Causal effects of HIV on labor supply in low-income settings

Abstract: This paper estimates the effect of HIV infection on employment. We find that being HIV positive reduces the probability of being employed by 5.5 percentage points among males in Uganda. To overcome endogeneity arising from simultaneity bias in the estimation of the effect of HIV on employment status, we used a recursive bivariate probit using male circumcision as the instrument. The effect size is larger for lower socioeconomic groups and for individuals employed in manual labor professions or with characteristics predictive of employment in such professions.