Informing Durable Solutions by Micro-Data A SKILLS SURVEY FOR REFUGEES IN ETHIOPIA 7/10/2018 1 Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. . Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750- 8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. 2 This report was led by Utz Pape (TTL; Senior Economist, GPV01) and written together with Benjamin Petrini (Consultant, GPV01) and Syedah Aroob Iqbal (Consultant, GPV01). The case-study represents a chapter of the regional study ‘Using micro-data to inform durable solutions for conflict-related displacement in Sub-Saharan Africa’. Inputs were received by Taies Nezam (Consultant, GPV01). Comments were received by Elsa Araya (Senior Public Sector Specialist, GGOAE), Menaal Fatima Ebrahim (Consultant, GPV01), Lennart Kaplan (Consultant, GPV01), Rebecca Lacroix (Social Development Specialist, GSU03), Simon Lange (Jr Professional Officer, GPV01), Alexander Meckelburg (Consultant, GPV01), Ambika Sharma (Consultant, GPV01) and Varalakshmi Vemuru (Lead Social Development Specialist, GSU07). The team would also like to thank Pierella Paci (Practice Manager, GPV01) and Tom Bundervoet (Senior Economist, GPV01) for guidance. Vice President Makhtar Diop Country Director Carolyn Turk Senior Director Carolina Sanchez-Paramo Practice Manager Pierella Paci Task Team Leader Utz Pape 3 Table of Contents Executive Summary ......................................................................................................................................................... 7 Introduction and rationale .............................................................................................................................................. 8 1. Country Context .................................................................................................................................................... 10 1.1. Conflict and Forced Displacement: Dynamics, Scale and Profile .................................................................. 10 1.2. Legal Framework, Policy and Program Responses........................................................................................ 14 2. Demographic Profile ............................................................................................................................................. 16 3. Standard of Living ................................................................................................................................................. 20 3.1. Poverty .......................................................................................................................................................... 20 3.2. Hunger ........................................................................................................................................................... 21 3.3. Housing ......................................................................................................................................................... 22 3.4. Service Delivery: Water and Sanitation, Health, and Education ................................................................... 26 4. Employment and Livelihood ................................................................................................................................. 31 5. Social Cohesion, Public Participation, Safety and Security ................................................................................... 38 6. Movement and Return Intentions ........................................................................................................................ 42 Conclusion ..................................................................................................................................................................... 44 Appendix I: Skills Profile Survey (SPS) 2017 .................................................................................................................. 46 Appendix II: Weights ..................................................................................................................................................... 49 References .................................................................................................................................................................... 50 4 List of Figures Figure 1: Refugee population in Ethiopia by country of origin ............................................................................................. 11 Figure 2: Map of Ethiopia, refugee camps and heat map of conflict events since 1997 ...................................................... 13 Figure 3: Reasons for displacement ...................................................................................................................................... 16 Figure 4: Reasons for settling in the current location........................................................................................................... 16 Figure 5: Demographics of Refugee population in Ethiopia ................................................................................................. 17 Figure 6: Gender of Household Head.................................................................................................................................... 18 Figure 7: Dependency ratio ................................................................................................................................................... 18 Figure 8: Percentage of refugee population separated from their household members due to displacement .................. 19 Figure 9: Reasons for separation .......................................................................................................................................... 19 Figure 10: Demographics of separated population. ............................................................................................................. 19 Figure 11: Poverty Incidence of Refugee Groups, Countries of Origin, and Host Communities .......................................... 20 Figure 12: Poverty Gap.......................................................................................................................................................... 21 Figure 13: Food insecurity ..................................................................................................................................................... 22 Figure 14: Aid as a share of food consumption .................................................................................................................... 22 Figure 15: Ownership of dwelling ......................................................................................................................................... 23 Figure 16: Housing conditions............................................................................................................................................... 24 Figure 17: Overcrowding (4 or more persons per room)...................................................................................................... 25 Figure 18: Source of lighting ................................................................................................................................................. 26 Figure 19: Access to Improved sources of water .................................................................................................................. 27 Figure 20: Access to Improved Sanitation ............................................................................................................................ 27 Figure 21: Net primary and secondary enrollment rate in comparison with national enrollment rates ............................. 28 Figure 22: Mean time (minutes) taken (one-way) to fetch water ........................................................................................ 29 Figure 23: Time (minutes) to walk one way to health facility............................................................................................... 30 Figure 24: Time (minutes) to walk one way to primary school ............................................................................................ 30 Figure 25: Source of livelihood currently and before displacement..................................................................................... 31 Figure 26: Labor force participation and employment status .............................................................................................. 32 Figure 27: Highest educational attainment for working age population............................................................................. 33 Figure 28: Reasons for not participating in the labor force (Refugees) ................................................................................ 34 Figure 29: Reasons for not securing employment (Refugees) .............................................................................................. 34 Figure 30: : Access to productive assets ............................................................................................................................... 35 Figure 31: Primary employment activity............................................................................................................................... 36 Figure 32: Sector of employment ......................................................................................................................................... 37 Figure 33: Inter-personal relations between refugees and host community ....................................................................... 38 Figure 34: Host Community Feelings: Good Relations.......................................................................................................... 39 Figure 35: Host Community Feelings: Other Measures ........................................................................................................ 39 Figure 36: Participation in public meetings in the last 12 months ....................................................................................... 40 Figure 37: Interaction with community leader in the last 12 months .................................................................................. 40 Figure 38: Feelings of safety and security ............................................................................................................................. 40 Figure 39: Movement and Return Plans of Refugees ........................................................................................................... 42 Figure 40: Reasons to Stay .................................................................................................................................................... 42 Figure 41: Main support needed to settle in preferred location .......................................................................................... 43 List of Tables 5 Table 1: Number of refugee and host community households interviewed by stratum ..................................................... 46 Table 2: Sampled population by Country of Nationality....................................................................................................... 47 Table 3: Refugee Camps in Sample Frame ............................................................................................................................ 47 6 Executive Summary The Ethiopia case study addresses multiple dimensions of poverty of the refugee populations and the Ethiopian host community living in the vicinity. The analysis is based on a household survey (Skills Profile Survey) conducted with the four main refugee groups in the country (Eritreans, Somalis, South Sudanese, and Sudanese) and with members of host communities. Ethiopia has been suffering from multiple refugee crises (some more protracted, some more recent) that put a strain on coping capacity of national and local authorities. In line with the overall study’s objective, the goal is to inform policies on durable solutions through an evidence-based approach. Ethiopia’s modern history has witnessed periods of conflict and violence and ensuing displacement, both internally and as a result of its proximity to some conflict-affected states. Displacement situations in the country result from a combination of protracted conflicts in neighboring countries (Somalia, Eritrea, and Sudan), more recent crises (South Sudan, Yemen), and endemic internal ethnic unrest in some peripheral regions (Oromia, Somali/Ogaden, Afar). As a result of these regional and domestic conflicts, Ethiopia has been one of the most important refugee hosting countries for decades. There are four main Ethiopian regions that host refugees, each of whom hosts a specific group and has a unique ethnic composition: Tigray and Afar (hosting Eritreans), Gambella (South Sudanese), Benishangul Gumuz (hosting mostly Sudanese, but also South Sudanese), and Somali (Somalis). Thus, the displacement contexts are remarkably diverse: the regions hosting refugees are all peripheral and relatively underserved. Eritreans, Somalis, South Sudanese and Sudanese were displaced due to different drivers related to conflict and fragility, and each group is integrated to different degrees within Ethiopian economy and host communities. Despite adequate protection mechanisms, refugees in Ethiopia still have limited socio-economic rights, although important changes are underway. Refugees mostly live in camps that are separated from the social and economic life of host communities and are mostly dependent on aid, which is the main source of livelihood for them. In 2016, the Government of Ethiopia articulated a strategic approach in a series of pledges aimed at improving rights and expanding services to benefit both refugees and host communities. The nine pledges include potential provisions to ease the refugees’ restrictions in matters of freedom of movement, labor rights, and access to services, livelihoods and resources. The micro-level study on refugees in Ethiopia and host communities is meant to support the implementation of the new strategy. The analysis found that not only refugees witnessed trauma and life disruption (e.g. family separation), and incurred into material loss during displacement (i.e. land, livestock, and assets), but also their standard of living, livelihood and access to services’ prospects are currently highly dependent on aid. Uncertainty and inability to plan for the future represent the common denominator among refugees, which is the reason why a conducive legal framework and related policies to enable refugees’ self-reliance are key. Predictably, refugees are worse off in terms of standard of living compared to host communities, although they have comparable access to services. In fact, according to nationality, refugees fare differently with respect to standard of living, livelihood and employment, and ties to host communities. Among refugee groups, Eritreans are the ones that enjoy more rights compared to others, and, as a result, display higher standard of living and much lower poverty rates. On the other hand, South Sudanese are the poorest group on many indicators, including food security, housing, labor force participation, and ties to host community. More concerning, indicators on livelihood and access