Socially Responsible Investing

The launch in the United States last month of the iShares KLD Select Social Index Fund (ticker KLD: .50% expense ratio) makes this an appropriate time to examine the logic behind what is known as "socially responsible investing." We will first consider the theoretical basis for SRI: What does it mean for a company to be "socially responsible?" Should we expect socially responsible companies to deliver superior risk adjusted returns? If not, then why invest in them? We will then look at the way SRI works in practice, starting with the companies that rate "social responsibility", the indexes that are derived from these ratings, and the performance of funds that only invest in socially responsible companies. Finally, we will answer a simple question: should you tilt your equity allocation towards socially responsible companies?

The Theoretical Basis for SRI

When examining the logic behind socially responsible investing, the first, and ultimately the most difficult challenge one encounters, is defining what constitutes a "socially responsible" company. Writing in 1776, Adam Smith quite clearly equated "socially responsible" with "profit maximizing" in his famous book, An Inquiry Into the Nature and Causes of the Wealth of Nations. Assuming that all externalities (e.g., the indirect costs of pollution) are taken into account (i.e., either included in prices or via taxes and regulations imposed by government), an economy in which firms pursue their own self-interest (by seeking to maximize their profits) will also be one that maximizes the overall welfare of its citizens. Either because they disagree with Smith's conclusion, or because they doubt that all externalities are fully taken into account, ever since 1776 people have advocated alternative views about the proper social role of the corporation. The most recent manifestation of this controversy is found in calls for "corporate social responsibility" and "socially responsible investing."

Unfortunately, beyond disagreeing with Smith, the people and organizations calling for CSR and SRI don't completely agree on what it means. In general, classification as a "socially responsible firm" is the result of a two step process. The first stage is a "negative screen" that eliminates companies that operate in certain, presumably "socially irresponsible" industries. However, the firms that assign "social responsibility" ratings don't agree on which industries belong in this category, which usually contains some mix of alcohol, gambling, tobacco, weapons manufacturing, nuclear power, and sometimes chemicals. As you can see, ultimately any such characterization has to be based on values about which people can and do disagree. Moreover, these "negative screens" also raise questions about where to draw the line, so to speak. For example, while tobacco companies are out, banks that lend to them can still be considered "socially responsible." The same is true for the difference between chemical companies and companies in other industries that use chemicals as inputs into their production process.

The second stage of the "socially responsible" classification is a so-called "positive screen" that assigns points for behaviors and practices that, based on some criteria, are considered "socially responsible." Unfortunately, there is no agreement on what these are. For example, consider the differences in criteria used by two firms that assign social responsibility ratings to firms. One's list includes community relations, employee relations, workforce diversity, environment, human rights, and product quality and safety; the other's includes governance and ethics, safe and health work environment, environment, product safety and impact, international operations and human rights, indigenous peoples' rights, and community relations.

Assuming one has settled on a definition of the characteristics of a "socially responsible" company, the next question is whether there is any evidence that social responsibility (so defined) is associated with superior risk adjusted returns. Perhaps the strongest evidence is found in the area of corporate governance. In theory, this is not surprising. The separation of ownership and management in a modern corporation creates what is called an "information asymmetry", which gives rise to "principal-agent conflicts." In practical terms, this means that managers can potentially take advantage of shareholders because they have better information about what is going on at the company. The purpose of "corporate governance" is to minimize these conflicts, through a variety of means (e.g., eliminating anti-takeover defenses, using more outside directors, improving accounting disclosures and the like). Multiple academic studies have found a significant positive relationship between good corporate governance and higher shareholder returns (e.g., see "Corporate Governance and Equity Prices" by Gompers, Ishii and Metrick). Yet even within this area, there is disagreement about the specific factors that are causing the observed beneficial effects. For example, in "Corporate Governance and the Returns on Investment" by Gugler, Mueller and Yurtoglu, the authors find that being located in a country that uses a common law (English) based legal system has a substantial impact. However, in another paper ("What Matters in Corporate Governance?" by Bebchuk, Cohen, and Ferrell), the authors find that only six factors (all related to the degree to which management is entrenched) have strong correlations with shareholder returns

Beyond corporate governance, there is also substantial evidence linking firm performance to some of the other factors included under some definitions of "social responsibility", including certain employee relations, health and safety, and quality practices. However, it is equally legitimate to argue that these practices are also the hallmarks of good management, which maximizes profits.

The real difference of opinion between Adam Smith and the leaders of the CSR/SRI school is about factors for which no strong link to shareholder value creation has been demonstrated. For example, one firm that rates companies on their "social responsibility" takes into consideration such factors as whether "the company's chief executive officer is a woman or a member of a minority group," whether the company "has implemented notably progressive policies toward its gay and lesbian employees," and whether "the company has as part of its basic mission the provision of products and services for the economically disadvantaged." To our knowledge, there is no body of research that demonstrates a statistically significant link between these factors and superior risk adjusted returns.

