Internal Migration in Ethiopia Evidence from a Quantitative and Qualitative Research Study Tom Bundervoet March 2018 © 2017 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution—Please cite the work as follows: “Bundervoet, Tom. 2018. Internal Migration in Ethiopia: Evidence from a Quantitative and Qualitative Research Study. © World Bank.� All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. SUMMARY Internal migration or labor mobility in Ethiopia remains limited. In the five years prior to the 2013 Labor Force Survey (the last survey with systematic migration data), six percent of Ethiopians changed zone of residence. Though the scale of internal migration has not increased between 1999 and 2013, its pattern has changed, with rural-urban migration becoming the dominant migration flow between 2008 and 2013. While as a share of their population, smaller cities and towns attracted most rural migrants, in absolute terms Addis has been the main destination, with close to 40 percent of all rural migrants moving to the capital. Next to urban areas, the Regions of Gambella, Afar, and Benishangul-Gumuz also experienced net population inflows, likely related to the availability of large swathes of land for commercial agriculture. Education and demographics drive migration. Younger and better-educated rural dwellers are more likely to migrate compared to older or less-educated villagers, and this is true both for rural-urban and within- rural migration. The effect of education is strong, with rural dwellers who obtained at least some secondary education being 26 percentage points more likely to migrate, all else equal (the bulk of rural-urban migrants however are educated at the primary level). Rural to urban migration had a dual nature, with young and relatively less educated women moving to Addis Ababa for domestic work and slightly older and better-educated rural dwellers moving to secondary urban centers to work in commerce, agriculture, and services. Characteristics of the origin zone also influence the propensity to migrate: People in rural zones with high population densities were more likely to migrate, consistent with a potential role of land shortages. Poverty and remoteness inhibited migration, with people in zones with a higher poverty rate and further away from an all-weather road being less likely to move. The qualitative research confirmed that most migrants move to urban areas in search for work and better opportunities (mainly education), and to escape rural areas they describe as bereft of hope and prospects. Young women also migrate to escape arranged marriages and traditional gender roles. The migration experience is described as tough and full of challenges, and risky for young women in particular. Despite the many challenges, the bulk of migrants rate their migration as positive, saying that it opened up opportunities that were unthinkable in rural areas. Life in the city however is hard, with migrants complaining of frequent harassment by local authorities and law enforcement (and for young women, sexual harassment by brokers and employers), the difficulty of obtaining ID cards, and, related, the difficulty of accessing different types of government support. Interviews with regional and city authorities revealed a staunchly negative view of rural-urban migration, labeling it as “unacceptable� or “illegal� and exacerbating problems in urban areas. SME bureaus and BOLSAs confirmed not providing services to migrants as they do not have ID cards. As Ethiopia continues to grow and develop and rural youth become more educated, internal migration will further increase. This is a positive development considering the large private welfare effects of migration in Ethiopia. To maximize the benefits of migration and minimize its costs, both for the migrant and the sending household and society as a whole, increased labor mobility and town development to avoid overconcentration in any given city would need to be considered as important elements of the development process. 1 Table of Contents 1. Introduction ................................................................................................................... 3 2. Internal migration in Ethiopia: Scale, patterns, and trends ............................................... 6 3. Who migrates where and why? ..................................................................................... 13 4. A qualitative study on rural migrants in urban areas ...................................................... 19 4.1 Why do rural youth migrate? ........................................................................................ 20 4.2 Transition to urban life and job search .......................................................................... 21 4.3 Migration experiences .................................................................................................. 22 4.4 Specific challenges faced by rural migrants .................................................................... 24 4.5 Migrant’s advice to rural youth and aspiring migrants ................................................... 25 4.6 Regional and local authorities’ view on internal migration ............................................ 25 4.7 Migrant’s suggestions to the government...................................................................... 27 5. Conclusions and policy perspectives .............................................................................. 28 References .......................................................................................................................... 30 Annex 1: Data and definitional limitations to the study of internal migration ....................... 32 Annex 2: Qualitative research study on rural-urban migration .............................................. 33 Annex 3: Origin zone by city ................................................................................................. 36 Annex 4: Migration motivations, disaggregated by migration type and gender ..................... 39 Annex 5: Estimating the drivers of migration ........................................................................ 40 Annex 6: Labor market indicators of migrants, by migration type ......................................... 43 1. Introduction In many developing countries, disparities in living standards between lagging and leading areas, or between rural and urban areas, are large1. In Ethiopia too, these disparities remain substantial, despite having narrowed over the past decade: in 2016, poverty rates in the poorest Region were four times higher than poverty in the least poor Region;2 poverty in rural areas was almost twice the rate in urban areas. These disparities in living standards largely correspond with disparities in economic density, tending to be higher in the places with high economic density (mainly cities and towns) and lower in places far away from it (mainly rural and relatively remote places). The spatial disparities in living standards offer a compelling motivation for people in lagging areas to move closer to economic density. Reducing “economic distance� (the distance between areas where economic activity is concentrated and areas that lag) is key for economic growth and poverty reduction in the lagging areas. The most straightforward way to reduce economic distance is for people to move closer to economic density. Unfortunately, many governments still see labor mobility and rural-urban migration as undesirable and try to restrict internal movement of people, typically out of fears of urban unemployment, overburdened city infrastructure, or a casual attribution of rising urban crime to rural migrants. By attempting to stem internal migration, governments impose the cost of forgone opportunities for economic growth, poverty reduction and convergence in living standards (Box 1). Ethiopia too has a history of trying to prevent or limit internal migration. In the 1980s, The Derg regime attempted to control population mobility by grouping, often forcefully, farmers into grid-plan villages, usually heavily monitored by the army. This villagization Land Reform Policy (1984) aimed to increase agricultural production and improve delivery of services such as education and health (but was unsuccessful in doing so)3. Since the early 2000s, Ethiopia’s successive development plans have tended to view internal migration as a problem that needs to be solved. The 2002 Sustainable Poverty Reduction Strategy (SPRSP) considered spontaneous migration as negative and noted that only well-planned migration organized by the government can be productive (p.56). The 2006 Plan for Accelerated and Sustained Development to End Poverty (PASDEP) continued that line, claiming that migration worsens urban development problems, such as poverty, unemployment, and the spread of HIV (p.41 and p.120). The more recent Growth and Transformation Plans (GTP-I and GTP-II) do not mention migration at all. A small but growing micro-economic empirical literature finds large welfare effects of internal migration. Tracking migrants over a period of 13 years in Tanzania, Beegle, De Weerdt and Dercon (2011) find that individuals who migrated within the country added 36 percentage points to their consumption growth. For Ethiopia, De Brauw, Mueller and Woldehanna (2017) find higher impacts, with both rural to rural and rural to urban migration leading to substantial gains in real consumption levels (adjusted for rural-urban differences in cost of living). Focusing on China, De Brauw and Giles (2018) find that the easing of migration restrictions in 1988 through the issue of a national ID led to large welfare gains among households remaining in migrant-sending villages, with the effects being particularly large for the poorest tercile. Internal migration however has disutility as well: Though a migration subsidy experiment in Bangladesh led to substantial increases in consumption, the broader welfare effects were smaller (but still 1 Leading areas are the most advanced on social and economic poverty, while lagging areas are the least advanced. 2 Tigray Region was the poorest Region in 2016, while Harari was the Region with the lowest poverty rate (NPC, 2017). 3 Ethiopia’s land tenure system can also be considered an implicit barrier to migration. The effects of the land tenure system on mobility appear however to be small (De Brauw and Mueller, 2012). positive) due to the disutility of migration. This disutility was mainly caused by poor housing conditions faced by migrants in urban areas (Bryan, Chowdhury and Mobarak, 2014; Lagakos, Mobarak and Waugh, 2018). This Note takes a closer look at the scale and nature of internal migration in Ethiopia. While a Population Census offers the best source of data to study internal migration, the most recent Census was conducted in 2007. The next Census is planned to happen towards end 2018, while the census microdata will likely not be available before 2020. As a result, this paper uses three rounds of the Labor Force Surveys (LFS) and the to look at trends and patterns of internal migration in Ethiopia and the characteristics of internal migrants. At the outset, it is worth mentioning that there are some data limitation and definitional issues that complicate the study of internal migration. These issues and limitations are presented in Annex 1. To contextualize the quantitative analysis, a qualitative research study with rural migrants in urban areas was conducted in May 2017. The design of the qualitative research is presented in Annex 2. This Note proceeds as follows: Section 2 presents the scale and pattern of internal migration, focusing particularly on rural-to-urban migration. This section also identifies the main receiving and sending areas and calculates net migration rates, identifying which areas and cities have net inflows of migrants and which have net outflows. Section 3 focuses on the characteristics and motivations of migrants, focusing on push and pull factors and disaggregating by type of migration. The main findings from the qualitative research are summarized in Section 4. The final section concludes and formulates some tentative recommendations. Box 1. From advocating restrictions to letting them move Economic theory now recognizes that governments should not try to hold on to people Many governments around the world actively attempt to discourage internal population movements from rural to urban areas. In a 2013 UN survey on Population and Development, 148 out of the 185 surveyed countries with data (80 percent) had government policies aimed at reducing internal migration from rural to urban areas. Such efforts are particularly prevalent among countries in Africa, where 85 percent of countries have policies aimed at reducing rural-to- urban migration. Only five countries in the world had policies aimed at speeding up migration to urban areas: China, Sri