Some Considerations for Investors Exploring ESG Strategies

PERSPECTIVES

Although investors’ focus on ESG issues is growing, investors nonetheless have diverse views about the role of ESG factors in portfolio management. In particular, investors may share an interest in ESG issues, but balancing those issues may take a variety of forms (Exhibit 1). Some investors aim to align their portfolios with their values. For example, a foundation established to promote sustainable agriculture might choose to exclude the stocks of companies that rely on genetically modified organisms (GMOs). Other investors may be exclusively (or at least primarily) concerned about performance. These investors, for instance, may favor companies involved in developing alternative energy technology, believing that such companies will deliver superior risk-adjusted returns over time. To be sure, investors interested in values alignment would almost certainly prefer higher returns rather than lower returns on their portfolios, but the main objective of these investors is aligning their portfolios with their values.

There is no universal approach to ESG investing. Some investors may choose to focus on a single area, say environmental risks, and put less emphasis on other areas, such as social or governance risks. For other investors with broader concerns, a combination of areas may be important. Even investors who target the same general area may prefer contrasting approaches. Consider climate change as an example.

Some investors may care most about companies’ current impact, preferring a portfolio that lowers greenhouse gas emissions; other investors may be most concerned about companies’ future impact, wishing to avoid companies with large fossil fuel reserves; and still other investors may be most interested in companies that help reduce climate change or mitigate its impact, aiming to allocate capital to companies that are developing innovative products or business practices.

In some cases, different groups of investors may focus on very similar issues but hold sharply divergent views about a company’s optimal policy. Activist hedge funds and many Taft–Hartley (union) pension plans, for example, have shown a keen interest in corporate governance. Activist hedge funds, which offer investors “pure play” exposure to governance issues, often encourage companies to streamline their businesses by divesting subsidiaries, increasing financial leverage, and expanding performance-based pay for senior management.¹ Taft–Hartley plans, however, given their connection to organized labor, tend to favor directly opposing policies, namely, less aggressive restructuring, less debt, and more uniform compensation levels across employees within a company. Both perspectives have merits, but reflect contrasting philosophies by investors wishing to put greater emphasis on governance.

Given the many perspectives regarding ESG issues, it is not surprising that there are many ways to incorporate ESG factors into a portfolio. One useful distinction is whether a strategy adopts an integrated approach or whether the strategy incorporates a screen and/or tilt on the basis of one or more ESG factors.

With an integrated investment approach, ESG factors are part of a manager’s regular investment process. A company’s ESG profile is not the only potential determinant of a company’s fundamental value, but it is an important consideration along with other company characteristics. Just as a company’s competitive positioning, regulatory environment, and quality of management could determine a company’s future performance, so too could a company’s exposure to ESG factors.

With screens and/or tilts, ESG factors, together with a corresponding set of rules, define the investment universe relative to a broader selection universe or help to determine the set of overweights and underweights for a portfolio relative to a parent index. For instance, a portfolio could exclude the stocks of companies involved in thermal coal (a screen) or underweight stocks of companies that generate high levels of greenhouse gas emissions (a tilt). The rationale for specific screens and/or tilts reflects the objectives of the investor. As mentioned earlier, some investors may wish to align their values (mission) with their investment portfolio (e.g., a health organization that wants to avoid tobacco stocks). Other investors may be motivated by performance considerations, emphasizing stocks that have a more favorable ESG profile as a way to boost returns and/or dampen risk (e.g., a pension plan that believes more energy efficient companies will deliver superior risk-adjusted returns over time). In the context of the specific set of rules, the ESG factors help to determine the composition and characteristics of a portfolio consistently and transparently.

The choice between the two approaches—an integrated investment process versus a process that involves screens and/or tilts—affects the tradeoff between ESG and non-ESG factors (Exhibit 2). With an integrated approach, it is entirely possible that a stock with a weak ESG profile will be included in a portfolio. ESG factors are a potentially important input, but they are not the only input. If a stock with poor ESG characteristics appears sufficiently attractive along other (non-ESG) dimensions, then a portfolio manager using an integrated approach could conclude that the stock is undervalued even after taking into account ESG factors. Hence, the manager could choose to hold the stock (and possibly even overweight the stock) in the portfolio.

• Download figure
• Open in new tab
• Download powerpoint

Exhibit 2

Interaction between ESG Factors and Non-ESG Fundamental Factors

Notes: The upper-left portion of the chart shows that some companies with favorable ESG characteristics may be unattractive investments. The lower-right portion of the chart shows that some companies with unfavorable ESG characteristics may be attractive investments.