to services highlight the nearly complete dependence of refugees from aid. Many of the findings directly feed into the need to support policies of job creation, enhanced and sustainable access to markets and services, and conducive regulatory environment and governance, among others. Virtuous practices – that organically take into consideration refugees as well as host communities – are functional to economic self-reliance and to move away from dependency on external sources of aid. 7 Introduction and rationale 1. The Ethiopia case study addresses multiple dimensions of poverty of the refugee populations and the Ethiopian host community living in the vicinity. The analysis is based on a household survey (Skills Profile Survey) conducted with the four main refugee groups in the country (Eritreans, Somalis, South Sudanese, and Sudanese) and with members of host communities (Annex I). In line with the overall study’s objective, the goal is to inform policies on durable solutions through an evidence-based approach. The emphasis on both refugees and host communities acknowledges the mutual – and in some cases, interdependent – development needs of both groups. The Ethiopia case study makes use of original micro-level data on (i) a comprehensive set of social and economic indicators that assess poverty incidence, standard of living, and livelihood sources, among others; and (ii) perceptions on the refugee-host community relationship, and the future intentions and prospects of refugees. It analyzes and compares the different context and situation of the four refugee groups and the regions that host them. Box 1: Skills Profile Survey (SPS) with Refugees and Host Communities Data on refugees and host communities in Ethiopia comes from the Skills Profile Survey (SPS) 2017. SPS is a household survey that was conducted with refugees from South Sudan, Somali, Eritrea and Sudan living in camps in Ethiopia, and with Ethiopian host community members within a 5-km radius of a camp. IDPs were not included and the survey is only representative of refugees living in camps – who, nonetheless, are the majority. Sample frame was the list of all refugee camps in the four main regions that host refugees: Tigray and Afar (hosting mostly Eritreans), Gambella (South Sudanese), Benishangul Gumuz (hosting mostly Sudanese, but also South Sudanese), and Somali (Somalis). Because each region hosts a predominant majority of one refugee nationality, the stratification of the sample is practically based on nationality. A total of 900 refugee households and 500 host community households per each region were to be surveyed, and all the refugee camps in the sample frame to be included. A volatile security situation in the country imposed some changes during fieldwork. In Gambella, host community households were not surveyed, and only 439 of the intended 900 refugee households were surveyed there. The remaining interviews with South Sudanese refugees in Gambella were substituted by oversampling in Benishangul Gumuz, as 25 percent of the refugee population in this region is South Sudanese. Similarly, the escalation of violent conflict in Oromia and Somali rendered some of the camps in Somali inaccessible. The result is that camps in peaceful areas were oversampled. In addition, due to sparse host population in Somali, the final number of host households surveyed in this region was 303 against the intended sample of 500. Despite the changes, the survey captured roughly similar number of refugee households of the four main refugee nationalities. 2. Ethiopia has been suffering from multiple refugee crises (some more protracted, some more recent) that put a strain on coping capacity of national and local authorities. Contextual factors and dynamics in the four main regions that host refugees are unique as they are geographically and ethnically distinct, although they are all peripheral and relatively underserved areas. The four refugee groups are also remarkably diverse. They have been displaced for different reasons and at different times, and are integrated to different degrees within Ethiopian economy and host communities. The tremendous surge in refugees in the country in the last decade from 85,000 in 2007 to almost 1 million refugees by mid-2018, coupled with the more recent spike in the number of internally displaced persons (IDPs), makes the present analysis timely. 1 In addition, the Government of Ethiopia is currently designing policies to implement a 2016 strategic initiative (“nine pledges”), which is aimed at expanding refugee rights and service delivery, and comprehensively addressing socio- economic issues of both refugees and host populations. 3. In the last decade and half, Ethiopia has achieved tremendous progress in terms of growth and poverty reduction. 2 Between 2004-2014, the economy recorded an annual average growth rate of 10.9 percent. The percentage of people living below the international poverty line of USD 1.25 PPP a day decreased substantially, from 56 percent to 30 1 UNHCR Statistics. http://popstats.unhcr.org/en/overview [Accessed on April 23, 2018]. 2 World Bank, 2015. 8 percent in 2000-2011. Remarkably, poverty reduction has been more pronounced in poorer regions, some of which host refugees (including Tigray and Benishangul Gumuz, among others), leading to some degree of spatial convergence in the overall level of economic development in the country. In addition, throughout this progress, Ethiopia largely maintained its traditional low levels of inequality. Ethiopian households improved their material condition, living standard and access to services on several economic and human development indicators, including health, education, nutrition, and livelihoods. Poverty eradication has been driven by agricultural growth, accompanied by state-led expansion of services and effective safety nets for the poorest. Nonetheless, Ethiopia’s economy has been diversifying: between 2004-2014, the share of agricultural within the economy fell from 52 to 40 percent, while the service sector rose from 37 to 46 percent. 9 1. Country Context 1.1. Conflict and Forced Displacement: Dynamics, Scale and Profile 4. Ethiopia’s modern history has witnessed periods of conflict and violence and ensuing displacement, both internally and as a result of its proximity to some conflict-affected states. Displacement situations in the country result from a combination of protracted conflicts in neighboring countries (Somalia, Eritrea, and Sudan), more recent crises (South Sudan, Yemen), and endemic internal ethnic unrest in some peripheral regions (Oromia, Somali/Ogaden, Afar). Internal conflicts both in Ethiopia and in neighboring countries, combined with some regional inter-state conflicts (Ethiopia vs Somalia, Ethiopia vs Eritrea), produced socio-economic marginalization and protracted instability in peripheral areas and borderlands, and continuous waves of displacement of IDPs and refugees – currently estimated at 1.1 million and nearly 920,000, respectively. 3 Conflicts in Somalia and Sudan, among others, are also inextricably linked to environmental degradation and migration patterns associated with livelihoods and food security. Domestically, vulnerability to droughts and floods coupled with ethnic-based violence, issues of access to and ownership of resources and related governance arrangements, are core conflict issues that are intertwined with displacement of refugees and IDPs 4. 5. Regionally, there are three main theaters of crises that have caused refugees to pour into Ethiopia over several decades (Figure 1). First, conflict and instability in Somalia are an example of how interstate dynamics conflate with domestic unrest in Ethiopia, and clan-based conflicts and high levels of violence in Somalia. Since the 1960s, Ethiopia has been engulfed in a border conflict with Somalia over the Ogaden region of Ethiopia (or ‘Somali National Regional State’, as per official name 5), including a full war being fought in 1977-78. The Somali region of Ethiopia has also been claimed by a low-level domestic insurgency led by the ethnic-based Ogaden National Liberation Front (ONLF) since 1994. In the last decade, Ethiopia and Somalia cooperated to counteract violent extremism in the region. From the east, Somali refugees escaped repeated cycles of internal violence occurring from the late 1980s on, including during the 2008 drought.6 Second, in the north, following Eritrea’s peaceful independence from Ethiopia (1993), an interstate war between the two countries over a disputed border took place (1998-2000), with instability and further violence occurring as recently as 2012. In the last decade, refugees from Eritrea mostly escaped political persecution, military conscription and economic hardship. 7 The third front of instability is to the west of Ethiopia, at the border with Sudan and South Sudan. The former has had several internal armed conflicts since the 1950s, resulting in continuous cycles of refugees. One such conflicts led to the independence of South Sudan (2011), which, in turn, has experienced a full blown civil war since December 2013. The war has had a severe toll in terms of refugees fleeing to Gambella (Ethiopia), a region with traditional ethnic tensions between the Nuer and Anuak groups. 8 6. As a result of these regional and domestic conflicts, Ethiopia has been one of the most important refugee hosting countries for decades. The country hosts the sixth largest refugee population in the world and the second-largest refugee population in Sub-Saharan Africa, after Uganda. As of May 2018, Ethiopia hosts 920,000 refugees. Nearly half of them arrived in the last four years from South Sudan (443,000), with the other half coming at different times from Somalia 3 IOM, 2018a; UNHCR, 2018a; Carter and Rohwerder, 2016. 4 “Ethiopia has been the scene of numerous non-state conflicts between different ethnic groups, some occurring between pastoralist communities and others pitting pastoralists against agriculturalists. Common for most of these conflicts have been that they have worsened as a result of continued droughts in the region.” Uppsala Conflict Data Program (Date of retrieval: 18/03/23); International Crisis Group, 2009. 5 In 1995, the Ethiopia’s state turned into an ethnic-based regional system by proclaiming a federal state composed of nine regional states and two federally-administered cities (Addis Ababa and Dire Dawa). Afar, Amhara, Oromia, Somali, and Tigray are single-ethnic states, while Benishangul Gumuz, Gambella, Harar, and Southern Nations, Nationalities and Peoples Region (SNNPR) are multi-ethnic states. International Crisis Group, 2009. 6 International Institute for Strategic Studies (IISS); International Crisis Group, 2013; Carter and Rohwerder, 2016; Richards and Bekele, 2011. 7 International Institute for Strategic Studies (IISS); Carter and Rohwerder, 2016. 8 Feyissa, 2014; Asfah Gemechu, 2016; World Bank, 2016; Carter and Rohwerder, 2016. 10 (256,000), Eritrea (169,000), and Sudan (44,000). Refugees from these four nations tend to settle close to their country of origin (Figure 2). The nearly 8,000 remaining refugees come from some 15 countries, including Yemen. 9 In 2017, nearly 110,000 new refugees arrived in Ethiopia, overwhelmingly from South Sudan (75,000) and Eritrea (25,000). 10 These trends (Figure 1) show how forced displacement in Ethiopia is both a humanitarian emergency and a protracted crisis. Figure 1: Refugee population in Ethiopia by country of origin 900 Thousands 800 700 Number of Refugees 600 500 400 300 200 100 0 Eritrea Somalia South Sudan Sudan Source: Authors’ calculations based on UNHCR data 7. Forced displacement of IDPs is a rising emergency too. As of end of 2016, Ethiopia had 258,000 IDPs originating mostly from the unrest in Ogaden/Somali. Following widespread anti-government protests and ensuing violence and instability, conflict-induced IDPs have skyrocketed to an estimated 1.1 million by April 2018 (plus another 488,000 IDPs due to environmental factors). 11 This dramatic spike is due to localized insecurity that has taken place since late 2015 in several Ethiopian regions, including Oromia, Somali, Tigray and Amhara. Current conflictive dynamics reinforced long- standing ethnic grievances of some ethnic groups and regions. The issue of internal forced displacement in Ethiopia is also a protracted one, which is intertwined with highly variable environmental conditions. 12 8. Since the conflict contexts are diverse, the displacement situations are similarly heterogenous. Most refugees settle in areas of Ethiopia that border their country of origin and reside in camps. Refugee populations have marked differences in terms of size, demographics, causes and length of displacement, as well as origins and prospects. In turn, development needs and durable solutions apply differently to the different refugee groups. Border regions that host refugees are often isolated and lagging, some of which suffer from domestic conflicts too (e.g. Somali, Oromia, Afar, and Gambella). 13 Similarly, there are ethnic, socio-economic and security-related differences among hosting regions in Ethiopia, in addition to the differences between refugee groups. Thus, the present report adopts a differential analytical approach to the displacement situations in the country. 9 UNHCR, 2018a. 10 UNHCR, 2018b. 11 IOM, 2018b; IDMC, 2017; IDMC, 2018a; IOM, 2018a; OCHA, 2018. See also, International Crisis Group, 2009; Tesfaye, 2016. 12 “Internal displacement in Ethiopia is multi-causal and complex. The confluence of numerous drivers and triggers of new displacement is so complex that any attempt to distinguish between displacement caused by conflict or disaster is rendered pointless. The interaction between high levels of existing vulnerability in rural populations; severe droughts, sometimes followed by heavy rains and floods; ongoing conflict; already high numbers of displaced people; and overstretched government capacity create a high-risk environment in which new displacements are likely to continue.” IDMC, 2018b. 13 While the four regions where fieldwork took place are traditionally poorer than national average, the host communities that were surveyed for this study were mainly peri-urban with poverty trends and development indicators close to the national level. 11 9. Over 90 percent of the 443,000 South Sudanese refugees settled in camps along the southwestern border of Ethiopia, in Gambella, a traditionally fragile and underserved region. 14 Here, the South Sudanese refugees (mostly ethnic Nuer) outnumber the host population (307,000 as of 2007 census), causing enormous strains on food security, service delivery and access to livelihoods. The two main Ethiopian ethnic groups in Gambella are the Nuer (47 percent, mainly pastoralists) and the Anuak (21 percent, mostly farmers), who at the national level represent a mere 0.2 and 0.1 percent, respectively. 15 The presence of refugees has exacerbated existing tensions between the two groups over land and water rights. However, the influx of refugees and the associated flow of humanitarian assistance have also benefitted host communities through infrastructure projects, and expanded services and local markets.16 10. Nearly all of the 256,000 Somali refugees reside in eight camps in the conflict-affected Somali region (i.e. Ogaden) of eastern Ethiopia. The largest group of Somali refugees (208,000) live confined in Dollo Ado, a small area with a host population of 141,000. 