Using factors like these, which lack a clear link to value creation, as criteria for identifying "socially responsible" companies, implies a value judgement that they are worth pursuing, even if they worsen risk adjusted returns. For some, this undermines the case for socially responsible investing (see, for example, "The Myth of Social Investing" by Jon Entine). Others, however, make two additional arguments in support of SRI

The first argument is often called "stakeholder theory." Its advocates propose that a company's managers owe allegiance not just to shareholders, but to other "stakeholders" as well, including customers, suppliers, employees, communities and governments. Unfortunately, stakeholder theory does not provide a criterion (like maximizing shareholder value) for making trade-offs between the competing interests these groups. Critics argue that this opens the door to bigger agency problems (and lower equity returns), as managers maximize their personal satisfaction by spending resources owned by shareholders on other stakeholder groups (see, for example, "Value Maximization, Stakeholder Theory, and the Corporate Objective Function" by Michael Jensen). They also point out that in many countries, the law is quite clear that the primary obligation of managers and people who invest other people's money is to maximize risk-adjusted returns.

The second argument used to justify corporate consideration of theoretically weak social responsibility criteria is based on a particular view of investors. Specifically it assumes that, rather than simply seeking to maximize their risk adjusted returns, investors actually derive their satisfaction from a mix of financial returns and the "warm glow" that comes from being a shareholder in a socially responsible firm (see, for example, "A Modigliani-Miller Theory of Corporate Social Responsibility" by Small and Zivin). If this argument is true, then these investors should be willing to accept lower returns in exchange for a higher level of social responsibility spending by the companies in which they invest. Indeed, the '"activist" school of socially responsible investing claims that investors (and others) should undertake political action to force companies to do this. While acknowledging the theoretical elegance of this argument, critics note that it does not appear to be supported by the data. For example, there has not been a great inflow of funds into socially responsible investment funds that deliver performance significantly below their less responsible benchmarks. In fact, poor performance by these funds has led to outflows of assets. Moreover, other research has found that investors in socially responsible funds are more, rather than less performance sensitive than other investors (see, for example, "Socially Responsible Investors and Performance Sensitivity" by Bollen and Cohen). This suggests that for many people, socially responsible investing is more of a luxury good than a necessity.

Investing in SRI Funds

Let's now move on to the practical issues associated with socially responsible investing. As previously noted, one key issue in this area is the profusion of firms that are rating companies on their "social responsibility", and the different criteria that they use. A closely related problem is the inherently subjective nature of this process, involving as it does "active management" type decisions about (1) the screening criteria to use, (2) the weight to give each criteria, and (3) the inherently subjective nature of a company's "score" on many of these criteria.

A recent report ("Values for Money: Reviewing the Quality of SRI Research") by Mistra, a foundation established by the government of Sweden, highlighted other problems: "Company disclosure was by far the most significant single source of information [for firms assigning SRI ratings], accounting for 40% to 80% of the information input [into the ratings process]." Not only was this information hard to verify, but "one of the most common comments made by companies about the SRI research organizations is the time it takes to respond to numerous requests for information." Many of these companies reported that they were suffering from "questionnaire fatigue." Another comment often raised by companies was that "many SRI research organizations do not understand [the company's] business, and that the [SRI ratings] methodologies are not focusing on key company and sector specific issues.

After they have been identified (on the basis of a set of screening criteria), socially responsible companies are often combined into a "socially responsible" index. There is no shortage of such indexes available today, from both global providers (e.g., the Dow Jones Sustainability Indexes and the FTSE4Good indexes), and national or regional ones (e.g., the Calvert and KLD/Domini indexes in the United States). The most common approach is to use market capitalization weighting. If one believes that socially responsible companies will generate superior risk adjusted returns, then market capitalization weighting maximizes the financial benefit to an investor. We note, however, that if you believe that corporate social responsibility is valuable in its own right, even if it results in lower financial returns, then one should logically use a weighting scheme based on companies' "social responsibility" scores. The KLD Select Social Index (upon which the iShares ETF is based) uses a variant of this approach. It is constructed using an optimization model (pioneered by TIAA-CREF) that maximizes the "social responsibility score" of the index, subject to a maximum tracking error constraint (in this case, 2% per year) versus an external benchmark (in this case, the Russell 1000 Index). This raises another key point about socially responsible indexes: most of them contain very different risk exposures than the non-socially responsible indexes against which they are often benchmarked. For example, at the end of December, 2004, the FTSE4Good global index had, relative to the FTSE Developed Countries Index, 5.4% more exposure to the financial sector and 3.6% more exposure to non-cyclical services, but 4.7% less to general industrials and 2.6% less to resources. It also had 7.9% less exposure to the United States, but 6.4% higher exposure to the UK, and 2.3% higher exposure to France. Other studies have shown similar differences in socially responsible indexes' exposure to other risk factors, including small versus large capitalization shares, and value versus growth style shares.

