VOLUNTARY MIGRATION IN ETHIOPIA In Search for Work and Better Opportunities Christina Wieser; Berhe Mekonnen; Lina Cardona-Sosa; Aisha Abubakar August 2022 © 2022 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street NW Washington, DC 20433 USA All rights reserved Photos: Shutterstock CONTENTS Acknowledgements ....................................................................................... iv 3.1 What drives people to migrate? .................................. 29 Abbreviations ................................................................................................... v Why do people migrate? ................................................... 29 Executive summary ................................................................................. vi Why do youth migrate? ..................................................... 31 What are government policies on migration? .................... vii Who migrates where and why? ................................... 33 Dispelling migration myths ............................................................ viii What are some other push and pull factors? .... 36 Myth 1. Migration flows in Ethiopia are too high ........ ix 3.2 What are institutional barriers migrants face? .... 36 Myth 2. All migrants come to Addis Ababa .................... x 3.3 What are the effects of migration on the Myth 3. Urban migration is unlikely to benefit different dimensions of welfare of migrants migrants or their families who stay back home .. xi and their families? ............................................................... 38 Myth 4. Rural-to-urban migration harms residents How do migrants fare in the labor market? ....... 38 at destination locations ............................................. xii How large are wage differentials for migrants? .. 40 Myth 5. Rural-to-urban migration stalls What are the effects on migrants’ families? ...... 42 development of the rural economy ................... xiv 3.4 What are the migration effects on Policy directions ..................................................................................... xv destination Areas? ............................................................... 43 . 1. Introduction .......................................................................................... 1 4. Conclusion and policy directions ........................................ 45 Shift policy focus from preventing to leveraging migration .. 45 2. Overview of migration in Ethiopia: Scale, pattern, Reduce barriers to migration .................................................... 45 and trends ............................................................................................. 7 Adapt urban areas to fast growing populations .......... 46 2.1 Evidence on internal migration ................................... 7 Improve process of international labor mobility ......... 47 To where do Ethiopians migrate? ............................... 11 References .......................................................................................................... 49 Who migrates? ........................................................................ 14 How did COVID-19 affect migration? ....................... 16 Annexes 2.2 Evidence on international labor migration 18 Annex 1. Mobility framework for international labor Where do Ethiopians migrate to abroad? ............ 20 migration in Ethiopia .......................................................... 53 Who migrates abroad? ...................................................... 21 Annex 2. Labor market outcomes: returnees vs. How does international migration affect non-returnees ......................................................................... 55 remittances? ............................................................................. 22 Annex 3. Propensity to migrate - Probit estimates ............. 57 2.3 Evidence on return migration ....................................... 24 Annex 4. Multinomial logistic regression model .................. 59 Annex 5. Oaxaca-Blinder decomposition methodology ... 66 3. Motives and impacts of migration ...................................... 29 Annex 6. Effects on destination areas .......................................... 70 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES ACKNOWLEDGEMENTS This report on “Voluntary Migration in Ethiopia: In Search for Work and Better Opportunities” was prepared by a team of the Poverty and Equity Global Practice, comprising Christina Wieser (Senior Economist, EAEPV), Berhe Mekonnen (Economist, EAEPV), Lina Cardona-Sosa (Economist, EAEPV), and Aisha Abubakar (Consultant, EAEPV). The report benefited from comments and guidance from Pierella Paci (Practice Manager, EAEPV) and Obert Pimhidzai (Senior Economist, EAEPV). The team also received feedback from peer reviewers Tom Bundervoet (Senior Economist, EECPV) and Emily Weedon (Senior Economist, MHNDR). The Migration Statistics team of the Ethiopia Statistics Service also provided valuable comments which helped further enrich the report. Last but not least, we also thank Aldo Mori for excellent editorial support and Cybil Maradza for the design and layout of the report. iv ABBREVIATIONS ABBREVIATIONS EAP East Asia and the Pacific ECA Europe and Central Asia ESS Ethiopia Socio-economic Survey ESS Ethiopian Statistics Service FAO Food and Agricultural Organization GCM Global Compact for Safe, Orderly and Regular Migration GoE Government of Ethiopia GCC Gulf Cooperation Council ILO International Labour Organization KNOMAD Global Knowledge Partnership on Migration and Development LFS Labor Force Survey LAC Latin America and the Caribbean MENA Middle East and North Africa MoFED Ministry of Finance and Economic Development NPC National Planning Commission (now Ministry of Planning and Development SNNP Southern Nations, Nationalities, and People’s SSA Sub-Saharan Africa SOE State of emergency UNDESA United Nations, Department of Economic and Social Affairs v VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES EXECUTIVE SUMMARY substantial growth and welfare payoffs, bringing sizable economic benefits to migrants and their families, as well as to their communities of origin through remittances that help to narrow gaps in living standards across the country (El Mufti 2019; Adugna 2021). Evidence on rural-to-urban migration shows that migration improves overall productivity as remittances have the potential to increase crop income and holding of land and livestock (Redehegn et al. 2019). Analysis based on the Ethiopia Socioeconomic Survey (ESS) shows that remittances in Ethiopia are an important source of livelihoods, equivalent to 31 percent of recipient household consumption expenditure nationally and 70 percent among the bottom quintile in 2016. Rural out-migration also reduces poverty through increased caloric consumption and diet (Abebaw et al. 2020; de Brauw, Mueller, and Woldehanna 2018) People migrate both within and between countries and increased non-food consumption (de Brauw, to improve their lives and the lives of families left Mueller, and Woldehanna 2018). Moreover, rural-to- back home. While both internal and international urban migration increases intensity of labor use in migration are common, internal migration is more migrant-origin households, agricultural output per common as it is relatively easier and cheaper to capita, and the amount of land rented out (World move within a country than across borders. While Bank 2022). The [limited] structural transformation reasons for migration are diverse and can include that took place in recent years was mainly the result moves following climatic or political shocks, of relatively better educated rural young people migrants typically move voluntarily from relatively migrating to urban areas. In fact, large investments less developed areas to more developed areas for in infrastructure, while demanding intensive use economic, family, or education reasons. This report of labor, have benefited from migrant labor (World focuses on these voluntary moves. In Ethiopia, in Bank 2020). the five years prior to the 2021 Labor Force and Migration Survey (LMS), about 6 percent of the Wage differentials incentivize people to cross Ethiopian adult population migrated voluntarily. borders and work abroad. Wage differentials— Migration has the potential to not only benefit gains from worker mobility—from international migrants and their families, but it can also have labor migration are large and could have positive effects for a country’s population overall by significant poverty reducing benefits in the country. increasing productivity and reducing poverty. Despite Remittances from international labor migration the benefits, migration can also place pressure on benefit migrants’ families as they improve destination locations and local labor markets. household welfare and economic conditions in origin countries. International migration also promotes Evidence is growing on the significant returns to economic activities in origin countries through voluntary internal and international migration. remittances, more efficient labor allocation, and Research shows that increased mobility has transfers of knowledge. Remittances can further vi EXECUTIVE SUMMARY increase savings for migrants’ household and states the importance of incorporating migration— give households the ability to buttress themselves from rural to urban areas and from small towns to against future shocks (UNCTAD 2018). bigger urban centers—with national and sectoral policies to ensure that migration has “positive Despite positive welfare effects, internal economic outcomes with reduced pressure on rural migration can also strain destination communities, areas and small towns and improved capacity of particularly urban areas, which can contribute towns/cities to accommodate migrants” (National to negative social externalities. Though little Planning Commission 2021). While encouraging, research on Ethiopia exists, rural migrants are in the Plan does not establish details on how to search of basic urban facilities, which can strain achieve these goals. Regulations for international socio-economic and environmental conditions in labor migration have also evolved in recent years. the host area. In Ethiopia, migrants are often blamed Ethiopia was a signatory to the Global Compact for creating shortages of housing, unemployment, for Safe, Orderly and Regular Migration (GCM) rising cost of living and crime, and expansion of form its inception. Since the adoption of the GCM, urban informal sectors. Analysis in this report did the Government of Ethiopia (GoE) has developed not confirm increased unemployment in areas and revised several practices, proclamations, with larger migrant inflows. We did, however, find directives, and policies to better govern labor that increased migration flows may cause slight mobility. For example, GoE is currently working on downward pressure on urban wage rates. a National Migration Policy, revised and passed the Proclamation on Prevention and Suppression of The benefits of internal and international labor Trafficking in Persons and Smuggling of Migrants migration, especially increasing household (No. 1178/2020) in 2020, amended the primary incomes and reducing poverty, are likely to instrument to govern international labor mobility in outweigh costs. Authorities should recognize Ethiopia (the Overseas Employment Proclamation that migration is a natural process, especially No. 923/2016), and signed it into law Ethiopian’s during fast economic growth and transformation. Overseas Employment (Amendment) Proclamation Migration can bring about positive welfare effects No. 1246/2021 (Federal Democratic Republic of and support economic growth, food security, and Ethiopia 2020) in 2021. overall poverty reduction. Policymakers should therefore encourage internal and international labor migration and focus on the overall positive welfare benefits of migration. Sound policies should promote the positive effects to flourish while compensating for the negative effects. WHAT ARE GOVERNMENT POLICIES ON MIGRATION? Policies in Ethiopia have focused on the negative aspects of migration, but perceptions are changing. Ethiopian government policies have historically restricted migration. Barriers to internal and international migration are still prevalent, though policies are evolving. The new Ten-Year Development Plan (2021-2030) explicitly vii VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES However, rural internal migrants continue to face general public and among government officials. a myriad of difficulties at their destinations. These While there are valid reasons for concern about some difficulties are related to lack of accommodations, of the pressures rural-to-urban population inflows job, and familiarity with urban life; challenging may create on urban services, these negative views relationships with local authorities; limited access are mostly grounded in misperceptions. This report to public services and support schemes (due to tries to test these misperceptions or “myths” about lack of ability to transfer Kebele IDs¹ with ease); internal migration by providing detailed analysis on and in some cases language and cultural barriers various aspects of migration, including migration (Bundervoet 2018). Potential migrants abroad also rates, relative importance of migration types, profile face challenges. International migration is lengthy of migrants, and reasons for migration over time. and expensive, discouraging migrants from using It also studies the potential effect of migration on formal channels to migrate. In addition, incentives migrants themselves, their families, and destination are badly aligned between recruitment agencies, areas. It takes advantage of the last three rounds jobseekers, and GoE (Smith et al. 2020). This results of the LFS/LMS, the latest of which was collected in a system of informal migration with high costs in 2021. It complements the analysis with findings and vulnerabilities to the migrant. from other data sources such as the Ethiopia Socio-economic Survey (ESS) data, a three-round, DISPELLING MIGRATION MYTHS country-level data panel collected between 2012 and 2016. Furthermore, it synthesizes existing Various common misperceptions or myths about studies on migration in Ethiopia. Using this evidence internal migration in Ethiopia exist, both in the base, we dispel five myths about internal migration: MYTH 1 MYTH 2 MYTH 3 Migration flows in Ethiopia All migrants come to Migration is unlikely to benefit are too high Addis Ababa migrants or their families who stay back home MYTH 4 MYTH 5 Rural-to-urban migration Rural-to-urban migration harms residents at stalls development of the destination locations rural economy ¹ The Kebele ID is the most important form of identification in Ethiopia. It is issued by local administrators in more than 16,000 different locales. The kebele ID provides legal identity and allows individuals to conduct public or private transactions, including obtaining a passport or voting in an election (World Bank 2016). viii EXECUTIVE SUMMARY MYTH 1: MIGRATION FLOWS IN ETHIOPIA ARE TOO HIGH Despite common belief, data show that internal migration rate was particularly limited in rural areas, migration remains low. Ethiopians and their with a mere 3 percent of the current rural population policymakers have a common belief that internal having migrated from other rural areas or urban migration in Ethiopia is high but the 2021 LMS data areas to their current location between 2016 and 2021 show that only 16 percent of the Ethiopia’s national (the five years preceding the 2021 LMS). Like lifetime population had migrated² at some point in their lives. migration rate, the recent migration rate is higher in The proportion of migrants is higher in urban (39 urban areas. In 2021, 16 percent of urban adults were percent) than rural areas (10 percent). Looking at only recent migrants (came to the city between 2016 and the adult population aged 15 and above,³ about one- 2021), with a decreasing trend over time. The pattern quarter migrated to their current location at some at the national level is driven by the decreasing role of point in their life; this is not surprising as adults are rural-to-rural migration. This is low compared to other more likely to migrate than children (Figure 1).⁴ These countries in the region; about 25 percent of Ugandans rates are also low in comparison with other countries. and 20 percent of Kenyans, for instance, had recently migrated. Despite a decreasing share of recent adults Recent migration also remains limited. Only about who migrated as a proportion of the overall adult 5 percent of the whole population and 6 percent of population, the absolute number of recent migrants is adults (Figure 2) migrated to their current location increasing. The number of people who changed their within five years preceding the 2021 LMS survey, residence within the past five years increased from a rate marginally lower than in 2005. The recent 3.6 to 4.7 million between 2013 and 2021. Figure 1: Migration for adults is low in Ethiopia Figure 2: Fewer recent adult migrants in cities (Share of lifetime adult migrants: national, rural and urban; (Share of recent adult migrants: national, rural and urban; 2005, 2005, 2013 and 2021) 2013 and 2021) 70 25 60 20 50 40 15 30 10 20 5 10 0 0 2005 2013 2021 2005 2013 2021 Whole country Rural Urban Whole country Rural Urban Note: Lifetime migrants are those who moved to their current Note: Recent migrants are those who lived in their current location location from another place in Ethiopia (town or rural woreda) at any (from another town or rural woreda in Ethiopia) for less than five point in their life. Adult population refers to people aged 15 years and years. Adult population refers to people aged 15 years and above. above. Source: Authors’ estimation based on LFS 2005, 2013, 2021. Source: Authors’ estimation based on LFS 2005, 2013, 2021. ² Migrants are defined as people who lived in areas other than their former woreda (for rural areas) or towns (for urban areas). ³ The definition of “adult population” used in this report is different from the official definition in Ethiopia which includes those who are aged 10 years and above. Migration patterns largely remain the same irrespective of the definition used. ⁴ The 2021 LMS does not cover the Tigray region and for comparability purposes, Tigray is also excluded from 2005 and 2013 LFS analysis throughout the report. However, migration patterns in 2005 and 2013 hardly change when Tigray is included. ix VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Significant regional differences exist in both the scale investments in public infrastructure and vulnerability and pattern of migration.⁵ In 2021, Gambela (11 percent) to weather-related events.⁶ Migration rates in the and Benishangul-Gumuz (9 percent) had the highest predominantly urban regions and city administrations— migration rates for recent adult migrants from the Harari, Addis Ababa, and Dire Dawa—are among the primarily rural regions. The two regions are located in the highest. The share of recent adult migrants in 2021 “moisture reliable lowland” ecological zone and their more was 9 percent in Harari, 8 percent in Addis Ababa, and favorable climate may attract migrants. From regions 7 percent in Dire Dawa. Consistent with the pattern at the that are predominantly rural, Somali had the lowest rate national level, and contrary to common belief, migration of migration (1 percent), which may be explained by low rates have decreased since 2005 in most regions. MYTH 2: MIGRATION FLOWS IN ETHIOPIA ARE TOO HIGH It is commonly believed that migrants from rural larger (Figure 3); secondary cities and small towns areas mostly go to Addis Ababa but 2021 LMS host twice as many migrants as a share of overall data paint a more nuanced picture. In 2021, only resident population. In fact, small towns host the 8 percent of Addis Ababa’s residents were recent majority of migrants to urban areas, underscoring the migrants. When comparing Addis Ababa to other importance of small towns as destinations of rural- Ethiopian cities, we see that the share of migrants in to-urban and urban-to-urban migration. In 2021, 71 secondary cities (which includes regional capital cities percent of rural-to-urban and 69 percent of urban- and other cities with a population of at least 100,000 to-urban migration was to small towns, larger than residents in 2007) and small towns (all urban areas the national population share in small towns of 63 other than Addis Ababa and secondary cities) is much percent (Figure 4). Addis Ababa hosted roughly 11 Figure 3: Migration rate for Addis Ababa was lower Figure 4: Small towns host proportionally more than for secondary cities and small towns⁷ migrants from rural areas (Share of recent adult migration in respective location) (Migration and population share of migrants to urban areas by city size) Urban Small towns type Secondary 100% Dire Dawa 80% Addis Ababa Harari 60% 70.6 69.2 62.5 Gambela Regions SNNP 40% Benishangul-Gumuz Somali 19.4 20% 19.0 21.3 Oromia 10.5 9.5 18.1 Amhara 0% Afar Rural-urban Urban-urban Population share 0.0 5.0 10.0 15.0 20.0 25.0 30.0 2021 2013 2005 Addis Ababa Secondary cities Small towns Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. ⁵ Tigray is not included in the 2021 LMS due to ongoing conflict in the region. Yet, people who came from Tigray in the five years preceding the survey are included. ⁶ Somali has a large nomadic/mobile population not considered to be migrant populations. The non-sedentary nature of Somali’s population in rural areas will naturally limit rural-to-urban migration, the most important migration type in 2021. ⁷ Sidama region was part of SNNP region in the earlier rounds of LFS (LFS 2005 and 2013). For comparison purpose, it is also included as part of SNNP region in the 2021 data. Statistics for Sidama based on LMS 2021 can be provided upon request. x EXECUTIVE SUMMARY percent of all rural-to-urban and 10 percent urban- 4 percent. Yet, Addis Ababa is an important destination to-urban migration while it accounts for 18 percent of for migration between regions; roughly one-quarter of the entire urban population. Moreover, considering all internal migrants who moved between regions moved types of internal migration between 2016 and 2021, to Addis Ababa between 2016 and 2021. Addis Ababa migration to the capital city is actually similar to that is a particularly important destination for recent rural- of other predominantly urban regions. Addis Ababa to-urban migrants who move between regions; the hosted about 6 percent of all internal migrants, a capital city received almost half of all recent migrants proportion slightly higher than its population share of who moved between regions. MYTH 3: URBAN MIGRATION IS UNLIKELY TO BENEFIT MIGRANTS OR THEIR FAMILIES WHO STAY BACK HOME Migration improves the lives of migrants and their unemployment rates (Figure 5 and Figure 6). In families. It is sometimes argued that migrants are fact, when we look at youth only, both labor force not likely to benefit from moving to urban areas participation and unemployment rates are lower because of their relatively lower human capital and for rural-to-urban migrants compared to urban the already crowded urban labor market. Evidence non-migrant youth. Labor market outcomes are shows that there are benefits to rural-to-urban even better for youth who migrated from rural to migration and labor market outcomes for rural- urban areas for economic reasons (Figure 7 and to-urban migrants do not differ significantly from Figure 8). Rural-to-urban migrants are equally urban non-migrants. Moreover, urban migrants likely to be economically active, employed, and in are likely to do better than their counterparts who non-agricultural work as urban residents; but are remain in rural areas. Rural-to-urban migrants more likely to work in formal wage employment. are more likely to be active in the labor force (74 A disproportionate share (17 percent), however, is percent) than urban non-migrants (71 percent) engaged in domestic wage work. This shows that and have no statistically significant difference in job prospects for rural migrants are promising. Figure 5: Rural-to-urban migrants are more active Figure 6: But they are also more likely to be unemployed than urban non-migrants (Unemployment rate by type of migration of working age (Labor Force Participation Rate by type of migration of working population in 2021) age population in 2021) 20.9 19.5 77.4 74.4 74.1 17.7 70.8 5.5 Urban Rural-urban Urban-urban Rural Urban Rural-urban Urban-urban Rural non-migrants migrants migrants non-migrants non-migrants migrants migrants non-migrants Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Only working age population (people aged 15 to 64 years) are years. Only working age population (people aged 15 to 64 years) are included. Bars indicate 95 percent confidence intervals. included. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. xi VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Thus, given the large income gap between rural study tracking a sample of internal migrants that and urban areas, rural-to-urban migration is left between 2004/2005 and were observed again likely to benefit migrants and their families. In in 2009, found that migrants increased their non- fact, analysis based on data from around 1,000 food consumption relative to non-migrants and rural households from Amhara, Oromia, and improved their diets (de Brauw, Mueller, and Southern Nations, Nationalities, and People’s Woldehanna 2018). Positive spillovers were also (SNNP) regions collected in 2016 and 2017 found in the rural agriculture sector, where land shows that rural out-migration increased the rent-outs increased among migrant households, households’ migrant calorie consumption by 22 thus increasing the efficiency of rental markets percent while the food poverty gap and severity and reducing disguised unemployment,⁸ resulting of food poverty decreased by 7 and 4 percent, in increased output per worker and labor respectively (Abebaw et al. 2020). Similarly, a productivity in rural areas (World Bank 2022). Figure 7: Rural-to-urban youth migrants are more Figure 8: They also have a lower unemployment rate active than urban non-migrant youth (Unemployment rate by type of migration of youth in 2021) (Labor force participation rate by type of migration of youth in 2021) 90.1 26.8 22.3 63.7 59.9 15.1 42.9 7.7 Rural Urban All rural-urban Rural-urban Rural Urban All rural-urban Rural-urban non-migrants non-migrants migrants economic non-migrants non-migrants migrants economic Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Only working age population (people aged 15 to 64 years) are years. Only working age population (people aged 15 to 64 years) are included. Bars indicate 95 percent confidence intervals. included. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. MYTH 4: RURAL-TO-URBAN MIGRATION HARMS RESIDENTS AT DESTINATION LOCATIONS No strong evidence supports the common belief than the general rural population. In 2021, the that rural-to-urban migration harms destination literacy rate for rural non-migrants was only areas. One reason for the belief of rural-to-urban 46 percent while it was 78 percent for rural-to- migrants not contributing positively to their urban migrants, similar to urban non-migrants destination areas is that they have lower levels of (79 percent) (Figure 9). Similarly, while only 16 education. In reality, rural dwellers who migrate percent of rural non-migrants completed primary to urban areas are significantly more educated education, the primary completion rate for those ⁸ Disguised unemployment occurs when productivity is low and too many workers are filling too few (or the same) jobs. xii EXECUTIVE SUMMARY who migrated to rural areas was 49 percent, which Industrial Parks come from rural areas—about 70 is only modestly lower than the rate for urban non- percent of workers in Bole-Lemi Industrial Park migrants (55 percent) (Figure 10). Thus, rural-to- and 52 percent of workers in Hawassa Industrial urban migrants who have better education than Park, for example, are migrants (Abebe, Buehren, the general rural population and are younger can and Goldstein 2020; Meyer et al. 2021)—and as potentially contribute to their destination areas. industry expands, rural-to-urban migration can There is also evidence that most workers in serve as a sustainable source of labor supply. Figure 9: Migrants originating in rural areas have Figure 10: Educational attainment is higher for higher literacy rates than rural non-migrants migrants than non-migrants (Literacy rates by type of migration in 2021) (Share of educational attainment by type of migration in 2021) 100 100 80 80 60 60 40 20 40 0 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban 20 0 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban Non- Rural origin Urban origin migrants migrants migrants No education Less than primary Completed primary Non- Rural origin Urban origin migrants migrants migrants Completed secondary Completed post-secondary Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. years. Adult population refers to people aged 15 years and above. Bars indicate 95 percent confidence intervals. