Measuring the Sensitivity of Parameter Estimates to Sample Statistics

Empirical papers in economics often describe heuristically how their estimates depend on intuitive features of the data. We propose two quantitative measures of this relationship that can be computed at negligible cost even for complex models. We show that our measures can be informative about robustness to model misspecification, and can complement the discussions of identification that have become common in applied work. We illustrate our measures with applications to industrial organization, macroeconomics, public economics, and finance.