Lanka, Poland, Tajikistan, and the Maldives.4 Governments’ efforts to control rural-urban migration have their roots in the early influential literature on the links between rural-urban migration and urban unemployment. In the well-known Harris-Todaro model (1970), differences in expected incomes between rural and urban sectors attract migrants from rural areas. Wages in the modern urban sector are fixed and exogenous, and jobs are rationed. Only a small fraction of rural migrants find employment in the modern urban sector, with the others unemployed or underemployed in an urban informal sector. Job creation programs in urban areas raise the expected urban income, stimulating further rural-urban migration and, if the labor demand elasticity in urban areas is large enough, increase the level of urban unemployment. This implication of the model was particularly important for policy because it argued against making cities attractive and implicitly endorsed measures to discourage or reverse migration (Commission on Growth and Development, 2009). Though the Harris-Todaro model has been and remains influential, evidence supporting the predicted link between migration and urban unemployment is weak. Many of the critical assumptions and predictions of the model have not been supported by subsequent empirical studies of labor markets in developing countries. More robust and more plausible alternative models of migration have since emerged, with very different policy implications.5 In particular, increasing returns to scale in the modern sector (vs. constant returns in agriculture) and spillovers from clustering imply that movement from lagging to leading places could have substantial growth and welfare payoffs. In addition, on average, migration brings sizable economic benefits to the migrant in terms of increased consumption levels. Also, migrants who move to cities tend to maintain strong links with their home communities, sending back remittances that boost consumption and investment in origin communities and help to converge living standards across space. In that sense, limiting migration comes down to slowing down development. As argued by the World Development Report on Reshaping Economic Geography, the policy challenge is not to keep people from moving, but how to keep them from moving for the wrong reasons. Agglomeration forces and economic opportunities will inevitably pull workers and families to cities, and the goal for policy is how to best accommodate these flows. To avoid migration for the wrong reasons, governments should work to eliminate or alleviate the factors that push people out of their origin areas, such as agricultural decline, due to pressures of population growth or environmental degradation, inefficient land tenure systems, and lack of adequate public services. Migration due to push factors is unlikely to add to agglomeration benefits but likely to exacerbate the urban congestion that policy-makers strive so hard to avoid. 4 World Population Policies Database (http://esa.un.org/poppolicy/about_policy_section.aspx). 5 See Lall, Selod and Shalizi (2006); Commission on Growth and Development (2009). 2. Internal migration in Ethiopia: Scale, patterns, and trends Internal migration in Ethiopia remains limited6. In the five years prior to the 2013 LFS, about 6.5 percent of the Ethiopian adult population moved zone of residence, marginally higher than the share in 1999 (5.7 percent). In rural areas in particular, mobility is limited, with a mere 3.5 percent of adults moving zones between 2008 and 2013 (the five years preceding the 2013 LFS-Table 1). Migrants account for a higher share of the population in urban areas: In 2013, 17 percent of urban dwellers were recent migrants (came to the city in the five years up to 2013). At the regional level, Gambella and Benishangul-Gumuz attracted most migrants in the five years up to 2013, presumably related to the availability of agricultural land in these regions. Table 1. Internal migrants as share of the population, recent migrants and lifetime migrants Survey year 1999 2005 2013 Recent All-time Recent All-time Recent All-time migrants migrants migrants migrants migrants migrants (%) (%) (%) (%) (%) (%) Whole country 5.70 32.34 7.52 28.17 6.49 22.59 Rural areas 3.61 25.88 4.93 20.58 3.49 13.42 Urban areas 16.87 66.51 19.99 64.64 17.25 55.41 Region Tigray 6.56 36.72 9.98 32.94 6.64 23.58 Afar 10.57 37.36 12.64 46.21 8.56 22.29 Amhara 5.02 27.11 5.58 22.41 6.46 20.05 Oromiya 5.71 32.32 9.27 29.58 6.47 22.06 Somali 5.51 42.24 4.82 35.45 2.19 8.51 Benishangul-Gumuz 9.77 52.67 8.55 45.08 10.37 38.53 SNNPR 4.76 27.59 5.67 23.00 5.55 19.74 Gambela 15.46 61.69 22.18 75.33 13.10 36.87 Harari 10.41 44.16 9.22 38.31 8.73 33.62 Addis Ababa 9.01 59.93 7.97 53.23 9.61 46.41 Dire Dawa 10.73 55.33 11.38 55.08 8.97 39.48 City Mekele 22.46 66.11 17.26 67.06 15.87 49.19 Adigrat 25.13 82.94 21.26 68.89 12.67 42.86 Gonder 39.73 73.97 22.40 66.54 11.99 52.58 Dessie - - 18.15 67.62 14.32 49.61 Bahir Dar - - 23.09 69.83 26.17 69.69 Debre Birhan - - 26.16 74.18 17.79 53.61 Adama - - 16.89 71.72 21.66 69.57 Bishoftu - - 13.52 59.39 20.01 58.25 Jimma - - 14.62 57.38 18.55 60.41 Nekemte - - 15.42 61.31 26.04 73.82 Shashemene 26.72 72.50 15.87 68.93 22.12 62.76 Assela - - 25.07 70.99 22.12 69.98 6Note however that the scale of internal migration will be underestimated as the LFS only picks up a change in zone of residence. Movements within zones, e.g. from one woreda to another, will not be considered as internal migration. Jijiga 17.26 68.69 13.32 55.95 10.82 37.72 Asosa 38.69 92.46 26.47 80.82 24.88 74.57 Awassa 31.43 78.80 25.66 75.81 22.75 71.63 Sodo 27.78 88.89 30.73 66.96 16.93 54.77 Arba Minch 25.55 82.11 26.23 77.95 19.03 64.73 Gambela 19.05 79.22 22.18 75.33 14.13 54.57 Harar 16.14 63.83 13.63 56.67 12.54 48.41 Addis Ababa 9.00 60.08 7.92 53.33 9.61 46.41 Dire Dawa 14.01 70.29 13.81 68.29 10.63 49.71 Notes. Based on LFS data. Recent migrants are individuals who moved less than five years prior to survey data collection. Based on the population aged 15 and over. World Bank Staff calculations. Looking at city-level, migrants account for a larger share of the population in smaller cities. In 2013, recent migrants accounted for 26 percent of total population in Bahir Dar and Nekemte, 25 percent in Asosa, and 22 percent in Shahemene, Adama, Assela and Awassa (Table 1). As a share of total population, Addis Ababa, Dire Dawa, and Jijiga have the lowest share of migrants. This means that smaller cities and towns are attracting relatively more migrants (relative to the size of the town or city). Looking at absolute numbers rather than shares paints a different picture, showing the preponderance of Addis as a destination city: Of all migrants moving to cities between 2008 and 2013, 39 percent moved to Addis. Adama follows as a distant second with 7.4 percent (Figure 1). Figure 1. Where do rural-urban migrants go (share of migrants going to a certain city)? 45 39.1 40 35 30 25 20 15 10 7.4 6.8 5.8 5.2 3.7 3.4 3.3 3.1 3 5 0 Note: The figure only shows recent migrants (individuals who moved less than five years prior to survey data collection). Based on the population aged 15 and over. Only the 10 main destination cities are shown. Source: LFS, 2013. Internal migration in Ethiopia has increasingly been directed towards urban areas. Of all internal population movements between 2008 and 2013, 34 percent went from rural to urban areas while 25 percent was within-urban migration (from one city to another - Figure 1).7 Intra-rural migration accounted for 23 percent of population movements between 2008 and 2014. This is in sharp contrast to earlier times, where the bulk of internal migration happened within rural areas. Between 2000 and 2005, for instance, 7 As both internal migration and population have increased since the early 2000s, a higher absolute number of people now come to urban areas than before. close to 40 percent of migration was within-rural, while only one in four migrants came to urban areas (Figure 2). Figure 2. Share of internal migration, by type and time period 45 40 35 30 Rural to rural 25 Rural to urban 20 Urban to rural 15 Urban to urban 10 5 0 1999 2005 2013 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: LFS, 1999; 2005; 2013. Urban areas and their fringes, and areas with large swathes of land available for commercial agriculture were the main destinations for internal migrants between 2008 and 2013. Within each of the four most populated regions, migrants tended to move to cities and their surrounding zones: Mekelle in Tigray, Bahir Dar in Amhara, Adama, Jimma and North Shoa in Oromia, and Hawassa and Derashe in SNNPR (Figure 3). The fringes of Addis Ababa in Oromia have also attracted many migrants because of proximity to the capital (Dukem, Burayu, Sebeta). Gambella (mainly zone 3 and Godere), Benishangul-Gumuz (Kamashi), and Afar (zone 1) also attracted substantial numbers of migrants. As usual in fast-growing developing countries, there has been a dual migration pattern, marked by a move towards density (to cities and towns) and a parallel move away from it (towards places with low population densities and land availability). Oromia and Amhara Regions are the main origin Regions of internal migrants in Ethiopia. On the zonal level, North Gondar zone (Amhara Region) and Arsi (Oromia) are the main origin areas of internal migration in Ethiopia (Figure 4). Other important origin areas are East and West Gojam, South Wollo, South Gondar, North Shoa, and the Addis Ababa city administration. Taken together, these origin zones (and Addis) accounted for over one-third of all internal migrants in 2013. Figure 3. Where do internal migrants go? (Internal migrants as a share of zone population) Source: LFS, 2013. Figure 4. And where do internal migrants come from? (share of migrants by origin zone) Source: LFS, 2013. World Bank Staff calculations The bulk of migration in Ethiopia happens within the boundaries of the regional states. For instance, of all migrants originating from Tigray, close to 70 percent went to another zone in Tigray (Table 2). This increases to over 70 percent for migrants originating from Amhara and SNNPR and over 80 percent for migrants originating in Oromia. Migration to and from city administrations is more diverse, with most migrants in Addis coming from Amhara (38 percent) and Oromia (31 percent), and most migrants in Dire Dawa coming from Oromia (44 percent) and Somali (14 percent). Cities also mainly attract migrants from the same region. Of all recent migrants in Mekelle, over 80 percent came from other zones in Tigray region (Annex Table 1). The same is true for Bahir Dar (mainly from West Gojam and South Gondar), Hawassa (mainly from Sidama and Welayita), and Adama (mainly from Arsi and East Shoa). Addis Ababa is the exception, drawing in migrants from all over the country, except from the emerging regions (Figure 5). Migrants in Addis mainly hail from Gurage, South Wollo, North Shoa and West Shoa. Figure 5. Where do migrants in Addis come from? (Origin zones of migrants in Addis Ababa) Source: LFS, 2013; World Bank Staff calculations Table 2. Destination region of internal migrants, by region of origin Origin Destination region (% of migrants) region Tigray Afar Amhara Oromiya Somali Benishan SNNP Gambela Harari Addis Dire Ababa Dawa Tigray 68.26 2.98 13.49 2.28 0.08 0.73 5.32 0.20 0.13 6.40 0.14 Afar 4.29 51.52 15.34 13.50 0.19 0.09 11.66 0.04 0.05 3.21 0.11 Amhara 1.87 2.15 71.32 8.25 0.22 1.81 3.27 0.29 0.17 10.53 0.13 Oromiya 0.20 0.28 3.07 81.46 0.42 1.01 4.13 0.47 0.54 7.40 1.01 Somali 0.72 0.00 0.00 46.49 33.54 0.34 1.65 0.07 1.71 5.44 10.04 Benishan 0.00 0.28 23.64 9.08 0.00 63.50 0.31 1.01 0.31 1.87 0.00 SNNP 0.09 1.30 1.20 12.46 0.35 0.08 72.53 1.53 0.29 9.59 0.58 Gambela 0.00 0.00 1.51 14.25 0.00 0.00 31.78 50.44 0.00 1.71 0.31 Harari 0.00 0.62 34.00 32.52 6.68 0.81 0.00 0.00 3.02 10.66 11.69 Addis Ababa 3.61 0.68 21.11 51.08 0.33 0.54 20.58 0.35 0.53 0.00 1.19 Dire Dawa 0.00 0.96 9.02 28.98 11.48 1.23 11.99 0.16 9.89 9.16 17.13 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: LFS, 2013; World Bank Staff calculations All Regions and zones in Ethiopia have been both the origin and recipient of migrants in the years preceding the 2013 LFS, if to varying extent. To see whether Regions or zones had an average net inflow or outflow of people, net migration rates were calculated based on the 2013 LFS. The net migration rate is calculated as the difference between the number of people entering (immigrants) and leaving (emigrants) a Region (or zone) per 1,000 individuals in the given period. A positive number indicates more immigration than emigration in the Region (zone) in the given period (in this case, 2008-2013), whereas a negative number indicates more people leaving the Region (zone) than coming into it. At the Regional level, Dire Dawa and Gambella had the highest net migration rates between 2008 and 