARGETED BY RESILIENCE PROGRAMS In Ethiopia, resilience is usually defined in terms of food security, and thus resilience programming often targets those who are food insecure during or because of climatic shocks. However, both climatic and livelihood-related vulnerabilities differ between highland (agrarian) and lowland (pastoralist) societies.

In the highlands, droughts are the key adverse climatic shocks, because most areas continue to be farmed using precipitation as the only source of water for crops; but floods and land degradation also cause people to experience crop failures. These vulnerabilities are compounded by general poverty and rapid population growth, leading to smaller land sizes and, in many cases, landlessness. According to

2. RESILIENCE BUILDING IN ETHIOPIA: ANALYSIS OF KEY INFORMANT INTERVIEWS 27

one key informant, landless farmers account for 30 percent of the population in Tigray. Nevertheless, if the climatic shock is short, highland farmers can start cultivating again when the rain comes, a few months or a year after the shock. However, in the lowlands, droughts can lead to much longer recovery periods, particularly if a large number of livestock die or are sold at low prices.

Pastoralists, who need to sell three to five goats per month per family to cover their basic needs, are the most vulnerable group in the areas under study, considering their high livestock dependency. Key challenges in the lowland areas (Afar Region, and Guji and Borana Zones of Oromia Region) are the lack of watering points in general and particularly during droughts, and the dramatic shrinkage or unavailability of grazing areas during droughts. Additionally, scarcity of land, due to encroachment of agricultural areas into traditional dry-season grazing lands, contributes to low resilience capacity. Further challenges include very low educational attainment and limited health and rural infrastructure services, including roads and access to markets, as well as the dearth of microfinance institutions for pastoralists, which makes it challenging to escape low-income, natural resource–dependent livelihoods in the lowlands. Agrarian communities in the highlands have more access to credit during shocks, as well as better-developed rural infrastructure, including health and education services.

In pastoral communities, the livelihood situation of some households can change drastically overnight. “Climate change can erode the adapta-tion capacity of not only the poor but also the better off,” said one interviewee; but the latter may not be able to get help because they are considered resilient. The erosion of indigenous mechanisms to cope with extreme weather events increases pasto-ralists’ vulnerability to climate change. For example, pastoralists used to borrow from each other when a climate shock hit, but during the last five years, recurrent droughts have affected everybody in the community, and many have nothing left to share during adverse times. Nevertheless, interventions

in lowland areas need to take into consideration the culture and traditions of pastoralists, where strong social capital and networks play a key role in addressing resilience at the community level.

HOW ARE TARGET AREAS AND HOUSEHOLDS IDENTIFIED?Each program, donor, and NGO generally uses its own approach to targeting households for inclusion in resilience interventions. Some programs rely on the PSNP, which already uses a screening approach for targeting vulnerable households. Others focus on households facing chronic food shortages, such as poor households with children and female-headed households. Some programs target those who can afford a certain technology, such as improved cook-stoves, or beneficiaries with specific features, such as those involved in livestock production. Some consult with the community to identify households for targeting through participatory rural assessment, and use wealth ranking, vulnerability tools, and screening committees to identify poor households.

To identify and select the areas and communities to be targeted by a program, most organizations work together with the existing government bodies at the zone, woreda (district), and kebele (ward) level. Programs focused on resilience aim to work in the poorest areas and the areas most vulnerable to extreme weather events such as droughts. Other criteria used to select woredas and kebeles are the level of food insecurity, access to natural resources, land degradation, access to basic services, and access to markets. However, one of the main chal-lenges identified by interviewees is that households in more remote areas, which are generally more vulnerable, remain underserved because most orga-nizations prefer working in easy-to-access areas.

Several tools are used to identify the main vulner-abilities and needs of communities, including index-based vulnerability assessments to identify the zones that are more vulnerable to climate shocks; consulta-tions with community leaders, representatives, and main stakeholders; situational analysis that identifies hazards, vulnerable areas, and existing community


capacities; and the community-managed disaster risk reduction approach, whereby the community assesses and maps out its vulnerabilities to develop an action plan with implementation priorities. Communities and government officials are usually actively involved in the needs assessment process, although this area could be further strengthened, as one interviewee stated, adding that donors sometimes do not consult the community enough.

KEY RESILIENCE PROGRAMS: MODALITIES, INTERVENTIONS, AND ACTIVITIES Many projects aiming to increase the resilience of the Ethiopian population take a multisectoral approach. Some interventions (usually natural resource interventions) aim to benefit the whole community, whereas others (usually livelihood inter-ventions) target specific households and individuals, frequently giving priority to women and the poorest of the poor.

According to most of the stakeholders inter-viewed, communities are actively involved in the design and implementation of interventions. However, one interviewee explained that sometimes in pastoralist areas, communities do not know about or do not agree with the activities implemented. Others stated that short-term projects—those that are active for just three to five years—are generally less beneficial and less trusted than longer-term projects, because beneficiaries know that benefits under the former will be short term.

Several program modalities were identified in Ethiopia: multidonor programs, PSNP-related and other food-for-work programs, direct support and emergency aid programs, and renewable energy programs. (See Appendix 1 for a complete table of the main resilience programs implemented in Ethiopia.)

• Multidonor programs or flagship programs. Several large programs, such as the PSNP and the Sustainable Land Management Program (SLMP), are implemented by the government and receive funding from various donors. For example, the

PSNP, launched in 2005, has so far benefited 7.8 million people through conditional cash and food transfers. Community members willing to partici-pate in labor-intensive public works (such as soil and water conservation activities and rangeland management, as well as the development of community roads, water infrastructure, schools, and healthcare centers) receive the transfers. Unconditional cash and food transfers benefit people who cannot work, such as the disabled and single mothers. In 2009, the program started integrating environment and climate change considerations through the Climate Smart Initiative, making climate resilience a key priority for its interventions.

• Programs that support the PSNP. Several programs target the PSNP households to help them graduate from the program by increasing their ability to cope with medium-level shocks over the long run without receiving support. Most projects implement activities in different sectors such as WASH, microfinance, and capacity building for livelihood improvement. Examples include the Overseas Development Institute’s BRACED (Building Resilience and Adaptation to Climate Change and Disasters) program and USAID’s Graduation with Resilience to Achieve Sustainable Development (GRAD) program.

• Other food-for-work programs. These programs (for example, those of World Vision) frequently aim to regenerate natural resources and increase the resilience of the whole community rather than of individual beneficiaries. Households receive food and other supplies in return for their work. For example, each community member partici-pating in the World Food Programme’s MERET (Managing Environmental Resources to Enable Transitions) program receives 3 kg of cereals per day upon completing tasks related to watershed management, such as digging pits or planting seedlings. The community contributes about 15 percent of the project cost and donors fund the remaining 85 percent.


• Direct support programs. Programs that provide direct support usually target household members who are unable to work, such as elders, pregnant women, lactating mothers, people with disabili-ties, and people with HIV/AIDS.

• Emergency aid programs. Emergency aid programs provide food aid, seeds, and medical help that usually benefit those unable to work. For example, the Oromia Agricultural Bureau fills food gaps whenever crop failures are anticipated. Preassessment for emergency aid is completed once or twice a year, after the harvest season. If the support needed exceeds the regional govern-ment capacity, emergency aid is requested from the federal government.

• Renewable energy programs. These programs target everybody in a given area who can afford the technology (for example, improved cook-stoves and baking ovens, and solar technologies). The programs start with building awareness of the potential benefits of these improved technologies, such as avoided deforestation, improved health, and reduced cooking time. They then create demand by displaying and demon-strating the use of these technologies in markets and bazaars, and through leaflets and mass media (radio), as well as establishing local-level producers and retailers. Beneficiaries are required to cover part of the investment and to have certain assets (such as cattle), and the federal government often covers part of the cost.

The interventions and activities promoted sometimes vary depending on the type of society (pastoralist versus agrarian), geographic area (lowland versus highland), and type of climate and hazards the population is exposed to.

Activities in pastoralist (lowland) societiesPastoralists reside in semi-arid and arid lowlands, which are frequently affected by rising temperatures and adverse climate shocks, particularly droughts.

Assessing the population dynamics and livestock production system, and understanding the culture and traditions of pastoralist societies, are very important to deliver successful interventions.

Several interviewees made reference to this challenge: “Government and NGOs push the pasto-ralist society to become agrarian to make them resilient without fully understanding their livelihood system.” A different point of view was expressed by another interviewee: “If you provide more water and food to the animals in a system which is still tradi-tional, it may not necessarily improve; you need to modernize the system itself.”

Pastoralist areas are usually remote, and access to basic services remains very limited. Improving access to water, health, education, electricity, and roads is the key focus of several resilience programs implemented in pastoralist areas, especially in Afar, with the aim of eventually settling the pastoralist communities in one place. One question is whether improved services can be provided only to pastoral-ists who settle down in a specific area or whether such services can also be made available to pastoral-ists who would like to pursue traditional nomadic ways of life.

The main activities implemented by projects working in pastoralist and agropastoralist areas include these:

• Rangeland regeneration activities, fodder cultiva-tion, and fodder banks, whereby pastoral commu-nities cultivate and store grass/hay to be used in times of drought

• Livelihood diversification activities, including capacity building for beekeeping and coop-eratives for gum and incense collection, with some programs giving start-up money to the cooperatives

• Livestock market access improvement activities, such as building new market centers that are closer and more accessible to pastoralists, as well as animal health and insurance services, such


as providing access to livestock vaccination and veterinary clinics

• Water resource management, involving construc-tion of micro-dams and catchments, rehabilitation of ponds, and construction of water supply and treatment stations

• Savings and credit cooperatives, often targeted at women organized in self-help groups

Activities in agrarian (highland) societiesResilience programs that target agrarian societies frequently aim to transform the agricultural system by introducing sustainable agricultural production and climate-smart agricultural practices. Introducing irrigation is also important to reduce the vulnerability of crops to rainfall variability.

Some specific activities implemented by programs working in agricultural areas follow. (See Appendix 2 for a full table of the activities promoted by different resilience programs and organizations.)

• Watershed management through capturing water runoff, increasing water availability (check dams, water harvesting, and moisture harvesting), and terracing

• Land rehabilitation and reforestation, including the enclosure of eroded land for protection from animals and humans, use of terracing, planting of grasses and agroforestry, soil fertility manage-ment, and promotion of climate-smart agricultural practices

• Promotion of vegetable production and crop diversity to promote balanced diets and improve the nutritional status of households

• Distribution of seeds, such as fast-maturing varieties and improved seeds (for example, improved teff and sorghum seeds)

• Crop insurance to protect farmers against the loss of their crops due to natural disasters such as drought, floods, frost, and snow

Gender considerationsWomen bear the brunt of climate vulnerability. Whenever water sources dry up, women are forced to travel long distances to collect water. Women also have a harder time migrating after a shock because they must take care of their families and lack savings or assets that could aid them in coping with shocks. Government policy states that programs must be gender sensitive and that at least 30 percent of the beneficiaries of any program should be women. It also requires that programs create job opportunities for the young community members.

Programs that promote green technologies—such as improved cookstoves, solar technologies, and biogas digesters—target men and women equally. However, because more women than men are involved in using improved cookstoves, women are prioritized for such technologies. They tend to be more successful with such technologies with dropout rates lower than for men. Women who adopt the technology save time by not having to collect firewood to cook, which has a positive impact on children’s school attendance and women’s empowerment.

Several organizations and programs, such as the Hundee Oromo Grassroots Development Initiative, GRAD-II, and DRSLP, prioritize support to women, because they believe that empowering women is the best way to build resilience in the community. One stakeholder interviewed explained that imple-menting a program through women is more produc-tive, especially in pastoralist areas, because women are responsible for every activity in the household.

Many programs report having special measures or activities that target women in particular, such as self-help groups, savings and lending groups, specific income-generating activities (for example, goat and chicken rearing), trainings on business and bookkeeping skills, and DRM. Similarly, many programs have provisions to include a minimum number of women on the different committees created. For example, some projects report that they try to ensure that water management committees,


early warning committees, and disaster risk reduction (DRR) committees comprise at least 40 percent women.

Other stakeholders reported difficulties in involving women. “The leader of the household is usually a male, which makes it difficult to reach 50 percent women’s participation,” explained one interviewee. “We try to benefit women intentionally, but the culture by itself does not give permission to women to go out in public in pastoralist areas,” explained another interviewee.

Some stakeholders interviewed explained some of the challenges they face listening to women’s needs. “Unless you talk to women alone, they will not tell you their problems,” explained one interviewee; thus some programs conduct discussions with men and women separately. Another great challenge programs face is in having women play decision-making roles.

Catholic Relief Services (CRS) explains that it uses role models to teach communities about gender equity and gender equality. It also encourages and rewards positive models. CRS brings spouses or families together for discussions on gender and gender-based labor division. Based on these discussions, families and communities develop a community- and household-level action plan to increase gender equality.

DISASTER RISK REDUCTION: CLIMATE FORECASTING AND EARLY WARNING SYSTEMSClimate shocks are increasingly unpredictable, and therefore improving climate prediction systems is becoming increasingly important to enhance resilience. Several programs incorporate a DRR component that aims to improve early warning systems and the preparedness of the population for drought and other shocks.

If communities know in advance when rainfall is forecast to begin or whether to expect a seasonal rainfall shortage, they can prepare for it. For example, agrarian communities can plant drought-tolerant and early-maturing seeds, which need less water and grow faster, or can choose when to plant. They can

also get agricultural advisory services that help them choose the right seed and fertilizer for each situation. If above-normal rainfall is anticipated, flood control activities can be prioritized. In pastoralist societies, it is critical to receive information about an expected shock beforehand. For example, if a drought is forecast for the coming months, pastoralists can sell their animals at the market and save the cash before the shock hits them.

Farmers and pastoralists draw on local knowledge that functions as an early warning system. They forecast weather based on wind direction, star positions, animal behavior, and soil moisture. Traditional knowledge can be amalgam-ated with scientific knowledge to reduce uncertainty and unpredictability through the co-production of climate information systems.

Several programs also work on climate informa-tion dissemination to increase climate understanding and prediction capacity at the community level. A consortium of several aid and development organi-zations works to make climate information available on the ground in Oromia and to disseminate down-scaled climate information for local communities.

“This program is very relevant and successful to help communities deal with climate variability and shocks,” explained one interviewee.

The government is planning to install weather stations at the woreda level beyond the zone level where they are usually found. Furthermore, early warning systems and information dissemination mechanisms from the woreda to the kebele level are very weak. Farm Africa is promoting an interesting approach to improve early warning procedures, installing 25 automatic weather stations in its program areas. A weekly resilience radio program is also broadcast at the national and regional levels and targeted to different groups, along with advice on what to do when a shock is predicted.

However, one interviewee stated the belief that “the existing climate modeling follows a top- down approach that is difficult to explain to local communities.” Effective and reliable early warning and information systems should be timely, relevant


to communities’ livelihoods, and communicated in local languages and through channels accessible to all members (including, for example, the illiterate or those with no access to radio or mobile phone).

PROGRAM IMPLEMENTATION AND INSTITUTIONAL CAPACITYGovernment bodies such as the Ministry of Agriculture, the National Risk and Disaster Management Commission, and others are imple-menting many of the programs on resilience building in Ethiopia (such as the PSNP, MERET, PCDP, SLMP2, and the Regional Pastoral Livelihoods Resilience Project, or RPLRP). Some international NGOs (for example, Save the Children) use government struc-tures to implement programs and largely work as facil-itators to support and enhance the quality of services provided by the government. In contrast, other inter-national NGOs (for example, CARE and CRS) work in partnership with local NGOs to implement their programs on the ground. A few international NGOs (for example, World Vision) act as direct implementers of programs, and their staff work together with community actors at the grassroots level.

Some stakeholders highlighted that the capacity of government structures at the lower (kebele) level is limited. Many resilience programs expect the government to help them implement their programs, but due to the government’s limited capacity, programs sometimes get delayed. “This challenge could be minimized if project owners looked for synergies and planned the implementation of the programs together,” explained one interviewee. Linkages between government structures from the kebele to the woreda level and from the woreda to the regional level are also weak.