17 While Somali refugees arrived in waves over several decades, most of the current refugees arrived after the 2008 drought and the Ethiopian-Somali joint military campaign against violent extremism (2006- 09). Somali refugees share the same ethnic group of host communities in the Somali region of Ethiopia: 97 percent of the 4.4 million Somali region’s population (2007) is ethnic Somali, who, in turn, are 6.2% at the national level. In the Somali, historic grievances against the central government fueled a protracted insurgency by the ONLF and brutal state responses, including a counterinsurgency campaign in 2007. In addition, inter- and intra-clan conflicts traditionally beset the Somali region, increasingly over ownership of resources and governance arrangements. Currently, the region remains unstable. Rural population (85 percent of the total) is mostly pastoralist (60 percent), with presence of agro-pastoralists (25 percent), and farmers (14 percent). 18 14 Ethiopia currently hosts 17 percent of the total 2.4 million South Sudanese refugees in the region. UNHCR, 2018c. In 2016, South Sudan had the third largest, after Syria and Afghanistan, and fastest growing refugee population in the world. UNHCR, 2017; UNHCR, 2018d; Sarzin, 2017; World Bank and UNHCR, 2015. 15 Conflict between Anuak and Nuer across Sudan/South Sudan – Ethiopia border dates back to the 19th century. Civil war in Sudan in the 1980s caused a wave of ethnic Nuer refugees from Sudan to pour into Ethiopia, permanently altering Gambella’s ethnic composition. World Bank, 2016; Feyissa, 2014; Asfah Gemechu, 2016. 16 Girma, 2016. 17 Sarzin, 2017; World Bank, 2016. 18 Richards and Bekele, 2011; Sarzin, 2017; International Crisis Group, 2009, 2013; World Bank, 2016, 2017. 12 Figure 2: Map of Ethiopia, refugee camps and heat map of conflict events since 1997 Source: Armed Conflict Location and Event Data 11. Half of the 169,000 Eritrean refugees have settled along the northern border, in the Afar and Tigray regions. Compared to other refugee groups, only one third of Eritrean refugees live in camps, with many thousands living in individual accommodations and/or benefitting from the ‘Out of Camp’ scheme (see below). Eritreans also account for most of the 21,000 refugees who reside in Addis Ababa. 19 Many Eritreans do not intend to permanently settle in Ethiopia, but seek to reach Europe via Sudan and Libya: 65 percent of Eritrean refugees arriving in camps in Tigray leave within the first year. The regions of Tigray and Afar are home to 4.3m and 1.4m individuals, respectively. Both are also ethnically homogenous: 97 percent of Tigray’s population is ethnically Tigray, and over 90 percent of Afar’s population is ethnically Afar. 20 12. The vast majority of the 44,000 Sudanese refugees reside in four camps in the Beneshangul-Gumuz region in western Ethiopia. Sudanese refugees come from protracted conflicts and chronically marginalized regions of Sudan, including Darfur, Abyei, Southern Kordofan, and Blue Nile State (from where most of the Sudanese refugees settling in Ethiopia come from). Beneshangul-Gumuz is a multi-ethnic region inhabited by 780,000 individuals (2007) with a 19 UNHCR, 2017; UNHCR, 2018e; Sarzin, 2017. 20 World Bank, 2017; Sarzin, 2017; Mallet et al. 2017. 13 predominantly farming economy. Despite favorable climatic conditions it suffers from food insecurity due to limited social and economic infrastructure, and more recently environmental degradation. 21 1.2. Legal Framework, Policy and Program Responses 13. Ethiopia formally adheres to international or regional instruments of refugee protection, and has adopted national policies consistent with international standards. It signed and ratified the 1951 UN Convention Relating to the Status of Refugees, the 1967 Protocol, and the 1969 Organization of African Unity Convention Governing Specific Aspects of Refugee Problems in Africa. Ethiopia is also a party to the 2010 Kampala Convention on the protection of IDPs. At the national level, the Refugee Proclamation No. 409/2004 – currently being reformed – establishes protection standards and related rights for refugees, providing the legal framework for Ethiopia’s traditional open-door policy to refugees (Art. 9). The proclamation expresses a commitment for safe reception, promoting peaceful coexistence, and returning refugees when conditions in countries of origin are safe. Institutionally, the Administration for Refugees and Returnee Affairs (ARRA) is responsible for the camps’ management, security and some services (health, education, WASH). Increasingly, some refugees access basic services through national providers, as well as some host communities benefit from ARRA-provided services. 22 14. Despite adequate protection mechanisms, refugees in Ethiopia have had so far limited socio-economic rights – a situation that is progressively improving. The proclamation allows competent authorities to restrict refugees’ freedom of movement to designated settlements (art. 21), which has been the legal basis for Ethiopia’s overt use of camps to host refugees. Camps are separated from the social and economic life of host communities and are mostly dependent on aid. In the last decade, the restriction to freedom of movement has been partially lifted. While Ethiopia remains one of the most camp-reliant refugee hosting countries globally, the introduction of an ‘Out-of-Camp’ scheme in 2010 allowed Eritrean refugees who had the means to support themselves to live in urban areas. 23 In terms of livelihoods, refugees rely primarily on humanitarian aid. Under the 1995 Constitution, only Ethiopian citizens are granted the right to work, and there are no provisions under Ethiopia’s law for local integration of refugees. While self-employment is limited to the few privileged and business licenses are not available to refugees, they can only seek limited opportunities in camps or in the informal sector of surrounding areas. 24 15. In 2016, the Government articulated a strategic approach in a series of nine pledges aimed at improving rights and expanding services to benefit both refugees and host communities. The nine pledges include potential provisions to ease the refugees’ restrictions in matters of freedom of movement, labor rights, and access to services, livelihoods and resources (Box 2). The groundbreaking plan has a long-term vision that includes interventions in several social and economic sectors beyond the assistance to refugees. In synergy, target key areas can strengthen refugees’ self- reliance by moving away from aid dependency and contribute to greater socio-economic inclusion of refugees. Pledges’ implementation in each area is at the inception stage and the government is due to soon pass a new Refugee Regulation to amend the 2004 Proclamation. The pledges are part of the GoE’s adoption of the Comprehensive Refugee Response Framework (CRRF) – an initiative launched at the UN General Assembly meetings in September 2016. Under the CRRF, the international community commits to support host countries and refugee groups through a more comprehensive response to displacement crises. Specifically, the GoE’s adoption of the CRRF entails out-of-camp support to refugees and refugee integration within host communities, gradually moving away from the current in-camp assistance to refugees. 21 UNHCR, 2018f; Sarzin, 2017; World Bank, 2016. 22 Sarzin, 2017; World Bank, 2016; World Bank, 2018a. 23 As part of the 2016 new strategy (see below), the Government aims to bring the percentage of urban refugees to 10 percent. The commitment to expand the out-of-camp policy recognizes that it has enhanced the self-reliance of Eritrean refugees who are living outside the camps, increased opportunities to pursue educational and employment opportunities, and reduced pressures to embark on perilous journeys to Europe. ARRA, 2017. 24 Zetter and Ruaudel, 2016. 14 Box 2: The Government of Ethiopia’s Nine Pledges The nine pledges are a GoE’s strategy to sustainably address refugee crises beyond the humanitarian level and with the involvement of a broader array of stakeholders. In line with the CRRF, they represent a ground-breaking approach that includes policy provisions to benefit both refugees and host communities. The nine pledges are the following: (i) Expand the Out of Camp policy to benefit 10% of refugees. (ii) Provide work permits to refugees and those with permanent residence identification. (iii) Provide work permits to refugees in the areas permitted for foreign workers (iv) Increase enrolment in all levels of education to all qualified refugees. (v) Make available 10,000 hectares of irrigable land to allow 100,000 people (refugees and local population) access to crop production. (vi) Allow refugees who have lived in Ethiopia for 20 or more years to locally integrate (vii) Build industrial parks, reserving some jobs for refugees. (viii) Strengthen provision of social services. (ix) Provide access to other benefits including birth certificates, bank accounts, and driving licenses. Source: ARRA, 2017. 16. The World Bank currently supports the GoE’s efforts to address the displacement crises. In addition to the roll out of the pledges, support to the implementation of the CRRF includes also some potential operational changes at ARRA, from an agency exclusively responsible for refugee protection to a potentially larger role of coordination of refugee policies within Ethiopia’s broader economic development agenda. The Bank’s portfolio on displacement in the country includes the ongoing Development Response to Displacement Impact Project in the Horn of Africa (P152822), aimed at improving access to services, enhancing environmental management of host areas, and expanding economic opportunities for both refugee groups and host communities in Ethiopia, Djibouti and Uganda. Two more projects under preparation include: the Economic Opportunities Program (P163829) with a refugee component to create jobs, build skills, and support work permits and business license for refugees; and the Additional Financing to Education Project. Some of the funding for these programs are drawn from the IDA18 USD 2-billion sub-window dedicated to refugees and to supporting hosting countries. 25 25 World Bank, 2018a; Sarzin, 2017; World Bank, 2016. 15 2. Demographic Profile 17. Differences in displacement dynamics between the four main refugee groups (South Sudanese, Somali, Eritreans, and Sudanese) are reflected in their demographic profiles. Refugee populations display unique demographic features based on the drivers and immediate causes of their displacement. Demographics of the four groups are also related to the length of displacement, to the severity of violence to which refugees were exposed, and more generally to the intensity and status of the conflicts in the four countries. 18. Sudanese, South Sudanese and Somalis overwhelmingly link their displacement to security issues, while for Eritreans state persecution is the most important driver of displacement (Figure 3). Armed conflict is the most prevalent reason for fleeing that Sudanese, South Sudanese and Somali refugees report. For Somalis, both rising crime and violence and drought are particularly relevant too. Most Eritrean refugees report being displaced to Ethiopia due to political persecution (48 percent) and rising crime (30 percent), while another combined 11 percent fled due to family reasons and lack of employment. Only 5 percent of Eritreans maintains that the opportunity to migrate to another country was the leading factor for displacement: this percentage seems low, as in other studies such reason has come up as a more frequent explanation. 26 The most important causes of displacement for Eritreans are either marginal or negligible for the other three groups. 19. All refugee groups list improved security as the leading reason for settling in the current location of displacement in Ethiopia (Figure 4). This point directly links the reasons for fleeing to the reasons for settling in a specific location as refugees. Unsurprisingly, conflict and security are the most recurrent factors of forced displacement dynamics both as push and pull factors. Conflict and security are related to forced displacement in terms of (i) what the causes are for an individual or group to flee (i.e. push factors), and (ii) what his/her motives are for choosing and settling in a new place to live (i.e. pull factors). In addition to security, Eritrean respondents provide more diverse answers. Due to the more open legislation and policies by Ethiopia toward Eritrean refugees compared to the other groups, more than 30 percent Eritreans state that the possibility to join family members and employment opportunities were factors in their decision to settle in the current location in Ethiopia. These two reasons are barely mentioned at all by the other three groups. In addition, approximately a quarter of Eritrean, South Sudanese and Somali refugees state that access to humanitarian aid was driving their decision to settle in a specific area, whereas this factor is not relevant for Sudanese. Figure 3: Reasons for displacement Figure 4: Reasons for settling in the current location 100 Migration to another country 90 80 To join 70 family/friends Percentage of Refugees 60 Better access to 50 livelihood 40 opportunities 30 Better access to assets/services 20 10 Access to 0 humanitarian aid Better Security Source: Authors’ calculations. Source: Author’s calculations. 20. Demographic profiles of the four groups present differences in terms of age and gender (figure 5). Eritrean refugees have a higher percentage of men who are in military age (over 31 percent combined adults and youth aged 15-24), 26 Mallet et al. 2017; Amnesty International, 2016. 16 compared to the other three groups (14 percent South Sudanese, 18 percent Somalis, and 20 percent Sudanese). This fact could be explained by the fact that Eritrean men at home are subject to a mandatory, indefinite and harsh military service, and have therefore a higher incentive to flee than similar age groups in other countries. 27 21. One more remarkable difference across groups concerns the pronounced gender imbalance in favor of women among South Sudanese refugees. While there are 16.4 percent South Sudanese refugees who are adult women, there are only 6 percent adult men (Figure 5). Other groups, including host community, have an approximately similar gender balance in their adult population. 28 In turn, the low percentage of adult men among South Sudanese refugees results in an overwhelming percentage of South Sudanese refugee women-headed households in Ethiopia (91 percent), compared to Eritrean, Somali and Sudanese refugee households. The majority of the latter three groups’ households are headed by men (Figure 6). Figure 5: Demographics of Refugee population in Ethiopia Children (0-14) Youth (15-24) Adults (25-64) Elderly (>65) 60 Percentage of population 50 13.5 16.4 10.7 11.5 8.8 14.2 17.0 6.0 12.7 40 14.5 11.7 11.6 8.6 9.2 8.2 7.4 7.2 30 8.1 7.3 17.8 8.3 7.1 8.5 12.4 20 30.3 28.4 30.8 30.8 32.1 28.8 32.1 24.9 24.8 23.5 26.3 10 19.9 0 Men Women Men Women Men Women Men Women Men Women Men Women Refugee Host Community South Sudanese Somalis Eritreans Sudanese Source: Authors’ calculations. 