Estimating the Dynamics of Consumption Growth

Abstract: We study the dynamics of consumption growth in a series of countries over a time span of 200 years. We seek to answer whether long-run risk or disasters are features of models that yield good fit to consumption data. To accomplish this goal, we develop a new methodology for the filtering and estimation of multivariate jump-diffusion processes in the presence of incomplete data and measurement errors. Our methodology is both statistically and computationally efficient, and enables the empirical analysis of previously intractable multidimensional models. Our estimates suggest that small and frequent disasters that arrive at a time-varying rate are a predominant feature of consumption data. Persistent stochastic volatility is also found to be a significant driver of consumption growth, especially in the United States.