Household preferences and child labor in rural Ethiopia

This paper revisits the causes behind child labor supply by focusing on an aspect that has received little attention: the link between the household head's risk and time preferences and observed child labor supply. We develop a theoretical model and empirically test for this link using data from the seventh round of the Ethiopian Rural Household Survey. We find child labor at the extensive and intensive margins to be increasing in both higher adult discount rates and higher degrees of risk aversion, and this finding is robust across alternative empirical specifications. While higher time discount rates are supposed to favor child labor at the expense of schooling, the explanation of the positive link between risk aversion of the household head and observed child labor supply is nuanced. The most plausible explanation is that risk averse parents react to uncertain returns to education by combining child labor and schooling as opposed to substituting schooling for child labor.