22. In terms of age group, refugees are predominantly young, leading to high dependency rates. All groups except Eritreans have higher percentages of children than host communities. About 60 percent of the refugee populations of Sudan, South Sudan and Somalia is under 15 years of age, as compared to 50 percent of the host community population (Figure 5). On the other hand, host communities have higher proportion of adults (especially men) than the four refugee groups. Young populations lead to high dependency ratios, which put an increased burden on working age individuals (Figure 7). In turn, relative shortage of men in working age among refugees may indicate higher rates of dependency from external aid, for example, and lower rates of economic self-reliance. Dependency ratio is highest for South Sudanese and Somali refugees: one working age member is responsible to support approximately two dependents compared to the 1:1 ratio among host populations. This situation speaks to the extreme vulnerability of refugees when considering the overwhelmingly prevalent women-headed households of South Sudanese refugees, for example. 27 Human Rights Watch, 2017: 201. 28 The gender imbalance among South Sudanese refugees could be partially attributed to the intensity and volatility of the civil war in South Sudan. There, adult men may be more systematically recruited by armed groups, compared to say Somalia and Sudan where conflict dynamics are protracted. Such hypothesis could be confirmed by survey respondents from South Sudan, who list “recruitment” as an important factor for family separation (30 percent), compared to the other three refugee groups, for which “recruitment” is negligible (figure 8). However, that is not the full picture. Such gender imbalance among South Sudanese refugees exists among South Sudanese population too (see South Sudan case study). 17 Figure 6: Gender of Household Head Figure 7: Dependency ratio Man Woman 2.1 1.9 1.9 Dependency Ratio 100 90 1.7 Percentage of households 80 70 1.1 1.2 60 50 40 30 20 10 Eritreans South Sudanese Somalis Sudanese Host Community Refugees 0 Eritrean Sudanese Host Community Refugee South Sudanese Somali Overall Refugee Overall Refugees Source: Authors’ calculations. Source: Author’s calculations 23. Rates of separation from family members during forced displacement are different for the four refugee groups. These differences may be related to the drivers and circumstances of displacement. More than one in four refugees report that they have been separated from immediate family members due to displacement (Figure 8). This estimate is higher for Eritrean and South Sudanese refugees (39 and 38 percent, respectively) than for Sudanese and Somalis (27 and 4 percent, respectively). Arguably, the drivers of forced displacement in the former groups (intense violent conflict in South Sudan and repressive regime in Eritrea) could help explain their higher rates of family separation than the latter groups. The intensity of the civil war in South Sudan, the emergency character of displacement, and recruitment by armed groups may cause refugees to flee faster than in the Somalia context, for example. As a result, more South Sudanese families may get separated during displacement, compared to Somalis and Sudanese who may have more time to organize and flee. For different reasons, Eritrean refugee families may suffer from higher rates of separation because either politically active men or men in military age are at higher risk of incarceration and conscription, respectively. Ultimately, Eritrean men may be forced to flee even without family members; the latter may be at lower risk and thus may have lower incentives to flee. 24. As the four refugee groups exhibit different percentages of family separation, the reasons why family members got separated are also different among the four refugee groups. Most of these separated household members were left behind in the country of origin at the time of displacement. Eritrean and Sudanese survey respondents list this as the only reason for separation. Almost half Somali and a quarter of South Sudanese refugees, instead, report that family members were displaced to another location. In South Sudan, recruitment by armed groups play also an important role in separating families, as nearly one in three South Sudanese male family members either joined or was recruited by armed groups (Figure 9). 29 In turn, separation contributes to refugees’ vulnerabilities. 29 For South Sudanese IDPs the data does not show that recruitment plays a role in separating families (see South Sudan case study). 18 Figure 8: Percentage of refugee population separated from their Figure 9: Reasons for separation household members due to displacement 100 Percentage of separated population 45 80 40 60 Percentage of population 35 40 20 30 0 25 Women All Women All Women All Women All Women All Men Men Men Men Men 20 15 Refugee - South Somalis Eritreans Sudanese 10 Overall Sudanese 5 0 Stayed behind in the country of origin Refugee - South Somali Eritreans Sudanese Displaced to another location Overall Sudanese Recruited by or joined armed forces Source: Authors’ calculations. Deceased 25. In Somalia, South Sudan and Sudan, a disproportionately high percentage of separated members are adult men compared to women. For South Sudan, 48 percent of the separated members are adult men and 30 percent are adult women, while for Sudan 36 percent of the separated members are adult men and 24 percent are adult women (Figure 10). For Somali refugees, the gender discrepancy of separated family members is even more tilted toward men: over 33 percent of the separated members are adult men compared to only 6 percent of adult women. For both Somali and Eritrean refugees, there is also a higher proportion of separated members who are children and youth than the other two groups. Eritrean refugees experience an opposite trend than the previous three groups. As Eritrean men cross into Ethiopia to avoid forced military service, a disproportionately higher percentage of separated members are adult women – 24 percent of adult women are separated as compared to 15 percent of adult men. Figure 10: Demographics of separated population. 80 household members/households 60 Percentage of separated 40 33.4 23.7 34.0 48.4 6.2 15.4 35.5 24.7 26.7 20 30.4 0 Men Women Men Women Men Women Men Women Men Women Refugees - Overall South Sudanese Somalis Eritreans Sudanese Children (0-14) Youth (15-24) Adults (25-64) Elderly (>64) Source: Authors’ calculations based on Skills Profile Survey, Ethiopia (2017). 19 3. Standard of Living 3.1. Poverty 26. Refugees in Ethiopia are much poorer than host community households, but poverty rates across the four refugee groups are quite heterogenous, ranging from 38 per cent (Eritreans) to 72 percent (Sudanese). Every 2 in 3 refugees live below the international poverty line of US$1.90 per day per person at 2011 PPP (Figure 11). 30 This compares with around 1 in 4 host community members living below the poverty line. Some refugee groups (South Sudanese, Eritreans) fare better compared to their respective country of origin at the national level, while some others (Sudanese, Somalis) fare worse. South Sudanese and Sudanese refugees in Ethiopia have the highest incidence of poverty (each group with over 70 percent), but the two countries of South Sudan and Sudan have remarkably different poverty rates (82 percent and 46 percent, respectively). Poverty incidence in Eritrea is also high (66 percent), although this figure and the one for Sudan are quite outdated. However, Eritrean refugees in Ethiopia have the lowest poverty rate amongst the refugee population in the country with a poverty rate that is closer to that of host population. Poverty incidence of women-headed households is higher than men for all groups, except for Somali refugees for whom the percentages between women and men are approximately equal. Figure 11: Poverty Incidence of Refugee Groups, Countries of Origin, and Host Communities 100 Poverty Incidence (below $1.9 capita per day) 80 Extreme Poverty Incidence (bel Poverty Incidence (%of population) $1.25 per capita per day) 60 South Sudan - 2016 40 Somali - 2016 20 Eritrea - 2003 0 Sudan - 2009 Host community Woman head Woman head Overall Woman head Overall Woman head Overall Woman head Refugee Man head Man head Overall Man head Man head Man head South Sudanese Somali Eritrean Sudanese Overall Refugee Source: Authors’ calculations. 27. Refugees will need a substantial increase in their consumption to overcome poverty. On average, refugees who are poor consume 28 percent below the poverty line (Figure 12). 31 Like the trends in poverty incidence, South Sudanese and Sudanese refugees suffer from the highest poverty gap followed by Somali and Eritrean refugees. On average, the former two groups consume 34 percent below the poverty line while Somalis consume 23 percent below the poverty line. Similar to the trends observed for poverty, Eritrean refugees are the better off group among the four: Eritreans who are poor on average consume 11 percent below the poverty line. In addition, households headed by women both among all refugees and within all four groups face a higher poverty gap than households headed by men. 30 The amount in USD equals 17.9 Birr per day per person in 2017. 31 This percentage amounts to around 14 Birr per person per day as compared to the poverty line of 17.9 Birr per person per day. 20 Figure 12: Poverty Gap Source: Authors’ calculations. 3.2. Hunger 28. Hunger trends mirror poverty trends for the four groups and their demographics, and vis-à-vis host population, resulting in high dependence on food aid. On average, 2 in 3 refugees are highly food insecure compared to 1 in 4 host community members experiencing high food insecurity (Figure 13). 32 Refugees in Ethiopia, specifically South Sudanese, Sudanese and Somali refugees, are in dire need of food aid to overcome high food insecurity in the camps, which is around 80 percent for the former two groups, and 57 percent for the Somalis. Mirroring poverty trends, Eritrean refugees fare similarly to host community members and have the lowest food insecurity rates among refugee groups. In addition to higher poverty incidence, women-headed households are also more food insecure than men, and as a result they are more vulnerable. Predictably, food security improves with income: 1 in 3 refugees among the richest quintile experiences low food insecurity, compared to 1 in 100 for the poorest quintile and an average 1.3 in 10 for all the other quintiles. These high rates of food insecurity are despite the available food aid in the refugee camps. On average, 44 percent of the food consumption of refugees come from food aid (Figure 14). This percentage is over 50 percent for the poorest refugee groups (South Sudanese and Sudanese), while it is lower for Eritreans and Somalis (36 and 26 percent, respectively). 32 Food insecurity is defined as an individual facing food shortage at least once in the previous 7 days and using a combination of coping strategies to overcome the shortage. It is calculated using Reduced Coping Strategies Index (rCSI) adapted by WFP/VAM (World Food Programme/Vulnerability Analysis Mapping Unit), FAO/FSNAU (UN Food and Agriculture Organization/Food Security and Nutrition Analysis Unit for Somalia), and the Global IPC (Integrated Phase Classification) team, among others. rCSI is a weighted index that combines information on frequency and severity of coping strategies used in a single score for household food security. 21 Figure 13: Food insecurity Figure 14: Aid as a share of food consumption 100 100 Percentage share of aid in food Percentage of population 80 80 60 60 consumption 40 40 20 20 0 Man Head Richest Quintile Somalis Poorest Quintile Q2 Q3 Q4 Eritreans Sudanese Host Community Refugee South Sudanese Woman Head 0 4. Eritrea Man Head Richest Quintile 1. South Sudan Poorest Quintile Q2 Q3 Q4 Host Community 2. Somalia Refugee Woman Head 5. Sudan Overall Refugee Overall Refugee High food insecurity Medium food insecurity Source: Author’s calculations. Low food insecurity Source: Authors’ calculations. 3.3. Housing 29. Housing situation of refugees in Ethiopia – assessed in terms of housing ownership, conditions, and overcrowding – goes hand in hand with poverty and hunger trends. Even more than food aid, housing needs of refugees in Ethiopia are almost entirely provided by United Nations or non-profit organizations through temporary shelters. Because these sources are unsustainable, strengthening self-reliance is key to achieve durable solutions. As for other standard of living measures, refugees fare worse than host community members on housing issues too. Ethiopia’s overreliance on camps is confirmed by survey respondents among refugees, 95 percent of whom live in temporary shelters, compared to nearly 80 percent host community members who live in owned dwellings (Figure 15). An overwhelming majority of refugee population (81 percent) live in unimproved housing 33 and 59 percent of the refugee population live in overcrowded housing. 34 In contrast, 44 percent of the host community population live in unimproved housing and 25 percent of them live in overcrowded housing (Figure 16 and Figure 17). Unimproved and crowded housing conditions increase the likelihood of disease transmission for a wide range of respiratory diseases including pneumonia, tuberculosis and many allergies. Both refugees and host community populations need improvement in housing conditions as well as in food security. 33 Unimproved housing is defined as structure that is not made of wood, concrete or block and/or that is not intended for habitation. 34 Housing is defined overcrowded when there are 4 or more individuals per room (UN-HABITAT 2016). 22 Figure 15: Ownership of dwelling 100 Percentage of population 80 60 40 20 0 Refugee- before displacement Richest Quintile Host community Poorest Quintile Q2 Q3 Q4 Refugee - current Woman head Before displacement Before displacement Before displacement Before displacement Man head Current Current Current Current South Somalis Eritreans Sudanese Sudanese Overall Refugee Temporary shelter provided by UN/INGOs Squatting Relatives/friends Work-provided Rented Owned Source: Author’s calculations based on Skills Profile Survey 2017 30. According to their nationality, refugee groups experience differences in terms of housing. Trends on ownership of dwellings are significantly low for all groups, ranging from 2 to 11 percent (Figure 15). This is simply explained by the fact that surveys were conducted in camps where temporary shelters are the norm. As more Eritreans live in urban settings, arguably, their rate of either rented or owned dwelling may be higher than data from survey results. Trends in overcrowding are fairly homogenous too, ranging from 2 in 3 South Sudanese living in overcrowding housing to 2 in 5 Sudanese living in such conditions (Figure 17). More differentiation concerns housing conditions. As for poverty and food security, Eritreans fare better than the other 3 groups, in line with lower poverty incidence, higher consumption rate and greater disposable income. One in two Eritrean refugees live in improved housing, compared to 26 percent of Somalis, 15 percent of South Sudanese, and 4 percent of Sudanese (Figure 16). 