2013. This means that, relative to the size of their population, they experienced the largest net population inflows (Table 3). Addis Ababa, Benishangul-Gumuz, and, to a lesser extent, Afar also experienced net population inflows. Net population outflows were highest in Tigray, Somali and Amhara. Net migration rates at the zonal level provide a more detailed picture (Figure 6). For instance, while Tigray as a Region experienced net population outflows, Mekelle experienced net inflows. In general, net migration rates between 2008 and 2013 were highest for secondary cities: Bahir Dar, Adama, Jimma, and Hawassa all had net migration rates above 100, meaning that for every 1,000 residents, over 100 new residents came to the city (Figure 6). While Addis Ababa attracted many immigrants, many emigrants also left the city, leading to a modest positive migration rate of 33. The fringes of Addis however experienced high net in- migration. Western Tigray, zone 4 of Afar, and West Wellega had the highest negative migration rates (more people leaving than coming). Table 3. Net migration rates by Region, 2008-2013 Recent immigrants Recent emigrants Difference Net migration (immigrants - rates emigrants) Rural areas 1,210,163 1,635,304 -425,141 -12.20 Urban areas 1,675,224 1,250,083 425,141 43.19 Regions Tigray 185,797 233,891 -48,094 -17.09 Afar 60,782 48,662 12,120 16.92 Amhara 721,028 839,310 -118,282 -10.54 Oromiya 1,055,837 985,109 70,728 4.30 Somali 21,313 32,236 -10,923 -11.21 Benishangul-Gumuz 51,914 36,840 15,074 29.80 Snnpr 497,238 511,316 -14,078 -1.56 Gambela 25,636 17,017 8,619 43.64 Harari 11,526 10,323 1,203 9.06 Addis Ababa 231,624 159,121 72,503 29.80 Dire Dawa 22,693 10,790 11,903 46.66 Notes. Based on LFS 2013 data. Recent migrants are individuals who moved less than five years prior to survey data collection. Net migration rates are calculated as the difference of immigrants and emigrants of an area over a period of time, divided (usually) per 1,000 inhabitants. Based on the population aged 15 and over. Figure 6. Net migration rates by zone, 2008-2013 Source: LFS, 2013; World Bank Staff calculations 3. Who migrates where and why? People migrate in response to both “push� and “pull� factors. Push factors refer to conditions that push people out of their original place of residence, such as land scarcity, poverty, or lack of public services in rural areas, or the high cost of living in urban areas. Pull factors refer to the availability of better opportunities elsewhere, mainly related to employment. This section examines the characteristics of migrants and the main push and pull factors for the different migration patterns (rural-to-urban, urban- to-rural, rural-to-rural, and urban-to-urban), based both on self-reported motivations and on analysis of the 2013 LFS. Age and education are main correlates of internal migration in Ethiopia. Regardless of whether the origin area is rural or urban, migrants are younger and better educated compared to non-migrants from the same origin area (Table 4). Rural dwellers who migrate, either to other rural or to urban areas, accumulated twice as many years of education than rural dwellers who stayed put, and were three times more likely to have enjoyed schooling at the secondary level (Table 4). A similar pattern is found for urban origin areas, where migrants (those who move to another urban areas) are younger and more educated to urban non- emigrants. In general, education drives migration: Rural dwellers who had at least some secondary education (not necessarily completed) had a likelihood of close to 30 percent to migrate to an urban area between 2008 and 2013, and this rose to close to 60 percent for people with post-secondary education (Figure 7)8. In general, the probability of each type of migration increased with education, with the exception of urban-rural migration, which was most common for individuals without any formal education. Table 4. Characteristics of emigrants and non-emigrants, by type of origin area Migrant characteristics Rural origin areas Urban origin areas Emigrants Non-emigrants Emigrants Non-emigrants Age (mean) 26.06 34.62 27.25 34.62 Gender (1 = male) 0.43 0.49 0.47 0.46 Marital status (1 = married) 0.46 0.63 0.46 0.48 Literacy (1 = yes) 0.59 0.37 0.80 0.75 Years of schooling (mean) 4.56 2.15 7.51 6.72 No schooling (1 = yes) 0.37 0.59 0.17 0.23 Primary school (1 = yes) 0.45 0.37 0.43 0.41 Secondary school (1 = yes) 0.17 0.05 0.34 0.30 Higher education (1 = yes) 0.01 0.00 0.06 0.06 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: LFS, 2013; World Bank Staff calculations Figure 7: Educational attainment and propensity to migrate, by migration type 0.60 0.50 0.40 No education 0.30 Primary education Secondary Post-secondary 0.20 0.10 0.00 Rural to rural Rural to urban Urban to rural Urban to urban 8Education also drives migration aspirations: Recent 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). Note: The estimated propensities are based on a multinomial logit model that models the decision to migrate based on demographic characteristics of the individual and the characteristics of his/her zone of origin. Recent migrants are individuals who moved less than five years prior to survey data collection. Based on the population aged 15 and over. Source: LFS, 2013; World Bank Staff calculations People mainly migrate to look for work. In the five years up to 2013, 36 percent of migrants mentioned search for work as the main motivation to move. Moving for work has become increasingly frequent over time, especially after 2005, which partly reflects the expansion in education in recent decades (educated people are more likely to move). Other important motivations to migrate are to live with relatives (especially for young people) and marriage arrangements (for young women). Disaggregating between type of migration, the search for work is the main motivation of every type of migration (Annex Table). Shortage of land is also an important motivation for rural-to-rural migration, especially for men. Disaggregating by gender, looking for work is by far the main migration motivation for men. For women too, searching for work is an important motivation, together with moving for marriage (Annex Table ). Table 6. Migration motivations of recent migrants by survey year Migration reasons Survey year of recent migrants 1999 2005 2013 Search for work 21.80 18.50 36.02 Education 9.08 15.92 9.35 Marriage arrangement 18.49 12.26 12.08 Marriage dissolution 4.13 2.52 1.93 Job transfer 7.20 7.97 7.66 Displacement 3.14 4.49 1.17 Health problems 0.77 2.22 2.35 Shortage of land 2.48 13.80 2.41 Return back to home 12.47 6.82 2.98 To live with relatives 7.23 3.67 16.12 Moving along with family 10.63 10.04 - Lost family 0.24 - - Displacement or work - 0.72 - Dismissed from work - - 0.76 Family/peer/ pressure - - 1.25 Other reasons 2.35 1.06 5.92 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: LFS, 1999; 2005; 2013; World Bank Staff calculations To examine the correlates of internal migration in more detail, we estimate a multinomial logistical regression. In the specification, migration is modeled as a function of the migrant’s demographic characteristics (age and sex) and educational attainment, as well as characteristics of the migrant’s origin zone that proxy potential push or pull factors (zonal-level poverty rates, population density as a proxy for land holdings, and zonal infrastructure). Origin region dummies are included to capture unobserved effects9. In a first step, we focus on migrant out of rural areas, regardless of whether the destination is another rural area or un urban area. The base category in the analysis are rural dwellers who did not move zones in the five years preceding LFS 2013. The second category are people who migrated from one rural area to another, while the third category consists of people who migrated from rural to urban areas. The regression analysis is presented in Annex 5. At the individual level, demographics and education drive migration. Education is the main correlate of migration, with better educated people more likely to move out of rural areas. While better educated people are also more likely to move within rural areas, education is particularly important for rural-urban migration. All else being equal, rural dwellers with secondary education (completed or not) were 26 percentage point more likely to migrate to urban areas relative to non-educated rural dwellers. Given the low education levels in rural Ethiopia in general, the bulk of rural-urban migrants are merely educated at the primary level. Demographic characteristics also influence migration: Younger people, and especially women, are more likely to migrate from rural to urban areas than older people and men10. There is no age- effect for within-rural migration, though the gender effects persist given that women are likely to move to another rural zone for marriage. Population density, access to road infrastructure, and poverty all influence the propensity to migrate out of the zone. All else equal, people living in places with a high population density were more likely to migrate out of their rural area towards an urban area. While this is consistent with a potential role of agricultural land shortage, the density variable is defined at too coarse a level (the zone) to confidently draw conclusions about this11. Zonal level access to road infrastructure and zonal poverty rates (as per the 2010 HCES and WMS) have the expected effects on migration: The higher the distance to an all-weather road (i.e. the more inaccessible the zone) and the higher the poverty rate, the less likely people from that zone migrate to urban areas. Absence of road infrastructure and poverty increase the costs of migration, and also decrease the availability of information on opportunities that may be available elsewhere12. Rural to rural migration is also linked to population density in the origin zone, but in a non-linear way: At low levels of density, an increase in density decreases the propensity to migrate out. At high density levels however, a further increase is linked to increased migration probabilities. At the Regional level, people from the four Developing Regional States are least likely to migrate, whether to urban or to other rural areas. Migration out of urban areas is structurally different from migration out of rural areas. People who move from urban to rural areas tend to be older and little educated, and are more likely to move to live with relatives. A fair share (46 percent) of urban to rural migrants are employed in agriculture, suggesting that older and less educated people leave urban areas to go farming in rural areas (Annex 6). Looking at the destination zones of urban to rural migrants (Figure 8), we observe that they overlap with the origin zones for rural-to-urban migration, hinting that a share of urban to rural migration is in fact return migration to the place of origin13. Urban to urban migration is largely the prerogative of the university-educated wage employees and consists for a good part of professional mutations within the public and private sector. 9 For non-migrants, the origin region is their region of residence at the time of the survey. 10 The qualitative research found that women are more likely to find employment in urban areas, which may be partly driving this finding. 11 The 2013 LFS only has information on the origin zone of migrants, not on the origin woreda or kebele. 