Abstract
Many studies have focused on the relationship between companies with strong environmental, social, and governance (ESG) characteristics and corporate financial performance. However, these have often struggled to show that positive correlations—when produced—can in fact explain the behavior. The authors of this article provide a link between ESG information and the valuation and performance of companies, by examining three transmission channels within a standard discounted cash flow model—which they call the cash-flow channel, the idiosyncratic risk channel, and the valuation channel. They tested each of these transmission channels using MSCI ESG Ratings data and financial variables. This showed that companies’ ESG information was transmitted to their valuation and performance, both through their systematic risk profile (lower costs of capital and higher valuations) and their idiosyncratic risk profile (higher profitability and lower exposures to tail risk). The research suggests that changes in a company’s ESG characteristics may be a useful financial indicator. ESG ratings may also be suitable for integration into policy benchmarks and financial analyses.
TOPICS: ESG investing, security analysis and valuation, risk management
Environmental, social, and governance (ESG) investing is a very broad field with many different investment approaches addressing various investment objectives. At a top level, we can break down ESG investing into three main areas that each have their own investment objective: First, ESG integration, in which the key objective is to improve the risk–return characteristics of a portfolio. Second, values-based investing, in which the investor seeks to align his portfolio with his norms and beliefs. Third, impact investing, in which investors want to use their capital to trigger change for social or environmental purposes, for example, to accelerate the decarbonization of the economy. In this article, we focus on the first investment objective—ESG as a means to achieve financial objectives in portfolio management.
In recent years, many researchers from both academia and the asset management industry have analyzed the relationship between the ESG profile of companies and their financial risk and performance characteristics. In fact, research has been so plentiful that several meta studies1 have summarized the results of over 1,000 research reports and found that the correlation between ESG characteristics and financial performance was inconclusive: The existing literature found positive, negative, and nonexistent correlations between ESG and financial performance, although the majority of researchers found a positive correlation.
The reasons for these inconclusive results likely stems from the different underlying ESG data used and the varying methodologies applied, especially in how far they control for common factor exposures. However, even researchers finding a positive correlation between ESG and financial performance often fail to explain the economic mechanism that led to better performance, as they typically focused on historical data analysis. A paper by Harvey, Liu, and Zhu (2016) highlights that this type of purely data-focused research entails the risk of correlation mining, that is, overfitting a financial model to a specific dataset to observe correlations that will not prevail when tested out of sample.
Another criticism mentioned in a paper by Krueger (2015) is the fact that many empirical studies analyzing the link between ESG and financial performance do not strictly differentiate between correlation and causality. Often, a correlation between ESG and financial variables is implicitly interpreted to mean that ESG is the cause and financial value the effect, although the transmission easily could also be reversed. For instance, one can argue that companies with high ESG scores are better at managing their risks, leading to higher valuations. Alternatively, companies with higher valuations might be in better financial shape and therefore able to invest more in measures that improve their ESG profile; such investments might lead to higher ESG scores.
To address these issues, this research article takes a different approach. Instead of simply looking for correlations between ESG characteristics and financial performance in historical data, we first analyze the transmission channels from ESG to financial performance and develop a fundamental understanding of how ESG characteristics affect corporations’ valuations and risk profiles. Afterwards, we verify these transmission mechanisms using empirical analysis.
The advantages of this type of approach are threefold: First, it mitigates the risk of correlation mining between ESG data and financial performance data. We use MSCI’s Barra Global Equity Model for all financial and risk data. Model data has not been fitted in any way to the underlying ESG dataset. Secondly, it reduces the risk of finding correlations that are caused by unintentional exposures to common factors. Thirdly, it better differentiates between correlation and causality by studying transmission channels.
In essence, our analysis is designed to help explain how ESG affects the financial profile of companies in a fundamental way, thus producing more convincing evidence than simple correlation studies. In the online supplement, we present key performance measures of the MSCI ESG Universal Index (which re-weights components in the MSCI ACWI Index according to their ESG profile) and the MSCI ESG Leaders Index (which performs a best-in-class selection based on ESG ratings). Both ESG indexes, which are based on MSCI ESG Ratings, showed lower levels of risk, improved in risk-adjusted returns and higher levels of valuation in line with this study’s findings.
