Household and Spatial Drivers of Migration Patterns in Africa: Evidence from Five Countries Valerie Mueller (IFPRI) Emily Schmidt (IFPRI) Nancy Lozano-Gracia (World Bank) Urbanization and Spatial Development of Countries Research Workshop July 13, 2015
Introduction Migration key force shaping both origins and destinations Economic and physical mobility as forces to achieve structural change But SSA seems to be different - no clear agricultural or industrial revolution as in other countries to encourage rural-urban migration (Jedwab and Vollrath, 2015) Migration in sub-saharan Africa characterized by labor chasing due to distress or risk-reduction strategies (Campbell, 2011) Much yet to be understood
Objectives and research questions Is migration in Africa different? What are the key factors determining migration in SSA? How do they compare to places that faced an industrial (China), agricultural (India) revolution? What are the types of flows that dominate? Is rural-urban migration the main patter of movement? Is migration being used to diversify employment?
Data LSMS-ISA Surveys Ethiopia (2011-2, 2013-4) Malawi (2010-1, 2012-3) Nigeria (2010-1, 2012-3) Tanzania (2008-9, 2010-1, 2012-3) Uganda (2009-1, 2010-1, 2011-2) Detailed questions about individuals and households over time But no perfect measure of migration, and short panels
Measures of Migration Temporary migration - All countries document if household member moved over last 12 months (but don t know where they go or why) Different measures of individual permanent migration out of the household Tracked migrants in waves after baseline (Malawi and Tanzania) Self-reported migration by proxy in waves after baseline (Ethiopia and Nigeria) Low long-distance household migration is low over two-year horizon Different place than birth location (NOT used here because we don t know anything about pre-migration characteristics)
Ethiopia requires careful interpretation large urban areas excluded from wave 1
Rural-Urban Migration Country Year Time Period Source Rural-Urban Migration Ethiopia 2013-4 2 years LSMS-ISA 25.3 Malawi 2012-3 2 years LSMS-ISA 6.1 Nigeria 2012-3 2 years LSMS-ISA 11.1 Tanzania 2012-3 2 years LSMS-ISA 14.1 China 2000 5 years Census, Cai, Park, and Zhao (2008) India 2001 9 years Census, Ministry of Home Affairs 40.8 21.1
Hypotheses patterns of migration Hypothesis: Rural-rural migration patterns is driven by land scarcity, big households, and high population densities. People are NOT diversifying out of agriculture, they are moving to find a job in agriculture or access land for their own farm Hypothesis: Rural-rural migration patterns dominate because roads, liquidity constraints, education deficits and other factors pose barriers to moving to a city (Stark and Bloom, 1985; Stark, 1991; Rozelle et al., 1999; Wouterse and Taylor, 2008; Dillon et al., 2011). People are diversifying out of agriculture into non-agriculture in rural areas (e.g., to satisfy rise in demand for services). Hypothesis: Rural-urban patterns are predominantly explained by an absence of insurance mechanisms (Barrios et al., 2006; Poelhekke, 2011). People are diversifying out of agriculture into non-agriculture in cities where climate or conflict risk is low.
Migrating for Land Hypothesis Rural Urban Non-Mig. Mig. t-test Non-Mig. Mig. t-test ETHIOPIA Household size 5.63 6.36 0.74*** 4.92 5.95 1.04*** Owned land 4.21 5.31 1.1 0.5 0.45-0.04 GRUMP Pop. Density 180.19 183.27 3.09 149.29 142.8-6.49 MALAWI Household size 5.51 6.26 0.75*** 5.3 5.88 0.58*** Owned land 3.74 1.84-1.9 0.45 0.31-0.13** GRUMP Pop. Density 208.23 198.72-9.51** 1794.59 1914.98 120.39** NIGERIA Household size 7.91 8.13 0.21 6.78 7.10 0.31* Owned land 2.63 3.52 0.88* 0.35 0.66 0.31*** GRUMP Pop. Density 325.39 402.23 76.84*** 1862.38 2007.89 145.51 TANZANIA Household size 6.24 7.62 1.38*** 5.75 6.90 1.15*** Owned land 7.48 8.34 0.86 1.34 1.56 0.22 GRUMP Pop. Density 207.28 161.18-46.11* 1242.25 1189.02-53.23
Constrained Migration Hypothesis Rural Urban Non-Mig. Mig. t-test Non-Mig. Mig. t-test ETHIOPIA Primary Education 0.07 0.21 0.14*** 0.32 0.44 0.12*** No. of rooms 1.70 1.86 0.16** 2.31 2.48 0.17 Travel Time 223.44 223.54 0.10 179.27 207.02 27.75** Distance to major road 16.44 14.61-1.83** 14.29 16.48 2.20 MALAWI Primary Education 0.10 0.11 0.00 0.15 0.13-0.02 No. of rooms 2.70 2.83 0.12*** 3.23 3.41 0.17** Travel Time 131.68 130.85-0.83 42.83 37.74-5.09*** Distance to major road 9.50 9.20-0.30 1.74 1.81 0.07 NIGERIA Primary Education 0.29 0.39 0.10*** 0.31 0.31 0.00 No. of rooms 4.43 4.70 0.27** 3.57 3.75 0.17 Travel Time 176.63 156.96-19.67*** 65.99 75.44 9.45* Distance to major road 17.76 15.69-2.07*** 5.78 5.77-0.00 TANZANIA Primary Education 0.51 0.57 0.06** 0.18 0.11-0.07*** No. of rooms 3.68 4.12 0.45*** 3.53 3.98 0.45*** Travel Time 210.50 219.06 8.56 52.70 60.46 7.76 Distance to major road 22.89 23.84 0.96 45.76 41.47-4.29
Migration for insurance hypothesis Rural Urban Non-Mig. Mig. t-test Non-Mig. Mig. t-test ETHIOPIA Temperature wettest Q 19.26 19.32 0.05 19.21 19.54 0.33 Rainfall wettest Q 557.61 573.15 15.54*** 601.88 616.30 14.42 Conflict fatalities 0.82 0.93 0.11 0.80 0.93 0.13 MALAWI Temperature wettest Q 23.08 23.21 0.13* 22.01 21.86-0.15** Rainfall wettest Q 674.62 657.51-17.11*** 668.74 662.09-6.65* Conflict fatalities 0.02 0.02 0.00 2.52 2.90 0.38 NIGERIA Temperature wettest Q 25.24 25.09-0.15*** 25.29 25.16-0.13*** Rainfall wettest Q 728.83 822.37 93.54*** 723.17 766.05 42.88*** Conflict fatalities 7.68 4.70-2.97* 39.43 21.06-18.36*** TANZANIA Temperature wettest Q 23.51 23.38-0.13 25.48 25.18-0.31* Rainfall wettest Q 582.72 538.48-44.24*** 607.59 573.77-33.83*** Conflict fatalities 0.14 0.14-0.01 2.94 3.39 0.45
Pre- and Post-Migration Employment Patterns
What s next? What is rural and what is urban? Are dynamics of peri-urban areas masked within the rural classification? Explore breakdown of spatial patterns Uncovering the patterns of migration across five African countries: Estimate multivariate regressions of migration to examine which hypotheses dominate Robustness to measure of migration (temporary versus permanent) Robustness to panel framework (temporary, Tanzania permanent)
Thanks! V.Mueller@cgiar.org E.Schmidt@cgiar.org nlozano@worldbank.org