23 Figure 16: Housing conditions 100 Percentage of population 80 60 40 20 0 Richest Quintile Refugee- Before displacement Host community Poorest Quintile Q2 Q3 Q4 Refugee - current Woman head Before displacement Before displacement Before displacement Before displacement Man head Current Current Current Current South Somalis Eritreans Sudanese Sudanese Overall Refugee Unimproved housing Improved housing Source: Author’s calculations based on Skills Profile Survey 2017 31. Housing situation of refugees was better before displacement than during it. Even though some refugee groups (South Sudanese, Eritreans) have lower poverty rates compared to their respective country of origin, all refugee groups have worse housing situations during displacement than before being displaced. While 92 percent of the refugee population owned their dwellings before displacement, only 5 percent of refugees own their dwellings in Ethiopia. Similarly, only 1 out of 4 refugees before displacement used to live in an overcrowded housing, while 3 out of 5 are currently in that situation. South Sudanese experience the greatest gap, as over 90 percent used to live in not overcrowded housing before displacement compared to 35 percent South Sudanese refugees who are currently able to live in not overcrowded housing. Among refugee groups, South Sudanese are also the ones with the highest percentage of refugees living in both overcrowded and unimproved housing. Interestingly, a higher percentage of Somali refugees live in improved housing than they used to before displacement – a trend that links with ownership of dwellings which, albeit low in absolute terms, is highest for Somali refugees when compared to other groups. 24 Figure 17: Overcrowding (4 or more persons per room) 100 Percentage of population 80 60 40 20 0 Richest Quintile Refugee - before displacement Host community Poorest Quintile Q2 Q3 Q4 Refugee - current Woman head Before displacement Before displacement Before displacement Before displacement Man head Current Current Current Current South Somalis Eritreans Sudanese Overcrowded Not Overcrowded Sudanese Source: Author’s calculations based on Skills Profile Survey 2017 32. Similarly, refugees have much lower access to electricity or solar power/biogas than host community households (37 percent versus 66 percent, respectively), but higher access compared to before being displaced. Access to electricity or solar power/biogas increases with household income. Interestingly, only the richest quintile has access to electricity (17 percent), while all other quintiles have exclusively access to solar power/biogas. As women headed households are poorer, a lower percentage of women headed households have access to electricity as compared to households headed by men (Figure 18). There are also substantial and significant differences between the average hours of electricity available to refugee and host community populations and whether the two populations can charge their phones using the electricity. While host community members receive around 16 hours of electricity per day, refugees receive electricity for only half as long. Similarly, while around 70 percent of host community members who receive electricity can charge their mobile phones using electricity, only half as many percent of refugees can charge their mobile phones. On the other hand, except for South Sudanese refugees, a higher percentage of refugees have currently access to electricity as compared to before being displaced. This finding is opposite to the trends in housing, where the situation before displacement was generally better than during displacement. Thus, improved access to electricity points to the relative good quality of Ethiopia’s refugee camps. 25 Figure 18: Source of lighting 100 80 Percentage of population 60 40 20 0 Richest Quintile Host community Poorest Quintile Q2 Q3 Q4 Before displacement Before displacement Before displacement Before displacement Before displacement Woman headed Current Current Current Current Current Man headed Refugee South Somalis Eritreans Sudanese Sudanese Overall Refugee No Lighting Grass / Firewood Lamp / Candle / Torch Solar Power / Biogas Electricity Source: Author’s calculations based on Skills Profile Survey 2017 3.4. Service Delivery: Water and Sanitation, Health, and Education 33. Access to services for refugees (water and sanitation, health, and education) compares, or in some cases even exceeds, access to the same services for host community members, pointing to a high quality of camps. Trends in service delivery run opposite to poverty incidence, food insecurity, and housing situation, all of which saw refugees faring worse than host communities. Both refugees and host community members have nearly complete access to improved water sources 35 (98 and 91 percent, respectively), which for refugees is better than it was in their respective countries of origin (62 percent, Figure 19). Refugees have also far better access to improved sanitation facilities (69 percent) as compared to host community members (37 percent) and as compared to their access before displacement (28 percent, Figure 20). 36 Similarly, refugees in Ethiopia have greater access to health services (in particular, hospitals) compared to host communities. For example, the majority of refugee women gave deliveries at hospitals (72 percent) or maternity clinics (18 percent) during the last 24 months, compared to women among host community who gave birth in hospitals (36 percent) and maternity clinics (52 percent). 35 Improved water sources are piped water supply into the dwelling; piped water to a yard/plot; a public tap/standpipe; a tube well/borehole; a protected dug well; a protected spring; and rainwater. Unimproved water sources are unprotected dug well; an unprotected spring; a cart with a small tank/drum; a water tanker-truck; and surface water. WHO 2006. 36 As improved sanitation is defined as household having some type of flush toilet or latrine, or ventilated improved pit or composting toilet, provided they are not shared, the additional criteria of not sharing the toilet with other households reduces the percentage of population with access to improved sanitation. Following this definition, a lower percentage of refugee population, 42 percent, have access to own improved sanitation i.e., do not share their toilet with other households. WHO 2006. 26 Percentage of population 0 20 40 60 80 100 Percentage of population Refugee - before displacement 0 20 40 60 80 100 Refugee - current Refugee - Before displacement Overall Refugee - current Host community Overall Host community Man Head Man head Woman Head Woman head Poorest Quintile Poorest Quintile Q2 Unimproved sources of water Q2 Q3 27 Q3 Q4 Q4 Richest Quintile Improved sanitation facility - Not shared Richest Quintile Refugee Improved sanitation facility - Shared/Not shared Before displacement Figure 20: Access to Improved Sanitation Refugee Figure 19: Access to Improved sources of water South Improved Sources of Water Current Sudanese Before displacement Author’s calculations based on Skills Profile Survey 2017 South Before displacement Current Sudanese Source: Author’s calculations based on Skills Profile Survey 2017 Somalis Current Before displacement Before displacement Somalis Current Current Before displacement Before displacement Eritreans Current Current Eritreans Sudanese Before displacement Current Sudanese Source: 34. Overwhelmingly, refugee children in Ethiopia attend primary school (79 percent), whereas enrolment rates for secondary school are remarkably low (13 percent). Refugee children have higher net primary enrollment rates than the national average in their countries of origin. In fact, net primary enrollment rates for refugee children (79 percent) are even slightly higher than net primary enrollment rates for children in the host communities (Figure 21). In contrast, net secondary enrollment rates for refugee children are lower both as compared to net enrollment rate for children of host community households and as compared to net secondary enrollment rates in their respective countries of origin (except for South Sudan). In addition, a majority (82 percent) of secondary age students who are not in secondary school are attending grades in primary school. While forced displacement clearly contributes to school disruption, the fact that also host communities have a similar percentage of children of secondary age attending primary school (65 percent) suggests that there are structural barriers to secondary school attendance in Ethiopia other than displacement.37 Figure 21: Net primary and secondary enrollment rate in comparison with national enrollment rates 100 Percentage of primary and secondary 80 60 40 age children 20 0 Refugee Girls Eritreans Host Community Boys South Sudanese Somalis Sudanese Overall Refugees Net primary enrollment Net secondary enrollment National - Primary National - Secondary Source: Author’s calculations based on Skills Profile Survey 2017, World Bank 2018b, MICS 2014 (Sudan), HFS South Sudan Wave 3 (2016) and HFS Somalia Wave 1 (2016) 35. Most refugee children between 6 and 18 years of age (72 percent) attend schools run by NGOs, but there is ample difference across the four refugee groups.38 While nearly all South Sudanese refugees attend NGOs-run schools (96 percent), most Somalis (65 percent) and Sudanese (57 percent), and approximately 1 in 3 Eritreans attend Ethiopian public schools. This estimate points to not only differences in terms of camp management and dynamics of displacement (i.e. South Sudanese represent the latest arrival of refugees and are the largest group), but also variation in attaining durable solutions by the different refugee groups. Durable solutions entail an incorporation of refugee children into the government or local school system. Thus, as far as education is concerned, Somalis and Sudanese are the two groups that are faring best. 36. Proximity to services also represents an important measure to assess quality of service delivery. In fact, the longer the time incurred to access services (i.e. proximity to water sources, health facilities etc.) the smaller the rate of 37 Interestingly, Eritrean children faced most difficulty to obtain documents to enroll in schools: 20 percent of Eritrean children between the ages of 6 and 18 who were currently out-of-school cited a lack of documentation as the reason for not attending school. This finding may be counterintuitive as Eritreans are the most integrated among refugee groups in Ethiopia on a number of indicators. Nonetheless, it could be the case that other refugees are not be required to produce documents to enroll into school (e.g. school is inside the refugee camp), whereas Eritreans who are more integrated and have more freedom of movement may have to go through more elaborate bureaucratic processes to enroll into school. A higher degree of integration may then involuntarily come with more barriers to access services (in this case education) 38 This contrasts with children in host community households who overwhelmingly attend government schools. 28 access to the specific service; limited access also puts strains on households and limits the time available for other economically productive activities. In Ethiopia, refugees have shorter access to services while in displacement than before being displaced, resulting in some improvement in quality of life. Such improvement is related to the quality of refugee camps in Ethiopia, which are meant to be temporary. Thus, greater proximity cannot be taken as a sustainable indicator of improvement in the life of refugees. Irrespective of income status, nationality and gender of household head, it takes refugees approximately the same amount of time than host community members to reach the nearest source of water, hospital and primary school (Figure 22, Figure 23, Figure 24). Figure 22: Mean time (minutes) taken (one-way) to fetch water 100 Mean time (minutes) taken to fetch water 80 60 40 20 0 Refugee - before displacement Man Head Richest Quintile Host community Poorest Quintile Q2 Q3 Q4 Refugee - current Before displacement Before displacement Before displacement Before displacement Current Current Current Current Woman Head South Somalis Eritreans Sudanese Sudanese Overall Refugee Source: Author’s calculations based on Skills Profile Survey 2017 29 Time (minutes) to walk one way to primary school Time (minutes) to walk one way to hospital 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 Refugee - before displacement Refugee - before displacement Refugee - current Refugee - current Overall Host community Overall Host community Male Head Man head Female Head Woman head Poorest Quintile Poorest Quintile Q2 Q2 Q3 Q3 Q4 30 Q4 Richest Quintile Richest Quintile Refugees Before displacement Refugees South Before displacement Current Sudanese South Sudanese Current Before displacement Somali Figure 23: Time (minutes) to walk one way to health facility Before displacement Figure 24: Time (minutes) to walk one way to primary school Current Source: Author’s calculations based on Skills Profile Survey 2017 Source: Author’s calculations based on Skills Profile Survey 2017 Somali Current Before displacement Before displacement Eritreans Current Eritreans Current Before displacement Current Sudanese Before displacement Sudanese Current 4. Employment and Livelihood 37. An overwhelming majority of refugees relies on aid (including cash, food and non-food) from the Government or humanitarian organizations for their livelihood. While refugees relied primarily on agriculture and wages and salaries for their livelihood before displacement, currently 83 percent of refugees obtain livelihood through aid (Figure 25). This is in stark contrast to the host populations, who mainly derive their livelihood from agriculture (39 percent), and wages and salaries (28 percent), with 1 in 5 being occupied in services or retail. Among refugees, richer households and households headed by men are slightly more likely to depend on wages and salaries for their livelihood (each group with an estimated 15 percent), than the poorest quintile and women-headed households (2 and 4 percent, respectively). With respect to nationality, all groups overwhelmingly depend on aid except for the Somalis: while 66 percent of Somali refugees are aid dependent, 1 in 5 obtains their livelihood from salaries and wages and 15 percent from other sources (including services, retail and agriculture). Thus, Somalis’ greater economic self-reliance compared to other groups may explain higher rate of home ownership and housing conditions enjoyed by Somali refugees compared to Sudanese, South Sudanese, and partially Eritreans. Figure 25: Source of livelihood currently and before displacement 100 Percentage of households 80 60 40 20 0 Refugee - Current Current Current Current Current Host community Poorest Quintile Richest Quintile Refugee - Before displacement Man Headed Woman Headed Before displacement Before displacement Before displacement Before displacement Q2 Q3 Q4 South Somali Eritrean Sudanese Sudanese Overall Refugee Agriculture Manufacturing Retail Services Wages and salaries Income from assets Aid Remittances Source: Author’s calculations based on Skills Profile Survey 2017. 