To enhance the implementing capacity of govern-ment bodies, some programs train government workers. GIZ (the German government’s develop-ment agency) prefers to give trainings at the kebele development agent level because the development agent can meet more easily with beneficiaries. The community and the main local stakeholders are

generally actively involved in the design and imple-mentation of resilience programs.

Other challenges to implementation pointed out by some interviewees are lack of human resources, logistical issues, droughts and conflict, limited budget, low planning capacity, and lack of commitment from government bodies. For example, some organiza-tions face challenges in finding staff willing to work in remote and marginal areas at the district level.

Challenges related to program implementa-tion are particularly significant in pastoralist areas, especially in Afar. “The work done in pastoral areas is very slow,” stated one interviewee. For example, an interviewee explained that implementation of RPLRP in Afar faced a two-year delay at the start due to financial constraints and lack of commitment from policy makers. Meanwhile, high inflation rates significantly increased the cost of planned interven-tions. As a result, only half of the micro-dams initially promised could be built, which created a conflict between the communities and the project. Creating a government body committed to monitoring and evaluating the proper management of a project can improve implementation.

Another stakeholder, from a regional govern-ment agency, explained that due to lack of a suffi-cient annual budget, it takes two to three years to complete a project. The informant explained that no contractors with the required capacity are available in the area. Many of the needed materials have to be ordered from Addis Ababa, and delivery can take several months, so many projects get delayed.

Complaint-handling mechanisms are important because they encourage feedback from communi-ties and can lead to improvements in implementa-tion. The main social accountability mechanisms set up by programs include providing the community with the contact details of the implementers or regional village office head so that community members can file a complaint or provide feedback, creating special committees where communities can address their complaints, using community scorecards for beneficiaries to rate their satisfac-tion with the goods and services provided, and


having civil society peer-review committees assess program implementation.

COORDINATION WITH OTHER PROGRAMS AND INSTITUTIONS, AND CHANGES OVER TIMEStakeholders reported both good and bad coordina-tion experiences between the different institutions implementing resilience programs. The overall sense is that coordination has improved but could be better.

Some stakeholders said they believe coordination is good, and explained that they share their experi-ence and learning with other institutions and mutually support each other. Several efforts are in place to integrate programs through the creation of consor-tiums formed by different organizations working in the same area, or in some woredas by holding periodic woreda-level coordination meetings between the implementers of different programs. Some organiza-tions talk to each other and carefully select project sites during needs assessments to avoid overlap between programs.

A good example of a joint venture between two projects and institutions can be found in the MERET project. The MERET project, implemented by the government, built water-harvesting infrastructure (bench traces and check dams) in Tigray. Afterward, World Vision and the Relief Society of Tigray developed irrigation infrastructure in the area, thus improving the resilience capacity of the community.

In contrast, other stakeholders said they believe that resilience programs are not implemented in an integrated way and that coordination needs to further improve, an issue that seems more significant in pastoralist areas and in Tigray. An example is the reported discrepancies between the Agency of Mines and Energy and the Agricultural Bureau in Tigray, where each entity promotes a different type of fertilizer.

Some stakeholders complained about projects interfering with their programs by giving free handouts, which goes against the “work-for-food” mentality they are trying to establish. This makes community members reluctant to participate in program that require taking a loan and returning the money afterward.

Negative interactions between the PSNP and a program on natural resource management were reported by one stakeholder due to the use of different approaches to pay beneficiaries for their work. Whereas PSNP beneficiaries are paid a daily amount for public works, the other program pays beneficiaries per public work completed. As a result, beneficiaries of the latter complained about not getting paid enough.

Interviewees identified the need to work with a larger range of partners. To avoid redundancy between programs, some stakeholders suggested working closely with the government, while others expressed their desire to cooperate with other NGOs and the private sector. Several institutions reported working in coordination with research institutes, while others expressed their desire to do so.

Multisectoral coordination is required for successful implementation of the CRGE Strategy, and several high-level committees were formed for that purpose. However, “these committees are not functioning at the level required, and as a result, different sectors are operating independently,” stated one stakeholder interviewed. The coordination and linkages between government structures and at the

different administrative levels are also weak.

Over the last 10 years, noticeable changes in programming have occurred. Important improvements have been observed in terms of planning for and fulfilling communities’ needs, as opposed

“Many resources are wasted when projects work in parallel and don’t communicate with each other. We have to learn to share experiences and show the way to others, so that they don’t pass on the problems we [experienced].”

—Key informant from the pastoralist areas


to previous practices of imposing practitioners’ views. The first interventions focused on providing emergency relief through direct support, which created dependency; these were followed by inter-ventions that took a food-for-work approach. The current modality adopted focuses on livelihood development. And some programs have shifted their focus from implementation of “reactive” measures after a disaster hits to a more proactive risk manage-ment approach. Several programs are now focusing on targeting lowland areas, while a few years ago most programs operated in the highlands.

Some programs have evolved based on learning acquired from their previous phases. Now in its fourth phase, the PSNP has added new components, such as a credit program and natural resource protection. PSNP IV incorporates a component that promotes climate-smart and nutrition-sensitive agriculture, a stronger youth focus, and institutional capacity building to cope with some of the weak-nesses of previous phases. The monitoring and evaluation process has also been strengthened. Some programs have had to rethink their activities due to the occurrence of drought during their imple-mentation. For example, due to lack of rainfall, the initial design of one project revolving around water harvesting had to be rethought, and the activity’s budget was reallocated to emergency aid.

MONITORING, EVALUATION, AND LEARNING STRATEGIESAlthough some projects are not evaluated at all and no learnings are available, most government agencies, NGOs, and donors monitor and evaluate their programs using their own system and focus. Some hire an external evaluator as well.

Large government programs have created committees at the different administrative levels to oversee implementation of their programs and conduct follow-ups. Some institutions conduct evaluations quarterly and annually. Others conduct baseline, midline, and endline surveys, and a few also conduct an impact evaluation with a

counterfactual, or use checklists to monitor and evaluate program results.

Every project develops its own output and outcome indicators to determine if the project has met its goals. Indicators in use include improved access to water, share of communities using climate services, percentage of land rehabilitated, presence of women in leadership positions, community food deficits, dietary diversity, and wasting and stunting rates. Most stakeholders affirmed that their programs collect sex-disaggregated data, but other differentia-tors, such as age group, are not generally collected.

The lack of unified resilience indicators presents a serious challenge that hinders harmonization of resilience activities and coordination between sectors. A national monitoring and evaluation system for resilience interventions should be developed to better harmonize and centralize data collection and analysis. Such a system might also enable more informed decisions about the needs of different regions.

Some stakeholders said they believe monitoring and evaluation activities are not well planned or thorough enough. For example, some stakeholders recognized that baseline surveys are sometimes not conducted before the start of the project or are of poor quality.

In Afar, the main challenge in monitoring and evaluating the work done is the lack of professionals in woredas and the difficulty of accessing many areas. The coordinator of one project implemented in Afar explained that he has not been able to visit some project sites to evaluate the work done because they are too hard to reach.

Among the three big government programs, the PSNP has undergone the largest number of contin-uous and large evaluations and impact assessments. The PSNP has mixed but overall positive results, showing households that participate in the program are more resilient to adverse climatic events, such as droughts. It is important for the government to ensure that all agencies involved in the PSNP are aware of the need to target the poorest households


and to not overly focus on graduation. The final goal of the PSNP should be a program that meets the basic food and nutrition needs of all household members to keep households or their members from falling into poverty traps or spirals as a result of adverse shocks and events, including El Niño–Southern Oscillation events.

EXPERIENCES OF PAST PROGRAMSMany programs report having increased the resilience capacity of their beneficiaries. Other programs note that they could not achieve resilience goals because their beneficiaries still depend on emergency aid.

Success storiesDuring the 2015 drought caused by El Niño, Oromia Region was able to respond to the disaster without external help. The following year, a serious drought affected Bale and Borena Zones in Oromia Region as well, and the federal government did not need to get involved or provide help. The greatest success, according to key informants, is that no one dies or is displaced as a result of a drought.

Another key area of success is increased awareness of natural resource management and the understanding that farmers can actively increase their resilience by adopting a series of climate-resilient approaches and technologies. During the 2015/16 drought, in Enderta woreda (Tigray), project beneficiaries at Lemlem village witnessed how the irrigation system helped them produce vegetables and sustain their livelihoods until the government reached them.

Fodder cultivation at the community and household levels has also proved very successful. People in pastoralist areas have learned to save grass and store it to cope with shocks. Other success stories have been reported for watershed manage-ment, asset building, increased government commit-ment, animal health, human health and nutrition, youth employment, fruit production and agriculture, and financial services.

FailuresThe stakeholders interviewed reported specific failures or elements their programs were not able to achieve in the different regions under study.

In Oromia, few people can afford green tech-nologies such as biogas and solar systems that the Environment, Forest and Climate Change Authority had funds to build. Also, a farmer-managed natural regeneration program promised the community it would develop carbon trading after the forest was fully recovered. However, due to budget cuts, it could not fulfill its promise, disappointing the community. Lack of commitment of staff and the government, and lack of rapport with the community were also mentioned as key failures in Oromia. Furthermore, key informants noted that graduation from the PSNP was sometimes rushed or fabri-cated, with the result that beneficiaries fall back into poverty traps shortly after graduating.

In Tigray, farmers faced challenges with improved cookstoves from the Agency of Energy and Mines. Users complained about the small size of their improved stoves because putting firewood in them was hard. “This clearly shows how much we failed to build awareness,” explained one interviewee. Further, biogas pits were not always completed even if reported as such, and the quality of construction was sometimes poor.

In Afar, some credit and savings cooperatives were not being adequately supported, resulting in weak structures. Also, due to lack of risk manage-ment in pastoral areas, early warning was not effec-tively implemented, and information dissemination was delayed.

MAIN CHALLENGES AND RECOMMENDATIONSSeveral recommendations emerge from both the synthesized literature and the KIIs. Some of the main challenges and recommendations provided are summarized below. Box 2 outlines the challenges and recommendations specific to agrarian and pastoralist areas.


Ensure that activities are not affected by drought and climatic shocks. Many stakeholders state that drought itself is the main challenge they need to handle while implementing resilience programs. Interventions should be better linked to addressing the key challenge at hand. Sometimes some activi-ties, such as fodder development or construction of water-harvesting infrastructure, cannot be imple-mented due to lack of water. Even the livestock provided by some programs have been lost due to the occurrence of drought. Contingency budgets that some programs use in times of severe droughts can help.

Ensure community involvement and ownership. Some stakeholders demand more effort on awareness creation. Understanding communities’ needs, listening to their problems, and involving them in the needs assessment, design, implemen-tation, and monitoring of the project is critical to designing successful interventions, ensuring smooth participation and commitment, and promoting sustainable outcomes. One stakeholder highlighted the importance of having the community do the work, rather than outsourcing it, to increase the sense of ownership among community members. Furthermore, the distribution of free handouts (for example, beehives) without any awareness building is seen as harmful for resilience programs because it promotes a dependency culture.

Program modality: Do not focus on singular interventions that might target many but are fit for few. Interventions in one activity area are not sufficient to achieve resilience. Programs need to promote multiple interventions in different areas to achieve resilience outcomes. Similarly, farmers and pastoralists need to engage in diversified income-generating activities to enhance their capacity for resilience to shocks.

Support sustainability of projects after comple-tion. Ensuring the sustainability of projects after completion is a great challenge of NGOs. Developing a policy to encourage government

supervision or follow-up and supporting community facilitators after a project phases out were presented as possible solutions. Longer-term programs are seen as more sustainable.

Increase institutional capacity and integration. Lack of institutional leadership, professionalism, commitment, and collaboration from the govern-ment was also highlighted by some stakeholders. Furthermore, communication between the federal, woreda, and kebele levels needs to improve. Capacity-building activities to enhance the technical background of some government workers can help improve their performance. It is also important to align project activities with government activities for synergies. Creating steering committees to enhance the level of coordination between sectors has proven to be effective. Slow budget release from the federal government to the local level was also pointed out as an area in need for improved coordination.

Make early warning systems and climate data more user-friendly. Limited availability of meteoro-logical data was mentioned as a constraint for the design of effective interventions. Improving climate forecasts at the regional level and producing longer-term estimates will help support early warning systems and disaster risk prevention.

Early warning systems are an important contrib-utor to any resilience program. Such programs could be improved by (1) choosing information that is relevant to people’s livelihoods and their local geography (for example, the NMA has certain calendar data associated with crop production that are not useful for lowland pastoralists, who use a different calendar), (2) transmitting information at the right time to be actionable, (3) transmitting information in an inclusive manner (for example, in local languages, accessible to illiterate people and to poor people who may not have access to a radio or mobile phone), (4) making sure that people comprehend the information and know whom to contact in case of questions, (5) including strategies to cope with predicted shocks, and (6) considering


BOx 2 CHALLeNgeS AND reCOMMeNDATiONS BY LiVeLiHOOD AreA

Agrarian areas

Address scarcity of seeds. Seeds are sometimes very expensive or very hard to get, which has created problems of supply for some projects in Oromia.

Improve market accessibility. Challenges related to market accessibility to support cereal, vegetable, and livestock produc-tion were encountered in Oromia.

Facilitate access to irrigation by the poor. Vulnerable farmers may not get access to irrigation systems because of their geo-graphic location. In contrast, irrigation systems frequently are sited next to farmers who are already relatively well-off.

Diversify rural income sources to reduce pressure on natural resources. Agriculture puts important pressure on land, which is almost at the tipping point in some areas. Promoting other forms of livelihood and amendments in the national eco-nomic policy can help prevent soil degradation in some areas.

Improve provision of agricultural information. Agricultural advisory services can contribute to increasing productivity by providing climate-adaptive seeds and context-specific advice related to the use of inputs.

Consider the cropping cycle when planning. Programs should consider the cropping cycle when planning activities such as land preparation, planting, and harvesting. For example, activities such as soil conservation should be implemented when farmers are not busy with other activities. Water conservation activities should also be planned before the rainy season starts.

Pastoralist areas

Address conflicts over natural resources. Conflict between different clans for natural resources is reported as an important obstacle to resilience building in some areas, particularly around borders.

Review pastoralist settlement policies. Donors have attempted to change the way of living of pastoralist communities with-out understanding the dynamism of these societies. The fact that people shift from one place to another as a livelihood strategy poses some challenges to resilience programs, especially with infrastructure construction activities. Thus, it is very important to design programs in collaboration with the beneficiary population and create a sense of belonging in pastoralist areas.

Incorporate pastoralist culture in resilience activities. Pastoralist culture influences the implementation capacity of some programs. Several stakeholders reported difficulties in involving women in interventions, because they are not allowed to go out in public. Furthermore, in times of shock, pastoralists share their resources with their neighbors, which may go against some project goals by dissipating the assets saved by the household. Bringing microcredit associations into pastoralist com-munities has been difficult because the associations initially did not trust pastoralists. At the same time, pastoralists believe that a person with a debt should not be buried until the debt is paid. This practice may ensure that microcredit associations get paid back.

Urgently improve basic services for pastoralists. Lack of infrastructure (such as roads, health centers, schools, and so on) is one of the biggest challenges that resilience programs face in Afar and other pastoralist areas. Furthermore, the materials required by some interventions can take a long time to reach project sites, which delays program implementation.

Provide livestock marketing and veterinary services and better markets to pastoralists. The price of livestock is not fixed by the owners but by brokers and traders. As a result, middlemen benefit at the expense of poor farmers. Similarly, veterinary services need to improve to prevent the spread of animal diseases.

Improve human resources and capacity. Due to the remoteness of pastoralist areas, it is hard to find capable, skilled person-nel to implement programs. Staff turnover is also high. Designing a reward system for people implementing projects in pasto-ralist areas was suggested as a mechanism to increase their morale and commitment.

