Nemmers Prize Lecture
"Linear Estimation of Structural and Causal Effects in Nonseparable Panel Data with Applications to Heterogeneous Taxable Income Elasticities and Demand Analysis"
Many questions in economics concern how people respond to changes in their economic environment, for example how taxable income responds to tax rates, or how purchases respond to prices. These responses, which economists study as structural or causal effects, can differ considerably across individuals.
The lecture presents a method for estimating the average of such individual-specific effects using panel data, which follow the same individuals over many periods. The approach estimates a separate regularized regression for each individual, averages the results, and then corrects the average for the resulting regularization bias. The method relies on a linear approximation to how outcomes depend on the variables of interest, and it comes with simple standard errors that are reliable when both the number of individuals and the number of periods are large.
Two applications are presented. The first estimates the average of heterogeneous taxable income elasticities in the Panel Study of Income Dynamics (PSID) when tax schedules are endogenous. The second estimates the bounds of the welfare effects of taxes from price scanner data and allows for prices to be endogenously determined.
