Testing Short-Term Alpha191 Signals in U.S. Stocks with Double-Selection LASSO

Summary

The study tests whether short-horizon price and volume signals developed for China’s A-share market add pricing information for S&P 500 stocks beyond established U.S. factors. It applies double-selection LASSO to assess signal loadings in a stochastic discount factor framework, then checks sensitivity to test-asset design and alternative control-selection methods.

The baseline analysis finds significant loadings for some signals, but robustness varies. Three signals remain significant across the reported finer-grid and alternative-method checks; others pass only some checks or depend on the specification. More robust results cluster in volume-price interactions and short-term mean reversion, while volatility signals are less stable. The evidence covers the study’s stated U.S. sample and model designs; signal significance does not by itself establish practical trading profitability or persistence beyond that setting.

Key ideas

• The study tests China-developed Alpha191 signals on U.S. equities beyond established U.S. factors.
• Double-selection LASSO is used to assess signals within a stochastic discount factor framework.
• Robustness checks vary the test-asset grid and control-selection method.
• Only a subset of signals remain significant across the reported checks.
• Volume-price and short-term mean-reversion signals are more robust than volatility signals.

Tags

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


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









We test whether 168 short-horizon price-volume signals from the Alpha191 library, originally developed for China's retail-dominated A-share market, contain pricing information for S&P 500 stocks from 2002 to 2022 beyond 153 established U.S. factors. Using the double-selection LASSO of Feng et al. (2020), 17 signals receive significant stochastic discount factor (SDF) loadings in the baseline test-asset design. Their robustness is uneven. Only three signals (a multi-horizon moving-average ratio, a directional-pressure ratio, and a price-gap correlation) remain significant with a finer test-asset grid and under Elastic Net and principal-component control selection; six more pass most checks, and the remaining eight depend on the specification. Robust signals are concentrated in volume-price interaction and short-term mean reversion, whereas volatility-based signals are fragile.

Shown in full with attribution under the source's licence. Licence: abstract CC0

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.