Labor Suply and Rainfall Shocks in Rural Ethiopia

This study investigates how labor supply of working-age individuals in rural Ethiopia changes in response to the shock of a deviation in rainfall during the past year from its long-term average. The study focuses on time allocation to three work activities and to two household production activities and on differences in labor supply responses by gender. The method used is nonparametric regression that is less restrictive than the parametric models commonly used in the previous literature. Parametric models rely on functional forms that may be unrealistic characterizations of the data. The nonparametric regressions, in contrast, may be a superior characterization of the data. The study uses information on individuals' time allocation and characteristics from the 2015-2016 Ethiopian Socioeconomic Survey. Time allocation during the past week in early 2016 at the extensive (participation) and intensive (hours) margins is related to the difference in the logarithm of average rainfall in 2015 and the logarithm of average rainfall in 2001-2015 at a household's geolocation. According to the findings, a positive rainfall shock is associated with an increased probability that men participate in work activity overall and in household agricultural work. Men's hours of work, overall and in household agriculture, are unaffected by a positive rainfall shock. Therefore, men who already worked reduce their hours of work. For women, a positive rainfall shock is associated with reduced hours of work overall, in household agriculture and in household nonagricultural business work. A positive rainfall shock is associated with an increased probability that women participate in wage work for others. This finding is consistent with a reduction in income risk for households in which women join the wage labor force. Following a positive rainfall shock all genders are more likely to take secondary jobs in addition to working in household agriculture. Taking a secondary job is a way of reducing income risk and potentially helping to smooth consumption. A positive rainfall shock is associated with a modest increase in the probability that women will collect water but also a modest reduction, for all genders, in the time spent collecting water.