Jeanne Commault: How Does Consumption Respond to a Transitory Income Shock? Reconciling Natural Experiments and Structural Estimations
Job talk. Jeanne Commault is a PhD candidate at Ecole Polytechnique. She will present "How Does Consumption Respond to a Transitory Income Shock? Reconciling Natural Experiments and Structural Estimations".
Jeanne Commault. Photo: personal website
Results from natural experiments show that consumption responds strongly and significantly to a transitory variation in income such as a tax rebate or a lottery gain. On the contrary, structural methods relying on theoretical restrictions to identify the consumption elasticity to a transitory shock in survey data find that this elasticity is small and not statistically significant. I show that this discrepancy is resolved when allowing a household’s log-consumption growth to depend on its history of income shocks, in structural methods. First, I establish that such a correlation exists even in the baseline life-cycle model because of precautionary saving, which simultaneously contributes to a households consumption growth and depends on its history of shocks, but is assumed away by existing structural methods. In more general frameworks, borrowing constraints, illiquid reserves of wealth or habit persistence induce a similar correlation with past shocks. Second, I show that this correlation undermines the validity of the identifying restrictions used in existing structural methods, giving rise to a downward bias. Third, I develop a robust structural estimator that allows past shocks to affect log-consumption growth. With it, the elasticity to a transitory shock is statistically significant and the results imply that, on average, 20% of a transitory gain in net income is consumed within the following year, which is comparable to the findings of the literature based on natural experiments.