A positive impact of MSCI ESG Ratings on performance and risk figures was also found in the quantitative analyses of Dunn, Fitzgibbons, and Pomorski (2016) as well as Porse et al. (2017). Therefore, this paper does not focus on reproducing these positive risk and performance results. Instead, we focus on explaining the economic reasons for these results.
WHY ESG MATTERS
To develop a fundamental understanding of how ESG characteristics affect corporations’ financial profiles, we rely on existing corporate finance models in establishing the transmission channels of ESG to the financial world.
Studies by El Ghoul et al. (2011) and Gregory, Tharyan, and Whittaker (2014) show that a discounted cash-flow (DCF) model framework (which describes a company’s value as the sum of future cash flows, discounted at the cost of capital) can be used to break down the influence of a corporation’s ESG profile on equity valuations, including cash flows, risk, and cost of capital. The authors argue that it is important to differentiate between the systematic and idiosyncratic risk of equities. Gregory, Tharyan, and Whittaker (2014) explain that systematic risk is macroeconomic in nature and describes the general market risk all companies are exposed to, for example, the risk of shocks in commodity prices, interest rates, or inflation rates. Systematic risk also includes industrywide issues such as regulatory changes, technological developments, and stranded assets.
In contrast, firm-specific risk is particular to a company. The distinction between systematic and firm-specific risk is highly important for analyzing the impact of ESG characteristics on corporate valuation because investors can typically diversify away firm-specific risk. Therefore, it is solely the systematic risk component that determines shareholders’ required rate of return as compensation for the risk to which they are exposed. Consequently, within a DCF model, systematic risk is typically captured through the cost of capital (i.e., the denominator in the DCF model), whereas firm-specific risk is linked to the numerator of the DCF model, that is, future cash flows. We follow this approach and use a standard DCF model as a starting point of our analysis. However, instead of simply analyzing the impact of ESG characteristics through the discounted cash flow model, we take the investor’s perspective and break down the influence of ESG characteristics on corporations into three transmission channels: the cash-flow channel, the idiosyncratic risk channel, and the valuation channel.
In the next two sections, we analyze the two idiosyncratic transmission channels and the systematic risk channel. Next, we assess the question of causality of ESG. Subsequently, we analyze the financial value of ESG rating changes before examining the intensity–longevity profile of ESG ratings compared to common factors. Finally, we show the impact of the various ESG transmission channels on financial performance.
DATA AND METHODOLOGY
We now validate the three transmission channels using MSCI ESG Ratings2 for the MSCI World Index universe for the January 2007 to May 2017 time period. The universe contains over 1,600 stocks and is therefore sufficiently diversified for the statistical analysis performed in this article. All risk and factor calculations are performed using the Barra Long-Term Global Equity Model (GEMLT).
All the results shown in this article are neutralized for industry exposure (through the use of industry-adjusted ESG scores) and size. We created size-adjusted ESG scores as the residuals from regressing standard MSCI ESG scores on the size exposure in the GEMLT model and an intercept variable.
In our analysis, we show the distribution of financial variables across five size-adjusted ESG score quintiles (Q1 to Q5), with Q1 indicating the companies with the lowest ESG rating and Q5 the highest-rated companies. Financial variables, such as beta or book-to-price ratio, are based on GEMLT and are therefore in the format of Z-scores.3 For each of these financial variables, we show in the exhibits the average value over the 10-year observation period as square dots, the current exposure as round dots, and the 5% to 95% range of observed values as vertical bars.
The results of all simulations, including the corresponding t-statistics that show their statistical significance, are summarized in the online supplement.
IDIOSYNCRATIC TRANSMISSION CHANNELS
In this section, we analyze the company-specific impact of ESG on risk and performance. The firm-specific risk profile of companies is transmitted through the numerator (future cash flows) in the DCF model framework and can be broken into two separate channels: the transmission of ESG into future opportunities and therefore into profitability on the one hand, and the transmission to firm-specific downside risk protection on the other.
Cash-Flow Channel
The cash-flow transmission channel can be summarized as follows:
The work of Gregory, Tharyan, and Whittaker (2014) explains the economic rationale of the cash-flow channel:
1. Companies with a strong ESG profile are more competitive than their peers. For instance, this competitive advantage can be due to the more efficient use of resources, better human capital development, or better innovation management. In addition to this, high ESG-rated companies are typically better at developing long-term business plans and long-term incentive plans for senior management.
