Causes and consequences of internal migration: evidence from Brazil and Ghana

This thesis investigates socio-economic drivers and impacts of internal migration in two countries, Brazil and Ghana. The first empirical chapter analyses the choice of Brazilian workers to move out of metropolitan cities. This direction of movement is substantial in the Brazilian context and leading against standard models of rural-to-urban migration. I estimate the role of living costs and local amenities in the determination of the destination choice of metropolitan out-migrants. Furthermore, I quantify the returns to migrating out of a metropolis by computing counterfactual wages applying matching techniques. The metropolitan out-migrants prefer to move to smaller towns where their real wage gain is positive. They minimize the physical and social costs of migration by moving to closer towns within their state of birth. Living costs in big cities appear to be a main driver for workers to leave these, especially if they are low-skilled. In the second empirical chapter, I investigate the effect of internal migration on homicide rates in Brazil in the period from 2005 to 2010. I construct a retrospective panel of migration rates between municipalities and use local labour demand shocks in the manufacturing sector at the origins of migrants as instrument for immigration rates. An increase in immigration rates of 1% translates in an increase of 1.2% in crime rates at the local level. The effect is predominant in municipalities with historically higher homicide rates and there is no effect in locations with a large informal sector. While internal migration puts pressure on destination labour markets, these results suggest that it is the presence of a criminal or lack of a flexible sector that channel this pressure into negative outcomes. The third empirical chapter explores dynamic patterns of internal migration from rural areas in Ghana. With a new household panel survey collected in 2013 and again in 2015, I document that many households have multiple migrants moving at different points in time and for various reasons. Conditional on having had a migrant in the past, I estimate the effect of having a new migrant on the asset welfare of origin households. The findings suggest that due to prior migration experience and consequently lower migration costs for new migrants, there is no decline in welfare from having a new migrant.