«Sustainable Investment Amsterdam, 31 August 2010 Commissioned by Duisenberg School of Finance and Holland Financial Centre Sustainable Investment ...»
A key assumption underlying the 2nd hypothesis is that conventional portfolio managers do not use all value-relevant information, which is at odds with the market efficiency theory: some claim that, since SRI portfolios are based on public information such as CSR issues, they cannot generate a better return than ‘normal’ funds (Barnett & Salomon, 2006; Bollen, 2007; Halkos & Sepetis, 2007; Harold et al., 2007; Renneboog et al., 2008a, 2008b; Soppe, 2004). This Efficient Market Hypothesis (EMH) refers to “a market where, given the available information, actual prices at every point in time represent very good estimates of intrinsic values” (Fama, 1970, p. 90).21 The third hypothesis claims that SRI portfolios have different risk exposures and therefore different expected returns than conventional portfolios.
Most of these mutual fund studies are unable to conclude that SRI underperform or outperform conventional funds, as most research brings forward statistically insignificant results (Benson, Brailsford, & Humphrey, 2006; Benson & Humphrey, 2008; Harold et al., 2007; Mulder, 2007;
Plinke, 2008; Renneboog et al., 2007, 2008b).
Usually a taxonomy of three EMHs are distinguished (Fama, 1991): the weak form of efficiency (the information set includes only the history of prices), the semi-strong form efficiency (the information set includes all information known to all market participants; i.e., all publicly available information) and the strong form of efficiency (the information set includes all information known to any market participant, including private information).
The Fama-French pricing (or three factor) model (1993) evaluates fund performance. It consists of the capital asset pricing model (CAPM) plus two additional factors: the market capitalization factor (SMB) and the book-to-market factor (HML).
Or, vice versa, from ‘sin stocks’ being underpriced due to a lack of demand.
Only a few show significant results that point either at underperformance or outperformance of SRI-funds, but this is only in specific regions/countries or dependent on performance period (the short, medium or long run) which is discussed in paragraph 3.3.3 respectively 18.104.22.168 Renneboog et al. (2008b) give an excellent and extensive overview of findings of studies on the performance of SRI funds/portfolios:25
The general conclusion that there is no (statistically significant) difference in performance between SRI funds and conventional funds, even though social and environmental standards restrict the investment universe of SRI funds, has spurred the debate on whether SRI funds really differ from conventional funds (i.e., do they invest according to social and environmental standards), or whether they are merely conventional funds in disguise. Kempf & Osthoff (2008) explore this subject. They find that US SRI equity funds indeed have higher ethical ranking and therefore are not conventional funds in disguise.
Plinke (2008) offers a similar synopsis. This can be found in Table 3 of Appendix A.
The quality of financial data used in SRI studies is subject of some debate. This is a research area in its own and first and foremost the expertise of methodologists. Therefore it is beyond the scope of this paper. A good starting point for further reading on the subject of date quality is Chatterji & Levine (2005) and Hoepner & McMillan (2009).
NSD = Differences between funds are not statistically different OP = SRI-fund outperformed conventional fund UP = SRI-fund underperformed conventional fund Source: Adapted from Renneboog et al. (2008b) Additional research on SRI fund performance includes a study by Galema et al. (2008), which is not included in the review by Renneboog et al. (2008b) and which points at a positive relationship between (positive) investment screens and fund performance. Galema et al. (2008) find that SRI has a significant impact on stock returns. This is in particular the case for portfolios that score positive on diversity, environment and product.
SEO ECONOMIC RESEARCHSOCIALLY RESPONSIBLE INVESTMENT 19
3.3.3 Regional Differences Survey reviews of performance differences between SRI and conventional funds note that whether (statistically significant) differences occur, is partly dependent on the geographical region that is studied (Renneboog et al., 2007, 2008b).
United States and United Kingdom In general, for SRI funds in the US and UK there is little evidence that the (risk-adjusted) returns of SRI funds are different from those of conventional funds. Statman (2000), for example, compares 31 SRI funds with 62 non-ethical funds of similar size. The outcome is measured against both the S&P 500 Index and the Domini 400 Social Index (DSI 400), the most wellknown SRI index. The performance measures are different for the two types of funds but these results are not significant. Therefore, it suggests that SRI funds and conventional funds do not differ in performance. Hamilton et al. (1993) reached a similar conclusion 7 years earlier.
Goldreyer & Diltz (1999) researched fund performance in the US of 49 SRI and 20 non-SRI funds between 1981 and 1997. The SRI funds include equity funds, 9 bonds funds and 11 balanced funds. The difference between the two type of funds is again not significant, which leads the authors to conclude that there is no evidence to conclude that SRI funds show a different performance than non-SRI funds. Within the group of SRI funds, it appeared that those selected by positive screening outperform those of other screening methods. Therefore, one may conclude that screening methods influence funds performance.27 This subject will be discussed further in the next section. One evident shortcoming of these findings is that they are based on a relatively small sample of 29 funds.
Bauer et al. (2005) find that in the United States and United Kingdom, one cannot conclude that SRI funds generated higher return. For this they compared the performance of 103 SRI funds with 4,384 non-SRI funds over the period 1990-2001.
Other UK studies also fail to find significant differences between ethical and conventional funds (Gregory et al., 1997; Luther & Matatko, 1994; Luther et al., 1992; Mallin et al., 1995; Renneboog et al., 2007).
