Conference Agenda

I propose a novel framework to study climate change adaptation strategies via buffer saving under uncertainty, where constrained agents dispose of assets as a self-insurance mechanism in order to protect themselves from increasingly erratic income shocks and smooth consumption. I frame the problem as a mixed classical-impulse stochastic control problem, where consumption and labor supply are continuous and asset transactions happen at discrete times. I prove optimality of the solution for the general problem in both finite and infinite time, and present a closed form solution for specific functional forms. I study the probabilistic properties of asset transactions as hitting times and show that buffer stock saving emerges as the combination of wealth and precautionary effects. I further show how market imperfections can generate Pareto tails in transaction timing. I apply this novel framework to study how farmers in West Africa adapt to climate shocks by using livestock as a buffer stock. Using a novel spatially explicit climate and household dataset for Burkina Faso, I show reduced form evidence of the model predictions and quantify the welfare loss that stems from the absence of crop insurance markets. This framework is general and can be used to explain a wide variety of climate change adaptation strategies throughout the developing world.