12 Note that this relationship is at the zonal level, not the individual level. Whether within a given zone the poor or the somewhat better-off are more likely to migrate is something that cannot be examined with the LFS data. 13 Figure 4 showed that many rural to urban migrants come from Arsi, Wollo, Gojam, North Shoa, Wolayta and Gamo Gofa. These zones are also main destinations for urban to rural migrants, consistent with a pattern of return migration. Figure 8. Where do migrants who leave urban areas go? (Destination zones of urban-rural migrants) Source: LFS, 2013; World Bank Staff calculations To summarize, the four migration patterns in Ethiopia are driven by distinct circumstances and characteristics. Intra-rural migration appears to be mainly driven by demographic factors and life-cycle effects and increasing pressure on land in some areas of the country, while rural-to-urban migration is driven by skills and linked to higher returns to education in urban areas. Urban-to-rural migration is correlated with age and the lack of education, with older and little educated people going to rural areas to farm. In many cases, urban to rural migration appears to be return migration. Urban to urban migration is the prerogative of well-educated public and private-sector wage employees who move for work. One common element in all migration patterns is the role of education: Education is the most robust correlate of migration, positively influencing rural-to-rural, rural-to-urban, and urban-to-urban migration and negatively influencing urban-to-rural migration. Younger and more educated people are more likely to migrate to urban areas. Another question is whether people who migrate to the capital have different characteristics from those who migrate to secondary cities and other urban centers. Using data from the 2013 LFS, we find that people who migrated to the capital were on average younger, much more likely to be female, and less educated compared to those migrating to other cities and urban areas (Table 7). The lower education outcomes of people who migrated to Addis does not seems to be due to their being younger, as the share of migrants still in school does not differ between Addis and other urban. Youth who migrated to Addis are more likely to be wage employed, especially in domestic work (young women). Migrants in other urban areas are split between wage- and self-employment, and are mainly employed in commerce, agriculture, and personal services (domestic work). While education drives rural-urban migration in general, it seems that the relatively better educated migrants move to secondary cities and other urban areas, while relatively less educated young women move to Addis for domestic work. Table 7. Characteristics of rural migrants by destination (Addis vs. other urban) Addis Other urban Demographic and education characteristics Age 22.3 26.1 Sex (% male) 30.8 43.6 No education (%) 22.6 24.4 Primary (%) 54.3 42.7 Secondary (%) 19.3 22.8 Post-secondary (%) 3.8 10.2 In school (%) 11.2 12.3 Labor market status Inactive (%) 20.3 21.7 Unemployed (%) 16.3 12.6 Public sector 4.8 16.5 Private sector (%) 95.2 83.5 Paid employee 85.6 46.7 Non-paid employee 0.6 14.8 Employer 0.2 0.7 Self-employed 13.1 36.9 Other 0.3 1 Industry of employment (%) Agriculture 1.3 16.7 Mining 1.4 0.8 Manufacturing 10.6 13.2 Publuc utilities 0 0.7 Construction 12.9 9.1 Commerce 6.6 19.7 Transport & communication 9 9 Financial & business-oriented services 1.5 3.4 Public administration/education/health 6.3 10.2 Community and family-oriented services 50 16.9 Other services 0.4 0.2 Poverty rate origin zone 24 28.6 Population density origin zone 380 438 Distance to road (minutes) 79.5 88.4 N 721 3731 4. A qualitative study on rural migrants in urban areas The survey-based analysis presented in the previous sections provides general and representative information on migration flows, migrant characteristics, and correlates of mobility. To get richer information on the specific motivations and experiences of migration and the challenges internal migrants face, a qualitative research study on young rural migrants in urban areas was conducted in April and May 2017. The focus was on young people who move from rural to urban areas in search of work or better opportunities in general. The aim of the study was to get a better understanding of migrants’ motivations, the patterns of migration and how migration is financed, the barriers and obstacles to migration and migration experiences. Formally, the research was structured along six themes: 1. Migration decision and motivation 2. Financing of migration 3. Stepwise migration 4. Job search 5. Migration experiences 6. Barriers to migration To shed light on the above questions, a qualitative research study was undertaken in Addis Ababa, the four regional cities of Adama, Bahir Dar, Hawassa and Mekelle, and eight selected towns in the four most populated regions (Oromiya, Amhara, SNNPR and Tigray). In each of the towns/cities, a series of focus group discussions (FGDs) were undertaken with young rural migrants, separately for young men and young women (“young� defined as being between 18 and 29 years-of-age). The FGDs included both young migrants that were already working and those who were still seeking employment. All youth had been in the town/city for less than four years. Overall, 28 FGDs were conducted with a total of 224 youth participants. In addition, key informant interviews (KIIs) were conducted with regional or city authorities in Addis Ababa and in the four regional cities. KIIs were conducted with the local bureaus of the Ministry of Labor and Social Affairs (BOLSA), the MSE Bureaus, and the Youth Bureaus or Associations. Annex Table 1 summarizes the sample, while Annex Figure 1 shows the geographic distribution of selected cities and towns. Though the qualitative study was not designed to be representative for the migrant population, some of the basic socio-economic information on the participants square remarkably well with the findings from the LFS analysis. First, the rural migrants who were included in the qualitative study were on average schooled at the “incomplete primary� level, corresponding to the average years of schooling of rural emigrants found in the 2013 LFS (Table – an average years-of-schooling of 4.6 or incomplete primary). Second, rural migrants, through still poorly educated for the most part, are nevertheless better educated that their age-peers who stay in rural areas: In our sample of 18-29-year-old rural migrants, 11 percent never went to school, compared to the national average of 31 percent of rural 18-29-year-olds who never went to school14. Third, the qualitative study also showed that migration mainly happens within regions, with the exception of Addis Ababa: Three-quarters of rural migrants in the study came from a rural area in the same region. 14 Based on the 2016 Ethiopia Demographic and Health Survey (EDHS). 4.1 Why do rural youth migrate? Reflecting the findings of the quantitative analysis, focus group participants said they migrated due to the lack of job opportunities in their hometown. Most viewed their migration as a chance to improve their financial wellbeing. “I came to Hawassa to work,� said a male participant, “since there is no work back home and I grew up without a father. I’m here to improve my economic situation.� Most described the rural areas where they grew up as bereft of job prospects and hope. “Everyone at home is poor,� said a male participant. “Frustrated by this, youth spend their time drinking alcohol and chewing chat. I migrated to avoid these problems.� A relatively well-educated female participant from Tigray agreed that migration is often the logical last resort. “When you complete 10th grade and you don’t have work and no one to help you, your final decision will be migration.� For many, moving to improve their life is intrinsically tied to supporting their large, resource-poor families. “I don’t have a father, so we lived in poverty and I dropped out of school,� said a female participant from Tigray. “When my younger brothers were also going to drop out of school, I decided to migrate and work to help them.� Another said: “My family’s big problem is that there are so many children in the house, and we do not have enough means to generate income. Because of this, I was forced to quit school when I was a child. I migrated to Mekelle because of my family’s poverty.� Some participants cited an abrupt family trauma that served as the catalyst for their decision to migrate. This included the sudden death of a parent or caretaker: (“After my mother’s death, my father became an alcoholic; he drank a lot and disturbed the family. Because of this I decided to move to Bahirdar�); or the materialization of an unwanted marriage: (“I worked with my family at the grain mill but they brought for me an old man to force me to marry him�). Others said they would not have migrated if they had access to fertile land to farm. Issues such as declining land productivity, a toilsome landscape, and diminished land sizes compelled those who would have otherwise preferred to farm to leave their family home. This finding squares with recent quantitative research on the role of land shortage in driving migration15. However, some disagreed that having access to land would have prevented their migration. A female participant noted the lack of opportunity to trade farmed goods in the rural area where she is from, while a male participant noted that he typically migrates during the off-farming season. A few said they migrated after hearing the success stories of others. “We have friends we went to primary school with, who after finishing their second semester, moved to Addis Ababa,� said a male participant. “When they came back to our village, our friends looked very attractive and were well dressed. That motivated us to migrate.� Some participants said they migrated due to the lack of educational facilities in their hometown. Youth said they had to travel a relatively long distance to attend school, especially at the secondary level, and their families could not afford to support them. They believed that migrating to a larger town and getting a job would allow them to support their education. 16 15 See, for instance, and Koseck et al. (2017). 16Female participants in Wukro said they came to Wukro for secondary education, as there are no high schools in their villages. But when they finished high school and failed to join universities, they found it difficult to return to their home village. Both their 4.2 Transition to urban life and job search Migrants mostly select their destination based on the city or town’s perceived job opportunities and the presence of a network of relatives, friends, or acquaintances. Upon their arrival in an urban area, rural migrants typically stayed with family members (e.g., siblings, relatives); agents; friends; or on their own. Hosts often covered migrants’ expenses until their first job. In some cases, females provided domestic services to their host without charge. Those who knew no one in the city before they arrived relied on a combination of their wits and luck. A male FGD participant said: “I wandered on the road looking for a job. Any kind of job. I also studied what kinds of jobs were easy to access, and how to communicate with people. During this process, I usually went to church and begged for food.� Migrants typically found the transition into city life difficult. One shared how the noise and movement of a bajaj in city traffic unnerved her. Youth described initially being unable to fit into social environments common to urban areas. Language is a major issue for those who travelled across regions, and those who didn’t often found themselves mocked for their rural accents and way of dress. Their conspicuousness made them easy targets for conmen and thieves, and many experienced theft and harassment. For females, the transition is especially tough. Upon their arrival, girls often face abuse by the agents they rely on for jobs, as the men frequently expect sexual favors in exchange for work placement. Those who do not have a safe space to sleep, end up spending the night in chat houses, where they are vulnerable to the whims of customers. According to discussants, chances are that upon arrival to her destination, a girl migrant will be coerced into having sex, and end up with an unwanted pregnancy or a venereal disease. Once employed, sexual harassment at the work place is also common17. Few migrants arrive at their destination with a pre-arranged job; most look for work only after they arrive. A male participant in Adama said: “When you leave your home, no one promises you a job. Instead, you gamble that you will get a job. So, I for instance searched for jobs on my own once I arrived.� Participants seemed divided as to whether they found the job search process difficult. Some transitioned into their new jobs easily, through the services of an agent or a relative, while others could not find employment for a while. A female participant from Addis Ababa said: “I paid the necessary commission to the broker and he got me a job. I didn’t experience any major ups and downs.� A male participant from Enticho said: “When I came here, I suffered to get work. I finally found work through a contact. I didn’t find work for three months. It is very difficult to find work, and in order to survive, we do everything.� Nepotism and bribery were raised as a frequent challenge in migrants’ job search. “Everything here is done through relatives,� said a participant from Enticho. “Nepotism is our biggest problem. We hear of vacancies in different offices but those who have relatives in that office will get the job. When you report this to the concerned body, they will say that they will correct it, but they will do nothing, because family and local kebele officials encourage them to remain in Wukro, while town officials want them to return. The position of town officials is reflected in the opinions of various key informant interviewees selected from the sector organizations of all four regions, such as labor affairs experts, small and micro enterprise officers, and youth and sports officers. These officials believe that rural youth should not leave their villages as they are the primary productive labor force in those areas, and government efforts should focus on creating rural jobs so that youth remain in their villages. 