Does Social Investing Generate Higher Returns?

by Denys Glushkov and Meir Statman

Editor’s note: This is based on the authors’ article, “The Wages of Social Responsibility,” copyright 2009, CFA Institute. Reproduced and republished from the Financial Analysts Journal, Volume 65, Number 4, with permission from the CFA Institute. All rights reserved.

Socially responsible investments have attracted much money, many investors, and many studies. We have studies of socially responsible mutual funds, socially responsible indexes, “sin” stocks, stocks with good and bad environmental records, and stocks with good and bad employee relations. But some parts of our knowledge are inconsistent with other parts and some gaps in our knowledge remain.

The Social Investment Forum, a national nonprofit organization promoting the concept, practice, and growth of socially responsible investing, describes socially responsible investing as “an investment process that considers the social and environmental consequences of investments, both positive and negative, within the context of rigorous financial analysis.”

Screening is the most prevalent form of socially responsible investing, followed by shareholder advocacy and community investing. Negative screening excludes or reduces the portfolio weights of companies with weak environmental, social, or governance records, and positive screening includes or increases the portfolio weights of companies with strong records.

Negative screens that exclude tobacco companies have been the most popular screens among socially responsible mutual funds, followed by screens that exclude companies associated with alcohol, gambling, and weapons. Negative and positive screens related to community relations come next in popularity, followed by screens related to the environment, labor relations, products and services, and equal employment.

Previous studies have shown no statistically significant difference between the returns of socially responsible mutual funds and those of conventional funds. While useful, these studies teach us little about the relative returns of stocks of socially responsible companies since managerial skills and expenses create gaps between the returns of stocks and the returns of mutual funds that contain these stocks, and these gaps vary from fund to fund.

Comparisons of the returns of indexes are free of the confounding effects of managerial skills and expenses that plague comparisons of the returns of mutual funds, but such comparisons do not provide a clear picture of differences between the returns of socially responsible stocks and conventional stocks since there is much overlap between the list of stocks in socially responsible indexes and conventional indexes. For example, differences between the returns of the Domini 400 Social index (called DS 400 here; now the KLD 400) and the S&P 500 index likely understate differences between the returns of stocks of socially responsible companies and stocks of conventional companies because the two indexes share approximately 250 companies.

Moreover, social responsibility criteria vary among indexes and so do their relative weights. The Calvert index assigns relatively high weight to corporate governance, while the DS 400 index assigns relatively high weight to the environment. The DS 400 index excludes companies with any interests in nuclear power plants, while the Calvert index excludes such companies only if their interests are substantial. And while the DS 400, Calvert and Citizens indexes exclude all tobacco companies, The Dow Jones Sustainability index does not.

We sought to close the gap of knowledge about returns associated with characteristics of social responsibility, such as diversity and employee relations. We found that stocks of companies with high ratings on social responsibility characteristics outperformed companies with low ratings. However, we also found that ‘shunned’ stocks outperformed stocks in other industries. ‘Shunned’ stocks are defined more broadly than ‘sin’ stocks as stocks of companies in the alcohol, tobacco, gambling, firearms, military, and nuclear industries. The two effects largely balance out, such that socially responsible indexes have returns that are approximately equal to those of conventional indexes.

Hypotheses About Stock Returns

There are three alternative hypotheses about the relative returns of the stocks of socially responsible companies and conventional companies. The first hypothesis is the “doing good but not well” hypothesis, where the expected returns of socially responsible stocks are lower than the expected returns of conventional stocks. This hypothesis might be true if the benefits of company actions that tilt it toward social responsibility fall short of the costs and investors, on average, know that.

One study (by Amir Barnea and Amir Rubin in 2006) described a scenario where ordinary shareholders are shortchanged by their companies’ socially responsible actions. But shareholders might not be shortchanged. Instead, socially responsible investors might be willing to sacrifice returns for social responsibility. Consider stocks of ‘sin’ companies associated with alcohol, tobacco and gambling. The activities of such companies are seen as violating social norm, and some socially responsible investors avoid them even if they yield higher returns than stocks in other industries.

The second hypothesis is the “doing good while doing well” hypothesis, where the expected returns of socially responsible stocks are higher than those of conventional stocks. This is possible if managers and investors consistently underestimate the benefits of being socially responsible or overestimate its costs. For example, a company with high employee satisfaction might generate higher profits over the long term, even if the perceived worker happiness increases costs over the short term.

