Multiscale Adaptive Inference on Conditional Moment Inequalities

This paper considers inference for conditional moment inequality models using a multiscale statistic. We derive the asymptotic distribution of this test statistic and use the result to propose feasible critical values that have a simple analytic formula. We also propose critical values based on a modified bootstrap procedure and prove their asymptotic validity. The asymptotic distribution is extreme value, and the proof uses new techniques to overcome several technical obstacles. We provide power results that show that our test detects local alternatives that approach the identified set at the best possible rate under a set of conditions that hold generically in the set identified case in a broad class of models, and that our test is adaptive to the smoothness properties of the data generating process. Our results also have implications for the use of moment selection procedures in this setting. We provide a monte carlo study and an empirical illustration to inference in a regression model with endogenously censored and missing data.