Development in academic literature
Capelle-Blancard and Monjon (2012) confirm in their research a positive trend for academic articles on SRI between 1982 and 2009, with as many as 100 articles published on this topic in the later years. They display in their paper the major trends in literature on SRI, by using a simple content analysis on relevant phrases. They find, that in 2009 nearly two-thirds of all published academic articles cover performance measurement, which they explain by the fact that this research area is very data driven, as well as by the availability of respective data sets. Furthermore, they confirm new trends in SRI articles, besides performance measurement. The phrases stakeholder, corporate governance, sustainable, activism, and human rights have seen extensive growth in coverage over the researched time period. We follow this perception with our literature review and dedicate a large part of the following section to an overview on performance measurement.
Performance of SR investments
When it comes to measuring performance of SR investments over conventional funds, there seems to be disagreement in academic literature whether SRI is profitable or not. An extensive set of meta-studies on the relationship between financial performance and SRI exists. Mill (2006) names Wood and Jones (1995) who reviewed 60 empirical studies between 1970 and 1994; Pava and Krausz (1996) who looked at 21 studies between 1972 and 1992; Margolis and Walsh (2003) who researched on 127 studies between 1972 and 2002; Orlitzky et al. (2003) who analyzed 52 studies between 1972 and 1997; finally Salzmann et al. (2005), who researched on 15 studies between 1975 and 2001. In addition to this, Hoepner and McMillan (2009) reviewed 51 studies between 1991 and 2007. From these meta-studies, Mill (2006) concludes a positive impact of SRI on financial performance. However, this result might be described as illusory, because a compilation of findings cannot produce a definite conclusion given the limitations of the underlying studies (Barnett and Salomon 2006). We therefore see a need to analyze the link between social and financial performance, by looking at individual studies. We took the meta-studies mentioned above as a first basis and enhanced them by studies published more recently. We further focus only on empirical studies based on market data for our analysis, and not studies based on accounting data or those that present a theoretical model approach. Overall, we examine 53 different empirical studies regarding SRI; of these, 35 use a performance analysis of ethical investment vehicles compared to conventional benchmarks and are therefore qualified to provide an answer to the question if SRI pays for investors. We also examine 18 studies that established relationships between a specific behavior perceived as ethical or unethical by most people and a single company’s financial performance, to show whether SR behavior pays for the company. When conducting our set of studies, we focus mainly on studies of the past 15 years to display the latest trends in empirical analysis. To provide a complete picture of the empirical landscape, we add eight studies published more than 15 years ago. Our review covers a period of 27 years between 1986 and 2012.
From a technical perspective, most studies—49 out of 53—perform regression analysis; of these, 27 provide an explanatory contribution on different factors that influence returns, besides ethical commitment. Here, one-, theee-, or four-factor models are used, such as CAPM, Fama–French, and Carhart. The CAPM considers market beta as an explanatory factor for asset price movements and is used in eight of the analyzed studies. For a detailed derivation on the CAPM, see Copeland et al. (2003:147 et seq.). Fama and French (1995) extend the CAPM by two additional factors, considering beta, size, and book-to-market values, but their approach is only used in two studies, although the size effect is cited by many as an explanatory factor. The Carhart model is based on Fama and French (1995), but with a fourth factor. This factor captures the Jegadeesh and Titman momentum anomaly, i.e., the return difference of past 12 months’ winners and past 12 months’ losers (Carhart 1997; Jegadeesh and Titman 1993). Eight studies use the Carhart model, 15 studies define their own factor models, and 26 do not use a factor model at all to determine different influence factors in their retrieved returns. Five studies use a matched pair approach in addition to a simple regression analysis. Thereby, SRI and non-SRI funds of similar characteristics are compared, to reduce effects in size or investment style when comparing returns. Here, important studies are: Mallin et al. (1995); Gregory et al. (1997); Statman (2000); Kreander et al. (2005). Five studies use other methods such as implementing trading strategies (Gompers et al. 2003) or event studies that focus on only one point in time.
Besides different comparison methods, the analyzed studies also differ in the choice of performance measures and benchmarks they use. Both factors are very significant when determining the performance of SRI vehicles and their conventional counterparts. It must be noted that many studies use more than one performance measure; Fig. 2 provides an overview of the measures that are used. Besides average portfolio returns, which are used in 14 studies, the following performance measures are used.
