Do ESG Mutual Funds Deliver on Their Promises?

ESG investing is growing explosively, and the interest in ESG investing by retail investors continues to increase. A substantial proportion of retail ESG investing occurs through ESG mutual funds. The number of ESG mutual funds, and the assets they hold have each doubled in the past three years, and the COVID-19 pandemic has done nothing to slow this trend. But does this rapidly growing sector of the investment industry actually deliver investment exposure to ESG goals, or has the demand for ESG investing led to overpriced, greenwashed funds that are merely marketed as ESG to chase the latest investment fad or extract higher fees from investors? While these concerns have attracted attention from both the Securities & Exchange Commission (“SEC”) and the Department of Labor (“DOL”), much of the regulatory conversation to date has relied on theoretical concerns and anecdotal evidence. This, combined with the rapidly evolving market for ESG funds, demonstrates the compelling need for greater empirical analysis directly targeting the regulators’ concerns.

Our paper provides that evidence. Using market-wide data on fund portfolios, voting, fees, and performance, we specifically target the concerns articulated by the SEC and the DOL. We combine detailed information on mutual funds with four proprietary datasets evaluating company-level ESG performance: ISS, S&P, Sustainalytics, and TruValue Labs. Using this unique and comprehensive dataset, we explore the practical differences between ESG and non-ESG funds as well as the differences among ESG funds along four dimensions—portfolio composition, voting behavior, costs, and performance. Our goal is to provide an overview of the market as it currently stands for the purpose of informing regulatory initiatives that have the potential to reshape the ESG landscape.

Our analysis seeks to answer two questions. The first question, which drives the SEC’s analysis, is what investors are getting for their “ESG dollars.” To answer this question, we survey the existing market and construct several categories of ESG mutual funds—funds with names that convey an ESG-oriented strategy, funds classified by Morningstar as ESG funds, and funds that purport to consider ESG factors in their investment criteria. We then analyze the portfolio composition and voting behavior of these funds to compare them across multiple dimensions. We find evidence that ESG funds differ from other funds along both dimensions—they hold portfolios tilted toward companies with higher ESG ratings and they are more supportive of ESG voting issues. The second question, critical to the DOL, is what, if anything, investors are giving up when they invest in ESG funds. The costs associated with investing in ESG funds can be direct, in the form of fees, or indirect, in the form of lower raw or risk-adjusted returns. Here, although our analysis is limited to a relatively short timeframe, we find no evidence that ESG funds cost investors more or that they underperform their non-ESG counterparts.

Descriptively, we uncover an evolving landscape of ESG funds. Today’s ESG funds range from single issue funds that address water conservation or religious values to those that incorporate screening criteria into the construction of a broad-based index. We find extensive disclosures of fund investment strategies—strategies that range from stock selection to impact investing—as well as the extent to which the fund incorporates ESG considerations into voting and engagement. We find, in short, a market that recognizes that ESG means different things to different investors and is responding to those differences with an array of distinct products.

Our analysis focuses on the 2018-2019 timeframe, the most recent years for which we have data available. We construct our sample of ESG funds in two ways, the first by screening funds according to their names, and the second through those funds identified as ESG funds by Morningstar. We focus, for this paper on equity funds. This results in a sample of 303 funds.

To evaluate fund composition, we use data from four separate rating providers to calculate ESG scores, to determine the ESG scores of the fund’s portfolio companies. Using these scores, we calculate what we term a fund’s “ESG tilt”—the asset-weighted average of the ESG scores. We find that funds that identify themselves as ESG funds hold portfolios that have a significantly larger ESG tilt than non-ESG funds. In other words, contrary to the SEC’s concern about “greenwashing,” ESG funds deliver on their promise to invest differently from other funds.

Although ESG mutual funds have been criticized for not casting their portfolio-company votes in accordance with their investment profiles, we document clear differences between the voting behavior of ESG and non-ESG funds. We observe, in particular, that ESG funds vote differently on environmental and social issues; although ESG funds do not automatically support every shareholder proposal related to ESG, they do vote more independently of management compared to other funds. With respect to certain governance issues, such as say on pay, we also find clear differences. In short, ESG funds appear to be considering ESG criteria in voting, as well as investment decisions.

We next look at what these differences cost investors. To do so, we investigate the expenses associated with ESG funds and the returns offered by these funds. Contrary to the concern articulated by the DOL, we find no evidence that ESG funds cost more than comparable non-ESG funds, or that they offer inferior performance during our sample period (either raw or risk adjusted). The results persist despite the inclusion of a battery of control variables intended to ensure that we are making “apples-to-apples” comparisons. While these tests are not intended to establish—nor can they establish—whether or not ESG funds are a “good” investment, we find no evidence that they perform worse than comparable funds.

A final empirical contribution of this paper is to address the impact of variation in ESG ratings. As the SEC has noted, there is little consensus on what falls within the definition of ESG or how to weigh various ESG considerations. There are over 600 ESG rating providers, and these providers rely on a range of different sources of data and employ a variety of methodologies to analyze that data. Commentators have highlighted the fact that these differences frequently lead to different ratings. Thus, for example, among automobile manufacturers, Tesla receives a top ESG rating from MCSI and a bottom rating from FTSE. Although we do not directly interrogate differences among providers in this paper, we take the unique approach of incorporating ESG rating data from four different and well-known providers to measure the ESG orientation of the mutual fund portfolios that we examine. We find that, although the providers take very different approaches to measuring ESG, the overall pattern is remarkably stable across providers.

In sum, we provide new data on the role of ESG in mutual fund investing and its effects. Our goal in this paper is modest. We do not seek to establish that ESG funds are good with respect to any specific benchmark. Rather, the goal of this paper is to address concerns that ESG funds present special regulatory concerns relative to the mutual fund market as a whole, either because (as the SEC fears) they are not doing what they purport to do or because (as the DOL fears) their economic performance is inferior to non-ESG funds. Our empirical results provide powerful evidence that ESG funds are offering investors something different from traditional funds with respect to both portfolio composition and voting, and that they are doing so without causing investors systematically to sacrifice economic performance. In short, the ESG fund market appears to be working relatively well. We therefore challenge the claim that ESG mutual funds warrant distinctive regulatory treatment and propose instead that regulators adopt a policy of “ESG neutrality.”

The complete paper is available for download here.