17A separate qualitative research study on internal migration conducted in 2014 in Amhara, Bahir Dar, and Addis also found that sexual harassment and abuse are common in domestic work in Ethiopia (Atnafu, Oucho, and Zeitlyn, 2014). they will be given a bribe.� Similarly, many FGD participants found the requirement to have a reference when applying for jobs to be problematic. Another participant noted that bureaucracy is a problem. “When we ask the woreda administration about work, they give us an appointment for the day after tomorrow.“ Another female participant from Enticho noted a lack of transparency: “They post vacancy announcements but we do not know who gets the job. We do not know the criteria.� A third female participant from Enticho noted the imbalance between the number of available jobs and job seekers. When jobs are found, they tend to be casual in nature. Examples include construction work (as a day laborer18), shoe shining, petty trade (selling gum, tissues, lemons, candy, firewood, tea, coffee, food, qolo, fruit, and coal), domestic work, ironing clothes, and working as a porter, security guard, barber, waiter, janitor, taxi driver, storekeeper, or, in some cases, sex work. Some participants mentioned taking up weeding on large commercial farms. Challenges associated with work include long hours and underpayment for hard physical labor. 19 Said one migrant: “We are paid very little yet we sacrifice our sleeping time to work.�20 Those who trade goods on the street are frequently harassed by security guards, police, and store owners, who resent migrants as they do not pay tax on the items they sell. A female FGD participant from Welkite said: “Working on the street is very challenging. People insult you, they force you to abandon your business, and consider you to be a beggar.� While some found work soon after arrival, others had to wait a long time before gaining meaningful employment. Migrants described the social isolation related to being jobless in a new city. One migrant who initially had a job, soon felt abandoned by his friends and relatives when that job expired. Most speak of suffering because of the unavailability of affordable food and harsh living conditions that they endured as a result. Some relied on meals scavenged from hotels and restaurants. One migrant described the lack of empathy the locals he encountered had, as they kept asking why someone of his physical capacity refused to work and insisted on eating leftovers. Migrants are pragmatic in their approach to taking employment. A female FGD participant from Addis Ababa said: “We do every possible job. We don’t stick to a particular job. I for instance have worked on a construction site, I’ve worked as a shoeshiner, and I’ve done different types of street vending.� A female FGD participant from Bahir Dar said: “I use a scale to weigh passersby on the street. I could not get a good job since my education level does not allow for it.� 4.3 Migration experiences Overall, and despite the many challenges, most migrants had a positive view of their decision to 18 Among FGD participants, working as a day laborer on a construction site was the most prominent occupation. 19 During a key informant interview, a small and micro enterprise officer noted that the overwhelming number of rural migrants to urban areas drives construction wages down, and youth “are not paid as equal as their labor.� (more examples needed) 20 Another participant shared the following anecdote: “I got a job through a broker, and was employed in a hotel. But when he took me there he didn’t tell me what I’d be doing and for how many hours it would be. I used to work long hours in the hotel during the day, and at night, I was forced to work as a security guard. It was also my duty to monitor chairs and tables, and to arrange them both day and night. After the third day, I was unable to shoulder my responsibilities, and I informed them that I want to leave the job. But the owner said he has already paid the broker and that I have to work more days until the payment is compensated.� migrate21. Positive views were mainly expressed in relation to three dimensions: Material, educational, and social. On the material side, migrants expressed that migration allowed them to earn a better monthly salary than they would have had they remained in their hometown. With this improved salary, they are now able to afford more material goods and conveniences, save, and better support their families back home. The ability to support their families in rural areas was a big source of satisfaction for migrants: “Oh...it has an advantage. I am buying clothes for my sister, mother, brother by working here. My life is also being changed and my family’s too. If I am sick and my family members are sick, I can afford to take them to a health center�, said one young women in Welkite. Similarly, a young woman in Hawassa who migrated because of his family’s poverty said: “It has an advantage since I am supporting my families to buy sheep, oxen and goats. This is because my families don’t have enough land to produce more and to be self-reliant�22. Obtaining more education and skills was another frequently mentioned positive effect of migration. Many migrants say it enabled them to attend school, and the gain a certain maturity and worldliness that staying in their home village would not have allowed. “I am happy about coming to this place. Because living in Merawi town enabled me to learn many things. It helped us to become civilized. If I was in a rural village, I could not attend to my education�, said a male participant. Female migrants stressed that migration allowed them to escape the traditional gender roles that prevail in rural areas and in some cases, early and arranged marriages. A young woman from Hawassa explained: If I were in my village I may not even speak to people and just live by myself. The only option living in the rural area is to serve my family or to marry a man and give birth to a child; then the burden of life is very high and you will die with nothing�. Referring to the dominant form of employment for women in rural areas (firewood fetching), another young woman from Welkite said: “Yes it has an advantage. Previously I was selling wood traveling a long distance even in the night and the day and living with wildlife. If you work in town even as house maid it is an advantage�. Those that regretted their decision to migrate cited long working hours, hard physical labor, and not enough pay. These migrants have not been able to accumulate money to save or use past day to day expenses, mostly due to high rents for housing. Yet they also feel that they cannot return to their home villages due to their unfulfilled goals. Some mentioned that they had underestimated how hard one must work in the city compared to life in the rural area, while others noted that the application of the same amount of effort in their hometown may have yielded significant results. Some migrants said they struggled especially because of language barriers, while others said their isolation from family and loved ones resulted in frequent feelings of depression and loneliness. Said one: “Yes you may dress well and look good. But the real problem lies in our heart. You are always alone, you don’t get the things you were getting in your home and especially during holidays, you will get depressed.� Migrants also noted that their lack of a meaningful social network in their new location exposes them to vulnerabilities. Though migration experiences were largely positive for both boys and girls, girls noted their particular vulnerability. While accessing jobs is relatively easier for girls, finding a safe place to stay was more challenging, and female migrants indicated that landlords frequently harassed them. Said a female participant from Asela: “Migration has affected us a lot. It is more difficult for women than men. Men can 21 Of the 73 references in the qualitative study on the evaluation of migration, 59 were positive (rated their decision to migrate as good), 8 were negative, and 6 were ambivalent. 22 The qualitative study by Atnafu, Oucho and Zeitlyn (2014) also found positive effects of migration (as expressed by migrants themselves): “Despite the hardships that migrants face in their jobs and lives in the cities, interviews with migrant domestic and construction workers revealed that they considered themselves better off in the cities� (p. 14). go out freely and protect themselves, but for women things are more difficult.� Said another: “Migration is more difficult for women. You may be harassed by men, you may come across bad people who will beat you, you may have to work bad jobs like selling chat. We cannot sleep outside of our home like men. So migration is a very risky thing for women.� 4.4 Specific challenges faced by rural migrants Rural migrants face numerous challenges in their new location. Apart from the difficulty in job search and the general state of precariousness especially in the early stages of migration and for those without a social network in the city or town, the challenges relate to their relationship with their employers and the local government; and their inability to organize into associations and obtain urban ID cards. Workplace harassment Participants discussed the various forms of abuse they have experienced from their employers, such as denial of salary, false accusation of theft, sexual harassment, untimely payment of salary, and exploitation of labor. A male participant from Hawassa said of construction work: “You will be forced to work hard, even beyond your capacity. Instead of carrying a single stone they force you to carry more than two, and if you refuse to do so they will fire you. There is no time to take a rest, no time to eat, and no time to visit the toilet.� Those who sell goods on the street also complained of facing harassment from police. A male FGD participant from Hawassa said: “We travel along roads to sell small items but the police forbid it. During a police raid, they beat us and confiscate our property. Sometimes it’s very difficult to survive...most of the time they confiscate our stock. If they didn’t, we may not have remained at this level of poverty; we may have grown our business and changed our life.� A female FGD participant from Assela said: “In my opinion, if I got the chance to work in my country, I don’t want to go elsewhere. I just want to improve my life here. But so many challenges stop you. For instance, you cannot get a working place, and even if you work on the street, police will take your items. They ask us to have a license. Look, how can you get a license when working for such a small business?� Inability to access government services and support Youth perceive that government services are not as available for migrants as they are for urban residents. Of primary significance is migrants’ inability to access an urban ID card, either due to migrants’ inability to furnish a leave letter from their home kebele, or the perceived reluctance of urban government officials to provide them with the ID card23. As a result, youth lament their inability to move freely within the area, obtain bank books, and join or form associations that would allow them to access loans.24 Local community members treat those without urban ID cards with suspicion, and migrants find it difficult to find affordable housing or buy subsidized goods like oil and sugar from kebele stores. A migrant said she was told she had to have a house and an urban ID card to access credit. 23 To solve the ID card issue, migrants would have to return to their home village to obtain a leave letter from kebele officials, or pay bribes. Most say they cannot afford to do so. In some cases, migrants returned to their home village only to be told that they’ve been away too long to qualify for it. 24 Tigray has a regional policy that stipulates that associations must be comprised of a minimum of five or more persons from the same kebele, and that association members must have no kinship relations. According to FGD participants, this adds difficulty to establishing the association, as they would prefer to collaborate with trusted relatives and friends they already have, who may be spread across kebeles. Youth identify the inaction of government bodies to support their associations as a sign of poor governance. In some places, well-established migrant trade associations, like the shoeshiners, face harassment from police. Other migrants said they frequently faced physical abuse from police officers, who sometimes held their rural IDs hostage for bribes. 