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. Another common belief about rural-to-urban migration is that migration pressures already crowded urban labor markets, leading to higher unemployment and lower wages. Analysis based on the 2021 LMS shows that larger inflows of rural-to-urban migrants do not increase urban unemployment but they do modestly decrease wages. As the fraction of recent rural migrants increases in urban woredas, the unemployment rate is not affected in a statistically significant way but wages seem to slightly decrease; a 10 percent increase in the proportion of recent rural- to-urban migrants in a destination woreda leads to a decrease in wage of less than 3 percent. Yet, such a large increase is unlikely considering the rather modest migration rates from rural to urban areas of merely 9 percent. xiii VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES MYTH 5: RURAL-TO-URBAN MIGRATION STALLS DEVEL- OPMENT OF THE RURAL ECONOMY Evidence also shows, contrary to common consumption of those recipients in the bottom belief, that rural-to-urban migration neither consumption quartile (Figure 11). Remittances causes rural productive labor shortages nor prevent households from falling into poverty harms agricultural productivity. Recent research during hard times. In fact, the likelihood of falling actually shows that even without taking account into poverty is 7 percentage points lower among of remittances, migration enhances welfare in households with a migrant compared with similar migrant-origin households in Ethiopia and increases households without a migrant (World Bank 2022). agricultural productivity. Given excess labor supply in rural areas, migration has a positive effect on Figure 12: Rural-to-urban migration is associated labor productivity. Migration increases the intensity with improvements in welfare (Share of households of rural-to-urban migrants within certain of family labor use and output per worker—family consumption quartile) labor days worked increased by 29 percent and 35 32.3 output per capita increased by 18 percent–in 30 28.2 migrant origin households, thus reducing disguised 25.3 26.6 23.4 24.1 25 unemployment. It also increases the share of land 19.0 21.1 20 rented out (World Bank 2022). 15 10 Moreover, migrant remittances contribute 5 to households’ incomes and bolster shock 0 resilience back home. Analysis based on the ESS 2012 (before migration) 2016 (after migration) shows that although only a small fraction of rural Poorest quartile 2nd quartile households receives remittances (about 5 percent), 3rd quartile Richest quartile the amount they receive is significant. For example, Note: Recent migrants are those who lived in their current location in 2016, remittances received from urban areas (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. was on average about 31 percent of all recipient Source: Authors’ estimation based on ESS 2012, 2016. households’ consumption and 70 percent of the Figure 11: Size of urban-rural remittance income Figure 13: Migration moves households out of the relative to total consumption of recipient households bottom 40 percent of the welfare distribution (Remittances as a percentage of consumption by quintile) (Share of households of rural-to-urban migrants in bottom 40 percent of consumption distribution) 80 60 70 47.8 50 Remittance as a % 60 of consumption 37.5 45 50 30 40 30 25 20 10 10 0 0 2012 (before migration) 2016 (after migration) All Poorest Second Third Fourth Richest rural quintile quintile quintile quintile quintile Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five 2012 2014 2016 years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on ESS 2012, 204, 2016. Source: Authors’ estimation based on ESS 2012, 2016. xiv EXECUTIVE SUMMARY Migration seems to improve the welfare of origin removing the requirement for a release letter households. Figure 12 shows that rural-to-urban from home areas. In addition to encouraging migration reduce the likelihood of migrant origin migration, reducing migration costs can households to be in the lower consumption quartiles. enhance returns to migration. Reducing In 2012, before migrants moved to urban destination, “frictions” in the job matching process—better 32 percent of migrant origin households were in the connecting migrant jobseekers to employment lowest consumption quartile, but after the migrants opportunities and job placement services—can moved (in 2016) only 24 percent of migrant origin reduce costs. Moreover, facilitating access to households were in the lowest consumption quartile. credit and financial services can help members Similarly, while 48 percent of the migrant households of poorer households—who suffer most from were in the bottom 40 percent based on consumption, liquidity constraints—to migrate. the proportion decreased to 38 percent in 2016 (Figure 13). • Third, adapting urban areas in advance of fast growing population inflows can reduce POLICY DIRECTIONS challenges migrants create in urban areas. Urban development strategies, particularly This report identifies several policy directions for cities with large shares of migrants, can related to strengthening economic and social avoid the negative consequences from large benefits from migration while, at the same time, population influxes. Urban adaptations to better reducing pressures migration may place on urban integrate migrants socially and economically areas and service provision: include: (i) continuing to expand public infrastructure and services in main migrant • First, shifting the policy focus from preventing destination cities, (ii) including migrants in to leveraging migration could increase urban social protection schemes, and (iii) mobility and help Ethiopia’s households and investing in housing infrastructure. economy to benefit from migration-induced growth and welfare benefits. Authorities • Fourth, improving the process for workers should recognize that migration is a natural to migrate internationally can expand process, especially during fast economic labor flows from Ethiopia to reap economic growth and transformation. Policies should benefits. Improving the regulatory framework therefore encourage labor migration rather for international labor migration and reducing than deter it; national government programs barriers to migrations can save time and money. should promote internal migration and urban- Currently, the process places significant time rural linkages and attempt to reverse negative and cost burden on the worker, incentivizing perception of migrants. Reshaping policy payments to an agent to take on this burden or perspectives and overall attitude towards even to migrate irregularly to avoid the process migration could encourage a positive policy altogether. This undermines the competitiveness shift towards increasing mobility. of Ethiopian workers compared to workers from other countries. Better aligning incentives • Second, reducing barriers to migration can between workers, agents, and government increase mobility and encourage migration to entities could promote formal recruitment locations with better economic opportunities. channels. Moreover, reducing gaps in protection To facilitate migrants’ freedom of movement systems while abroad, increasing the skills of and integration, policymakers could remove prospective migrants, and reintegrating migrants restrictions on urban ID requirements, such upon return could improve the development as requiring a minimum length of stay and potential of international labor migration. xv VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES 1. INTRODUCTION Ethiopia has historically restricted migration, by grouping, often forcefully, farmers into grid- it is taking place nonetheless. In recent years, plan villages. This villagization aimed at increasing migration takes place through two main channels: agricultural production and improving delivery of internal migration (that is, migration within Ethiopia) services such as education and health (Bundervoet and temporary international labor migration. 2018). Since the early 2000s, national programs Government policies have controlled or moderated such as the Sustainable Development and Poverty both types of migration. In the case of international Reduction Program of 2002/2003, focused on the migration, the Government lifted the ban on work- negative aspects of migration, relating it with urban related migration to Gulf States in 2018, while poverty, HIV, and crime (Ministry of Finance and enacting new legislation aimed to protect citizens Economic Development [MoFED] 2002). In 2006, by ensuring that immigrants fulfill a minimum set the message started to change with the Plan for of requirements related to education and training Accelerated and Sustained Development to End (Smith et al. 2020). Internal migration, on the Poverty, which emphasized the need to strengthen other hand, has been constrained by language rural-urban links to develop small towns and and cultural barriers, land redistribution policies— generate employment (MoFED 2006). In 2010, the including the possibility of losing the ownership of Growth and Transformation Plan mentioned the land when working in non-farm employment—and need to create urban-rural linkages to promote requirements for migrants to have a Kebele ID to urban development, without specifying the role of access services as well as providing evidence of migration (National Planning Commission 2010). land ownership (Bundervoet 2018). The new Ten-Year Development Plan (2021-2030) explicitly states the importance of incorporating National policies have focused on the negative migration to bigger urban centers with national aspects of migration, but perceptions are changing. and sectoral policies to ensure that migration has In the 1980s, the Derg regime implemented a Land “positive economic outcomes with reduced pressure Reform Policy (1984) aimed at controlling mobility on rural areas and small towns and improved 1 1. INTRODUCTION capacity of towns/cities to accommodate migrants” economic growth and structural transformation. (National Planning Commission 2021). While this is Migration and urbanization can increase education a positive step, the Plan does not include details on levels in rural areas, harness the demographic how to achieve its goals. Despite these restrictions, dividend, and manage conflict and climate shock internal migration continues to take place, albeit in risks. These also represent opportunities to low proportions. increase non-farm employment and contribute to urban development by participating in sectors International labor migration laws have also such as construction. Despite starting from a evolved in recent years. Ethiopia was a signatory low base, Ethiopia is urbanizing quickly. In 2019, to the Global Compact for Safe, Orderly and Ethiopia had one of the lowest urban population Regular Migration (GCM) from its inception. Since shares in the world, with only 21 percent of the adoption of the GCM, the GoE has developed Ethiopians living in urban areas—well below the and revised several practices, proclamations, Sub-Saharan Africa (SSA) average of 37 percent. directives, and policies to better govern labor However, with economic growth and structural mobility. For example, GoE is currently working on transformation, this is set to change dramatically a National Migration Policy; revised and passed the as off-farm employment opportunities in urban Proclamation on Prevention and Suppression of areas increase. According to official figures from Trafficking in Persons and Smuggling of Migrants the Ethiopian Statistics Service (formerly Central (No. 1178/2020) in 2020; amended the primary Statistics Agency), the urbanization rate has been instrument to govern international labor mobility in growing at an average of 5.2 percent per year since Ethiopia (the Overseas Employment Proclamation 2018. If these trends continue, the urban population No. 923/2016) and signed into law Ethiopian’s is projected to reach 50 million by 2034 (United Overseas Employment (Amendment) Proclamation Nations, Department of Economic and Social Affairs No. 1246/2021 (Federal Democratic Republic of [UNDESA] and Population Division 2019). Natural Ethiopia 2020) in 2021. increase—population growth rather than rural- to-urban migration—has been the main driver of However, barriers to internal and international urban population growth up to 2018, with rural-to- migration are still prevalent. Qualitative research urban migration being the main driver since 2018 (Bundervoet 2018) suggests that rural migrants (World Bank 2020; Ethiopian Economics Association face a myriad of difficulties in destination towns 2021). As population density increases, combined and cities related to accommodations, jobs, lack of familiarity with urban life, harassment by local authorities, limited access to public services and support schemes (due to lack of ability to transfer Kebele IDs with ease), and sometimes language and cultural differences (Bundervoet 2018). Potential migrants abroad also face challenges. The current process for international migration is lengthy and expensive, discouraging potential migrants from using formal channels to migrate. In addition, incentives are badly aligned between recruitment agencies, jobseekers, and the GoE (Smith et al. 2020). This results in a system of informal migration with high costs and vulnerabilities to migrants. Rural-to-urban migration and rapid urbanization have characterized countries undergoing rapid 2 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES with continued land fragmentation, large cohorts of households from rural areas with a migrant were young people will increasingly become functionally 15 percent less likely to be food insecure and landless. In addition, as youth education levels rise, migration showed a positive effect on household and access to social media reveals aspirational asset accumulation (Kangmennaang, Bezner- lifestyles in urban areas, migration will become Kerr, and Luginaah 2018). Analysis based on data a livelihood strategy as youth look for off-farm from around 1,000 rural households from Amhara, employment in urban areas (Food and Agricultural Oromia, and SNNP regions collected in 2016 and Organization [FAO] 2016). 2017 shows that rural out-migration increased the households’ migrant calorie consumption by 22 Moreover, climate change-induced natural percent while the food poverty gap and severity disasters will likely accelerate large population of food poverty were reduced by seven and four inflows into urban areas. Drought increases percent, respectively (Abebaw et al. 2020). Similarly, mobility, primarily through triggering short-term a study tracking a sample of internal migrants that migration to closer destinations to cover immediate left their homes between 2004/2005, observed needs, such as food in cases of shortages (Hermans again in 2009, found that migrants increased their and Garbe 2019). Improving absorption capacity non-food consumption relative to non-migrants of these population inflows into urban areas and improved their diets (de Brauw, Mueller, and will be increasingly challenging and will require Woldehanna 2018). spatial planning based on sustainable urban-led development. This will require employing migrants Increased mobility is potentially a powerful tool productively and generating enough economic for boosting productivity and reducing poverty. In growth without jeopardizing living conditions terms of aggregate productivity effects, removing of migrants and local residents. In Ethiopia, migration barriers in Indonesia increased labor environmental changes have already contributed productivity by 22 percent, and reducing migration to migration, due to pressures on livelihoods and costs lead to a 7 percent productivity boost (Bryan unfavorable effects on agricultural activities (Groth and Morten 2019). In Vietnam, remittances from et al. 2020). internal migrants played an active role in reducing the probability of receiving households falling into Despite increasing evidence for significant poverty and it reduced the depth of poverty (Quoc returns from voluntary internal migration, many Hoi Le and Thi Hoai Thu Nguyen 2019). Research governments actively discourage rural-to-urban also shows ample welfare benefits of increased migration. In a 2013 UN survey on Population and internal migration in Ethiopia (de Brauw, Mueller, Development, 148 out of the 185 surveyed countries and Woldehanna 2018). (Limited) structural with data had government policies aimed at reducing transformation over recent years in Ethiopia was internal migration from rural to urban areas. Such mainly the result of relatively better educated efforts are particularly prevalent in Africa, including rural young people migrating to urban areas. In Ethiopia (UNDESA 2022). Yet, research shows that fact, large investments in building labor-intensive migration has substantial growth and welfare infrastructure have benefited from migrant labor. payoffs, bringing sizable economic benefits to the migrant and their families, and to their communities Yet, internal migration in Ethiopia is low. According of origin through remittances that help to converge to the 2021 LMS, 5 percent of Ethiopians moved to living standards across the country (El Mufti 2019; their current residence between 2016 and 2021. The Adugna 2021). There is also evidence that rural- internal migration rate among youth aged 15-24 to-urban migration improves overall productivity. years was higher at 9 percent. The scale of reported Indeed, remittances in Ethiopia have shown to internal migration even decreased between 2005 increase the crop income and the holdings of land and 2021, and its pattern has changed in important and livestock (Redehegn et al. 2019). Also, Malawian ways, with rural-to-urban migration becoming 3 1. INTRODUCTION the dominant migration form in 2013 and further earnings even after they return, by helping migrants increasing in 2021 (Figure 1.1). Migration to and establish start-up capital to facilitate entrepreneurial within urban areas (either rural-to-urban or urban- activities back home (UNCTAD 2018). to-urban migration) is the most prominent form of migration, accounting for roughly two-thirds of all International labor migration can be an effective internal migration. About 34 percent of all migrants policy instrument for Ethiopia to capitalize on move from rural to urban areas but urban-to-urban its demographic transition and growing youth migration is also on the rise. Hence the welfare population. Youth migrating to urban areas could effects of internal migration can support domestic challenge the absorption capacity of the labor economic growth and significantly benefit food market. Slow structural transformation in Ethiopia security and overall poverty reduction. is unlikely to keep pace with the demographic transition of a growing youth population. Every year, International temporary labor migration has the the working-age population is expected to grow potential to increase incomes. Wage differentials— by two million people (World Bank 2020). Given gains from worker mobility—from international the limited absorption capacity of the domestic labor migration are large and could have significant labor market, international labor migration can be poverty reducing benefits in the country. For example, vital to employ a young labor force. Yet, Ethiopian a migrant construction worker in Saudi Arabia earns international migration pales in comparison to approximately 8 times more in monthly earnings similar SSA countries (Figure 1.2). The low outflow compared to a construction worker in Ethiopia (Smith in part reflects previous restrictions imposed on et al. 2020). In addition, the minimum monthly wage working overseas, as well as the lack of information of a domestic worker in Saudi Arabia is 21 times and transparency from recruitment agencies, and compared to the median monthly wage of a domestic bad experiences related to exploitation. Hence, worker in Ethiopia (Pritchett and Hani 2020). The large government strategies to increase training for income benefits and consequent remittances could skills demanded abroad, together with promotion of increase savings for the migrants’ household giving formal recruitment channels, might help increase them the ability to prepare against future shocks. the number of migrants and positive outcomes In addition, it could increase a migrants’ lifetime from international migration. Figure 1.1: Rural-to-urban migration has become Figure 1.2: Migrant stock in Ethiopia is low in comparison the most important type of migration to other countries (Share of recent migration by type and time period) (Migrant stock as a percentage of population) 45 40 7.2 35 30 25 5.2 20 15 3.9 10 5 2.2 0 2.0 1.7 2005 2013 2021 1.1 1.3 Rural-rural Rural-urban Urban-rural Urban-urban Ethiopia Kenya Uganda South Africa SSA Average Bangladesh Nepal Thailand Note: Recent migrants are individuals who moved less than five years prior to survey data collection. Based on the population aged 15 and over. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: UN DESA Population Division (Migrant Stock 2019). 4 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Despite many benefits, internal migration can This report expands our understanding of voluntary also pressure urban areas and local labor economic migration in Ethiopia.¹⁰ The possible markets. Based on the 2021 LMS, as a share of adverse effects arising from pressures on urban their population, smaller towns attracted the largest areas requires making migration part of the overall share of rural migrants. Contrary to common development plan at different administrative levels perception, Addis Ababa hosts proportionately (national, regional, and woreda) and establishing fewer migrants than its population share and less mechanisms to support migrants in their transition than 4 percent of all migrants went to Addis Ababa to urban or overseas labor markets. This, in turn, between 2016 and 2021. Yet, Addis Ababa has been requires a thorough understanding of migration, an important destination for migrants; close to 41 including the profile of migrants, destinations, percent of its residents were not born in the capital, benefits (and possible harms), and barriers to but this share is still lower than in small towns migration, among other things. This report presents (53 percent) and secondary cities (55 percent).⁹ If a comprehensive picture on migration in Ethiopia by not managed well, rapid urban population growth synthesizing previous research and complementing will pose challenges as cities struggle to provide existing evidence with new analysis using more jobs, infrastructure, services, and housing. Growing recent data, including the latest available 2021 LMS. urban boundaries and stretched municipal budgets already undermine infrastructure and service This report is structured around two broad delivery in many cities, while formal labor markets sections, which aim to provide a comprehensive are failing to keep up with demand for jobs. Moreover, picture of voluntary internal and international the large number of new labor market entrants migration in Ethiopia, as well as a section each year adds to existing urban unemployment, highlighting broad policy implications. which, as of 2020, stands at 19 percent of the urban population (Urban Employment and Unemployment Chapter two provides an overview of migration in Survey, 2020). While migration to nearby towns Ethiopia and the latest trends on migration. “Myths” can be an alternative to migration to large towns/ outlined in the Executive Summary on rural-to- cities, small towns lag in urban services like water, urban migration in Ethiopia informed the evidence sanitation, and housing (World Bank Group 2015). in this section by focusing on migration scale, spatial patterns, and latest trends. Sub-section one Benefits of internal and international labor provides evidence shedding light on the following migration, especially increasing household questions for internal migration: incomes and reducing poverty, are likely to outweigh costs. Authorities should recognize that • To where do Ethiopians migrate? migration is a natural process, especially during • Who migrates? fast economic growth and transformation. Internal • How did COVID-19 affect migration? and international labor migration should therefore be encouraged rather than deterred, and national Sub-section two looks at the evidence on government programs should explicitly promote international migration, including spatial patterns internal migration or urban-rural linkages and aim of international labor migrants, migrant profiles, to change the negative perceptions of migrants. and remittances trends. ⁹ Secondary cities include regional capital cities and other cities with a population of at least 100,000 residents in 2007. Small towns include all other urban areas. ¹⁰ This report does not cover forced displacement, including displacement from violence and conflict or climate-induced migration. 5 1. INTRODUCTION Chapter three discusses migration motives and • What are institutional barriers migrants face? effects. It focuses on the characteristics and • What are the effects of migration on different motivations of migrants, concentrating on push and dimension migrant s’ and their families’ welfare? pull factors, disaggregated by type of migration. • What are the effects of migration on Chapter three provides evidence to help answer the destination areas? following questions: Chapter four highlights policy directions to • What drives people to migrate? maximize the benefit of migration while minimizing • Why do youth migrate? the costs. 6 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS This section provides an overview of recent 2.1 EVIDENCE ON INTERNAL MIGRATION Ethiopian migration trends. We focus on internal migration (with an emphasis on rural-to-urban In Ethiopia, as in many developing countries, migration), international labor migration, as well as disparities in living standards between rural and return migration. The evidence of internal migration urban areas are large. Disparities across rural and is based on the Labor Force Survey (LFS)/Labor Force urban areas as well as across leading and lagging and Migration Survey (LMS) series produced by the regions remain substantial in Ethiopia. The country Ethiopian Statistics Service, the latest of which was has shown strong poverty reduction nationally collected in 2021. Evidence on international labor between 2011 and 2016, from 30 percent in 2011 to 24 migration is based on the Ethiopia Labor Mobility percent in 2016, but poverty reduction concentrated Diagnostic (Smith et al. 2020)¹¹ and analysis based in the urban areas of almost all regions, while in on the 2021 LMS. The evidence on return migration general weak in rural areas (Mekonnen et al. 2020). is based on the 2021 LMS. The analysis presented in For example, in 2016, poverty rates in rural areas were this report aims at providing a fresh perspective on almost double those of urban areas at 26 percent migration in Ethiopia, dispelling some of the myths compared to 15 percent. Moreover, the poverty rate on the negative aspects of migration, and providing in the poorest region (Tigray) was four times higher a sound analytical base for decision making on than poverty in the least poor region (Harari) (World migration policies. Bank Group 2021). Disparities in living standards ¹¹ The Ethiopia Labor Mobility Diagnostic was produced as part of the Jobs Advisory Services and Analytics which closed in 2021 and new stock data on migrants abroad is not available. 