Source: Authors.


the co-production of climate information services (whereby, for example, communities tell climate forecasters when the rain has actually started).

Improve monitoring, evaluation, learning, and technological innovation. Resilience is difficult to measure because it is a multidimensional concept observed in the long term and cannot be easily quantified. Harmonization and sharing of lessons learned by different programs is very much needed. Monitoring, evaluation, and learning activities need to be geared more toward resilience building and resilience thinking.

Improve resilience services: Improve microfi-nance and create a national insurance company. Key services that need improving include microfi-nance, insurance, and rural infrastructure, including markets. In Ethiopia, the field of insurance providers is very limited. More than 80 percent of the farmers who have insurance in Ethiopia are insured with international insurance companies, which have high premium rates. Creating a national insurance company for farmers would help lower the premiums and encourage more farmers to invest in insurance products. Another way to increase the uptake of climate insurance is through promoting group

Figure 5 eNABLiNg eNVirONMeNT KeY iNTerVeNTiONS

Key indirect interventions in the enabling environment

Infrastructure

Focus on better infrastructure that helps facilitate resilience and can help pro-vide a buffer in times of water stress

• Improve road networks (market and input access)

• Develop improved storage infrastructure (to reduce post-harvest losses)

• Create infrastructure to store water, such as dams or on-farm tanks and ponds (water access)

Risk and volatility

Focus on increasing the stability and reducing the vulnerability of household indicators to different types of risks

• Link with private insurance against natural hazards

• Improve collection of and access to early warning and weather information

• Improve extension support

Institutions and markets

Focus on providing stronger institu-tions and better policies to help people improve their resilience

• Continue reform of land rights

• Support the functioning of the Ethiopia Commodity Exchange

• Enable farmer access to plant and livestock disease control methods

• Improve access to finance to increase adoption of technologies

• Strengthen trust in cooperatives and their capacity

Key direct interventions for households and communities

Food production

Focus on different techniques and technologies that help increase agricultural productivity

• Promote a functioning and dynamic private sector for fertilizers, seeds, credit, and market information systems

• Increase farmers' access to improved germplasm and more resistant livestock breeds

• Improve food security under climate change through promoting diversified income

Natural resources

Focus on management and policies that conserve and enhance natural resources

• Strengthen current efforts to reduce deforestation and land degradation

• Move away from biomass fuels and toward cleaner stove technologies

• Enrich understanding of how climate change impacts the interactions among crops, livestock, and other productive uses, such as forests

Source: Authors.


insurance. Investment risks can be mitigated through risk sharing, which has also been shown to increase the demand for insurance products among farmers (Dercon et al. 2014).

Focus on youth unemployment and migration. Many programs have special provisions (skills training and local investment opportunities) to enhance the livelihood opportunities of youth, who often migrate to cities in the absence of alternative livelihoods. In pastoralist areas, working with youth is especially important because, if they lose their animals due to lack of rain, they often lack the skills for other work without proper support.

Protect communal land. Limitations on protecting communal land were reported. Communities who

live on communal land are sometimes afraid that restored land would be redistributed among indi-viduals. Some stakeholders requested changes to the existing land use policy to consider the reality of communal land tenure systems.

KEY INTERVENTIONS IN THE ENABLING ENVIRONMENT Several interventions in the enabling environment appear necessary to achieve increased levels of economic resilience and food security. Some are already the target of programs and projects, and are mentioned to reiterate their importance. Figure 5 (see prior page) summarizes these investments.


3. EL NIÑO–SOUTHERN OSCILLATION IMPACTS ON AGRICULTURE AND THE NATIONAL ECONOMY Jawoo Koo, James Thurlow, Hua Xie, Ricky Robertson, Carlo Azzarri, Ho-Young Kwon, and Beliyou Haile

a rangE of intErvEntions havE bEEn idEntifiEd that, if implemented, could help mitigate the adverse effects of climate shocks, such as El

Niño–Southern Oscillation (ENSO) events, on the Ethiopian economy and the food

security of its population. As outlined in Chapter 2, these interventions include, among

others, on-farm investments in technology and irrigation infrastructure, investments

in roads and grain storage facilities to expand and stabilize food markets, and

social transfers to provide households with a cushion against immediate crises and

opportunities for longer-term recovery. However, resource constraints and competing

interests mean that there are sometimes trade-offs associated with pursuing policies

aimed at building resilience to climate shocks versus policies aimed at achieving

other development objectives. Therefore, to motivate resilience-building policies, it

is necessary to assess the costs of inaction; to measure policy effectiveness using

recognized outcome indicators; and to identify synergies between, say, resilience and

development interventions and objectives.

To this end, this chapter measures the conse-quences of severe climate shocks for Ethiopia’s economy and its population. More specifically, the chapter provides new quantitative estimates of the impacts of the 2015/16 El Niño event on the agricultural sector and the national economy. The 2015/16 El Niño was selected as a representative El Niño event in the modeling framework because of its severe droughts and consequent economywide impacts. As discussed in Chapter 1, this is not the first study to evaluate climate shocks in Ethiopia. However, previous studies have typically focused on agriculture in isolation from the broader economy, or

have measured the incremental costs from long-term climate change while overlooking the substantial economic costs caused by historical climate vari-ability. These limitations make it more difficult to explain (1) how climate shocks that originate within agriculture can have economywide implications, (2) why action is needed today to build resilience to existing climate variability, and (3) that resilience-building interventions are often consistent with broader development goals.

The Gender, Climate Change, and Nutrition Integration Initiative (GCAN) framework introduced in Chapter 1 describes the many complex factors

41

determining resilience to climate shocks. Within that framework, this chapter focuses on tracking the pathways between climate shocks and household welfare, and measures outcomes (well-being) at both national and household levels. ENSO is associated with extreme weather events throughout Africa South of the Sahara, but the economic impacts caused by these events are difficult to disentangle from those of other climate shocks, natural disasters, and economic cycles. This complicates the design of policies and response mechanisms that could help mitigate economic damages and welfare losses.

The ENSO cycle encompasses fluctuations in temperature between the ocean and atmosphere across the east-central Equatorial Pacific and consists of two opposite phases: El Niño and La Niña. El Niño and La Niña are the warm and cold phases of ENSO, respectively. Deviations from normal surface temperatures can have large-scale impacts, not only on ocean processes but also on global weather. El Niño and La Niña episodes occur every 2 to 7 years and typically last 9 to 12 months. ENSO’s effect on weather patterns can be forecast, albeit with varying degrees of accuracy, and this makes it possible to design policies and prepare emergency responses in advance of extreme shocks.

ENSO’s adverse effect on agriculture is most concerning, given the sector’s linkages to national poverty and food security. Agriculture is an important economic sector in Ethiopia, providing over one-sixth of national employment and one-half of GDP. Most poor people live in rural areas and work in agriculture, and so are vulnerable to climate shocks. Moreover, climate shocks such as floods and droughts have had severe adverse impacts on the overall agriculture–food system (AFS) in Ethiopia. This includes farmers as well as workers in down-stream sectors and consumers purchasing foods in local and national markets. ENSO’s impacts on agri-culture therefore have economywide implications. Whereas a growing body of empirical evidence measures the effects of climate change and vari-ability on agricultural and national economies, few studies have focused on the measurement of

economywide outcomes. Although there is an expanding literature on how natural disasters are managed in Ethiopia, quantitative evaluation of the policies and investments that could help avoid at least some of the damages caused by ENSO events is limited.

ESTIMATING AGRICULTURAL AND ECONOMIC LOSSES

Methods for evaluating climate impactsNumerous studies have evaluated climate impacts on developing economies (see Tol 2009 for an early review). Most studies still focus on agricultural impacts, because this is where damages to low-income countries are expected to be greatest and most immediate, and where there is the strongest evidence on the link between climate signals and outcomes. Agriculture-focused studies can be divided into two groups based on the methods they use. The first and larger group uses biophysical crop models to capture local agroecological condi-tions (for example, soil type and elevation), climate patterns (for example, rainfall, temperature, and solar radiation), and crop-specific plant physiology (see, for example, Parry et al. 2004). Changes in weather patterns, either historical or projected, are imposed on these crop models to estimate how crop yields are affected. An alternative approach uses statistical or econometric analysis (rather than process-based modeling) to estimate the historical relationship between weather patterns and crop production (Lobell et al. 2011). These ex post relationships are then used to project ex ante the impact of climate changes or shocks. Statistical approaches have the advantage of being empirically grounded, but they are limited by how well the historical record captures climate events of interest (that is, they are susceptible to conducting “out-of-sample” projections). Crop models, on the other hand, can be used to examine hypothetical climate scenarios, because they are less dependent on historical data, but they are generally more data demanding and more reliant on model assumptions. Given their respective limitations, our analysis uses both crop models and econometric


analysis to estimate the impacts of severe climate shocks on Ethiopia’s agriculture sector (see detailed discussion below). Combining these approaches also broadens our impact assessment to include both crops and livestock.

Biophysical analysis may be of limited use for decision-makers, who need to understand the economic and social impacts of climate changes. Most studies at the country level therefore combine biophysical and econometric models to link climate signals to economic outcomes. Again, the literature can be divided into two groups. The first group uses agriculture-sector models to capture how supply shocks caused by climate variability affect agricultural prices and demand (see, for example, Block et al. 2008; Nelson et al. 2009). These models track changes in demand and supply for detailed crops, livestock products, or both, and sometimes include subnational regions, although rarely at the same level of detail as crop models. The main short-coming of these models, however, is that they do not capture spillovers or linkages between agriculture and the rest of the economy. These can be important effects, especially in low-income countries, where agriculture is a large part of the economy and where much of the nonagricultural sector is still agriculture related. A growing number of studies capture these spillovers by using economywide models, such as computable general equilibrium (CGE) models (Hertel et al. 2010; Yu et al. 2010). Like agriculture-focused models, these models capture the direct impacts of climate shocks on agricultural production, as well as the indirect effects on downstream sectors and nonfarm/urban households. Our analysis uses an economywide model because of its broader capture of the economy and food system.

Ours is not the first study to combine crop and CGE models to assess the economic effects of climate change in low-income African countries. Most studies, however, focus on the incremental effects of long-term climate change on agriculture rather than on historical climate shocks (see, for example, Arndt et al. 2012; Calzadilla et al. 2013). More recent studies assess the effects of climate

uncertainty using stochastic simulation techniques, but although these studies capture the effects of droughts and floods on the agriculture, transport, and energy sectors, they still do not separate histor-ical climate variability from long-term climate change (Arndt et al. 2014; Arndt and Thurlow 2015). Few published studies focus on historical events. Pauw and colleagues (2011) combined statistical analysis and a CGE model to estimate the impact of droughts and floods on Malawi’s economy, and Thurlow, Diao, and Zhu (2012) combined crop and CGE models to estimate the damages caused by historical droughts on the Zambian economy. These authors paid partic-ular attention to the ENSO-related droughts of the early 1990s. We will use an approach similar to that of these studies in our integrated assessment of the 2015/2016 drought in Ethiopia.

Integrated analytical framework for EthiopiaThis report employs a framework that combines an analysis of ENSO impacts on crop and livestock agriculture, as well as the spillover impacts from agriculture to the rest of the economy (Figure 6). This integrated approach first examines historical climate data, including variability in rainfall and tempera-ture. It focuses exclusively on the weather pattern that occurred during the latest severe El Niño event, in 2015/16, as a representation of ENSO-induced climate variability, because data are available to validate the estimated impacts. Short-term climate fluctuations during this event year are compared with

“neutral” weather years that did not experience ENSO shocks, in order to isolate ENSO-related deviations in rainfall and temperature. Changes in weather variables during 2015/16’s main crop-growing season (meher) are translated into crop productivity outcomes using a process-based model. More specifically, crop models in the Decision Support System for Agrotechnology Transfer (DSSAT) are used (Hoogenboom et al. 2017; Jones et al. 2003). DSSAT is widely used by agricultural researchers to understand crop production system dynamics and simulate different farm management and environ-mental changes in Ethiopia (for example, Araya et

3. EL NIñO–SOUTHERN OSCILLATION IMPACTS ON AGRICULTURE AND THE NATIONAL ECONOMY 43

al. 2015; Kassie et al. 2014, 2015; Mohammed et al. 2017) and climate variability impacts associated with ENSO (for example, MacCarthy et al. 2017; Sarkar, Ortiz, and Balkcom 2017).

The spatial crop models in the integrated analyt-ical framework (Figure 6) estimate ENSO-affected seasonal yield deviations of major crops in a grid-based spatial analysis framework. Daily historical weather data are spatially interpolated from weather

station data, linked with the corresponding ENSO phase, and then used as inputs to the crop models that estimate yield changes for two key crops: maize and wheat. The crop models also estimate how yield responses differ when using improved or traditional seed varieties, with and without chemical fertilizer, and with and without irrigation infrastructure. Crop-specific yield deviations are estimated for 5 arc-minute x 5 arc-minute (about 10 km x 10 km) grids

Figure 6 iNTegrATeD ANALYTiCAL FrAMeWOrK

Source: Authors.

Note: CGE = computable general equilibrium; DSSAT = Decision Support System for Agrotechnology Transfer; GDP = gross domestic product.

Weather and climate dataRainfall, temperature (maximum, minimum)

Crop management optionsSeed varieties, fertilizer, irrigation, crop calendar

Crop and livestock dataHerd size by type of livestock

Social accounting matrixProducers, households, government

Market policy optionsForeign trade taxes and subsidies, social safety nets

Crop yield impactsSpatial crop models (DSSAT)

Production impactsStatistical trends analysis

Economywide impactsDynamic spatial CGE and microsimulation model

Economic outcomesNational and agricultural GDP, rural and urban poverty, and so on

Supply outcomesQuantities by crop, livestock type, and region

Yield outcomesYields by crop and region under alternative

technologies and practices


covering the cropland extent of the entire country. Finally, gridded results are aggregated into major subnational regions using cropland use patterns (Figure 7). The crop modeling provides insights into the potential effectiveness of adopting improved farming technologies and practices in offsetting the productivity impacts caused by ENSO.

This process-based crop modeling is comple-mented with statistical analysis of secondary crop and livestock production data for major subnational regions. The impact of the representative 2015/16 El Niño event on crop and livestock production is estimated. For the former, the focus is exclusively on the meher cropping season. Like the crop modeling, the statistical analysis distinguishes between Ethiopia’s five subnational agroclimatic regions. However, it examines a wider range of crops, namely teff, barley, wheat, maize, and other cereals (that is, rice and oats), as well as noncereal grains. Impacts on other crops are derived from historical correlation of production patterns. Finally, and most important,

the statistical analysis considers changes in produc-tion quantities, not just yields. In other words, it includes the effect of ENSO on harvested land area, thereby capturing the fact that farmers may abandon fields due to complete crop failure.

The estimated impacts of ENSO events on crop and livestock production are then imposed on a dynamic CGE model. This class of model captures all producers and consumers in an economy, including the government, as well as interactions with the rest of the world (for example, imports and exports). All sectors and households are disaggregated across major subnational regions. Region- and crop-specific productivity shocks thus translate into changes in agricultural and national GDP, employment, and prices. The model reacts to crop- and sector-specific productivity changes by reallocating resources and products between sectors and households to minimize overall losses to the economy (that is, it takes autonomous adaptation into account). The CGE model is also linked to a survey-based micro-simulation module that tracks changes in national and subnational poverty rates. Ethiopia’s version of the CGE model developed by the International Food Policy Research Institute (IFPRI) has been used to examine a range of development policy issues, including providing background information to the country’s Growth and Transformation Plan II, and is an ideal tool for assessing the economywide impacts of large shocks to the agriculture sector.

This integrated approach to measuring the econo-mywide impacts of climate shocks is similar to that often used for long-term climate change impact studies. The DSSAT and CGE models represent some of the most sophisticated tools available for such analysis, and the high-resolution spatial databases used in both types of models are quite unique, for both Ethiopia and developing countries in general. The framework allows not only for isolation of the expected impacts of ENSO events, but also for experiments with alternative policy responses. Whereas the crop models focus on changes in on-farm technologies and practices, the CGE model focuses on market policies and safety nets.

