Migration and remittances in Haiti : their welfare impact on poor and non-poor households

The purpose of this study is to identify a person's likelihood of emigrating to another country and to identify a household's likelihood of receiving remittances. We also compute average treatment effects as well as the marginal impact of receiving remittances on household welfare, across welfare quantiles. The novelty of our approach is to control for omitted variable bias by including the difference between individual and average propensity scores obtained in an auxiliary regression. The fact that an individual or household is above or below the average propensity score can thus be considered as a proxy of being different from the average for a variety of characteristics that might also be unobservable or unquantifiable. Based on Haitian household survey data from 2012, we find that non-poor individuals are more likely to emigrate but the welfare level of a household per se does not trigger the receipt of remittances. The receipt of remittances favors non-poor households in absolute terms but not in relative terms. While remittances can help overcome extreme poverty (for the poorest 10% but not for the poorest 1%), they do not help people escape moderate poverty.