Topic:

Illusion of Stock Return Predictability

Time&Date: 

10:30 am-12:00 pm, April 18, 2024 (Thursday)

Venue

Room 625, Zhiren Building

Speaker:

Prof. Jingyu He (City University of Hong Kong)

The existing body of literature on stock return predictability has predominantly focused on aggregate performance, often overlooking the inherent heterogeneity in stock return behavior across different securities and time periods. This paper sheds light from a different angle by examining the heterogeneous nature of stock return predictability using high-dimensional stock characteristics and macroeconomic predictors. We introduce a novel tree-based asset clustering methodology designed to partition the panel of stock-return observations in the cross section and time series that aims to differentiate their predictability. Studying U.S. equity for the past five decades, we find that some characteristics-based or macro-based clusters are more predictable and further linked to higher risk-adjusted investment performance and risk anomalies.
Key Words: asset clustering, heterogeneity, high-dimensional characteristics, macroeconomic regimes, return predictability.

Biography:

Jingyu He is an assistant professor of business statistics at the City University of Hong Kong. He received Ph.D., M.B.A., and M.S. from The University of Chicago, and B.S. from University of Science and Technology of China. His research interests include machine learning in finance, empirical asset pricing, Bayesian statistics and econometrics. His research work has appeared in leading statistics, finance and econometrics journals.