Figure 7 eTHiOPiA’S FiVe AgrOCLiMATiC regiONS

Source: Schmidt and Thomas (2017).

Drought-prone highland

Addis Ababa

AfarAmhara

Benishangul-Gumuz

DireDawa

Gambella

Harari

Oromia

SNNP

Somali

Tigray

Lakes

Drought-prone lowland—pastoralistHumid, moisture-reliable lowlandMoisture-reliable highland—cerealMoisture-reliable highland—enset (Ethiopian banana)


It estimates how these would have mitigated the economic damages caused by a severe ENSO event, such as the 2015/16 El Niño.

This chapter therefore examines many of the interventions that were prioritized at the end of Chapter 2, including investments in infrastructure (for example, roads, storage, and irrigation), rural services (for example, extension), on-farm technolo-gies (for example, improved seeds and farm manage-ment practices), and trade and market efficiency. The chapter’s analysis examines the short- to medium-term effects of ENSO, rather than capturing the full recovery period. The latter could extend over many years and is subject to uncertainties that may be driven by social, economic, and political factors that are difficult to quantify or are beyond the scope of this study. The chapter aims to provide quantitative

evidence to identify policy options and motivate implementation.

ENSO’S IMPACTS ON WEATHER PATTERNS AND AGRICULTURAL PRODUCTION

Historical climate analysisENSO alters weather patterns in Ethiopia. On average, the warm El Niño phase is drier than La Niña during July–September (Figure 8), coinciding with the main crop-growing period of meher. Focusing on 1950–2016, statistical analysis of historical rainfall indicates that, compared with the neutral phase, average annual rainfall across the country was 8 percent lower during El Niño over this period. During La Niña, rainfall was 10 percent higher than during the neutral phase, with greater variability (Figure 9). However,

Figure 8 AVerAge MONTHLY rAiNFALL BY eL NiÑO–SOuTHerN OSCiLLATiON PHASe

Source: Authors’ reanalysis of the University of East Anglia’s CRU-TS v4.01 dataset (CRU, UEA 2018).

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec0

50

100

150

200

Ave

rag

e ra

infa

ll (m

m/m

ont

h)

El Niño Neutral La Niña


within the country, climate variability patterns are highly heterogeneous.

As mentioned above, this study separates Ethiopia into five zones depending on rainfall regime and predominant use of farmlands. This includes the drought-prone areas that generally lie toward the eastern half of the country. When compared with the neutral average, the distribution of the differ-ence in rainfall from historical climate data for July–September showed more separation between El Niño and La Niña in the drought-prone highland, humid, moisture-reliable lowland, and moisture-reli-able highland—cereal zones than elsewhere (Figure 10). Other zones did not show meaningful differences in rainfall patterns.

Crop modeling analysisCrop production data can be confounded by a wide variety of on- and off-farm factors. Before turning

to the statistical analysis, crop models are first used to simulate seasonal crop growth and produc-tivity. This allows for isolation of the impact of the ENSO-induced weather variability effects discussed above. The DSSAT crop models simulate the farming situation systematically across the entire country, rather than assessing at a few isolated locations.

IFPRI’s grid-based, spatially explicit crop modeling framework was used to generate yield estimates for every 5 arc-minute grid cell (about 10 km on a side) for maize and wheat under both rainfed and irrigated conditions. The flexibility of the crop models allowed for simulation of all planting months and several levels of fertilizer application. Using the historical daily weather data, this modeling framework generated site-specific crop yield vari-ability over the period 2009–2017. Grid cell–specific daily historical weather data were retrieved from the Weather Terrain™ database by aWhere (2018).

The assessment of impacts focused on recent years. The analysis was grouped into three periods: 2015/16 El Niño, 2016/17 La Niña, and neutral years (that is, 2012/13, 2013/14, 2014/15, and 2017/18). The simulated yield consequences of climate variability associated with the ENSO phases on each grid cell were aggregated at the agroecological zone level. Only the meher season was considered, because it accounts for around 96 percent of Ethiopia’s total annual production and was the worst-affected season during the 2015/16 ENSO event. These data were statistically analyzed to probabilistically examine the extent of potential crop yield changes (benefits and losses) under each ENSO phase. Figure 11 shows how the simulated yields within each phase—El Niño and La Niña—combined to determine the final yield impact. Potential areas suitable for irrigation at a small or large scale were considered in the analysis, based on a previous analysis by Xie and others (2014). In addition, the adoption of comple-mentary management practices and technologies, such as the doubling of nitrogen fertilizer, no-till practices, and integrated soil fertility management, was simulated to assess their potential impacts on yields under ENSO.

Figure 9 DeViATiON iN NATiONAL rAiNFALL PATTerNS DuriNg eL NiÑO–SOuTHerN OSCiLLATiON PeriODS


Note: Percentage difference in the mean and standard deviation of monthly rainfall during July–September in El Niño and La Niña years relative to the neutral phase.

El Niño La Niña

0

20

0

20

Perc

enta

ge

Diff

eren

ce fr

om

Neu

tral

10.27

-7.66

-11.05

27.43

Stan

dar

d d

evia

tion

(%)

Ave

rag

e ra

infa

ll

(%)


Figure 10 AVerAge MONTHLY rAiNFALL DuriNg eL NiÑO–SOuTHerN OSCiLLATiON PHASeS ACrOSS AgrOCLiMATiC ZONeS


Note: Historical monthly average rainfall aggregated at the agroclimatic zone level.


Drought-prone lowland—pastoralist

Humid, moisture-reliable lowland

0

100

200

300

Rainfall (mm/month)

0

100

200

300

Rainfall (mm/month)

0

100

200

300

Rainfall (mm/month)

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Moisture-reliable highland—cereal

Moisture-reliable highland—enset (Ethiopian banana)

0

100

200

300

Rainfall (mm/month)

0

100

200

300

Rainfall (mm/month)

El Niño Neutral La Niña


When the simulated yields are aggregated to the zonal level, no clear yield patterns emerge for El Niño (2015/16) and La Niña (2016/17) (Figure 11). Some yield losses during El Niño and gains during La Niña are estimated for two zones (drought-prone highlands and humid lowlands), coinciding with historical rainfall variability patterns. Other zones experienced yield increases ranging from 1 percent to 46 percent. However, when the simulated yields for the three zones with the largest rainfall vari-abilities are further disaggregated, estimated yield

losses during the 2015/16 El Niño are concentrated in just a few subregions, namely Afar, Amhara, Harari People’s National Regional State, Oromia, Tigray, and Addis Ababa (Figure 12).

When improved agricultural technologies and intensification management practices are adopted, the yield losses during El Niño are reduced. For instance, in Harari Region, maize yields are reduced by 25 percent (420 kg/ha) under the 2015/16 El Niño. Yields would remain at this low level, even during such a severe El Niño event, if the application

Figure 11 SiMuLATeD YieLD DeViATiONS FOr MAiZe AND WHeAT BY eL NiÑO–SOuTHerN OSCiLLATiON PHASe AND AgrOCLiMATiC ZONe

Source: Authors’ calculations using the International Food Policy Research Institute’s Decision Support System for Agrotechnology Transfer crop modeling framework (Hoogenboom et al. 2017; Jones et al. 2003).

Note: Yield deviations during the 2015/2016 El Niño event are relative to average yields from recent neutral years. SNNPR = Southern Nations, Nationalities, and Peoples’ Region.

El Niño

La Niña

MAIZEWHEAT

0 50

-10

6

3

3

-1

4

10

-1

10

1

-1

3

22

20

-7

9

13

10

46

20

0 50


Drought-prone lowland

Humid lowland


Moisture-reliable highland—enset

El Niño

La Niña

El Niño

La Niña

El Niño

La Niña

El Niño

La Niña

Percentagedifference in yieldfrom neutral (%)


Figure 12 YieLD DeViATiONS WiTHiN THree AgrOCLiMATiC ZONeS DuriNg THe 2015/16 eL NiÑO

Source: Authors’ calculations using the International Food Policy Research Institute’s Decision Support System for Agrotechnology Transfer crop modeling framework (Hoogenboom et al. 2017; Jones et al. 2003).

Note: Yield deviations during the 2015/2016 El Niño event are relative to average yields from recent neutral years. SNNPR = Southern Nations, Nationalities, and Peoples’ Region.

Dro

ught

-pro

ne

hig

hlan

d

Afar

Amhara

HarariPeople

Oromia

Tigray

Hum

idlo

wla

nd

Amhara

Benishangul-Gumuz

Tigray

Mo

istu

re-r

elia

ble

hig

hlan

d—

cere

al

AddisAbaba

Amhara

Benishangul-Gumuz

Oromia

SNNPR

MAIZEWHEAT

-40 -20 0 20 40 60-40 -20 0 20 40 60

-25

-2

-25

-12

-3

-18

12

-12

-15

20

18

6

10

4

-28

-5

0

-3

-2

-19

-10

19

11

42




rate of nitrogen fertilizer doubled. Figure 13 shows the relative yield increases under the various farm management practices considered in this report. The crop modeling analysis concludes that investing in the kinds of interventions that raise on-farm

productivity can be highly effective in preventing cereal yields from declining to the low levels expected under current management practices during a severe El Niño event.

Figure 13 eFFeCTS OF FArM TeCHNOLOgieS iN rAiSiNg YieLDS uNDer eL NiÑO AND LA NiÑA BY AgrOCLiMATiC ZONe

Source: Authors’ calculations using the International Food Policy Research Institute’s Decision Support System for Agrotechnology Transfer crop modeling framework (Hoogen-boom et al. 2017; Jones et al. 2003).

Note: ISFM = integrated soil fertility management; N = nitrogen, BAU = business as usual.

El Niño La Niña

N, B

AU

, RA

INFE

D

2xN

, No

-Till

, RA

INFE

D

2xN

, ISF

M, R

AIN

FED

2xN

, No

-Till

, IR

RIG

ATE

D

2xN

, ISF

M, I

RR

IGA

TED

N, B

AU

, RA

INFE

D

2xN

, No

-Till

, RA

INFE

D

2xN

, ISF

M, R

AIN

FED

2xN

, No

-Till

, IR

RIG

ATE

D

2xN

, ISF

M, I

RR

IGA

TED

MA

IZE

WH

EA

T

0

50

100

0

50

100

Yie

ld d

iffe

renc

e fr

om

bas

elin

e (%

)


Drought-prone lowland—pastoralist

Humid lowland


Moisture-reliable highland—enset

Yie

ld d

iffe

renc

e fr

om

bas

elin

e (%

)


Statistical crop and livestock analysisThe crop modeling analysis revealed wide variation in ENSO’s impacts across subnational regions and across El Niño and La Niña events. Historical produc-tion data are now used to estimate the production losses that occurred during the 2015/16 El Niño event. Information on crop production quantities, harvested area, and yields was extracted from annual Agricultural Sample Surveys (Ethiopia, CSA, various years). As with the above weather analysis, produc-tion outcomes during 2015/16 are estimated relative to production trends from recent neutral years for which data are available (that is, 2010/11 to 2014/15 and 2016/17 to 2017/18).

Figure 14 illustrates this approach using national production quantities for teff—Ethiopia’s main staple crop—during the meher season. The simplest approach would be to compare the production quantity in 2015/16 (that is, 4.47 million tons) with the production quantity in the preceding year, 2014/15 (that is, 4.75 million tons). This would suggest that El Niño reduced the national average teff yield by

5.9 percent (that is, 4.47/4.75 = 0.941). However, this ignores the expected increase in production that would have occurred without the ENSO shock, and so underestimates the losses that occurred due to El Niño. To generate a more accurate counterfactual, production levels are first projected using historical data for the non-ENSO years. As shown in Figure 14, the trend line slopes upward, indicating that produc-tion levels were increasing year over year and would likely continue to do so in the absence of the ENSO shock. More specifically, national teff production would have been an estimated 4.82 million tons in 2015/16 without ENSO, somewhat higher than observed production levels in 2014/15. The revised estimate is that El Niño caused national teff produc-tion to fall by 7.3 percent rather than by 5.9 percent (that is, 4.47/4.82 = 0.927).

The approach above is replicated for other grain crops; Table 2 shows the resulting yield deviations. The final column reports the estimated percentage change in yields caused by the 2015/2016 El Niño event, with -7.3 percent for national teff production.

Figure 14 eSTiMATiNg NATiONAL TeFF PrODuCTiON LOSSeS DuriNg THe 2015/16 eL NiÑO

Source: Authors’ calculations using crop production data from Agricultural Sample Surveys (Ethiopia, CSA, various years).

4.75

4.47

4.82

3.0

3.5

4.0

4.5

5.0

5.5

6.0

2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18

Pro

du

ctio

n q

uan

tity

(mill

ion

s o

f me

tric

to

ns)

Observed production Projected trend


Large production losses were also estimated for barley, maize, and other cereals—the latter includes rice and oats, which are far less important for Ethiopia than the other cereals. These changes in production quantities are caused by either declines in crop yields or reduced area harvested. As shown in the first and second columns of the table, reductions in harvested land area can be a major driver of overall produc-tion losses. For example, teff production fell by 7.3 percent, but most of these losses were caused by a reduction in harvested area (4.0 percent) rather than a reduction in yield (3.3 percent). Area losses also exceeded yield losses for barley and other cereals. (Note that the sum of yield and area losses need not equal production losses due to interaction effects.) Overall, yield losses dominate the decline in total cereals production caused by El Niño.

The estimated yield losses are smaller for wheat than for maize, broadly consistent with the findings from the crop models. For both crops, however, the statistical analysis indicates that declining yields were offset by increased harvested land area at the national level. This underscores the need to assess ENSO’s impact at the subnational level.

Following the same approach, Table 3 reports the estimated ENSO-related changes in cereal production levels during 2015/16 across the five agroclimatic zones. The first column in Table 3 is the national weighted average loss and so corresponds to the final column of Table 2. The results show that El Niño’s negative impacts on production quantities are consistently larger in the drought-prone lowlands (DPL-P) and humid or moisture-reliable lowlands (MRL). Total cereal production fell by an estimated 10 percent in both of these regions. Losses were smaller, but still significant, in the other three regions.

The regional results from the statistical analysis differ from those of the crop modeling analysis. The latter reported yield gains in the moisture-reliable highland regions (MRH-C and MRH-E in Table 3) during the 2015/16 El Niño event. One explana-tion for this difference is that, as discussed above, production quantities are determined by both yield and area, and so the yield deviations estimated by

TABLe 3 eSTiMATeD grAiN PrODuCTiON LOSSeS BY SuBNATiONAL regiON DuriNg THe 2015/16 eL NiÑO

Quantity deviation from trend projection by region (%)

National DPH DPL-P MRL MRH-C MRH-E

Cereals -4.75 -5.69 -10.36 -9.76 -3.80 -7.26

Teff -7.30 -9.81 -18.78 -17.90 -5.47 -18.08

Barley -5.59 -6.67 -0.20 -5.85 -6.13 3.05

Wheat -1.30 -5.13 10.85 -13.37 -1.11 3.35

Maize -5.26 -3.28 -12.31 -8.18 -4.73 -6.14

Sorghum and millet

-4.03 -4.61 -15.48 -7.93 -1.96 -16.35

Other cereals -9.69 -1.31 -9.21 -45.69 -10.00 -7.43

Noncereals -2.87 -5.67 -2.02 -5.61 -2.85 2.55


Note: DPH = drought-prone highland; DPL-P = drought-prone lowland—pastoralist; MRL = humid, moisture-reliable lowland; MRH-C = moisture-reliable highland—cereal; and MRH-E = moisture-reliable highland—enset (Ethiopian banana).