The Jensen alpha (Jensen 1967) is used in 31 of the analyzed studies. The Treynor ratio (Treynor 1965) is used six times, and the Sharpe ratio (Sharpe 1966), is used in 14 studies. The excess standard deviation adjusted return (eSDAR), as calculated by Statman (2000), quantifies the extra return at a specific point in time that could be earned by a portfolio when using the same amount of risk as the specified benchmark (Cengiz et al. 2010). The measure is used in three studies. Finally, Tobin’s Q (Tobin 1969) is also used in three studies.
In addition to the choice of performance measure, one of the primary goals of performance measurement is to find an appropriate benchmark to compare one’s portfolio to (Henningsen 1992). In their paper Fowler and Hope (2007) provide an overview of the major sustainability indices, which screening methods they use and which benchmark index underlies them. The best-known indices are introduced in detail by Sparkes (2002:295). He discusses the US-based Domini Social Index by KLD and the Dow Jones Sustainability Index, as well as the UK-based FTSE4Good index. The Domini 400 social index (sometimes also referred to as the KLD 400 index, because it was established by Kinder, Lydenberg, Domini and Company) is one of the best-known indices regarding social topics. Launched in March 1990, it is referred to as the “first broad market, common stock index in the United States designed to measure the performance of portfolios subject to multiple social constraints” (Kurtz et al. 1992). When launched, it applied primary screens on military contracting, alcohol and tobacco, gambling, nuclear power, and South Africa, as well as secondary screens on corporate responses to environment, product quality, and attitudes toward consumers, corporate citizenship, employee relations, women, and minorities. In a first step, the screen is applied to the S&P 500, which delivers 257 remaining companies that are included in the index. Furthermore, 40 smaller companies with strong social positives are added to the index and, finally, 100 large capitalization stocks in underrepresented industries that surpass all screens (Kurtz et al. 1992). In the studies, we analyze a very broad range of different indices used as benchmarks.
However, indices have more uses beyond benchmarking. Already, a few studies have compared the performance of SRI indices to conventional indices: Sauer (1997), DiBartolomeo and Kurtz (1999), Statman (2000), and Schröder (2004). As Sauer (1997) points out, the advantage of using indices is that transaction costs, management fees, and differences in investment policy can be eliminated so as to isolate any potential performance implications more clearly.
The studies we analyze can further be distinguished according to sample sizes. Here, one should distinguish between performance studies, which focus on a mostly smaller number of mutual funds or indices and those that establish relationships with single companies to prove a relationship between social behavior and company performance, because these studies usually work with a larger sample size. The average performance study sample size is 125 samples (see Fig. 3). The average relationship study sample size is 198 (see Fig. 4).
When asked on his view on SRI, the 1976 Nobel Prize winner, Prof. Milton Friedman, said: “If people want to invest that way that’s their business. In most cases such investing is neither harmful nor helpful” (Laufer 2003). Numerous studies have been performed solely to determine how SR investments perform compared to conventional investments. While Henningsen (1992) provides a basic introduction to how performance should be measured in a social portfolio, there is as yet no agreement in the academic literature as to how this relationship can be described (Camey 1994; Cowton 1998). Hamilton et al. (1993) formulate three different hypotheses on the relationship between SR investment returns and conventional investment vehicle returns. Ethical funds or indices can outperform, underperform, or perform as well as conventional funds or indices. As Fig. 5 shows, of the 35 studies that compare SRI vehicle performance to conventional benchmarks, 15 conclude that SRI vehicles perform in the same way as their conventional benchmarks, 6 find underperformance, and 14 exhibit outperformance compared to their various benchmarks. We performed a simple sign test on these results, finding that the fraction of studies that find outperformance of SR investments over their conventional counterparties is significantly larger, at a 10 % level, than the portion finding underperformance. The main studies that provide proof of the three relationships are presented below.
SR = conventional portfolio returns
Hamilton et al. (1993) find this first hypothesis consistent with a world in which social responsibility is not priced in the market, i.e., SR investors who want to sell their shares find enough conventional buyers for them, so share pricing is not affected. As Guerard (1997b) phrases it, being a socially conscious investors is at least not “dumb”. This also means that SR companies do not acquire any benefit by acting accordingly, because their cost of capital is not reduced, compared to conventional companies. Within their study, Hamilton et al. (1993) confirm this hypothesis by analyzing the Jensen alpha of 32 SRI mutual funds compared to value-weighted NYSE returns. The following papers provide empirical evidence for the hypothesis that SR investment returns are equal to the returns of conventional portfolios: Guerard (1997a) shows that no significant outperformance can be achieved by means of ethical screening. Sauer (1997) compares a well-diversified SRI portfolio—the Domini Social Index 400—with two benchmark portfolios. His studies find that the SRI portfolio does not underperform the benchmark portfolios; this is confirmed by the use of different performance measures: the Jensen alpha, the Sharpe ratio, and average monthly raw returns and variability. In their study, Teoh et al. (1999) analyze companies divesting from South Africa between 1986 and 1989 and did not find significant different relative performance compared to their benchmark portfolio. For the Australian market, Cummings (2000) compares seven local ethical mutual funds, without finding significant over- or underperformance of those compared to their local benchmark indices. Statman (2000) compares the Domini Social Index between 1990 and 1998 with the S&P500, and also finds no underperformance. When comparing SR mutual funds, he finds that they underperform both indices, but are not worse than conventional mutual funds. In his paper, Schröder (2004) compares US, German, and Swiss SRI funds and also finds no significant underperformance compared to their specific benchmarks. He finds that US investors are overinvested in blue chip stocks, whereas German and Swiss SRI funds invest more in smaller stocks. Bello (2005) uses a matching approach, and shows that between 1994 and 2001 ethical funds did not underperform or outperform conventional funds regarding the effect of assets held, portfolio diversification, and variable effects of diversification on investment performance. The compared portfolios both stay well behind their benchmarks (S&P500 and the Domini 400 Social Index). He uses Morningstar data to determine 42 SR mutual funds (including 3 dead funds) to heal survivorship bias, and matches those with two randomly selected conventional funds of approximately the same net asset size. He uses three different measures to define portfolio performance: Jensen alpha, Sharpe information ratio, and eSDAR. Bauer et al. (2005) researched German, UK, and US ethical mutual funds and find no evidence of significant differences in risk-adjusted returns between ethical and conventional funds between 1990 and 2001. They use the CAPM and Carhart’s four-factor asset pricing model for return calculations. They find that distinct investment styles are used—ethical funds seem to be less exposed to market variability than conventional funds. They also find a catch-up period regarding returns since 1990, possibly due to learning effects, but only average returns between 1998 and 2001. They extend their study to the Australian market using the same methods. They investigate the performance of 25 ethical mutual funds compared to the Worldscope Australian Index, showing equal performance for the period between 1996 and 2003, after a catch-up period between 1992 and 1996 (Bauer et al. 2006); a further analysis performed on the Canadian market comes up with the same results (Bauer et al. 2007). Kreander et al. (2005) study 60 European funds 30 ethical, and 30 non-ethical over the period 1995–2001 using a matched pair analysis, and find no signs of significant outperformance of the ethical funds over the non-ethical funds. They match the funds on the basis of age, size, and investment universe. Their study might contain survivorship bias, but since non-ethical funds and ethical funds should be affected in the same way, this was not seen as problematic. They use log returns to reduce the effect of skewness in the return distribution. Mill (2006) also finds no difference in performance between SRI and conventional funds. What is exceptional in his study is that he compares data for only one fund that has switched its investment style from conventional to SRI. Although he finds no significant underperformance or outperformance, the fund’s variance has been higher for 4 years after the style switch, which might be due to portfolio manager learning effects. Amenc and Le Sourd (2008), focusing on France, look at 2002–2007 and cannot find any outperformance by ethical indices and ethical mutual funds over their conventional alternatives. Researching European ethical fund portfolios between 1991 and 2009, Cengiz et al. (2010) find that, due to title selection restrictions, a worsening of the risk–return profile occurs. Their fund portfolios are split into three clusters: principle-oriented (negative screening approach combined with positive criteria), best-in-class approach and ecology–climate–environment. They conclude that none of their analyzed clusters could beat the benchmark. Only the principle-oriented cluster is little behind its benchmark, while the other two clusters fall well behind. Besides the Hamilton et al. (1993) argument that SRI is simply not priced in the market, another reason why the performance of social and conventional funds is closely correlated might be that the specific holdings in the two different portfolios do not differ from their conventional counterparts as much as expected. Haigh and Hazelton (2004), for example, find that Australian SRI funds are invested in 171 of the 200 largest companies. Table 1 summarizes the discussed studies.
SR < conventional portfolio returns
The second hypothesis by Hamilton et al. (1993) expects SRI portfolios to deliver smaller returns than conventional portfolios. This is supported by Rudd (1981), who argues along the lines of classical portfolio theory. Because social responsible criteria limit fund managers’ allocation possibilities, they lead to additional costs and investment risk, negatively impacting the portfolio’s performance. The possibility of uncompensated risk in a socially screened portfolio is also mentioned by Kurtz (1997); it is perceived to be one of the largest obstacles to SRI implementation. This argument is supported by Luther et al. (1992), who hold that SRI portfolios’ returns might be smaller, due to additional monitoring costs, a restricted investment portfolio, and fewer diversification possibilities. Cowton (1998) argues that it seems likely that SRI fund returns are smaller than those of the various mainstream funds, because mainstream investors could build the same portfolio as SR investors, but not vice versa. Michelson et al. (2004) as well as Tippet (2001) refer to the lower returns of SRI funds as an “ethical penalty.” Although the reasoning above is consistent with classical portfolio theory, only six studies find empirical support for the hypothesis for lower than expected returns of SRI portfolios compared to conventional funds. Mueller (1991) tests ten mutual funds with ethical restrictions and finds significant underperformance compared to conventional funds in this category. An investor loses on average approximately 1 % point of return per year by incorporating ethical considerations when making investments. In his study, Teper (1992:343 et seq.) compares the KLD 400 between 1985 and 1989 with the S&P 500 and finds significant underperformance of the SRI approach. Furthermore he compares a South Africa-free portfolio, a sin-free portfolio (without alcohol, tobacco and gambling stocks), a portfolio without major defense contractors, and a portfolio that eliminates birth control manufacturers. For all portfolios except the major defense contractors, he finds underperformance of the SR investment compared to the S&P 500 between 1979 and 1989. Gregory et al. (1997) conduct a matched pair and a cross-sectional analysis, and concerning both find a tendency to underperform in ethical funds compared to their benchmarks. Kahn et al. (1997) compare the performance of a tobacco-free S&P 500 portfolio with the complete S&P 500 portfolio and find underperformance for the tobacco-free portfolio between 1986 and 1996. In his study, Tippet (2001) shows significant underperformance of the three major Australian ethical mutual funds between 1991 and 1998 compared to their benchmarks. He holds higher transaction costs and management fees responsible for this underperformance. Finally, Geczy et al. (2005) connect the costs of SRI with the investment views of different investors: while investors following a market index realize only a view basis points of underperformance per month, investors looking closer at fund manager skills have much larger costs of up to 30 basis points per month. The six relevant studies in this section are summarized in Table 2.
SR > conventional portfolio returns
The reasoning behind the third hypothesis by Hamilton et al. (1993) that SRI firms’ returns are higher than those of their conventional counterparts is that investors might underestimate the impact that negative news due to irresponsible behavior might have on conventional fund performance. Such negative news would lead to conventional portfolio underperformances, and, vice versa, to SRI portfolio outperformance. This argument is supported by Moskowitz (1972), who finds that good environmental screening does decrease the likelihood of high costs owing to environmental disasters that would decrease conventional portfolio returns. He finds that a good social and environmental performance is a signal of good managerial quality, which might lead to an increase in SRI portfolio returns. Further, he provides a first portfolio of 14 SR companies, but does not draw a conclusion regarding their performance. Finally, SRI mutual funds might yield higher returns because they are subject to more scrutiny—or at least should be—than conventional funds (Schwartz 2003).