4.5 Migrant’s advice to rural youth and aspiring migrants Participants differed in the advice they would give to youth or aspiring migrants in their rural origin area. While some migrants would encourage rural youth to migrate, others, even the ones who themselves reported a positive migration experience, would advise youth to stay in rural areas. A male participant from Mersa said: “Rather than leaving their village, rural youth have to try to develop their area by rearing sheep.� A female participant from Adama said: “You always think that migration is a good thing but that is not always the case. So they should think twice before they leave their home.� In contrast, a female participant from Welkite said: “I would tell her to leave the challenges of the rural area. I might even help her here until she adjusts.� Migrants who generally encouraged other people to migrate nevertheless urged aspiring migrants to be realistic: “There is no big or small job. If we start from the smaller businesses, we will be changed. What matters is working hard and saving, […] and the change will come this way�. Participants said they would advise aspiring migrants to be prepared, and have some money before they leave home. Participants also stressed the need to work hard, and to build wealth slowly. A female participant from Hawassa advised: “When you start a business it is impossible to begin from big things or big ideas; it is better to start from small businesses that are assumed to be affordable. Have short- and long-term plans. First you will start vending tea and coffee, and when your economy advances you will grow into a restaurant, and then when you advance further, you will think of other businesses.� Lastly, they encouraged aspiring migrants to persist. A male participant from Wukro said: “They may not find a job when they first come, but they should not feel bad or give up hope since any healthy person can face poverty and may become wealthy once they overcome it.� 4.6 Regional and local authorities’ view on internal migration Officials from all surveyed regions had a generally negative view of internal migration. Some cited the depletion of productive forces in rural communities, resulting in underutilized potential resources. Others cited migrants’ likelihood of remaining jobless once they reached their destination, or their probable engagement in laborious and risky work, or lack of access to decent sanitation, food, and shelter. One official spoke to migrants’ likely exposure to addictive substances such as alcohol and drugs. While some officials noted “the positive value of labor power when industries are opened in towns,� as well as the recent expansion of cities’ construction sector that allows migrants to easily acquire work as daily laborers, most do not believe these occupations will substantively improve migrants’ wellbeing. Several quotes highlight the adamantly negative view on migration: “We stand against migration. It is problematic in different ways�, said a participant from the Amhara Youth and Sports Office. This was echoed by the same Office in Tigray: “According to our organization migrations from countryside to towns are unacceptable�. The Oromia BOLSA went one step further: “As I told you earlier, this type of migration from rural areas to urban centers is illegal�. Despite the negative view on migration, regional officials understood why young rural dwellers were leaving their villages. They stressed the search for better jobs and the lack of a conducive work environment in migrants’ home villages. They noted that some migrants want to complement their work by pursuing the educational opportunities that are more readily available in urban areas. They also said that youth want to benefit from the development and modernity of urban cities, and may idolize relatives who had previously migrated successfully. Overall, the local authorities exhibited a good grasp of the different pull and push factors: “The push factor is lack of good infrastructure in their origin. Lack of educational opportunities again it can be because of problems related to food, shelter. Consequently, the rural youth migrate to city to get good livelihood. Looking for modernity especially standing from what they see from their old migrants clothing style in urban center“ (Amhara MSE Office), or, in Amhara BOLSA: “In the urban center there are certain factors that can attract the rural youth to migrate from their area. For example, they need to change in their lives. After they look at their friends and relatives those who came to the urban and have changed, the youth move to the urban center for the sake of that change. We call this pull factor�. Local offices reported not assisting rural migrants in any way, which seemed mainly related to the lack of an ID card and the fear of attracting more migrants if they would assist them. According to the Addis MSE bureau, �our organization’s focus is on youth of Addis Ababa and on those who have identification card. Therefore, we cannot address youth rural urban migrants’ issue. Amhara and Tigray MSE Offices had similar opinions: “They also lack identification card and forbidden from employment� (Amhara), and, “the main problem is most of them didn’t fulfill the first formality of the organization� [i.e. ID card] (Tigray). Conscious of the pull factors, the Amhara SME office also said: “If we give them all job, the youth will leave the rural land and concentrate at urban center and create urban problems�. Similarly, In Tigray: “Youth who are in towns if they have identity card they can be supported to work in the studied job activities otherwise if they don’t have identity card we advise them to go to their areas and work on the natural wealth. If we are not treating them in such away we are going to reinforce migration�. The bureaus and offices also stressed the very real challenge of finance and how to budget for rapidly growing urban populations. An officer from Tigray’s MSE Office explained: “There is also a shortage of budget. We are allocated a budget based on an employment survey taken by the Central Statistical Agency, but we always get unexpectedly high numbers of unemployed. When we compare the project that we’ve established, and the unemployment numbers, there is great variation because of the scarcity of budget�. Offices in other regions also raised this issue, and stressed that the financing they are allocated is not supposed to cater to new arrivals in the city. Officials unanimously agree that rural youth should stay in their villages and be supported there. Said an official from the Tigray Youth and Sports Office: “As a rule, the youth are encouraged to work in their villages. In the region, most of the time we believe we should provide job opportunities around their area of origin to assure the equitable and fair distribution of wealth.� Officials from Amhara, SNNPR, and Tigray described some of their region’s job creation initiatives, which mostly consist of organizing youth into groups and train them to engage in different activities. A key informant interview from the Tigray Labor and Social Affairs Office also said there is a loan opportunity of up to ETB 45,000 for those rural youth who form associations to work in their village. None of the regions have a policy that explicitly addresses migrants. Said an official in Oromia: “This type of migration from rural areas to urban centers is illegal and we are doing nothing related to it in that regard. We work on the legal dwellers of the towns. We have no plan for migrants because we have no jurisdiction to do so�. So far, there are no plans at the regional level to help migrants. A Tigray Youth and Sports officer said: “We don’t have any plan specifically to guess to the number of migrants who migrate from the countryside to towns. But we will see the situation...our focus is already known. We’ve been using the Youth Development Program for towns and rural area as a guide. We don’t have any additional plan other than that.� In Amhara, a BOLSA officer pointed to a regional plan to create job search centers in Dessie, Bahir Dar, and Gondar, through which users (some of whom may potentially be migrants) can seek counseling services and be linked to employers. 4.7 Migrant’s suggestions to the government FGD participants shared their views on what government could do to make their migration more successful and beneficial and less difficult. Answers converged around issues of assistance with job search and ID cards or, more generally, equal treatment: Migrants believe that the government should allow them to qualify for urban ID cards so that they can organize into associations to access finance, and a work place. A female participant from Bahirdar said: “If we are given ID cards and a work place, we can organize into teams and work like any other Bahirdar city resident.� For some, the right to work where they have relocated to is simply a matter of equity. Said a male participant from Assela: “We should have the right to be involved in the associations that residents in different cities form. Even though we are migrants, we are still citizens of this country.� A female participant said: “The government shall not separate and divide us by saying you are from rural or urban parts. People are coming from the rural parts of the region to search for jobs. We have to be granted equal rights.� To this effect, migrant youth want the police to stop harassing them. A male participant from Welkite said: “We came here to change our life and support our families but the police are breaking our shoe shining tools, throwing our bottled water. They mostly call us illegal, but my question is, why they don’t give us legality, why don’t they offer us opportunities? We are ready to even pay taxes, and to organize ourselves into micro and small-scale organizations.� Migrants believe the government should include migrants in urban job creation schemes. Said an FGD participant: “I want to say that our government should consider internal migrants in job creation activities. We want to work and improve ourselves. We are leaving our country because we are in poverty in rural areas25. We are leaving our country because we are seeking better job opportunity there. So if we are given job here, we won’t want to leave our home.� Migrant youth believe that the government’s initial support will eventually lead to their self-sufficiency. A female participant from Merawi said: “We youth migrants are living with a shortage of capital, and the government needs to support us in organizing, training, and financing for starting a job. After that, we can support ourselves rather than being a burden to the town.� 25 “Country� in this citation was likely used in a colloquial way, referring to the area the migrant came from (home place). 5. Conclusions and policy perspectives Internal migration or labor mobility in Ethiopia remains fairly limited. In the five years prior to 2013, the latest year for which data are available, slightly over six percent of the adult population changed zone of residence, marginally higher than the share in 199926. The structure of internal migration has however changed, with rural-to-urban movements becoming the dominant migration flow in the most recent period (2008-2013). Since the start of the urban renewal and construction boom in 2008, young rural dwellers have increasingly moved to urban areas in search for employment and a better life. Internal migration in Ethiopia has had positive economic effects, both for aggregate growth and individual migrants’ welfare. On the macro-level, the strong increase in non-farm employment since 2005 has been tightly linked to increased rural to urban labor mobility. As productivity levels in non-farm sectors are higher than in farming, increased internal migration has been an important driver of structural transformation and economic growth, especially between 2005 and 201327. On the micro-level, rural dwellers in Ethiopia who migrated to another rural area or to an urban area experienced large increases in living standards, relative to comparable rural dwellers who did not move28. In addition, rural out- migration does not seem to lead to losses for sending households due to a lost labor effect, as is often assumed29. Despite the positive effects of migration, barriers seem to exist for people to move to places where they would be objectively better-off. In general, migration can be limited by several factors (related to policy, credit, and information), and the evidence in Ethiopia is consistent with several of those. Migrants in the qualitative study advised aspiring migrants to save some money before they leave, to be able to finance urban job search and sustain themselves in the process. Quantitative research also found evidence of credit constraints, with households with higher agricultural productivity more likely to send out a migrant30. In addition, though internal mobility is not restricted by any explicit laws or policies, the migration experience is substantially complicated by the difficulty of obtaining urban ID cards and, related to this, of accessing certain public services, and the negative attitude of local authorities towards migrants. Official views on internal migration at the city and regional levels are negative. Many offices and bureaus consider rural to urban migration as unacceptable, or even illegal. The dominant view is that rural youth should stay in rural areas and work on the land and be supported there. This stands in sharp contrast to experiences of migrants themselves, who overwhelmingly considered their migration as a good thing- despite the severe challenges-and described the rural areas they left as bereft of any prospects and hope. In general, migration is often the most profitable investment for rural youth in low-income countries (Clemens and Ogden, 2013). As Ethiopia continues to grow and develop, internal labor mobility will further increase. The expansion of education in rural areas will further add to this, as education drives mobility, as will the investments in rural connectivity. Currently, many rural migrants face a host of difficulties partly linked to policies and 26 Factoring in rapid population growth, the absolute number of internal migrants has of course increased substantially. 