7 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS largely correspond to disparities in economic density market outcomes in cities led to attempts to restrict and thus labor market opportunities; locations with internal movements of people (Bundervoet 2018). higher economic density (typically cities) tend to This resulted in a relatively low and decreasing have higher living standards, while locations far rate of internal migration. In 2021, 16 percent of the from economic activity (mainly rural and relatively national population had migrated¹² at some point in remote areas) tend to have lower living standards. their lives (Figure 2.1). The proportion of migrants These spatial disparities in living standards provide is higher in urban areas (39 percent) compared compelling motivation for people in lagging regions to rural areas (10 percent). Looking at only the to move closer to economic activity. adult population aged 15 years and above,¹³ about one-quarter of Ethiopian adults, and half Yet, internal migration continues to be low in the urban adult residents, have migrated at some Ethiopia. Ethiopians and their policymakers have a point in their lifetime as adults are more likely common belief that internal migration in Ethiopia to migrate than children (Figure 2.2). The rate of is high and until recently, labor mobility and rural- overall migration (including all types of migration) to-urban migration was perceived as undesirable slightly decreased since 2005 and remains low by many Ethiopians, and fears of overburdening at the national level, particularly to comparator infrastructure and services and undesirable labor countries (Figure 2.3).¹⁴ Figure 2.1: Internal migration remained low for all Figure 2.2: Adults are more likely to migrate Ethiopians (Share of lifetime adult migrants: national, rural and urban; 2005, (Share of lifetime migrants: national, rural and urban; 2005, 2013 and 2021) 2013 and 2021) 60 70 50 60 50 40 40 30 30 20 20 10 10 0 0 2005 2013 2021 2005 2013 2021 Whole country Rural Urban Whole country Rural Urban Note: Lifetime migrants are those who moved to their current Note: Lifetime migrants are those who moved to their current location from another place in Ethiopia (town or rural woreda) at location from another place in Ethiopia (town or rural woreda) at any point in their life. Adult population refers to people aged 15 any point in their life. years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. ¹² Migrants are defined as people who lived in areas other than their former woreda (for rural areas) or towns (for urban areas). ¹³ The definition of “adult population” used in this report is different from the official definition which includes those who are 10 years and above. Migration patterns remain the same irrespective of which definition is used. ¹⁴ The 2021 LMS does not cover the Tigray region and for comparability purposes, Tigray is also excluded from 2005 and 2013 LFS analysis throughout the report. However, migration patterns in 2005 and 2013 hardly change when Tigray is included. 8 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Figure 2.3: Internal migration is low in Ethiopia compared to other countries in the region (Internal migration as a share of population in comparator countries) 30% 25% 20% 16% 14% 9% Uganda Ethiopia Kenya Rwanda India Vietnam (2001) (2021) (2014) (2011-2014) (2015) SSA Outside SSA Note: For Ethiopia, Uganda, India and Vietnam, internal migrants refer to any individuals who resides in a different woreda/city different to the one of their birth. For Uganda, this corresponds also to people aged 25-49. For Kenya, the statistic corresponds to the share of working-age men that moved to their current residence in the last four years. For Rwanda, the figure corresponds to the share of internal migrants between 2011-2014. Source: World Bank Group. 2015 (Ethiopia Poverty Assessment 2014); World Bank. 2018 (Kenya Gender and Poverty Assessment 2015- 2016: Reflecting on a Decade of Progress and the Road Ahead); World Bank Group. 2017 (Reshaping Urbanization in Rwanda: Profiling Secondary Cities in Rwanda—Dynamics and Opportunities). UNFPA, 2016. (The 2015 National Internal Migration Survey: Major Findings). Even in the last five years, internal migration years preceding the 2021 LMS). Migrants account within Ethiopia remains limited. In the five years for a higher share of the population in urban areas. prior to the 2021 LMS, about 5 percent of Ethiopians In 2021, 16 percent of urban adults were recent (Figure 2.4) and 6 percent of the Ethiopian adult migrants (came to the city between 2016 and population (Figure 2.5) migrated, marginally lower 2021). Despite a decreasing share of migrants, the than the shares in 2005. In rural areas, mobility absolute number of migrants increased. The number is particularly limited, with a mere 3 percent of of internal migrants increased from 3.6 million in 2013 adults migrating between 2016 and 2021 (the five to 4.7 million in 2021, with a faster increase in urban Figure 2.4: Internal migration remained low for all Figure 2.5: Fewer recent adult migrants in cities Ethiopians (Share of recent adult migrants: national, rural and urban; 2005, (Share of recent migrants: national, rural and urban; 2005, 2013 and 2021) 2013 and 2021) 18 25 16 14 20 12 15 10 8 10 6 4 5 2 0 0 2005 2013 2021 2005 2013 2021 Whole country Rural Urban Whole country Rural Urban Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five years. (from another town or rural woreda in Ethiopia) for less than five Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. 9 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS areas. Decreasing rural-to-rural migration (Figure 2.6) Moreover, climate change-induced natural disasters mainly relates to an increase in primary and secondary will likely accelerate large population inflows into education levels that generated higher aspirations urban areas. Of all internal population movements to migrate, both nationally and internationally. This between 2016 and 2021, about two-thirds went to shows a clear relationship between development and urban areas—either from rural or from other urban migration (Schewel and Fransen 2018). areas. Relative to other migration types, rural-to- urban migration as a share of overall migration has For the remaining analyses, we focus on recent adult increased over time to become the most important migrants and not migration of the overall population type of migration in 2013 and 2021; it accounts or overall lifetime migration (unless otherwise stated), for roughly 34 percent of recent adult migrants though results for the whole population typically mirror in 2021. This means that close to 2 million people, results for recent adult migrants. representing one-third of all recent migrants, moved from rural to urban areas between 2016 and 2021. Internal migration in Ethiopia has increasingly Within-urban migration (from one city to another) is been directed towards urban areas. Migration also increasing, accounting for roughly 27 percent to urban areas is the most important form of of all internal migration in 2021. This is in sharp internal migration. This is not surprising considering contrast to earlier times where the bulk of internal continued land fragmentation in Ethiopia (Figure 2. 6). migration happened within rural areas. Migration Average landholdings are small with the poor owning to rural areas has been decreasing over time, with less land than the rich (World Bank 2022), and large rural-to-rural migration halving from 42 percent in cohorts of young people are becoming functionally 2005 to 21 percent in 2021. landless. In addition, as youth education levels rise and access to social media reveals aspirational The general migration patterns are similar for urban lifestyles, migration will increasingly become men and women, but important differences exist. a livelihood strategy for the growing numbers of Rural-to-rural migration decreased while urban- rural youth as they look for off-farm employment. to-urban migration increased for both recent male Figure 2.6: Migration to urban areas is the dominant Figure 2.7: Migration patterns are similar for men form of recent migration, particularly from rural areas and women (Share of recent adult migration by type and time period) (Share of type of migration (recent adult migration) by sex) 45 100 40 19.8 25.4 17.3 23.0 27.7 80 27.7 35 10.8 16.0 30 60 15.1 12.4 23.4 21.6 23.9 28.8 25 40 36.9 37.7 29.6 31.2 20 15 20 45.5 38.0 22.6 22.3 24.0 19.5 10 0 5 2005 2013 2021 2005 2013 2021 0 2005 2013 2021 Female adult Male adult Rural-rural Rural-urban Rural-rural Rural-urban Urban-rural Urban-urban Urban-rural Urban-urban Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. 10 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES and female migrants (Figure 2.7). Between 2005 and To where do Ethiopians migrate? 2021, the share of rural-to-rural migration more than halved from 46 to 22 percent for females and from In Ethiopia, migration within regions is important, 38 to 20 percent for males. The share of urban-to- especially for large regions.¹⁵ Regions with the urban migration increased by roughly 10 percentage largest population shares also had the majority of points for both females and males between 2005 and their migrants come from within the region. In 2021, 2021. Rural-to-urban migration was relatively more 81 percent of all migrants in Oromia, the largest important for males in 2005 but this pattern reversed region according to population, originated from within in 2013 and 2021. Yet, for females, rural-to-urban the region. In Amhara, the second largest region, 87 migration has become the dominant type of migration. percent of its migrants originated from within the The share of women who recently migrated from region. Afar (70 percent), SNNP (70 percent), and rural to urban areas increased from 29 percent in Somali (54 percent) also had more than half of their 2005 to 38 percent in 2021. These changes in pattern migrants originate from within the region. of rural-to-urban migration likely relate to changes in the reasons for why Ethiopians migrate. Oromia, Amhara, and Addis Ababa are important destinations for migrants who move between regions and these three regions alone account for roughly three quarters of all between-region migration. Looking at migration only between regions but not within regions, Amhara and Oromia host roughly 26 percent of all internal migrants respectively and Addis Ababa hosts 23 percent (Table 2.1). On the other hand, considering all migration or gross migration (including migration between regions as well as migration within regions), Amhara hosts 38 percent of all recent internal migrants in 2021, even though only 24 percent of all Ethiopians live in Amhara region. Oromia hosts about 33 percent of all recent internal migrants. ¹⁵ Tigray is not included in the 2021 LMS due to ongoing conflict in the region. Yet, people who came from Tigray in the five years preceding the survey are included. 11 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS Table 2.1: Regional share of destination of recent adult migrants in 2021¹⁶ (Share of gross and net migration of recent adult migrants by type of migration) Type of migration All migration Population share Rural-Rural Rural-Urban Urban-Rural Urban-Urban Gross migration (within region and between region) Region Afar 1.4 0.6 2.1 0.7 1.6 2.0 Amhara 37.6 42.2 34.9 48.9 30.8 23.7 Oromia 33.4 30.9 32.1 28.1 40.0 39.4 Somali 1.1 1.7 0.7 0.3 1.6 6.7 Benishangul-Gumuz 1.7 2.2 2.3 0.7 1.3 1.2 SNNP 16.4 20.8 15.5 20.1 12.0 21.7 Gambela 1.0 1.2 0.6 0.5 1.4 0.5 Harari 0.5 0.1 0.5 0.2 0.9 0.3 Addis Ababa 6.3 10.5 9.4 3.9 Dire Dawa 0.7 0.3 0.9 0.4 1.0 0.5 Only between region Region Afar 2.0 1.3 2.6 0.6 2.5 2.0 Amhara 26.3 47.3 10.2 41.9 23.4 23.7 Oromia 26.1 23.4 20.7 22.3 32.4 39.4 Somali 0.8 0.0 0.9 0.0 1.4 6.7 Benishangul-Gumuz 2.8 7.1 4.5 0.4 1.5 1.2 SNNP 13.5 16.1 5.2 33.3 9.0 21.7 Gambela 1.7 2.7 2.1 0.3 1.8 0.5 Harari 1.6 0.6 2.0 0.5 2.1 0.3 Addis Ababa 23.2 48.4 24.0 3.9 Dire Dawa 2.1 1.6 3.5 0.6 2.0 0.5 Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. Population shares are calculated from survey data and Tigray is excluded. Source: Authors’ estimation based on LMS 2021. ¹⁶ Sidama region was part of SNNP region in the earlier rounds (LFS 2005 and 2013) and for comparison purpose, it is also included as part of SNNPR in the 2021 data here and in subsequent regional analyses. Statistics for Sidama in 2021 can be provided upon request. 12 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Despite the common belief that migration to almost half of all recent migrants who moved Addis Ababa is high and increasing, 2021 LMS between regions. data paints a more nuanced picture. 2021 LMS data show that migration to the capital city is Looking at the entire population of Addis Ababa, only actually similar to that of other predominantly 8 percent of Addis Ababa’s residents were recent urban regions. Considering all recent adult migrants in 2021. When comparing Addis Ababa to migrants, Addis Ababa hosted about 6 percent of other city types (small towns and cities), we see that all internal migrants, a proportion slightly higher the share of migrants in small towns and secondary than its population share of 4 percent (Table 2.1). cities is much larger. Figure 2.8 shows that small towns Even when considering rural-to-urban migration, and secondary cities host twice as many migrants as Addis Ababa only hosts about 10 percent of all the share of their overall resident population. In fact, recent rural-to-urban migrants. On the other small towns host the majority of migrants to urban hand, and as already mentioned, Addis Ababa is areas, underscoring the importance of small towns an important destination for migrants who move as destinations of rural-to-urban and urban-to-urban between regions. Yet, Addis Ababa is an important migration. In 2021, 71 percent of rural-to-urban and destination for migrants who migrate between 69 percent of urban-to-urban migration was to small regions; roughly one-quarter of internal migrants towns, larger than their population share of 63 percent between regions moved to Addis Ababa between (Figure 2.9). Addis Ababa, on the other hand, hosted 2016 and 2021. Addis Ababa is a particularly roughly 11 percent of all rural-to-urban and 10 percent important destination for recent rural-to-urban of urban-to-urban migration, while accounting for 18 migrants who move between regions, receiving percent of Ethiopia’s entire urban population. Figure 2.8: Migration rate for Addis Ababa was lower Figure 2.9: Small towns host proportionately more than for secondary cities and small towns migrants (Share of recent adult migration in respective location) (Share of rural-to-urban and urban-to-urban migration by urban type) 100 Urban Small towns type Secondary 80 Dire Dawa Addis Ababa 70.6 Harari 60 69.2 Gambela 62.5 Regions SNNP 40 Benishangul-Gumuz Somali 19.4 Oromia 20 19.0 21.3 Amhara 10.5 18.1 Afar 0 9.5 0,0 5,0 10,0 15,0 20,0 25,0 30,0 Rural-urban Urban-urban Population share 2021 2013 2005 Addis Ababa Secondary cities Small towns Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years than five years. Adult population refers to people aged 15 years and above. and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. 13 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS While varying, all Ethiopian regions have both to their population size, many more people come originated and received migrants in the years to urban areas than leave them. The opposite preceding the 2021 LMS. We calculated net is true for rural areas, where more people leave migration rates based on the 2021 LMS to see than flow in. Second, there are large differences whether regions had an average net inflow or between regions in terms of net migration, with outflow of migrants.¹⁷ A positive number indicates Somali, Benishangul-Gumuz, SNNP, and Gambela more immigration than emigration to the region having negative net migration between 2016 between 2016 and 2021, whereas a negative and 2021. This means that, relative to the size number indicates more people leaving the region of their population, they experienced the largest than coming to it. First, we observe that net net population outflows. Harari and Dire Dawa migration flows into urban areas are positive experienced large net population inflows, indicating and large (Figure 2.10). This means that, relative that more entered than left the region. Figure 2.10: There are large differences in net- Figure 2.11: Migrants are on average younger migration to regions (Average age by type of migration in 2021) (Net migration flows by region in 2021) 50 40 40 30 30 20 20 10 0 10 -10 -20 0 -30 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban -40 Rural Urban Afar Amhara Oromia Somali B.Gumuz SNNPR Gambela Harari Addis Ababa Dire Dawa Non- Rural origin Urban origin Rural/urban Region migrants migrants migrants Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years than five years. Adult population refers to people aged 15 years and above. and above. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. Who migrates? non-migrants from the same origin area. Migrants are younger than non-migrants; the average age Better understanding migrants’ sociodemographic for migrants is 28 while that of non-migrants profiles can help understand challenges and is 35 (Figure 2.11). This is true for all types of opportunities they face in new destinations. migration, whether to urban or rural areas. While Internal migration is driven by education and women are overrepresented among migrants demographics. The 2021 LMS shows, regardless of (compared to non-migrants) in general, urban-to- whether the origin area is rural or urban, migrants rural migrants are more likely to be men compared are younger and better educated compared to to non-migrants (Figure 2.12). There are no large ¹⁷ The net migration rate is calculated as the difference between the number of people entering (immigrants) and leaving (migrants abroad) a region per 1,000 individuals in the given period. 14 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES differences in terms of the proportion of migrants is found for urban origin areas, where migrants who are married but rural-to-urban and urban- are more literate and better educated than urban to-rural migrants are less likely to be married non-migrants. Migrants originating from urban and compared to other migrants (Figure 2.13). going to rural areas have similar literacy rates than urban non-migrants (roughly 80 percent) but those Literacy rates are relatively high among migrants going to urban areas have higher literacy rates (88 and migrants have much higher education percent). The completed level of education by type of outcomes than non-migrants from the same migration shows similar results. For example, only origin area. In general, rural dwellers who migrate, about 20 percent of rural non-migrants completed either to other rural or to urban areas, have much primary education or more but many more rural higher literacy rates (roughly 60 percent) than migrants completed primary education or more rural non-migrants (46 percent) (Figure 2.14) and (Figure 2.15). About 40 percent of urban-to-urban higher educational attainment. A similar pattern migrants completed at least secondary education, Figure 2.12: Migrants are on average more likely Figure 2.13: There are no large differences in the to be female share of married migrants (Share of females by type of migration in 2021) (Share of married migrants by type of migration in 2021) 70 70 60 60 50 50 40 40 30 20 30 10 20 0 10 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban Non-migrants Rural origin migrants Urban origin migrants Non-migrants Rural origin migrants Urban origin migrants Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years than five years. Adult population refers to people aged 15 years and above. Bars indicate 95 percent confidence intervals. and above. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. Figure 2.14: Migrants originating in rural areas Figure 2.15: Educational attainment is higher for have higher literacy rates than rural non-migrants migrants than non-migrants (Literacy rates by type of migration in 2021) (Share of educational attainment by type of migration in 2021) 100 100 80 80 60 60 40 40 20 20 0 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban 0 Rural Urban Rural-rural Rural-urban Urban-rural Urban-urban Non-migrants Rural origin migrants Urban origin migrants Non-migrants Rural origin migrants Urban origin migrants No education Less than primary Completed primary Completed secondary Completed post-secondary Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years than five years. Adult population refers to people aged 15 years and above. Bars indicate 95 percent confidence intervals. and above. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. 15 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS a share higher than that of urban non-migrants. 2019 (Figure 2.16). The onset of the COVID-19 pandemic In general, education drives migration and higher in Ethiopia in March 2020 caused a significant increase literacy rates and educational attainment for in the arrival of new household members, according to migrants likely mean a relatively easy transition information collected through the World Bank’s Ethiopia into the local labor market. High-Frequency Phone Surveys (HFPS).¹⁹ In April 2020, 25 percent of rural households said that they received How did COVID-19 affect migration? new household members since roughly June 2019²⁰ (Figure 2.17), a proportion higher than pre-pandemic.²¹ COVID-19 seems to have accelerated migration to Moreover, fewer household members left the household rural areas. The COVID-19 pandemic continues to pose (6 percent) compared to pre-pandemic. Nevertheless, serious risks to the health and economic wellbeing in the months following the COVID outbreak, relatively of Ethiopians. Early on, Ethiopia took several steps, few households reported new arrivals month-on- including declaring a state of emergency (SOE) in April month (around 5 percent). Individuals joining the 2020 to halt COVID-19 infections and to stave off the household during the pandemic were similar in their negative repercussions on the economy. Before the sociodemographic characteristics than pre-COVID and COVID-19 pandemic, the share of rural households¹⁸ included primarily sons or daughters of the household that received new household members in the past head under the age of 15 who had little education, most year was 15 percent and the share of households with likely a result of school closings and children moving members leaving was 20 percent, based on the ESS back to their household of origin. Figure 2.16: The share of migrants in the past year Figure 2.17: The share of migrants in the past pre-pandemic roughly 6-9 months is higher than pre-pandemic (Share of rural household from which one household member (Migration flow during COVID-19 in rural Ethiopia) moved away or joined in ESS4) 30 25 25 25 20 20 20 15 10 6 8 65 7 65 6 15 43 42 15 5 3 1 32 21 24 3 0 10 R1 (Apr 20) R2 (May 20) R3 (Jun 20) R4 (Jul/Aug 20) R5 (Sep 20) R6 (Oct 20) R7 (Nov 20) R8 (Dec20) R9 (Jan 21) R10 (Feb 21) R11 (Apr 21) 5 0 Rural % HH with new members joining the household % HH with new members joining the household % HH with members leaving the household % HH with members leaving the household Note: The reference period for April 2020 is since June 2019. The Note: The reference period is the past year. reference period for the following rounds is the previous month/round. Source: Authors’ estimation based on ESS4. Source: Authors’ estimation based on HFPS. ¹⁸ We can only look at rural households as the ESS4 which provides pre-COVID information only asks questions on new household members for rural households. ¹⁹ The World Bank conducted the Ethiopia High Frequency Phone Survey (HFPS) between April 2020 and May 2021; regularly collection of timely data helps monitor the economic and social effects of COVID-19 on households, addressing the effects, and protecting the welfare of the least-well-off Ethiopians. The HFPS is a subsample of the national longitudinal Ethiopia Socioeconomic Survey (ESS), which the Ethiopian Statistics Service and the World Bank carried out in 2019; it is representative of households with access to a working phone. The HFPS tracked the same households over 14 months, with selected respondents, typically household heads, completing phone-based interviews every four to six weeks. Frequent follow-up allows for better understanding of how the pandemic affects households in near real time to support prompt, evidence-based responses. ²⁰ In April 2020, households were asked if new household members joined since they participated in the data collection for the Ethiopia Socioeconomic Survey (ESS4) in June through September 2019. The reference period is therefore somewhere between June or September 2019 through April 2020. ²¹ New births made up roughly 35 percent of new household members. 16 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Compared to pre-pandemic, the share of pandemic; less than 2 percent of migrants left household members who left the household due the country altogether. to economic reasons more than doubled during the pandemic. Before COVID-19, more than half Migration to rural areas was the dominant form of of all household members left for family reasons migration during the pandemic, a stark difference (Figure 2.18). During the pandemic,²² this share to typical migration trends. More than one-third reduced dramatically to roughly 18 percent. On the of new household members moved between rural other hand, one-quarter of household members areas to join the household, and 17 percent of all left the household for economic reasons. During migrants went from an urban to a rural area (Figure the pandemic, this share more than doubled to 2.19). Furthermore, 26 percent moved between urban 55 percent. Moreover, security-related reasons areas and 22 percent moved from a rural to an urban increased during the pandemic. However, this household. During COVID-19, more people still moved could also be related to the increased violence from rural to urban areas than vice versa; however, related to the conflict in Tigray and spillovers this is in stark contrast to recent trends observed using to other areas. Finally, only 2 percent of the 2021 LMS. Though not strictly comparable,²³ many household members left due to health-related more people seemed to have moved to rural areas (52 issues. Furthermore, internal migration was percent) compared to 2021 (37 percent), mainly due to the main type of migration observed during the a larger share of people moving between rural areas. Figure 2.18: Before the pandemic, people mainly Figure 2.19: The largest share of new household members left their household for family reasons moved between rural areas during the pandemic (Share of household members leaving the household by reason (Share of type of migration during COVID-19 for household for migration) members who joined the household) 60 55 100 53 50 26 80 40 17 60 30 22 23 40 20 18 18 16 10 7 9 20 35 0 1 1 0 0 0 Family Job Security Health Education Other National reasons reasons Rural-rural Rural-urban Rural Pre-COVID Rural COVID-19 Urban-rural Urban-urban Note: The pre-COVID estimate refers to June/September 2019. Note: Reasons for joining the household were asked only in The COVID-19 estimate refers to round 8 of the HFPS conducted round 8 of the HFPS conducted in December 2020. in December 2020. Source: Authors’ estimation based on HFPS. Source: Authors’ estimation based on ESS4 and HFPS. ²² Round 8 of the HFPS included reasons for migration, conducted in December 2020. ²³ HFPS looks at new household members over 6-9 months while LFS trends look at migrants joining the household in the past 5 years. 17 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS Family-related issues were the main reason for reduce poverty and increase shared prosperity. joining a household during the pandemic. This was The differences in earnings between countries especially the case for rural-to-rural migration, and are large and provide a strong incentive for to a lesser extent for those moving from urban-to- people to migrate from one country to another. urban areas (Figure 2. 