Understanding how internal labour migration affects the agricultural sector is important for all developing countries whose markets do not work well or are non-existent. In fact, even if the movement out of the agricultural sector can be viewed as a process to reach development for many African countries, this could lead to a negative effect on the rural economy. The availability of labour and the cost of hiring people to work on farms is an example of a problem that farmers may face in the presence of a critical level of labour migration. This paper investigates the effect of internal labour migration on agricultural productivity of rural households in Uganda. Since households select themselves into migration this raises the endogeneity problem. In order to account for the endogeneity of the migration decision and the fact that the effect might be different from one household to another, I model the households' decisions to participate in migration along with their investment in agricultural productivity using the Bayesian treatment analysis. This approach allows me to self-match each household and to estimate a distribution for the counterfactual outcome. The results show that even if on average internal labour migration positively affects agricultural productivity, there are some households for which the effect is negative. Those households for which the effect is negative are mostly small farmers and are therefore more likely to be poor and thus more sensitive to the local price volatility. Moreover, the average effect of the labour migration tends to increase with the likelihood of participating in the internal labour migration. In parallel, I also examine to what extent previous migration rates, widely used in the literature as instrument for the migration decision, are exogenous to the agricultural productivity. It turns out that previous households' decisions to participate in migration are intimately correlated with their current agricultural productivity.
labour migration agriculture Bayesian treatment analysis instrumental variables rural Uganda