Harvests and Hooky in the Hills: Crop Yield Variability and Gendered School Enrollment in Rwanda

This paper investigates the trade-off that households in agrarian economies face between immediate production needs and long-term human capital investment. We ask how exogenous agricultural productivity shocks affect primary and secondary school enrollment in Rwanda, a country characterized by a heavy reliance on rain-fed agriculture alongside ambitious development goals. Using a district-level panel dataset for the years 2010-2021, we employ a two-stage least squares (2SLS) instrumental variable strategy. Plausibly exogenous variation in annual rainfall is used to instrument for a satellite-derived measure of vegetation health and agricultural productivity, the Normalized Difference Vegetation Index (NDVI), allowing for a causal interpretation of the results. For primary education, where direct costs are low, enrollment is countercyclical: positive productivity shocks are associated with lower enrollment. Notably, boys' primary enrollment is found to be significantly more elastic to these shocks than girls' enrollment. Conversely, for secondary education, which entails additional financial outlays, enrollment is strongly procyclical. Positive productivity shocks lead to significant increases in enrollment. Further, a sustained positive shock is associated with a subsequent decline in female secondary enrollment. The results challenge previous regional findings supporting the "girls as a buffer" hypothesis and investigate the dynamic and gendered responses to the persistence of economic shocks.