27 The move from farm to nonfarm occupations explains 25 percent of growth in labor productivity between 2005 and 2013 (World Bank, 2016). This was strongly linked to migration. 28 In the Ethiopia study, internal migrants experienced large increases in non-food consumption and income, and also improved their diets (De Brauw, Mueller, and Woldehanna, 2017). 29 De Brauw, 2014. 30 De Brauw, 2014. attitudes that explicitly or implicitly disadvantage them, and partly linked to the absence of protection mechanisms. Facilitating and managing mobility through various interventions would improve outcomes for internal migrants and their families (though remittances) and support Government policy objectives in the realm of poverty reduction, rural development, and structural transformation. Potential measures could include: • Facilitating migrants’ access to urban ID cards to allow them to access social services and legally start and register a business. This would also allow rural migrants to be considered for MSE support and urban safety nets if they meet all other criteria. • Providing job search assistance, such as the provision of easily accessible information on jobs available (or where and how to seek jobs) and related support services (accommodation, training opportunities, one-stop shops, credit facilities, etc.). • Formulating an internal labor mobility strategy and policy that acknowledges labor mobility as a positive livelihood option and that aims to maximize the benefits of internal migration while mitigating the risks. This strategy would also need to include sensitization of local authorities and law enforcement to better understand the concept of mobility and migration and the reasons behind it. Given the risk of exploitation young migrant women face in urban areas, especially in domestic work and hospitality services and through contacts with brokers, special focus need to be given to make urban migration safer for women. Though facilitating internal labor mobility would likely come with macro- and micro-benefits, it would also pose challenges. It would require a careful reallocation of the budget for MSE and job-creation support from rural to urban areas, in function of the anticipated migration flows to towns and cities. Ideally, facilitating migration would also go hand in hand with investments to make small cities and towns more attractive. While migrants, also in our qualitative study, prefer to move to a town or city that is close to their origin areas and families, many end up migrating to regional capitals and Addis where economic opportunities and services are better. Investments in smaller cities and towns can avoid an over- concentration of people in large cities. In addition, research has shown that development of towns and secondary cities has a bigger poverty-reduction effect than concentration in mega-cities31. 31 Christiaensen and Todo, 2014; Christiaensen, De Weerdt and Todo, 2013. References Lall, S., Selod, H., and Shalizi, Z. (2006). Rural-urban migration in developing countries: A survey of theoretical predictions and empirical findings. Policy Research Working paper 3915. Washington DC: The World Bank. Annez, P. and Buckley, R. (2009). “Urbanization and growth: Setting the context�. In: Spence, M., Annez, P., and Bucley, R. (eds). Urbanization and Growth. Washington DC: The World Bank. World Bank (2008). World Development Report 2009: Reshaping Economic Geography. Washington DC: The World Bank. Atnafu, A., Oucho, L. and Zeitlyn, B. (2014). “Poverty, youth and rural-urban migration in Ethiopia�. Working Paper. Migrating out of Poverty RPC. Beegle, K., De Weerdt, J. and Dercon, S. (2011). “Migration and Economic Mobility in Tanzania: Evidence from A Tracking Survey�. Review of Economics and Statistics 93(3): 1010-1033. Bezu, S. and Holden, S. (2014). “Are Rural Youth in Ethiopia Abandoning Agriculture?�. World Development 64: 259-272. Bryan, G., Chowdury, S. and Mobarak, M. (2014). “Underinvestment in a Profitable Technology: The Case of Seasonal Migration in Bangladesh�. Econometrics 82(5): 1671-1748. De Brauw, A. (2015). “Migration, Youth and Agricultural Productivity in Ethiopia�. International Food Policy Research Institute, Conference Paper. De Brauw, A. and Mueller, V. (2012). “Do Limitations in Land Rights Transferability Influence Mobility Rates in Ethiopia?’. Journal of African Economies 21(4): 548-579. De Brauw, A., Mueller, V. and Woldehanna, T. (2017). “Does Internal Migration Improve Overall Well-Being in Ethiopia?�. Journal of African Economies: 1-19. Kosec, K. Ghebru, H., Holtemeyer, B., Mueller, V. and Schmidt, E. (2016). “The Effect of Land Inheritance on Youth Employment and Migration Decisions: Evidence from Rural Ethiopia�. International Food Policy Research Institute discussion paper 1594. World Bank. (2016). “Ethiopia’s Great Run. The Growth Acceleration and How to Pace It.� Washington DC: The World Bank. Christiaensen, L. and Todo, Y. (2014). “Poverty Reduction During the Rural-Urban Transformation – The Role of the Missing Middle�. World Development 63: 43-58. Christiaensen, L., De Weerdt, J. and Todo, Y. (2013). “Urbanization and Poverty Reduction: The Role of Rural Diversification and Secondary Towns�. Agricultural Economics 44(4-5): 435-447. Clemens, M. and Ogden, T. (2014). “Migration as a Strategy for Household Finance: A Research Agenda on Remittances, Payment, and Development�. Working Paper 354. Center for Global Development. Lagakos, D., Mobarak, M. and Waugh, M. (2018). “The Welfare Effects of Encouraging Rural-Urban Migration�. Working Paper 24193, National Bureau of Economic Research. De Brauw, A. and Giles, J. (2018). “Migrant Labor Markets and the Welfare of Rural Households in the Developing World: Evidence from China�. World Bank Research Observer 32(1): 1-18. Ministry of Finance and Economic Development. (2006). Plan for Accelerated and Sustained Development to End Poverty (PASDEP). Federal Democratic Republic of Ethiopia: MoFED. Annex 1: Data and definitional limitations to the study of internal migration Internal migration is measured in the LFS by asking all members of the household about the number of years the household members has continuously lived in the town or rural part of the wereda. If the answer was less than five years, follow-up questions are asked about when the person came to the town or rural part of the wereda, in which zone the person lived before leaving, and the reason for leaving the previous residence. From the LFS data it is therefore possible to detect individuals who moved at one point in their lives, hereafter referred to as all-time migrants, and recent migrants (those who moved less than five years prior to the survey data collection). Throughout the report a distinction is consequently made between all-time and recent migrants. For recent migrants we can identify previous zones of residence, type of previous area (rural/urban) and reasons for migration. These data provide important information on migration patterns and trends over time. For all-time migrants we know that they had moved at one point in their lives. The data do not include information on previous migration episodes or the place of birth for all-time migrants. Most of the statistics presented in this report are therefore based on recent migrations, i.e. those that have taken place five years prior to data collection. Although the LFS contains interesting information on internal migration in Ethiopia, there are important limitations as well. The main limitation is that the administrative unit at which migration is measured – the zone - represents a relatively large area, which makes it impossible to detect short-distance migration or migrations between weredas within zones. It is likely that many internal migrations in Ethiopia take place between wereda within zones, particularly in the case of rural-urban migrations. This means that the magnitude of internal migration will be underestimated. Second, another important limitation of the data was that the survey questions only referred to the latest migration episode of the individual. This makes it impossible to derive information on recurrent or multiple migration episodes and the exact directions of internal migrations in Ethiopia. From the survey data we only know where people were born and where they moved to in their most recent migration, but we do not have information on potential migration episodes that took place in between these two moments. Annex 2: Qualitative research study on rural-urban migration In the framework of a broader research agenda on youth, employment, and migration, a qualitative research study on young rural migrants in urban areas was conducted in April and May 2017. The focus was on young people who move from rural to urban areas in search of work or better opportunities in general. The aim of the study was to get a better understanding of migrants’ motivations, the patterns of migration and how migration is financed, the barriers and obstacles to migration and migration experiences. Formally, the research was structured along six themes: 1. Migration decision and motivation 2. Financing of migration 3. Stepwise migration 4. Job search 5. Migration experiences 6. Barriers to migration To shed light on the above questions, a qualitative research study was undertaken in Addis Ababa, the four regional cities of Adama, Bahir Dar, Hawassa and Mekelle, and eight selected towns in the four most populated regions (Oromiya, Amhara, SNNPR and Tigray). In each of the towns/cities, a series of focus group discussions (FGDs) were undertaken with young rural migrants, separately for young men and young women (“young� defined as being between 18 and 29 years-of-age). The FGDs included both young migrants that were already working and those who were still seeking employment. All youth had been in the town/city for less than four years. Overall, 28 FGDs were conducted with a total of 224 youth participants. In addition, key informant interviews (KIIs) were conducted with regional or city authorities in Addis Ababa and in the four regional cities. KIIs were conducted with the local bureaus of the Ministry of Labor and Social Affairs (BOLSA), the MSE Bureaus, and the Youth Bureaus or Associations. Annex Table 1 summarizes the sample, while Annex Figure 1 shows the geographic distribution of selected cities and towns. Annex Table 1: Cities and towns selected for the FGDs and KIIs City FGD Interview Addis Ababa 4 (2 female, 2 male) 3 (Youth bureau, BOLSA, MSE) Adama 2 (one female one male) 3 (Youth Assn, BOLSA, MSE) Bahir Dar 2 (one female one male) 3 (Youth bureau, BOLSA, MSE) Hawassa 2 (one female one male) 3 (Youth bureau, BOLSA, MSE) Mekelle 2 (one female one male) 3 (Youth Assn, BOLSA, MSE) Region Town FGD Tigray Wukro 2 FGD (one female one male) Tigray Entcho 2 FGD (one female one male) Amhara Habru/Mersa 2 FGD (one female one male) Amhara Merawi 2 FGD (one female one male) Oromia Aweday 2 FGD (one female one male) Oromia Asela 2 FGD (one female one male) SNNPR Arba Minch 2 FGD (one female one male) SNNPR Welkite 2 FGD (one female one male) Annex Figure 1: Cities and towns selected for the qualitative research FGD participants had an average age of 22 and were mostly schooled at the primary level. 