Continental Europe and Asia-Pacific SRI funds in Continental Europe and the Asia-Pacific region show mixed performance results relative to benchmark portfolios.
Renneboog et al. (2008b) find that during the period of January 1991 until December 2003, in Continental Europe and the Asia-Pacific region, the SRI funds underperformed non-SRI funds.
Bauer et al. (2005) analyzed SRI funds in Germany. In Germany it appeared that SRI funds went through a learning phase. In the beginning of the 1990s, SRI funds showed underperformance compared to conventional funds, but caught up and over the period of 1998-2001, both type of Interestingly, the previously mentioned studies that were omitted by Renneboog et al. (2008b) also suggest a positive correlation between positive investment screens and fund performance.
funds generated similar returns. Another outcome of the research was that SRI funds launched before the end of 1997, performed better than SRI funds launched since 1998.28 Kreander et al. (2005) are unable to find statistically significant underperformance or outperformance of SRI funds in Belgium, Germany, The Netherlands, Norway, Sweden, Switzerland and the UK. Schröder (2004) also found no significant difference between ethical funds and a benchmark portfolio.
In Australia, international SRI funds outperformed conventional funds, while domestic SRI funds underperformed their counterparts (Bauer et al., 2006).
3.3.4 Short Run Versus Medium and Long Run Results Hill et al. (2007) studied SRI funds in the United States, Europe and Asia. They found different results for funds in the short (3-year), medium (5 year) and long run (10 year) until the year 2005 (this has a statistical origin: the data set is composed of weekly security prices from January 1, 1995 through August 8, 2005). In the medium term, the different measures of performance between SRI and non-SRI funds are not significant for all three regions. In the short run, only Europe’s SRI funds outperformed their conventional counterparts and in the long run this is the case for both Europe and the United States. The results for Asia in the long run are almost significant which might be interpreted as Asia catching up with the other two regions.
3.3.5 Multifactor Models In recent years, a series of authors have focused on the marginal effect of ESG-related variables, thereby trying to disentangle the effect of SRI screens from other portfolio management decisions. This latter wave of studies uses multifactor models, thereby using more sophisticated econometric models in order to incorporate non-quantifiable fund aspects. A detailed discussion of underlying econometric methodologies is beyond the scope of this paper. Of relevance here, is that these studies generally reach a more positive verdict on the question whether SRI funds outperform conventional funds.
The primary advantage of multifactor models is that they control for non-quantifiable aspects such as momentum effects, management skill and mutual fund style (Bauer et al., 2005, p. 1765;
Derwall et al., 2005, p. 52; Kempf & Osthoff, 2007, p. 913, 2008; Renneboog et al., 2007, p. 25).
Bauer et al. (2007, p. 112) argue that “not using a multifactor model to evaluate ethical funds can lead to an erroneous assessment of mutual fund performance [since without] multifactor models, we cannot separate returns associated with social investment policies from the returns on common investment styles that do not incorporate those policies”.
Derwall et al. (2005), for instance, measure the performance of portfolios that are selected by means of positive screening (based on environmental performance criteria).29 Portfolios comprising shares with a positive sustainability rating outperform a portfolio with companies with low environmental scores by 6% per annum, over the period of 1997-2003. The authors Bollen (2007, p. 685) hints that rational learning could explain the difference in performance between young and mature funds, although his empirical findings do not support this hypothesis.
They compare 30 % of US companies with the best CSR ratings with 30 % of companies with the worst CSR ratings, using Innovest ratings.
conclude that financial institutions can improve their profitability by taking into account the environmental information of a portfolio.
Kempf et al. (2007) perform a similar portfolio analysis, comparing 10 % of companies with the best CSR ratings with 10 % of companies with the worst CSR ratings, using a 4-factor financial model and socially responsible ratings from the KLD Research & Analytics. A strategy of buying stocks with high socially responsible ratings and selling stocks with low socially responsible ratings leads to high abnormal returns of up to 8.7% per year. In other words, portfolios with a negative sustainability rating produced a weaker performance than portfolios with a positive sustainability rating, even after taking into account reasonable transaction costs.
Edmans (2010) analyzes the relationship between employee satisfaction and long-run stock returns. He finds that a portfolio of the “100 Best Companies to Work For in America” (companies with a good working environment) exhibits significantly higher returns than the (adjusted) market portfolio.30 This leads him to conclude that the stock market does not fully value intangibles, and that certain SRI screens may improve investment returns.
3.4 SRI Success Drivers Performance is not the same for all SRI funds and some studies indicate there might be differences between conventional funds and specific SRI funds (see previous paragraph). Possible factors that might influence the profitability of individual SRI funds are the type of investments screens, the deduction of fees before or after the funds return and fund management.
Investment Screens SRI investors invest in companies that show corporate social responsibility. They select these companies based on investment screens (section 3.1). The type of screens impacts performance of SRI-funds.
Goldreyer & Diltz (1999) conclude that SRI funds with positive screening outperform SRI funds that do not use these types of screens, thereby supporting the hypothesis that investments screens affect the performance of SRI funds. However, as discussed previously, these results are based on a small sample. Renneboog et al. (2008a) also conclude that the screening activities have impact on the return: funds adopting a community involvement policy (excluding firms that have a poor record of accountability to local community stakeholders) or employing an in-house SRI research team to screen portfolios, have better returns than SRI funds without such process policies.