Source: GSAM.

Similarly, with a process that uses screens and/or tilts, it is likely that an undervalued stock will be excluded from a portfolio, or included as an underweight position.² The market may form an overly negative assessment of a stock’s future prospects independently of—or because of—the stock’s poor ESG characteristics; as a result, the stock may trade below its fundamental value and thereby represent a compelling investment (for which maximizing return is the primary objective).

The possibility that a weak ESG stock is an attractive investment or that a strong ESG stock is an unattractive investment parallels the familiar contrast between a “good company” and a “good stock,” where a profitable and well-managed business may be overpriced (“good company” but “bad stock”) and a struggling business may trade below its fundamental value (“bad company” but “good stock”). Investors in ESG-oriented Portfolios may encounter this distinction just as investors in other equity strategies have previously.

Another consideration for investors designing an ESG program is whether they wish to engage with the management of investee companies regarding ESG issues and, if they do, how they should do so (Exhibit 3). Some investors may prefer to focus exclusively on identifying attractive stocks and building efficient portfolios. Investors who have discretion to sell (or not hold) stocks with a poor ESG profile may feel that talking to company management about ESG issues is a distraction. To be sure, engagement requires relevant expertise and involves considerable time and effort.

Nonetheless, some investors may feel that engagement is the best way to bring about improvements in a company’s disclosures, operations, or business model.

Investors who conclude that talking with a company’s management would be beneficial need to develop an engagement strategy by determining which issues to emphasize, how best to motivate any recommendations regarding these issues, and how to respond if a company is unwilling to change. For instance, should investors focus on improving a company’s disclosures or go further by encouraging a company to alter its business practices? Should the rationale for addressing any ESG issues be based on potential business opportunities, underappreciated risks, or improvements to social welfare (corporate citizenship)? If a company resists making any changes, should investors simply be patient and continue advocating change, or does it make sense to sell the position or act via the annual proxy voting process? The strategy that an investor chooses depends on several factors, including the investor’s philosophy, expertise, resources, and time frame.

IMPLEMENTATION

As highlighted above, investors will generally want to form specific views about ESG issues if they pursue an ESG program. Investors will also want to determine the best implementation of a strategy that is compatible with their views. A few implementation-related questions follow.

First, should the ESG strategy depend on current performance (specifically, the level of one or more ESG factors) or the rate of improvement (the change in ESG factors), a potential indication of future performance? For example, an investor concerned about worker safety could choose to focus on the magnitude of a company’s injury rate (the number of injuries scaled by total hours worked), whereas another investor might emphasize the year-to-year reduction in the injury rate. The level reflects a company’s performance at a point in time, which may be relevant for investors who wish to assess the ESG “footprint” of their portfolio. In contrast, the trend may be appropriate for investors who want to anticipate future levels or who want to encourage steady progress among companies and recognize that companies may need several years to show improvement (a grace period).

Second, should the ESG strategy evaluate each company relative to its peers in the same industry, or should a company be compared with a broader universe? Industry is an important (and natural) determinant of many ESG metrics because companies’ types of raw materials (processed vs. unprocessed inputs), production processes (use of capital vs. labor), and byproducts (recyclable materials vs. toxic waste) vary widely across industries. The potential risks arising from a company’s supply chain (e.g., tracking the employment of children) may be much greater for some industries (e.g., retailers) than they are for others (e.g., insurers). Therefore, it may be misguided to apply the same performance benchmark across industries. In other cases, however, cross-industry comparisons may be appropriate, even if there are systematic differences among industries. As an example, the risk associated with bribery and corruption may differ across industries, but that fact does not mean that investors should apply different standards across industries. Moreover, adopting a universal set of principles may increase pressure on companies to continually review and evolve their business model over time.

Third, should the ESG strategy rely on estimated data when reported (actual) data are not available, or should the strategy use reported data exclusively? Specific ESG metrics may be missing because the metrics are not relevant for a company. For instance, some companies may produce so little waste that they do not even bother to track the quantity. Other metrics may be relevant but difficult to measure accurately. In those cases, third-party data vendors may provide estimated values. Consider an electric utility: Even if the utility does not report its greenhouse gas emissions, it may be possible to estimate its greenhouse gas emissions on the basis of the quantity of electricity produced together with information about the utility’s fuel mix (coal vs. oil vs. natural gas). However, although estimated data may be available, some investors may choose not to use those data. Investors may be concerned about the accuracy of the estimated data. Alternatively, some investors may believe that they are “rewarding” a nonreporting company by purchasing its common stock, and that by not holding the company’s stock, they are sending a strong signal to the company’s management. Of course, investors that are able and willing to engage with company management (as discussed above) may find that their engagement efforts are more effective if they hold the company’s stock.

