Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
5096912 | Journal of Econometrics | 2010 | 11 Pages |
Abstract
This paper studies conditional moment restrictions that contain unknown nonparametric functions, and proposes a general method of obtaining asymptotically distribution-free tests via martingale transforms. Examples of such conditional moment restrictions are single index restrictions, partially parametric regressions, and partially parametric quantile regressions. This paper introduces a conditional martingale transform that is conditioned on the variable in the nonparametric function, and shows that we can generate distribution-free tests of various semiparametric conditional moment restrictions using this martingale transform. The paper proposes feasible martingale transforms using series estimation and establishes their asymptotic validity. Some results from a Monte Carlo simulation study are presented and discussed.
Related Topics
Physical Sciences and Engineering
Mathematics
Statistics and Probability
Authors
Kyungchul Song,