Insuring Against a Poverty Trap

A growing evidence base suggests that in some environments a critical asset threshold exists, around which equilibrium outcomes and optimal behaviors bifurcate. This creates a poverty trap, such that whenever a household's asset holdings fall below this critical threshold, the household becomes trapped in poverty. This dissertation considers the value of insurance to poor households in this type of environment.;I first develop a theoretical model of demand for asset insurance in the context of a poverty trap. I use stochastic dynamic programming techniques to show that households holding assets near a critical threshold will not insure, because the opportunity cost of insurance is high. Paradoxically, these households have the most to gain from insurance. Despite low initial demand, an insurance-induced behavioral response and the ability to protect assets in the future reduces vulnerability for these households, significantly altering the expected poverty dynamics.;I then take advantage of a unique empirical application in Kenya, where pastoralists have insured against drought-related livestock losses since 2010. Previous research in this context has provided evidence of nonlinear asset dynamics indicative of a poverty trap, making this an appropriate setting for understanding how households react to insurance when nonlinear asset dynamics exist. The empirical chapters analyze both impact and demand in this setting.;The impact analysis results suggest that a drought-induced insurance payout prevents households prone to selling productive assets from doing so, otherwise harming their future income-earning potential. In addition, insurance helps prevent those households inclined to reduce consumption from doing so, thereby protecting against undernutrition and malnutrition, and improving the human capital of future generations. Together, these results show that insurance can act as a safety net, allowing smoothing of consumption and nutrition, while preserving productive assets.;The structural demand analysis uses the theoretical framework to analyze demand for drought insurance in Kenya. I use both parametric and semiparametric techniques for panel data, and find some evidence in support of demand that is responsive to a critical asset threshold. However, the analysis suggests that other constraints, including understanding and trust in the insurance product, may be more important in explaining demand.