Fourth, how should ESG metrics be scaled to allow meaningful comparisons across companies? Many ESG metrics represent quantities or counts that are a function of a company’s production level. For instance, the emissions for an airline depend not only on the characteristics of its fleet of aircraft, but also on the number of passengers flown and the length of their flights. Similarly, the emissions for an automobile manufacturer are related to the number of vehicles assembled. While scaling ESG metrics by a measure of production makes conceptual sense, it may not make practical sense because measures of production are not defined consistently across industries (revenue passenger miles vs. number of automobiles). Instead, investors and/or portfolio managers typically scale by market capitalization or revenues, measures of size that are consistently defined for most companies.³

Fifth, how much tracking error (active risk) should the ESG strategy target relative to its performance benchmark? The level of tracking error affects the extent to which the strategy can achieve its ESG goals (along with other investment objectives). The tracking error also affects the risk of underperformance relative to the strategy’s benchmark.⁴ If the tracking error is “large,” then the difference between the portfolio and the benchmark due to ESG considerations can be “large” (a more aggressive ESG tilt), but then the potential for the portfolio to lag its benchmark is also “large.” In contrast, if the tracking error is “small,” the difference between the portfolio and the benchmark due to ESG considerations will be “small” (a modest ESG tilt), but then the potential for the portfolio to lag its benchmark is “small.”

The link between ESG factors and predicted tracking error will likely vary by ESG factor, and the link will generally be nonlinear (concave) for a long-only portfolio (Exhibit 4).⁵ For instance, reducing a portfolio’s water usage intensity (cubic meters of water scaled by market capitalization) by 25% may require deviations from the benchmark that result in, say, 20 basis points (0.20%) of tracking error, whereas reducing a portfolio’s water usage intensity by 50% (twice as much) may require a tracking error of, say, 80 basis points (four times as much). At the same time, the link between ESG factors and predicted tracking error will typically differ across ESG factors. For instance, reducing a portfolio’s waste intensity (metric tonnes divided by market capitalization) by 25% may involve a tracking error of 30 basis points (compared with 20 basis points of tracking error for the same 25% reduction in water usage intensity above).

• Download figure
• Open in new tab
• Download powerpoint

Exhibit 4

Relationship between ESG Improvement and Tracking Error for Long-Only Portfolio

Note: The relationship, which can differ for different factors, is nonlinear because of the no-shorting constraint.

Source: GSAM. For illustrative purposes only.

MONITORING

Before completing the design or selection of an ESG strategy—and before allocating assets—an investor and/or portfolio manager should decide how to monitor the strategy after it is launched. Adopting a robust monitoring framework helps ensure the strategy is performing as intended: Are the ESG characteristics, risk exposures, and performance results consistent with expectations? Monitoring an ESG strategy is challenging for at least four primary reasons.

First, in some cases, how to interpret any differences in ESG characteristics between a portfolio and its benchmark may not be obvious. To illustrate, suppose the greenhouse gas emission intensity of a US large-cap portfolio is lower than its benchmark, the S&P 500 index, yet the portfolio actually prefers higher emission stocks within sectors. The overall portfolio nonetheless scores well on greenhouse gas emission intensity because it overweights low-emission sectors (e.g., technology and financial services) and underweights high-emission sectors (e.g., energy and materials). Is this a well-structured portfolio from an environmental perspective? Although the portfolio’s emission intensity appears attractive compared with its benchmark, the improvement is due entirely to the portfolio’s sector allocations rather than its stock selection. Some investors may be satisfied with the portfolio, especially if the sector allocations take into account each sector’s emission intensity. Other investors may be less impressed given the negative impact from individual stock positions, an explicit concern for most portfolio managers.