38. High dependency on aid is a result of low labor force participation rates among refugees. Only 22 percent of working-age refugees (15-64) are currently employed, compared to 66 percent of the host community working-age population who is employed (Figure 26). Among refugees, working age population in the richest quintile is more likely to be employed than the poorest (38 versus 10 percent), confirming the trend that refugees with higher income are more economically self-reliant. Refugees in Ethiopia are not officially allowed to work, which explains the low labor force participation rate as over 70 percent of them are inactive (neither employed, nor unemployed, nor actively looking for employment). In contrast, 1 in 3 host community members in working age is inactive. On a positive note, 27 percent of all inactive refugees are currently enrolled in school or college, and are thereby developing skills. There are significant differences in the proportion of working age population inactive but enrolled in schools across nationalities. Unexpectedly, Sudanese and South Sudanese – the two groups with higher poverty incidence – have higher rates of inactive refugees who 31 are attending school (41 and 34 percent, respectively) compared to Somali and Eritreans (12 and 11 percent, respectively), both of whom instead have over 60% rate of inactivity without school enrollment. For Eritreans, this trend may be explained by the fact that most members of this group only stay in Ethiopia on a temporary basis, and thus do not invest in education – an explanation that cannot be applied to Somalis. On the other hand, Somali refugees have the highest rate of employment among refugees (28 percent). 39. There are important differences between men and women participating in the labor force, which are likely to persist in the foreseeable future. Men of working-age population have higher rate of employment than women (27 versus 19 percent, respectively). More importantly, most women are inactive and are not pursuing an education opportunity (55 percent) compared to men in the same situation (31 percent). In fact, this trend is true for all 4 nationalities of refugee groups, ranging from 47 percent of South Sudanese refugee women who are inactive and not attending school or college to 72 percent of Eritrean refugee women in the same condition. While there are more women than men who are not currently part of the labor force, it is conceivable that more women than men will not be part of it in the future either, because they are not working toward attaining an education. Thus, there are specific gender-related barriers to women in accessing the labor force both in the present and in the foreseeable future. Figure 26: Labor force participation and employment status 100 Percentage of working age population 80 60 40 20 0 Women Women Women Women Women Poorest Quintile Richest Quintile Refugee Host community Overall Overall Overall Overall Q2 Q3 Q4 Men Men Men Men Men South Somali Eritrean Sudanese Sudanese Overall Refugee Inactive and Not Enrolled in School/College Inactive and Enrolled in School/College Employed and Enrolled in School/College Employed and Not Enrolled in School/College Unemployed Source: Author’s calculations based on Skills Profile Survey 2017 32 40. On average, refugees have similar levels of educational attainment comparable to those of host populations. Both refugees and host community population have a large and comparable percentage of working-age population without any education (41 percent and 44 percent, respectively, Figure 27). However, Ethiopians in the host community are more likely to have secondary and university education (19 percent and 6 percent respectively) as compared to refugee population (13 and 2 percent, respectively). Also, among refugees, a higher percentage of men have some education as compared to women across all nationalities. There are significant differences between the educational attainment of refugees of different nationalities, with Eritreans being the most likely to have some education, followed by South Sudanese. Somali refugees have the lowest educational attainment but as observed above have the highest labor force participation and employment rates. Figure 27: Highest educational attainment for working age population 100 Percentage of working age population 90 80 70 60 50 40 30 20 10 0 Women Women Women Women Women Poorest Quintile Richest Quintile Refugee Host community Overall Overall Overall Overall Q2 Q3 Q4 Men Men Men Men Men South Somali Eritrean Sudanese Sudanese Overall Refugee No education Primary & Intermediate Secondary University Technical and Vocational Education Others Source: Author’s calculations based on Skills Profile Survey 2017 41. Barriers to labor force participation and employment for refugees need to be removed. 28 percent of the working-age population does not participate in the labor force because they are enrolled in school or college and are therefore not currently active in the labor market (Figure 28). Another 21 percent of working-age refugees (predominantly women) are not participating in the labor force because they are taking care of their households. Even those who enter the labor market face barriers in securing employment. Around half of the unemployed population report a lack of regular work opportunities and around a quarter of the unemployed population feel that they lack adequate skills and experience (Figure 29). Around half of the unemployed have been looking for job for more than a year and almost all (97 percent) of the refugee population not participating in the labor force in the last 7 days was not engaged in any economic activity for more than a year. These protracted periods of displacement and economic inactivity potentially erode skills and make it harder for refugees to find employment. Refugees report that lack of connections and discrimination also played a role in their unemployment. 33 Figure 28: Reasons for not participating in the labor force (Refugees) Figure 29: Reasons for not securing employment (Refugees) 30 35 Percentage of population inactive in Percentage of unemployed population 25 30 25 20 20 15 15 labor force 10 10 5 5 0 0 documentation/work… Too young / too old for In school / college Taking care of household Does not expect to find a Not allowed to work as Ill / sick Limited or irregular work opportunities Lack of experience religious discrimination Lack of or inadequate Lack of family/clan or political connections Lack of work Ethnic / political / Lack of proper opportunities refugee skills / family work job Source: Author’s calculations based on Skills Profile Survey 2017 Source: Author’s calculations based on Skills Profile Survey 2017 42. Refugees need access to productive assets to develop livelihood opportunities for themselves. Before displacement, most refugees had access to agricultural land, at least one productive asset 39 and at least one livestock. 40 Currently, a very small percentage of refugees have access to any of the above (Figure 30). Refugees of all nationalities have faced a significant decline in their access to agricultural land, productive assets and livestock since displacement and South Sudanese refugees have faced the greatest decline. Only Somali refugees fare slightly better – a trend that mirrors Somalis’ higher labor force participation and lower dependence on aid compared to other nationalities. While 51 percent of Somali refugees held any livestock before displacement, 48 percent of them still hold at least one livestock. However, the quantity of livestock held has declined significantly: while Somali refugees held 6.68 Livestock Units 41 on average before displacement, the current livestock that is held is estimated at 0.2 Livestock Units on average. This lack of access to productive assets further limits the ability of refugees to create employment opportunities for themselves and hampers self- reliance. 39 Productive assets include Moefer and Kember, Axe/sickle, Plough, Weaving equipment, Builder’s equipment, Carpenter’s equipment, welding equipment, wood cutting equipment, block production equipment, refrigerator, private car and Bajaj. 40 Livestock include cattle, horses, donkey/mules, pigs, sheep, goats, poultry, camels and beehives. 41 Livestock units are used for aggregating the numbers of different categories of livestock for regional and global comparisons and are obtained by converting the body weight into the metabolic weight. The livestock unit coefficients used are those corresponding to the regions of Near East North Africa: Cattle – 0.70, Buffalo - 0.70, Sheep - 0.10, Goats – 0.10, Pigs – 0.20, Asses – 0.50, Horses – 0.40, Mules – 0.60, Camels – 0.75, Chickens – 0.01 (Chilonda and Otte 2006). 34 Figure 30: : Access to productive assets 100 Percentage of population 80 60 40 20 0 Current Current Current Current Current Before displacement Before displacement Before displacement Before displacement Before displacement Refugee Host Community South Sudanese Somali Eritreans Sudanese Overall Refugees Agricultural land Productive assets Livestock Source: Author’s calculations based on Skills Profile Survey 2017 43. Refugees who are employed are working primarily as salaried laborers or are engaged in non-farm self- employment. While self-employment in agriculture was a predominant activity for most refugees in their countries of origin, especially Somalis and Sudanese, only 7 percent of refugees are currently engaged in own-account farming (Figure 31). There are remarkable differences between refugee groups according to their nationality. More than half of all South Sudanese refugees that are in employment are merely helping in the non-farm businesses of their families. Instead, the main employment activity of the other three groups is salaried labor, followed by non-farming self-employment. 35 Figure 31: Primary employment activity 100 Percentage of employed population 90 80 70 60 50 40 30 20 10 0 Current Current Current Current Current Current Current Poorest Quintile Richest Quintile Host community Before displacement Before displacement Before displacement Before displacement Before displacement Before displacement Before displacement Q2 Q3 Q4 Refugee Men Women . South Somali Eritrean Sudanese Sudanese Overall Refugee Unpaid apprenticeship or training Farming or hunting or fishing at own account / help on family farm Help in any kind of non-farm business owned by the household (Non farm) Own-account work, for example as a business owner Salaried labor or labor paid in kind (including agricultural salaried job) Source: Author’s calculations based on Skills Profile Survey 2017 44. Both salaried and self-employed refugees are predominantly engaged in the services sector. A remarkable 71 percent of employed refugees are working in the services sector with only 12 percent working in agriculture, 10 percent in manufacturing and 6 percent in education. This contrasts significantly with host community population as 48 percent of employed host community population is currently employed in services and 41 percent in agriculture (Figure 32). When asked details about the work refugees and host community members would want to do in the future, an overwhelming majority of both refugees and host community members aspire to continue or start working in the services sector. After services, more men amongst refugees aspire to work in the manufacturing sector as compared to women (16 percent versus 7 percent). Similarly, after services, 1 in 4 Eritreans aspire to work in the manufacturing sector, an aspiration which is not largely shared by refugees of other nationalities. 36 Percentage of employed population 0 10 20 30 40 50 60 70 80 90 100 Before displacement Current Aspiration Refugees Overall Current Aspiration Host Agriculture Community Before displacement Current Men Aspiration Manufacturing Before displacement Current Women Aspiration Services 37 Before displacement Current Aspiration Figure 32: Sector of employment Education South Sudanese Refugees Before displacement Current Somali Aspiration Source: Author’s calculations based on Skills Profile Survey 2017 Before displacement Current Public sector / defence Aspiration Eritrean Before displacement Current Sudanese Aspiration 5. Social Cohesion, Public Participation, Safety and Security 45. Both refugees and host community perceive that they are in good relationship. Only a relatively small percentage of both groups perceive that relations are bad: 17 percent of refugees and 11 percent of host communities report so (Figure 33). Somalis have much better relations with host community members followed by Eritreans and Sudanese, although over 85 percent of all 3 groups report to have good relations. South Sudanese refugees are the group reporting the lowest percentage in terms of good relations, perhaps because they are the newest group and have not yet adapted: 1 in 3 thinks that refugees and host community population have bad relations with each other. 42 The more positive perceptions by Somalis tie with their longer permanence as refugees, common ethnic identity and similar clan system. These commonalities entail a higher degree of integration, from economic self-reliance and higher participation in the labor force to better housing condition and lower poverty incidence. In addition, the higher the income level for all refugees the more positive relations are, ranging from 67 percent of poorest quintile who report good relations to 96 percent of the richest quintile who believe relations are good. Women are more likely than men to report negative relations: nearly 1 in 4 women perceive relations are not good, compared to 7 percent of men. Figure 33: Inter-personal relations between refugees and host community Refugees and Ethiopians have good relations with each other 100% Percentage of population 50% 0% Poorest Quintile Richest Quintile Host community Eritreans Man Headed Somali Refugee - current Woman Headed South Sudanese Sudanese Q2 Q3 Q4 Overall Refugee Strongly disagree Slightly disagree Neither agree nor disagree Slightly agree Strongly agree 42 Another reason could be that there is some saturation effect: problems become apparent when more and more refugees arrive. Thus, late-comers are perceived in negative terms as they are linked to these experiences. 