ESG Rating Disagreement: An Investigation of Alpha in the Presence of ESG Rating Disagreement

Abstract

This thesis demonstrates the variations in ESG scores across five well-established rating providers and analyses the implications of ESG rating disagreement on stock returns. Effectively, the study highlights the adverse effects of conflicting ESG ratings. By deriving a proxy for rating disagreement based on ESG ratings from Bloomberg, S&P Global, Refinitiv, Sustainalytics, and RepRisk, this study empirically tests the performance of long-short strategies in high- and low-disagreement portfolios, respectively. The CAPM, the Fama & French three-factor model, the Carhart four-factor model, and the Fama & French five-factor model are applied for this assessment. The portfolios consist of stocks in the S&P 500 and the period analyzed is January 2010 to December 2021. Regarding the Total ESG score, the authors find that when going long in a high-disagreement portfolio and short in a low-disagreement portfolio, a significant and positive alpha is generated. Specifically, the positive alpha is presumed to reflect an uncertainty premium due to uncertainty about a company’s “true” ESG performance. However, results diverge when testing the same strategy for the Environmental, Social, and Governance scores. Specifically, the analysis conducted on disagreement in the Environmental score returns an insignificant alpha, the analysis on Social returns a negative and significant alpha, and the analysis on Governance returns a positive and significant alpha. Furthermore, the authors reveal a diminishing alpha by applying a rolling window regression, suggesting that ESG ratings move towards lower levels of ESG rating disagreement (i.e., higher agreement). The authors thereby advocate concerns related to the sensitivity of the selected rating provider in academic studies and stress the necessity for a unified reporting framework, facilitating allocation into “truly” sustainable firms. By uncovering gaps, inconsistencies, and data limitations in the literature, this thesis builds grounds for further research on ESG rating disagree.