To ban or not to ban? ESG is the question

In April 2023, Florida passed a bill banning state and local governments from using Environmental, Social and Governance (ESG) criteria when investing public money. The bill requires investment managers to restrict their investment decisions solely to pecuniary factors. Presumably, the framers of the legislation believe that ESG criteria have no pecuniary relevance because if they had some pecuniary relevance they wouldn’t ban them. In addition to Florida, 18 other states (Alabama, Alaska, Arkansas, Georgia, Idaho, Iowa, Mississippi, Missouri, Montana, Nebraska, New Hampshire, North Dakota, Oklahoma, South Dakota, Tennessee, Utah, West Virginia and Wyoming) have passed, or are considering passing, similar legislation. This is no small deal – these states have aggregate public pension assets of just under $1 trillion. 

The sheer abundance of available data means that the ability to demonstrate or to execute a technical investment edge is much less than has historically been the case. Therefore the ability to create investment portfolios beyond the constraints of the 16 factors in the pecuniary-based Barra Global Equity Factor Risk Model increasingly becomes a key performance factor for asset managers. This is not to say that pecuniary factors cannot create out-performance against a broad market index. They can. It is well understood that a portfolio quantitatively based on quality factors has historically outperformed broad indices. This is simply due to a quality factor index capturing more of the upside of a broad index and less of the downside. For example, the MSCI World Quality Net Total Return Index (M1WOQU) has historically outperformed the MSCI World Net Total Return Index (NDDUWI) fairly consistently.

The problem with pecuniary factors is that they are based on historical data. And as the US Securities and Exchange Commission (SEC) and UK’s Financial Conduct Authority (FCA) remind us “past performance is not a reliable indicator of future results”[1] . Future results are driven by a multitude of factors, some of them qualitative, not pecuniary or quantitative. An example is brand value; a corporate with a better brand is likely to produce better results than a peer with a lesser brand. For example, the food and beverage company with the highest brand value[2], The Coca-Cola Co (XNYS: KO) has better basic metrics than those peers with lower brand value, e.g. National Beverage Corp (XNGS: FIZZ).

Brand value contains multi-dimensional data covering both brand awareness and brand relevance, both of which are largely based on customer perceptions and therefore a future intent to buy and an expectation of customer satisfaction. While brand value contains many such qualitative factors we are not aware of any movement to exclude these as valid criteria for making investment decisions. Therefore we can only assume that the pushback against the use of ESG criteria for making investment decisions is to protect pension assets from greenwashing or an assumption that ESG factors have no pecuniary value. (We understand the political differences of opinion around ESG activism and engagement (wokeness), but regard this more as a social and electoral issue than a financial matter.) We have written about greenwashing before[3] and the value of ESG metrics in the construction of sector based investment portfolios[4] . In this paper, we want to address the challenge that ESG factors have no pecuniary value in stock selection.

Our hypothesis is that better companies will outperform their peers. What is “better”? Better management, better customer service, better product quality, better governance and standards, better customer and market responsiveness and so on. It is those “better” companies that will be more able to meet the challenges of the future and, better yet, able to embed meeting those challenges into their corporate vision and strategy.

…the greatest opportunities lie in solving the greatest challenges…

Identifying those better companies requires a comprehensive investment process incorporating both quantitative and qualitative metrics. Among those qualitative metrics are the 26 Sustainability Accounting Standards Board (SASB) Materiality factors and other ESG factors. Being better is not static, it is a dynamic metric. When a credit rating agency puts an issuer on review for a downgrade it reflects a belief of a higher potential probability of default, and, consistent with that, we would expect that issuer’s credit spread to widen (all other things being equal). There is therefore an immediate and direct relationship between a changed credit outlook and the price of a security. (See also: Kume and Weir: Causality Relationship between Bond Ratings and Credit Spreads[5]). Thus despite criticism[6] of credit ratings during the global financial crisis, the legislative response was not to ban the use of credit ratings but rather to strengthen oversight of the credit rating agencies[7]. Likewise the SEC has not reacted to growing numbers of mutual funds focused on ESG principles by looking to restrict or eliminate the use of ESG factors. Rather the SEC encourages market participants to evaluate whether their ESG disclosures and marketing claims are accurate[8].

It is important not to conflate credit ratings and ESG ratings (or scores). While there are multiple providers for both, credit ratings measure only one thing: probability of default. ESG scores measure a multitude of things.

The fact that there are multiple factors behind an ESG score and multiple factors behind the price of a security simply means that it is hard to isolate the relevance of a discrete ESG factor to a discrete security price or a security price move. We have observed before the dangers of drawing false conclusions. For instance, overweighting a global equity portfolio with higher ESG-scored constituents will have the consequence of a higher US weight and a lower China weight relative to a global benchmark simply because more US issuers report ESG data than their Chinese counterparts. Thus the portfolio results reflect this macro over/under-weight much more than the higher ESG score composition. This is obvious given the stark historic performance differences between the US and Chinese markets.

In our prior work,[9] we observed that static ESG data cannot be used in isolation to predict asset return. We continue to hold that to be true.

Thus while National Beverage and Napco Security have similar ESG scores the total equity return (dividends reinvested) is dramatically divergent. At the same time, while Coca-Cola and Ford have similar ESG scores and similar 5 year total returns their relative stock price performance during the past 5 years has varied considerably.

Our prior work identified momentum as generating superior absolute and risk-adjusted returns relative to global equity benchmarks. Intuitively, we continue to suggest that issuers with improving ESG metrics will demonstrate better-than-benchmark equity returns over time and, conversely, those with deteriorating ESG metrics will underperform benchmarks over time. But when we look at global data[10], the correlation (R2) between improving ESG metrics and relative equity performance is very low, 0.0006.

