Poverty Traps and Growth in a Model of Endogenous Time Preference

Abstract We introduce endogenous probability of survival in the Keynes-Ramsey optimal growth model. An individual's probability of survival is assumed to be dependent on past levels of consumption. Endogenous probability of survival implies that the rate of time preference (or degree of patience) of an individual is endogenously determined. We solve the dynamic optimization problem facing an agent and provide a complete characterization of the steady states and their stability properties. We find that with endogenous rate of time preference an economy may have multiple steady state equilibria. The equilibrium an economy converges to depends on its initial conditions. The results are interpreted in light of the growth experiences of developing economies. The model can explain why two economies that have identical production technologies and identical preferences may converge to different levels of income depending on initial conditions. We estimate the relationship between adult probability of survival and lagged consumption for a cross section of countries. Our estimation results and subsequent simulations of the model suggest that if we interpret capital in our model broadly to include both physical and human capital, poverty traps are empirically plausible.