Cross-Market Alpha: Testing Short-Term Trading Factors in the U.S. Market via Double-Selection LASSO

TL;DR

This paper addresses cross-market validation of short-term trading-based factors by testing the Alpha191 library, originally developed in the retail-dominated Chinese market, within the U.S. market. It introduces a Double-Selection LASSO (DS-LASSO) framework to mitigate high-dimensionality and omitted-variable bias when screening hundreds of potential signals. The study finds that $17$ Alpha191 factors have significant incremental explanatory power after controlling for $151$ traditional factors, with signals clustering around Volume-Flow, Mean Reversion, and Volatility-Risk themes, supporting a Behavioral Universality across markets. The results suggest that fast trading signals contain non-redundant information beyond the conventional factor zoo and highlight the value of cross-market, dual-horizon asset-pricing frameworks, while acknowledging limitations such as monthly data aggregation and the need for higher-frequency analyses and non-linear extensions.

Abstract

Current asset pricing research exhibits a significant gap: a lack of sufficient cross-market validation regarding short-term trading-based factors. Against this backdrop, the development of the Chinese A-share market which is characterized by its retail-investor dominance, policy sensitivity, and high-frequency active trading -- has given rise to specific short-term trading-based factors. This study systematically examines the universality of factors from the Alpha191 library in the U.S. market, addressing the challenge of high-dimensional factor screening through the double-selection LASSO algorithm an established method for cross-market, high-dimensional research. After controlling for 151 fundamental factors from the U.S. equity factor zoo, 17 Alpha191 factors selected by this procedure exhibit significant incremental explanatory power for the cross-section of U.S. stock returns at the 5% level. Together these findings demonstrate that short-term trading-based factors, originating from the unique structure of the Chinese A-share market, provide incremental information not captured by existing mainstream pricing models, thereby enhancing the explanation of cross-sectional return differences.