中国股票行业与全球企业ESG表现的多重互相关分析

Multifractal Cross-Correlations Analysis between Chinese Stock Sectors and Global Corporate ESG Performance

摘要: 本文研究了中国股票行业与全球企业ESG表现的多重互相关关系，使用MF-DFA法和MF-DCCA法进行多重分形特征和互相关关系分析。该分析包括沪深300行业指数及S&P全球ESG指数的多重分形特征、沪深300行业指数与S&P全球ESG指数之间的多重互相关关系，以及相位重排和随机重组后的序列对分析。结果表明，所有序列均存在显著的多尺度多重分形特征，信息指数的多重分形强度最强，消费指数的多重分形强度最弱；中国股票各行业与全球企业ESG表现之间存在显著的多重分形互相关关系，较大时间尺度下的材料行业与全球ESG表现之间的互相关关系最强，而所有时间尺度下的消费、医疗行业与全球ESG表现之间的互相关关系较弱；信息行业与全球ESG表现的互相关关系最复杂，而公用事业行业与全球ESG表现的互相关关系最简单；相位重排和随机重组后的序列对仍具有显著的多重分形特征，但多重分形程度有所降低，厚尾分布是序列对多重分形特征的主要来源。本文的工作可为ESG投资者提供行业选择的策略建议，为监管机构提供行业规范依据，并推动企业的可持续发展实现。

Abstract: This research utilizes multifractal detrended fluctuation analysis (MF-DFA) and multifractal detrended cross-correlation analysis (MF-DCCA) to examine the multifractal properties and cross-correlation relationships between Chinese sector indices and global corporate Environmental, Social, and Governance (ESG) performance. The analytical scope includes: (1) multifractal properties exhibited by CSI 300 sector indices and the S&P World ESG Index;(2) the multifractal cross-correlations between these Chinese sector indices and the global ESG benchmark; (3) the analysis of phase-randomized and shuffled time series. Findings reveal significant multi-scale multifractal features across all sequences. Specifically, the Information Technology Index (ITI) demonstrates the most pronounced multifractal strength, while the Consumer Staples Index (CSI) exhibits the weakest characteristics. A notable multifractal cross-correlation exists between Chinese stock sectors and global ESG performance. The material sector shows the strongest cross-correlation at greater time scales, whereas the consumer and medical sectors display weaker linkages across all time frames. The information sector exhibits the most complex cross-correlation pattern with global ESG performance, in contrast to the relatively simple relationship observed in the utilities sector. Analyses of phase-randomized and shuffled series indicate that although multifractality is reduced, significant multifractal features persist. Fat-tailed distributions are identified as the primary contributor to the multifractal nature of these sequences. This study offers regulatory bodies targeted policy insights and provides practical guidance for enhancing corporate sustainability practices.