2. High ESG-rated companies use their competitive advantage to generate abnormal returns, which ultimately leads to higher profitability.
3. Higher profitability results in higher dividends.
The competitive advantage of high ESG-rated companies cannot be readily observed. Therefore, our empirical analysis focuses on the second and third steps of the cash-flow channel, that is, higher profitability and higher dividends.
Higher profitability and dividends. We found data supporting the assertion that high ESG-rated companies (Q5) were more profitable and paid higher dividends, especially when compared to bottom quintile (Q1) companies, as can be seen in Exhibits 1 and 2.
Gross Profitability of ESG Quintiles
Notes: Gross profitability (Z-score) of size-adjusted ESG quintiles is computed as most recently reported sales less cost of goods sold, divided by most recently reported company total assets. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Trailing Dividend Yield of ESG Quintiles
Notes: Trailing dividend yield (Z-score) of size-adjusted ESG quintiles is computed by dividing the trailing 12-month dividend per share by the price at the last month end. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
High-dividend yields play an essential role in our analysis because sustainability investors typically have a long-time investment horizon (Eccles and Kastrapeli 2017). Gupta, Melas, and Suryanarayanan (2016) analyzed the importance of dividends for long-term performance. Exhibit 3 illustrates their breakdown of the total return for the MSCI ACWI Index into stock price increases, dividends, and dividend growth, over different time periods. The performance contribution of dividends to portfolio returns was increasingly important as time horizons lengthened.
In the Long Run, Cash Flows to Shareholders Drive Most of the Portfolio Return
Notes: The graph presents a decomposition of the total return of the MSCI ACWI Index for the time period from December 1994 to September 2015. Total return is decomposed into dividend yield and price return components, and price return is further decomposed into dividend per share (company fundamental) growth and price-to-dividend (valuation ratio) growth.
Therefore, the apparent tilt of high ESG-rated strategies such as the ESG Universal Index toward high dividend-paying companies may have helped enhance medium- to long-term improvement of risk-adjusted returns. Performance data can be found in the online supplement.
Idiosyncratic Risk Channel
The second company-specific transmission channel relates how well high ESG-rated companies manage their business and operational risks. Their stock prices typically have shown lower idiosyncratic tail risk, as outlined below:
The economic rationale for this transmission channel is explained in research by Godfrey, Merrill, and Hansen (2009); Jo and Na (2012); and Oikonomou, Brooks, and Pavelin (2012). It is summarized as follows:
1. Companies with strong ESG characteristics typically have above-average risk control and compliance standards across the company and within their supply chain management.
2. Because of better risk control standards, high ESG-rated companies suffer less frequently from severe incidents such as fraud, embezzlement, corruption, or litigation cases (cf. Hong and Kacperczyk 2009) that can seriously impact the value of the company and therefore the company’s stock price. Hoepner, Rezec, and Siegl (2017) call this an “insurance-like protection of firm value against negative events.”
3. Less-frequent risk incidents ultimately lead to less stock-specific downside or tail risk in the company’s stock price.
The authors also support this transmission channel by empirical analysis. For instance, Hoepner, Rezec, and Siegl (2017) observe that high ESG-rated companies showed statistically significant lower downside risk measures such as volatility, lower partial moments, and worst-case loss.
We will now verify each step of the idiosyncratic risk channel.
Better risk management practices. The analysis of companies’ exposure and management techniques in relation to environmental, social, and governance risks is the backbone of MSCI ESG Research’s framework: The MSCI ESG Rating model measures both risk exposure to and risk management of a company’s key ESG issues. The online supplement shows an overview of the key risk issues that are assessed as part of the MSCI ESG Rating research process.
To score well on a key issue, management needs to be commensurate with the level of exposure: A company with high ESG risk exposure must also have very strong management, whereas a company with limited exposure can have a more modest approach. Conversely, a highly exposed company with poor management will score worse than a company with the same management practices but lower exposure to the risk.
In each of 157 Global Industry Classification Standard (GICS) sub-industries, the MSCI ESG Rating model incorporates only a handful of key issues that it determines are the most financially significant for the specific industry. That is, not all ESG issues are considered important; those that are not deemed significant do not carry a weight in a company’s rating. In essence, the MSCI ESG Rating is a reflection of companies’ residual risk exposure to their industry’s most significant key issues after taking into account companies’ risk-mitigation techniques. Therefore, these ESG ratings are a suitable starting point for our analysis.
