Classification of Extremal Dependence in Financial Markets via Bootstrap Inference

2025-06-05

AI Summary

This study addresses the challenge of identifying extreme dependence structures in multivariate heavy-tailed financial data. We propose a bootstrap-based statistical inference framework integrating extreme value theory (EVT), pairwise extremal dependence measures ($hat{chi}$, $hat{eta}$), clustering analysis, and significance testing. To our knowledge, this is the first method enabling statistically testable, cross-market and sector-wise clustering of extremal dependence. Empirical analysis on absolute log-returns of U.S. and Chinese A-share equities reveals: (i) U.S. equities exhibit isolated dependence clusters, whereas A-shares show stronger intra-market extremal dependence; (ii) Materials, Consumer Staples, and Consumer Discretionary sectors display significant cross-market extremal co-movements. By moving beyond conventional static dependence modeling, our approach provides a rigorously testable, quantitative tool for systemic risk monitoring and cross-market risk management.

Technology Category

• Machine Learning: Multi-class/Multi-label Learning & Extreme Classification
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Game Theory and Economic Paradigms: Other Foundations of Game Theory & Economic Paradigms

Application Category

• Web Mining and Content Analysis: Normalization, clustering, classification, and summarization of Web text
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Security and Privacy: Large-scale security measurements

Abstract

Accurately identifying the extremal dependence structure in multivariate heavy-tailed data is a fundamental yet challenging task, particularly in financial applications. Following a recently proposed bootstrap-based testing procedure, we apply the methodology to absolute log returns of U.S. S&P 500 and Chinese A-share stocks over a time period well before the U.S. election in 2024. The procedure reveals more isolated clustering of dependent assets in the U.S. economy compared with China which exhibits different characteristics and a more interconnected pattern of extremal dependence. Cross-market analysis identifies strong extremal linkages in sectors such as materials, consumer staples and consumer discretionary, highlighting the effectiveness of the testing procedure for large-scale empirical applications.

Problem

• Identifying extremal dependence in heavy-tailed financial data
• Comparing US and China stock market dependence structures
• Analyzing cross-market extremal linkages in specific sectors

Innovation

• Bootstrap-based testing for extremal dependence
• Analysis of U.S. and Chinese stock markets
• Identifies cross-market sectoral extremal linkages