Why is Corporate Virtue in the Eye of the Beholder? The Case of ESG Ratings

Dane Christensen is an Associate Professor in the School of Accounting and Charles E. Kern Research Scholar at the University of Oregon Lundquist College of Business; George Serafeim is the Charles M. Williams Professor of Business Administration and the Faculty Chair of the Impact-Weighted Accounts Project at Harvard Business School; and Anywhere Sikochi is an Assistant Professor of Business Administration at Harvard Business School. This post is based on their recent paper, forthcoming in the Accounting Review. Related research from the Program on Corporate Governance includes The Illusory Promise of Stakeholder Governance by Lucian A. Bebchuk and Roberto Tallarita (discussed on the Forum here); For Whom Corporate Leaders Bargain by Lucian A. Bebchuk, Kobi Kastiel, and Roberto Tallarita (discussed on the Forum here); and Restoration: The Role Stakeholder Governance Must Play in Recreating a Fair and Sustainable American Economy—A Reply to Professor Rock by Leo E. Strine, Jr. (discussed on the Forum here).

Despite the rising use of environmental, social, and governance (ESG) ratings, there is substantial disagreement across rating agencies regarding what rating to give to individual firms. This is highly problematic because in the absence of agreement on what good ESG performance constitutes, market participants might be misled by ESG ratings. We also have very little evidence on why providers disagree so much. Without understanding the reasons for this disagreement, it is difficult to understand not only what the potential remedies could be, but also the plausible consequences of this disagreement.

To address this, we focus on a key firm attribute that we hypothesize is likely to be of first-order importance in causing providers to disagree on ESG ratings. Specifically, we focus on the extent of a firm’s ESG disclosure, as theory suggests that disagreement arises due to people having different information sets and/or different interpretations of information. Conventional wisdom and a plethora of evidence in other settings suggest that higher disclosure helps lower disagreement by reducing differences in the information that people have. We argue that in our setting, due to the subjective nature of ESG information, the opposite would be true, where higher disclosure would increase disagreement, as disclosure expands opportunities for different interpretations of information.

For companies with higher levels of disclosure, data providers need to make a judgment about whether the disclosure means good or bad performance. For example, a company that discloses lost-time injury rates needs to be judged based on this disclosure. This gives rise to a level of subjectivity that leads to higher levels of disagreement. We argue that this subjectivity increases as firms expand their disclosures. This prediction is consistent with arguments from sociology, which theorize that a plurality of evaluations is likely to occur in newly emerging fields where rules and norms for evaluation are less developed. Higher disclosure also increases the likelihood that the ESG rating agencies might be able to use different metrics to evaluate a firm’s performance on the same issue and therefore lead to greater rating disagreement. Moreover, as in financial markets, evaluators may disagree over which measures are more relevant to assessing ESG performance. Collectively, these arguments suggest that greater ESG disclosure would result in greater ESG rating disagreement.

Primary Results

Using data on firms from 69 countries over the years 2004 to 2016, we find strong support for our prediction that greater ESG disclosure leads to greater ESG rating disagreement. We analyzed data on ESG disclosure scores from Bloomberg and ESG performance from three of the largest providers of ESG ratings to investors: MSCI, Thomson Reuters, and Sustainalytics. To help corroborate our findings, we exploited the adoption of broad mandatory ESG disclosure requirements across countries. We find that after a country or stock exchange implements mandatory ESG disclosure requirements, the affected firms increase their ESG disclosures and experience greater ESG rating disagreement.

Input and Outcome Metrics

We enhance our understanding of what ESG ratings capture and how disclosure contributes to disagreement by examining individual metrics that ESG raters use to construct their overall ESG ratings. We categorize the metrics into inputs and outcomes. Inputs refer to efforts that a company is making to achieve a desired outcome (e.g., the presence of a diversity policy) and outcomes refer to actual performance outcomes (e.g., the percentage of women employed at the company). We predict and find that ESG raters tend to disagree less about ESG inputs and more about ESG outcomes. This is consistent with the notion that the evaluation of outcomes is more subjective and relies on having a shared understanding of what a good versus a bad outcome might be. Further, we predict and find that greater ESG disclosure exacerbates these disagreements, especially in the case of outcomes, consistent with our argument that more pieces of information that require subjective evaluation should lead to greater disagreement.

Stock Market and Financing Consequences

We also explore market consequences of ESG rating disagreement. Using short window tests around the release of revised ESG ratings, we find greater ESG disagreement is associated with higher stock return volatility and larger absolute price movements. These findings suggest that ESG disagreement is relevant to market participants and influences stock prices. We also find some evidence that these results are becoming even stronger over time, which suggests that ESG disagreement is having an increasing impact on markets. In addition, we explore the influence of ESG disagreement on firms’ financing choices. Consistent with ESG disagreement creating market frictions by introducing uncertainty regarding a firm’s long-term sustainability, we find that firms with greater ESG disagreement are less likely to raise external financing and instead tend to rely more on internal financing.

Conclusions and Implications

Overall, our results indicate that ESG disagreement is most pronounced for firms with high levels of ESG disclosure, thereby shedding light on a key driver of ESG rating disagreement. Thus, contrary to evidence in these settings that disclosure reduces disagreement, in our setting, disagreement is larger when firms have higher levels of disclosure. In our view, this highlights the importance of developing a shared understanding of a) what constitutes good or bad ESG performance, and b) what metrics to use to capture ESG performance, as preconditions for ESG disclosure to decrease disagreement.

Our findings also illuminate some of the challenges ESG rating agencies face as information intermediaries. While thousands of companies now claim to integrate ESG issues in their business strategy and operations, it is not clear whether those claims are merely ‘cheap talk’. In the presence of incomplete information about a firm’s ESG performance, ESG ratings can perform a significant information intermediary function by helping investors and other stakeholders to choose companies that exhibit their preferred ESG outcomes. In particular, having rating agencies focus on ESG outcomes might be desirable to mitigate ‘cheap talk’ by companies, as a company would need to show real effects (e.g., reductions in carbon emissions, improvements in lost time injury rates) instead of disclosing the adoption of a policy or initiative that might generate no real effects. However, our results suggest that ESG rating agencies have a more difficult time agreeing when evaluating a company on outcomes rather than input metrics, and that disclosure exacerbates disagreement on outcomes even more than it does on inputs. This lack of consensus about how to interpret outcomes might be an obstacle that encourages raters to focus more on inputs and thereby causes potential damage to the corporate accountability function that ratings could perform. In this context, we regard disagreement as inhibiting the accountability process. Moreover, while disclosure may have many positive effects, it likely needs to be placed in a framework that allows analysts to evaluate those outcomes with clear benchmarks.

Overall, given concerns over ESG rating disagreement, our findings suggest a lot of work still needs to be done to develop rules and norms to determine what characterizes good ESG performance.

The complete paper is available for download here.

Both comments and trackbacks are currently closed.