Second, for some ESG strategies, it may be appropriate to use a primary and a secondary performance benchmark, and depending on their liquidity characteristics, attempting to beat both benchmarks simultaneously may involve a tradeoff. To help illustrate that point, consider a strategy that incorporates screens and/or tilts that are chosen by an investor client and embodied in a specific ESG index, either an “off-the-shelf” index or a custom index designed by the investor. In this case, the investor, having selected the strategy, is largely responsible for the performance impact of the ESG screens and/or tilts, the difference between the ESG index and the investor’s policy benchmark (typically, a standard capitalization-weighted index).⁶ In contrast, the portfolio manager is responsible for the implementation of the ESG portfolio, including how closely the portfolio tracks the ESG index.⁷ These two goals (good ESG performance and low tracking error to an ESG index) are compatible if the ESG index has favorable liquidity characteristics relative to the size of the portfolio (index turnover and trading volume of index constituents). However, if liquidity of the ESG index is insufficient, then it may be necessary to balance the two goals. The compromise between ESG performance and tracking error results from market impact (transaction costs): trading the portfolio aggressively to match the return of the ESG index has the potential to hurt the return of the portfolio relative to the policy benchmark if such trading leads to price reversions once the portfolio and ESG index have been rebalanced.⁸

Third, monitoring an ESG strategy can be a challenge because most widely used models for evaluating portfolio risk and performance do not include ESG factors. That is not surprising given the wide variety and limited coverage of many ESG factors, making it difficult to build a standard, one-size-fits-all risk and attribution model. But if the model does not include any ESG factors, it is not possible to fully quantify their impact using the model. For that reason, investors may want to consider developing a custom risk and attribution model that includes the relevant ESG factors. That approach would allow investors to determine the contribution to risk and return from a portfolio’s exposure to the set of ESG factors, along with industries and common style factors such as size, value, growth, and so forth. Building a custom risk and attribution model, however, can be a significant undertaking that may not be feasible for all investors.

Another way to evaluate the performance of an ESG strategy, which we favor, is to create one or more “paper” (i.e., hypothetical) portfolios that are designed to isolate the impact of any ESG components (Exhibit 5). Compared with building a custom risk and attribution model, constructing paper portfolios is often easier and may produce more intuitive results.

• Download figure
• Open in new tab
• Download powerpoint

Exhibit 5

ESG Performance Attribution of a Managed Portfolio versus a Policy Benchmark Using a Series of “Paper” (model) Portfolios

Source: GSAM.

A final—and potentially much bigger—challenge of monitoring ESG strategies is the long and uncertain time horizon associated with numerous ESG risks. The performance impact of many ESG factors may not occur for several years. Given the complexity surrounding government policy, competitive pressures, and environmental consequences related to ESG issues, it may be a decade or more before climate change, as an example, has a material effect on the stock market. Similar challenges exist for other ESG-related risks such as pollution (improper disposal of toxic waste), worker safety (prolonged exposure to industrial chemicals), or poor corporate controls (manipulation of accounting reports). For that reason, some investors may choose to monitor a portfolio manager’s investment process rather than the investment outcome (i.e., performance). For strategies that use screens and/or tilts, validating an investment process may be as simple as examining the set of stock positions (e.g., confirming that a sharia portfolio does not own any pork producers). For strategies that use an integrated approach, verifying the soundness of the investment process may be more difficult. In particular, investors may need to evaluate the way in which a portfolio manager incorporates ESG factors into the investment process by, for instance, asking about the sources of any ESG data used, the weights placed on different inputs, and the experience of the investment team.

Investors with a horizon of three to five years may find it difficult to maintain confidence in an ESG strategy that experiences a few years of underperformance, even if it retains the potential to beat the broader market for the longer term. Ideally, an investor in an ESG strategy (comprising members of the investment staff and/or board), any constituents (beneficiaries, tax payers, donors, etc.), and the portfolio managers have compatible views about how best to define, measure, and reward success.

GROWTH

Strategies that incorporate ESG factors have existed for many years, but their recent growth suggests that many investors are assessing a broad spectrum of ESG risks for the first time. We believe the number of investors concerned about ESG issues will continue to increase in the coming years, as will the total amount of assets in ESG-oriented portfolios. Nonetheless, the actual pattern of growth is uncertain. At least three issues have the potential to influence the type, speed, and scale of ESG adoption among investors.

First, the availability of ESG data, which are often provided by companies voluntarily, is limited. To be sure, data availability has improved dramatically in recent years.⁹ However, there are still gaps in the number of companies with ESG information, the quantity of metrics with valid values, the timeliness of disclosures, and the accuracy of the reported figures. Because ESG data are critical to the development and management of ESG strategies, deficiencies in ESG data have the potential to lower the appeal of ESG strategies and thereby slow their growth.

