Why M&As Cannot Cheat on Sustainability Anymore
用机器学习模型预测并购后企业的可持续性表现，发现并购前ESG评分越高，并购后反而更难维持，且可持续性评分会随社会规范、技术等变化而漂移。
ABSTRACT The paper develops a model based on advanced techniques, such as machine learning, capable of predicting ex‐ante the sustainable outcomes that a company can achieve through the acquisition of a specific target. A comprehensive analysis about various indicators related to bidders' and targets' sustainable performance is carried out to determine which have the most significant impact on M&A deals' sustainable outcomes. The findings show that higher predeal ESG, E, S, and G scores tend to negatively impact postdeal ESG performance, suggesting that high sustainability scores may lead to challenges in maintaining these levels after a significant structural change. Furthermore, our results indicate the presence of a concept drift, determined by the intrinsic nature of sustainability scores, which are inherently influenced by evolving societal norms, technological advancements, economic shifts, and regulatory updates. Adding more recent observations to the training set enhances predictive capability.