Yuma Noritomo

Andrew W. Mellon Student Initiated Research Grant, Yamada Academic Research Encouragement Fund (Public Interest Trust), Chester O. McCorkle, Jr. Student Scholarship (AAEA), International Research Travel Grants (Einaudi Center), Research Travel Grant (Graduate School, Cornell University)*2, Academic Enrichment Program (Cornell CALS), Graduate Research Funding (The Center for the Study of Inequality)

AEARCTR-0017654

We study the long-run effects of catastrophic drought insurance on pastoralist households in Kenya and Ethiopia. Instrumenting with randomized premium discounts, we estimate the impacts of insurance coverage a decade later. Insurance coverage induced households to herd fewer small livestock like goats, typically used as precautionary savings in the short run, an effect which persisted in the long-run. Insurance also generated a significant long-run increase in children’s education. These effects are driven by households with small baseline herds, reflecting reduced child labor demand. We provide suggestive evidence that these impacts stem from reduced \textit{ex ante} risk exposure and the behavioral change it induces, rather than from \textit{ex post} indemnity payments.

Poor households in developing countries often face trade-offs between children’s schooling and labor. Using data on pastoralist households in rural Kenya and Ethiopia–where livestock require care and generate within-household demand for child labor–I exploit quasi-random variation in rangeland grazing conditions that affects both household income and labor demand to estimate the impact of productivity shocks during childhood on educational attainment. Positive productivity shocks during preschool ages increase completed schooling, primarily by raising the probability of primary school entry, consistent with the relaxation of short-term liquidity constraints. In contrast, negative productivity shocks during primary school ages–when children are most likely to exit school and child labor productivity increases–are associated with higher educational attainment. This effect appears to operate through significant reductions in livestock holdings that lower subsequent demand for boys’ labor in animal husbandry and probably reinforce the effect by inducing sedentarization. These results highlight the importance of non-separable household production and human capital decisions in shaping educational outcomes during critical stages of childhood in low-income settings.