DeltaLag: Learning Dynamic Lead-Lag Patterns in Financial Markets

2025-11-01

AI Summary

Traditional lead–lag detection methods assume static, time-invariant relationships, failing to capture the dynamically evolving lead–lag structures inherent in financial markets. To address this, we propose the first end-to-end deep learning framework that adaptively identifies time-varying lag lengths between asset pairs via a sparsified cross-attention mechanism—thereby relaxing the restrictive static-lag assumption. Subsequently, features of leading assets are temporally aligned according to the estimated lags to model their predictive power for followers’ future returns. Our approach achieves both strong forecasting performance and interpretability: empirical results across multiple markets and horizons demonstrate statistically significant outperformance over fixed-lag baselines, vector autoregression (VAR), and graph neural networks; trading strategies based on our model yield Sharpe ratio improvements of 18%–32%. Moreover, the framework explicitly reveals key driver–follower asset pairs and their evolving lag trajectories.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Search and Optimization: Learning to Search
• Computer Vision: Large Vision Models

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

The lead-lag effect, where the price movement of one asset systematically precedes that of another, has been widely observed in financial markets and conveys valuable predictive signals for trading. However, traditional lead-lag detection methods are limited by their reliance on statistical analysis methods and by the assumption of persistent lead-lag patterns, which are often invalid in dynamic market conditions. In this paper, we propose extbf{DeltaLag}, the first end-to-end deep learning method that discovers and exploits dynamic lead-lag structures with pair-specific lag values in financial markets for portfolio construction. Specifically, DeltaLag employs a sparsified cross-attention mechanism to identify relevant lead-lag pairs. These lead-lag signals are then leveraged to extract lag-aligned raw features from the leading stocks for predicting the lagger stock's future return. Empirical evaluations show that DeltaLag substantially outperforms both fixed-lag and self-lead-lag baselines. In addition, its adaptive mechanism for identifying lead-lag relationships consistently surpasses precomputed lead-lag graphs based on statistical methods. Furthermore, DeltaLag outperforms a wide range of temporal and spatio-temporal deep learning models designed for stock prediction or time series forecasting, offering both better trading performance and enhanced interpretability.

Problem

• Detecting dynamic lead-lag patterns in financial markets
• Overcoming limitations of traditional statistical analysis methods
• Exploiting pair-specific lag values for portfolio construction

Innovation

• Uses sparsified cross-attention for lead-lag pairs
• Extracts lag-aligned features from leading stocks
• Learns dynamic lead-lag patterns end-to-end