The following papers find empirical evidence for higher-than-expected SRI portfolio returns: One of the initial motivations for SRI was the late apartheid regime in South Africa. This led to a global discussion of whether investments in companies doing business in or with South Africa should be excluded. Grossman and Sharpe (1986) examine the effect of divestments from South African companies to receive more ethical investment results. They find that portfolios without South Africa-related companies show superior returns, at the same level of risk, compared to conventional funds. Luther et al. (1992) investigate UK ethical unit trusts and find weak evidence of outperformance of ethical funds over their conventional counterparts on a risk-adjusted basis. However, they consider their results as limited on the basis of being too varied, as well as too closely correlated with low yields to allow for any relation between returns and ethical effects in SRI portfolios. They also find a small company bias and low dividend yields for their screened portfolios. The reason for the small company bias might be that it is very likely to find at least one department in a very large diversified company that might be considered unethical, whereas small companies are much less likely to be allocated to the “unethical” section. In their study covering the years 1986–1993, Mallin et al. (1995) find weak signs of outperformance when comparing ethical funds to non-ethical funds. They find that ethical trust funds outperform non-ethical trust funds, but both perform worse than the market. The study by D’Antonio et al. (1997) identifies 140 SR companies out of the KLD 400 index, and compares their bond performance with the general Lehman Brothers Corporate Bond Index (LCB); they find that the SRI bonds outperform their benchmark, but they credit this result to the differences in credit risk of the investigated portfolio and its LCB benchmark. In 2000, they reevaluated their previous study, looking at differences in allocation strategies and compared mixed equity and debt SRI portfolios with their conservative benchmarks. They look at different investment strategies—such as the buy and hold approach, constant mix, and constant proportion—and find significant outperformance on a strict return basis for all strategies. Considering risk, most analyzed portfolios outperformed their benchmarks, but portfolios with more than 70 % allocated to equity underperformed (D’Antonio et al. 2000). In his study, Travers (1997) focuses on SRI outside the USA. He finds outperformance when comparing 23 selected SRI mutual funds from Europe, Australasia, and Asia with their benchmark, the MSCI EAFA Index (Morgan Stanley Capital International Europe, Australasia and Far East Index). One drawback of this study is the relatively short time frame. DiBartolomeo and Kurtz (1999) show outperformance of the Domini Social Index 400 over the S&P 500 between 1990 and 1999, but find that this outperformance is not generated by some “social factor” but is merely due to macroeconomic effects, the DSI 400’s high exposure to growth-oriented stocks, as well as industry-specific risks. Epstein and Schnietz (2002) divide the Fortune 500 index into three separate groups—environmentally abusive firms, labor-abusive firms and the rest—and look at a specific point in time: the failure of the 1999 WTO talks. They find that, around this event, the first two groups performed significantly poorer than the remaining portfolio. For 1997–2001, Bragdon and Karash (2002) also find outperformance when comparing their self-generated social index with the MSCI World Index as well as the S&P 500 Index. In a study of 1995–2003, Derwall et al. (2005) specifically examine non-environmentally friendly and environmentally friendly stock portfolios and find significant outperformance by the environmentally friendly portfolio. Furthermore, Gompers et al. (2003) show a positive effect of strong corporate governance on company performance. In their study, they set up a governance index, ranking 1,500 firms according to their shareholder rights. They then implement a trading strategy, selling the firms with lowest shareholder rights and buying those with most shareholder rights. With this simple strategy, they earn an abnormally high return of 8.5 % per year. These findings correspond well with the explanation by Tippet (2001) that excluding firms where management behavior is considered unethical should lead to significant outperformance, because firms with unnecessary costs are avoided. Shank et al. (2005) compare 11 firms that are well known for their SR behavior with selected SRI mutual funds and a conventional benchmark. They find that—for the short-term, 3- and 5-year time horizons—neither the selected single firms nor the SRI mutual funds are able to outperform the market. However, in terms of long-term (10-year) performance, the 11 selected firms produce significant positive alpha, thereby outperforming the market. In their study, Hill et al. (2007) researched on SR firms in Europe, Asia, and the USA. They identify a set of SR companies and compare their returns to their conventional benchmarks (S&P 500 for the US, Nikkei 225 for Asia, and FTSE 300 for Europe). They conclude that the examined European companies outperformed their benchmark in the short term, whereas Asian and US companies did not perform significantly differently to their benchmarks. An explanation of this might be differences in national culture’s influence on SRI. Kempf and Osthoff (2007) implement a simple trading strategy: buy stocks with a high SR investment rating, and sell those with a low rating. With this strategy, they achieved an abnormally high performance of up to 8.7 % per year. Especially good results are obtained for application of the best-in-class approach, stocks with extreme social ratings, and a combination of several social screens at a time. The results also hold if high transaction costs are applied. All discussed studies are displayed in Table 3.