The third and last hypothesis is the “no effect” hypothesis, where expected returns of socially responsible stocks are equal to the expected returns of conventional stocks. The “no effect” hypothesis might be true if company actions that tilt it toward social responsibility are costless, such as when actions amount to no more than words. The “no effect” hypothesis might also be true when costly company actions, such as better employee relations, increase benefits by as much as they increase costs, such that company profitability is not affected. This can happen, for instance, when the extra costs of higher employee pay are equal to the extra productivity benefits of more satisfied employees.

The “no effect” hypothesis might be true even if, in fact, the increase in costs exceeds the increase in benefits, as long as investors, on average, continue to overestimate the benefits of social responsibility actions or underestimate their costs. Such overestimation of benefits or underestimation of costs might come, for instance, from ‘rosy scenarios’ in the minds of socially responsible investors who are overly optimistic about the positive effects of employee satisfaction on employee productivity.

Last, the “no effect” hypothesis might be true if aspects of social responsibility that are consistent with the “doing good while doing well” hypothesis are counterbalanced by other aspects that are consistent with the “doing good but not well” hypothesis. This is indeed what we found.

Stocks of companies with good records on employee relations and similar social responsibility criteria have earned higher returns than stocks of companies with poor records. This is consistent with the “doing good while doing well” hypothesis. But we also find that stocks of ‘shunned’ companies—including those of companies associated with alcohol, tobacco and gambling—also earned higher returns than stocks of companies in other industries. This is consistent with the “doing good but not well” hypothesis. The two effects counterbalance each other such that the “no effect” hypothesis prevails.

Ranking Companies on Social Characteristics

KLD Research and Analysis, Inc. assigns a score of “1” when a company demonstrates strength on an indicator on its list of social characteristics (e.g., charitable giving) and zero if it does not. Similarly, it assigns a score of “1” when a company’s record raises concern on an indicator on its list (e.g., investment controversies) and zero otherwise. The score of a company on a given characteristic is the difference between the number of its strength indicators and the number of its concern indicators.

The scores of companies in the DS 400 index of socially responsible companies were generally better than those in the S&P 500 index of conventional companies. The mean score of the DS 400 index companies was 0.65, higher than the –1.55 mean score of the S&P 500 index companies.

Investors are likely to continue to debate which characteristics make a company socially responsible, but the distribution of the KLD scores of the companies in the S&P 500 index and the DS 400 index highlights the observation that companies are arrayed in a range; no company is perfectly socially responsible or irresponsible. Moreover, companies with the same overall KLD score differ in their characteristics scores.
For each company each year, we calculated the company’s score in each of the seven characteristics of social responsibility (community, diversity, employee relations, environment, products, human rights, and governance) as the difference between the number of strength indicators and the number of concern indicators. Our sample included scores between 1991 and 2006.

Among the 2,955 companies are 198 companies ‘shunned’ because of an association with tobacco, alcohol, gambling, firearms, military or nuclear operations. They include some companies that are members of the S&P 500 index, but not the DS 400 index.

Our classification of ‘shunned’ companies is broader than the stereotypical classification of ‘sin’ companies which is the triumvirate of alcohol, tobacco and gaming. Our classification of shunned companies follows KLD and includes not only companies associated with defense or military operations but also companies associated with nuclear operations. Shunned companies are those that KLD classifies as associated with one or more of tobacco, alcohol, gambling, firearms, military or nuclear operations.

Performance of Socially Responsible Portfolios

We find, in general, that stocks of companies with high social responsibility scores yielded higher returns than stocks of companies with low scores. Excess returns in equally weighted top-bottom portfolios are positive and statistically significant for the community, employee relations and environment characteristics but not for the diversity and products characteristics. Excess returns in the human rights and governance characteristics are negative.

Evidence elsewhere on the relation between corporate governance and stock returns is conflicting. We find no statistically significant relation between governance and stock returns, consistent with the “no effect” hypothesis.

Companies that rank high by community, employee relations, environment and products tend to be growth companies, while those that rank high by diversity, human rights, and governance tend to be value companies. Companies that rank high on community, employee relations and diversity tend to be relatively large, while those that rank high on environment, products, human rights and governance tend to be small. Companies that rank high on all social responsibility characteristics tend to tilt toward high momentum, but such tilts are far from statistical significance.

The generally higher returns of stocks of companies with high social responsibility scores are especially evident in a long-short portfolio of companies that are ‘top-overall’ and ‘bottom-overall.’ We define a top-overall company as one that is in the top third of companies by two or more social responsibility characteristics and not in the bottom third by any characteristic. Similarly, we define a bottom-overall company as one that is in the bottom third of companies by two or more social responsibility characteristics and not in the top third by any characteristic. We exclude governance because it was added to the KLD list only in 2002. The portfolio is tilted toward growth stocks and stocks with high momentum. But there is no significant tilt toward large or small-cap stocks.