Lower risk of severe incidents. To assess the ability of companies’ risk management functions to successfully mitigate severe incidents that can lead to financial losses, we looked at the frequency of large, adverse idiosyncratic stock price moves. To be precise, for the 10-year observation period, we identified companies in the MSCI World Index that have had a drawdown of more than 95% or went bankrupt in the 3-year period after the company was categorized in either the top or bottom ESG rating quintile, which we consider to be an idiosyncratic risk incident. For each of these incidents, we look at each company’s ESG rating before the respective 3-year drawdown period started. Exhibit 4 shows how frequently such incidents occurred (the incident frequency) for the top and bottom ESG quintiles over the full 10-year time period.
Idiosyncratic Incident Frequency of Top and Bottom ESG Quintiles
Note: For each month, we report the number of stocks that realized a more than 95% cumulative loss over the next 3 years, taking the price at month end as the reference point for return calculation.
Over the past 10 years, higher ESG-rated companies showed a lower frequency of idiosyncratic risk incidents, suggesting that high ESG-rated companies were better at mitigating serious business risks. We have also tested the robustness of this result by using different drawdown thresholds (25%, 50%, and 95%) and drawdown periods (three years and five years). In each parameter setup, companies with high ESG ratings had a significantly lower incident frequency than companies with poor ESG ratings.
Lower idiosyncratic tail risks. Better risk management practices should ultimately be visible in the form of reduced stock-specific risk of the corresponding stock price—especially stock-specific tail risk as discussed in the work of Hoepner, Rezec, and Siegl (2017).
To understand how ESG characteristics are linked to tail risks, Exhibit 5 compares the residual volatility of companies across ESG quintiles, that is, the volatility that is not explained by the common factors in the MSCI Barra Global Equity Model. Exhibit 6 shows the kurtosis of stock returns across ESG quintiles; kurtosis is a commonly used measure for tail risks.
Residual CAPM Volatility of ESG Quintiles
Notes: Residual capital asset pricing model (CAPM) volatility (Z-score) here is defined as the volatility of the residual returns from the CAPM regression used in the calculation of historical beta (see Exhibit 3). Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Kurtosis of ESG Quintiles
Notes: Kurtosis (Z-score) is computed as the ratio of the fourth central moment of daily returns divided by the square of daily return variance. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Both stock-specific risk measures show lower idiosyncratic risk for high ESG-rated companies, in particular with respect to tail risks.
SYSTEMATIC RISK TRANSMISSION CHANNEL
We now analyze how companies’ ESG profile impacts their exposure to systematic risk and how this impact may ultimately lead to financially significant effects. In a DCF model framework, the systematic risk exposure affects the denominator of the DCF model.
Valuation Channel
Studies by Eccles, Ioannou, and Serafeim (2014); El Ghoul et al. (2011); and Gregory, Tharyan, and Whittaker (2014) argue that a strong ESG profile leads to higher valuations through the following transmission process:
Their economic rationale is as follows:
1. Companies with a strong ESG profile are less vulnerable to systematic market shocks and therefore show lower systematic risk. For instance, energy- or commodity-efficient companies are less vulnerable to changes in energy or commodity prices than less-efficient companies and therefore their share price tends to show less systematic market risk with respect to these risk factors.
2. In a capital asset pricing model (CAPM) model framework (cf. Ruefli, Collins, and Lacugna 1999), the beta of a company has two important functions: First, beta measures the systematic risk exposure of companies (i.e., lower beta means less systematic risk), and second, it translates the equity risk premium into the required rate of return for the individual company. Therefore, lower systematic risk means a company’s equity has a lower value for beta, and therefore investors require a lower rate of return. Ultimately, this translates into a lower cost of capital for a company. This argument can be extended to multifactor models, in which the systematic risk exposure of a company is measured by several factor loadings instead of one beta.
3. Finally, a lower cost of capital leads directly to the last step of the transmission mechanism: In a DCF model framework, a company with lower cost of capital would have a higher valuation.
In addition to this, research by Hong and Kacperczyk (2009) together with a study by El Ghoul et al. (2011) shows that the transmission channel from lower systematic risk to higher valuations can also be explained through the relative size of the investor base. That is, the authors argue that companies with low ESG ratings have a relatively small investor base because of two effects:
• Investor preferences: Many risk-averse investors and socially conscious investors avoid exposure to low ESG-ranked companies.
