Does Consumption Respond to Transitory Shocks? Reconciling Natural Experiments and Semistructural Methods†

Studies based on natural experiments find that consumption responds strongly and significantly to a transitory variation in income, while semistructural estimations find no pass-through of transitory shocks to consumption. I develop a more robust semistructural estimator that relaxes the assumption that log consumption is a random walk. The robust pass-through estimate is significant and large, implying a yearly marginal propensity to consume of 0.32, close to the natural experiment findings. The robust estimator performs well in numerical simulations of a life cycle model, while nonrobust estimators do not. The difference between the two in the simulations is similar to their difference in the survey data. (JEL D15, E21, J11)