现代产业经济学系列讲座第292期

时    间：2026年4月7日15:30-17:00

地    点： 复旦管院史带楼303室

题    目：A Spatial Autoregressive Type-III Tobit Model with an Application to Labor Supply and Wages

主讲人：周小焱 助理教授 北师香港浸会大学 工商管理学院 金融经济系

主持人：俞学文 副教授

摘要: Traditional labor supply models usually assume that spouses make sequential decisions, while modern families often exhibit simultaneous decision-making and peer effects such that labor hours and wages are correlated across family members and social networks. Failure to account for these cross-sectional dependencies may bias estimates of wage gaps stemming from discrimination. This paper addresses this concern by proposing a Spatial Autoregressive Type-III Tobit (SART-III) model under a latent specification, which incorporates spatial correlations in both the selection into employment and wage outcomes. We develop a Bayesian approach to estimating the SART-III model. Results from Monte Carlo experiments demonstrate satisfactory finite-sample performance of the proposed algorithm. An empirical study using the 2023 American Community Survey (ACS) data indicates that adopting a baseline Tobit-III model and ignoring cross-sectional dependencies would lead to overestimation of racial and gender wage gaps due to failure of disentangling network effects from impact of discrimination.

个人简介：Xiaoyan Zhou is an assistant professor at Beijing Normal-Hong Kong Baptist University. He received his Ph.D. in Economics from Purdue University (2025), M.A in Economics from Duke University (2019) and B.A. in Economics from Renmin University of China (2017). He has a broad research interest in econometrics and applied microeconomics. His current research mainly focuses on spatial econometric models with high dimensions or censored dependent variable, with applications to real estate economics, urban and regional economics, labor economics, and so on. Besides, he also has research work in Bayesian time series analysis and applied macroeconomics.