Effects of Rainfall Shocks on Child Labour: Panel Data Evidence from Ethiopia

Abstract The majority of economies in Sub-Saharan Africa are heavily dependent on rain-fed agriculture. This paper examines the impact of positive and negative rainfall shocks on child labour in agricultural households in Ethiopia. We use the timing of survey collection and rainfall in Ethiopia's two growing seasons to investigate both immediate and more long-term (after several months, in the harvest/post-harvest period) effects of rainfall variation. Using household panel data matched with geospatial rainfall data, we find that child labour is higher in the harvest/post-harvest period after a positive main-season rainfall shock and lower after a negative one. This effect is found both for agricultural activities and time spent collecting water and firewood and for both boys and girls, suggesting that households adjust child labour supply in response to changes in labour demand caused by rainfall shocks. The immediate impact of a negative Belg-season rainfall shock is also to decrease child labour. Time spent on agricultural activities decrease for both boys and girls, but the effect is stronger for boys. Furthermore, child labour responses to rainfall shocks are accompanied by effects on school attendance and absenteeism. We also investigate whether Ethiopia's public works program or access to local credit markets appear to mitigate the impact of rainfall shocks on child labour. Our results suggest that they do not.