• Information asymmetry: The problem of asymmetric information between companies and their investors is less severe for high ESG-rated companies, since high ESG-rated companies are typically more transparent, in particular with respect to their risk exposures and their risk management and governance standards.
Although the impact of ESG ratings on a company’s investor base is fairly difficult to measure in practice, it can be a key motivation for large asset owners to integrate ESG in their portfolios. For instance, a presentation by SwissRe (2017) mentions that
a shift to ESG benchmarks would lead to a smaller investment universe and hence lower demand for the excluded securities. Over the long term, we expect that such movements will motivate these companies to further include ESG aspects into their business approach and extend their ESG-related disclosure. Due to the improved resilience to long-term risks, this is beneficial for investors as well as for the company itself. Consequently, ESG factors will have an impact on company valuation and cost of capital, and as such become an integral part of financial analysis.
Next, we examine how ESG ratings have affected systematic risk, the cost of capital, and equity valuations.
Lower Systematic Risk
Exhibit 7 compares the average systematic volatility of ESG rating quintiles within the MSCI World Index over a 10-year period, whereas Exhibit 8 compares the earnings variability of the ESG-rated quintiles. Exhibits 7 and 8 illustrate that companies with high ESG ratings have shown less volatile earnings and less systematic volatility, in line with the conjecture that companies with high ESG ratings show lower systematic risk exposure.
Systematic Volatility of ESG Quintiles
Notes: Systematic volatility (or common factor risk) is calculated as the volatility predicted by all the factors of the GEMLT model. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Variability in Earnings of ESG Quintiles
Notes: Variability in earnings (Z-score) is computed as the standard deviation of company reported annual earnings over the last five fiscal years, divided by the average annual earnings. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Lower Cost of Capital
In a CAPM model framework, the cost of capital is determined by the expected return, which is calculated as the risk-free rate plus the stock’s beta times the market’s excess return. Therefore, the stock’s beta is a one-to-one measure for the cost of capital, that is, higher cost of capital coincides with higher values of beta. Exhibit 9, which compares the average beta of ESG quintiles, demonstrates that high ESG-rated companies experienced lower levels of beta and therefore—in the context of the CAPM—lower costs of capital.
Historical Beta of ESG Quintiles
Notes: Historical beta (Z-score) is computed as the slope coefficient from a time-series regression of stock excess returns, against the cap-weighted excess returns of the estimation universe over a trailing window of 504 trading days, with a 252-day half-life. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Higher valuation. Ultimately, we expect lower costs of capital to result in higher company valuations. Exhibits 10 and 11 compare the average book-to-price and predicted earnings-to-price ratios of ESG quintiles, respectively. The two exhibits show that high ESG ratings coincided with higher valuations in terms of both book-to-price and earnings-to-price ratios.4
Book-to-Price Ratio of ESG Quintiles
Notes: Book-to-price ratio is computed as the last reported book value of common equity divided by current market capitalization. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Predicted Earnings-to-Price Ratios of ESG Quintiles
Notes: Predicted earnings-to-price ratio is computed by dividing the average analyst estimate of 12-month forward-looking earnings by the current market capitalization. Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
It is worth noting that the valuation channel has been supported by both academic and industry researchers. The paper by El Ghoul et al. (2011) shows that higher ESG-rated companies had lower costs of capital according to four different measures while controlling for common factor exposures. Research by Dunn, Fitzgibbons, and Pomorski (2016) observes that high ESG ratings coincided with lower systematic risk and higher valuations. Furthermore, the work of Melas, Nagy, and Kulkarni (2016) shows that ESG ratings exhibited a negative correlation to the value factor, which is in line with the observation that high ESG-rated stocks have carried higher valuations (and consequently less exposure to the value factor).
In addition, Desclee et al. (2015) found a similar transmission channel in the corporate bond market: In their analysis, they show that higher ESG-rated corporate bonds had lower systematic risk, lower spreads, and therefore higher valuations while controlling for common corporate bond factors.
