Speaker: Xiaoxia Shi, University of Wisconsin

Title: Testing Inequalities Linear in nuisance parameters

Abstract: This paper proposes a new test for inequalities linear in possibly partially identified nuisance parameters. It extends the subvector conditional chi-squared (CC) test in Cox and Shi (2022, CS22) to a setting where the nuisance parameter is multiplied to an unknown and estimable matrix. Properly accounting for the estimation noise in this matrix while maintaining the simplicity of the CC test is the main innovation of this paper. As such, the paper provides a simple solution to a broad set of problems including subvector inference for models represented by linear programs, nonparametric instrumental variable models with discrete regressor and instruments, and linear unconditional moment inequality models. We also derive a simplified formula for computing the critical value that makes the computation of the CC test elementary.

Area of research: Econometrics

Organisers: Luke Taylor and Leopoldo Catania.