Sensitivity Analysis of Inference in GMM Estimation With Possibly-Flawed Moment Conditions

Credible inference requires attention to the possible fragility of the results (p-values for key hypothesis tests) to flaws in the model assumptions, notably including the validity of any instruments used. Past sensitivity analysis has mainly consisted of experimentation with alternative model specifications and with tests of over-identifying restrictions. We provide a feasible sensitivity analysis of GMM estimation, quantifying the fragility/robustness of inference with respect to pos- sible flaws in population moment conditions of any sort and also indicating which conditions are most crucial. Two illustrative empirical applications are given: one to possible instrument flaws and another to potential explanatory variable endogeneity.