FROM CORRELATION TO CAUSALITY
When it comes to understanding whether higher ESG ratings can lead to higher valuations or whether higher valuations lead to higher ESG ratings, we have a chicken-and-egg problem. The work of Krueger (2015) emphasized that the direction of causality between positive correlations for ESG rating and corporate valuation is not clear: Higher ESG ratings can—through lower systematic risk and lower cost of capital—lead to higher valuations. Alternatively, higher valuations can indicate successful companies that have more money to invest in sustainability-related areas, leading to a higher ESG rating.5
Understanding the causal relationship between ESG characteristics and financial values is crucial in showing the benefits of ESG investing. To start with, the fundamental basis of our transmission channels is the observation that ESG characteristics influence both the systematic and idiosyncratic risk profiles of corporates. Consequently, we examine the extent to which a change in a company’s ESG characteristics has been a leading indicator for changes in systematic and idiosyncratic risk, and how far these changes in the risk profile have led to a change in the financial target variable of the transmission channels. This analysis helps us understand how ESG rating changes have affected ESG strategies.
Testing and verifying causality empirically is more difficult than looking at correlations alone. Although correlations can be assessed by simply analyzing datasets at a given point in time, understanding causality requires an analysis of changes over time, which limits the dataset to events in which actual changes have occurred. Another limiting factor is the length of time series that is available for observing rating changes, which in our case is the 10-year time period used for the empirical analysis of the transmission channels.
As a consequence, we expect the statistical significance of our causality analysis to be lower than for the statistical analysis of the transmission channels, in particular because of the relatively short time series. For that reason, we focus the causality analysis on those channels in which we were able to find empirical support for causality, namely, the valuation channel (in which the empirical evidence was strongest) and the idiosyncratic risk channel.
Causality in the Systematic Risk Channel
To assess causality in the valuation channel, we analyze the extent to which changes in the ESG profile of companies have led to changes in the systematic risk profile, changes in the cost of capital, and finally, changes in the valuation of companies.
The economic rationale can be derived directly from the previous arguments for the valuation channel:
• An improving ESG profile means a company is becoming less susceptible to systematic risks.
• Lower systematic risk leads to a reduction in a company’s cost of capital.
• The reduction in cost of capital leads to an increase in valuation.
We now analyze the same financial variables as in the valuation channel, with one key difference: We assess to what extent changes in companies’ ESG profiles predicted changes in these financial variables. Therefore, we plot the change of key variables over three buckets of changes in ESG ratings—downgrades, neutral (no change), and upgrades. As we show in the online supplement, rating changes of more than one notch were relatively rare and are therefore aggregated into one combined upgrade and downgrade bucket. Since ESG characteristics can be expected to influence corporations’ financial profile mainly in the medium- to long-term (i.e., over a multiyear period), we look at changes in financial variables over a 3-year time period after the change in ESG rating occurred.
Decreasing systematic risk. Exhibit 12 shows the change in companies’ systematic volatility (as above) over the rating change buckets. Companies with a rating upgrade demonstrated a relative improvement in their systematic risk profile compared to neutral or downgraded companies. As expected, the statistical significance of the result (shown in the online supplement) is lower than in the corresponding analysis of the valuation channel.
Change in Companies’ Systematic Volatility
Notes: Change in systematic volatility (also known as common factor risk) is computed using the volatility predicted by all the factors of the responsive variant of the GEMLT model. The difference is calculated between the volatility 36 months after the analysis date and the volatility as of the analysis date. (See also Exhibit 1.) Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Decrease in cost of capital. Analogous to the valuation channel, we use companies’ betas as proxies for their cost of capital. In Exhibit 13, we assess how far a rating upgrade or downgrade predicted a fall or rise in a company’ beta, respectively. These changes can be used as a proxy for a change in their cost of capital. We observed a relative decrease in beta for companies whose ESG rating improved compared to downgraded companies. As for systematic volatility, the statistical significance was lower than in the analysis of the valuation channel.
Change in Companies’ Beta
Notes: Change in historical beta (Z-score) is computed as the difference between historical beta 36 months after the analysis date and beta as of the analysis date. (See also Exhibit 3.) Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
Increase in valuation. The final step of the causality analysis is to study the impact a change in the cost of capital had on a company’s valuation. We analyze this relationship in Exhibit 14, looking at the predicted earning-to-price ratio. We observe that rating upgrades led to a relative decrease in the predicted earning-to-price ratio compared to rating downgrades, indicating an increase in valuation.
