Bad ESG News Predicts Poor Stock Performance

New research shows that a company’s history of bad environmental, social and governance (ESG) events foretells poor stock performance. That contrasts with other research showing that ESG ratings lack such predictive power.

In his November 2017 study, “ESG Risks and the Cross-Section of Stock Returns,” Simon Glossner found that weak corporate social responsibility destroys shareholder value, and stock prices fail to fully incorporate the consequences of intangible risks – there is persistence in ESG scores, as firms with high ESG risks have more ESG issues in the next year than firms with low or medium ESG risks, and shorting firms with high ESG risks is a profitable strategy. His finding that investors persistently underreact to negative news is a well-documented behavioral bias, with the explanations for slowly incorporating information into prices including optimism, anchoring and confirmation biases. Each of these can cause investors to underweight or ignore contrarian information. Glossner noted that the literature also demonstrates that anomalies often occur in intangibles (such as accruals, R&D research and patents) – and ESG risks are intangible risks.

Glossner followed up his work with the February 2021 study, “ESG Incidents and Shareholder Value.” He used a dataset of negative incident news to examine poor ESG practices and answer the following questions: “Is a firm’s incident history predictive of future ESG incidents? How do high incident rates impact firm value over time? Do stock markets properly price incident-based ESG information? If not, what friction prevents investors from doing so?” He measured a firm’s past incident rate to quantify the frequency and severity of past incident news (e.g., environmental pollution, poor employment conditions or anticompetitive practices). The incident news came from RepRisk. His sample had over 80,000 incident new reports for about 2,900 unique firms traded on U.S. stock markets between 2007 and 2017. Glossner created portfolios with low (0-25), medium (26-50) and high (51-100) incident scores.

He began by noting that incident-based ESG measures have two main advantages over conventional ESG ratings used in previous research: “First and conceptually, an incident measure captures poor ESG practices in a direct way, through past realizations of ESG-related business risks and criticism raised by stakeholders themselves. An incident measure could therefore be a better predictor of future incidents than conventional ESG ratings. The problem with conventional ESG ratings is that it is unclear what they measure. Conventional ratings aggregate hundreds of ESG criteria into one company score using very different approaches. As a result, those ratings often disagree with each other because the news character of incidents allows conducting event studies. Second, and empirically, with an incident measure, it is easier to identify its value implications.” Following is a summary of his findings: