Climate Shocks and Teenage Fertility

In communities highly dependent on rainfed agriculture for their livelihoods, the common occurrence of climatic shocks can lower the marginal cost of a child and raise fertility. We test this hypothesis using longitudinal data from Madagascar. Exploiting exogenous within-district year-to-year variation in rainfall deficits in combination with individual fixed effects, we find that drought occurring in the agricultural season increases the fertility of young women living in agricultural households. This effect is long-lasting, as it is not reversed within four years after the drought occurrence. Analyzing mechanisms, we find that drought does not affect common factors of high fertility such as marriage timing. It operates mainly through a reduction of female agricultural income. Indeed, agricultural drought reduces the number of hours worked by women in agriculture but not men. It has no effect on the fertility of young women living in non-agricultural households, or in non-agrarian communities. Moreover, it does not affect fertility if it occurs during the non-agricultural season. These findings validate the marginal cost hypothesis whereby drought, by reducing the value of women's agricultural labor, lowers the marginal cost of a child, thus raising fertility.