The lack of any meaningful correlation between a change in ESG scores over 5 years and security performance relative to benchmark (MSCI World) is not surprising given the sheer number of other factors at play during an extended period. The lack of correlation is starkly evident in the data for shorter periods as well.

Within that, there are nonetheless corporate specific examples of historic positive equity return performance and improved ESG scores. During the period June 2018 to June 2023, Danaher Corp (XNYS: DHR) has had its MSCI ESG rating upgraded by 4 notches from B to AA. At the same time, its stock has outperformed the MSCI World Index.

Are the two related? There’s no denying Danaher’s commitment to sustainability and resultant ESG score upgrades[12]. There’s also no denying the fact that Danaher’s EBITDA has grown 170% from 2018 to 2023, from $4.24 billion to $11.45 billion (Source: Bloomberg). There’s also no denying that other US pharmaceutical companies have had similar 5 year total equity returns, e.g. Natera (XNGS: NTRA) 20.90% pa compared to Danaher 19.93%pa although with marked differences in stock price volatility.

At the end of the day, the two companies are very different in terms of size, profitability, business focus and ESG rating (one leading (DHR) and one lagging (NTRA))[13]. No single factor is the sole determinant of the relative value of one versus the other. 

We conclude that ESG factors, or more precisely changes in ESG factors, have more pecuniary relevance for some issuers than others. 

We can see this when we look at the historic data broken out by ESG rating change.

The proportion of global constituents with a positive return over the global benchmark tends to increase as ESG ratings are upgraded. 

…ESG metrics have pecuniary relevance… 

The data for the 5 year period 2018 to 2023 also holds true for the individual one year periods over this stretch of time. The only outlier period is during the COVID pandemic (2019-20) when many other factors were at play.

By looking at companies by peer group we reduce the number of factors at play by eliminating those factors that do not impact their specific industry group. For the period 2018 to 2023, we observe the same trend: the greater the change in ESG rating, typically the higher the proportion of constituents with higher rates of return above industry benchmarks.

What is significant is that all the sectors that do not exhibit the same trend (utilities, financials and real estate) are those sectors most affected by rising interest rates. ESG factors are not the sole determinant of security performance. This does not mean they should not be considered or banned, but rather that they be considered in a thoughtful way. As a result, investors and investment managers really can find an investment edge.

…ESG factors can provide an investment edge…

What are we trying to identify? 

1. Any correlation between a change in ESG metric to a change in security return 
2. Any relevance to security performance based on factoring in a change in ESG metrics

We expect to see: 

More companies with positive performance and positive ESG rating changes and less companies with negative performance and negative ESG ratings. We view ESG factors as one of many other investment factors that matter. This is not an attempt prove that ESG scores in of themselves cause positive return.

How we measure ESG relevance: 

We use changes in MSCI ESG scores to gauge market participants' perceptions of a firm's ESG quality. For instance, a 5-point increase in ESG scores over five years indicates the market's belief that a company's environmental, social, and governance performance is improving. Therefore, we use ESG score changes as a proxy for market sentiment of a company’s ESG quality.

Timeframe and data:

We collected MSCI ACWI ESG data and total return data for securities across the following time frames: 

• June 29, 2018 - June 28, 2019 
• June 28, 2019 - June 30, 2020 
• June 30, 2020 - June 30, 2021 
• June 30, 2021 - June 30, 2022 
• June 30, 2022 - June 30, 2023 
• And finally, June 2018 - June 2023 for a comprehensive 5-year view 

The research commenced in February 2023, so the initial data collection focused on securities included in the index and their corresponding ESG scores in February 2023. We then retrospectively eliminated securities with missing ESG scores and those that exited the index within the June 2018-2023 timeframe, resulting in a final sample of 1,697 securities.

The MSCI ESG scores range from AAA (strongest) to CCC (weakest). We assigned numerical values to the rating changes as follows: 

• 0 for no change 
• +1 for each upward change 
• -1 for each downward change 

Since the research began before June 2023, the June 2022-2023 and 2018-2023 data were added later. This timing difference necessitated excluding an additional 19 stocks from our research universe for the 2022-2023 and, consequently, 2018-2023 periods. These exclusions were due to the 19 stocks being demoted to the small cap index or being acquired and/or merged. 

The stocks removed from the June 2022-2023 dataset include the following 19 stocks:

The exclusion of these 19 stocks is not considered a significant limitation, as it represents a negligible portion (19 out of 1,697) of our research universe. Moreover, this data was only missing for the periods 2022-2023 and 2018- 2023, not affecting the analysis for other time frames.

We calculated two returns: 

1. Total net relative return: This is the total return (weighting x the return) of the 1,697 stocks for the years 2018-2022 and the 1,697 stocks minus the 19 stocks for 2022-2023 and 2018-2023, minus the total return of the entire portfolio. 
2. Sector net relative return: This is the same as the total net relative return, but with the sector beta removed from all of the securities in each time frame (Return of security X - Return of the sector of security X). 

We compared each securities' return year on year and across the full 5 years, to their rating change in the respective time period. To ensure comparability, we reweighted the holdings in our portfolio to align with the characteristics of this select universe. We acknowledge that sector classifications have changed over time, and the MSCI methodology has also undergone revisions. Additionally, our research is limited to the MSCI ACWI universe, which may introduce survivorship bias. Furthermore, some data was missing due to disclosure issues, which could potentially impact our findings.