【学术讲座】An Alternative Estimator for the Quantile Sample Selection Model with Endogeneity

发布者:统计与数据科学学院发布时间:2022-04-07浏览次数:1349


专家简介:周亚虹,香港科技大学经济学博士,现任上海财经大学经济学院院长,常任教授、博导,教育部特聘教授,研究领域为微观计量经济学、微观经济学。

报告摘要:Arellano and Bonhomme (2017) studied quantile sample selection model with endogeneity in the spirit of inverse quantile regression approach motivated by Chernozhukov and Hansen (2006). This paper corrects for endogeneity by adopting a control function approach. Similarly with Arellano and Bonhomme (2017), we correct for sample selection by modeling the error terms in outcome and selection equation via the copula structure. We facilitate the estimation procedure by carrying out the series approximation of the functional of control variable included as an additional explanatory variable in the primary outcome equation. Compared with the inverse quantile regression, the control function approach is easy to implement and avoids solving high-dimensional optimization problem of non-convex objection function. We establish the asymptotic normality for our estimator. We also propose a test for the existence of endogeneity. Finite sample results of our estimator and the comparison with inverse quantile regression are presented. As an illustration, we apply our method to estimate the impacts of woman's education on the wage offer.

腾讯会议号:750-978-539

  间:202248上午 9:00