HIV/AIDS AND POVERTY IN SOUTH AFRICA: A BAYESIAN ESTIMATION OF SELECTION MODELS WITH CORRELATED FIXED‐EFFECTS

In this paper, we estimate the causal impact of human immunodeficiency virus/acquired immune deficiency syndrome (HIV/AIDS) on monetary poverty using a panel database from South Africa. We treat endogeneity and selection problems associated with HIV/AIDS by a selection model that includes correlated fixed effects both in the level and in the participation equations, which are estimated simultaneously via original Bayesian methods. We model the consequences of the illness on both labour income and income transfers, and disentangle between urban and rural households. While no significant impact of HIV/AIDS on labour income is found because of households' recomposition, we find a substantial fall in received transfers among rural population and a dramatic increase in chronic poverty.