报告题目：Testing trending time-varying coefficient spatial panel data models

报 告 人：宋晓军 教授（北京大学 光华管理学院）

报告时间：2024年9月9日 14：30

报告地点：格物楼数学研究中心528室

报告摘要：This paper proposes nonparametric tests for trending time-varying coefficient spatial panel data models. The residual-based U-statistic tests whether the coefficients of the spatial panel data models are time-varying. The proposed test includes three cases: (i) whether the spatial coefficient, the regression coefficient, and the time effects are all time-varying; (ii) whether the spatial coefficient and the regression coefficient are time-varying while the time effects follow a trend function; (iii) whether the spatial coefficient changes over time while the regression coefficient and the time effects are time-varying. All statistics follow an asymptotic standard normal distribution under the null hypothesis of parameter constancy and diverge to infinity in probability under the corresponding alternatives. The asymptotic properties of the proposed tests under two types of local alternatives are also investigated.