TABLe 2 eSTiMATeD NATiONAL grAiN PrODuCTiON LOSSeS DuriNg THe 2015/16 eL NiÑO

Deviation from trend projection (%)

Crop yield (MT/ha)

Harvested area (ha)

Production quantity (MT)

Cereals -3.72 -0.90 -4.75

Teff -3.25 -4.02 -7.30

Barley -2.33 -3.48 -5.59

Wheat -1.40 0.29 -1.30

Maize -6.06 1.01 -5.26

Sorghum and millet

-5.58 1.96 -4.03

Other cereals 0.53 -9.92 -9.69

Noncereals -3.73 0.66 -2.87


Note: MT = metric ton.


the crop yield models may not adequately capture the factors that led to a decline in harvested land area. The historical production data appear to confirm this explanation.

For example, small increases in average yields were reported in Amhara and Oromia Regions during the 2015/16 meher season, but overall production quantities fell, implying that a decline in harvested land area was the sole driver of the decline in production levels. It should also be recog-nized that the historical data include more than just the effects of ENSO-related weather patterns—they also include market and other economic responses to the food shortages in specific regions and throughout the economy. This might include lower

demand for food products caused by falling farm incomes, as well as an increase in food prices that would encourage unaffected regions to increase production to supply deficit regions. To untangle the interactions of these complex economic forces, it is necessary to use the CGE model, which captures the workings of the AFS and the national economy.

Finally, the impact of the 2015/16 El Niño on livestock herd sizes is estimated following the same approach used to estimate crop production losses (Figure 15). Livestock numbers from recent neutral years (that is, 2010/11 to 2014/15 and 2016/17 to 2017/18) are used to project what herd sizes would have been in the absence of the ENSO shock. Historical information was drawn from the livestock

Figure 15 eSTiMATeD LiVeSTOCK LOSSeS BY SuBNATiONAL regiON DuriNg THe 2015/16 eL NiÑO


Note: DPH = drought-prone highland; DPL-P = drought-prone lowland—pastoralist; MRL = humid, moisture-reliable lowland; MRH-C = moisture-reliable highland—cereal; and MRH-E = moisture-reliable highland—enset (Ethiopian banana).

-0.2

-2.4

-22.7

-5.8

3.9

0.7

-8.9

-5.3

-24.4

-9.5

-3.3

0.6

-30

-25

-20

-15

-10

-5

0

5

10

National DPH DPL-P MRL MRH-C MRH-E

Cha

nge

in h

erd

siz

e (p

erce

nta

ge)

Cattle Goats & sheep


section of the country’s Agricultural Sample Surveys, which measures herd sizes at a single point during the year. In other words, the data do not fully capture seasonal variations, although stock-based analysis, as opposed to flow-based analysis, is less affected by seasonality.

The figure reports the estimated deviations in herd sizes caused by the 2015/16 El Niño. At the national level, cattle herds were only slightly lower than is typical, at -0.2 percent. However, this hides considerable variation at the subnational level. Cattle herds are estimated to have fallen by 22.7 percent in the drought-prone lowlands (DPL-P in Figure 15), where most pastoralists reside. Cattle herds also declined in the humid lowlands (MRL) and drought-prone highlands (DPH), but they were slightly larger in the two moisture-reliable highland regions (MRH-C and MRH-E), relative to what would have been expected without the effects of ENSO. The figure shows similar regional variation for small ruminant herds. The same analysis was conducted for poultry and other livestock, not shown in the figure. Results indicate large declines in poultry flocks, although this may reflect some substitution between animal-sourced foods resulting from cattle and small ruminant deaths.

In summary, the 2015/16 ENSO event is found to have had a significant impact on both weather patterns and agricultural production in Ethiopia. National weather conditions were generally drier during El Niño and wetter during La Niña, although the extent of these trends varied considerably across subnational regions. The crop modeling analysis showed how ENSO’s impacts varied across regions, even within agroclimatic zones. That said, crop simulations indicated that maize and wheat yields declined in the more drought-prone parts of Ethiopia during El Niño. This was confirmed by the statistical analysis, which estimated substantial declines in production of cereals and livestock herds, particularly in the drought-prone lowlands. The next section imposes these estimated production losses on an economywide model to translate productivity

changes into changes in economic indicators, taking account of spillovers between agriculture and other sectors of the economy.

ENSO’S IMPACTS ON THE NATIONAL ECONOMY

Agriculture in the national economyEl Niño’s impacts may be felt throughout Ethiopia, although this analysis has focused only on changes in agricultural production. This is because agriculture’s direct contribution to production and employment hides the sector’s indirect importance for many other parts of the economy. Using a spatial economywide model, it is possible to measure the direct and indirect effects of the agricultural production shocks estimated in the previous section. IFPRI’s Rural Investment and Policy Analysis model (2017) is used; it has, as its core database, a social accounting matrix (SAM) that captures all income and expenditure flows between all economic actors in the country, including producers, consumers, government, and the rest of the world (Box 3).

The economywide model of Ethiopia is bench-marked to a SAM for 2010/11. It is necessary to adopt a base year that predates the ENSO event of interest and that represents a neutral or non-ENSO-affected year. As mentioned above, the most recent neutral year prior to the 2015/16 ENSO event was 2014/15. Unfortunately, the data needed to construct a detailed spatial SAM for that year were not available at the time of the analysis. In using a 2010/11 benchmark, we are effectively assessing the economic impacts of the 2015/16 ENSO event, assuming that an event of similar magnitude and profile had occurred in 2010/11. Since the Ethiopian economy grew between 2010/11 and 2014/15, the dollar-dominated damages reported in this chapter should be treated as lower-bound estimates. However, despite its rapid economic growth, Ethiopia has yet to experience significant structural change, and so the sectoral linkages between agri-culture and the rest of the economy are unlikely to have changed dramatically over this four-year period.


Using a 2010/11 SAM for Ethiopia, the size of the country’s AFS can be estimated (Table 4). Agriculture was only 12.5 percent of total GDP in 2011. However, when downstream agricultural processing, input production, and agriculture-related trading and transporting are included, the contribution of AFS rises to 30.3 percent of GDP (that is, 2.5 times agriculture’s direct GDP share). Shocks to agriculture can therefore have important economywide implications.

The SAM separates the Ethiopian economy into 49 sectors (producers) and 30 household groups (consumers), both of which are further separated across the five agroclimatic zones (Figure 7). Together, these household groups capture the entire population, separated across rural and urban areas and per capita expenditure levels. Similarly, the

TABLe 4 AgriCuLTure–FOOD SYSTeM SHAre OF gDP AND eMPLOYMeNT, 2010/11

Share of national total (%)

GDP Employment

National economy 100 100

Agriculture–food system 52.9 84.4

Direct production 44.1 79.8

Agriculture 42.1 79.0

Agroprocessing 2.1 0.7

Input production 1.3 0.4

Trade and transport 7.5 4.2

Source: Authors’ calculations based on computable general equilibrium model using 2010/2011 Ethiopia social accounting matrix data from the International Food Policy Research Institute’s Rural Investment and Policy Analysis Model (Benfica and Thurlow 2017).

Notes: Employment is defined as workers in primary jobs. GDP = gross domestic product.

BOx 3 iFPri’S rurAL iNVeSTMeNT AND POLiCY ANALYSiS MODeL

Rural Investment and Policy Analysis (RIAPA) is a recursive dynamic computable general equilibrium model that simu-lates the functioning of a market economy, including markets for products and factors (that is, land, labor, and capi-tal). RIAPA measures how impacts are mediated through prices and resource reallocations, and ensures that resource and macroeconomic constraints, such as limits on inputs or foreign exchange, are respected. RIAPA provides a con-sistent “simulation laboratory” for quantitatively examining value-chain interactions and spillovers at the national, sub-national, and household levels.

RIAPA divides the national economy into different sectors and household groups that act as individual economic agents. Producers maximize profits and supply output to national markets, where it may be exported, combined with imports, or both, depending on relative prices, with foreign prices affected by exchange rate movements. Producers combine factors and intermediate inputs using sector-specific technologies. Maize farmers, for example, use a unique combination of land, labor, machinery, fertilizer, and purchased seeds. Workers are divided by education level, and agricultural capital is separated into crop and livestock categories. Labor and capital are in fixed supply, but less-edu-cated workers are treated as underemployed. Producers and households pay taxes to the government, which uses these and other revenues to finance public services and social transfers. Remaining revenues are added to private savings and foreign capital inflows to finance investment; that is, investment is driven by levels of savings. RIAPA is dynamic, with past investment determining current capital availability.

RIAPA tracks changes in incomes and expenditures for different household groups, including changes in food and nonfood consumption patterns. Poverty impacts are measured using survey-based microsimulation analysis. Indi-vidual survey households map to the model’s household groups. Estimated consumption changes in the model are applied proportionally to survey households, and post-simulation consumption values are recalculated and com-pared with a poverty line to determine households’ poverty status.

Source: Benfica and Thurlow (2017).


various sectors together capture all production that occurs in the country, including home production of agricultural products consumed by the household.

Agricultural production shocks in the model affect farmers as well as workers in downstream sectors and consumers purchasing food in local and national markets. The model captures how changes in agricul-tural production lead to changes in farm incomes as well as producer and consumer prices. For example, ENSO’s direct impact may be to reduce teff produc-tion in a region, leading to higher consumer prices for teff. This may encourage households to reduce teff consumption or shift toward cheaper foods, including those produced in regions less affected by ENSO. Farmers in such regions may also respond to higher prices by increasing teff production, although this is constrained by land-planting decisions that may have already been made. Finally, the country may respond to production losses by increasing food imports,

although this requires foreign exchange, which must be either borrowed or earned from exports. A country’s production and trade structure is therefore a key determinant of the overall impact of ENSO events. Table 5 shows the broad structure of the Ethiopian economy.

The model includes household groups that have distinct income and expenditure patterns (Table 6). The population consumes, on average, US$457 of goods and services per person each year (at market exchange rates unadjusted for purchasing power parity). Consumption levels are much lower in rural areas (US$403), especially among the rural poor (US$188). ENSO’s effects on cereals, for example, will have serious implications for poorer rural consumers,

TABLe 6 HOuSeHOLD iNCOMe AND CONSuMPTiON PATTerNS, eTHiOPiA, 2010/11

National Rural Rural poor Urban

Population (millions) 53.4 45.0 9.4 8.4

Consumption per capita (uS$) 456.6 403.0 187.8 742.4

Food consumption share (%) 100 100 100 100

Cereals and roots 28.6 29.5 38.8 25.3

Vegetables 6.8 6.6 7.1 7.8

Fruits 4.8 4.9 5.4 4.5

Meat, fish, and eggs 27.9 26.0 10.3 34.3

Milk and dairy 7.4 7.9 10.4 5.6

Pulses and oilseeds 9.5 9.7 11.6 9.0

Other foods 14.9 15.3 16.4 13.5

Household income share (%) 100 100 100 100

Cropland returns 12.4 16.7 26.7 0.0

Labor remuneration 39.5 42.6 56.1 30.7

Capital profits 46.2 38.6 16.8 67.6

Other sources 2.0 2.1 0.4 1.7


Note: Food consumption excludes meals prepared outside the household. Pro-cessed foods exclude products processed and consumed within the household. Other income sources include social and foreign remittances.

TABLe 5 eTHiOPiA’S NATiONAL eCONOMiC STruCTure, 2010/11

Share of total (%)

GDP Employment Exports Imports

All sectors 100 100 100 100

Agriculture 42.1 79.0 49.0 4.2

Crops 30.7 56.9 44.3 4.2

Livestock 7.9 12.6 4.7 0.0

Forestry and fishing 3.4 9.6 0.0 0.0

Industry 11.5 3.2 7.6 73.1

Mining 1.6 0.4 2.6 0.0

Manufacturing 3.7 1.3 5.0 69.0

Agroprocessing 2.1 0.7 1.7 6.0

Other manufactures 1.6 0.6 3.3 63.0

Other industry 6.3 1.4 0.0 4.0

Services 46.4 17.8 43.4 22.8


Note: Employment is defined as workers in primary jobs. GDP = gross domestic product.


who spend a large share of their income on cereals. Finally, the poorest households, even in rural areas, rely on incomes from cropland and labor—the latter dominated by less-educated workers. Urban house-holds, in contrast, rely more on the profits earned by nonfarm enterprises. Other sources of income include remittances and social transfers, neither of which contributes much to average household incomes.

ENSO’s economywide impactsThe CGE model is used to estimate the economy-wide outcomes associated with the agricultural production losses during El Niño presented in the previous section. The model also simulates various nonfarm policy options that could mitigate damages during ENSO events. Some scenarios reflect existing policies in the country, such as expanding social transfers for poorer households, which is similar in intent to Ethiopia’s Productive Safety Net Programme (PSNP). Other scenarios consider policies that may not exist today, such as major investment in and use of grain storage and distribution systems. The scenarios are scaled, however, to reflect existing conditions. For example, the extent to which social transfers can be used to mitigate welfare losses is informed by the scale and distribution system of current public safety net programs. The scenarios are therefore a combination of expanding current interventions and introducing new policies.

Whereas the crop modeling focused on on-farm interventions, the CGE modeling focuses on the kinds of market and social policy interventions identi-fied in Chapter 2:

• Trade policy (food import subsidies). Introduce a 25 percent price subsidy for imported cereals and processed foods during El Niño years. The aim of an import subsidy is to offset any increases in food prices caused by ENSO’s disruption of domestic production. In the model, demand shifts toward imported foods, and consumers would benefit from lower prices (relative to a situation without the subsidy). Note that for farmers who are hurt by an ENSO shock, the subsequent

increase in market prices offsets some of their losses. Providing an import subsidy limits any price increases, and so can make farm revenue losses larger. Moreover, subsidies have fiscal implications. The fiscal burden of providing import subsidies is “internalized” in the CGE model through lower government revenues and larger deficits, which in turn has economywide implications. Net impacts are reported here.

• Market infrastructure (grain storage). Distribute 1 million metric tons of cereals from public and private stocks. Depleting stocks addresses short-term supply shortfalls during ENSO events and offsets some of the price increases caused by production losses. Like import subsidies, depleting grain stores benefits consumers but may prevent market forces from limiting farm revenue losses via higher prices for agricultural products. The scenario assumes that storage facil-ities already have been or can be expanded to achieve this capacity. Food balance sheets of the Food and Agriculture Organization of the United Nations (FAO) indicate that over the last decade, Ethiopia depleted collective grain stocks by an average 1 million metric tons per year (FAO 2017). This suggests that the grain storage scenario is consistent with the country’s average capacity to achieve. That said, the reliability of the FAO data can be questioned, especially because grain stocks have declined every year between 2009 and 2013. The CGE analysis does not consider the financial cost of restocking public and private grain stores in the years following an ENSO event. Note that the grain storage scenario is equivalent to an alternative scenario in which the govern-ment procures grain in foreign markets (financed by foreign borrowing) and distributes the grain in domestic markets.

• Social protection (cash transfers). Provide short-term cash transfers to poorer households (income quintiles 1–3) equal to US$4.50 per capita. Currently, the government’s various social protec-tion and pension programs transfer, on average,


US$9.60 per person per year. The social transfer scenario roughly doubles this national average for poor households during an El Niño year. Households in the model use these funds to either offset higher food costs or purchase nonfood products, whose prices may also rise during ENSO events as economic shocks spill over from agriculture to nonfarm sectors. The fiscal cost of expanding social transfers is internalized through higher direct taxes (for example, pay-as-you-earn and corporate taxes). The scenario assumes that the distribution of new social transfers occurs through existing social protection systems and does not increase the administrative cost of this system. This is equivalent to assuming that addi-tional administrative costs (not actual transfers) are borne by foreign development partners.

• Combined. Implement all three of the above policy scenarios concurrently.

A feature of economywide models is that they capture the economywide benefits and costs of different policies. First, the model captures both direct on-farm impacts and indirect downstream impacts on food processing, trading, and the rest of the economy. In this way, all sectors in the economy are affected, albeit to varying extents. Second, the model measures the trade-offs associated with each policy scenario. Subsidizing food imports or distributing stored grains is expected to benefit consumers more than farmers. Similarly, allocating more productive resources to agriculture reduces the resources available to nonagriculture sectors. As a result, winners and losers often arise from a given policy change. Third, as mentioned, the model internalizes the fiscal cost of certain policy options. Import subsidies (cash transfers) require either an increase in taxes to offset revenue losses (to pay for higher spending) or an increase in the fiscal deficit. The latter means that the government borrows more from private financial corporations, and this reduces the amount of loanable funds available for private

investment. Simply put, there is no “free lunch” in this class of economic model.

