报告题目：Panel Endogenous Interactive Effects Models
嘉宾简介：Cheng Hsiao is a Professor of Economics at the University of Southern California. He is a Fellow of Academia Sinica, a Fellow of the Econometric Society, and a Fellow of the Journal of Econometrics. In 2018, he became a Founding Fellow of the International Association for Applied Econometrics and was also awarded the Multa Scripsit Award by Econometric Theory in 2012. He has served on the advisory boards of Pacific Economic Review and Singapore Economic Review, and as the Editor of Journal of Econometrics for over two decades.
Professor Hsiao's research interests include theoretical and applied econometrics. He has published extensively in these areas, with numerous research articles in leading journals such as Econometric Theory, Econometrica, Journal of Business and Economic Statistics, Journal of Econometrics, Journal of the American Statistical Association, Journal of Monetary Economics, and Review of Economic Studies, and several books, including Analysis of Panel Data and Panel Data Econometrics.
报告摘要：Many micro econometric modeling takes a single equation approach. However, economists typically consider the general (or partial) equilibrium (or coherence) of (a sector of) the economy is achieved through the automatic operation (say, free competition) of the parts that compose the system. We argue that without considering if an equation is identified in the system, the estimation of such an equation could lead to meaningless curve fitting. Panel data provide the possibility to decompose the stochastic errors into components, thus introducing restrictions on the error variance covariance. We show that incorporating such restrictions provide the possibility to relax the rank and order conditions for the identification of an equation developed by the Cowles Commission. We review inference methods for a single equation panel interactive model including the least squares or two stage least squares, profile generalized method of moments (PGMM) or transformed GMM in light of the identification conditions in an interconnected system. Limited Monte Carlos are also conducted to demonstrate finite sample properties of those methods.
报告时间：2025年11月4日(周二)，16:30-18:00
线下地点：厦门大学经济楼C108
腾讯会议：883 676 549