38 Source: Author’s calculations based on Skills Profile Survey 2017 46. Host communities’ sentiments towards refugees are remarkably varied when the 3 surveyed regions that host refugees are considered (Figure 34, Figure 35). In the Somali region, relations between host communities and Somali refugees are the best. Over 90 percent “agree” that relations are good, 70 percent of which “strongly agree” that relations are good. This confirms that Somali refugees are well integrated with their host community population. In the Tigray and Afar regions that host Eritreans the picture is slightly more mixed, although relations are overall perceived positively. In contrast, host community sentiments in Benishangul Gumuz region (which hosts 75 percent Sudanese and 25 percent South Sudanese refugees) are not very positive: more than 1 in 3 reports that relations are not good, and only 1 in 4 “strongly agrees” that they are good. In addition, approximately 60 percent of host community members in Benishangul Gumuz report that Ethiopians feel that refugees should be repatriated, that crime has risen, and it is more difficult to secure employment. Figure 34: Host Community Feelings: Good Relations Refugees and Ethiopians have good relations with each other 100 Percentage of host population 80 60 40 20 0 Somali Tigray/Afar Beninshangule Strongly agree Slightly agree Neither agree nor disagree Slightly disagree Strongly disagree Source: Authors’ calculations based on Skills Profile Survey 2017 Figure 35: Host Community Feelings: Other Measures 100 Percentage of host population 80 60 40 20 0 Somali Tigray/Afar Beninshangule Somali Somali Tigray/Afar Beninshangule Tigray/Afar Beninshangule Ethiopians want refugees to return to their The arrival of refugees has made it more The arrival of refugees has homes difficult for people in this community to brought insecurity to the area find work Strongly disagree Slightly disagree Neither agree nor disagree Slightly agree Strongly agree Source: Authors’ calculations based on Skills Profile Survey 2017 47. Public participation of refugees is not high as 1 in 2 refugees never participates in the public sphere nor has interactions with community leaders (Figure 36, Figure 37). However, there are remarkable differences between refugee 39 groups according to nationality. Eritreans and Sudanese have a 90 percent rate of participation in public meeting, and nearly 80 and 60 percent rate of interaction with community leaders, respectively. Somalis and South Sudanese report an opposite picture: over 75 percent of the former and nearly 50 percent of the latter never participate with similar trends in interaction with leadership. The low public participation of Somali refugees is striking: by several measures, Somalis are the better off group and the most integrated one economically. This finding shows how the relationship between community participation and social and economic integration is not uniform, but multifaceted. Host community have higher rates of participation than refugees as only 1 in 5 fails to participate publicly and 30 percent never interacts with community leaders. In terms of gender, both men and women equally participate in the public sphere. Figure 36: Participation in public meetings in the last 12 months Figure 37: Interaction with community leader in the last 12 months 100% 100% Percentage of population Percentage of population 50% 50% 0% 0% Host community Man headed Woman headed Eritreans Host community Eritreans Somali Somali South Sudanese South Sudanese Refugee - current Sudanese Refugee - current Sudanese Man Head Woman Head Overall Refugee Overall Refugee Never 1 to 4 times 5 or more times Never 1 to 4 times 5 or more times I am a community leader Source: Authors’ calculations based on Skills Profile Survey 2017 Source: Author’s calculations based on Skills Profile Survey 2017 48. Over half of the refugee population in Ethiopia feel very safe or moderately safe in the refugee camps, while host communities feel overwhelmingly safe. While 30 percent of the refugees feel neither safe nor unsafe, 14 percent of the refugee population feel unsafe at home or walking around in the refugee camps during day or night (Figure 38). 43 This sense of security varies significantly across gender of household head, household income and nationality. A higher percentage of refugees living in households headed by women feel unsafe as compared to refugees living in households headed by men (19 percent vs. 5 percent). Similarly, wealthier refugees feel safer. According to nationality, differences are striking. A large majority of Somali refugees (77 percent) feels very safe in the camps, compared with 20 percent of Eritreans, 9 percent of Sudanese and only 1 percent of South Sudanese. While a majority of Eritrean and Sudanese refugees also feel moderately safe in the camps, a large percentage of South Sudanese refugees feels either unsafe (26 percent) or neither safe nor unsafe in the camps (49 percent). Figure 38: Feelings of safety and security 43 Skills Profile Survey (2017) asked three questions: “In general, how safe from crime and violence do you feel when you are alone at home?”, “How safe do you feel when walking around alone after dark?”, “How safe do you feel walking around during the day?”. A combined scale for safety was created using these three questions. Cronbach alpha for the scale was 0.70. 40 Feeling of safety 100% Percentage of population 50% 0% Poorest Quintile Host community Eritreans Man Headed Somali South Sudanese Refugee - current Woman Headed Sudanese Q2 Q3 Q4 Q5 Overall Refugee Very unsafe Moderately unsafe Neither safe nor unsafe Moderately safe Very safe Source: Author’s calculations based on Skills Profile Survey 2017 41 6. Movement and Return Intentions 49. Based on their nationality, refugees have different ideas about their preferred durable solution. 44 Almost half of the refugee population does not plan to move from their current locations in camps at any point in the future – although they do not specifically mention to wanting to integrate locally (Figure 39). Nearly 1 in 3 reports wanting to be resettled, while 16 percent prefer to return. Most South Sudanese refugees intend to stay in their current location followed by Sudanese and Somalis, while only 1 in 4 Eritreans want to stay where they are. In contrast, Sudanese refugees’ preferred solution is to return to Sudan (1 in 2), while only a small minority of the other 3 groups intend to return. As it is evident from the literature, most Eritreans want to move to another country (over 60 percent), an intention shared by 40 percent of Somalis, 22 percent of South Sudanese and by virtually no Sudanese (3 percent). The U.S. is the ideal destination for 70 percent of the refugees who want to move to a new country. In addition, about 60 percent of all refugees would still want to move to a new country even if given the right to settle and work freely in Ethiopia was granted to them. The nationality of the refugee group plays an important role here too in differentiating refugees’ intentions: the rate of those who would prefer to resettle anyway, even in presence of local integration mechanisms ranges widely from 4 in 5 Somalis to 1 in 5 Sudanese. Figure 39: Movement and Return Plans of Refugees 100 Percentage of population 90 80 70 60 50 40 30 20 10 0 Eritrean Somali Overall South Sudanese Stay here Move to another camp Move out of camp Return Sudanese Move to a new country Source: Authors’ calculations based on Skills Profile Survey 2017 50. Refugees highlight the main reasons for their preferred intention and what factors may affect their decision. With respect to return, the same reasons why refugees were displaced in the first place from their respective country of origin apply as reasons for not wanting to return. These include armed conflict and security concerns, political or identity- based persecution, as well as lack of employment opportunities (Eritreans) and drought and associated famine (Somalis). Irrespective of nationality, the refugees who want to stay list the following reasons not to move: better security, better access to education and health services, and access to humanitarian aid (Figure 40). In contrast, family reasons and other economic indicators do not feature prominently. Concerning refugees who want to move from the current location, lack of employment opportunities (40 percent) is the main reason for wanting to move, followed by lack of access to home/land/livestock (26 percent) and to humanitarian aid (17 percent). Refugees also mention the political situation and the security conditions on transit routes as the most important information they would need to make an informed decision. Finally, refugees largely point to security as the most important factor to settle in the preferred location (Figure 41). Figure 40: Reasons to Stay 44 Refugees were not specifically asked about their preferred durable solution, but only what their intention was. Conceivably, reported intentions can be framed as refugees’ preference for durable solution although this caveat should be noted when analyzing and using this data. 42 60 Percentage of population 50 40 30 20 10 0 Overall South Sudanese Somali Eritrean Sudanese Better Security Better Access to Home/Land/Livestock Better Access to Education and Health Services Better Access to Livelihood Family Reasons Access to Humanitarian Aid Source: Authors’ calculations based on Skills Profile Survey 2017 Figure 41: Main support needed to settle in preferred location Security 90 Mine clearance 80 Access to original house/land Percentage of refugee Provision of a new shelter/housing 70 Access to a farm plot and / or grazing land 60 Agricultural tools and seeds/livestock population 50 Funds/productive assets for re-establishing business Skills training for job 40 Connections for job 30 Access to markets Access to educational markets 20 Access to health services 10 Access to water and / or sanitation Access to aid 0 Transportation arrangement Overall South Somali Eritrean Sudanese Regrouping of family Sudanese Renewal/replacement of documentation Source: Authors’ calculations based on Skills Profile Survey 2017 43 Conclusion 51. Over the last few years, Ethiopia witnessed a dramatic increase of refugees (and IDPs), which led to a shift in its approach to forced displacement. Refugees’ number has risen tenfold in the last decade, to almost 1 million refugees. Nearly half million South Sudanese refugees settled in Ethiopia in the last 5 years alone. IDP trends are booming too: from an official 258,000 in 2016 the number of conflict-induced IDPs climbed up to an estimated 1.1 million. While IDPs are not part of the present case study, nonetheless Ethiopia faces compounding humanitarian crises and protracted displacement that put a strain on domestic actors’ coping capacities and question the sustainability of the current reception system. As a result, the Government of Ethiopia unveiled a strategic plan (“nine pledges”) to address the development aspects of the refugee situations in the medium to long term. Currently under implementation, the multi-sectoral initiative aims to improve the material condition of refugees, to expand their rights, and to support synergies and local integration mechanism with host communities. 52. The micro-level study on refugees in Ethiopia and host communities is meant to support the implementation of the new strategy. It provides comprehensive findings on several measures of poverty at the household level, putting in relationship different groups of refugees in Ethiopia and their respective host community. The analysis found that not only refugees witnessed trauma and life disruption (e.g. family separation), and incurred into material loss during displacement (i.e. land, livestock, and assets), but also their standard of living, livelihood and access to services’ prospects are currently highly dependent on aid. Uncertainty and inability to plan for the future represent the common denominator among refugees, which is the reason why a conducive legal framework and related policies to enable refugees’ self-reliance are key. 53. Eritreans, Somalis, South Sudanese and Sudanese were displaced due to different drivers related to conflict and fragility, and each group experiences different displacement dynamics while in Ethiopia. While all groups except Eritreans reported that armed violence and a general sense of insecurity were the chief drivers of displacement, specific conflict drivers include localized conflicts and a fallout in law and order (Somalis and Sudanese), environmental degradation leading to strain in resources (Somalis), political persecution (Eritreans), and a full blown civil war (South Sudanese). Timeline of displacement is quite different from group to group: while South Sudanese are the newest group (as they are also coming from the newest country), the other three groups have had some repeated patterns of displacement in the last three decades. Geographically, refugees are compartmentalized according to nationality: each group settled in the border area in proximity of the respective country of origin. As a result, refugees have specific development needs that should be put in relationship with the respective Ethiopian region of displacement. Border regions in Ethiopia are traditionally poorer and more marginalized than the center, although the economic gap center-periphery has been closing during the last decade. 54. Overall, Eritreans are better off compared to other refugee groups, with indicators that are similar to host community ones. In fact, according to nationality, refugees fare differently with respect to standard of living, livelihood and employment, and ties to host communities. Among refugee groups, Eritreans are the ones that enjoy more rights compared to others, and, as a result, display higher standard of living and much lower poverty rates. In turn, less than 30 percent Eritreans experience high food insecurity, while the other three groups have rates ranging between 60 and over 80 percent. Similarly, trends in housing condition and overcrowding highlight Eritreans’ better situation with respect to other groups. On the other hand, South Sudanese are the poorest group on many indicators. Arguably, the length of displacement plays an important role when assessing not only refugees’ standard of living, but also social capital and relationships with host communities. Longer interactions refugees-host communities seem to predict better relationships going forward. 