To summarize to date the relationship between social responsibility and returns has not been conclusive. Fowler and Hope (2007) conclude that “Despite the considerable research, there is no consensus in the academic or practitioner communities on the relative performance of SRI mutual funds.” A possible explanation for the different results discussed above is provided by Barnett and Salomon (2006), who put forward a possible solution on why the debate on the relationship between SRI and financial performance has gone on for so long. They find that one must decide either to wholeheartedly follow an intensive SR investment screening process and gain extra returns as better managed and more stable firms are selected, or exclude very few firms to retain the ability to diversify. In their approach they do not—as many studies have done before—compare SRI and conventional funds, but address differences within SRI funds and look at different screening methods. They conclude that the relationship between financial and social performance is neither purely positive nor purely negative, but curvilinear, i.e., the best financial performance is at the lowest and highest social responsibility levels. But those funds with a maximum screening of 12 still underperformed by roughly 2.4 % per year compared to those funds that performed only one screen. Therefore, it cannot be concluded that social screening does not come at a cost. The same reasoning has been used before by David Diltz (1995), who finds different performance results, depending on the type of screen he uses on a predefined portfolio. The influence of different screening methods is also held responsible by Tippet (2001), for the different findings in the academic literature regarding the relationship between ethical investments and financial returns: if the screening approach is based simply on a company’s product (e.g., alcohol or tobacco), it is likely that profitable companies are excluded from the portfolios and that underperformance is reported. On the other hand, if the screening is based on management’s ethical behavior (i.e., if independence of auditors or management remuneration are considered), excluding firms behaving unethically in this understanding will most likely exclude companies from the portfolios that bear additional costs and therefore most likely show outperformance compared to their conventional benchmarks. Another explanation for the different results produced by the various studies is that there are so many different definitions of SRI, which might affect the outcome of the research (Sandberg et al. 2009).
But it might be questioned whether SRI funds’ performance is at all relevant for certain investors. In their study, Benson and Humphrey (2008) find that SRI fund flows are overall less sensitive to past returns than conventional fund flows and that SR investors are more likely to reinvest in funds they already own. The flow–performance relationship is asymmetric, i.e., the best-performing SRI funds receive the largest share of the inflow, but the poorest performing SRI funds do not experience the similarly large outflows. They conclude this is due to SRI fund investors’ difficulty to find adequate ethical funds that match their non-financial goals. They have higher search costs to determine a fitting alternative fund. Furthermore, Bollen (2007) finds that cash inflows into SRI funds are more stable and independent from the fund’s past performance, than regular mutual funds.
The assessment of the empirical studies above shows the heterogeneity of results when comparing the connection between social and financial performance. Garcia-Castro et al. (2010) name three potential reasons for the contradictory results. First, they point out that social performance is difficult to assess, and sophisticated measures to do so are still not widely accepted. Further, they state that certain circumstances might influence the relationship between social and financial performance in a way that is not yet completely understood. Finally, the long- and short-term analysis of the relationship between social and financial performance might account for the different results. In their article, they find support for a fourth argument. They show that endogenous effects influence this relationship, as a company’s management makes its strategic decisions not randomly, but based on specific internal information not available to the broad market. Derwall et al. (2011) argue that the different investment styles of investors might be the reason for the variety of results. Regarding the perceived underperformance of SR investments, they argue with the “shunned-stock hypothesis”, whereby value-driven investors use negative screening approaches to derive their investments, thereby screening out non-SR investments which leads to an overall decreasing demand in the market, resulting in relatively lower stock prices. On the contrary, they state that the profit-driven investors who use positive screening approaches can explain the outperformance of these investments. The “error-in-expectations theory” might be accountable for the perceived outperformance of SR investments, as the market continuously seems to undervalue the positive effects CSR might have on a company. Although both theories cannot hold true in the long run, the combination of both screening approaches in practice at the moment might be an explanation for those performance studies, finding similar performance for SR and conventional funds. As Fig. 2 shows, the vast majority of the analyzed performance studies use Jensen’s alpha to determine the relative performance of the portfolio; Galema et al. (2008), however, argue that SRI lowers the book-to-market ratio and therefore alpha is not suited to capture the positive SRI effects. While confirming a positive relationship between financial and social performance, they also provide a possible explanation why many studies find different results when trying to establish a link between alpha and SRI.
