Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
5062103 | Economics Letters | 2008 | 9 Pages |
Abstract
This paper proposes a new specification test for discrete choice models based on nonparametric estimation of the likelihood ratio. The test is robust against any possible misspecification. A bootstrap procedure is proposed to obtain critical values in small samples. Monte Carlo simulations show that the proposed test can have much higher power than some of the well known parametric tests.
Keywords
Related Topics
Social Sciences and Humanities
Economics, Econometrics and Finance
Economics and Econometrics
Authors
Xu Zheng,