A Multidimensional Machine Learning Study of Environmental Innovation and ESG Integration in BRICS Economies
利用2009-2023年金砖国家企业数据，通过弹性网络回归和集成学习模型，发现环境创新正向影响ESG绩效，且制度质量和CEO权力起调节作用，为企业与政策制定者提供参考。
ABSTRACT Environmental, social, and governance (ESG) performance has developed as a critical axis of business strategy, specifically within countries enduring institutional transformation and coping with extreme environmental exposures. This empirical study considers the degree to which environmental innovation improves ESG outcomes in BRICS economies (Brazil, Russia, India, China, and South Africa). Using a dataset from Refinitiv Eikon covering the 2009 to 2023 data, this study applies elastic net regression and a comprehensive stacked ensemble learning framework combining random forest , gradient boosting , ridge regression , and support vector machine (SVM) models to evaluate the predictive capability of environmental innovation with governance and firm‐specific determinants. We find that environmental innovation positively influences ESG performance, suggesting environmental innovation is not only strategic, but also an important driver of corporate sustainability initiatives in emerging markets. Results reveal that institutional quality and CEO power moderate the relationship between environmental innovation and ESG. These insights feature the economically significant and contingent value of sustainability investments and offer actionable inferences for corporate leaders and policymakers.