Many socially responsible investors shun stocks of companies associated with alcohol, tobacco, or gambling, but many also shun stocks of companies associated with firearms, military or nuclear operations. We refer to companies in these six groups as ‘shunned’ companies and refer to companies outside these groups as ‘accepted’ companies. We constructed a long-short portfolio of stocks of ‘accepted’ and ‘shunned’ companies at the end of 1991, reconstituted it at the end of each subsequent year and recorded its monthly returns through the end of September 2007. We find that the ‘accepted minus shunned’ portfolio had a negative annualized excess return. The portfolio is tilted toward small growth stocks. But there is not much tilt toward momentum stocks or away from them.

The effect on returns of the ‘positive screen’ of tilting toward stocks of companies with high scores on social responsibility characteristics offsets somewhat the effect on returns of the ‘negative screen’ of excluding stocks of ‘shunned’ companies. We see that offset in the performance of a portfolio long in the socially responsible DS 400 index and short in the S&P 500 index. The tilt toward stocks of companies with high scores on social responsibility characteristics increases the return of the DS 400 index relative to the return of the S&P 500 index by more than the exclusion of ‘shunned’ companies from the DS 400 index decreases it. But the excess returns of the DS 400–S&P 500 long-short portfolio have smaller magnitudes than the excess returns of the top-overall minus bottom-overall portfolio or the ‘accepted’ minus ‘shunned’ portfolio.

Management of Socially Responsible Portfolios

Our findings indicate that portfolio managers who want to construct high-performing socially responsible portfolios should construct best-in-class portfolios tilted toward stocks with high social responsibility ratings. However, practical portfolios deviate from the portfolios we analyzed in several ways.

First, practical portfolios are likely to be value weighted or close to value weighted, while the portfolios we analyzed are equally weighted.

Second, practical portfolios deviate from the portfolios we analyzed because managers want portfolios with a high likelihood of positive excess returns, but they usually do not insist that such likelihood be higher than common statistical significance levels of 95% or even 90%. A 60% likelihood of positive excess returns is good, a 70% likelihood is better, and a 95% likelihood is even better.

Third, practical portfolio managers want to assure themselves that excess returns are robust, not due to positive excess returns during the first half of an overall period and negative excess returns during the second half.

The signs of excess returns are stable during the overall period, positive for the ‘top-overall minus bottom-overall’ and the ‘DS 400 minus S&P 500’ portfolios, and negative for the ‘accepted minus shunned’ portfolios.
In sum, the results indicate that excess returns in value-weighted portfolios are lower and less reliable than excess returns in the equally weighted portfolio. However, excess returns continue to favor best-in-class portfolios tilted toward stocks with high social responsibility ratings.

Conclusion

Typical socially responsible portfolios, such as the DS 400 index, are tilted toward stocks of companies with high scores on social responsibility characteristics such as community, employee relations and the environment. We analyzed returns during 1992–2007 of stocks rated on social responsibility by KLD and find that this tilt gave socially responsible portfolios a return advantage relative to conventional portfolios. This finding is consistent with the “doing good while doing well” hypothesis, where the expected returns of stocks of socially responsible companies are higher than those of conventional companies.

However, typical socially responsible portfolios also shun stocks of companies associated with tobacco, alcohol, gambling, firearms, military, and nuclear operations. We find that such shunning brings to socially responsible portfolios a return disadvantage relative to conventional portfolios. This finding is consistent with the “doing good but not well” hypothesis, where the expected returns of socially responsible stocks are lower than the expected returns of conventional stocks.

The return advantage that comes to socially responsible portfolios from the tilt toward stocks of companies with high scores on social responsibility characteristics is largely offset by the return disadvantage that comes to them by the exclusion of stocks of ‘shunned’ companies. The net effect is consistent with the “no effect” hypothesis where expected returns of socially responsible stocks are approximately equal to the expected returns of conventional stocks. This is consistent with a world where the social responsibility feature of stocks has no effect on returns. But it is also consistent with the world we find, where return advantages of some social responsibility criteria are offset by return disadvantages of other social criteria.

Socially responsible investors can do both well and good by adopting the best-in-class method for the construction of their portfolios. That method calls for tilts toward stocks of companies with high scores on social responsibility characteristics, such as community, employee relations and the environment, but refrains from calls to shun the stocks of any company, even one that produces tobacco.

Denys Glushkov is a research support director at Wharton Research Data Services, University of Pennsylvania, Philadelphia.
Meir Statman is a Glenn Klimek professor of finance at Santa Clara University, Santa Clara, California.