20). In contrast, among those The income disparity is over 20-fold between who went from urban-to-rural or rural-to-urban, countries at the top and bottom decile of the global the main reason was economic. Furthermore, income distribution (e.g., Australia vs Tanzania). domestic work motivated migration to urban areas, The incidence of poverty and deprivation is also especially if coming from rural areas. Nevertheless, vastly different across countries, providing strong after joining a household, only one-third of new motivation for people to move across countries to household members reported working in the same improve their wellbeing. type of job as in the previous location. Finally, health-related issues were important factors to join One goal of Ethiopia’s Job Creation Commission an urban household during the pandemic. (now part of the Ministry of Labour and Skills) was to facilitate 150,000 Ethiopians into good Figure 2.20: The largest share of new household jobs abroad by the end of the Ethiopian fiscal members moved for family-related reasons year 2019/2020. This is a significant shift, given (Share of household members joining a household by reason for migration and type of migration) that historically the GoE’s primary policy stance towards labor mobility was one of deterrence. Urban to urban 26 23 18 19 10 Indeed, from 2013 to 2018, GoE banned out- Urban to rural 27 41 11 4 13 migration of labor migrants to the Gulf Cooperation Council (GCC) countries, Ethiopia’s primary Rural to urban 9 35 5 21 17 11 destination markets.²⁴ In response to the growing Rural to rural 54 18 10 11 4 3 youth population seeking work, alongside lifting the 0% 20% 40% 60% 80% 100% ban, the GoE is now interested in promoting quality Family Job reasons Security employment opportunities for Ethiopian jobseekers Domestic work Health Education Other abroad, but labor migrant outflows remain small relative to the need. Outflows largely halted during Note: The reference period for April 2020 is since June 2019. The reference period for the following rounds is the previous the COVID-19 pandemic, and while GoE still has a month/round. policy of promoting labor mobility as a job strategy, Source: Authors’ estimation based on HFPS. it paused on active steps towards this goal (Smith et al. 2020).²⁵ 2.2 EVIDENCE ON INTERNATIONAL LABOR MIGRATION Ethiopian mobility has a complex history, with many factors motivating migration decisions as International labor migration leads to immediate well as the modality and choice of destination. and large wage increases dwarfing benefits Historically, Ethiopian migration has been from any other development intervention. For predominantly due to displacement,²⁶ beginning migrants and their families, gains in wages can with the 1985 famine; in recent years, however, ²⁴ The 2013 ban was prompted by a deportation campaign, in which Saudi Arabia began to deport high numbers of Ethiopian migrants. ²⁵ For highlight on the mobility framework for international labor migration in Ethiopia, please refer to Annex 1. ²⁶ The 1980s to early 2000s was a period of high out-migration from Ethiopia; however, these flows were predominantly composed of displaced persons. In the late 1980s and 1990, Ethiopian emigration was at its highest level; 1.6 million Ethiopians were living abroad, with 1.3 million of these being Ethiopians who were involuntarily displaced according to UNHCR, as of 1990. At this time, Sudan was the primary destination, hosting over 50 percent of Ethiopians abroad (over 900,000 in 1990). By 2000, the stock of Ethiopian migrants declined significantly. This appears to be due to Ethiopian refugees and asylum seekers returning to Ethiopia, as UNHCR reported a significant decline in the stock of refugee and asylum seekers at that time. 18 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES voluntary migration has become increasingly rate of emigration is still higher for urban areas important. After a period of migration of primarily (14 migrants per thousand residents) due to the refugees and displaced persons, both stocks and relatively small urban population share compared flows of migrants from Ethiopia increased in the to rural areas (7 migrants per thousand residents).²⁸ years following 2000. UNDESA data reports that There is significant variation in emigration rates between 2000 and 2005, outflows rose to 15,000 between regions with predominantly urban regions migrants per year on average; these outflows then (Addis Ababa, Dire Dawa, and Harari) showing the peaked between 2005 to 2010 with an average highest emigration rates. From the pre-dominantly of ~28,000 migrants per year before declining to rural regions, SNNP and Gambela have the highest 23,000 migrants annually between 2010 and 2015 rates (around 10 migrants abroad per each thousand and declining further to 18,000 annually between resident) while Somali has the lowest rates of 2015 and 2019. Yet, these outflows are small and 3. Considering recent migrants abroad—those pale in comparison to the two million additional migrating between 2016 and 2021—the stock of working-age Ethiopians annually the labor market migrants is lower but still sizable at 470,000. must absorb. These flows also pale in comparison to similar countries (Figure 2.21) that have actively pursued migration as an employment strategy to access jobs, improve wages, and increase remittances (Smith et al. 2020). Close to 850,000 Ethiopians lived abroad in 2021, most of them originated from rural areas. Based on the 2021 LMS, an estimated 840,000 Ethiopians are living abroad,²⁷ which translates to an emigration rate of 9 migrants abroad for every thousand Ethiopians (Table 2.2). About two-third of the migrants abroad came from rural households (Figure 2.22) but the Figure 2.21: Migrant stock in Ethiopia is low in international comparison (Migrants as as share of population) 7.2 5.2 3.9 2.0 1.7 1.1 1.3 Ethiopia Kenya Uganda South Africa Bangladesh Nepal Thailand Source: UNDESA Population Division (Migrant Stock – 2019). ²⁷ The 2021 LMS does not include migrants abroad from Tigray. The total number of migrants abroad is therefore an underestimate. ²⁸ The emigration rate is likely to be under-estimated because the survey only includes those migrants abroad that left their households/ family behind and not those who emigrated with the whole family (the survey asks if anyone left the household and went abroad). This understatement is expected to be particularly high for urban areas where is it common for the whole household to emigrate, thus the share of urban areas and the rate is particularly likely to be underestimated. 19 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS Table 2.2: For every thousand Ethiopian residents, there are nine lifetime migrants abroad (Rate of emigration by location) Number of Population Rate of emigration migrants abroad size per 1000 residents National 839,224 98,000,000 8.6 Urban 288,409 21,000,000 13.7 Rural 550,815 77,100,000 7.1 Afar 11,098 1,989,167 5.6 Amhara 211,604 23,200,000 9.1 Oromia 309,700 38,600,000 8.0 Somali 17,077 6,605,361 2.6 Benishangul-Gumuz 6,442 1,200,471 5.4 SNNP (incl. Sidama) 174,283 21,365,655 8.2 Gambela 4,779 498,848 9.6 Harari 4,290 273,640 15.7 Addis Ababa 91,061 3,804,071 23.9 Dire Dawa 8,891 537,345 16.5 Note: The 2021 LMS does not include Tigray due to conflict in the country and migrants abroad from Tigray are not captured. Source: Authors’ estimation based on LMS 2021. Where do Ethiopians migrate to abroad? average length of stay for migrants abroad is six years but only four and a half years for Saudi Arabia In 2021, Ethiopian migrants abroad concentrated and four years for other Middle Eastern countries. in a few locations, mostly in the Middle East. More Considering recent migrants abroad, the spatial than half of all migrants abroad were in the Middle pattern is similar but the share of migrants in the East (31 percent in Saudi Arabia and 25 percent Middle East is much higher; almost 70 percent of all in other Middle Eastern countries), 12 percent in recent migrants went to Saudi Arabia (34 percent) South Africa, 8 percent in the United States, and 7 or other Middle Eastern countries (34 percent). percent in neighboring countries (Figure 2.23). The Considering the importance of the Middle East for Figure 2.22: Two thirds of lifetime migrants abroad Figure 2.23: Ethiopian lifetime migrants abroad originate from rural areas are concentrated in few locations (Number of migrants abroad by source location) (Number of recent migrants abroad by destination country) 550,815 South Others, Africa, 127,967 103,799 Other Middle 288,409 Eastern Saudi Arabia, countries, USA, Neighbors, 257,220 213,468 69,985 62,741 Urban Rural Source: Authors’ estimation based on LMFS 2021. Source: Authors’ estimation based on LMS 2021. 20 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES recent migrants, we will report results separately with that of current household members of origin for Saudi Arabia and other Middle Eastern households (households who have at least one countries. Many Ethiopian women migrate to the migrant abroad), migrants abroad are more educated Middle East to improve their living conditions, and than households members remaining in Ethiopia. they are typically employed in low-skilled jobs, such While almost all migrants abroad of working age as domestic work (Zewdu 2018). have at least some primary education, 32 percent of the working age non-migrants abroad have no Who migrates abroad? education at all. The percentages of those who completed primary school or above is 47 percent for About an equal number of men and women migrate migrants abroad but only 32 percent for household abroad, but a larger share of women migrates to members who stay behind. Similarly, while recent Middle Eastern countries (other than Saudi Arabia). migrants abroad average seven years of education, International labor mobility is equally important for non-emigrants averaged only five years (Figure men and women, with about 49 percent of all recent 2.25). However, those migrating to the Middle East, migrants being women (Figure 2.24). There is one particularly to Saudi Arabia, are less educated than notable exception. A larger share of women (60 other migrants abroad. percent) migrates to Middle Eastern countries (other than Saudi Arabia), likely associated with the large Out-migration from Ethiopia is unlikely to result in number of women who migrates to the Middle East “brain drain”. There is an important concern about as domestic workers. Migrants abroad are younger “brain drain”, arguing that out-migration of skilled than the average Ethiopian. The average age is about individuals (typically university graduates) depletes 24 years for recent migrants abroad but 34 years for the stock of human capital, results in shortages of all Ethiopians. skilled works, and hurts prospects of economic development in sending countries. While roughly The most educated household members migrate one-third of Ethiopian migrants to OECD countries abroad but migrants to Saudi Arabia are less was highly-skilled in 2015 (d’Aiglepierre et al. 2020), educated than other migrants abroad. Based on migration to OECD countries is low. As reported, the intra-household evidence, there seems to be a clear largest number of migrants goes to GCC countries, selection in terms of the households’ choices on which is dominated by low-skilled workers. Education whom to send abroad. When comparing education levels for migrants abroad, even if higher than attainment of migrants abroad upon leaving Ethiopia residents, are typically below secondary education. Figure 2.24: About an equal number of men and Figure 2.25: Migrants abroad have more education women migrate abroad than non-emigrants (Share of female migrants abroad) (Mean years of schooling for recent migrants abroad and non-emigrants) 60.2 8,0 7.3 48.6 7,0 6.4 6.6 47.1 6,0 5.1 5,0 4,0 3,0 2,0 1,0 0,0 All Saudi Other Middle Non-emigrants All Saudi Other Middle Countries Arabia Eastern Countries Arabia Eastern Countries Countries Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. than five years. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. 21 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS The overwhelming majority of migrants How does international abroad—nine out of ten—left for economic migration affect remittances? reasons, leading to high employment rates while abroad. Nine of ten migrants left Ethiopia International labor migration is an important source for a different country for economic reasons. of remittances, which improve development outcomes Migrants to the Middle East almost exclusively in origin countries. Part of the increase in income of leave for economic reasons (Figure 2.26). Other migrants abroad is shared with family and community reasons to migrate are family reasons (4 percent) members left behind through remittances. The benefits and education (2 percent). Not surprisingly, the from labor migration stem from remittances, more vast majority of migrants abroad are employed, efficient labor allocation, and knowledge transfers. given their intent of finding employment abroad; Remittances enhance migrant households’ welfare – about 80 percent of the migrants abroad are raising income and consumption, allowing more spending employed,²⁹ 7 percent are unemployed, and 8 on children’s education, and increasing opportunities percent are studying (Figure 2.27). The proportion to start businesses. Remittances can also enhance of employed migrants abroad is even higher macroeconomic stability, buoy economic dynamism, for those based in the Middle East, particularly and reduce poverty. But the benefits vary. As migration is in countries other than Saudi Arabia. The costly, the poorest are less likely to migrate internationally percentages of employed are 83 percent and 88 and remittances to better-off households could worsen percent, respectively, for Saudi Arabia and other existing disparities. However, even the poor left behind Middle Eastern countries. benefit through economy-wide effects of migration. Figure 2.26: The overwhelming majority of migrants Figure 2.27: The vast majority of migrants abroad is abroad for economic reasons is employed (Share of recent migrants abroad migrating for a specific reason (Share of migrants abroad by activity status and location) by location) 100% 100% 80% 80% 60% 60% 40% 40% 20% 20% 0% 0% All Saudi Other Middle All Saudi Other Middle Countries Arabia Eastern Countries Arabia Eastern Countries Countries Economic Education Family related Others Employed Unemployed Student Others Don't Know Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. than five years. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. ²⁹ Employment status is based on proxy response by household members who stayed in Ethiopia. 22 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Remittances are an important part of the Ethiopian Nevertheless, remittances started to recover in the economy, contributing 5 percent of GDP. According second half of 2020, and in 2021 SSA remittances to the National Bank of Ethiopia,³⁰ in 2020/21, the increased by 14 percent, a proportion below only country received around US$5 billion of remittances Latin America and the Caribbean (LAC) (25.3 (Figure 2.28), contributing to around 4.5 percent to percent) and higher than in Europe and Central Asia GDP (Figure 2.29). This is considerably above the (ECA) (7.8 percent), Middle East and North Africa remittances reported by the Central Banks of Kenya (MENA) (7.6 percent), South Asia (6.9 percent), and (US$3 billion) and Uganda (US$1 billion), where East Asia and the Pacific (EAP) (excluding China) remittances contributed to around 3 percent of GDP. (2.5 percent) (KNOMAD 2022). Ethiopian remittances grew consistently between 2011 and 2018, but decreased slightly thereafter as a result of the COVID-19 pandemic. In 2018/19, remittances reached an all-time high of US$5.3 billion, but subsequently decreased to US$4.3 billion in 2019/20—when remittances declined at the start of the COVID-19 outbreak for many countries— before recovering in 2020/21 to roughly US$5 billion. In Ethiopia, remittances fell by around 14 percent, a greater decline than the SSA average (12 percent) (KNOMAD 2022), and countries like South Africa 9 percent), but below Uganda (25 percent). Figure 2.28: Remittances in Ethiopia are increasing Figure 2.29: The contribution of remittances to (Personal remittances in current US$ million) Ethiopia’s GDP is roughly 5 percent (Personal remittances as a share of GDP) 6,000 4.5% 5,000 4,000 3,000 3.1% 2.9% 2,000 1,000 0 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 Ethiopia Uganda Kenya Ethiopia Uganda Kenya Source: Respective Central Bank using information on Source: Authors’ estimation based on Central Bank information. personal transfers as reported in the Balance of Payments. The annual amount of remittances corresponds to the Fiscal The annual amount of remittances corresponds to the Fiscal Year in each country. Year in each country. ³⁰ The remittances reported here corresponds to personal transfers made by individuals as reported in the balance of payment of the National Bank of Ethiopia (NBE). For Ethiopia, numbers on remittances differ considerably between NBE and those reported in the World Development Indicators (WDI) database. This report includes figures based on NBE’ balance of payment reports. 23 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS Four-fifths of Ethiopia’s remittances come Figure 2.30: Four fifths of Ethiopia’s remittances from just ten countries. Not surprisingly given come from just ten countries (Top 10 Economies sending remittances to Ethiopia, percent) its history of emigration to the United States in the 1980s, Ethiopia receives one-third of all United States 34 remittances from the United States (Figure 2.30). Saudi Arabia 21 Yet, Saudi Arabia is also an important source of Israel 9 remittances since so many Ethiopians go there Sudan 7 as international labor migrants; one-fifth of all Canada 4 remittances come from Saudi Arabia. Other Sweden 3 countries from which a relatively large share of Italy 3 Australia 2 remittances originate are Israel (8 percent) and Netherlands 2 Sudan (7 percent) (UNCDF 2021). Germany 2 0 10 20 30 40 2.3 EVIDENCE ON RETURN MIGRATION Source: UNCDF (2021), using data from IMF Balance of Payments The vast majority of international labor migration Statistics database and data releases from central banks, national statistical agencies, and World Bank country desks, is temporary and migrants struggle to reintegrate November 2020. upon return to Ethiopia. As noted, migrants abroad remain, on average, in their country of destination The GoE projects that the number of returnees, for six years before returning to Ethiopia. Programs particularly from Saudi Arabia, will be large in the or efforts to facilitate migration typically focus on coming months. Since 2017, Ethiopia has received the earlier stages of the migration cycle, including large flows of returnees from Saudi Arabia due to establishment of effective labor migration channels a Saudi initiative to curb unemployment. However, and pre-departure training, but much less attention by the start of 2022, it was estimated that about is paid on the process of return and reintegration 750,000 Ethiopians were still in Saudi Arabia, with of migrant workers. Many migrants struggle to 450,000 of them needing help in their return to reintegrate upon return in their home country. A Ethiopia. As a response, starting on March 30, 2022, prolonged stay abroad can create challenges for the GoE plans to repatriate over 100,000 Ethiopians returnees to reintegrate into the labor market, living in Saudi Arabia (IOM 2022). About 30,000 reattach to prior networks, and using human and Ethiopians were already repatriated between the financial resources. In urban and rural areas alike, start of the campaign and the end of May (Ethiopian returning migrants encounter significant challenges Monitor 2022). finding productive employment in Ethiopia’s better educated, young labor market. A growing workforce— Based on the 2021 LMS, about 580,000 possibly increased by returnees—pressures the Ethiopians returned from abroad and settled absorption capacity of the Ethiopian labor market. equally in urban and rural areas. The pool of Even though Ethiopia has sustained high economic migrants abroad who returned was large with growth and labor market conditions have improved six returnees per 1,000 residents. The number of over time, challenges remain, including lack of labor migrants returning to urban areas is almost the market inclusion and productive employment for same as the number of migrants returning to rural youth³¹ and women; lack of structural transformation areas (Table 2.3). Given a much larger population out of agricultural employment in rural areas; and in rural areas—roughly 80 percent of Ethiopians high unemployment in urban areas (Wieser and live in rural areas—the rate of return in urban Mesfin 2021). ³¹ Youth is defined as those aged 15-24. 24 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES areas is much higher at 14 returnees per thousand 1,000 residents and lowest in Somali with just 1 residents compared to 4 returnees per thousand returnee per 1,000 residents. However, it is also residents in rural areas. Close to 80 percent of worth noting that Addis Ababa’s rate of return is all returnees settled back into the two largest lower than the urban average, which implies urban regions, Oromia and Amhara. Yet, the rate of return areas other than Addis Ababa host more returnees was highest in Addis Ababa at 11 returnees per relative to their population share. Table 2.3: About 580,000 Ethiopians returned from abroad between 2016 and 2021 (Rate of return by location) Returnees Population Rate of return (per 1000 residents) National 575,442 98,000,000 5.9 Urban 285,529 21,000,000 13.6 Rural 289,913 77,100,000 3.8 Afar 5,217 1,989,167 2.6 Amhara 210,075 23,200,000 9.1 Oromia 241,810 38,600,000 6.3 Somali 3,403 6,605,361 0.5 Benishangul-Gumuz 2,646 1,200,471 2.2 SNNP (incl. Sidama) 64,330 21,365,655 3.0 Gambela 1,882 498,848 3.8 Harari 734 273,640 2.7 Addis Ababa 43,143 3,804,071 11.3 Dire Dawa 2,203 537,345 4.1 Note: The 2021 LMS does not include Tigray due to the conflict in the country and migrants abroad from Tigray are not captured. Source: Authors’ estimation based on LMS 2021. Close to 80 percent of returnees are from the Even when voluntary, we have limited evidence Middle East. Just under half of all returnees come on successful reintegration of returnees, from Saudi Arabia alone, 9 percent of the returnees particularly as many Ethiopians migrate abroad come from United Arab Emirates (UAE), and 23 informally. Reintegration success, as defined on percent come from other Middle Eastern countries whether a returnee successfully integrates into (Figure 2.31). About 13 percent come from the labor market and society, depends on the neighboring countries. The potential for returnees sociodemographic characteristics of the returnee; to reintegrate into Ethiopian society and the labor how well off they were prior to return; the extent market vary depending on sociodemographic to which returnees maintained social networks characteristics. Returnees are more likely to be in the country of origin while abroad; but also on female (63 percent) and relatively younger (31 the way in which they return. A large proportion average age) than the overall population. Yet, of Ethiopians returnees migrated without required important differences exist depending on the documentation and deportation is therefore source country. For example, more than 90 percent prevalent; only 61 percent of all returnees had a of returnees from the UAE are female, given the passport before migrating, of which 70 percent large share of domestic work in UAE. had the required visa. In other words, only 42 25 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS percent had a valid passport and visa for their Successful reintegration of returnees could travel. This indicates the prevalence of informal/ be an engine for positive change in Ethiopia illegal migration, which increases the likelihood of a but returnees tend to have poor labor market challenging experience during their stay abroad and outcomes. To better understand effectiveness of possible challenges in reintegration upon return. returnee reintegrating into the Ethiopian labor Returnees from Saudi Arabia are particularly prone market, we look at returnees’ employment status to not having the right documents. before and after international labor migration.³² A majority of returnees were not employed prior to Given the prevalence of informal/illegal migrating abroad; about 45 percent of returnees migration, it is not surprising that deportation is reported they were unemployed and almost 20 the most common reason for Ethiopian migrants percent were studying prior to migration. Of those to return home. About 31 percent of returnees to who were employed (36 percent), only 20 percent Ethiopia were due to deportation, an additional were engaged in wage employment, of which more 28 percent returned due to end of contract or than half worked as domestic workers. The majority lack of employment, and 9 percent retuned due of self-employed worked in agriculture. Figure 2.33 to health-related reasons, including pregnancy shows returnees’ activity status pre-migration (left- (Figure 2.32). This means that more than half of hand side) and post-migration (right-hand side). We all returnees return with “unfavorable” conditions note several key findings: for reintegration. Moreover, reintegration programs, which can support returnees in their i. Many returnees moved in and out of categories. reintegration effort, are essentially non-existent. ii. A large number of returnees to Ethiopia were Only 2 percent of all the returnees received unemployed.³³ support upon return. iii. Very few returnees engaged in self-employed, Figure 2.31: Close to 80 percent of returnees came Figure 2.32: About one third of returnees was deported back from the Middle East (Main reason for return) (Share of returnees by source country) 100% 80% 60% 40% 20% Neighbors, UAE, 0% All Countries Saudi Arabia UAE 12.6 8.8 Deportation End of study Other ME, End of contract/unemployed Social/political Saudi Arabia, 47.2 22.7 Others, 8.7 Family reunion Health related Others Note: Returnees are individuals who returned to Ethiopia Note: Returnees are individuals who returned to Ethiopia between 2016 and 2021. between 2016 and 2021. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. ³² Using 2021 LMS, we can look at the employment status before migrating to the country of destination (based on information in the returnee module) and compare with the employment status at time of interview in 2021. ³³ Activity status pre-migration is based on self-reports and not based on the “standard” labor market module which gathers employment and unemployment indicators used in the LFS for the current labor market status (i.e. post-migration). 26 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES non-agriculture work after their return, In most cases they received regular payments and indicating difficulties in using resources earned more than half of migrants received housing and abroad to set up businesses. food that was not deducted from the usual payment. iv. A similar number of returnees engaged in wage Nevertheless, returnees reported long duration of employment before and after employment, but work, with one of five mentioned working all days in they are not the same people as movements a week. Yet, half of returnees said some of their rights from wage employment to other forms of were restricted during work: one-third said they employment and vice versa are prevalent.. could not communicate their views freely; another v. Many returnees continue their education as 8 percent were unable to talk to people outside the students upon return. job; 6 percent said that job security was not ensured; and 4 percent felt that their religious beliefs were Many returnees from Saudi Arabia worked in curtailed. Furthermore, only 30 percent had at least elementary occupations and expressed that one rest day a week, and half of those injured received some of their rights were limited during their payment for the days they could not work. Moreover, work abroad. Though the 2021 LMS does not one-third of returnees made informal payments provide information on type of employment during to get a job. Considering all costs incurred to work migration, according to joint 2015 Global Knowledge abroad—visa, passport, transportation, recruitment Partnership on Migration and Development agency, informal payments, among others—women (KNOMAD) and International Labour Organization reported having spent around ETB 14,000 in 2015 (ILO) Migration Costs Surveys, once in Saudi Arabia, (roughly US$ 770 today), while men spent roughly many returnees worked in elementary occupations. ETB 22,000 (roughly US$ 1,200). Figure 2.33: A majority of returnees were not employed prior to migrating abroad (Activity status of returnees pre-migration (left-hand axis) and post migration (right-hand axis) Note: Returnees are individuals who returned to Ethiopia between 2016 and 2021. Source: Authors’ estimation based on LMS 2021. 