Numerous studies have examined the performance record of socially responsible investment funds. Unfortunately, many of them have been criticized for employing questionable methodologies. The stronger studies have taken into account the differing risk factor exposures (e.g., to the market, small caps, value, or momentum) of socially responsible funds. This makes it possible to see whether the socially responsible screens themselves (as opposed to the differing factor risk exposures they produce) had any impact on the funds' performance (technically, they test to see if the SRI screens produced any alpha). In most cases, they find that they do not. For example, in "International Evidence on Ethical Mutual Fund Performance and Investment Style", Bauer, Koedijk, and Otten review performance in Germany, the U.K. and the U.S. and "find little evidence of significant differences in risk adjusted returns between ethical and conventional funds for the 1990 - 2001 period." They also find that actively managed socially responsible funds failed to outperform the socially responsible indexes over this period. In "Socially Responsible Investments in Germany, Switizerland and the United States," Michael Schroeder reaches the same conclusion.

In "Canadian Ethical Mutual Funds: Performance and Ethical Style Analysis," Bauer, Derwall and Otten find that their "Canadian results are consistent with the perception that the performance differential between ethical mutual funds and their non-ethical peers is statistically insignificant." And in "Ethical Investing in Australia," Bauer, Otten and Rad find that "Australian ethical mutual funds underperform [on a risk adjusted basis] both their relevant indices and their conventional peers between 1992 and 2003." Finally, in "The Performance of Socially Responsible Bond Funds", Derwall and Koedijk find "a positive but statistically insignificant performance differential between socially responsible bond funds and their conventional peers over the period 1987 to 2003" in the United States.

While not adjusted for their differing factor exposures, the following comparison between different socially responsible U.S. index funds and the Vanguard Total Market Index Fund is also interesting. The data cover the three years ended December 31, 2004.

As you can see, the Vanguard Total Market Index Fund outperformed all of the socially responsive index funds. However, the difference between it and the TIAA-CREF Social Choice Equity Fund does not appear statistically significant. This is not surprising. Both CSXAX and DSEFX are market capitalization weighted funds. In contrast, TCSCX employs an optimization methodology intended to maximize the fund's weighted social responsibility score within the constraint that its return and risk don't deviate too far from the Russell 3000 (which includes about 98% of U.S. equity market capitalization, versus 100% in the Wilshire 5000). This is the same approach used by the new iShares KLD Select Social Index ETF. However, rather than using the broad Russell 3000 as its benchmark, it takes a large cap tilt and uses the Russell 1000. Also, at .50%, its expense ratio is also somewhat higher than the TIAA-CREF fund.

Conclusion

Should you tilt your equity (or even your bond) allocation toward socially responsible investments? For better or worse, the answer to that question seems to lie more in the realm of values than it does financial economics. To be sure, some of the criteria used to identify "socially responsible" companies seem to be associated with higher shareholder returns. Foremost among these seem to be some of the criteria associated with corporate governance. Unfortunately, as with all successful active management screening criteria, now that this has been publicized an efficient market should eliminate the future potential for excess returns, by bidding up the price of well-governed companies. However, the majority of criteria used to identify "socially responsible" companies do not seem to have a clear link to shareholder value creation. If you believe that a tilt toward "socially responsible" companies will produce superior risk adjusted returns, you will likely be disappointed.
This does not necessarily mean that you should avoid funds that invest in socially responsible funds. If you derive non-monetary satisfaction from investing in companies identified as "socially responsible" under some set of criteria, and are willing to sacrifice some financial returns to obtain this satisfaction, there is nothing wrong with that. However, if you choose to go this route, we would strongly suggest doing it using socially responsible index funds. Being socially responsible does not make active management any less challenging. Fortunately, there is a growing number of socially responsible index products available around the world. We prefer the ones that employ an optimization methodology that tries to keep the fund's returns and risk close to those of an underlying conventional benchmark index (e.g., the Russell 1000 for the new KLD ETF, and the Russell 3000 for the TCSCX mutual fund in the United States). However, the maximum tracking error versus the benchmark index that these funds allow (2% in the case of KLD) is very large in the world of index funds. Also, most actively managed funds take on tracking error risk in order to generate higher returns (alpha). In this case, tracking error is being taken on to generate a higher social responsibility score. This raises an interesting question: how many people would take the risk of paying 2% more in interest just to get a mortgage from a socially responsible bank?

When all is said and done, we remain unconvinced that for most people the level of non-monetary satisfaction that accompanies socially responsible investing exceeds the amount that comes from investing in a conventional broad market index fund, and contributing the higher earnings on this investment (ideally along with your time) to a charitable cause whose benefits you can observe and enjoy first-hand.