How climate policy uncertainty influences corporate social responsibility: empirical evidence from China’s listed companies
摘要
Against the backdrop of deepening global climate governance, how climate policy uncertainty (CPU) affects firms’ sustainable behavior has become a critical issue for understanding low-carbon transition at the micro level. While corporate social responsibility (CSR) is widely viewed as an essential vehicle for promoting sustainable development, existing studies have yet to reach a consensus on how climate policy uncertainty (CPU) influences firms’ CSR engagement, particularly in developing economies. Moreover, empirical evidence on the underlying transmission mechanisms remains limited. Using a panel dataset of Chinese A-share listed firms covering the period from 2010 to 2023, we investigate the impact of CPU on corporate CSR performance, utilizing industry and year fixed effects. The results indicate that heightened CPU is associated with a significant decline in firms’ CSR performance, and this relationship remains robust to a battery of endogeneity and robustness checks. Further mechanism analyses suggest that CPU constrains CSR engagement primarily by increasing agency costs, crowding out research and development (R&D) investment, and weakening regional environmental regulation intensity. In addition, heterogeneity analyses reveal that the adverse effect of CPU is more pronounced for non-state-owned enterprises, firms operating in high-carbon-emission industries, and firms located in regions with relatively underdeveloped financial systems. This study extends the literature on the economic consequences of climate policy uncertainty to corporate sustainable behavior, offering new insights into how policy-related uncertainty shapes firms’ long-term responsibility commitments. The findings imply that greater policy consistency and regulatory credibility are essential for alleviating agency frictions, preserving firms’ innovation capacity, and supporting their long-term sustainable engagement.