55. Indicators on livelihood and access to services highlight the nearly complete dependence of refugees from aid. Over 90 percent of Eritrean, South Sudanese and Sudanese households rely on aid as their greatest source of livelihood, while Somalis are the least aid-dependent and they have the lowest rate of food consumption coming from aid too. Arguably, among the four refugee groups, Somalis are the most integrated group with host communities. From higher rates of labor force participation to slightly higher rates of home ownership to stronger feelings of safety and security and better host populations’ perceptions, Somalis display specific displacement dynamics that are different from those of other groups. Access to services for all refugees compares and, in some cases, is even better than host community’s – a situation that underlines the relative high quality of camp management vis-a-vis domestic service delivery mechanism. These findings 44 directly feed into the need to support policies of job creation, enhanced and sustainable access to markets and services, and conducive regulatory environment and governance, among others. Virtuous practices – that organically take into consideration refugees as well as host communities – are functional to economic self-reliance and to move away from dependency on external sources of aid. 56. Ultimately, refugees showcase perceptions of uncertainty which leads them to inability to plan for the future. As the main driver of displacement was personal security, likewise, all refugee groups regard security as the utmost priority when making considerations about future intentions, prospects and/or durable solutions. Overall, 1 in 2 refugees prefers to stay in the current location, which is an indication of their inability to plan. Interestingly, South Sudanese – the poorest among the four – have the highest percentage of respondents who do not want to move. This fact suggests that the trauma they have suffered from is still fresh and ongoing, and that stay put is the most favorable option despite poor standard of living. Eritreans and Somalis have, instead, large percentages of refugees who want to move to a new country, while over 1 in 2 Sudanese wants to return to Sudan. These contradicting intentions – which run opposite to standard of living across the four groups – are indicative of the fundamentally different situations that refugee groups currently face. 57. More analysis is needed to understand the ad hoc development needs and durable solutions prospects of the four refugee groups against host communities. Given the above findings, it is key to better understand each refugee groups’ dynamics vis-à-vis the respective host population, who, in turn, live in regions of Ethiopia that have remarkable differences and are far away from one another. The Skills Profile Survey (SPS) collected data on refugees and host communities in Tigray and Afar, Somali, and Benishangul Gumuz. The present paper analyzed and compared refugee groups among them, and against an average host community. There is scope to zoom in on each specific host community in order to make more meaningful comparisons between each refugee group and its respective host community. Such findings would give more explanatory power to each of the four displacement situations in Ethiopia, and offer more specific policy insights. 45 Appendix I: Skills Profile Survey (SPS) 2017 The Skills Profile Survey (SPS) 2017 is a household survey that was administered in 2017. It aimed to survey refugees of four main nationalities – South Sudanese, Somalis, Eritreans and Sudanese – living in camps in Ethiopia, and the Ethiopian host communities living in the proximity of refugee camps. 45 The list of refugee camps, sites and locations provided by UNHCR-Ethiopia as of January 2017 was used as the sample frame. Of all refugee households 46 in Ethiopia, 33 percent live out of camps. An overwhelming majority (85 percent) of these out of camp refugees are Eritrean refugees who either spontaneously settled in different locations of Ethiopia, or who benefited from the 2010 ‘Out of Camp’ policy and as a result moved to Addis Ababa or other locations outside camps. The refugee households living out of camp were excluded from the sample frame because tracing these households for this survey exercise was not feasible. The SPS is therefore only representative of refugees living in camps in Ethiopia. Table 3 below lists all the refugee camps in the sample frame. The sample frame was then divided into four strata based on four regions (main refugee group in parenthesis): Tigray Afar (Eritreans), Gambella (South Sudanese), Benishangul Gumuz (Sudanese), and Somali (Somalis). Each region hosts a predominant majority of one refugee nationality leading to an implicit stratification based on nationality. The sample design uses a multi-stage stratified random sample. Camps in each stratum were divided into enumeration areas (EAs) of 150 by 150 meters using GIS technology. The number of EAs to be selected from each camp was obtained proportional to the size of the camp. In this way, all the camps in the sample frame were selected in the sample and were surveyed. Within camps, EAs were selected using equal probability to make up the required number of EAs for that camp. In total, 82 enumeration areas were selected from each stratum. All the households in the selected EAs were listed and 12 households were randomly selected and surveyed per enumeration area making up to a total of 900 refugee households per stratum. Table 1: Number of refugee and host community households interviewed by stratum Stratum Tigray Afar Gambella Benishangul Gumuz Somali Total Refugees 894 439 1423 871 3627 Host Community 412 0 975 303 1690 For host community sample, all households within a 5-km radius of a camp were classified as host community households. Areas within 5 km radius of camps were divided into EAs of 300 by 300 meters using GIS technology. Of these, EAs marked as residential by Open Street Maps were included in the sample frame. EAs within a stratum were then selected using proportional probability sampling with the probability of selection of an EA equal to the area of the Enumeration Area outside the camp. In total, 42 EAs were selected for each stratum. Like EAs within camps, all the households in the EAs selected for host community sampling were listed and 12 households were selected randomly and surveyed per EA making up to a total of 500 host community households per stratum. Due to security concerns, major revisions were made to the sample during fieldwork. Enumerators in Gambella region faced repeated security threats and could survey only 439 of the intended 900 refugee households in the region. As the survey team was withdrawn from Gambella region, host community in Gambella region was not surveyed at all. The remaining interviews with refugees in Gambella region were substituted by oversampling enumeration areas in Benishangul Gumuz, as 25 percent of the refugee population in this region is South Sudanese. Similarly, in early September 2017, violent conflict in Oromia and Somali regions escalated rendering some of the camps in Somali stratum inaccessible. The enumeration areas of Jijiga sub-region were replaced by enumeration areas in non- violent areas of Somali stratum. Also, as most refugee camps are in remote areas with sparse host population, the final number of host households surveyed fell short of the original intended sample of 500 host households per stratum. However, 45 The case study looks specifically at refugees and does not include Internally Displaced Persons (IDPs) in the country. 46 Household is here defined as all people living in the same dwelling and sharing all meals and finance. 46 despite the changes in sample, the survey captured roughly similar number of refugee households of the four main refugee nationalities. Table 2: Sampled population by Country of Nationality Country Number of households Percentage of households in surveyed surveyed population South Sudanese 837 16% Somalis 871 16% Eritreans 893 17% Sudanese 1016 19% Ethiopians - Host community 1690 32% Other Country 10 0% Total 5317 100% Table 3: Refugee Camps in Sample Frame Nationality of Region Camp refugees Mai-Aini Adi Harush Tigray Shimelba Eritrean Hitsats Aysaita Afar Barahle Pugnido Kule Jewi South Gambella Okugo Sudanese Tierkidi Pugnido II NGUENYYIEL camp (new) Sherkole 75% Sudanese, Bambasi Benishangul 25% South Gumuz Tongo Sudanese Tsore Ken-Borena Kebribeyah Somali Somali Aw-barre Sheder 47 Bokolmanyo Melkadida Kobe Hilaweyn Buramino Source: UNHCR 48 Appendix II: Weights Sampling weights are used to make survey observations representative for sample. As the sample was divided by stratum, weights for refugee surveys are constructed to be representative of refugee population in the different regions and of the overall refugee population living in camps in Ethiopia. Similarly, weights for host population are constructed to be representative of the host households living within 5 km radius of refugee camps. The sampling weight is the inverse probability of selection. The selection probability P for a household can be decomposed into the selection probability P1 of the EA and the selection probability P2 of the household within the EA: (1) = 1 2 As refugee population in the different strata lived in different camps, the selection probability P1 of an EA k is calculated as the number of households within the EA divided by the number of households within the stratum multiplied by the number of selected EAs in the stratum � ∗ (2) 1 = � denotes the number of households in EA k (obtained by multiplying the percentage of camp area covered by the where EA with the number of households in the camp as information on number of households in an EA was not available prior to listing), K is the number of EAs selected in the corresponding stratum and N is the total number of households in the stratum. For host community sampling, as information on number of host households living within 5 km of camps in a stratum was not available, the selection probability of an EA for host sampling is calculated as the number of EAs selected divided by the total number of EAs in the stratum. (3) 1 = Where K is the number of EAs selected in a stratum and T is the total number of EAs in the corresponding stratum. Replacement enumeration areas were assigned the sampling weight of the enumeration area that they were replacing. Due to changes in sample during fieldwork the number of enumeration areas surveyed in each stratum differed from the original sample. The weights were therefore scaled at the end to correct for the change in the value of K. The selection probability P2 for a household within an EA k is constant across households and can be expressed as | | (4) 2 = where |H| is the number of households selected in the EA and nk denoting the number of listed households in EA k. Usually, the number of households per EA is 12 while a few exceptions exist due to invalid interviews. Sampling weights were scaled to equal the number of households per strata using the information for number of households provided by UNHCR. There was no source of information on number of host households living within 5 km distance of the camps. The weights for host community surveys were therefore not scaled. 49 References Asfah Gemechu, E. 2016. “The Recent Attack in Gambela and its Implications for Humanitarian Operations.” Horn of Africa Bulletin (March-April), Volume 28 Issue 2, Life and Peace Institute. Administration for Refugees and Returnee Affairs (ARRA). 2017. “Roadmap for the Implementation of the Federal Democratic Republic of Ethiopia Government Pledges and the Practical Application of the CRRF in Ethiopia.” Addis Ababa. Amnesty International. 2016. “Tackling the Global Refugee Crisis: From Shirking to Sharing Responsibilities.” New York: Amnesty International. Carter, B. and Rohwerder, B. 2016. “Rapid Fragility and Migration Assessment for Ethiopia.” Rapid Literature Review. Birmingham, UK: GSDRC, University of Birmingham. Central Bureau f Statistics (CBS), UNICEF Sudan. 2014, Multiple Indicator Cluster Survey 2014 of fieldwork, Key Findings. Khartoum, Sudan: UNICEF and Central Bureau of Statistics (CBS) Chilonda, P. and J Otte. Indicators to Monitor Trends in Livestock Production at National, Regional and International Levels. Livestock Research for Rural Development, v.18, n.8, 2006. Available at: http://www.lrrd.org/lrrd18/8/chil18117.htm Feyissa, D. 2014. “The Spillover Effect of South Sudan in Gambella.” Horn of Africa Bulletin (September-October), Life and Peace Institute. Girma, E. 2016. “Assessing the Impact of South Sudanese Refugees on the Host Communities of Itang Woreda.” Addis Ababa University. Human Rights Watch. 2017. “World Report 2018. Events of 2017.” New York IDMC. 2017. “Global Report on Internal Displacement.” Geneva. IDMC. 2018a. “Mid-year Update”. Available at http://www.internal-displacement.org/assets/country-profiles/Mid-Year- update-2017/20170818-Ethiopia-edited-MYF-.pdf [Accessed on March 22, 2018]. IDMC. 2018b. “Ethiopia.” Available at http://www.internal-displacement.org/countries/ethiopia/ [Accessed on March 28, 2018]. International Crisis Group. 2009. “Ethiopia: Ethnic Federalism and its Discontents.” Africa Report No. 153, Nairobi/Brussels. International Crisis Group. 2013. “Ethiopia: Prospects for Peace in Ogaden.” Africa Report No. 207, Nairobi/Brussels. International Institute for Strategic Studies (IISS). “IISS Armed Conflict Database: Ethiopia.” https://acd.iiss.org/member/conflicts.aspx?id=274&type=2&archived=False&page=7 [Accessed on March 23, 2018). IOM. 2018a. “Displacement Tracking Matrix (DTM) – Ethiopia. Round 10: March to April 2018.” IOM-DTM, Addis Ababa. IOM. 2018b. “Ethiopia 2018: Humanitarian Compendium.” https://humanitariancompendium.iom.int/appeals/ethiopia [Accessed on May 14, 2018]. Mallet, R., J. Hagen-Zanker, N. Majidi, and C. Cummings. 2017. “Journeys on Hold: How Policy Influences the Migration Decisions of Eritreans in Ethiopia.” Working Paper 506. ODI, London. OCHA. 2018. “Ethiopia: Conflict Displacement Situation Report #3.” April 17. Richards, S. and Bekele, G. 2011. “Conflict in the Somali Region of Ethiopia: Can Education Promote Peace-Building?” Feinstein International Center, Tufts University. Sarzin, Z. 2017. “Background Paper for the Ethiopia Forced Displacement Strategy Note.” World Bank. 50 Tesfaye, B. 2016. “The Point of No Return in Ethiopia.” Africa is a Country. Available at: https://africasacountry.com/2016/10/the-point-of-no-return-in-ethiopia UN-Habitat. 2016. “Sustainable Cities and Communities. SDG Goal 11. Monitoring Framework. A Guide to Assist National and Local Governments to Monitor and Report on SDG Goal 11 Indicators.” Available at: https://unhabitat.org/wp-content/uploads/2016/02/SDG-Goal%2011%20Monitoring%20Framework%2025-02-16.pdf UNHCR Statistics. http://popstats.unhcr.org/en/overview. UNHCR. 2017. “Global Trends: Forced Displacement in 2016.” UNHCR. 2018a. “Ethiopia Fact Sheet (May 2018).” UNHCR. 2018b. “Ethiopia Fact Sheet (December 2017).” UNHCR. 2018c. “South Sudan Situation (31 December 2017).” UNHCR. 2018d. “South Sudan Situation: Refugee Population in Gambella Region.” 31 January. UNHCR. 2018e. “Ethiopia Urban Refugees Fact Sheet”. December 2017. UNHCR. 2018f. “Ethiopia Fact Sheet (February 2018).” Uppsala Conflict Data Program (Date of retrieval: 18/03/23) UCDP Conflict Encyclopedia: www.ucdp.uu.se, Uppsala University. World Bank. 2015. “Ethiopia Poverty Assessment 2014.” Poverty Global Practice, Africa Region. Washington, DC: World Bank Group. World Bank. 2016. “Social Assessment Report: Development Response to Displacement Impact Project in the Horn of Africa (P152822).” SFG1888. Ministry of Agriculture and Natural Resources, Government of Federal Republic of Ethiopia. World Bank. 2017. Ethiopia: Country Partnership Framework for the Period FY18-FY22. Washington, DC: World Bank Group. World Bank. 2018a. “Forced Displacement in Ethiopia: Update on Current Situation and WBG Engagement.” Washington, DC: World Bank. World Bank. 2018b. World Development Indicators 2018. Washington, DC. World Bank. 2018. South Sudan Poverty Assessment 2018. [THIS SOURCE DOES NOT APPEAR IN CITATION] World Bank. 2018. Somali Poverty Profile 2018. [THIS SOURCE DOES NOT APPEAR IN CITATION] World Bank and UNHCR. 2015. “Forced Displacement and Mixed Migration in the Horn of Africa.” Report No. ACS14361. Geneva and Washington, DC. World Health Organization and UNICEF. 2006. “Core Questions on Drinking Water and Sanitation for Household Survey.” Zetter, R. and H. Ruaudel. 2016. “KNOMAD Study Part-II Refugees Rights to Work - An Assessment.” 51