Website CV Paper
Fields
Econometrics
Job market paper
How Robust are Robustness Checks?
Robustness checks are routine in empirical work, but there is no standard statistical procedure to formally measure what one can learn from them. I propose the robustness tolerance: the smallest distance b at which the hypothesis that every robustness-check estimand lies within b of the main-specification estimand is not rejected. It is a lower confidence bound for the maximal discrepancy, obtained by inverting a moment-inequality test, and it accounts for the sampling uncertainty and the correlation of the estimates. I show the robustness tolerance satisfies desirable theoretical properties. An application shows that, although assessing overall robustness is context-specific, the robustness tolerance guides those judgments and improves transparency.