讲座题目:Regime Switching Conditional Factor Models
时间:2026年6月4日(周四)下午13:30-15:00
地点:博学楼509
主讲人:王法

北京大学经济学院金融系 王法 副教授
主讲人简介:王法,北京大学经济学院金融系副教授,博士生导师,美国雪城大学经济学博士。加入北京大学前,曾在伦敦大学卡斯商学院和上海财经大学担任助理教授。研究领域为金融计量,因子模型,因果推断和高维计量经济学,主讲时间序列、固定收益证券等课程。研究成果多次发表于Journal of Econometrics和Econometric Reviews等期刊,并多次担任经济学国际一流期刊的审稿人。
讲座内容简介:To capture abrupt changes in the risk exposures and the prices of risk, we develop a regime-switching generalization of the Instrumented Principal Component Analysis (IPCA) model. Our framework allows regime-switching dependence of both the risk exposures and the prices of risk on the observable instruments. We estimate the model using the Expectation-Maximization (EM) algorithm for maximum quasi-likelihood at the managed-portfolio level. Within the approximate factor model framework, we establish the consistency of the regime identification and the asymptotic properties of our estimators for large N and large T. Regime switching can be detected consistently right after the turning point with only one observation period. In our empirical application to the U.S. stock returns, the estimated regimes align closely with the business cycle fluctuations, and our model can achieve significantly higher out-of-sample predictive power than that of the IPCA model and its auto-encoder generalization.
作者 | 郭如玥
一审 | 闵小珊 二审 | 吴虹利 三审 | 王茂斌
排版推送 | 何佳苗