Comparison of single companies
As noted, most studies construct SRI portfolios and compare them on a portfolio or index basis to their conventional counterpart. Another approach is to look at single company values to determine the effects of SR behavior, mostly referred to as CSR. A good overview of the status of CSR in Europe is provided by Habisch et al. (2005). They provide qualitative insight into CSR history as well as the status quo in 23 European countries. For an overview on the impact the financial crisis of 2008 had on CSR projects, see Jacob (2012). The major studies on CSR effects on single company values are shown in Table 4. There are two ways to measure company value: by using accounting or market data. Most studies focus on market values, determined by company stock price times shares outstanding, since they want to determine CSR’s effect on shareholder wealth (Mackey et al. 2007). Hill et al. (2007) find differences in European CSR-conscious companies’ performance, compared to Asian and US ones. They conclude that there might be differences in the national culture between these countries and that European investors appear to value SRI higher than Asian and US investors. Mackey et al. (2007) conduct a theoretical study and find a positive correlation between firms participating in CSR and company value. However, this seems rather straightforward, since they use the basic assumption that managers will only take on those SR decisions that improve company value.
They present a supply and demand model, which does not provide consistent results on the relationship between CSR activities and company value. This model shows that if the demand for SR investment opportunities generated by respective investors is greater than the supply, then such investments could generate economic value for the company. Nevertheless, the model also suggests, that if supply and demand conditions are unfavorable, the company’s market value might also be reduced by such measures. There are two different approaches to determine if SR firms have a higher value than conventional firms. It can be shown by event studies, for example, that stock prices rise when a company receives an environmental award (Klaasen and McLaughlin (1996); or vice versa, if information on toxic release, or an environmental crisis is published, the stock price drops significantly (see Hamilton (1995) and Klaasen and McLaughlin (1996)). Dasgupta et al. (2001) confirm these findings for developing countries. Already, the explicit commitment to ethical behavior can have a positive effect on a company’s financial performance, as shown by Verschoor (1998). This finding is contradicted—at least for the defense sector—by Boyle et al. (1997), who report a negative relationship between a company’s performance and its participation in the ethical Defense Industries Initiative. Many event studies do not study CSR as a whole, but focus on specific characteristics such as environment or stakeholders. Some studies show that higher environmental standards are associated with higher market value, measured by Tobin’s Q [for example, Dowell et al. (2000) as well as Konar and Cohen (2001)]. A positive relationship is also shown between a company’s financial performance and its level of integration of environmental issues into its strategic planning process (Judge and Douglas 1998), its Fortune rating on responsibility for community and environment (Brown 1998); there has even been a positive effect on the company value being determined if a company is not named in the context of Toxic Release Inventory (Konar and Cohen 1997). Blacconiere and Northcut (1997) also find a positive relationship between the level of environmental information provided by a company and its financial performance. Apart from environmental relationships, the effects between financial performance and the treatment of a company’s different stakeholders (such as employees, managers, shareholders, customers, creditors, society, government, or suppliers) is also subject of several studies. In their study, Hillman and Keim (2001) show a positive relationship between a company’s market value added and its stakeholder management level, but find a negative relationship between market value added and social issue participation. Waddock and Graves (1997) look at this from another angle and confirm a positive relationship between a company’s quality of stakeholder management and the quality of its social performance. Looking at customers as company stakeholders, Ogden and Watson (1999) prove that there is a positive relationship between customer satisfaction and financial performance, by analyzing the ten largest privatized water companies in the UK. Further, Jones and Murrell (2001) focus their research on a company’s employee stakeholder group and prove a positive relationship between financial performance and whether or not a company is among the top family-friendly rated companies in the USA. This is supported by Edmans (2011), who shows a positive relationship between a company’s financial returns and whether it is ranked under the 100 best companies to work for in the USA. Wright and Ferris (1997) find a negative relationship between divestments in South Africa between 1984 and 1990 due to the apartheid regime and companies’ financial performance; they argue that the political pressure on the company managers was not well perceived by the capital markets. This negative relationship should be challenged, as Teoh et al. (1999) have conducted the same type of event study between 1986 and 1989 and find no negative or positive effect between divestiture announcements concerning South Africa and a company’s financial performance. Finally, Humphrey et al. (2012) use SAM Rating data to describe the relationship of CSR behavior and financial costs. They do neither find a positive or negative relationship, confirming that at least in the UK market, the implementation of CSR strategies should have no financial consequences for a company. To complete the picture, it has to be stated that CSR behavior does not only affect a company’s cost of equity, as determined by the studies mentioned above. Goss and Roberts (2011) follow a different approach by determining the effect CSR consistent behavior has on the interest rates companies pay for their bank debt. They argue that the bank has a specific duty in monitoring a company’s behavior before signing a loan agreement. They find that companies with a low perceived CSR behavior pay up to 18 basis points more for their loans than companies that are perceived to have a positive CSR behavior.