27 2. OVERVIEW OF MIGRATION IN ETHIOPIA: SCALE, PATTERN, AND TRENDS Among other reintegration challenges, returnees underlying structural disadvantages, or this could are twice as likely to be unemployed relative to be related to high reservation wages³⁶ as their wage non-returnees. About 81 percent of all working- expectations may be elevated given their experience age returnees are active in the labor market abroad. Among those employed, the quality of upon return, higher when compared to the whole employment is also mixed (Figure 2.35);³⁷ returnees population at 74 percent. Yet, returnees are are less likely to be in wage employment but more nearly twice as likely to be unemployed than non- likely to be engaged in non-farm self-employment returnees (28 percent compared to 17 percent) compared to non-returnees, particularly in urban (Figure 2.34).³⁴ These differences, which we observe areas. These outcomes are most likely related to the in both urban and rural areas, suggests that unfavorable labor market conditions and about one- returnees are disadvantaged in the Ethiopian labor quarter of returnees say that they intend to go back market. This could be related to their labor market abroad. In rural areas however, both returnees and status prior to migrating—40 percent reported non-returnees are equally likely to work in similar being unemployed prior to migration³⁵—indicating types of employment. Figure 2.34: Returnees are nearly twice as likely to Figure 2.35: The quality of employment for returnees be unemployed is mixed (Predicted probabilities of unemployment: returnees vs. non-returnees) (Predicted probabilities by type of employment) 40 100% 35 80% 30 60% 40% 25 20% 20 0% 15 Non- Returnees Non- Returnees Non- Returnees returnees returnees returnees 10 National Urban Rural 5 0 Wage-employed Self-employed agriculture National Urban Rural Self-employed non-agriculture Employer Non-returnees Returnees Unpaid family worker Other Note: Returnees are individuals who returned to Ethiopia Note: Returnees are individuals who returned to Ethiopia between 2016 and 2021. between 2016 and 2021. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. ³⁴ We estimate the probability of being unemployed, controlling for individual, household and locational characteristics. Details are reported in Annex 2 Table A 1. ³⁵ Though their current unemployment rate, which is based on detailed labor market questions, is not comparable with their self- reported unemployment rate before migrating, the result suggests that unemployment rate among returnees was high even before they migrated abroad. ³⁶ Reservation wage is the lowest wage rate at which a worker is willing to accept a certain type of job. ³⁷ We estimate a multinomial logit of the type of employment returnees and non-returnees are engaged in, controlling for individual, household and locational characteristics as before. Details are reported in Annex 2 Table A 2. Note that the estimates which consider the urban and rural sample separately have been omitted from this report but available upon request. 28 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES 3. MOTIVES AND IMPACTS OF MIGRATION This section of the report highlights drivers Large differences in earnings and productivity of migration and discusses migration effects across areas within a country and across countries on migrants, their families, and destination represent strong motivators for people to move, as locations with a focus on internal migration. This well as significant differences in amenities, such as section answers four key questions by taking an access to infrastructure and services. in-depth look at reasons for migration as well as synthesizing literature on the motives of migration People migrate in response to both “push” and and institutional barriers of internal migration: “pull” factors. “Push” factors are conditions that impel people to move out of their places of residence. i. What drives people to migrate? These can include land scarcity, poverty, lack of ii. What are the main institutional barriers public services and infrastructure, or high cost of migrants face? living. “Pull” factors are conditions that refer to the iii. How does migration effect different dimensions availability of better opportunities or circumstances of welfare of migrants and their families? that attract migrants to a specific place. This can iv. What are the effects of migration on areas of include favorable labor market outcomes or higher migrants’ destinations? incomes. This section looks at the characteristics of migrants and the main push and pull factors based 3.1 WHAT DRIVES PEOPLE TO MIGRATE? on analysis of the 2021 LMS and the literature on migration in Ethiopia. Mobility represents a key part of the history of humankind. Some people move within their Why do people migrate? own country, others cross a border, but similar forces drive the two patterns. Migration stems Ethiopians migrate mainly for economic from people’s desire to improve their wellbeing. reasons.³⁸ Irrespective of the type of migration, ³⁸ “Economic reasons” include job search/offer and shortage of land where the dominant on is the first one; in 2021 and 2013, shortage of land accounted only for two percent and three percent of all migration, respectively. The role of shortage of land was higher in 2005 (15 percent of all migration and 36 percent of economic migration). 29 3. MOTIVES AND IMPACTS OF MIGRATION moving for economic reasons has become the main migration motivator for men, while both increasingly frequent over time, especially after searching for work and marriage are important 2005, which partly reflects expansion in education motivators for women. in recent decades (educated people are more likely to move). In the five years up to 2021, 43 percent of Many rural-to-urban migrants cross regional migrants mentioned economic reasons (including boundaries when migrating for economic reasons. job search, job loss, and shortage of land) as their Between 2016 and 2021, rural-to-urban migrants main reason for migrating (Figure 3.1). Over half who migrate to another region than their region of of rural-to-urban migrants moved for economic origin, migrate mainly for economic reasons (57 reasons between 2016 and 2021, while shortage percent) (Figure 3.2). On the other hand, rural-to- of land is an important motivator for rural-to- urban migrants who migrate to another location rural migration, especially for men. Migrating for within the same region, are almost equally likely education has become less important over time, to migrate for family (38 percent) or economic (39 while migrating due to shocks, such as conflicts percent) reasons. For men, economic migration is and natural disasters, became more important in the most important type of rural-to-urban migration. 2021. Between 2016 and 2021, approximately 7 For women, economic reasons are more important percent of all internal migrants left their homes for between-region migration, while family reasons because of shocks. Other important motivations are more important for within-region migration. to migrate are family reasons—31 percent of adult Irrespective of sex, education is more important for migrants moved for family reasons between 2016 within-region migration. and 2021—which include marriage-related reasons and moving along with or to join family, especially Large cities tend to be a magnet for economic among young people. Looking for work is by far migrants from rural areas. Addis Ababa is the main Figure 3.1: People mainly migrate for economic reasons Figure 3.2: Many rural-to-urban migrants cross regional (Reasons for migration for recent migrants by year and sex) boundaries when migrating for economic reasons (Reasons for migration for rural-to-urban migrants between 100% and within regions by year and sex) 80% 100% 80% 60% 60% 40% 40% 20% 20% 0% 2013 2021 2013 2021 2013 2021 2013 2021 2013 2021 2013 2021 0% All adults Female Male All Adult Female Male 2005 2013 2021 2005 2013 2021 2005 2013 2021 Between regions Within regions All adults Female Male Economic Education Family related Economic Education Family related Shocks Others Shocks Others Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. 30 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES destination when rural-to-urban migrants move for Though the share of youth migrants is significantly economic reasons, followed by secondary cities. higher than their population share, the relative Between 2016 and 2021, 73 percent of rural-to- importance of youth migration has decreased urban migration to Addis Ababa was for economic over time. In 2005, the share of migrants defined reasons and 46 percent to secondary cities. For as “youth”, those aged 15 to 24, accounted for 55 rural-to-urban migration to small towns, the share percent of all recent internal migrants, larger than of economic and family-related migration were the the share of the population who was youth which same (39 percent). was 35 percent (Figure 3.3). The share of migrants who were youth dropped to 45 percent in 2021— Why do youth migrate? though the absolute number of youth migrants is higher in 2021 than in 2005 due to population Youth are more likely to migrate, particularly growth—while their population share decreased to urban areas. As youth education levels rise only marginally to 32 percent. We observe a similar and access to social media reveals aspirational pattern for rural-to-urban migration; about three- lifestyles in urban areas, migration will become fourth of the rural-to-urban migrants were youth in a livelihood strategy among the growing rural 2005 but their share fell to just about 50 percent in youth as they look for off-farm employment in 2021. This is mainly due to a decrease in migration urban areas. Youth move for a myriad of reasons, for education reasons since 2005, resulting from including lack of job opportunities in their home school expansion in rural areas over the past residences, hopes for improving their lives, fifteen years. Rural-to-urban youth migration is search for economic opportunities to support usually to small towns, representing 65 percent of families, lack of access to land, and lack of youth rural-to-urban migration in 2021, while 21 educational facilities in their place of residence percent of youth migrated to secondary cities, and (Bundervoet 2018). 14 percent to Addis Ababa (Figure 3.4). Figure 3.3: Youth’s migration share is higher than their Figure 3.4: Rural-to-urban migrant youth are more population share though the gap narrowed over time educated than rural non-migrant youth (Share of youth among recent migrants and among population) (Share of rural-to-urban migration for youth by educational attainment) 80 70 100% 2.3 60 12.3 7.0 23.4 80% 50 40.0 60% 45.7 40 40% 52.9 30 37.7 20% 29.7 20 20.8 0% 5.7 10.4 10 Rural Urban Rural to urban 0 non-migrants non-migrants migration 2005 2013 2021 No education Less than primary All migration types Rural-urban migration Completed primary Completed secondary Population share Completed post-secondary Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years than five years. Adult population refers to people aged 15 years and above. and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LMS 2021. 31 3. MOTIVES AND IMPACTS OF MIGRATION Rural-to-urban youth migrants in 2021 were more the role of education as a reason for migration has likely to be female and married compared to non- decreased since 2005, it is still significant, with 16 migrant youth. Migrant youth, while having more percent of youth migrating for education reasons education compared to rural youth, are still less in 2021. Young men tend to migrate for economic educated than urban youth. In 2021, 72 percent of reasons while young women tend to migrate for young rural-to-urban migrants were female, higher family reasons. Looking at types of migration for than for rural non-migrant youth at 48 percent, youth, economic reasons are most important for and urban non-migrant youth at 55 percent. About rural-to-urban migration in 2021. Not surprisingly, 29 percent of rural-to-urban youth migrants are youth who migrated from rural to urban areas married, compared to 23 percent for rural youth for economic reasons have more favorable labor and 17 percent of urban youth, respectively. In 2021, market outcomes than those who migrated for other educational attainment improved for all groups reasons. The labor force participation rate in 2021 but gaps in primary school completion remain was 90 percent for young migrants who migrated significant between rural and urban youth. Migrant for economic reasons. Unemployment rates were youth have much higher primary school completion also lower for rural-to-urban migrants who left their rates (52 percent) than their counterparts who homes for economic reasons. Though still relatively remain in rural areas but this is still lower than for high at 15 percent and higher than for rural youth urban youth (65 percent) (Figure 3.4). at 8 percent, the unemployment rate of all rural-to- urban migrant youth was 22 percent and for urban Youth migrate mainly for family-related or non-migrant youth was 27 percent (Figure 3.6). economic reasons and unemployment rates for There is a common belief that young migrants rural-to-urban migrants are lower than those of arrive in cities only to be unemployed, triggering natives. In 2021, more than two-thirds of youth social issues. Yet, we see that unemployment rates migrated for family-related reasons (42 percent) or for young rural-to-urban migrants are lower than economic reasons (36 percent) (Figure 3.5). Though those of natives. Figure 3.5: Youth migrate mainly for family and Figure 3.6: Economic migration for the youth is associated economic reasons with higher activity and lower unemployment rates (Share of reasons of migration for adults and youth) (Youth unemployment rate for migrant and non-migrant youth who migrated for economic reasons) 100% 26.8 90% 80% 22.3 70% 60% 50% 15.1 40% 30% 20% 10% 7.7 0% 2005 2013 2021 2005 2013 2021 All adults Youth Economic Education Family related Urban Rural-urban Urban-urban Rural non- non-migrants migrants migrants migrants Shocks Others Note: Recent migrants are those who lived in their current Note: Recent migrants are those who lived in their current location location (from another town or rural woreda in Ethiopia) for less (from another town or rural woreda in Ethiopia) for less than five than five years. Adult population refers to people aged 15 years years. Only working age population (people aged 15 to 64 years) and above. are included. Bars indicate 95 percent confidence intervals. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LMS 2021. 32 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Who migrates where and why? increased with education (Figure 3.7).⁴⁰ In terms of labor market characteristics, the likelihood of Having more education is a driver of migration. migrating to another rural or urban location within As noted, regardless of whether the origin area is Ethiopia is highest among the unemployed and rural or urban, migrants are younger and better wage employed (Figure 3.8). Regionally, people educated compared to non-migrants from the same are more likely to out-migrate from Benishangul- origin area. Rural dwellers who migrate, either to Gumuz or Amhara compared to Oromia. other rural or to urban areas, were more literate and accumulated more educational attainment. To examine drivers of types of migration, we A similar pattern is found for urban-origin areas, estimate a multinomial logit regression.⁴¹ The where migrants (that is, those who moved to other determinants of the propensity to migrate may urban areas) are younger and more educated depend on the type of migration. In the specification, compared to urban non-migrants. To examine we model migration as a function of the migrant’s the overall picture of the determinants of internal demographic characteristics (age and sex), migration (irrespective of the type of migration), we educational attainment, and the migrant’s origin used a probit model to estimate the propensity to zone to proxy for potential push or pull factors (zone- migrate while controlling for individual, household, level poverty rates, population density as a proxy for labor market, and locational indicators that could land holdings, and zonal infrastructure). We include reflect push and pull factors.³⁹ Looking at the origin region dummies to capture unobserved entire sample, the probability of internal migration effects that may reflect regional variations.⁴² In the Figure 3.7: In general, education drives migration Figure 3.8: The propensity to migrate is highest (Educational attainment and propensity to migrate, percent) among the unemployed and wage employed (Labor force status and propensity to migrate, percent) 9 8 9 7 8 6 7 5 6 5 4 4 3 3 2 2 1 1 0 0 Pre-school/No education Less than primary Unemployed Wage-employed Self-employed Completed primary Completed secondary Unpaid family worker Others Inactive Completed post-secondary Note: The estimated propensities are based on a probit model that Note: The estimated propensities are based on a probit model that models the decision to migrate based on demographic characteristics models the decision to migrate based on demographic characteristics of the individual and the characteristics of his/her origin. Figures of the individual and the characteristics of his/her origin. Figures of predicted probabilities were rescaled to percentages. Recent of predicted probabilities were rescaled to percentages. Recent migrants are those who lived in their current location (from another migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five years. Adult town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. population refers to people aged 15 years and above. Source: Authors’ estimation based on LMS 2021. Source: Authors’ estimation based on LMS 2021. ³⁹ For details on the probit estimates on the propensity to migrate, please see Annex 2. ⁴⁰ Education also drives migration aspirations: Research on the Young Lives data find that over 70 percent of young people who completed primary education or more aspired to migrate to urban areas or abroad (Schewel and Fransen 2018). ⁴¹ See Annex 4 for details on the methodology which is based on the resource provided by Katchova, 2013. ⁴² For non-migrants, the origin region is their region of residence at the time of the survey. 33 3. MOTIVES AND IMPACTS OF MIGRATION first step, we focus on migration out of rural areas, migrate. The likelihood of rural-to-rural migration regardless of whether the destination is another is highest among people with a post-secondary rural area or an urban area. The base category for education while people who have completed the analysis is rural dwellers who did not migrate primary education exhibit the highest likelihood in the five years preceding 2021 LMS. We estimated of rural-to-urban migration between 2016 and the regression separately for migrants from rural 2021 (Figure 3.9). Rural dwellers who had at least and urban areas. Annex 4 Table A 4 presents completed primary education have a 4-percentage detailed results for migrants from rural areas using point higher likelihood of migrating to an urban rural non-migrants as the reference category, with area compared to people with no education. For those who migrated to other rural areas as one urban-to-urban migration, completed secondary category and those who migrated to urban areas education appears to be the most important factor. as another category. In a second step, we focus on Wage-employees are more likely to have migrated migration out of urban areas, regardless of whether from a rural setting or from one urban area to the destination is another urban area or a rural another compared to the unemployed (Figure 3.10). area. Annex 4 Table A 5 includes results for the This is partly influenced by movement of public determinants of urban migrants using urban non- servants; many wage employees who relocate migrants as a base category. within rural or urban boundaries, or from urban to rural areas, tend to work in the public sector. The better educated, wage employed, and women However, a noticeable share of wage-employees are more likely to move out of rural areas.⁴³ who migrate from rural to urban areas work in the Considering migration out of rural areas, we private sector as opposed to the public sector (67 found that educational attainment tends to play percent versus 33 percent). In terms of sex, women a distinctive role in determining the propensity to are generally more likely to migrate from rural-to- Figure 3.9: In general, education drives migration Figure 3.10: The propensity to migrate is highest (Educational attainment and propensity to migrate by type among the unemployed and wage employed of migration, percent) (Labor force status and propensity to migrate by type of migration, percent) 4,5 7 4 6 3,5 5 3 2,5 4 2 3 1,5 2 1 0,5 1 0 0 Rural to rural Rural to urban Urban to rural Urban to urban Rural to rural Rural to urban Urban to rural Urban to urban Pre-school/No education Less than primary Unemployed Wage-employed Self-employed Completed primary Completed secondary Unpaid family worker Others Inactive Completed post-secondary Note: The estimated propensities are based on a multinomial logit Note: The estimated propensities are based on a multinomial logit model that models the decision to migrate based on demographic model that models the decision to migrate based on demographic characteristics of the individual and the characteristics of his/ characteristics of the individual and the characteristics of his/ her origin. Figures of predicted probabilities were rescaled to her origin. Figures of predicted probabilities were rescaled to percentages. Recent migrants are those who lived in their current percentages. Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than location (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. five years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. Source: Authors’ estimation based on LFS 2005, 2013, LMS 2021. ⁴³ Generally, the propensity to migrate from urban-to-rural areas is rare so we cannot clearly distinguish the determinants associated with this type of migration. However, we find that the likelihood of urban-to-rural migration tends to be higher among people who originated from Addis Ababa. 34 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES urban or from urban-to-urban areas, while men are inactive individual is, on average, 29 percentage slightly more likely to migrate from rural-to-rural points less likely to have migrated for economic or urban-to-rural areas. In line with findings for reasons compared to someone who is unemployed, the entire sample, people are more likely to out- with comparable effects for an unpaid worker. migrate from Benishangul-Gumuz relative to other Furthermore, those who completed primary or origin regions, particularly if the form of migration secondary education (relative to no education or was rural-to-rural or rural-to-urban. preschool), being born in destination zone, and living in a household with high dependency ratio Next, we employed a regression approach to better exhibit lower likelihoods of economic migration understand factors determining motives for rural- than their respective counterparts. Surprisingly, we to-urban migration. We adopt another multinomial find no evidence to show that local destination labor logit specification to analyze the determinants of market conditions provide strong pull incentives for the five reasons for migration (economic, education, people to migrate to urban areas. Similarly, lifetime family, shocks, and other reasons) using data from the migration inflows into destination and origin-zone 2021 LMS. The analysis only considers recent rural- poverty rates have little push influence on migration. to-urban adult migrants between 2016 and 2021.⁴⁴ We include individual-level characteristics such as As expected, people with higher education age, sex, marital status, educational attainment, ICT (relative to no education) tend to migrate for access, and a person’s current labor force status educational purposes. This is particularly true for as independent variables. We also controlled for young men between the age of 15 and 24 compared some household level variables such as sex of the to older counterparts. An unpaid family worker or an household head and dependency ratio (the share economically inactive individual is likely to move for of children and elderly persons in the household educational or family reasons, while people living in in relation to household size). Furthermore, the households with a high dependency ratio migrate analysis includes pull factors such as woreda-level for the sake of the family or due to shocks. Migrants employment rate at destination and whether the originally born in their destination zones are likely migrant was originally born in their destination zone to return for educational or family-related reasons. of residence. Push factors in the regression include migrant’s origin region and origin-zone poverty Regional differences play a role in the purpose rates. Annex 4 Table A 6 summarize all results. for why people migrate. Considering Oromia as the reference group, people who previously lived in Perhaps not surprisingly, age, sex, employment Afar, Benishangul-Gumuz and Gambela regions are status, and education all increase the probability less likely to have migrated because of economic of economic migration. In line with the descriptive reasons. The likelihood of migrating for educational statistics, female non-youth are significantly less purposes is lower among those who previously likely to migrate for economic reasons but more resided in Tigray,⁴⁵ Afar, and Harari regions while likely to migrate due to family reasons compared people who had previously lived in Gambela most to males. All else equal, older (or rather non-youth) likely migrated for family-related reasons. Also, men, the wage-employed, employers, and other people who previously resided in Tigray, Amhara types of employees are more likely to migrate and Dire Dawa have a lower probability of migrating for economic reasons. However, an economically as a result of shocks while in contrast, those who ⁴⁴ Note that we could not estimate the regressions for males and females separately given the small sample sizes and the number of independent variables we have included in the regressions. ⁴⁵ Note that although Tigray was not sampled in the 2021 LMS, there are migrants who said that their previous regions of residence was Tigray. 35 3. MOTIVES AND IMPACTS OF MIGRATION formerly lived in Afar or Benishangul-Gumuz livelihoods of migrants (Teweldebrihan, Pande, and regions are more likely to migrate due to shocks. McClain 2020). What are some other push and pull factors? Places with better physical and social infrastructure attract migrants. The existence Potential gains in income draws people to migrate of better opportunities, services, and amenities in from rural to urban areas. Rural areas provide other locations serve as pull factors of migration, few economic opportunities and rural-to-urban especially to urban areas. Better opportunities and migration is therefore a pathway for improving infrastructure in terms of education, health, and access to better income-generating opportunities. electricity particularly attract migrants (Grover, The Rural Income Diagnostics Study (World Bank Lall, and Maloney 2022). 