Model results indicate that the scale of the ENSO shock experienced during 2015/16 causes significant economic losses (Figure 16). National GDP falls by 1.6 percent relative to a neutral climate year. Losses are larger in agriculture itself, where GDP falls by 3.6 percent, and particularly within the drought-prone lowlands, where agricultural GDP falls by 11.1 percent. However, it should be noted that even small percentage reductions in national GDP can imply substantial monetary losses. For example, a 1.6 percent drop in national GDP is equal to US$438 million in lost value-added or national income (measured in 2010/11 prices).

CGE models track economic spillovers or linkages between sectors. Figure 16 summarizes the GDP losses occurring in different sectors of the economy during a strong El Niño year. Percentage GDP losses are largest in agriculture (3.6 percent). This reflects this study’s focus on direct impacts to agriculture. Percentage losses fall when considering the broader AFS (3.2 percent) and the national economy (1.6 percent). However, absolute (dollar value) losses increase as the focus broadens beyond agriculture to include the entire AFS. This is due to negative spillovers between sectors. Lower agricultural production, for example, constrains the supply of raw materials to agriculture-related processing and trading. As a result, GDP losses in the AFS are larger (US$463 million) than those in primary agriculture (US$414 million). Overall, the CGE model estimates that more than one-tenth of the damages to the AFS caused by El Niño occur outside of agricul-ture. That said, losses in agriculture and the food system encourage workers to migrate to nonfarm sectors in search of employment and income. This inflow of new workers into the non-AFS parts of the economy offsets some of the production losses occurring within the AFS. The final decline in national GDP caused by El Niño is lower than AFS losses, at US$438 million.


Figure 16 gDP LOSSeS DuriNg THe 2015/16 eL NiÑO eVeNT

Source: Simulation results from computable general equilibrium model using 2010/2011 Ethiopia social accounting matrix data from the International Food Policy Research Institute’s Rural Investment and Policy Analysis Model (Benfica and Thurlow 2017).

Note: AFS = agriculture–food system; DPH = drought-prone highland; DPL-P = drought-prone lowland—pastoralist; MRL = humid, moisture-reliable lowland; MRH-C = moisture-reliable highland—cereal; and MRH-E = moisture-reliable highland—enset (Ethiopian banana).

1.6

3.23.6 3.5

11.1

6.9

1.8

3.5

438463

414

89119

45

9665

0

50

100

150

200

250

300

350

400

450

500

0

0.02

0.04

0.06

0.08

0.1

0.12

All DPH DPL-P MRL MRH-C MRH-E

National AFS Agriculture

Perc

enta

ge

red

ucti

on

GD

P losses in U

S$millio

ns

TABLe 7 gDP CHANgeS DuriNg STrONg eL NiÑO eVeNTS, AND iNTerVeNTiON SCeNAriOS

Without interventions

With interventions

Import subsidies Stored grains Social transfers Combined

Percentage change in GDP (%)

National -1.60 -1.61 -1.67 -1.60 -1.68

Agriculture–food system -3.21 -3.23 -3.11 -3.21 -3.13

Agriculture -3.60 -3.61 -3.43 -3.60 -3.43

Absolute change in GDP (US$ millions)

National -437.6 -438.3 -457.2 -437.6 -457.7

Agriculture–food system -463.4 -465.9 -449.4 -463.2 -451.9

Agriculture -413.8 -414.5 -393.9 -413.7 -394.5

Source: Simulation results from computable general equilibrium model and 2010/2011 Ethiopia social accounting matrix from the International Food Policy Research Institute’s Rural Investment and Policy Analysis Model (Benfica and Thurlow 2017).

Note: GDP = gross domestic product.


The policy scenarios focus on mitigating the adverse effects of El Niño. Table 7 reports GDP losses (in absolute and percentage terms) with and without the effects of the policy interventions. The first column repeats the national GDP losses shown in Figure 16. None of the market or social policy options considered here are effective at limiting GDP losses. Depleting grain stocks offsets the effects of domestic production shortfalls on downstream processing and eases upward pressure on consumer food prices. However, it does not curb the produc-tion losses that occur on the farm during the ENSO event, and by outcompeting local producers, distrib-uting stored grains further reduces agricultural production.

The effects of restocking grain storage are not captured in this analysis, because the simulations focus on the immediate effects of ENSO events. Similarly, cash transfers to the poor increase their demand for domestically produced goods and

services, but these transfers are financed by higher taxes, and hence lower consumption spending by higher-income households. Ultimately, on-farm investments, such as those identified by the crop modeling analysis, are needed to reduce agricul-tural GDP impacts and prevent negative knock-on impacts to the rest of the economy.

Changes in GDP are distributed across house-holds, investors, and the government. Impacts on private household consumption (or welfare) are larger than the impacts on national GDP (Figure 17). This partly reflects this study’s focus on direct damages to agriculture and food prices, which is of greater importance for household spending than for investment or government spending. The figure reports consumption losses for all households and for households in the poorest per capita expendi-ture quintile. All households experience a decline in consumption, or welfare, during a strong El Niño event. Consumption losses are smaller for poorer

Figure 17 HOuSeHOLD CONSuMPTiON LOSSeS DuriNg eL NiÑO BY POLiCY SCeNAriO (PerCeNTAge)

Source: Simulation results from computable general equilibrium model and 2010/2011 Ethiopia social accounting matrix from the International Food Policy Research Insti-tute’s Rural Investment and Policy Analysis Model (Benfica and Thurlow 2017).

2.62 2.48

3.55

2.62

3.42

1.03 0.92

0.40

0.89

0.17

Food importsubsidies

Distribute storedgrains

Cash transfers Combined

Withoutinterventions

With interventions

All households Poorest quintile


households, who tend to be smallholder farmers, because they are more likely to rely on labor wages and earnings from nonfarm self-employment, and less likely to own farmland (Table 6). Note that whereas lower-income households may be less affected by El Niño, their ability to smooth consump-tion, such as by selling assets, is more limited than that of higher-income households. In the absence of supporting evidence, the modeling analysis presented here does not include the short-term benefits and longer-term costs of disposing of household assets.

Policy interventions are effective at reducing household welfare losses. Food import subsidies reduce total consumption losses (that is, from a 2.62 percent decline to a 2.48 percent decline). Figure 17 shows that cash transfers reduce losses for poorer households but leave total consumption outcomes unchanged, at 2.62 percent. This is because the additional cash transfers are financed through

greater taxation of higher-income households, which make up most of the national tax base. In other words, the simulated program is designed to be a transfer from higher- to lower-income households, and this explains why there is little change in national GDP in the transfer scenario. Distributing stored grains is very effective at reducing welfare losses for poorer households, which spend a larger share of their incomes on cereals, but this comes at the cost of larger welfare losses for higher-income house-holds. Finally, when all policy scenarios are imple-mented at the same time, the consumption losses of poor households are largely eliminated, but losses rise for higher-income households (Figure 18).

The impact of severe El Niño events on poor households can be directly measured by changes in the national poverty headcount rate, which shows the share of the population living below the official poverty line. In the base year of the model, the national poverty rate was about 30 percent, meaning

Figure 18 HOuSeHOLD CONSuMPTiON LOSSeS BY QuiNTiLe, WiTH ALL iNTerVeNTiONS COMBiNeD (PerCeNTAge)


1.0

1.51.9

2.3

3.7

0.2

1.2

2.0

2.8

5.6

Q1 Q2 Q3 Q4 Q5

National per capita consumption quintiles

El Niño impacts

Combinedinterventions


that changes in consumption for households in the second-lowest quintile will determine changes in the poverty rate. The CGE model measures these changes using a survey-based microsimulation module that links changes in consumption for house-holds in the CGE model to changes in consumption for a more detailed set of households captured in the survey. Figure 19 shows the change in both the poverty headcount rate and the number of poor people. Without interventions to mitigate impacts, an El Niño event like the one that occurred during 2015/16 causes the national poverty rate to increase by 1.2 percentage points. This is equivalent to an additional 656,200 people living below the poverty line during the event period. Policy interventions can

be effective in helping avoid some of the increase in the incidence of poverty, although they are more effective in reducing the poverty gap (that is, the numbers of the poorest).

RECOMMENDATIONSIn response to the interest of Ethiopia’s government and development partners, this study reviewed ongoing resilience programming, consulted stake-holders on opportunities for strengthening such programming, and provided quantitative modelling of a series of alternative resilience strategies.

The analysis of historical weather data indicated that El Niño causes drier conditions in most of the country, whereas La Niña causes wetter conditions.

Figure 19 CHANgeS iN NATiONAL POVerTY rATe AND POOr POPuLATiON DuriNg eL NiÑO BY POLiCY SCeNAriO


1.21.1

1.4

1.21.1

656.2605.8

769.5

631.1596.3

Food importsubsidies

Distribute storedgrains

Cash transfers Combined

Withoutinterventions

With interventions

Number of poor people Poverty rate (Percentage)


High-resolution crop modeling analysis showed that for maize and wheat, yield deviations caused by El Niño vary substantially across crops and subnational regions. In general, cereal yields in drought-prone areas and the lowlands fall during El Niño but rise during La Niña.

The former was confirmed by in-depth statistical analysis of historical production data, which showed that additional production losses for key staple crops were caused by reduced harvested land area, rather than just lower crop yields. The analysis also showed that substantial crop and livestock production losses occurred in the lowlands and drought-prone regions.

Based on the estimated deviations in agricultural production, the economywide assessment showed that El Niño causes major losses in gross domestic product (GDP), and that a significant share of the agriculture–food system (AFS) GDP losses occur outside of agriculture itself (via intersectoral and demand linkages). Welfare losses are found to be larger for net food-consuming urban households than for rural smallholder households. A large increase in the number of poor people also occurs following ENSO events.

Based on these analyses, we propose the following recommendations:

Build on the strength of Ethiopia’s current resilience programmingA wealth of experience in past and current work on resilience in Ethiopia clearly exists. Efforts to improve resilience in the agriculture sector should build upon this expertise in the context of the evolving Ethiopian conditions. Although some of these efforts are recorded in the peer-reviewed and gray literature, forming a clear picture of where activities have been concentrated is of enormous importance. As several program reviews and key informants pointed out, developing a clear under-standing of where and how interventions relate one to another in terms of thematic and geographic areas is essential to achieving effective coordination among the various agencies and pursuing potential synergies. Strengthening regional cooperation

between national and international development partners rests on the availability of this information.

Develop new monitoring tools for resilience and improve coordinationThe coordination should start with the government of Ethiopia and partners developing and applying a resilience framework according to which its many development programs can be structured and monitored for progress and outcomes. This can be supplemented by a database with an up-to-date record of the programs and organizations that have worked on and are working on resilience-building projects and interventions. With this information agencies will be able to identify common avenues for future collaboration and potential gaps in investments.

Furthermore, it is essential to strengthen existing coordination mechanisms, such as the Disaster Risk Management Technical Working Group, facili-tated by the National Disaster Risk Management Commission; the intercluster coordination meeting, facilitated by UNOCHA (the United Nations Office for the Coordination of Humanitarian Affairs); and the national and regional experience sharing forums on coordination for resilience, facilitated by the European Union and the Food and Agriculture Organization of the United Nations. Platforms such as these provide a place to share resources and expertise through mutual engagement, as well as to support humanitarian organizations, development partners, and government agencies to strengthen their respective resilience-building agendas. Bridging diverse institutional interests, mandates, policies, and resources is a challenging but necessary task, and coordination efforts will reduce the risk of duplication in programs and strategies. Integrating selected resilience-related priorities, such as reproductive health and rights, and gender equality, into the various sectors would provide a large step toward strengthening resilience in the long term.

The early warning system and assessment results are not always consistent with regional beneficiary


numbers. The system’s performance needs to be strengthened in lowland areas, where shocks are not always identified fast enough. It is also important to develop long-term climate early warning systems with the participation of local communities to limit the unfavorable impacts of ENSO on livestock popu-lations and productivity.

Impact assessments, monitoring, and evaluation need to be strengthened in programs beyond the Productive Safety Net Programme (PSNP) to ensure that the intended benefits, impacts, and outcomes are achieved and that programs that do not perform as well can be corrected in time.

Improve targeting and linkages between short- and long-term programming The accumulated experience indicates that programs and policies need to be customized and targeted toward communities and regions based on their geographic location, needs, and circumstances. Coordination and information sharing across agencies need to be improved to integrate efforts to improve resilience and move away from one-size-fits-all types of policies and interventions. Agencies need to jointly determine where and how resources need to be allocated. Beyond immediate responses to shocks, funds should also be used for recovery and long-term rehabilitation.

At the same time, it is important that food security initiatives be linked to poverty reduction, shorter-term disaster response, and longer-term disaster risk reduction. Examples of improved targeting include tailoring of the Targeted Supplementary Feeding Programme allotments to the nutrient requirements of different groups. Potential recipients who might adopt negative coping strategies that deplete their natural resource and asset base should be particu-larly identified and targeted.

Because lowland pastoralist areas are particu-larly vulnerable, a strategy specifically tailored to their livelihoods should be developed. The strategy should focus on the timing of assessments, types of livelihood systems in the area, public works projects appropriate for lowland areas, and management of

food resources in the lowlands context, focusing on climatic extremes and targeting of communities. The focus should be on interventions targeted toward building livestock infrastructure to improve the resilience of these communities. Dividing the respon-sibility for livestock between two ministries might not be ideal for coordination in this area.

Substantial investments are still needed for access to basic health and education services in the lowlands. Public works should be integrated into comprehensive regional plans, and the PSNP component tailored more appropriately to the pastoral and lowlands context, in terms of the timing and types of public works and who participates in the program, to reduce inclusion and exclusion errors.

Humanitarian organizations should systematically assess and analyze the impacts of crises on liveli-hoods as a basis for designing, implementing, and monitoring livelihood interventions. Supporting live-lihoods requires an in-depth understanding of and support for affected populations’ assets, capabilities, and activities to ensure their means of living. Special attention is needed during mass internal displace-ments, such that pregnant women have access to safe and clean deliveries attended by skilled birth attendants, and all children, including separated and unaccompanied children, and women are protected.

Strengthen the enabling environment Ethiopia has made substantial progress in building the resilience of its agriculture–food system. It has successfully raised the productivity of cereal producers and provided social protection to poor rural households. The government has also invested heavily in rural roads and other infrastructure needed to ensure that markets operate to smooth production and consumption shocks. Despite this progress, smallholder productivity remains low and scope remains to further expand modern input use and water management practices.

The development of microfinance institutions and village savings and loan associations should be enhanced to allow communities to diversify


their livelihoods or invest additional resources in current productive activities to increase efficiency, productivity, and profitability while improving resilience. Female-headed households should be prioritized as targets for financial and cash support.

Develop multipronged investment strategiesNo single intervention can fully eliminate all GDP and welfare losses in all places and for all people in Ethiopia. Instead, a portfolio of farm, market, and social policies will be needed to cushion ENSO’s economywide impacts. In terms of investments in agricultural technologies, the analysis finds that an increase in irrigation combined with integrated soil fertility management and increased application of nitrogen fertilizer can dramatically reduce adverse ENSO impacts in all agroecological zones and during both ENSO events (El Niño and La Niña).

In terms of complementarity investments, although food import subsidies reduce total consumption losses to some extent, cash transfers reduce losses for poorer households but tax higher-income households. Distributing stored grains is very effective for reducing welfare losses for poorer households, which spend a larger share of their incomes on cereals, but this comes at the cost of larger welfare losses for higher-income

households. A combination of these three policies largely eliminates the consumption losses of poor households but makes losses rise for higher-income households.

