Paper: arXiv 2603.10559

Authors: Jing Liu, Maria Grith, Xiaowen Dong, Mihai Cucuringu

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-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.

Complexity vs Empirical Score

  • Math Complexity: 5.5/10
  • Empirical Rigor: 6.0/10 (capped from 7.5: transaction costs not modelled; see penalty flags)
  • Quadrant: Holy Grail — high math complexity, high empirical rigor

Why this score: The paper provides a compelling empirical framework applying directed bipartite graphs to capture cross-market return predictability under realistic, non-overlapping trading session constraints. The math is clear and statistically grounded without being overly abstract, while the empirical validation across ten ML models demonstrates strong rigor and clear findings on U.S.-China market informational asymmetry.

Research Flowchart

  flowchart TD
    A[Research Goal: Forecast U.S.-China Cross-Market Intraday Returns] --> B[Input Data: Non-Overlapping U.S. & Chinese Stock Market Returns]
    B --> C[Construct Directed Bipartite Graph via Rolling Hypothesis Testing]
    C --> D[Feature Selection: Extract Sparse, Economically Interpretable Linkages]
    D --> E[Apply ML Models: Regularized Regression & Ensemble Methods]
    E --> F[Key Outcome: Strong Asymmetric Predictability U.S. to China]

Cited References (65)

  1. … and the Cross-Section of Expected Returns (2015) — link
  2. Sure independence screening for ultrahigh dimensional feature space (2008) — link
  3. Liquidity Constraints and the Illusion of Anomaly Profitability (2025)
  4. Fluctuation scaling and covariance matrix of constituents’ flows on a bipartite graph (2010)
  5. International stock return predictability: on the role of the United States in bad and good times (2017)
  6. International stock return predictability: what is the role of the United States? (2013)
  7. A Bipartite Graph-Based Recommender for Crowdfunding (2020)
  8. Modelling extremal dependence for operational risk by a bipartite graph (2020)
and 57 more
  1. Bipartite network influence analysis of a two-mode network (2024)
  2. Networks (2018)
  3. Bipartite graphs as models of complex networks (2006)
  4. The deflated Sharpe ratio: Correcting for selection bias, backtest overfitting and non-normality (2014)
  5. The elements of statistical learning: data mining, inference, and prediction (2009)
  6. Support vector machines (1998)
  7. Xgboost: A scalable tree boosting system (2016)
  8. Lightgbm: A highly efficient gradient boosting decision tree (2017)
  9. Random forests (2001)
  10. A decision-theoretic generalization of on-line learning and an application to boosting (1997)
  11. Portfolio selection in quantile decision models (2022)
  12. A systematic review of stock market prediction using machine learning and statistical techniques (2022)
  13. A LSTM-based method for stock returns prediction: A case study of China stock market (2015)
  14. Stock return prediction with multiple measures using neural network models (2024)
  15. Threshold Overnight Comovement Analysis of Intraday and Overnight Returns (2024)
  16. US stock market uncertainty and cross-market European stock returns (2014)
  17. International transmission of stock market movements (1989)
  18. Return and volatility linkages between the US and the German stock market (2006)
  19. Dynamic graph neural networks for enhanced volatility prediction in financial markets — → 2410.16858
  20. Multimodal deep learning for finance: integrating and forecasting international stock markets (2020)
  21. Financial machine learning (2023)
  22. Machine learning in business and finance: a literature review and research opportunities (2024)
  23. Machine learning for quantitative finance applications: A survey (2019)
  24. A review on graph neural network methods in financial applications (2021) — link
  25. Symmetry-Aware Graph Neural Approaches for Data-Efficient Return Prediction in International Financial Market Indices (2025)
  26. Financial time series forecasting with multi-modality graph neural network (2022)
  27. Cross-market volatility forecasting with attention-based spatial-temporal graph convolutional networks (2025)
  28. Cross-market risk spillovers among sovereign CDS, stock, foreign exchange and commodity markets: An interacting network perspective (2023)
  29. CSPO: Cross-Market Synergistic Stock Price Movement Forecasting with Pseudo-volatility Optimization (2025)
  30. Forecasting realized volatility with spillover effects: Perspectives from graph neural networks (2025)
  31. Selected Indian stock predictions using a hybrid ARIMA-GARCH model (2014)
  32. Stock price prediction using linear regression based on sentiment analysis (2015)
  33. Forecasting stock market movement direction with support vector machine (2005)
  34. Application of support vector machines in financial time series forecasting (2001)
  35. Forecasting stock returns based on information transmission across global markets using support vector machines (2016)
  36. Forecasting stock index movement: A comparison of support vector machines and random forest (2006)
  37. Integrating stock features and global information via large language models for enhanced stock return prediction — → 2310.05627
  38. What does chatgpt make of historical stock returns? extrapolation and miscalibration in llm stock return forecasts — → 2409.11540
  39. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead (2019)
  40. Deep learning in finance and banking: A literature review and classification (2020)
  41. Modeling the stock relation with graph network for overnight stock movement prediction (2021)
  42. Incorporating corporation relationship via graph convolutional neural networks for stock price prediction (2018)
  43. Multimodal multiscale dynamic graph convolution networks for stock price prediction (2024)
  44. Financial connectedness of BRICS and global sovereign bond markets (2018)
  45. A spatial analysis of international stock market linkages (2013)
  46. Volatility transmission in global financial markets (2015)
  47. Information transmission in informationally linked markets: Evidence from US and Chinese commodity futures markets (2011)
  48. Return and volatility spillovers across equity markets in mainland China, Hong Kong and the United States (2015)
  49. Chatgpt informed graph neural network for stock movement prediction — → 2306.03763
  50. A survey of large language models for financial applications: Progress, prospects and challenges — → 2406.11903
  51. Predicting stock returns: a risk measurement perspective (2021)
  52. Graph Machine Learning for Asset Pricing: Traversing the Supply Chain and Factor Zoo (2024)
  53. Spatial-temporal stock movement prediction and portfolio selection based on the semantic company relationship graph (2025)
  54. Modeling the momentum spillover effect for stock prediction via attribute-driven graph attention networks (2021)
  55. The physics of financial networks (2021)
  56. Financial risk forecasting with RGCT-prerisk: a relational graph and cross-temporal contrastive pretraining framework (2025)
  57. Enhancing Intraday Momentum Prediction: The Role of Volume-Based Information Uncertainty in the Chinese Stock Market (2026)

5 of 65 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.