2022) shows that migrants can expect substantial gains in average monthly wages from rural-to- Social networks play an important role in both urban migration. Once adjusting average monthly internal and international migration decisions. wages for hours worked, hourly wages in urban The existence of migrant networks can reduce areas are three times higher than rural agricultural costs and risks of migration (Groth et al. 2020). wages and significantly higher than average For example, Ethiopian women who migrate to the agricultural incomes. Middle East often do so by utilizing social networks, which reduces the cost of migration (Zewdu 2018). Lack of access to land accelerates migration in Similarly, the existence of family members in the Ethiopia. The literature explores land extensively as destination of migration motivates other family a reason for migration. The amount of land a young members to join them as it reduces migration costs. person expects to inherit can be a determining The presence of an existing support network can factor in the migration decision for rural Ethiopian help migrants establish themselves, for example by youth. Larger expected land inheritances are found helping them find places to live, work, and socialize to significantly reduce the probability of permanent (Grover, Lall, and Maloney 2022). The Ethiopia Rural long-distance migration and permanent migration Income Diagnostics (World Bank 2022) also shows to urban areas. Lack of inheriting land is a much that the probability for a rural household member stronger predictor of permanent rural-to-urban to migrate increases by 7.5 percentage points if migration and non-agricultural employment in areas networks are present. Hence both the decision to with less dynamic land markets, in relatively remote migrate and the success of migrants are closely areas, and in areas with poorer soil quality (Kosec linked to established networks. et al. 2018). For example, youth in rural southern Ethiopia have limited access to agricultural land 3.2 WHAT ARE INSTITUTIONAL due to land scarcity and land market restrictions, BARRIERS MIGRANTS FACE? forcing young people to abandon agriculture in search of other livelihoods. Bezu and Holden 2014 Actions in destination locations can stifle migration show that only 9 percent of rural Ethiopian youth even if no explicit laws or policies restrict plan to engage in agriculture as a livelihood. They mobility. In Ethiopia, no official restrictions exist also find a strong increase in migration of young but strong barriers to migration such as language, people resulting from a lack of access to land which xenophobia, uncertain employment prospects, forces them to quit farming. Absence of land rights and institutional barriers significantly complicate slightly encourages migration in Ethiopia (Brauw migration (de Brauw, Mueller, and Woldehanna and Mueller 2012). Moreover, migration away from 2018). Many migrants face a host of difficulties eastern Ethiopia in the Abbay River basin between linked to policies and attitudes that explicitly or 1984 to 1986 was due to scarcity of water, land, implicitly disadvantage them, and are partly linked to and rain, and migration positively affected the the absence of protection mechanisms. Despite the 36 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES positive effects of migration, institutional barriers even when they manage to find a job, it is usually exist that hinder mobility and easy integration. casual in nature with low and/or irregular payment Contrary to their expectations, migrants often find it and unfavorable working conditions. In general, difficult to integrate into the new place of residence migrants do not have access to employment support and find alternative sources of livelihood other than services, and other initiatives designed to reduce casual labor and petty trade (selling and buying urban unemployment—such as the “youth revolving goods at small scale) upon arrival in the destination. fund”—usually do not cover migrants. Migrants also Institutional barriers prevent people from capitalizing mention nepotism, the need for bribes, and lack of on the differences in wellbeing across countries and transparency related to employment opportunities geographic areas within Ethiopia. in general, and government jobs in particular. These challenges intertwine to complicate the situation; due Rural-to-urban migrants face various challenges to difficulties covering living expenses, the longer it and costs due to lack of institutional support and/ takes a migrant to find a job, the harder it becomes. or policies that discourage or exclude migration. This sub-section draws its main findings from a Rural-to-urban migrants face difficulties qualitative research study of young rural migrants acquiring a resident ID and accessing associated in urban areas in April and May 2017 as summarized services. Many migrants do not have any form of in Bundervoet 2018. Challenges highlighted in ID from their place of origin. At their destination, the study include: (i) difficulty to settle and cover it is difficult to acquire a Kebele ID for various living expenses; (ii) difficulty in finding a job; (iii) reasons including not satisfying the requirement difficulty in acquiring a resident ID and access to of living in the Kebele for a minimum period associated services; and (iv) negative perceptions (usually six months), lack of release letter from of government officials on internal migration. We their place of origin, and lack of housing address.⁴⁶ discuss each separately. Lacking a Kebele ID also implies that they cannot access public and other subsidized services, such Rural-to-urban migrants face difficulties settling as government food and employment-related and covering living expenses. After migrants arrive benefits. For example, a resident is required to have from rural areas, they find it difficult and expensive to an ID to get a certificate of being “unemployed”, a find a place to stay and cover other basic needs such prerequisite for government employment support as food. There is no support, financial or otherwise, including access to loans, training, and a workplace to new migrants. To the contrary, migrants are not for self-employment. Moreover, not having an ID eligible to access existing government support such can also lead to police harassment and arrest as a as subsidized housing, food, or access to social migrant may be considered an illegal resident. protection programs. This problem is more serious for those who do not have any family members or Rural-to-urban migrants face negative government friends at the destination. officials’ perceptions towards them. Official views on internal migration at the city and regional levels Rural-to-urban migrants face difficulties finding are negative. Many offices and bureaus consider employment. Finding employment is difficult for rural-to-urban migration as unacceptable, or even a typical migrant, and more so for those who are illegal, as it pressures urban job and services in urban not well connected. Lack of education, training, areas. Negative attitudes officials have towards and reference letters make finding a job even migrants are grounded in the belief that migration is more difficult for many migrants. As a result, not beneficial in general (including for the migrants ⁴⁶ Usually, someone who has a house has to go with the person who request an ID and inform the Kebele administration that the person is either part of her/his family or is a tenant. 37 3. MOTIVES AND IMPACTS OF MIGRATION themselves) and that rural youth should remain Hence this section contributes to the evidence on in rural areas and work on their land and receive the effects of migration not only on the individual but government assistance in their places of origin. also on their household and destination areas. This perception is in sharp contrast to experiences of migrants themselves, who typically migrate due How do migrants fare in the labor market? to lack of opportunity in their place of origin. Local government officials often do not consider migrants Profiles of labor market outcomes between as constituencies in terms of service provision and migrants moving to urban areas and local employment support. residents show few differences in unemployment but differences in types of employment. It is Migrants abroad also face challenges in their sometimes argued that migrants are not likely migration process. Smith et al. (2020) outlines the to benefit from moving to urban areas because current process for regular international migration of their relatively lower human capital and the from Ethiopia and its barriers. The process of already crowded urban labor market. Evidence migrating internationally is lengthy and frequently shows that there are benefits to rural-to-urban disincentivizes migration. The current managed migration and labor market outcomes for rural- labor migration process in Ethiopia has a number of to-urban migrants do not differ significantly from redundancies and unnecessary steps and is lengthy urban non-migrants. Table 3. 1 reports summary and not well-aligned with the needs of destination statistics on labor market outcomes of the adult markets. The process is further complicated by poor working-age population (15 to 64 years old). On coordination between relevant ministries and bodies, average, urban non-migrants are slightly less as well as an apparent disconnect between migration likely to be unemployed (18 percent) than rural-to- management at the federal and regional levels. The urban (20 percent) or urban-to-urban migrants (21 lengthy process places time and cost constraints percent).⁴⁷ T-testing shows that these differences on the worker, incentivizing them to pay an agent to are significant at the 1 percent level. However, we take on this burden, or even to migrate irregularly observe important differences when looking at the to avoid the process altogether. This undermines the quality of employment. Migrants are more likely competitiveness of Ethiopian workers compared to to be wage-employed but urban non-migrants workers from other countries. Beyond these process are more likely to be self-employed, including as problems, significant gaps in protection systems employers or unpaid family workers. Typically, while abroad, training prospective migrants, and wage-employment is associated with higher reintegrating migrants upon return disincentivize wages, job security, and more regular working international migration (Smith et al. 2020).. hours. Yet, we see mixed results on wages. Monthly wages for rural-to-urban migrants are 3.3 WHAT ARE THE EFFECTS OF considerably lower compared to those of urban MIGRATION ON THE DIFFERENT non-migrants and urban-to-urban migrants, DIMENSIONS OF WELFARE OF partly linked to the type and sector of jobs; many MIGRANTS AND THEIR FAMILIES? female rural-to-urban migrants work as low paid domestic workers. Rural-to-urban migrants are Measuring migration effects on households’ and more likely employed in non-permanent jobs countries’ welfare represents an important first (either contractual or casual works) in the private step to designing policies to promote migration. sector, which may partly undermine their earnings ⁴⁷ Lower unemployment rate for rural-to-urban migrants compared to urban-to-urban migrants could be related to higher reservation wage and better education for urban-to-urban migrants. 38 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES potential. Differences in earnings may also depend to be engaged in elementary positions compared on the type of occupation; rural-to-urban migrants to urban-to-urban migrants who are more likely and non-migrant natives are significantly more like engaged in highly-paid professional work. Table 3.1: Labor market outcomes, recent migrants vs. non-migrants (adults) Rural-to-urban Urban-to-urban Urban migrants migrants non-migrants Labour force status: Inactive 25.6%*** 22.6%*** 29.2% Active 74.4%*** 77.4%*** 70.8% Employment status (active population): Unemployed 19.5%*** 20.9%*** 17.7% Employed 80.5%*** 79.1%*** 82.3% Type of employment: Wage-employed 53.3%*** 61.3%*** 45.4% Self-employed 35.3%*** 29.5%*** 41.9% Employer 0%*** 0.5% 0.5% Unpaid family worker 9.9%** 7.4%*** 11.1% Others 1.4%* 1.3% 1.1% Income from wage employment: Monthly wages 3,099*** 5,387 5,457 Terms of employment: Permanent 43.8%*** 66.7%** 63.9% Temporary/contract 45.1%*** 26.6% 26.5% Casual 10.7%** 6.5%*** 9.3% Other 0.4% 0.2% 0.3% Occupation: Managers 0.6%*** 2.0% 2.0% Professionals 6.5%*** 20.5%*** 10.6% Technicians & associate professionals 7.2% 9.8%*** 7.3% Clerical support 1.0%*** 2.5% 2.3% Service & sales 37.7%*** 30.0% 29.2% Skilled agricultural, forestry & fishery 9.9%* 3.3%*** 10.8% Craft & related trades 6.1% 7.2%*** 6.0% Plant/machine operators & assemblers 3.1%*** 5.1%*** 6.5% Elementary occupations 27.4%*** 18.6%*** 24.6% Other occupations (e.g., armed forces) 0.5% 1.0%* 0.7% Sector of employment for wage employed: Public sector 18.2%*** 35.9%*** 23.4% Private sector 81.8%*** 64.1%*** 76.6% Industry of employment (Broad): Agriculture, mining and extractives 13.1% 5.0%*** 13.8% Manufacturing 19.1%*** 17.3%* 16.0% Services 67.8%*** 77.6%*** 70.3% Observations 5,847 5,222 55,689 Note: World Bank staff calculations using LMS 2021. The table reports summary statistics for only working age adults between 15 and 64 years. Migrants in this context refers to only recent internal migrants. Note that all non-wage employees are concentrated in the private sector while only 48% of wage-employees work in the public sector. T-tests of significant differences are reported for the respective migrant groups with reference to urban non-migrants * p<0.05 ** p<0.01 *** p<0.001. Source: Authors’ estimation based on 2021 LMS. 39 3. MOTIVES AND IMPACTS OF MIGRATION How large are wage differentials for migrants? with a difference of 0.59 log points (Figure 3.11); equivalent to a difference in wages between the two We explored whether wage gaps between groups of approximately ETB 1,553. The estimates rural-to-urban migrants and non-migrants suggest that about 78 percent of the wage gap is due in urban destinations are primarily due to to differences in differences in characteristics (or differences in human capital endowments, such endowments, such as education or type of job). The as education, or due to differences in returns remaining 22 percent is explained by differences to these endowments, from factors such as in returns to these endowments, this could include discrimination. We apply a simple Oaxaca-Blinder factors such as discrimination. We estimate that decomposition for linear regression models, increasing migrant workers’ endowments to those which is often used to study differences in wages of non-migrants would increase migrants’ wages across groups (for example, sex and race). The by 58 percent, while a gap of 14 percent would analysis partitions the wage differential between remain unexplained. two groups: a part explained by group differences in characteristics (endowments such as education or work experience) and a part that is unexplained by the differences in endowments, often used as a measure of discrimination.⁴⁸ Rural-to-urban migrants have lower wages compared to urban non-migrants and differences are mainly due to differences in characteristics such as education or work experience. Annex 5 Table A 7 reports detailed results from our decomposition analysis. The mean predictions of (log) wages are significantly higher for urban non- migrants than rural-to-urban migrant workers, Figure 3.11: Wages are significantly higher for urban non-migrants (Oaxaca-Blinder decomposition analysis for rural-to-urban migrants vs. urban non-migrants) Decomposition Contribution to total wage gap Difference Endowments Environments Returns Returns 0 .2 .4 .6 .8 0 10 20 30 40 50 60 70 80 Log difference in wages Percent Note: The decomposition analysis focusses on only the sample of recent rural-to-urban migrants and urban non-migrants who are wage employed and of working age. Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five years. Working age population refers to people aged 15 to 64 years. Source: Authors’ estimation based on LMS 2021. ⁴⁸ See Annex 5 Annex 4for details on the methodology. 40 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Sex, age, and level of education drive the permanent job and certain occupations contribute endowment effect on wage gaps between urban- to observed endowment effects (Figure 3.12, Panel to-rural migrants and urban non-migrants. A). Urban-to-rural migrants are more likely to have Using educational attainment as an example, our temporary or casual jobs; only 44 percent of migrants analysis shows that migrants who have less than have permanent jobs compared to 64 percent of primary education (relative to no education) earn non-migrants. If migrants had the same share of 5 percent less than their urban non-migrants with permanent jobs, wage earnings would increase by less than primary education (Figure 3.12, Panel A). 5 percent. Relative to managerial roles, working as Conversely, estimates for completed secondary professionals/technicians or machinery assemblers or post-secondary education are both positive have lower wages and working in clerical support, and significant. Hence, if migrant workers had sales, and elementary occupations have positive the same level of completed secondary education endowment effects on the wage differential in favor as their non-migrant counterparts (compared to of migrants. This is not surprising given that urban being uneducated), their wage earnings would have non-migrant workers are more likely to be in high- been 5 percent higher. Moreover, if migrants had skilled jobs compared to migrant workers. Again, we the same level of post-secondary education, their find little evidence to support that returns to these wages would increase by 36 percent. Though we worker-specific endowments explain the wage observe that urban-to-rural migrants earn less, gap, meaning that the majority of the wage gap can we also see that returns to education would be the be explained by differences in the type of job and same if they had higher levels of education. occupation. However, some discrimination seems to be at play as can be seen in differences in the Differences in job type, temporary or permanent, woreda unemployment rates. If we were to assume and occupation contribute to endowment effects. that urban-to-rural migrants live in woredas with the Worker-specific endowments such as job type same unemployment rates as urban non-migrants, or occupation show that having a temporary or wage gaps would increase (Figure 3.12, Panel B). Figure 3.12: The endowment effect on the wage gap is driven by sex, age, and the level of education (The contributions of endowments (panel A) and returns (panel B) to the wage gap) Female Youth Non-permanent job Less than primary Woreda Completed primary unemployment rates Completed secondary Completed post-secondary Professionals & technicians Clerical, services & sales Plant/machine operators & assemblers Elementary occupations Amhara Somalia Somalia Benishangul-Gumuz Addis Ababa -10 0 10 20 30 40 -150 -100 -50 0 Percent Percent Note The decomposition analysis focusses on only the sample of recent rural-to-urban migrants and urban non-migrants who are wage employed and of working age. Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five years. Working age population refers to people aged 15 to 64 years. Source: Authors’ estimation based on LMS 2021. 41 3. MOTIVES AND IMPACTS OF MIGRATION What are the effects on migrants' families? Rural-to-urban migration has positive effects on the productivity and unemployment in rural Migration of one household member can improve origin areas. There is a common belief that rural- the wellbeing of the entire household left behind. to-urban migration causes rural productive labor The main channel through which improvements in shortages and harms agricultural productivity. welfare manifest themselves is through improved Contrary to this belief, Ethiopia’s Rural Income consumption on food and non-food goods. Rural-to- Diagnostics (World Bank 2022) showed that rural- urban migration can positively influence household to-urban migration increases the intensity of labor food security in Ethiopia. Using data from a panel use in migrant-origin households by 29 percent survey of 573 households and a differences-in- (Table 3.2). This implies that migration reduces differences approach, Abebaw et al. (2020) found disguised unemployment in rural areas. Moreover, that, on average, rural out-migration significantly rural-to-urban migration increases agricultural increased the number of daily calories an adult output per worker by 18 percent and thus enables household member left behind consumes by 22 the remaining household members in rural areas percent. The study also finds positive effects on the to adequately feed off their land. Migration also poverty gap and the severity of poverty. Rural-to- increases the share of households that rented urban migration has positive benefits on originating out land by 1.2 percentage points, translating households and communities through remittances, into a 6.6 percent increases in the amount of an important source of livelihoods equivalent to land rented out. All of these aspects point to the 31 percent of recipient household consumption positive effects that urban-to-rural migration not expenditure nationally, and 70 percent among the only has on migrants and their families but also on bottom quintile in 2016 (World Bank 2020). productivity in rural agriculture. Table 3.2: Urban-to-rural migration has positive impacts on factor markets in origin communities Statistics Cultivated land Land Family labor supply Value of crop harvest (ha per capita) rented out (days per capita) (Birr per capita) ATT 0.065** 0.012** 120** 938.5** Standard Errors 0.001 0.017 21.302 369.5 97.5% Confidence Interval (0.045; 0.083) (-0.024; 0.042) (79.958; 161.455) (235.67; 1678.94) Note: ATT is the Average Treatment Effects on Treated. (a) Cultivated land per capita is the area in hectare that the household utilized for crop production. (b) Land rented out is the probability of households renting/sharing out agriculture land. (c). Value crop harvest is the Ethiopia birr value of the total production. ** result statistically significant at the 5 percent level. Source: Rural Income Diagnostics (World Bank 2022), estimation based on ESS 2011/12; 2013/14; 2015/16. 42 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Migration can contribute to family income at the spend on food decreased after migration, a clear place of origin and lead to development spillovers indication of improvements in welfare. for their communities. We will use the 2012 and 2016 rounds of ESS panel data to answer the 3.4 WHAT ARE THE MIGRATION EFFECTS question on whether migration (from rural-to-urban ON DESTINATION AREAS? or international) positively affects the welfare of family members back home by comparing pre and Rural-to-urban migration is at the core of economic post-migration outcomes. We divide the dataset development, reallocating population from villages into households that have at least one household to cities and across sectors (Kuznets 1964; Harris member who migrated between the two interviews and Todaro 1970). Once in urban areas, migrants in 2012 and 2016 and those households who do not do not allocate themselves randomly across have a migrant. We then look at household outcomes neighborhoods or at the cities’ frontier. They choose in 2021, before the household member migrates, and where to live, typically in low-rent, low-amenities in 2016, after the household member migrates. areas and segregated migrant neighborhoods with limited public services (Huang and Tao 2015; Rural-to-urban migration is associated with welfare Jedwab, Christiaensen, and Gindelsky 2017; Bharathi improvements for household members who remain et al. 2021). Climate change may reinforce this in rural areas. Before migration, 32 percent of dynamic: climate shocks-induced migrants may be households with a migrant were in the lowest welfare more likely to crowd into unplanned settlements, quartile; while after migration, only 24 percent of thus contributing to deteriorating living conditions households that have a member who migrated were and cities’ ability to generate economic growth. To in the lowest welfare quartile (Figure 3.13). Similarly, capture the multifaceted effects of migration on city 48 percent of households with a migrant were in growth and economic development, it is important the bottom 40 percent before migration but only 38 to understand and address challenges from rapid percent were in the bottom 40 percent after migration urbanization, including urban sprawl, segregation, (Figure 3.14). We also observe welfare effects on living conditions (congestion and pollution), and migrant households; the share of consumption they structural transformation (industrialization). Figure 3.13: Rural-to-urban migration is associated Figure 3.14: Migration moves households out of with improvements in welfare the bottom 40 percent of the welfare distribution (Share of households of rural-to-urban migrants within certain (Share of households of rural-to-urban migrants in bottom 40 consumption quartile) percent of consumption distribution) 32.2 47.8 28.2 25.3 26.6 23.4 24.1 37.5 21.1 19.0 2012 (before migration) 2016 (after migration) Poorest quartile 2nd quartile 3rd quartile Richest quartile 2012 (before migration) 2016 (after migration) Note: Recent migrants are those who lived in their current location Note: Recent migrants are those who lived in their current location (from another town or rural woreda in Ethiopia) for less than five (from another town or rural woreda in Ethiopia) for less than five years. Adult population refers to people aged 15 years and above. years. Adult population refers to people aged 15 years and above. Source: Authors’ estimation based on ESS 2012, 2016. Source: Authors’ estimation based on ESS 2012, 2016. 43 3. MOTIVES AND IMPACTS OF MIGRATION Although how migration and migrants’ location regional climate change mean that migration as a choices affect urbanization and development risk-coping mechanism will significantly influence is unknown, it is possible for rural-to-urban Ethiopia’s development. Migration should thus migration to have negative environmental, constitute a major policy concern for leaders and social, and labor consequences. Rural migrants urban planners for decades to come. go to urban areas in search of employment and to use urban services and facilities. Since the Critics often claim that migrant workers displace continuous outflow of rural migrants seek basic native workers from jobs and reduce wage rates, urban facilities, the migrants pressure the socio- harming destination areas. To explore this point, economic and environmental conditions of the we estimate the relationship between the size of host area already experiencing housing shortages, migration flows and local labor market conditions for unemployment, rising cost of living, lack of access rural-to-urban migrants and urban non-migrants in to social services, rising crime, and expanding the working-age population. We control for the share informal sectors (Habtamu 2015). Ethiopia is of individuals by educational attainment, access to experiencing important urbanization challenges. information, occupation, sector or employment, First, if rapid urban growth—with an urbanization and town size at the woreda level. Annex 6 Table rate of 5.2 percent per year since 2018—continues, A 8 indicates that a larger share of rural-to-urban the United Nations, Department of Economic and migrants in a woreda does not significantly increase Social Affairs, and Population Division (2019) unemployment. However, it may push down wages projects that Ethiopia’s urban population will reach at the destination but this effect is small; a 1 percent 50 million by 2034. Second, Ethiopian cities face increase in the share of rural-to-urban migrants acute urban sustainability issues with many urban induces a reduction in wages by 0.34 percent. dwellers suffering extremely limited access to basic Additionally, woredas with a larger share of workers infrastructures and services (Lall et al. 2017) and with post-secondary education (high-skilled) are cities seemingly failing to match labor demand with strongly negatively associated with unemployment supply (Franklin 2018; Wieser and Mesfin 2021). rates and weakly positive associated with wage Third, its large, fragile, and diverse agricultural rates. Put more clearly, woredas with a higher sector sets the stage for large, unexpected migrant concentration of migrants are more likely to have flows triggered by agricultural shocks. Ethiopia’s lower unemployment overall and slightly higher history of food insecurity and the consequences of wages than places with people who are unskilled. 