Information Based Inference in Models with Set-Valued Predictions and Misspecification

• Econometrics

Date and Location

Monday, November 14, 2022, 3:40 PM - 5:00 PM
Blue Room, 1113 Social Sciences and Humanities

Abstract

This paper proposes an information-based inference method for partially identified parameters in incomplete models that might be misspecified. Key features of the method are: (i) it is based on minimizing a suitable definition of Kullback-Leibler information criterion that accounts for incompleteness of the model and delivers a non-empty pseudo-true set; (ii) tests and confidence sets are based on Rao’s score statistic, which is shown to be asymptotically pivotal; (iii) implementation is the same for both correctly and incorrectly specified models; (iv) it is computationally tractable; (v) all information provided by variation in discrete and continuous covariates is exploited.