Design-Based Estimation of Structural Parameters, with an Application to Demand
We develop a design-based methodology for using as-good-as-random shocks to estimate key parameters in a class of structural models. Our approach uses recentered instruments to avoid restrictions on how model unobservables relate to predetermined variables. In the classical setting of estimating demand for differentiated products, this approach avoids the strong assumption of exogenous product characteristics. We illustrate these ideas by leveraging exchange rate shocks to estimate the demand for automobiles.