44 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES 4. CONCLUSION AND POLICY DIRECTIONS In this section we look at policy directions related Reduce barriers to migration to strengthening benefits from migration while reducing challenges. Removing free labor movement constraints can increase migration and encourage migration to Shift policy focus from preventing to locations with better economic opportunities. In leveraging migration Ethiopia, government intervention has controlled the movement of labor by requiring households Reshaping policy perspectives and overall to register and the need to obtain a Kebele ID. But attitude towards migration could encourage a migrant is required to live in a certain area for a positive policy shift. Policymakers’ concerns six months before being able to register their new relate to both urban destinations and rural origin address and request a Kebele ID that reflects his areas. For cities, the current belief is that migrants or her residence. While Kebele IDs are required to expand urban populations and overwhelm urban obtain access to government services, migrants services. For rural origin areas, concerns on the often cannot obtain ID cards, either because they damaging effects on the agriculture sector related are unable to obtain a leave letter from their home to the outflow of labor. Recent analysis, however, kebele or they face urban government officials’ suggests the opposite effects: rural-to-urban reluctance to provide them with the ID card. As a migration is an important pathway for facilitating result, it is difficult or impossible for migrants to both agricultural transformation and linking move freely within the area of destination, obtain rural youth to off-farm opportunities (World Bank bank books, or join associations that would allow 2022). Creating opportunities for dialog around the them to access services such as housing, credit, positive and negative aspects of migration could or to buy subsidized foodstuffs (Bundervoet 2018). contribute to changing attitudes and encouraging To facilitate migrants’ freedom of movement less restrictive migration policies and migration- and facilitate their integration, policymakers can focused development approaches. remove restrictions on urban ID requirements 45 4. CONCLUSION AND POLICY DIRECTIONS such as requiring a minimum length of stay and Kebele IDs (see above) and streamlining ID reforms removing the requirement for a release letter. and household registration would be helpful. Reducing migration cost can encourage migration Facilitating access to credit and financial services and enhance returns to migration. Migrants face can enable migration from poor households. high migration costs due to challenges related Liquidity constraints currently limit migration to integrating into destination areas, potentially in Ethiopia. The migration process can be costly discouraging migration and not allowing liquidity- (Bundervoet 2018) and poor (rural) households constrained households to reap returns to migration. can often not send migrants and reap benefits Returns from migration are positive in the medium from migration due to liquidity constraints. Lack term—the average monthly wage (adjusted for of access to liquidity and credit therefore means hours worked) in urban areas is three times higher poor households are unable to finance the upfront than rural agricultural wages and significantly costs of migration. Recent research shows that higher than average agricultural incomes—but households with more access to credit and those non-wage factors make integration into destination receiving cash transfers are more likely to send areas challenging and costly for migrants (World migrants (World Bank 2022). Policies can reduce Bank 2022). For example, migrants face high job liquidity constraints by facilitating poor households’ search costs due to lack of local administration access to credit or provide them with cash transfers. support and networks. Many migrants also struggle to transition into urban life, with female migrants Adapt urban areas to fast growing populations facing additional challenges (Bundervoet 2018). Costs are mainly driven by: (i) “frictions” in the job Ethiopia is experiencing important urbanization matching process, and (ii) barriers to accessing challenges, and migration will constitute a major public services. policy concern for leaders and urban planners for decades to come. Authorities should recognize Job-matching “frictions” can lead to high job that migration is a natural process, especially search costs, unaffordable for poor migrants during fast economic growth and transformation. without savings. Connecting migrant jobseekers Labor migration should therefore be encouraged to employment opportunities can reduce social rather than deterred and national government and economic costs. Interventions that help programs should promote internal migration and connect migrant jobseekers with jobs, such as job urban-rural linkages and attempt to reverse the intermediation services and youth apprenticeship negative perception of migrants. Ethiopia needs programs, and signal migrants’ skills to prospective to actively pursue urban development strategies, employers can improve job-matching. Though the particularly for cities with large shares of migrants, GoE is currently strengthening its labor market to avoid the negative consequences from large information systems, migrants can typically population influxes. not access these services. Enhancing public employment services and allowing migrant Adapting urban areas for population growth workers to participate in these systems can reduce can better integrate migrants socially and job-matching frictions. economically. As mentioned, there is a strong belief that migrant-fueled expanding urban population Streamlining administrative procedures can overwhelms urban services. Analysis in this report reduce barriers to accessing public services. showed that, as a share of their population, small Streamlined administrative procedures can improve towns and secondary cities attracted most rural migrants’ access to public services and facilitate migrants. Yet, small towns and secondary cities have their integration and adjustment to urban life. As larger informal employment and insufficient public discussed, minimizing burdens related to obtaining infrastructure and services (World Bank Group 46 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES 2015). Despite progress in building infrastructure displacement, beginning with the 1985 famine; and services, Ethiopian cities have low coverage however, in recent years, voluntary migration has for water and sanitation services, solid waste become increasingly important. Both stocks and management, and road density (World Bank Group flows of migrants from Ethiopia rose after 2000, 2015). Since an influx of migrants may exacerbate concentrating in a few main destination countries, existing challenges, it is vital to continue expanding but they remain very low. Migration dynamics seem public infrastructure and services in main migrant to vary significantly by destination region, with destination cities. Moreover, existing urban social labor migrants comprising the majority of migrants protection schemes should include migrants. to the MENA countries (Smith et al. 2020). Investing in housing infrastructure can relieve In addition to improving the regulatory framework some pressure migration poses to existing for international labor migration, reducing housing challenges. Addis Ababa faces massive barriers to the process of migration can expand housing shortages fueled by rapid in-migration from labor flows from Ethiopia. The current process rural areas and a high natural population growth. for regular migration is lengthy and disincentivizes Households often tend to trade-off housing quality for migrating legally. The Ethiopia Labor Mobility easier access to jobs, or move to city outskirts with Diagnostic (Smith et al. 2020) outlines the regular limited connectivity and services (World Bank 2021). migration process and highlights that it is long An influx of migrants exacerbates this situation. and complex, involving the securing and validating Policies focusing on increasing housing supply and of 14 different documents (including a passport, upgrading informal settlements—where migrants birth certificate, emergency contact, police record, typically live—can ease housing market pressures. contract, health certificate, certificate of eighth Moreover, improving access to infrastructure and grade completion, and certificate of competence). basic services could improve living conditions for The process has a number of redundancies and migrants and local residents, Enhancing water and unnecessary steps, as evidenced by the fact that sanitation, solid waste management, and electricity only eight of the 14 documents are widely known services and reliability are especially important by actors within the process itself. Moreover, (World Bank 2021). Careful reallocation of budgets some requirements on the Ethiopian side are not towards locations with the largest anticipated necessary for worker to receive authorization in migration inflows, including investments to make destination markets. Removing redundancies, small towns and secondary cities more attractive, facilitating the process, and improving alignment must accompany sound urbanization policies. of destination market needs could promote larger Investments in smaller cities and towns can avoid migration flows abroad. Moreover, improving over-concentration of people in large cities. In coordination between relevant ministries and addition, research has shown that development of bodies, and fostering communication federal and towns and secondary cities has a bigger poverty- regional migration management could better reduction effect than concentration of populations support outgoing migrants. Currently, the process in mega-cities (Christiaensen and Todo 2014; places a significant time and cost burden on the Christiaensen, De Weerdt, and Todo 2013). worker, incentivizing payments to an agent to take on this burden or even to migrate irregularly to Improve process of international labor mobility avoid the process altogether. This undermines the competitiveness of Ethiopian workers compared to International mobility from Ethiopia has a workers from other countries. complex history, with many factors motivating the decision to migrate as well as the modality Better aligning incentives between workers, and choice of destination. Historically, Ethiopian agents, and government entities could improve migration has been predominantly due to the development potential of labor mobility. The 47 4. CONCLUSION AND POLICY DIRECTIONS intermediation system—the processes through number of migrants and benefits from international which vacancies abroad and jobseekers are migration (Smith et al. 2020). identified, vetted, and matched—is rebuilding following a ban by GoE on overseas recruitment Reducing gaps in protection systems while from 2013 to 2018 to Gulf Cooperation Countries abroad, increasing the skills of prospective (the primary destination countries for Ethiopia). migrants, and reintegrating migrants upon Approximately 700 intermediation agencies are return could improve development outcomes. registered with GoE, although only 100 of have active Employment abroad entails a number of risks for operations. The lengthy agency administrative the migrant worker not adequately mitigated in processing time due to complex is a key reason for the existing system. Most of these risks are tied to the low migration outflows and threaten further the quality of employment and that employment decrease. For example, Saudi Arabia reallocated a terms and conditions are not those stipulated in significant portion of the visa allocation it had given the contract: pay is not as stipulated, employment to Ethiopia to other countries as Ethiopia was not circumstances are abusive, or violation of labor able to process workers visas in sufficient time to laws. These risks increase for female migrants, facilitate overseas employment. Agencies seeking particularly domestic workers, as a result of being to abide by existing regulations are not able to earn employed in a private or semi-private sphere. sufficient revenue to finance their operations. This This frequently results in abuse, including sexual incentivizes workers to step outside of regulation assault. Non-employment risks include passport and pay informally for intermediation and brokering retention, violations of immigration or criminal law, to avoid burdensome administrative processes and and health and insurance concerns. Undertaking reach employment abroad more quickly. Better steps to strengthen protection and increase quality aligning incentives of all actors and promoting assurance throughout the process can improve formal recruitment channels could enhance the outcomes for migrants (Smith et al. 2020). 48 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES REFERENCES Abebaw, Degnet, Assefa Admassie, Habtemariam Influence Mobility Rates in Ethiopia?” Journal Kassa, and Christine Padoch. 2020. “Can of African Economies 21 (4): 548–79. https://doi. Rural Outmigration Improve Household Food org/10.1093/jae/ejs007. Security? Empirical Evidence from Ethiopia.” Brauw, Alan de, Valerie Mueller, and Tassew World Development 129 (May): 104879. https:// Woldehanna. 2018. “Does Internal Migration doi.org/10.1016/j.worlddev.2020.104879. Improve Overall Well-Being in Ethiopia?†.” Abebe, Girum, Niklas Buehren, and Markus Journal of African Economies 27 (3): 367. https:// Goldstein. 2020. Short-Run Welfare Impacts of doi.org/10.1093/jae/ejx039. Factory Jobs: Experimental Evidence from Ethiopia. Bryan, Gharard, and Melanie Morten. 2019. “The World Bank, Washington, DC. https://doi. Aggregate Productivity Effects of Internal org/10.1596/1813-9450-9325. Migration: Evidence from Indonesia.” Journal Adugna, Girmachew. 2021. “Once Primarily an Origin for of Political Economy 127 (5). https://doi. Refugees, Ethiopia Experiences Evolving Migration org/10.1086/701810. Patterns.” Migrationpolicy.Org (blog). October 4, 2021. Bundervoet, Tom. 2018. “Internal Migration in https://www.migrationpolicy.org/article/ethiopia- Ethiopia Evidence from a Quantitative and origin-refugees-evolving-migration. Qualitative Research Study.” Washington, DC: Aiglepierre, Rohen d’, Anda David, Charlotte World Bank Group. https://openknowledge. Levionnois, Gilles Spielvogel, Michele Tuccio, worldbank.org/handle/10986/32097. and Erik Vickstrom. 2020. “A Global Profile of Christiaensen, Luc, Joachim De Weerdt, and Emigrants to OECD Countries: Younger and More Yasuyuki Todo. 2013. “Urbanization and Poverty Skilled Migrants from More Diverse Countries.” Reduction: The Role of Rural Diversification and OECD Social, Employment and Migration Working Secondary Towns1.” Agricultural Economics 44 (4– Papers 239. Vol. 239. OECD Social, Employment 5): 435–47. https://doi.org/10.1111/agec.12028. and Migration Working Papers. https://doi. org/10.1787/0cb305d3-en. Christiaensen, Luc, and Yasuyuki Todo. 2014. “Poverty Reduction During the Rural–Urban Bezu, Sosina, and Stein Holden. 2014. “Are Rural Transformation – The Role of the Missing Middle.” Youth in Ethiopia Abandoning Agriculture?” World Development 63 (November): 43–58. World Development 64 (December): 259–72. https://doi.org/10.1016/j.worlddev.2013.10.002. https://doi.org/10.1016/j.worlddev.2014.06.013. El Mufti, Karim. 2019. “Migration, Remittances and Bharathi, Naveen, Deepak Malghan, Sumit Mishra, Possible Return of Ethiopian Domestic Workers and Andaleeb Rahman. 2021. “Fractal Urbanism: in Lebanon.” Securing Women Migration Cycle. City Size and Residential Segregation in India.” World Development 141 (May): 105397. https:// Ethiopian Economics Association. 2021. State doi.org/10.1016/j.worlddev.2021.105397. of the Ethiopian Economy 2020/21: Economic Development, Population Dynamics, and Welfare. Blinder, Alan S. 1973. “Wage Discrimination: State of the Ethiopian Economy. Addis Ababa. Reduced Form and Structural Estimates.” The Journal of Human Resources 8 (4): 436–55. https:// Ethiopian Monitor. 2022. “Repatriates from Saudi doi.org/10.2307/144855. Arabia Top 30, 000.” Ethiopian Monitor, 2022. https://ethiopianmonitor.com/2022/05/30/ Brauw, Alan de, and Valerie Mueller. 2012. “Do repatriates-from-saudi-arabia-top-30-000/. Limitations in Land Rights Transferability 49 REFERENCES FAO. 2016. “Migration, Agriculture and Rural org/10.1068/c12120. Development: Addressing the Root Causes IOM. 2017. “National Labour Migration Management: of Migration and Harnessing Its Potential for Ethiopia.” Geneva: International Organization for Development.” Rome: Food and Agriculture Migration. Organization of the United Nations (FAO). ———. 2022. “Funding Needed to Assist Over 100,000 Federal Democratic Republic of Ethiopia. 2020. Ethiopian Migrants Returning from the Kingdom of “National Voluntary Report on the Implementation Saudi Arabia.” International Organization for Migration of the Global Compact on Migration.” National Press Release, 2022. https://www.iom.int/news/ Partnership Coalition (NPC) on Migration. funding-needed-assist-over-100000-ethiopian- Franklin, Simon. 2018. “Slums to Failed Estates? migrants-returning-kingdom-saudi-arabia. Evidence from a Lottery for Low-Cost Housing in Jedwab, Remi, Luc Christiaensen, and Marina Ethiopia,” Job Market Paper, , 51. Gindelsky. 2017. “Demography, Urbanization Groth, Juliane, Tobias Ide, Patrick Sakdapolrak, and Development: Rural Push, Urban Pull Endeshaw Kassa, and Kathleen Hermans. and … Urban Push?” Journal of Urban Economics 2020. “Deciphering Interwoven Drivers of 98 (March): 6–16. https://doi.org/10.1016/j. Environment-Related Migration – A Multisite jue.2015.09.002. Case Study from the Ethiopian Highlands.” Global Kangmennaang, Joseph, Rachel Bezner-Kerr, and Environmental Change 63 (July): 102094. https:// Isaac Luginaah. 2018. “Impact of Migration and doi.org/10.1016/j.gloenvcha.2020.102094. Remittances on Household Welfare among Rural Grover, Arti, Somik V. Lall, and William F. Maloney. Households in Northern and Central Malawi.” 2022. Place, Productivity, and Prosperity: Revisiting Migration and Development 7 (1): 55–71. https:// Spatially Targeted Policies for Regional Development. doi.org/10.1080/21632324.2017.1325551. Washington, DC: World Bank. https://doi. Katchova, Ani. 2013. “Multinomial Probit and Logit org/10.1596/978-1-4648-1670-3. Models Conditional Logit Model Mixed Logit Habtamu, Bimerew. 2015. “Rural-Urban Migration Model.” Econometrics Academy. 2013. https:// and Its Consequence on Urban Living: The sites.google.com/site/econometricsacademy/ Case in Hawassa City Southern Ethiopia.” Global econometrics-models/multinomial-probit-and- Journal of Human-Social Science Research, June. logit-models. https://socialscienceresearch.org/index.php/ KNOMAD. 2022. “A War in a Pandemic: Implications GJHSS/article/view/1426. of the Ukraine Crisis and COVID-19 on Global Harris, John R., and Michael P. Todaro. 1970. Governance of Migration and Remittance “Migration, Unemployment and Development: Flows.” 36. Migration and Development Brief. A Two-Sector Analysis.” The American Economic Washington, DC: World Bank. Review 60 (1): 126–42. Kosec, Katrina, Hosaena Ghebru, Brian Holtemeyer, Hermans, Kathleen, and Lisa Garbe. 2019. Valerie Mueller, and Emily Schmidt. 2018. “The “Droughts, Livelihoods, and Human Migration Effect of Land Access on Youth Employment in Northern Ethiopia.” Regional Environmental and Migration Decisions: Evidence from Rural Change 19 (4): 1101–11. https://doi.org/10.1007/ Ethiopia.” American Journal of Agricultural Economics s10113-019-01473-z. 100 (3): 931–54. https://doi.org/10.1093/ajae/ aax087. Huang, Youqin, and Ran Tao. 2015. “Housing Migrants in Chinese Cities: Current Status Kuznets, Simon. 1964. “Quantitative Aspects of and Policy Design.” Environment and Planning C: the Economic Growth of Nations: IX. Level and Government and Policy 33 (3): 640–60. https://doi. Structure of Foreign Trade: Comparisons for 50 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Recent Years.” Economic Development and Cultural https://doi.org/10.2307/2525981. Change 13 (1): 1–106. Pritchett, Lant, and Farah Hani. 2020. “The Mekonnen, Berhe, Arden Finn, Christina Wieser, and Economics of International Wage Differentials Manex Bule. 2020. “Ethiopia Regional Poverty and Migration.” Oxford Research Encyclopedia of Report: Promoting Equitable Growth for All Economics and Finance. July 30, 2020. https://doi. Regions.” Regional Poverty Report. Addis Ababa: org/10.1093/acrefore/9780190625979.013.353. World Bank Group.https://documents.worldbank. Quoc Hoi Le and Thi Hoai Thu Nguyen. 2019. org/en/publication/documents-reports/ “Impacts of Migration on Poverty Reduction in documentdetail/627681605631546436/ Vietnam: A Household Level Study.” Business & Ethiopia-Regional-Poverty-Report-Promoting- Economic Horizons 15 (2): 261–75. https://doi. Equitable-Growth-for-All-Regions. org/10.15208/beh.2019.16. Meyer, Christian Johannes, Morgan Hardy, Marc Redehegn, Misgina Asmelash, x Dingqiang Sun, Witte, Gisella Kagy, and Eyoual Demeke. 2021. Aseres Mamo Eshete, and Castro N. Gichuki. “The Market-Reach of Pandemics: Evidence 2019. “Development Impacts of Migration and from Female Workers in Ethiopia’s Ready- Remittances on Migrant-Sending Communities: Made Garment Industry.” World Development 137 Evidence from Ethiopia.” PLoS One 14 (2): (January): 105179. https://doi.org/10.1016/j. e0210034. https://doi.org/10.1371/journal. worlddev.2020.105179. pone.0210034. Ministry of Finance and Economic Development. Schewel, Kerilyn, and Sonja Fransen. 2018. “Formal 2002. “Ethiopia: Sustainable Development and Education and Migration Aspirations in Ethiopia.” Poverty Reduction Program.” Addis Ababa: Population and Development Review 44 (3): 555–87. Federal Democratic Republic of Ethiopia. https://doi.org/10.1111/padr.12159. https://reliefweb.int/report/ethiopia/ethiopia- s u s t a i n a b l e - d e v e l o p m e n t- a n d - p o v e r t y- Smith, Rebekah, Zuzana Cepla, Daisy Demirag, Farah reduction-program. Hani, and Christina Wieser. 2020. “Ethiopia Labor Mobility Diagnostic.” 0000A8056. Washington, ———. 2006. “Policy - Plan for Accelerated DC: World Bank Group. and Sustained Development to End Poverty (PASDEP).” Addis Ababa: Federal Democratic Teweldebrihan, Meseret Dawit, Saket Pande, and Republic of Ethiopia. https://extranet.who.int/ Michael McClain. 2020. “The Dynamics of Farmer nutrition/gina/en/node/7935. Migration and Resettlement in the Dhidhessa River Basin, Ethiopia.” Hydrological Sciences National Planning Commission. 2010. “Ethiopia Journal 65 (12): 1985–93. https://doi.org/10.108 Growth and Transformation Plan I.” Addis 0/02626667.2020.1789145. Ababa: Federal Democratic Republic of Ethiopia. https://www.greengrowthknowledge.org/ UNCDF. 2021. “Country Monitor on Migration & national-documents/ethiopia-growth-and- Remittance: Ethiopia.” New York: United Nations transformation-plan-i. Capital Development Fund (UNCDF). ———. 2021. “Ethiopia 2030: The Pathway to UNCTAD. 2018. “Policy Guide on Entrepreneurship Prosperity.” Addis Ababa: Federal Democratic for Migrants and Refugees.” United Nations Republic of Ethiopia. https://climate-laws. Conference on Trade and Development. https:// org/geographies/ethiopia/policies/10-year- unctad.org/system/files/official-document/ development-plan-2020-21-2029-30. diae2018d2_en.pdf. Oaxaca, Ronald. 1973. “Male-Female Wage UNDESA. 2022. “World Population Policies Differentials in Urban Labor Markets.” Database.” 2022. http://esa.un.org/poppolicy/ International Economic Review 14 (3): 693–709. about_policy_section.aspx. 51 REFERENCES United Nations, Department of Economic and Social System and Global Value Chains to Expand Rural Affairs, and Population Division. 2019. World Incomes.” Washington, DC: World Bank. Urbanization Prospects: The 2018 Revision. World Bank Group. 2015. “Ethiopia Urbanization Wieser, Christina, and Wondimagegn Mesfin. 2021. Review: Urban Institutions for a Middle- “Employment in Urban and Rural Ethiopia.” Income Ethiopia.” Washington, DC: World Bank Washington, DC: World Bank. Group. https://doi.org/10/25166113/ethiopia- urbanization-review-urban-institutions-middle- World Bank. 2016. “ID4D Country Diagnostic: income-ethiopia. Ethiopia.” Washington, DC: World Bank. ———. 2021. “Ethiopia Socioeconomic ———. 2020. “Ethiopia Poverty Assessment: Dashboard.” 2021. https://www.worldbank. Harnessing Continued Growth for Accelerated org/en/data/interactive/2020/06/24/ethiopia- Poverty Reduction.” Washington, DC: World Bank. socioeconomic-dashboards. https://doi.org/10.1596/33544. Zewdu, Girmachew Adugna. 2018. “Ethiopian Female ———. 2021. “Future of Addis Ababa 2021: Addis Domestic Labour Migration to the Middle East: Ababa’s Strategic Development Framework.” Patterns, Trends, and Drivers.” African and Black Washington, DC: World Bank Group. https:// Diaspora: An International Journal 11 (1): 6–19. imagebank2.worldbank.org/search/33160413. https://doi.org/10.1080/17528631.2017.1342976. ———. 2022. “Ethiopia Rural Income Diagnostics: Leveraging the Transformation in the Agri-Food 52 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES ANNEX 1: MOBILITY FRAMEWORK FOR INTERNATIONAL LABOR MIGRATION IN ETHIOPIA Laws for international labor migration evolved work abroad. The ban was, however, not successful in recent years. Ethiopia was a signatory to the in its aims as Ethiopians retorted to illegal channels Global Compact for Safe, Orderly and Regular for migration. Within two years of Ethiopia banning Migration (GCM) form its inception. To implement labor mobility to the UAE, as many as 30,000 the GCM, the Government of Ethiopia (GoE) works Ethiopians were detained there for irregular closely in partnership with various stakeholders migration (Smith et al. 2020). No. 923 was adopted who are directly or indirectly involved in migration in 2016 in order to prepare the regulatory structure management interventions such as IOM, AU, IGAD, to resume managed labor migration, prior to UNHCR, and non-governmental organizations lifting the ban which was legally lifted in 2018 and (Federal Democratic Republic of Ethiopia 2020). operationalized in 2019. The principal objectives of Since the adoption of the GCM, GoE has developed No. 923 are to cover the establishment of bilateral and revised several practices, proclamations, agreements, the fight to human trafficking, and directives, and policies to better govern migration. clearly define and regulate the role of the private For example, GoE is currently working on a National sector in overseas employment exchange service. Migration Policy. Proclamation on the Prevention It provides for a more streamlined governance and Suppression of Trafficking in Persons and than the previous regulation and allows for Smuggling of Migrants (No. 909/2015) was three recruitment channels: (a) government-to- replaced by a revised Proclamation on Prevention government; (b) via private employment agencies; and Suppression of Trafficking in Persons and and (c) direct recruitment by a foreign employer Smuggling of Migrants which was passed in 2020 (IOM 2017). It was only recently, through the (No. 1178/2020) (Federal Democratic Republic of formation of the Jobs Creation Commission (now Ethiopia 2020). part of the Ministry of Labour and Skills), that GoE policy shifted towards actively facilitating labor The primary instrument to govern international mobility rather than deterring it. labor mobility in Ethiopia is the Overseas Employment Proclamation No. 923/2016 No. 923 was amended and Ethiopian’s Overseas (hereafter referred to as No. 923). Before No. Employment (Amendment) Proclamation No. 923 was passed in 2016, labor migration was 1246/2021 was approved and signed into law in regulated by legislation first introduced in 1998 to 2021. The amendments aim at widening the scope regulate private recruitment agencies, which was of implementation as well as mitigate challenges augmented in 2009 by the Employment Exchange faced during implementation and are related to Services Proclamation No. 632/2009 (Smith et al. educational requirements to access regular labor 2020). Under these regulations, irregular migration migration channels, bilateral labour agreement was common, and migrants frequently ended up (BLA) requirements, contract approval, licensing in vulnerable employment situations. As a result, and registration of private employment agencies GoE banned the migration of Ethiopian workers (PEAs). Other improvements in the labor mobility to employment in Gulf Cooperation Countries framework include a directive issued in 2019 to (the primary destination countries for Ethiopia). determine how private employment agencies This ban both blocked intermediation activity of should provide employment services in the country private employment agencies to source vacancies (including those for overseas employment and and connect them with Ethiopian workers, and to a National Reintegration Directive (No. 65/2018) approvals for Ethiopian workers applications to issued in 2018 to place a legal framework around 53 ANNEX 1: MOBILITY FRAMEWORK FOR INTERNATIONAL LABOR MIGRATION IN ETHIOPIA the reintegration of returnees (Federal Democratic migrating regularly. This process seems to have a Republic of Ethiopia 2020). number of redundancies and unnecessary steps and is not well-aligned with the needs of destination To form the mobility framework, international markets. MoLS is currently putting Proclamation agreements with partner countries are critical No. 1246/2021 into action, adapting the current to complementing domestic regulations. Ethiopia labor mobility process, and digitizing the overseas currently has four Memorandums of Understanding employment administration, the process is lacking (MoUs) on the movement of Ethiopian workers coordination between relevant ministries and bodies, with Saudi Arabia, the UAE, Jordan, and Qatar. as well as an apparent disconnect between migration These MoUs appear to focus largely on migration management at the federal and regional levels. The of Ethiopian domestic workers to these markets, lengthy process places a very real time and cost leaving little scope for managed migration outside burden on the worker, incentivizing them to pay of these narrow corridors (Smith et al. 2020). an agent to take on this burden or even to migrate irregularly to avoid the process altogether and further Inefficiencies in the international labor mobility undermines the competitiveness of Ethiopian workers process remain. Numerous barriers to expanding with workers from other countries. Beyond this labor mobility flows from Ethiopia persist and are process, there are also significant gaps in systems for outlined in Smith et al. (2020). The current process offering protection while abroad, skilling prospective for regular migration is lengthy and disincentivizes migrants, and reintegrating migrants upon return. 