Toward ESG Alpha: Analyzing ESG Exposures through a Factor Lens

Abstract

Using data on 1,312 active US equity mutual funds with $3.9 trillion in assets under management, we analyzed the link between funds’ bottom-up, holdings-based environmental, social, and governance (ESG) scores and funds’ active returns, style factor loadings, and alphas. We found that funds with high ESG scores have profiles of factor loadings that are different from those of low-scoring ESG funds. In particular, funds with high environmental scores tend to have high quality and momentum factor loadings. In partitioning the ESG scores into components that are related to factors and idiosyncratic components, we found strong positive relationships between fund alphas and factor ESG scores.

Disclosure: The authors report no conflicts of interest. The views expressed here are ours alone. This material is not intended to be relied upon as a forecast, research, or investment advice and is not a recommendation, offer, or solicitation to buy or sell any securities or to adopt any investment strategy. The authors have not received any outside funding for this research.

Editor’s Note

Submitted 1 June 2020

Accepted 18 August 2020 by Stephen J. Brown

This article was externally reviewed using our double-blind peer-review process. When the article was accepted for publication, the authors thanked the reviewers in their acknowledgments. Keith H. Black, CFA, and one anonymous reviewer were the reviewers for this article.

Acknowledgements

We thank Ron Kahn, Brian Deese, Andre Bertolotti, Mark Carhart, Giorgio DeSantis, and the editors Stephen J. Brown, Steven Thorley, CFA, and Heidi Raubenheimer, CFA, for their helpful comments.

Notes

¹ Biehl, Hoepner, and Liu (2012) documented several notable examples of ESG investing at the beginning of the 20th century—in particular, the Methodist Church and Quakers avoiding “sin” stocks. A wave of ESG investing occurred during the 1960s as investors debated and disinvested from companies engaging with apartheid, experiencing labor issues, and violating civil rights. Thus, ESG factors have long been considered by investors, but only recently have these issues been endorsed by the largest investors and corporations.

² See, for example, the 2020 letter by BlackRock’s CEO at www.blackrock.com/corporate/ investor-relations/larry-fink-ceo-letter. The 2019 statement by the Business Roundtable is available at https://opportunity.businessroundtable.org/ourcommitment. Hartzmark and Sussman (2019) studied the natural experiment of Morningstar’s adoption of ESG scores in active mutual funds for flows and performance.

³ See Clark, Feiner, and Viehs (2015), Friede, Busch, and Bassen (2015), and Gerard (2018) for reviews of past research. In addition to the references in the main text, Hoepner, Oikonomou, Sautner, Starks, and Zhou (2019) showed that ESG engagement reduces company downside risk and exposure to a downside-risk factor, and Ilhan, Sautner, and Vilkov (2019) showed that companies with higher carbon emissions have greater tail risk and are more volatile than companies with low emissions. Several individual metrics often used in the construction of ESG metrics have been linked to returns and risk. Companies with higher employee satisfaction (Edmans 2011), strong shareholder rights (Gompers, Ishii, and Metrick 2003), and strong corporate culture (Guiso, Sapienza, and Zingales 2015) outperform. Bolton and Kacperczyk (2020) found that stocks of firms with higher total CO₂ emissions earn higher returns after controlling for size and other factors and concluded that investors demand a risk premium for their exposure to carbon emission risk.

⁴ Seminal studies are Banz (1981) for size, Basu (1977) for value, Jegadeesh and Titman (1993) for momentum, Sloan (1996) for quality, and Ang, Hodrick, Xing, and Zhang (2006) for minimum volatility.

⁵ Ang (2014) provided a detailed summary of this research.

⁶ See also Brinson and Fachler (1985), Brinson, Hood, and Beebower (1995), Grinold (2006), and Cremers and Petajisto (2009). Dynamic timing by managers can confound traditional regression-based approaches that treat factor loadings as constant, as shown by Henriksson (1984) and Henriksson and Merton (1981). More recently, Chaudhuri and Lo (2018) used spectral analysis to decompose the return from factor timing further into its frequency components.

⁷ For further details, see www.msci.com/documents/1296102/15388113/MSCI+ESG+Fund+Ratings+Exec+Summary+ Methodology.pdf/ec622acc-42a7-158f-6a47-ed7aa4503d4f?t=1562690846881.

⁸ The French data library is found at https://mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html.

⁹ The Herfindahl–Hirschman Index (HHI) is widely used in economics to measure industry concentration and the distribution of individual attributes, such as income inequality. Its inverse reflects breadth. For example, suppose the HHI for market share in a given industry is 0.25. One way to interpret this number is that it corresponds to that of an industry with just 4 (= 1/0.25) equally sized companies (because 4 × 0.25² = 0.25).

¹⁰ Holdings-based characteristics are well suited to capturing dynamic effects—especially as many authors, including Henriksson and Merton (1981), have documented the existence of skilled managers who dynamically change factor loadings in response to changing economic environments. Return-based attribution methods can be adjusted for dynamic portfolio changes: Treynor and Mazuy (1966), Henriksson and Merton, and Henriksson (1984) added nonlinear terms to a time-series regression to capture market-timing components, but these approaches do not yield estimates of time-varying factor loadings.