In the medium to long term, enhancing rural resilience and achieving middle-income status by 2025 will require a permanent cash transfer program, similar to the PSNP. Rural resilience can be achieved only if infrastructure development accelerates and if investments in agricultural research and development are increased. Even today only one-half of all Ethiopian farmers use chemical fertilizers, and recent growth in irrigated area, though commendable, remains too low to dramatically affect resilience outcomes in the next two decades.

Achieving climate resilience in rural areas will also require much faster generation of economic livelihood options outside of agriculture, which in turn will necessitate a much faster opening to and support of foreign direct investment. Despite an expected gradual decline in the share of people employed in the agriculture sector, an expected increase of 60 percent in the number of rural households in arid and semi-arid land areas suggests that rural investments need to substantially increase to ensure that rural–urban income gaps and income and food poverty do not increase.


APPENDIX 1: MAIN RESILIENCE PROGRAMS IMPLEMENTED IN ETHIOPIA AND KEY FEATURES

Note: SNNPR = Southern Nations, Nationalities, and Peoples’ Region.

Program / institution ModalityImplementing organization Donor

Community or household focus

Target population

Targets pastoralist or agropastoralists exclusively?

Targets mainly women? Regions

Public Safety Net Programme (PSNP)

Multidonor/flagship program

Ministry of Agriculture

European Union, World Bank, and others

Household Food-insecure households

No No All regions except Gambella and Benishangul-Gumuz (349 woredas)

Sustainable Land Management Program (SLMP2)


Agricultural office at woreda level

Kreditanstalt für Wieder-aufbau and World Bank

Community Communities with degraded land

No No Many regions: Oromia, Amhara, Tigray, SNNPR, Gambela, and Benishangul-Gumuz

Managing environmental resources to enable Transitions (MereT)

Food-for-work program

Bureau of Agriculture and Rural Development

World Food Programme

Community Communities with degraded land

No Yes Tigray and 5 other regions (451 communities)

Drought resilience and Sustainable Livelihood Program ii (DrSLP ii)


Pastoral Agriculture Development Bureau

African Development Bank

Community and household

House-holds not benefiting from other interventions

Yes No Afar, Somali, Oromia, and SNNPR

Livelihood for resilience (grAD-ii)

Programs that support PSNP

Catholic Relief Services and CARE

US Agency for International Development

Household PSNP beneficiaries, youth

No Yes Oromia,

Tigray

Building resilience and Adaptation to Climate Change and Disasters (BrACeD)

Programs that support PSNP

Hundee Oromo Grassroots Development Initiative

Christian Aid, UK Depart-ment for International Development


PSNP beneficiaries

Yes No Oromia

Pastoral Community Development Project (PCDP)


Ministry of Federal and Pastoral Affairs / Afar Regional Cooperation Promotion Office

World Bank Community Pastoralist communities

Yes Yes (priority is given to women)

Afar, Oromia, Somali, and SNNPR

resilience Building and Creation of economic Opportunities (reSeT)

Cordaid

Danish Church Aid

European Union


Most vulner-able com-munities

No No Oromia, Amhara, and Somalia

regional Pastoral Livelihoods resilience Project (rPLrP)

Ministry of Livestock and Fisheries

World Bank and African Development Bank


Pastoralist communities

Yes NoAfar, Somali, Oromia, and SNNPR


APPENDIX 2: ACTIVITIES PROMOTED BY DIFFERENT RESILIENCE PROGRAMS AND ORGANIZATIONS

Activity Organization /Program

Agriculture

Distribution of seeds (fast-maturing varieties, climate shock–resilient varieties) suitable for a specific climate and environment; biofuels

Oromia Agricultural Bureau; CARE; Environment, Forest and Climate Change Authority; farmer-managed natural regeneration (FMNR); CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS); Agency of Mines and Energy; Feed the Future Liveli-hoods for Resilience Activity (GRAD-II)

Crop insurance system Relief Society of Tigray (REST); GRAD-II

Capacity building on climate-smart agriculture to enhance productivityEnvironment, Forest and Climate Change Authority; FMNR; CARE; Livelihoods for Resilience—Oromia (LRO)

Livestock

Market link creation, livestock market access, road construction, value chains

Environment, Forest and Climate Change Authority; Drought Resilience and Sustainable Livelihoods Program (DRSLP); CARE; LRO; Pastoral Community Development Project (PCDP); Regional Pastoral Livelihoods Resilience Project (RPLRP)

Asset-building activities through provision of goats and heifersBuilding Resilience and Adaptation to Climate Change and Disasters (BRACED); Ethiopian Orthodox Church Development and Inter-church Aid Commission (EOC-DICAC)

Animal health (vaccinations, awareness building) DRSLP; Save the Children

Fodder development, animal fattening, goat rearing Resilience-building in Ethiopia (RESET); FMNR; CCAFS; World Vision; RPLRP; DRSLP

Livestock insurance service BRACED / Market Approaches to Resilience (MAR)

Water management

Construction of water-harvesting infrastructure such as ponds and cisterns, pond rehabilitation, groundwater extraction

Oromia Agricultural Bureau; DRSLP; RESET; BRACED; CCAFS; PCDP; RPLRP; Productive Safety Net Programme (PSNP)

Irrigation (irrigation diversion, check dams and channels)Oromia Agricultural Bureau; Water, Mineral and Energy Bureau; DRSLP; World Vision; Catholic Church social and development program; EOC-DICAC

Water supply, sanitation, and hygiene (WASH)RESET; CARE; LRO; World Vision; Catholic Church social and development program; Save the Children; Water Resources Office; Building Resilience and Creating Economic Opportunities

Finance

Financial services: Improving access to credit and savings, microcredit associations

Oromia Agricultural Bureau; CARE; RESET; CCAFS; BRACED/MAR; LRO; REST; GRAD-II; PCDP; EOC-DICAC

Capacity building for financial literacy REST; GRAD-II

Financial support: Cash transfers and revolving funds Environment, Forest and Climate Change Authority; EOC-DICAC

Women’s socioeconomic empowerment through self-help groups, savings groups, and training (bookkeeping, disaster risk management)

BRACED; RESET; CCAFS

Livelihood diversification

Livelihood diversification, livelihood support (for example, beekeeping, gum and incense collection, brick production, petty trading, horticulture production, sewing and embroidery, dairy production), creation of coop-eratives, capacity building on income-generating activities

Environment, Forest and Climate Change Authority and microenterprises; DRSLP; RESET; FMNR; BRACED; CARE; LRO; CCAFS; BRACED/MAR; REST; World Vision; SNV (Netherlands Development Organisation); GRAD-II

Youth employment creation to minimize migration RESET; LRO; REST; SNV; GRAD-II; EOC-DICAC

APPENDIX 2: ACTIVITIES PROMOTED BY DIFFERENT RESILIENCE PROGRAMS AND ORGANIZATIONS 67

Activity Organization /Program

environmental conservation

Rehabilitation of degraded lands, reforestation, bamboo plantation, terracing, soil erosion reduction, rangeland management, pasture productivity enhancement, selective bush clearing, rangeland enclosure or paddock system

Environment, Forest and Climate Change Authority; FMNR; BRACED; CCAFS; BRACED/MAR; LRO; DRSLP; RESET; Catholic Church social and development program; Save the Children; Sustainable Land Management Program; RPLRP; PSNP

Solar appliances, improved cookstoves, biodigesters (biogas)Environment, Forest and Climate Change Authority; Water, Mineral and Energy Bureau; CARE; BRACED/MAR; Agency of Mines and Energy; GIZ (German government development agency); SNV

Conflict resolution around natural resource use and so on DRSLP; RPLRP

early warning systems

Disaster risk reductionCARE; Save the Children; RPLRP; Building Resilience and Creating Economic Opportunities; BRACED; EOC-DICAC; RPLRP

Early warning systems, climate information dissemination, climate services activities

BRACED; CARE; EOC-DICAC; PCPD

Others

Food (grain, oil, and pulses) and nonfood relief aid Oromia Disaster Risk Management Commission; PSNP

Health and nutritionRESET; CARE; World Vision; Save the Children; Building Resilience and Creating Economic Opportunities

Indigenous Dabarre support system Hundee Oromo Grassroots Development Initiative

Capacity building for leadership and socioeconomic empowerment, institutional capacity building

CARE; LRO; GRAD-II

Urban resilience component (development of city plan) BRACED/MAR

Child sponsorship World Vision

Infrastructure development: Electricity, roads, schools PCDP; PSNP

Carbon trading FMNR

Education Building Resilience and Creating Economic Opportunities


REFERENCESAbate, T., A. Ebro, and L. Nigatu. 2010. “Traditional

Rangeland Resource Utilisation Practices and Pastoralists’ Perceptions on Land Degradation in South-East Ethiopia.” Tropical Grasslands 44 (3): 202–212.

Abebe, A., L. O. Eik, O. Holand, T. Ådnøy, and A. Tolera. 2012. “Pastoralists’ Perceptions of Feed Scarcity and Livestock Poisoning in Southern Rangelands, Ethiopia.” Tropical Animal Health and Production 44 (1): 149–157.

AKLDP (Agriculture Knowledge, Learning, Documentation and Policy) Project. 2016. El Niño in Ethiopia: Early Impacts of Drought in Amhara National Regional State. Field notes. Washington, DC: Feed the Future, United States Agency for International Development.

Alemu, A., D. Korecha, and M. Mohamod. 2017. “Impacts of Various ENSO Phases on Cereal Crop Productivity in the Upper Awash Basin, Central High Land of Ethiopia.” In Proceedings of the International Conference on Impact of El Niño on Biodiversity, Agriculture, and Food Security, 23–24 February 2017, Harayama University, Ethiopia, edited by S. Workeneh, T. Toru, F. Mekbib, N. Dechassa, K. Kibret, F. Beyene, M. Ketema, M. Urge, A. Debella, and K. Sebesibe, 3–18. Dire Dawa, Ethiopia: Haramaya University, Research and Extension Office.

Angassa, A., and G. Oba. 2007. “Relating Long-Term Rainfall Variability to Cattle Population Dynamics in Communal Rangelands and a Government Ranch in Southern Ethiopia.” Agricultural Systems 94 (3): 715–725.

Araya, N. S. 2011. “Weather Insurance for Farmers: Experience from Ethiopia.” Paper presented at “New Directions for Smallholder Agriculture,” January 24–25, Rome. Cited in Platform for Agricultural Risk Assessment. 2016. Managing Risks to Improve Farmers’ Livelihoods: Ethiopia Agricultural Risk Assessment Study. Greenwich, UK.

Araya, A., G. Hoogenboom, E. Luedeling, K. M. Hadgu, I. Kisekka, and L. G. Martorano. 2015. “Assessment of Maize Growth and Yield Using Crop Models under Present and Future Climate in Southwestern Ethiopia.” Agricultural and Forest Meteorology 214:252–265.

Arndt, C., W. Farmer, K. Strzepek, and J. Thurlow. 2012. “Climate Change, Agriculture and Food Security in Tanzania.” Review of Development Economics 16 (3): 378–393.

Arndt, C., A. Schlosser, K. Strzepek, and J. Thurlow. 2014. “Climate Change and Economic Growth Prospects for Malawi: An Uncertainty Approach.” Journal of African Economies 23 (S2): 83–107.

Arndt, C., and J. Thurlow. 2015. “Climate Uncertainty and Economic Development: Evaluating the Case of Mozambique to 2050.” Climatic Change 130 (1): 63–75.

Attia, B. B., and A. B. Abulhoda. 1992. “The ENSO Phenomenon and Its Impacts on the Nile’s Hydrology.” Paper presented at International Seminar on Climatic Fluctuations and Water Resources, Cairo, December 11–14.

aWhere. 2018. Weather Terrain database. Accessed November 2, 2018. https://developer.awhere.com/api/about-our-data/weather-data.

REFERENCES 69

https://developer.awhere.com/api/about-our-data/weather-data

https://developer.awhere.com/api/about-our-data/weather-data

Bachewe, F., G. Berhane, B. Minten, and A. S. Taffesse. 2015. Agricultural Growth in Ethiopia (2004–2014): Evidence and Drivers. Ethiopia Strategy Support Program Working Paper 81. Washington, DC: IFPRI.

Benfica, R., and J. Thurlow. 2017. “Identifying Priority Value Chains in Ethiopia.” Working Paper 110, Ethiopia Strategy Support Program. (Washington, DC, IFPRI).

Block, P., K. Strzepek, M. W. Rosegrant, and X. Diao. 2008. “Impacts of Considering Climate Variability on Investment Decisions in Ethiopia.” Agricultural Economics 39 (2): 171–181.

Bryan, E., S. Theis, J. Choufani, A. De Pinto, R. S. Meinzen-Dick, and C. Ringler. 2017a. Conceptual Framework: Gender, Climate Change, and Nutrition Integration Initiative. GCAN Policy Note 7. Washington, DC: IFPRI.

Bryan, E., S. Theis, J. Choufani, A. De Pinto, R. S. Meinzen-Dick, and C. Ringler. 2017b. “Gender-Sensitive, Climate-Smart Agriculture for Improved Nutrition in Africa South of the Sahara.” In A Thriving Agricultural Sector in a Changing Climate: Meeting Malabo Declaration Goals through Climate-Smart Agriculture, edited by A. De Pinto and J. M. Ulimwengu, 114–135. Washington, DC: IFPRI.

Calzadilla, A., T. Zhu, K. Rehdanza, R. S. J. Tol, and C. Ringler. 2013. “Economywide Impacts of Climate Change on Agriculture in Sub-Saharan Africa.” Ecological Economics 93:150–165.

Cervigni, R., and M. Morris, eds. 2016. Confronting Drought in Africa’s Drylands: Opportunities for Enhancing Resilience. Washington, DC: World Bank.

CRU, UEA (Climatic Research Unit, University of East Anglia). 2019. CRU-TS v4.1 dataset. https://crudata.uea.ac.uk/cru/data/hrg/.

Conway, D., and E. L. F. Schipper. 2011. “Adaptation to Climate Change in Africa: Challenges and Opportunities Identified from Ethiopia.” Global Environmental Change 21 (1): 227–237.

Davison, W. 2015. “Ethiopia Faces Worst Drought in 50 Years, Millions Affected.” Bloomberg News, December 7.

Degefu, W. 1987. “Some Aspects of Meteorological Drought in Ethiopia.” In Drought and Hunger in Africa: Denying Famine a Future, edited by M. H. Glantz, 23–36. Cambridge, UK: Cambridge University Press.

Deininger, K., D. A. Ali, and T. Alemu. 2011. “Impacts of Land Certification on Tenure Security, Investment and Land Market Participation: Evidence from Ethiopia.” Land Economics 87:312–334,

Dercon, S., R. V. Hill, D. Clarke, I. Outes-Leon, and A. S. Taffesse. 2014. “Offering Rainfall Insurance to Informal Insurance Groups: Evidence from a Field Experiment in Ethiopia.” Journal of Development Economics 106:132–143.

Desta, S., and D. L. Coppock. 2002. “Cattle Population Dynamics in the Southern Ethiopian Rangelands, 1980–97.” Journal of Range Management 55 (5): 439–451.

Eltahir, E. A. B. 1996. “El Niño and the Natural Variability in the Flow of the Nile River.” Water Resources Research 32 (1): 131–137.


http://ebrary.ifpri.org/cdm/ref/collection/p15738coll2/id/129782


http://gcan.ifpri.info/files/2017/01/GCAN-Framework-Note-7_web-1.pdf

http://gcan.ifpri.info/files/2017/01/GCAN-Framework-Note-7_web-1.pdf




https://crudata.uea.ac.uk/cru/data/hrg/

https://crudata.uea.ac.uk/cru/data/hrg/

Ethiopia, CSA (Central Statistical Agency, Ministry of Finance and Economic Development). Various years. Agricultural Sample Survey. Volume I: Report on Area and Production for Major Crops (Private Peasant Holdings, Meher Season). Addis Ababa.

—. 2013. Ethiopia—Household Consumption Expenditure Survey 2010–2011. Addis Ababa.