Second, the value relevance of various ESG metrics is not always clear. A subset of ESG factors, along with numerous other company characteristics, determines a company’s economic worth. It is likely that many other ESG factors, however, may have relatively little effect on a company’s fundamental value. The contrast between numeric metrics and policy metrics illustrates that point. Coverage of many numeric ESG metrics is often poor, whereas coverage of policy metrics is generally much better. For instance, a company may not produce any hazardous waste and hence fail to report on its hazardous waste intensity (a numeric metric). A data vendor may nonetheless be able to determine whether the company has a hazardous waste policy (likely “no” in this example). For that reason, it may be easier to create an ESG strategy using policy metrics rather than numeric metrics. Policy metrics, however, may provide less information about a company’s fundamental value compared with numeric metrics.¹⁰

The value relevance of an ESG metric can also be unclear because the directional impact of some ESG factors is ambiguous. In particular, it is not necessarily the case that a bigger value for a negative characteristic is, in fact, unfavorable. As an example, consider labor violations. A large number would seem undesirable, yet a high number of reported incidents could indicate a strong work environment to the extent that workers are more comfortable reporting any problems.¹¹

A third issue that has the potential to influence the growth of ESG investing is uncertainty regarding the information content of some ESG metrics. Even if a given metric is value relevant (i.e., helps determine a company’s fundamental value), the metric may have little predictability for the company’s future stock return if the information conveyed by the metric is already impounded in the company’s stock price. Information content, along with value relevance, provides much of the justification for including an ESG variable in an investment strategy.

To assess whether the market is efficient with respect to a given ESG variable—that is, whether there is a lag between the public release of the variable and the stock price reaction to the variable—equity researchers often rely on a backtest. A backtest measures the historical performance of an investment strategy that uses the variable through time to construct portfolios. If the simulated strategy delivers a positive excess return for the sample period, then the signal is both value relevant and predictive (based on the backtest). If the strategy’s excess return is not reliably greater than zero, then either the signal is not value relevant or the signal is value relevant but not predictive, which suggests the market is informationally efficient.¹²

Interpreting the results of a backtest through a statistical lens may be straightforward, but determining the implications of a backtest for future performance can be difficult. To see why, suppose a given ESG variable was useful in predicting stock returns in the recent past (recent because the availability of ESG data is limited). Some investors might conclude the variable should be included in an ESG strategy going forward. But suppose the market realized that the variable was value relevant sometime during the period of the backtest, having previously ignored the variable when setting prices. By trading on the basis of the ESG variable, the market drove stock prices to their true fundamental value during the backtest period, generating outperformance for investors who had positive exposure to the variable in their portfolios.

Although investors could have earned positive excess returns had they used the variable in the past, investors who use the variable in the future are likely to be disappointed. In fact, it is possible that the most effective variables going forward will be variables that failed to predict stock returns in the past (because their information was not yet reflected in stock prices). This is true especially for slow-moving signals like many ESG factors, which produce little turnover in portfolio holdings.

CONCLUSIONS

Given the limited availability of ESG data, establishing the value relevance and information content of ESG metrics is a big challenge. Overcoming this challenge is important because value relevance and information content, when supported by clean data, are used by investors and portfolio managers to form a robust investment thesis (Exhibit 6). Why should a given ESG variable be related to a company’s stock price, and why should the variable be useful in predicting the change in a stock’s price going forward? Most investors are uncomfortable relying on intuition alone to create an investment thesis, yet relying too heavily on empirical research, as highlighted previously, can be unsound as well, given the sometimes counterintuitive implications of historical simulations. For those reasons, forging a strong basis for an ESG strategy, in which investors and their portfolio managers have clearly fulfilled their fiduciary duty, can be a long and complex process. Fortunately, no single investor or portfolio manager has to embark on this process alone. An increasing number of market professionals—asset consultants, portfolio managers, data vendors, and like-minded investors—are analyzing the role of ESG factors in many different types of portfolios and developing a common framework for doing so. In our view, the question is not whether investors will embrace ESG strategies, but rather why, how, and when they will do so.

• Download figure
• Open in new tab
• Download powerpoint

Exhibit 6

Questions Underlying the Formation of an Investment Thesis for an ESG Factor

Source: GSAM.