Change in Predicted Earnings-to-Price Ratio
Notes: Change in predicted earnings-to-price ratio is computed as the difference between predicted earnings-to-price ratio 36 months after the analysis date and earnings-to-price ratio as of the analysis date (see also Exhibit 5). Data from January 2007 to May 2017. Average value over the period is represented by square dots; current exposure by round dots. The vertical bars represent the 5% to 95% range of observed values.
To conclude, we have analyzed the causality of the valuation channel analogously to the valuation channel itself, using changes in financial variables relative to changes in ESG ratings. However, the statistical significance of the obtained results is clearly lower than for the valuation channel, which can be explained by the relatively short time period of our analysis. To improve the statistical significance of our analysis, a longer time series is needed; this will be an important focus of future research.
Causality in the Idiosyncratic Risk Channel
Obtaining statistically significant evidence of causality for all the different steps in the cash-flow channel and the idiosyncratic risk channel is more challenging than for the valuation channel. We therefore focus on the empirically strongest result that we found, that is, the relationship between a change in ESG ratings and the impact on the incident frequency.
Exhibit 15 shows the idiosyncratic risk profile (measured as incident frequency) of rating upgrades and rating downgrades. Over the 10-year observation period, rating upgrades had a lower incident frequency than rating downgrades, supporting the assertion that rating changes are a leading indicator for idiosyncratic risks.
Proportion of Large Losses by Rating Notch Changes
Notes: For each month, stocks are sorted into three groups—rating upgrades, neutral, and downgrades over the previous 12 months. For each group, we then compute the proportion of stocks that realized a more than 95% cumulative loss over the next three years, assuming the price at month end as the reference point for return calculation. The exhibit shows only the time series of upgrades and downgrades (see also Exhibit 12).
To conclude, in our causality analysis of the systemic and idiosyncratic risk channels, we found that downgraded companies experienced a relative increase in both systematic and stock-specific risk compared to companies whose ESG rating was upgraded.
ESG MOMENTUM
Assessing the transmission of a change in a company’s ESG profile to a change in financial indicators such as valuation or profitability is not only important for testing causality: It is also important because the change in financial variables such as valuation can be a source of alpha. For instance, the paper by Gregory, Tharyan, and Whittaker (2014) argues from a conceptual point of view that since ESG characteristics have impacted corporations’ valuation through systematic risk, a change in a company’s ESG profile should be a predictor for a change in valuation and therefore for stock returns.
In a more practical analysis, Nagy, Kassam, and Lee (2016) show that an investment strategy that tilts a hypothetical standard market cap–weighted portfolio toward companies that show a positive ESG rating trend significantly outperformed both the benchmark and a comparable strategy that tilted the portfolio weights toward companies with high ESG ratings.
To verify the conjecture that ESG rating changes—which we call ESG momentum—can be a financially significant indictor and a potential link to returns, we compare the historical performance of the top ESG momentum quintile to the bottom ESG quintile. The ESG momentum indicator is calculated as the year-on-year change in the industry-adjusted ESG score.
The results in Exhibit 16 show significant outperformance of the top ESG momentum quintile over the bottom quintile, corresponding with the findings from our transmission-channel analysis: An improvement in ESG characteristics has led to increasing valuations over time.
Financial Performance of Top versus Bottom ESG Momentum Quintile
Notes: Cumulative performance differential of the top ESG momentum quintile versus the bottom ESG momentum quintile. ESG momentum is defined as the 12-month change in ESG score.
We conclude that ESG momentum can be a useful financial indicator in its own right and may be used in addition to the actual ESG rating in index or portfolio construction methodologies.
It is important to emphasize that the financial value of ESG momentum is also supported in existing literature, for instance a paper by Khan, Serafeim, and Yoon (2015). The authors used MSCI ESG Ratings data to create customized ESG scores and performed a regression analysis of stock returns to ESG score changes (i.e., ESG momentum), neutralized with respect to changes in size, market-to-book ratio, leverage, profitability, R&D intensity, advertising intensity, institutional ownership, and sector membership. The authors found statistically significant predictive power of ESG momentum for stock returns.
FACTOR INTENSITY AND LONGEVITY
The verification of the different transmission channels and the causality analysis also points to other important conclusions about the differences between ESG ratings and more traditional factors. Those differences relate to the intensity of the signal (i.e., financial impact per unit of time) and the longevity of the signal (i.e., how long the signal persists):
• Intensity: Looking at the impact ESG ratings have shown in the transmission channels, especially on systematic and idiosyncratic risk figures, we consider that the intensity of ESG ratings is lower than common factors such as momentum or low volatility.