Regional focus and time frame of SRI studies
Besides categorizing the performance of our sample, we also analyze it with regard to the regional areas the studies cover to determine potential new research areas. The vast majority of studies focus only on the USA. Four studies compare the USA to selected European countries, six focus on the UK, and only two compare selected European countries (see Fig. 6).
When focusing on Europe, the UK is the primary country that is analyzed, which is due to the long history of SRI in the UK. Other studies compare European SRI vehicles in general without naming a specific country. As Fig. 7 shows, if Europe is considered in SRI studies at all, the vast majority researches the UK followed by other core European countries. Southern or Northern European countries are mostly neglected, and we find no studies on Eastern European countries (as displayed in Fig. 7).
Sandberg et al. (2009) state that cultural differences might be one explanation for heterogeneity in the field of SRI and that it is quite unlikely that these differences might be resolved in the near future. Most empirical studies discussed above focus on the US and the UK, as both countries have the longest SRI traditions as well as significant assets under investment. Recently, some studies have been published that compare different SRI aspects in different countries. (Maignan and Ralston 2002) compare the web sites of French, Dutch, UK, and US firms, and find that the Anglo countries tend to focus more on SRI topics then their French and Dutch counterparts. One reason they provide for this might be the traditional responsible role of government in Europe regarding social issues, whereas in the Anglo countries, companies have had more autonomy regarding social issues. Bauer et al. (2005) find that in the USA, compared to the UK and Germany, different investment styles are used. While US ethical mutual funds are mainly invested in large caps compared to their conventional peers, UK and German ethical mutual funds are heavily exposed to small caps compared to conventional mutual funds. Schröder (2004) finds the same differences between US and German funds, but can also confirm that in Swiss SRI funds, like in German funds, small caps are overweighed. Chapple and Moon (2005) conduct a study of seven Asian countries (India, Indonesia, Malaysia, the Philippines, South Korea, Singapore, and Thailand) and find that social responsibility varies between these countries, which might be explained by national factors. They also confirm an enhancing effect of globalization on CSR in Asia. It is at least questionable, whether their study is representative as their main source of data are company website reports. Louche and Lydenberg (2006) compare SRI in the European Union and the USA and find that although the concepts are the same for both areas, the implementation differs. They conclude that SRI will become a mainstream investment pattern much quicker in Europe than in the USA owing to governmental influence and attention relating to SRI in the former. In the USA, proxy voting is mainly used to influence companies to act in greater accordance with SRI principles, which seems to be an active and more confrontational approach. Other differences include community development investing, a hot topic in the USA, but not yet relevant in Europe, as well as some differences in definitions. However, the main concepts and religious roots are the same for both regions.
Furthermore, we took a closer look at the timely distribution of the analyzed studies, as Fig. 8 outlines. We found that the vast majority of the performance studies were published between 1997 and 2005, which account for more than 65 % of all analyzed studies, with a clear maximum of 11 studies in 1997. The declining number of performance studies from 2006 onwards points to a shift to new fields of research. This is in line with Capelle-Blancard and Monjon (2012) who, in their paper on trends in SRI literature point out that the topic of performance measurement has been somewhat overemphasized in the academic world. They also address the need for more conceptual and theoretical work in this research area.