11 percent of FGD participants had never been to school, while 60 percent had either completed primary school or dropped out in primary. The remaining 29 percent attained at least some secondary education or more (Annex Table 2). Male migrants were slightly better educated than female migrants. Comparing the educational profile of FGD participants (rural migrants in urban areas) with that of the overall rural population of the same age range (18-29), we see that the FGD sample is on average better schooled than the overall rural population: Based on the 2016 DHS, 31 percent of the rural population between 18 and 29 never went to school, compared to 11 percent of FGD participants. FGD participants are however lower schooled than the native urban population of the same age range. Most of the rural migrants had their origins either in Amhara or SNNPR. Annex Table 2: Characteristics of FGD participants Age 22 No education (%) 10.7 Primary (%) 59.7 More than primary (%) 29.6 Origin Tigray (%) 22.6 Origin Amhara (%) 30.9 Origin Oromia (%) 14.2 Origin SNNPR (%) 29.4 The bulk of participants were employed at the time of the study. Only 9 persons were unemployed but actively looking for a job. Street vending and shoe-shining were the most common activities, followed by construction work and casual labor. Young women were mostly active as street vendors, waitresses, and construction workers. Young men mainly engaged in shoe-shining, street vending, and casual labor. With the exception of Addis, inter-regional migration is rare. Almost all FGD participants came from a rural area within the same region (Annex Table 3). For instance, of the 46 FGD participants in urban Tigray, 45 came from a rural area in Tigray while one came from Amhara Regional State. Migrants with an origin in SNNPR appear to be the most mobile, with substantial numbers going to an urban area in Oromia and Addis Ababa (though most still stay within the SNNPR state boundary). Annex Table 3: Origin region of participants by current region Origin region Current region Tigray Amhara Oromia SNNP Others N Tigray 45 1 0 0 0 46 Amhara 0 48 0 0 0 48 Oromia 1 8 23 10 6 48 SNNP 0 2 1 37 0 40 Addis Ababa 0 4 5 13 0 22 N 46 63 29 60 6 204 Annex 3: Origin zone by city Annex Table 4. Origin zone of recent migrants in selected cities, % Other Origin Region Origin zone Addis Awassa Adama Dire Dawa Mekelle urban Tigray north western tigray 0.80 0.00 0.00 0.38 2.11 1.33 Tigray central tigray 0.94 0.00 0.00 0.00 17.52 2.26 Tigray eastern tigray 1.60 0.47 0.23 0.81 32.79 1.14 Tigray southern tigray 0.83 0.38 0.00 0.28 22.53 1.30 Tigray western tigray 0.65 0.00 0.00 0.00 4.34 1.05 Tigray mekele special 1.64 0.00 0.23 0.27 1.71 0.57 Afar zone 1 0.44 0.00 0.38 0.00 0.35 0.72 Afar zone 2 0.00 0.00 0.00 0.00 0.00 0.55 Afar zone 3 0.18 0.00 1.22 0.28 0.35 0.13 Afar zone 4 0.05 0.00 0.00 0.00 0.00 0.40 Afar zone 5 0.00 0.00 0.00 0.00 0.00 0.03 Amhara north gondar 4.87 0.29 0.75 0.77 5.30 5.38 Amhara south gondar 3.52 0.48 0.69 0.92 0.00 3.41 Amhara north wello 5.63 0.56 0.60 0.00 3.51 2.51 Amhara south wello 8.31 0.45 1.45 0.00 0.00 4.25 Amhara north shewa 7.49 1.07 4.59 1.84 0.00 3.11 Amhara east gojjam 5.26 1.16 1.02 0.89 0.00 3.26 Amhara west gojjam 1.87 0.00 0.00 0.54 0.00 3.86 Amhara waghemira 0.18 0.00 0.00 0.00 0.35 0.65 Amhara awi 0.27 0.00 0.64 0.00 0.00 1.84 Amhara oromiya 0.22 0.32 0.00 0.00 0.00 0.47 Amhara bahir dar special 0.48 0.63 0.23 0.83 0.35 0.59 Amhara argoba special wereda 0.04 0.00 0.00 0.00 0.00 0.02 Oromia west wellega 1.78 0.18 0.23 0.55 0.00 2.00 Oromia east wellega 0.77 0.00 0.85 1.47 0.00 1.95 Oromia ilu aba bora 0.19 0.00 0.00 0.00 0.00 1.44 Oromia jimma 2.31 0.00 1.77 1.30 0.70 2.65 Oromia west shewa 7.02 0.89 2.70 1.55 0.00 2.95 Oromia north shewa 6.83 0.00 1.42 2.16 0.00 1.66 Oromia east shewa 2.94 5.31 14.68 1.89 0.00 2.96 Oromia arsi 2.41 0.89 21.05 1.99 0.00 5.28 Oromia west hararge 0.36 0.00 0.55 8.54 0.00 2.34 Oromia east hararge 0.71 0.32 3.90 20.74 0.00 2.38 Oromia bale 1.16 0.66 1.17 0.28 0.00 1.76 Oromia borena 0.04 0.00 0.57 0.00 0.00 0.33 Oromia south west shewa 1.70 0.23 2.48 1.48 0.00 0.99 Oromia guji 0.24 2.77 0.23 0.56 0.00 0.42 Oromia adama special 1.14 1.04 0.23 0.54 0.00 0.29 Oromia jimma special 0.91 0.00 0.88 0.27 0.00 0.15 Oromia west arsi 0.43 1.83 3.44 0.00 0.00 1.91 Oromia kelem welega 0.20 0.00 0.91 0.26 0.00 1.49 Oromia horo gudru welega 0.10 0.00 0.23 0.00 0.00 1.43 Oromia burayu special 0.23 0.00 0.00 0.00 0.00 0.20 Somali shinile 0.00 0.00 0.00 3.79 0.00 0.03 Somali jijiga 0.63 0.00 0.51 4.08 0.00 0.86 Somali degehabur 0.00 0.00 0.00 0.27 0.00 0.02 Somali warder 0.00 0.00 0.00 0.00 0.00 0.06 Somali korahe 0.00 0.00 0.00 3.61 0.00 0.02 Somali fik 0.05 0.00 0.00 0.83 0.00 0.02 Somali gode 0.03 0.00 0.00 2.60 0.00 0.01 Somali afder 0.05 0.00 0.46 0.00 0.00 0.00 Somali liben 0.00 0.00 0.00 0.00 0.00 0.06 Benishangul metekel 0.05 0.00 0.00 0.00 0.00 0.68 Benishangul asossa 0.21 0.00 0.20 0.00 0.00 0.40 Benishangul kemashi 0.00 0.00 0.00 0.00 0.00 0.18 Benishangul pawe special wereda 0.04 0.00 0.00 0.00 0.00 0.09 Benishangul mao komo special wereda 0.00 0.00 0.00 0.00 0.00 0.08 SNNPR gurage zone 8.87 4.83 7.06 9.04 0.00 1.18 SNNPR hadiya zone 2.65 1.56 0.00 1.97 0.00 1.74 SNNPR kembata timbaro zone 0.24 4.11 0.00 0.27 0.00 0.85 SNNPR sidama zone 0.82 29.78 0.46 0.76 0.00 1.84 SNNPR gedeo zone 0.47 2.21 0.24 0.38 0.00 0.60 SNNPR wolyita zone 1.51 22.41 3.36 0.27 0.00 2.26 SNNPR south omo zone 0.00 0.43 0.00 0.28 0.79 0.33 SNNPR sheka zone 0.04 0.23 0.00 0.00 0.00 0.41 SNNPR keffa zone 0.14 0.18 0.00 0.00 0.00 1.93 SNNPR gamo gofa zone 1.15 0.99 0.20 1.30 0.00 2.54 SNNPR bench maji zone 0.04 0.00 0.00 0.00 0.00 0.97 SNNPR yem special wereda 0.06 0.00 0.00 0.00 0.00 0.08 SNNPR amaro special wereda 0.00 0.00 0.00 0.00 0.00 0.10 SNNPR burji special wereda 0.00 0.99 0.00 0.00 0.00 0.00 SNNPR konso special wereda 0.00 0.00 0.00 0.00 0.00 0.12 SNNPR derashe special wereda 0.00 0.00 0.00 0.00 0.00 0.07 SNNPR dawuro zone 0.21 0.21 0.00 0.00 0.00 0.34 SNNPR basketo special wereda 0.00 0.00 0.00 0.00 0.00 0.05 SNNPR konta special wereda 0.05 0.00 0.00 0.00 0.00 0.10 SNNPR silti zone 3.65 4.61 7.18 0.00 0.00 1.00 SNNPR alaba special wereda 0.00 0.23 0.00 0.00 0.00 0.08 SNNPR hawassa city administration 1.28 0.00 0.00 0.66 0.35 0.28 Gambella agnewak zone 0.07 0.00 0.00 0.27 0.00 0.32 Gambella nuwer zone 0.05 0.00 0.00 0.00 0.00 0.14 Gambella mezhenger zone 0.00 0.00 0.00 0.00 0.00 0.20 Gambella etang special wereda 0.00 0.00 0.00 0.00 0.00 0.03 Harari harari zone 0.47 0.00 0.38 5.45 0.00 0.32 addis ababa city Addis administration 0.00 6.27 9.23 9.74 6.94 5.90 Dire Dawa dire dawa city administration 0.43 1.02 1.39 3.07 0.00 0.33 Note: Migrants aged 15 and over. Each number in the table shows the percentage of recent migrants in the city that came from that specific origin zone. Source: LFS, 2013; World Bank staff calculations. Annex 4: Migration motivations, disaggregated by migration type and gender Annex Table 5. Migration motivations, by migration type and gender (2013) Migration reasons Rural to rural Rural to urban Urban to rural Urban to urban of recent migrants M F M F M F M F Search for work 41.28 11.14 50.77 38.96 29.59 25.24 46.64 36.12 Education 1.66 1.45 14.93 14.76 7.81 6.83 9.67 11.55 Marriage arrangement 3.22 45.12 1.40 14.46 1.72 16.69 1.40 13.47 Marriage dissolution 0.93 7.05 0.38 2.49 0.03 2.13 0.08 2.10 Job transfer 8.64 1.71 7.64 2.42 7.80 6.38 22.21 8.12 Dismissed from work 1.12 0.23 0.26 0.13 2.36 0.42 1.73 0.40 Displacement 3.47 2.62 1.11 0.86 0.24 0.54 0.07 0.53 To live with relatives 15.22 13.80 12.12 13.22 29.03 23.69 9.68 17.68 Return back to home 4.68 5.20 1.44 0.94 5.64 3.25 2.17 2.52 Shortage of land 9.83 4.71 2.14 1.42 0.57 0.93 0.11 0.06 Health problems 2.49 1.88 1.75 2.88 3.71 2.45 1.23 2.26 Family/peer/ pressure 1.90 2.10 1.28 1.78 0.29 0.95 0.67 0.57 Other reasons 5.55 2.98 4.76 5.70 11.22 10.50 4.34 4.61 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: LFS, 2013; World Bank Staff calculations Annex 5: Estimating the drivers of migration To estimate the drivers of migration, recent migrants were defined as those who changed zone of residence in the five years prior to the 2013 LFS. The LFS sample was split into two parts, depending on whether respondents lived in rural areas in the five years preceding the survey (regardless of their current zone of residence) or in urban areas. The rural sample consists of individuals who (i) did not migrate and still live in their same rural zone of residence, (ii) did migrate from one rural zone to another, and (iii) migrated from rural to urban areas in the five years preceding the LFS. The urban sample consists of individuals who (i) did not migrate and still live in their same urban area of residence, (ii) did migrate from one urban location to another, and (iii) migrated from urban to rural areas in the five years preceding the 2013 LFS. To examine the drivers of rural migration, multinomial logit regression was estimated with rural dwellers who did not migrate as base category, those who migrated to other rural areas as category 1, and those who migrated to urban areas as category 2. Results are presented in Annex Table 6. A similar estimation is performed for the urban sample (base category are urban dwellers who did not migrate Annex Table 7). All coefficients need to be interpreted relative to the base category of people who did not move (sedentary rural dwellers in Annex Table 6 and sedentary urban dwellers in Annex Table 7). Annex Table 6. Drivers of rural migration (1) (2) VARIABLES Rural to rural Rural to urban age -0.005 -0.074*** (0.018) (0.007) Age squared -0.000 0.001*** (0.000) (0.000) Male (1 if yes) -0.542*** -0.976*** (0.165) (0.072) Age*Male -0.018*** -0.012*** (0.006) (0.003) Primary education -0.159** 0.734*** (0.076) (0.038) Secondary education 0.407*** 2.204*** (0.118) (0.048) Post-secondary education 2.536*** 3.755*** (0.143) (0.088) Origin density -0.001*** 0.000 (0.000) (0.000) Origin density squared 0.000*** 0.000*** (0.000) (0.000) Time to go to nearest 0.000 -0.004*** all weather road (0.001) (0.000) Zone poverty rate -0.346 -0.365*** (0.241) (0.115) Afar -1.537*** -0.174 (0.345) (0.153) Amhara -0.263* 0.974*** (0.150) (0.070) Oromiya -0.349*** 0.390*** (0.126) (0.062) Somali -2.168*** -1.330*** (0.297) (0.175) Benishangul-Gumuz -1.225*** -0.538*** (0.238) (0.123) SNNPR -0.514*** 0.385*** (0.123) (0.063) Gambela -0.862*** -0.881*** (0.245) (0.141) Constant -1.643*** -0.321** (0.296) (0.132) Observations 61,570 61,570 Source: LFS, 2013; World Bank staff calculations. Note: “Origin density� refers to density in zone of origin and is taken from the 2007 Census. Poverty at the zonal level is calculated based on the 2010/11 HCES. Only people aged 15 or more in 2013 LFS are included in the regression. ***: statistically significant at 1%; **: statistically significant at 5%.; *: statistically significant at 10%. Annex Table 7. Drivers of urban migration (3) (4) VARIABLES Urban to rural Urban to urban Age 0.040* 0.026** (0.023) (0.013) Age squared -0.000* -0.000** (0.000) (0.000) Male (1 if yes) -0.231 -0.931*** (0.216) (0.097) Age*Male -0.031*** -0.032*** (0.008) (0.004) Primary education -1.167*** -0.338*** (0.108) (0.054) Secondary education -1.452*** -0.493*** (0.114) (0.056) Post-secondary education -0.997*** 0.428*** (0.122) (0.054) Afar -0.152 -0.247** (0.232) (0.108) Amhara -0.471*** 0.110** (0.146) (0.056) Oromiya 0.088 0.128** (0.133) (0.055) Somali -1.053*** -1.182*** (0.219) (0.106) Benishangul-Gumuz 0.470** 0.068 (0.210) (0.100) SNNPR -0.285* -0.192*** (0.146) (0.059) Gambela 0.133 -0.495*** (0.199) (0.103) Harari -17.982*** -17.980*** (0.120) (0.058) Addis Ababa -0.988*** -1.395*** (0.153) (0.066) Dire Dawa -18.216*** -18.149*** (0.121) (0.057) Constant -1.873*** -0.234 (0.301) (0.144) Observations 56,687 56,687 Source: LFS, 2013; World Bank staff calculations. Note: Only people aged 15 or more in the LFS are included in the regression. ***: statistically significant at 1%; **: statistically significant at 5%.; *: statistically significant at 10%. Annex 6: Labor market indicators of migrants, by migration type Employment characteristics Rural to rural Rural to urban Urban to rural Urban to urban Employment status Employed 87.46 68.22 88.93 68.28 Unemployed 3.28 10.28 4.87 13.69 Non-LF 9.26 21.50 6.19 18.02 Employment type Paid employee 25.65 51.75 32.56 70.60 Non-paid employee 41.03 12.94 29.70 5.02 Employer 0.18 0.59 0.16 0.57 Self-employed 32.62 33.82 37.04 23.44 Other 0.51 0.91 0.53 0.37 Sector of employment Public sector 10.70 14.92 21.47 39.10 Private sector 89.30 85.08 78.53 60.90 Industry of employment Agriculture 70.24 14.70 46.24 4.96 Mining 0.33 0.87 0.50 0.77 Manufacturing 3.84 12.83 5.31 10.34 Public utilities 0.36 0.60 0.38 1.82 Construction 2.77 9.58 7.09 8.41 Commerce 3.12 18.05 8.27 11.92 Transports & comm. 0.97 9.00 3.82 10.88 Financial & business-oriented service 0.24 3.18 1.18 10.63 Public adm./education/health 9.27 9.70 18.05 25.12 Community & family oriented services 8.86 21.27 8.97 14.73 Other services, unspecified 0.00 0.21 0.17 0.42 Contract type for paid employees Permanent contract 43.83 37.35 65.68 63.99 Temporary contract 52.61 55.27 29.48 33.22 Casual labor 3.41 6.64 3.63 2.62 Other 0.15 0.74 1.21 0.17 Notes. Based on LFS 2013 data. Recent migrants are individuals who moved less than five years prior to survey data collection. Based on the population aged 15 and over. 43 44