• Longevity: The existence of a risk-reduction effect, even three years after an ESG rating upgrade, indicates a relatively long timespan for ESG as an investment signal.
Now, we assess the differences between ESG and common factors in terms of intensity and longevity in a more quantitative way. We use the information ratio (IR) of a factor as a measure of its intensity and factor stability as a measure of the longevity of the factor (Exhibit 17).
Intensity and Longevity of Common Factors, ESG Ratings and ESG Momentum
Notes: Factor stability is computed as the average cross-sectional correlation between factor exposures (or ESG scores and ESG momentum scores) at month end and three months later. For GEMLT factors, factor performance is computed as the annualized information ratio of monthly factor returns; for ESG, it is computed as the annualized Sharpe ratio of the equal-weighted top minus bottom quintile portfolio created from ESG scores.
We observe that dynamic factors, such as momentum, are quite intense (i.e., high IR), but their lifespan is relatively short. Defensive factors such as dividend yield are in the middle of the intensity–longevity spectrum. The two factors with the longest lifespan are in fact ESG and size, but their intensity as measured by their IRs is relatively low. ESG momentum shows higher intensity levels than ESG itself, but also a shorter lifespan.
In essence, ESG ratings have displayed a different intensity–longevity profile from most other factors, with important implications for how and where more traditional factors and ESG ratings can be used, based on our historical analysis. Whereas these factors have become increasingly popular in quantitative strategies and strategies replicating factor indexes that have experienced high turnover, the longevity of ESG ratings makes them especially suited for integration into indexes that serve as benchmarks.
In addition, combining traditional factors with ESG could have resulted in both the short-term performance benefits of quantitative factors and the medium- to long-term risk reduction potential of ESG ratings.
CONCLUSION
By creating transmission channels, we have shown how ESG has affected the valuation and performance of companies, both through their systematic risk profile (lower costs of capital and higher valuations) and their idiosyncratic risk profile (higher profitability and lower exposures to tail risk). Thus, the transmission from ESG characteristics to financial value is a multichannel process, as opposed to factor investing in which the transmission mechanism is typically simpler and one dimensional.
In addition, ESG ratings were lower in intensity than traditional factors such as momentum or low volatility (i.e., the financial impact per unit of time for ESG ratings is relatively low), but typically lasted for several years. Classical factors such as momentum typically have lasted for a few months only, making them suitable for factor investing but not necessarily as long-term policy benchmarks.
ENDNOTES
↵1 For example, see Carpenter and Wyman (2009) and Fulton, Kahn, and Sharples (2012).
↵2 For more information see http://www.msci.com/products/esg/about_msci_esg_research.html; https://www.msci.com/esg-ratings.
↵3 Z-scores are normalized values, calculated by first subtracting the cross-sectional mean from all values and then dividing the difference by the cross-sectional standard deviation. Z-scores have zero mean and unit standard deviation. Following the GEMLT methodology, for risk-related variables, we subtract cross-sectional global means; for fundamental data-related variables, we subtract cross-sectional country means to control for potential country biases in fundamental data. Standard deviation is calculated globally.
↵4 A higher valuation means lower book-to-price and lower earnings-to-price ratios.
↵5 The question of causality between ESG ratings and systematic and idiosyncratic risk is also analyzed in research by Dunn, Fitzgibbons, and Pomorski (2016). The authors show by using a regression model that ESG ratings are predictive of future systematic and idiosyncratic risk figures, while controlling for the current level of risk and common factor exposures. The authors’ empirical results show ESG ratings to be predictive for both future systematic and future stock-specific risk for up to five years.
- © 2019 Pageant Media Ltd
Jump to section
Similar Articles
Cited By...
- The Impact of ESG Risk on Stocks
- Sustainable and Impact Investing: A Taxonomy of Approaches and Considerations for Fiduciaries
- Does Sustainable Investing Deprive Unsustainable Firms of Fresh Capital?
- Get Green or Die Trying? Carbon Risk Integration into Portfolio Management
- Deconstructing ESG Ratings Performance: Risk and Return for E, S, and G by Time Horizon, Sector, and Weighting
- Board Effectiveness and Firm Risk
- Combining E, S, and G Scores: An Exploration of Alternative Weighting Schemes