A Bipartite Graph Approach to US-China Cross-Market Return Forecasting
A new study reveals a strong one-way link in stock return prediction between the U.S. and Chinese markets. Using a time-ordered bipartite graph and machine learning, researchers found that recent U.S. market data, especially previous-close-to-close returns, can reliably forecast Chinese intraday returns. The reverse effect is much weaker. The method, tested on data from 2014 to 2021 for 500 top stocks in each market, improves accuracy but does not consider trading costs or real-world constraints.
What it examines
This paper uses a machine learning approach with a directed bipartite graph to predict stock returns between the U.S. and Chinese markets. It focuses on how information from one market can help forecast returns in the other, especially given their non-overlapping trading hours.
What it concludes
The study finds U.S. stock returns are much more useful for predicting Chinese stock returns than the reverse. This method can help investors and analysts understand cross-market influences. Future research could expand to other markets or use advanced models, but practical trading applications need further testing.
Evidence objects
A groundbreaking study reveals U.S. stock returns, especially previous-close-to-close data, strongly predict Chinese intraday returns, while the reverse effect is much weakerhighlighting the U.S. markets dominant global influence.
key_findings bullet 1 · key_findings · validation V0
Researchers introduce a novel time-ordered bipartite graph to select cross-market predictors, feeding them into ten machine learning models. This approach boosts forecasting accuracy and interpretability, a rare achievement in financial modeling.
key_findings bullet 2 · key_findings · validation V0
Using data from 2014--2021 on 500 top stocks per market, the study finds predictive power peaks with recent U.S. data and proper graph structure, but real-world trading constraints remain unaddressed.
key_findings bullet 3 · key_findings · validation V0
This paper introduces a novel directed bipartite graph framework for stock-level cross-market return prediction between US and Chinese equities, exploiting non-overlapping trading hours for clean predictive linkages. Its originality lies in combining economic interpretability with machine learning feature selection, revealing economically meaningful directional asymmetry, making it compelling and distinct from prior index-level studies.
key_findings bullet 4 · key_findings · validation V0
Raw abstract and provenance
Abstract: This paper studies cross-market return predictability through a machine learning framework that preserves economic structure. Exploiting the non-overlapping trading hours of the U.S. and Chinese equity markets, we construct a directed bipartite graph that captures time-ordered predictive linkages between stocks across markets. Edges are selected via rolling-… ▽ More This paper studies cross-market return predictability through a machine learning framework that preserves economic structure. Exploiting the non-overlapping trading hours of the U.S. and Chinese equity markets, we construct a directed bipartite graph that captures time-ordered predictive linkages between stocks across markets. Edges are selected via rolling-window hypothesis testing, and the resulting graph serves as a sparse, economically interpretable feature-selection layer for downstream machine learning models. We apply a range of regularized and ensemble methods to forecast open-to-close returns using lagged foreign-market information. Our results reveal a pronounced directional asymmetry: U.S. previous-close-to-close returns contain substantial predictive information for Chinese intraday returns, whereas the reverse effect is limited. This informational asymmetry translates into economically meaningful performance differences and highlights how structured machine learning frameworks can uncover cross-market dependencies while maintaining interpretability. △ Less
Source row: 4 · abstract type: unknown