Ethiopia, Environmental Protection Authority. 2012. United Nations Conference on Sustainable Development (Rio+20): National Report of Ethiopia. Addis Ababa.

Ethiopia, NMA (National Meteorological Agency, Ministry of Water Resources). 2007. Climate Change National Adaptation Programme of Action (NAPA) of Ethiopia. Addis Ababa.

FAO (Food and Agriculture Organization of the United Nations). 2015. “The Global Administrative Unit Layers (GAUL).” http://www.fao.org/geonetwork/srv/en/graphover.show?id=12691&fname=GAUL.gif&access=public.

—. 2017. FAOSTAT: Food Balance Sheets. http://www.fao.org/faostat/en/#data/FBS.

Feyssa, D., and D. Gemeda. 2015. “Impact of Climate Change on Production Systems in Semiarid Regions of Ethiopia.” Journal of Biological and Chemical Research 32 (2): 755–764.

Funk, C., J. Rowland, G. Eilerts, E. Kebebe, N. Biru, L. White, and G. Galu. 2012. A Climate Trend Analysis of Ethiopia. USGS Fact Sheet 3053. Reston, VA: United States Geological Survey.

Gissila, T., E. Black, D. I. F. Grimes, and J. M. Slingo. 2004. “Seasonal Forecasting of the Ethiopian Summer Rains.” International Journal of Climatology 24:1345–1358.

Global Yield Gap and Water Productivity Atlas. 2019. Ethiopia. http://www.yieldgap.org/ethiopia.

Haile, T. 1988. “Causes and Characteristics of Drought in Ethiopia.” Ethiopian Journal of Agricultural Science 10:85–97.

Hassen, A., T. Zewdu, A. Tolera, and D. Korecha. 2017. “Interconnection between El-Niño-Southern Oscillation Induced Rainfall Variability, Livestock Population Dynamics and Pastoralists Adaptation Strategies in Eastern Ethiopia.” In Proceedings of the International Conference on Impact of El Niño on Biodiversity, Agriculture, and Food Security, 23–24 February, Harayama University, Ethiopia, edited by S. Workeneh, T. Toru, F. Mekbib, N. Dechassa, K. Kibret, F. Beyene, M. Ketema, M. Urge, A. Debella, and K. Sebesibe. Dire Dawa, Ethiopia: Research and Extension Office, Haramaya University.

Hertel, T. W., M. B. Burke, and D. B. Lobell. 2010. “The Poverty Implications of Climate-Induced Crop Yield Changes by 2030.” Global Environmental Change 20:577–585.

Hoogenboom, G., C. H. Porter, V. Shelia, K. J. Boote, U. Singh, J. W. White, L. A. Hunt, R. Ogoshi, J. I. Lizaso, J. Koo, S. Asseng, A. Singels, L. P. Moreno, and J. W. Jones. 2017. Decision Support System for Agrotechnology Transfer (DSSAT). Version 4.7. Gainesville, FL, US: DSSAT Foundation. https://DSSAT.net.

IFPRI (International Food Policy Research Institute). 2017. 2017 Global Food Policy Report. Washington, DC.

REFERENCES 71

http://www.fao.org/geonetwork/srv/en/graphover.show?id=12691&fname=GAUL.gif&access=public



http://www.fao.org/faostat/en/#data/FBS

http://www.fao.org/faostat/en/#data/FBS

http://www.yieldgap.org/ethiopia

https://DSSAT.net


Jones, J. W., G. Hoogenboom, C. H. Porter, K. J. Boote, W. D. Batchelor, L. A. Hunt, P. W. Wilkens, U. Singh, A. J. Gijsman, and J. T. Ritchie. 2003. “The DSSAT Cropping System Model.” European Journal of Agronomy 18:235–265.

Jones, P. G., and P. K. Thornton. 2003. “The Potential Impacts of Climate Change on Maize Production in Africa and Latin America in 2055.” Global Environmental Change 13 (1): 51–59.

Kandji, S. T., L. Verchot, and J. Mackensen. 2006. Climate Change and Variability in Southern Africa: Impacts and Adaptation Strategies in the Agricultural Sector. Nairobi: Word Agroforestry Centre and United Nations Environment Programme.

Kassie, B. T., S. Asseng, R. P. Rotter, H. Hengsdijk, A. C. Ruane, and M. K. Van Ittersum. 2015. “Exploring Climate Change Impacts and Adaptation Options for Maize Production in the Central Rift Valley of Ethiopia Using Different Climate Change Scenarios and Crop Models.” Climatic Change 129 (1–2): 145–158.

Kassie, B. T., M. K. Van Ittersum, H. Hengsdijk, S. Asseng, J. Wolf, and R. P. Rötter. 2014. “Climate-Induced Yield Variability and Yield Gaps of Maize (Zea mays L.) in the Central Rift Valley of Ethiopia.” Field Crops Research, 160:41–53.

Knippenberg, E., and J. Hoddinott. 2017. Shocks, Social Protection, and Resilience: Evidence from Ethiopia. Ethiopia Strategy Support Program Working Paper 109. Addis Ababa: IFPRI.

Korecha, D., and A. Sorteberg. 2013. “Validation of Operational Seasonal Rainfall Forecast in Ethiopia.” Water Resources Research 49 (11): 7681–7697.

Lobell, D. B., M. B. Burke, C. Tebaldi, M. D. Mastrandrea, W. P. Falcon, and R. L. Naylor. 2008. “Prioritizing Climate Change Adaptation Needs for Food Security in 2030.” Science 319 (5863): 607–610.

Lobell, D. B., W. Schlenker, and J. Costa-Roberts. 2011. “Climate Trends and Global Crop Production Since 1980.” Science 333 (6042): 616–620.

MacCarthy, D. S., S. G. Adiku, B. S. Freduah, and F. Gbefo. 2017. “Using CERES-Maize and ENSO as Decision Support Tools to Evaluate Climate-Sensitive Farm Management Practices for Maize Production in the Northern Regions of Ghana.” Frontiers in Plant Science 8:31.

Megersa, B., A. Markemann, A. Angassa, and A. Valle Zarate. 2014. “The Role of Livestock Diversification in Ensuring Household Food Security under a Changing Climate in Borana, Ethiopia.” Food Science 6 (1): 15–28.

Mohammed, A., T. Tana, P. Singh, A. Molla, and A. Seid. 2017. “Identifying Best Crop Management Practices for Chickpea (Cicer arietinum L.) in Northeastern Ethiopia under Climate Change Condition.” Agricultural Water Management 194:68–77.

NASA (US National Aeronautics and Space Administration). 2019. AgMIP Climate Forcing Datasets. Accessed December 16, 2016. https://data.giss.nasa.gov/impacts/agmipcf/agmerra/.

Nelson, G. C., M. W. Rosegrant, J. Koo, R. D. Robertson, T. Sulser, T. Zhu, C. Ringler, S. Msangi, A. Palazzo, M. Batka, M. Magalhaes, R. Valmonte-Santos, M. Ewing, and D. R. Lee. 2009. Climate Change: Impact on Agriculture and Costs of Adaptation. Washington, DC: IFPRI.


https://data.giss.nasa.gov/impacts/agmipcf/agmerra/

https://data.giss.nasa.gov/impacts/agmipcf/agmerra/

Nicholls, N. 1994. “What Are the Potential Contributions of El Niño–Southern Oscillation Research to Early Warning of Potential Acute Food-Deficit Situations?” In Usable Science: Food Security, Early Warning, and El Niño: Proceedings of the Workshop on ENSO/FEWS, Budapest, Hungary, October 25–28, 1993, edited by M. H. Glantz, 169–178. Boulder, CO, US: National Center for Atmospheric Research.

OPM (Oxford Policy Management) and HESPI (Horn Economic and Social Policy Institute). 2018. Generation El Niño: Long-Term Impacts on Children’s Well-Being. Policy Brief. Addis Ababa: UNICEF.

Parry, M. L., C. Rosenzweig, A. Iglesias, M. Livermore, and G. Fischer (2004). “Effects of Climate Change on Global Food Production under SRES Emissions and Socio-economic Scenarios.” Global Environmental Change 14 (1): 53–67.

Pauw, K., J. Thurlow, M. Bachu, and D. E. van Seventer. 2011. “The Economic Costs of Extreme Weather Events: A Hydro-meteorological CGE Analysis for Malawi.” Environment and Development Economics 16 (2): 177–198.

Quisumbing, A. R., and N. Kumar. 2014. Land Rights Knowledge and Conservation in Rural Ethiopia: Mind the Gender Gap. IFPRI Discussion Paper 1386. Washington, DC: IFPRI.

Sarkar, R., B. V. Ortiz, and K. S. Balkcom. 2017. “Strategic Adaptation of Nitrogen Management for El Niño Southern Oscillation–Induced Winter Wheat System.” Mitigation and Adaptation Strategies for Global Change 22 (3): 369–398.

Schmidt, E., and T. Thomas. 2017. “Discovering the Economic and Climatic Boundaries for Cropland Expansion in Ethiopia.” PowerPoint presentation at “The Future of Ethiopia’s Agriculture: Towards a Resilient System to End Hunger and Undernutrition,” Addis Ababa, December 15.

Seaward, C. 2016. El Niño in Ethiopia. Oxfam El Niño Briefings. Oxford, UK: Oxfam GB.

Singh, R., M. Worku, S. Bogale, A. Cullis, A. Adem, B. Irwin, S. Lim, L. Bosi, and C. Cabot Venton. 2016. Reality of Resilience: Perspectives of the 2015–16 Drought in Ethiopia. Resilience Intel No. 6. London: BRACED (Building Resilience and Adaptation to Climate Extremes and Disasters), Overseas Development Institute. http://www.braced.org/contentAsset/raw-data/18256c98-2a10-4586-9317-17a68b45c1a7/attachmentFile.

Tache, B., and E. Sjaastad. 2010. “Pastoralists’ Conceptions of Poverty: An Analysis of Traditional and Conventional Indicators from Borana, Ethiopia.” World Development 38 (8): 1168–1178.

Taffesse, A. S., P. Dorosh, and S. A. Gemessa. 2012. “Crop Production in Ethiopia: Regional Patterns and Trends.” In Food and Agriculture in Ethiopia: Progress and Policy Challenges, edited by P. Dorosh and S. Rashid, 53–83. Philadelphia: University of Pennsylvania Press.

Tall, A., et al. 2014. Scaling Up Climate Services for Farmers: Mission Possible. Learning from Good Practice in Africa and South Asia. CCAFS Report No. 13. Copenhagen: CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS).

REFERENCES 73




https://www.slideshare.net/essp2/discovering-the-economic-and-climatic-boundaries-for-cropland-expansion-in-ethiopia-84347262



http://www.braced.org/contentAsset/raw-data/18256c98-2a10-4586-9317-17a68b45c1a7/attachmentFile






Thornton, P. K., P. G. Jones, G. Alagarswamy, J. Andresen, and M. Herrero. 2010. “Adapting to Climate Change: Agricultural System and Household Impacts in East Africa.” Agricultural Systems 103 (2): 73–82.

Thurlow, J., X. Diao, and T. Zhu. 2012. “Current Climate Variability and Future Climate Change: Estimated Growth and Poverty Impacts for Zambia.” Review of Development Economics 16 (3): 394–411.

Tilahun, M., A. Angassa, A. Abebe, and A. Mengistu. 2016. “Perception and Attitude of Pastoralists on the Use and Conservation of Rangeland Resources in Afar Region, Ethiopia.” Ecological Processes 5:18.

Tol, R. S. J. 2009. “The Economic Effects of Climate Change.” Journal of Economic Perspectives 23:29–51.

Twyman, J., et al. 2014. “Gender and Climate Change Perceptions, Adaptation Strategies, and Information Needs Preliminary Results from Four Sites in Africa.” CCAFS Working Paper no. 83. CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Copenhagen, Denmark.

UNDESA (United Nations Department of Economic and Social Affairs). 2015. World Population Prospects: 2015 Revision. New York.

UNDP (United Nations Human Development Programme). 2015. Human Development Report 2015: Work for Human Development. New York.

UNISDR (United Nations Office for Disaster Risk Reduction). 2007. “Terminology on Disaster Risk Reduction.” Accessed on February 20, 2017. https://www.unisdr.org/we/inform/terminology.

UNICEF (United Nations Children’s Fund). 2016. “Ethiopia El Niño Emergency Fast Facts—June 2016.” https://reliefweb.int/report/ethiopia/ethiopia-el-ni-o-emergency-fast-facts-june-2016.

UNOCHA (United Nations Office for the Coordination of Humanitarian Affairs) Ethiopia. 2016. Humanitarian Requirements Document. New York.

USAID (US Agency for International Development). 2016. El Niño in Ethiopia, 2015–2016: A Real-Time Review of Impacts and Responses. Washington, DC: USAID/Ethiopia Agriculture Knowledge, Learning, Documentation and Policy Project.

Viste, E., D. Korecha, and A. Sorteberg. 2013. “Recent Drought and Precipitation Tendencies in Ethiopia.” Theoretical and Applied Climatology 112 (3–4): 535–551.

Walker, B. 2016. Seasonal Weather Assessment for Ethiopia during March–July 2016. London: Government of the United Kingdom.

Webb, P., and J. von Braun. 1994. Famine and Food Security in Ethiopia: Lessons for Africa. Chichester, UK: John Wiley & Sons.

Wolde-Georgis, T. 1997. “The Use of El Niño Information as Drought Early Warning in Ethiopia.” In “Using Science against Famine: Food Security, Famine Early Warning and El Niño,” ed. M. H. Glantz, special issue, Internet Journal for African Studies 2:114–121.

World Bank. 2017. World Bank Open Data. Accessed January 17. http://data.worldbank.org/.


https://www.unisdr.org/we/inform/terminology

https://reliefweb.int/report/ethiopia/ethiopia-el-ni-o-emergency-fast-facts-june-2016



http://data.worldbank.org/

Xie, H., L. You, B. Wielgosz, and C. Ringler. 2014. “Estimating the Potential for Expanding Smallholder Irrigation in Sub-Saharan Africa.” Agricultural Water Management 131:183–193.

Yu, W., M. Alam, A. Hassan, A. S. Khan, A. Ruane, C. Rosenzweig, D. C. Major, and J. Thurlow. 2010. Climate Change Risks and Food Security in Bangladesh. London: Earthscan.

REFERENCES 75

ABOUT THE AUTHORSCarlo Azzarri is a senior research fellow in the Environment and Production Technology Division of the International Food Policy Research Institute (IFPRI), Washington, DC.

Prapti Bhandary was a senior research analyst in the Environment and Production Technology Division of IFPRI, Washington, DC, at the time of writing the report.

Alessandro De Pinto is a senior research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC.

Laia Domenèch is an independent consultant.

Hagar ElDidi is a research analyst in the Environment and Production Technology Division of IFPRI, Washington, DC.

Beliyou Haile is a research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC.

Jawoo Koo is a senior research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC.

Ho-Young Kwon was a research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC, at the time of writing the report.

Claudia Ringler is deputy division director of the Environment and Production Technology Division of IFPRI, Washington, DC.

Ricky Robertson is a research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC.

Semhar Tesfatsion is a consultant with a focus on natural resource management based in Addis Ababa, Ethiopia.

Sophie Theis was a senior rearch analyst in the Environment and Production Technology Division of IFPRI, Washington, DC, at the time of writing the report.

James Thurlow is a senior research fellow in the Development Strategy and Governance Division of IFPRI, Washington, DC.

Hua Xie is a research fellow in the Environment and Production Technology Division of IFPRI, Washington, DC.


INTERNATIONAL FOOD POLICY RESEARCH INSTITUTEA world free of hunger and malnutrition

1201 Eye Street, NW, Washington, DC 20005 USA | T. +1-202-862-5600 | F. +1-202-862-5606 | Email: [email protected] | www.ifpri.org

mailto:ifpri%40cgiar.org?subject=

http://www.ifpri.org

building resilience to climate shocks in ethiopia · 2020-03-05 · building resilience to climate...

Documents