ESG ratings are a new type of performance evaluation like analyst forecasts or bond ratings Image: Shutterstock
Investment funds that prioritize environmental, social, or governance considerations have proliferated in recent decades.
Investment managers who employ these strategies (which are often gathered under the acronym “ESG”) might, for example, base investment decisions on companies’ carbon output, their employees’ safety on the job, or the percentage of women and minority leaders in top governance positions.
Fund managers who want to build strategies around such granular details need to interpret reams of data and parse ever-updating news and corporate reports. So more and more research firms are now addressing this growing need.
Some of these firms generate “ESG ratings”—a number (typically between 0 and 100) applied to individual companies that attempts to quantify their overall ESG performance. Each rating takes into account a large amount of information across various categories, and each firm producing such ratings uses a different formula. Sometimes, the result is that one company will be stamped with wildly divergent ESG ratings, leading some in the industry to complain that the ratings may be meaningless.
Others argue that, used correctly, the ratings could be a key to smarter investing.
Aaron Yoon, an assistant professor of accounting and information management at the Kellogg School, wanted to challenge the claims on both sides. On one hand, it was possible the ESG ratings didn’t hold water; on the other, it could be that investment managers were giving up on the ratings—and the broader ESG strategy—too easily.
“ESG ratings are a new type of performance evaluation like analyst forecasts or bond ratings,” Yoon says, “and they’ve never been really assessed because there hasn’t been a clear outcome variable.” In other words, while bond ratings and analyst forecasts make straightforward predictions that will eventually be judged against a specific future outcome, ESG ratings aren’t as easy to view as accurate or dead wrong, even with the benefit of hindsight.
In a new paper, Yoon and coauthor George Serafeim, at Harvard Business School, find a way to hold ESG ratings accountable. They test whether the ratings predict future ESG-relevant news about companies—whether this be good news, like a new eco-friendly initiative, or bad news, like a factory accident that injures workers.
They find that consensus ESG ratings do predict good and bad future ESG news, though this predictive ability weakens as disagreement between raters on a given company rises. However, the researchers come up with a profitable investment strategy that takes advantage of these disagreements.
In his previous research into the fast-growing ESG investment space, Yoon’s findings haven’t always reflected well on the industry. This newest paper, however, lends some credibility to the practice of applying and using ESG ratings.
“There’s a huge wave of money flowing into the ESG space,” Yoon says. “My research agenda is, how do we account for ESG? How do we quantify it so that it can be useful to investors? Or, to take it a step further, how can ESG ratings be used in a systematic way, to generate strong returns?”
Examining ESG Ratings’ predictive power
The researchers concentrate on the three ESG ratings providers with the most comprehensive coverage—MSCI, Sustainalytics, and Thomson Reuters.
To test whether the ratings predict future news, they average the three firms’ ratings to obtain a single ESG score for each individual company. Yoon and Serafeim then turn to a dataset from TruValue labs, a firm that tracks ESG-related news across thousands of companies and uses natural language processing to place the news on a spectrum of positive to negative. The TrueValue dataset summarized 31,854 firm-level ESG news observations between January 2010 and June 2018.
Comparing the two sets of data revealed that, indeed, ESG ratings were strongly correlated to the sentiment of future ESG news, with high ESG ratings predicting positive news stories and low ratings predicting negative ones.
However, that correlation started to diminish as soon as consensus among raters did; in the presence of significant disagreement, the predictive power of the ratings was much weaker.
The researchers then examined whether ESG ratings—and the level of disagreement among raters—had any effect on stock prices. To explore this, they first looked at reactions to ESG news by comparing a company’s share price one day prior to the news to the price one day after. As they expected, stock prices did respond—positively to good ESG news, and negatively to bad ESG news.
The next step was investigating whether ESG ratings seemed to change the way the good or bad ESG news was absorbed into prices. The researchers found that firms with low average ESG scores saw a bigger stock-price jump following positive ESG news (75 basis points, on average) than those with high ratings (only 34 basis points).
The researchers theorize that this is because companies with high ESG scores already had this good sentiment reflected in their stock prices, and more good news simply confirmed the market’s preconceptions.
The researchers took this finding a step further, by asking how an investment strategy might be designed to take advantage of the predictive abilities of ESG ratings.
First, they established which of the three ESG raters in their sample managed to most accurately forecast future ESG news—and found that this distinction went to MSCI. Next, they created a hypothetical investment portfolio in which they bought companies whose MSCI ESG rating was high and whose ESG ratings from Sustainalytics and Thomson Reuters were comparatively low. (They also did the inverse, selling firms whose MSCI rating was low and whose average of the other two ratings was high.) This high level of disagreement among raters—in tandem with a strong opinion from MSCI— was crucial to the strategy because it suggested that these may be the companies whose ESG performance wasn’t yet reflected in their stock prices.
The portfolio did indeed generate annualized alpha (that is, above-market returns) of 4.27 percent, indicating that the researchers’ strategy unearthed a real pattern surrounding ESG ratings and market response.
A step toward more credibility
Investors who can swiftly take advantage of these discrepancies are poised to profit.
“The general complaint in the industry and in academia is that ESG ratings can have huge discrepancies among them,” Yoon says. “We’re saying, ‘this is how you can potentially take advantage of that.’”
But the even bigger takeaway, Yoon adds, concerns the broader debate about how to use ESG ratings—and whether they’re worth using at all.
“We find that ESG ratings do predict future ESG news,” he says, speculating that the findings will increase the ratings’ credibility.