ESG Returns – As a Matter of Factor…

Mainstream ESG equity strategies have had a poor start to the year. In our view, this doesn’t disprove the existence of ESG alpha; rather, it proves the importance of understanding the underlying drivers of returns.

One would be hard pressed to imagine a more difficult storm for a typical ESG portfolio.

There has been some panic recently in the ESG community. From the war in Ukraine driving up both the oil price and defence spending, to the collapse of growth stocks in unison, one would be hard pressed to imagine a more difficult storm for a typical ESG portfolio low on commodities and high on tech.

For doubters of ESG alpha, the recent weak returns perhaps offer validation and an “I told you so” moment. Whether the recent episode is a one-off, or whether there is long-term “alpha” in ESG signals, are topics for another day. However, we believe the answer is complex, nuanced and very much debatable. In this article, we will focus more specifically on decomposing the experience of the past two years to better inform this broader discussion.

Dissecting ESG Performance

It is ever convenient to attribute recent ESG performance simply to rising oil prices and increased defence spending. In reality, ESG encompasses much more than that. When dissecting ESG performance to understand its drivers, we need to start by stripping the unintentional bets from the intentional.

Generic ESG factors are often influenced by unintended factor exposures. When interpretating ESG returns, one therefore needs to take special care. For example, if we regress a sample of ESG factor returns on generic factor returns (such as Value, Momentum, or Growth) we can often explain a good proportion of its cross-sectional returns. When we perform this exercise on MSCI ESG scores, we get an R-squared that hovers around 60-70% (Figure 1). And when we do the same with Sustainalytics ESG scores, we get an even higher fit of 80-90% R-squared. We include our own model for comparison.

Figure 1. R-squared Regression of ESG Returns on Barra Factor Returns1, April 2021 to March 2022 (Top 3000 Global Universe)

Problems loading this infographic? - Please click here

Source: Man Numeric, MSCI, Sustainalytics, Barra; as of 31 March 2022.

These results do not immediately imply that there is no idiosyncratic return to ESG. However, when we look at raw ESG returns, we should be aware that the returns may be overwhelmed by generic factor returns. It is not the returns of ESG attributes that we are measuring here, but the consequence of unintended biases. In the case of Sustainalytics ESG scores, they are subsumed by Value, Quality and Size (Figure 2). In the case of MSCI ESG scores, we see an anti-Quality, pro-Profitability and anti-specific-risk exposure.

Figure 2. T-stats - Regression of ESG Returns on Barra Factor Returns2, April 2020 to March 2022 (Top 3000 Global Universe)

Problems loading this infographic? - Please click here

Source: Man Numeric, MSCI, Sustainalytics, Barra; as of 31 March 2022.

A good majority of raw ESG returns can be explained by what is going on with the factor environment.

By using these two popular ESG methodologies as an example above, we see that a good majority of raw ESG returns can be explained by what is going on with the factor environment. This is often determined by the macroeconomic environment, so in our view is less about whether investors are rewarding or punishing ESG attributes.

This also contributes to the divergence in returns between different ESG approaches. For Q1 2022, long-short spread returns for Sustainalytics ESG scores outperformed MSCI ESG scores by a wide margin (Figure 3). However, a lot of this outperformance can be attributed to the greater Value and Quality exposures within the Sustainalytics ESG scores. When stripped of its generic factor exposures, all this outperformance goes away in the residual spread shown in Figure 3. Conversely, MSCI ESG returns were muted in Q1 2022, but the idiosyncratic ESG content was meaningfully positive after accounting for factor biases.

Figure 3. Q1 2022 – Long-Short Decile Performance Spreads (Top 3000 Global Universe)

Problems loading this infographic? - Please click here

Source: Man Numeric, MSCI, Sustainalytics, Barra; as of 31 March 2022.

For these reasons, in our own Man Numeric ESG alpha model (which relies on some of the underlying components of the Sustainalytics and MSCI ESG scores), we seek to ensure that we mostly cleanse style tilts and truly capture idiosyncratic ESG returns. Figure 1 illustrates the lower R-squared that results.

In a nutshell, what you get as an investor is often determined by the factor exposures you carry, intentionally or unintentionally. When deciphering ESG returns, one certainly needs to take that into consideration.

 

1. Regression performed daily with a rolling window of 252 days. Factors are Value, Momentum, Growth, Earnings Quality, Investment Quality, Profitability, Specific Risk and Size.
2. Regression performed daily with a rolling window of 252 days. Factors are Value, Momentum, Growth, Earnings Quality, Investment Quality, Profitability, Specific Risk and Size.

User Country: United States (237)
User Role: Public (Guest) (1)
User Access Groups:
Node Access Groups: 1