ESG (Environmental, Social, and Governance) investing—alternatively, socially responsible investing or sustainable investing—is growing rapidly in popularity. In the fourth quarter of 2019, the assets of ESG-based mutual funds stood globally at nearly $1 trillion. Just two years later, in the fourth quarter of 2021, those assets stood at nearly $3 trillion. That growth has created controversy: In the states of Florida and Texas, limits have been put forth on ESG investing, with the Florida resolution stating that the state would invest “without consideration for nonpecuniary beliefs.”
This controversy raises a natural question: is ESG investing only about aligning investments with nonpecuniary beliefs or is ESG investing spotlighting oft ignored, but financially material, information?
The existing research on ESG with respect to mutual funds has primarily focused on whether funds incorporate ESG considerations into their investment process and, if they do, whether investing in stocks with high or low ESG ratings impacts their performance (i.e., whether ESG ratings themselves are financially material). The primary contribution of our work is to investigate more broadly—focusing on the materiality of the information underlying ESG ratings. We do so by measuring the extent to which fund managers incorporate ESG information into their portfolio decisions.
We measure the extent of ESG information incorporation through a novel metric we label ‘Active ESG Share.’ Importantly, a fund can be active with respect to ESG information by investing more in stocks with higher ESG ratings or by investing more in stocks with lower ESG ratings. Therefore, a fund manager with a high Active ESG Share—that is, one aggressively using ESG information—could have a portfolio of stocks that tend to have high or low ESG ratings.
Our hypothesis is not that high Active ESG Share is unconditionally beneficial. Rather, we contend that the material component of ESG information is complex, such that processing it effectively should require specialization. Accordingly, we expect that the impact on fund performance of Active ESG Share will be concentrated among ESG funds, who presumably have managers specialized in processing ESG information. In addition, ESG rating providers often disagree on their ratings. We expect the impact of Active ESG Share on performance to be stronger among funds that tend to hold stocks with greater rating disagreement, since those stocks offer greater opportunity to identify and utilize material information.
We measure Active ESG Share over the period 2004 through 2021 for a large sample of actively managed U.S. equity mutual funds. We classify a fund as ESG focused if Morningstar identifies it as such or if a fund has certain key terms in its name (e.g., ‘climate’ or ‘social’). Our sample of actively managed funds contains 243 ESG funds and 1,875 non-ESG funds. Consistent with the growth we discussed earlier, the proportion of ESG funds in our sample increases over time, from 10% of funds and 15% of assets in 2004 to 18% of funds and 20% of assets in 2021.
Our empirical evidence within this sample is consistent with our expectations. On average, Active ESG Share does not have a significant relation with future fund performance. Higher Active ESG Share does, however, predict better future performance among ESG funds. A one standard deviation increase in Active ESG Share for ESG funds predicts a performance increase of about 0.57% per year. Furthermore, that impact is concentrated among those ESG funds that tend to buy stocks with a high level of ratings disagreement. If the Active ESG Share of an ESG fund with that tendency increases by one standard deviation, our model predicts performance will improve, on average, by about 0.87% per year.
A feature of ESG investing that adds to its complexity is its aggregation of disparate components. How a firm approaches their carbon emissions (environmental) and whether a firm staggers its board (governance) have little relation, but those decisions both fall under the heading of ESG investing. We further detail the relation between Active ESG Share and performance by disaggregating E, S, and G. We find that, among ESG funds, the environmental component is the most impactful. A one standard deviation increase in Active E Share predicts a performance improvement of about 0.65% per year, while analogous increases for Active S Share and Active G Share predict performance improvements of 0.37% and 0.32% per year, respectively.
Our results with respect to the average ESG ratings of the stocks held by these funds are also consistent with the idea that ESG information is complex. For non-ESG funds, there is no relation between average ESG rating and future fund performance. For ESG funds, a one standard deviation increase in average rating predicts that performance will decrease by about 0.55% per year. From a financial perspective, ESG ratings themselves appear to not provide useful information, but rather serve as a means of coordinating the trading of ESG funds, leading to stock overpricing and fund underperformance.
In conclusion, our results from studying actively managed mutual funds support the hypothesis that ESG information is financially material, but complex. Specialized fund managers can incorporate such information into their investment process to the benefit of their investors, especially when investing in stocks with a high level of disagreement in ESG ratings. Thus, regardless of an investor’s nonpecuniary beliefs, ESG information should not be ignored.