54 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES ANNEX 2: LABOR MARKET OUTCOMES: RETURNEES VS. NON-RETURNEES Table A 1: The probability of unemployment - returnees vs. non-returnees (1) (2) (3) National Urban Rural Returnee 0.425*** 0.430*** 0.476*** (0.069) (0.056) (0.140) Youth 0.196** 0.287*** 0.141 (0.077) (0.066) (0.156) Female 0.788*** 0.603*** 0.922*** (0.067) (0.059) (0.132) Education: Ref. – No education Less than primary 0.197* 0.112 0.234 (0.101) (0.102) (0.166) Completed primary 0.285*** 0.166* 0.192 (0.097) (0.101) (0.184) Completed secondary 0.344*** 0.127 0.634 (0.120) (0.121) (0.453) Completed post-secondary 0.048 -0.154 0.084 (0.113) (0.112) (0.525) Married -0.170*** -0.062 -0.390*** (0.066) (0.057) (0.137) Access to/use of any ICT 0.210* -0.132 0.177 (0.125) (0.141) (0.175) Dependency ratio -0.177 -0.102 0.009 (0.147) (0.115) (0.298) Zone population density -0.000 -0.000* 0.000** (0.000) (0.000) (0.000) Zone population density squared 0.000 0.000 -0.000** (0.000) (0.000) (0.000) Region: Ref: - Oromia Afar 0.169 -0.238** 1.246*** (0.133) (0.107) (0.321) Amhara 0.036 0.064 -0.048 (0.084) (0.069) (0.170) Somali -0.083 -0.535*** 0.683** (0.154) (0.175) (0.269) Benishangul-Gumuz -0.978*** -0.426*** -0.828* (0.107) (0.115) (0.477) SNNP -0.197** -0.196*** -0.006 (0.094) (0.069) (0.197) Gambela -0.065 -0.156 1.214*** (0.170) (0.112) (0.444) Harari -0.555*** -0.423*** 0.083 (0.118) (0.132) (0.416) Addis Ababa 0.164** 0.107* (0.066) (0.061) Dire Dawa -0.010 -0.075 0.812*** (0.094) (0.108) (0.272) Constant -1.478*** -0.920*** -2.955*** (0.164) (0.161) (0.525) Observations 12328 10409 1919 Note: The analysis focuses on only the working age population. Robust standard errors are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Source: Authors' estimation based on 2021 LMS. 55 ANNEX 2: LABOR MARKET OUTCOMES: RETURNEES VS. NON-RETURNEES Table A 2: Multinomial logit estimates of the type of employment Wage-employed Self-employed Self-employed Unpaid agriculture non-agriculture family worker Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Returnee -0.0531** 0.552* 0.0225 0.681*** 0.0554** 0.277 -0.0313 (-3.19) (2.40) (0.89) (3.65) (3.04) (1.31) (-1.53) Youth 0.0211 -0.133 -0.0119 -0.699*** -0.122*** 0.527** 0.119*** (1.42) (-0.54) (-0.34) (-3.95) (-4.93) (2.75) (5.04) Female -0.0233 -0.663*** -0.195*** 0.417** 0.0708*** 0.846*** 0.154*** (-1.84) (-3.47) (-8.35) (2.99) (4.45) (5.15) (8.07) Less than primary 0.0347 -0.836*** -0.107*** -0.0764 0.0582* -0.376 0.00913 (1.69) (-3.42) (-3.78) (-0.34) (2.33) (-1.54) (0.36) Completed primary 0.0775*** -1.670*** -0.197*** -0.249 0.107*** -0.823*** 0.0136 (4.09) (-6.65) (-6.34) (-1.17) (4.54) (-3.42) (0.52) Completed secondary 0.141*** -2.997*** -0.362*** -0.670** 0.144*** -1.257** 0.0649 (5.64) (-6.87) (-5.30) (-2.70) (4.02) (-3.29) (1.16) Completed post-secondary 0.333*** -5.302*** -0.490*** -2.528*** 0.0881* -3.093*** 0.0495 (14.08) (-10.14) (-5.96) (-10.44) (2.03) (-9.17) (0.93) Married -0.0365** 0.877*** 0.124*** 0.307* -0.00891 -0.0404 -0.0849*** (-2.93) (4.16) (4.12) (2.28) (-0.46) (-0.24) (-3.87) Access to/use of any ICT -0.0268 -0.434 -0.126** 0.920** 0.177*** -0.567* -0.102*** (-1.05) (-1.41) (-2.99) (2.86) (4.49) (-2.02) (-3.94) Dependency ratio -0.0691** 1.550*** 0.169** -0.174 -0.187*** 1.187*** 0.0649 (-2.70) (3.80) (3.01) (-0.58) (-4.99) (3.37) (1.51) Zone population density -0.000000221*** 0.00000392*** 0.000000397*** 0.00000131*** -0.000000152*** 0.00000235*** -6.02e-10 (-12.03) (11.00) (7.29) (6.15) (-5.26) (8.87) (-0.02) Zone population density 4.88e-14*** -8.65e-13*** -8.71e-14*** -2.79e-13*** 3.58e-14*** -5.39e-13*** -3.45e-15 squared (8.36) (-7.74) (-5.07) (-4.23) (3.79) (-6.62) (-0.28) Afar -0.0458* 1.154*** 0.137** -0.569* -0.150*** 0.969** 0.0647 (-2.52) (3.66) (2.80) (-2.00) (-7.69) (3.28) (1.36) Amhara 0.0181 -0.330 -0.0393 -0.0939 0.0142 -0.217 -0.00546 (0.93) (-1.43) (-1.32) (-0.49) (0.63) (-1.03) (-0.23) Somali -0.0786*** 1.692*** 0.168*** 0.339 -0.102*** 1.221*** 0.0226 (-5.22) (5.59) (3.64) (1.12) (-3.71) (4.19) (0.53) Benishangul-Gumuz -0.126*** 2.909*** 0.227*** 1.552*** -0.0709** 2.030*** -0.0203 (-12.54) (11.10) (5.77) (6.92) (-2.76) (8.52) (-0.66) SNNP -0.0263 0.00600 -0.0613* 0.162 -0.00927 0.663*** 0.101*** (-1.70) (0.03) (-2.12) (0.88) (-0.42) (3.32) (3.92) Gambela -0.0902*** 2.237*** 0.270*** 0.590 -0.100*** 0.921* -0.0737* (-4.81) (5.21) (5.72) (1.57) (-3.49) (2.26) (-2.14) Harari -0.0349* 1.137*** 0.191*** -0.0487 -0.0834*** 0.146 -0.0630* (-2.18) (4.05) (4.53) (-0.21) (-3.42) (0.56) (-2.05) Addis Ababa 0.231*** -3.680*** -0.317*** -0.694*** 0.156*** -1.935*** -0.0800** (7.67) (-6.56) (-10.03) (-4.66) (4.56) (-6.98) (-2.67) Dire Dawa 0.0748*** -0.692** -0.0290 -0.436* 0.0150 -1.004*** -0.0708* (3.30) (-2.67) (-0.74) (-2.24) (0.52) (-3.95) (-2.51) Constant -2.001*** -1.393*** -1.288** (-4.20) (-3.33) (-3.02) Observations 8837 8837 8837 8837 8837 8837 8837 Note: Only employed individuals of working age are included the regressions. For brevity, we have excluded the estimates for employers and other types of employees due to small sample sizes and non-significance. t statistics are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Source: Authors' estimation based on 2021 LMS. 56 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES ANNEX 3: PROPENSITY TO MIGRATE - PROBIT ESTIMATES Table A 3: The propensity to migrate (1) (2) All Rural-to-urban Individual characteristics Female=1 -0.082** 0.161** (0.042) (0.073) Youth=1 0.516*** 0.371*** (0.061) (0.110) Female=1 # Youth=1 0.425*** 0.378*** (0.071) (0.128) Married=1 -0.187*** -0.103 (0.039) (0.074) Education: Ref. – Pre-school/No education Less than primary 0.128*** 0.235*** (0.048) (0.067) Completed primary 0.148*** 0.582*** (0.050) (0.081) Completed secondary 0.128* 0.457*** (0.070) (0.170) Completed post-secondary 0.330*** 0.123 (0.057) (0.127) Access to/use of any ICT=1 0.245*** 0.659*** (0.049) (0.069) LM outcomes: Ref – Unemployed Wage-employed -0.015 0.515*** (0.052) (0.129) Self-employed -0.326*** -0.417*** (0.055) (0.115) Unpaid family worker -0.365*** -0.839*** (0.070) (0.130) Others -0.112 0.333 (0.143) (0.297) Inactive -0.300*** -0.305*** (0.054) (0.118) Household characteristics: Female-Headed HH 0.040 0.200** (0.043) (0.085) Dependency ratio -0.573*** -1.059*** (0.068) (0.114) Destination pull factors: Lifetime migrant inflow at destination ('000) -0.001*** 0.000* (0.000) (0.000) Woreda emp. rates at destination 0.080 -3.953*** (0.146) (0.277) Woreda wages at destination (log) -0.055* 0.489*** (0.030) (0.055) Origin push factors: Origin zone poverty rate -0.046 -1.267*** (0.139) (0.213) Origin zone density -0.000*** 0.000 (0.000) (0.000) Origin zone density squared 0.000*** -0.000 (0.000) (0.000) 57 ANNEX 3: PROPENSITY TO MIGRATE - PROBIT ESTIMATES (1) (2) All Rural-to-urban Previous region of residence: Ref. – Oromia Tigray -0.183 -0.404* (0.114) (0.210) Afar 0.061 0.692*** (0.080) (0.173) Amhara 0.145*** 0.248*** (0.044) (0.076) Somali -0.485*** -0.948*** (0.091) (0.123) Benishangul-Gumuz 0.298*** 0.713*** (0.078) (0.131) SNNP -0.040 0.072 (0.042) (0.072) Gambela -0.117 -0.635*** (0.082) (0.120) Harari -0.119 -1.314** (0.217) (0.572) Addis Ababa -0.091 (0.074) Dire Dawa -0.057 -0.711*** (0.226) (0.263) Constant 0.059 -1.420*** (0.295) (0.502) Observations 79954 32391 Note: World Bank staff calculations using LMS 2021. Only people aged 15 or more are included in the regressions. Data on zonal poverty rates, rural remoteness index and ecological zones were extracted from Ethiopia's PTI data. Robust standard errors are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. 58 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES ANNEX 4: MULTINOMIAL LOGISTIC REGRESSION MODEL The multinomial logistic regression model is a generalization of the binary logistic model such that the probability that person i will chose alternative j is: exp ( X' i βj ) Prij = Pr ( yi = j ) = (1) ∑m k=1 exp ( X'i βk ) Where the dependent variable y is an unordered categorical variable, the alternative choices (e.g., type of migration or reasons for migration) are represented by j=1,2,…,m, while Xi' is a vector of independent variables with its associated vector of β parameters to be estimated. The likelihood of selecting each alternative sums up to one (i.e., ∑m i=1 Prij = 1). There are j-1 sets of coefficients to be estimated because one set of coefficients needs to be normalized to zero to estimate the models (usually β1=0). This implies that the coefficients of other alternatives are interpreted in reference to the base outcome (in this case, being a non-migrant or economic reasons for migration). The interpretation of alternative j in comparison to the base outcome is ‘a change in an independent variable Xi makes the selection of alternative j more or less likely’ (depending on the sign found on the associated β coefficient). Hence the marginal effect of a change in an independent variable on the probability of selecting alternative j is written as: ∂Prij/∂X_i = Prij (βj ─ βi) (2) where βi is an average of all the coefficients. For this reason, the marginal effects in a multinomial logistic model do not necessarily coincide with the sign and significance of the coefficients. Although there may be j-1 set of coefficients because one set is normalized to zero, there are j sets of marginal effects to be estimated. Additionally, the coefficients that are obtained depend on the choice of the base category while the marginal effects are the same regardless of the base outcome selected (and this is because the marginal effects are not interpreted with respect to a base category). Marginal effects can be interpreted as a unit change in an independent variable which changes the probability of selecting alternative j by the marginal effect expressed as a percent. The marginal effects sum up to zero because a person is likely to select as many choices as there are; hence if an individual is more likely to select the first two alternatives, they are going to be less likely to choose the last alternative which makes these effects cancel out and sum up to zero. 59 ANNEX 4: MULTINOMIAL LOGISTIC REGRESSION MODEL Table A 4: Drivers of rural migration (1) (2) Rural-to-rural Rural-to-urban Individual characteristics: Female=1 -0.204 0.223 (-0.94) (1.54) Youth=1 0.893*** 0.747*** (3.10) (3.66) Female=1 # Youth=1 0.866*** 0.658** (2.67) (2.76) Married=1 0.0722 -0.145 (0.34) (-1.04) Education: Ref. – Pre-school/No education Less than primary 0.206 0.426** (1.06) (3.22) Completed primary -0.0885 1.036*** (-0.34) (6.66) Completed secondary -0.356 0.686* (-0.59) (2.06) Completed post-secondary 0.453 0.120 (1.31) (0.50) Access to/use of any ICT=1 0.291* 1.293*** (1.65) (8.92) LM outcomes: Ref – Unemployed Wage-employed 0.672* 0.997*** (1.84) (3.80) Self-employed -0.817** -0.732*** (-2.52) (-3.34) Unpaid family worker -0.748** -1.474*** (-2.33) (-5.92) Others -2.617*** 0.440 (-2.62) (0.72) Inactive -0.755** -0.498* (-2.31) (-2.20) Household characteristics: Female-Headed HH 0.652*** 0.548*** (2.65) (3.37) Dependency ratio -1.489*** -1.805*** (-4.83) (-8.55) Destination pull factors: Lifetime migrant inflow at destination ('000) -0.000595 -0.0000286 (-0.68) (-0.07) Woreda emp. rates at destination 2.121 -8.773*** (1.59) (-11.79) Woreda wages at destination (log) 0.0385 0.955*** (0.35) (7.98) Drought prone, highland 0.795*** -0.0569 (3.61) (-0.36) Drought prone, lowland -0.0115 0.686** (-0.04) (3.28) Moisture reliable, lowland -0.490* -0.449** (-1.79) (-2.64) Pastoralist -1.065** -2.316*** (-2.06) (-6.19) Origin push factors: Origin zone poverty rate 0.603 -2.167*** (1.08) (-4.84) 60 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES (1) (2) Rural-to-rural Rural-to-urban Origin zone density -0.00000110** -1.26e-08 (-2.25) (-0.04) Origin zone density squared 1.08e-14* -2.61e-15 (1.81) (-0.72) Origin rural remoteness index 0.00145 -0.0119 (0.17) (-1.89) Previous region of residence: Ref. – Oromia Tigray -1.258 -1.375** (-1.51) (-2.72) Afar -0.603 1.067** (-1.42) (2.64) Amhara 0.328 0.453** (1.48) (3.13) Somali -0.211 0.208 (-0.34) (0.57) Benishangul-Gumuz 1.167*** 1.646*** (2.81) (5.25) SNNP -0.198 0.0351 (-1.01) (0.23) Gambela 0.161 -0.745** (0.50) (-2.81) Harari -2.020*** -3.780*** (-3.67) (-3.29) Dire Dawa -3.236*** -0.273 (-2.92) (-0.36) Constant -3.791** -1.137 (-2.07) (-1.03) Observations 32725 32725 Note: World Bank staff calculations using LMS 2021. Only people aged 15 or more are included in the regressions. Data on zonal poverty rates, and rural remoteness index and ecological zones were extracted from Ethiopia's PTI data. t statistics are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. 61 ANNEX 4: MULTINOMIAL LOGISTIC REGRESSION MODEL Table A 5: Drivers of urban migration (3) (4) Urban to rural Urban to urban Female=1 0.0590 0.449* (0.21) (2.49) Youth =1 1.411*** 0.302 (4.66) (1.20) Female=1 # Youth =1 -0.314 0.643* (-0.81) (2.14) Married=1 -0.877*** -0.128 (-3.91) (-0.79) Education: Ref. – Pre-school/No education Less than primary 0.660* 0.846*** (2.37) (4.39) Completed primary 0.938** 1.713*** (3.02) (8.10) Completed secondary 2.356*** 2.558*** (5.40) (7.49) Completed post-secondary 2.621*** 2.314*** (7.08) (8.87) Access to/use of any ICT=1 1.072*** 1.974*** (3.65) (8.43) LM outcomes: Ref – Unemployed Wage-employed -0.0995 0.678* (-0.26) (2.45) Self-employed -0.541 -0.656* (-1.44) (-2.54) Unpaid family worker -0.417 -1.112*** (-1.13) (-3.57) Others -0.770 2.070** (-0.65) (3.18) Inactive -0.755 -0.332 (-1.94) (-1.30) Household characteristics: Female-Headed HH 0.0359 0.496** (0.14) (2.90) Dependency ratio -0.270 -2.777*** (-0.70) (-9.68) Destination pull factors: Lifetime migrant inflow at destination ('000) -0.00206* 0.00344*** (-2.13) (6.68) Woreda emp. rates at destination 0.886 -9.494*** (0.74) (-12.97) Woreda wages at destination (log) -0.590*** 1.311*** (-4.69) (9.36) Drought prone, highland 0.135 -0.770*** (0.56) (-3.96) Drought prone, lowland -0.416 0.132 (-1.12) (0.42) Moisture reliable, lowland -1.520*** -1.539*** (-3.31) (-7.08) Pastoralist -2.683*** -2.957*** (-4.07) (-8.51) Origin push factors: Origin zone poverty rate -0.590 -3.222*** (-0.90) (-6.68) 62 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES (3) (4) Urban to rural Urban to urban Origin zone density -0.00000360*** -0.00000255*** (-5.79) (-6.46) Origin zone density squared 4.11e-14*** 2.84e-14*** (5.44) (5.92) Previous region of residence: Ref. – Oromia Tigray 0.120 -0.457 (0.28) (-1.22) Afar -0.415 0.122 (-0.72) (0.23) Amhara 0.531 0.459* (1.86) (2.44) Somali 0.476 1.467*** (0.70) (4.05) Benishangul-Gumuz 0.758 1.400*** (1.16) (3.41) SNNP -0.227 -0.377* (-0.87) (-2.02) Gambela -0.311 0.435 (-0.63) (1.26) Harari 0.621 -1.082 (0.93) (-0.84) Addis Ababa 1.817*** 1.677*** (4.39) (4.39) Dire Dawa 0.123 1.375 (0.06) (1.69) Constant 2.010 -4.626*** (1.24) (-3.79) Observations 31941 31941 Note: World Bank staff calculations using LMS 2021. Only people aged 15 or more are included in the regressions. Data on zonal poverty rates, and rural remoteness index and ecological zones were extracted from Ethiopia's PTI data. t statistics are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. 63 ANNEX 4: MULTINOMIAL LOGISTIC REGRESSION MODEL Table A 6: Determinants of migration motivations for rural-to-urban adult migrants Economic Education Family Shocks Other motives Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Female=1 -0.136*** 0.428 -0.00150 1.470*** 0.138*** 0.273 -0.00219 0.112 0.00173 (-4.64) (0.88) (-0.11) (5.82) (5.53) (0.87) (-0.41) (0.31) (0.19) Youth=1 -0.187*** 1.931*** 0.0726*** 1.834*** 0.176*** -0.446 -0.0207** -1.614 -0.0412*** (-5.89) (4.78) (4.95) (6.18) (6.32) (-1.01) (-2.81) (-1.79) (-3.81) Female=1 # Youth=1 -0.500 -1.229*** -0.337 0.331 (-0.94) (-3.72) (-0.65) (0.35) Married=1 -0.0942** -0.958*** -0.0598*** 0.790*** 0.150*** 0.472 0.00712 0.0000770 -0.00312 (-2.67) (-3.86) (-5.19) (4.33) (4.76) (1.30) (0.97) (0.00) (-0.50) Less than primary -0.0496 0.926* 0.0174* 0.237 0.0348 -0.221 -0.00812 0.215 0.00548 (-1.28) (2.14) (2.38) (1.10) (1.04) (-0.70) (-0.95) (0.59) (0.42) Completed primary -0.113* 2.212*** 0.0866*** 0.428 0.0572 -0.155 -0.00886 -0.838 -0.0224* (-2.56) (5.08) (4.86) (1.86) (1.56) (-0.41) (-0.93) (-1.69) (-1.98) Completed secondary -0.331*** 3.516*** 0.223*** 1.140** 0.136* -0.0355 -0.0152 0.116 -0.0137 (-4.31) (6.71) (4.54) (3.05) (2.12) (-0.04) (-1.02) (0.16) (-0.83) Completed post-secondary -0.104 2.979*** 0.204*** -0.103 -0.0436 -1.590* -0.0265** -1.846* -0.0305** (-1.61) (5.78) (3.99) (-0.30) (-0.96) (-2.42) (-3.13) (-2.42) (-2.83) Access to/use of any ICT=1 0.0950 -0.0330 0.00602 -0.483* -0.0873 -0.730* -0.0161 -0.0382 0.00238 (1.87) (-0.10) (0.41) (-2.06) (-1.93) (-2.02) (-1.44) (-0.10) (0.32) Wage-employed 0.454*** -0.699 -0.00116 -2.638*** -0.453*** -1.565** -0.0195 0.219 0.0199 (10.16) (-1.94) (-0.06) (-10.16) (-11.17) (-3.21) (-1.84) (0.40) (1.60) Self-employed 0.138** -0.184 0.00114 -0.862*** -0.152*** -0.415 -0.00460 0.578 0.0174 (3.26) (-0.42) (0.05) (-4.01) (-4.11) (-1.14) (-0.60) (1.22) (1.65) Unpaid family worker -0.178** 1.890*** 0.0874** 0.634* 0.0852 -0.196 -0.0101 0.965 0.0156 (-2.87) (3.78) (3.21) (2.07) (1.64) (-0.38) (-0.97) (1.50) (1.12) Others 0.483*** 0.335 0.0567 -0.632 0.0203 -14.42*** -0.301*** -12.36*** -0.259*** (4.08) (0.41) (1.37) (-1.09) (0.20) (-18.56) (-5.34) (-20.70) (-4.27) Inactive -0.291*** 2.850*** 0.130*** 1.011*** 0.133*** 0.142 -0.00643 1.984*** 0.0345** (-6.24) (8.20) (5.65) (4.70) (3.61) (0.39) (-0.87) (4.07) (2.92) Female-Headed HH 0.0200 -0.120 -0.00486 -0.178 -0.0344 0.310 0.00748 0.499 0.0118 (0.58) (-0.57) (-0.48) (-0.99) (-1.12) (0.87) (1.01) (1.51) (1.57) Dependency ratio -0.211*** -0.121 -0.0235 1.123*** 0.187*** 1.315** 0.0220* 1.471* 0.0257 (-3.34) (-0.28) (-1.16) (3.38) (3.32) (2.76) (2.05) (2.28) (1.92) Lifetime migrant inflow at 0.000164 0.00144 0.0000905* -0.000957 -0.000165 -0.00379** -0.0000781* -0.000759 -0.0000115 destination ('000) (1.17) (1.68) (2.11) (-1.29) (-1.29) (-3.02) (-2.45) (-0.54) (-0.38) Born in destination zone -0.188*** 1.270*** 0.0533*** 0.781*** 0.116** 0.403 0.00271 0.999** 0.0160 (-4.58) (4.75) (3.56) (3.60) (3.13) (1.26) (0.40) (2.60) (1.91) Woreda emp. rates -0.287* -1.636 -0.111* 1.788* 0.314* 2.819* 0.0526 1.817 0.0309 at destination (-1.97) (-1.54) (-2.04) (2.38) (2.50) (2.19) (1.92) (1.32) (1.04)h Origin zone poverty rate 0.1000 0.787 0.0503 -0.918 -0.178 -1.041 -0.0196 1.971 0.0471 (0.79) (0.95) (1.26) (-1.42) (-1.63) (-0.98) (-0.85) (1.60) (1.83) Tigray -0.0651 -3.214** -0.0743*** 0.731 0.186 -13.32*** -0.0317*** -0.998 -0.0154 (-0.34) (-2.65) (-5.10) (0.90) (0.97) (-28.42) (-5.27) (-0.78) (-1.31) Afar -0.497*** -1.026 -0.0704*** 0.851 -0.0937* 4.529*** 0.673*** 0.737 -0.0115 (-9.31) (-1.34) (-4.86) (1.89) (-2.22) (11.33) (11.85) (0.97) (-1.19) Amhara 0.0386 -0.508* -0.0279* -0.0677 -0.00222 -1.160** -0.0212** 0.383 0.0127 (1.07) (-2.12) (-1.97) (-0.36) (-0.07) (-2.79) (-3.06) (1.09) (1.30) Somali -0.183 0.217 -0.00813 0.943 0.182 0.876 0.0231 -0.538 -0.0133 (-1.56) (0.33) (-0.20) (1.74) (1.55) (1.62) (1.02) (-0.57) (-1.33) Benishangul-Gumuz -0.192* -0.189 -0.0319 0.381 0.00750 2.476*** 0.235*** -1.353* -0.0188** (-2.45) (-0.47) (-1.86) (1.00) (0.13) (6.31) (4.10) (-2.17) (-2.71) 64 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Economic Education Family Shocks Other motives Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. Coeff. Marg.Eff. SNNP -0.000284 -0.483* -0.0295* 0.162 0.0376 0.134 0.00453 -0.764 -0.0123 (-0.01) (-1.99) (-2.13) (0.86) (1.18) (0.43) (0.44) (-1.59) (-1.63) Gambela -0.330*** 0.833 0.0108 1.590*** 0.306*** 1.101 0.0156 0.589 -0.00246 (-3.88) (1.78) (0.34) (3.80) (3.57) (1.66) (0.59) (0.78) (-0.17) Harari -0.182 -14.01*** -0.0770*** 0.417 0.0204 1.819 0.109 2.218* 0.130 (-0.99) (-18.11) (-5.33) (0.56) (0.19) (1.45) (0.78) (2.34) (1.27) Dire Dawa 0.124 -1.864 -0.0628*** -1.301 -0.145* -14.11*** -0.0317*** 1.616 0.115 (0.98) (-1.94) (-3.42) (-1.34) (-2.16) (-22.64) (-5.27) (1.53) (0.94) Constant -4.637*** -4.018*** -3.781** -5.959*** (-4.47) (-5.16) (-2.83) (-3.88) Observations 4186 4186 4186 4186 4186 4186 4186 4186 4186 Note: Interaction terms do not have marginal effects because mathematically, Female#Youth cannot change while both Female and Youth are held fixed which is a requirement for calculating the marginal effects. t statistics in parentheses * p<0.05 ** p<0.01 *** p<0.001. 65 ANNEX 5: OAXACA-BLINDER DECOMPOSITION METHODOLOGY ANNEX 5: OAXACA-BLINDER DECOMPOSITION METHODOLOGY Here we conduct the Oaxaca-Blinder decomposition (Blinder 1973; Oaxaca 1973) to identify the source of disparities between migrants and native workers in urban areas. The decomposition allows the division of the observed wage differential between migrants and native workers, into the explained component (i.e., differences in characteristics or endowments such as education or work experience), and the unexplained component which is the differences in coefficients or returns to these endowments). YiM = αiM + βM XM M (3) i + εi YiN = αiN + β N XN N (4) j + εi Where Y is the dependent variable (log wages) and α is a constant, while X is a vector of independent variables and β the associated vector of parameters to be estimated. The superscripts M and N in equation 1 and 2 represent the distinguishes the wage equation for migrant and native workers, respectively. Subtracting the two equations above yields a two-part decomposition of: YiN ⎻ YiM = β N XiN ⎻ XiM + αiN ⎻ αiM + XiM βN ⎻ βM (5) The explained portion of the wage differential (YiN ⎻ YiM) in equation 5 is captured by the term βN(XiN ⎻ XM i ), which is the differential attributable to the differences in the mean of observable characteristics between migrant (M) and native (N) workers. The unexplained component is the differences in constant and coefficient estimates that are respectively captured by the terms (αi N⎻ αiM ) + XiM (β N⎻ β M ) which is the wage differential that would remain if native workers had the average characteristics of migrant workers. The sum of the explained and unexplained components represents the total wage differential at means between migrant and native workers. 66 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Table A 7: Oaxaca wage decomposition between recent rural-to-urban migrants and urban non-migrants Coeff. overall 8.158*** Urban non-migrants (0.011) 7.569*** Rural-urban migrants (0.033) 0.589*** difference (0.035) 0.458*** explained (0.028) 0.131*** unexplained (0.025) explained 0.050*** Female (0.007) 0.088*** youth (0.010) -0.023*** Less than primary (0.006) -0.014* Completed primary (0.007) 0.024*** Completed secondary (0.007) 0.166*** Completed post-secondary (0.019) 0.006 Public sector (0.004) 0.024*** Non-permanent job (0.005) -0.019*** Professionals & technicians (0.006) 0.101*** Clerical, services & sales (0.014) 0.001 Skilled agricultural, forestry & fishery (0.002) -0.002 Craft & related trades (0.003) -0.013*** Plant/machine operators & assemblers (0.003) 0.022** Elementary occupations (0.010) -0.002 Other occupations (0.001) -0.000 Woreda unemp. rates at destination (0.001) -0.000 Afar (0.000) 0.008*** Amhara (0.003) 0.006*** Somali (0.001) 0.001** Benishangul-Gumuz (0.000) 0.005* SNNP (0.003) 0.000 67 ANNEX 5: OAXACA-BLINDER DECOMPOSITION METHODOLOGY Coeff. Gambela (0.000) 0.000 Harari (0.000) 0.029*** Addis Ababa (0.005) -0.000 Dire Dawa (0.000) unexplained Female -0.050* (0.027) youth 0.033 (0.025) Less than primary -0.020 (0.023) Completed primary -0.003 (0.023) Completed secondary 0.010 (0.009) Completed post-secondary 0.039 (0.037) Public sector -0.058 (0.042) Non-permanent job 0.009 (0.037) Professionals & technicians -0.044 (0.035) Clerical, services & sales 0.026 (0.052) Skilled agricultural, forestry & fishery 0.007* (0.004) Craft & related trades -0.015 (0.010) Plant/machine operators & assemblers 0.009 (0.008) Elementary occupations -0.049 (0.041) Other occupations 0.001 (0.002) Woreda unemp. rates at destination -0.100** (0.043) Afar -0.000 (0.001) Amhara 0.011 (0.017) Somali -0.003*** (0.001) Benishangul-Gumuz -0.002 (0.002) SNNP -0.006 (0.011) Gambela -0.001 (0.001) Harari -0.001 (0.001) 68 VOLUNTARY MIGRATION IN ETHIOPIA: IN SEARCH FOR WORK AND BETTER OPPORTUNITIES Coeff. Addis Ababa -0.024* (0.013) Dire Dawa -0.003* (0.001) Constant 0.364* (0.189) Observations 17010 Note: Authors’ estimation based on 2021 LMS. The results are only for the working age population (between 15 and 64 years old) in wage employment. Robust standard errors are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. 69 ANNEX 6: EFFECTS ON DESTINATION AREAS ANNEX 6: EFFECTS ON DESTINATION AREAS Table A 8: The effect of rural-to-urban migrant share on labor market indicators at woreda level (1) Unemployment (2) Wages Share of rural-urban migrants 0.003 -0.340** (0.038) (0.166) Share of females 0.108 0.146 (0.100) (0.442) Share of youth -0.041 -0.271 (0.076) (0.285) Education: Ref. – No education Share of less than primary 0.014 -0.337 (0.063) (0.301) Share of completed primary 0.100 0.051 (0.075) (0.282) Share of completed secondary -0.020 -0.127 (0.099) (0.401) Share of completed post-secondary -0.325*** 0.513 (0.110) (0.331) Occupational segregations: Ref. Managers Share of professionals/technicians -0.064 -0.346 (0.200) (0.618) Share of clerical, services & sales -0.236 -1.433** (0.204) (0.595) Share of skilled agricultural workers -0.295 -1.393** (0.197) (0.640) Share of crafts/related trades -0.118 -0.808 (0.208) (0.639) Share of plant/machinery workers 0.006 -0.984 (0.259) (0.684) Share of elementary occupations -0.273 -1.367** (0.194) (0.581) Share of other occupations (e.g., armed forces) -0.228 -1.698** (0.324) (0.802) Sector: Ref. – Agriculture Industry 0.024 -0.404 (0.078) (0.270) Services 0.026 -0.134 (0.057) (0.214) Town size 0.000* 0.000 (0.000) (0.000) Town size squared -0.000 -0.000 (0.000) (0.000) Addis Ababa woredas 0.029 0.040 (0.018) (0.066) Constant 0.298 9.700*** (0.207) (0.640) Observations 411 401 R² 0.144 0.249 Note: The analysis focuses on only the working age population. Robust standard errors are reported in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01. Source: Authors’ estimation based on 2021 LMS. 70