Performance of Socially Responsible Investing : evidence from indexes

(1)

(2)

Performance of Socially Responsible Investing:

evidence from indexes

BSc Thesis

Finance & Organization

Yuri Roukes - 10994122

June 2018

Abstract

This study investigates the performance of Socially Responsible Investing (SRI) using evidence from SRI Indexes. The performance is measured using two multiple factor models in a period between 2008 and 2018, using the monthly data sets. The Fama and French three-factor model and five-factor model are applied for the regression of 22 different SRI indexes. This study shows that both of the models explain a lot of the variation in the monthly returns of the indexes. The results from this study suggest that SRI investing has under-performed the benchmark in the time period analyzed.

(3)

1 Statement of Originality

This document is written by Yuri Roukes who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document are original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(4)

2 Introduction

2.1 Socially Responsible Investing

Socially responsible investing, often named sustainable investments or ethical invest-ments, has become a major trend in the financial industry and has grown rapidly around the world. According to the data of Bank of America, Socially Responsible Investing has grown by more than 97 percent globally in the past 20 years. The value of dollar invested has grown from tens of billions to hundreds of trillions in mutual funds and Exchange Traded Funds.

Socially Responsible Investing (SRI) apply a set of screens and criteria to include or exclude certain stocks based on their environmental, social and governance (ESG) impact. Because of this rapid growth in SRI interest, it has become a key topic in financial research. The ongoing debate in all studies on SRI is the economic viability of socially responsible investing. Many critics believe that socially responsible activ-ities may cost resources of a firm, putting it in an economic disadvantage to firms that are less socially active (Aupperle, Carroll, & Hatfield, 1985). Managerial focus on the ESG factors of a company may distract the directors from the key aspect of a company, value maximization for the shareholders. Another argument is that if investors exclude some non-ethical companies from their portfolios, it will result in a segmented market. Finance theory suggest that the effects of equity market segmen-tation are a higher cost of equity (Merton, 1987). Higher cost of equity will reduce the economic profit from a firm’s activities (Angel & Rivoli, 1997).

According to Renneboog, Ter Horst and Zhang (2008), SRI investors are less concerned about the financial performance of their investments and also strive for non-financial utility they gain from investing in companies that are consistent with their own values. The discussion about the performance of socially responsible investing gives us the following research question:

“Does Socially Responsible Investing diminish investment returns.”

Performance of Socially Responsible Investing mutual funds has been studied mul-tiple times in the 21st century. (Hamilton, Joe & Statman, 1993; Statman, 2000; Bauer, Koedijk & Otten, 2005; Bauer, Derwall & Otten, 2007; Galema, Plantinga & Scholtens, 2008). However, the number of studies on the performance of the con-structed SRI indexes is small. One of the reasons for the lack of study is that most of the SRI indexes are constructed after 2007. In this study the research is solely based on the SRI constructed indexes.

Having the study solely based on SRI indexes will provide a clearer view in the performance of the stocks selected by the ESG criteria. When using investment funds for the regression, the study has to take into account possible management fees of

(6)

mutual funds. The performance of mutual funds is also more focused on market timing instead of gaining an advantage through ethical screening.

Using a sample of 22 SRI or equivalent ESG indexes, the study addresses the question if there are any differences in performance between SRI indexes and their conventional benchmark. The sample for this study will be discussed in the data section of this paper. Regarding the motivations for an under- or out-performance of the socially responsible investing indexes, it is important to test this empirically. The performance will be tested using two mulitple factor models: the Fama-French (1995) three-factor model and the Fama-French (2015) five-factor model. Accord-ing to Fama & French (2015), this last model is better in predictAccord-ing a reliable risk adjusted return because of the additional firm characteristics that are taken into ac-count. After the regression, the results of this research will be presented. Finally, the results will provide a contributing conclusion with any limitations of this study and recommendations for future research.

2.2 Relevance of the study

The relevance of this subject is the fact that there are signs about major investors who are increasingly concerned that Socially Responsible Investing (SRI) hurts their potential returns. Although Socially Responsible Investing is widely discussed and fundamentally important, the real answer is not really clear for many of the individu-als whom are interested in investing. Critics and regulators question whether Socially Responsible Investing benefits investors, a common concern about SRI is that there is a premium to be paid for being socially responsible that necessarily diminishes the investment returns. Contribution of the conclusion drawn from this research is that individual investors may have more information about the importance of SRI and its potential costs in terms of investment returns. If the conclusion is drawn that SRI does hurt investment returns, then SRI will stay a niche market for those who also strive for non-financial utility (Renneboog, Ter Horst and Zhang 2008). If the conclusion is positive in the sense that SRI does not reduce investment returns or even boost returns, SRI will potentially become easier to integrate in the portfolios of investors.

2.3 Literature review

According to Friedman (1970), a company should only care about profit maximization in the interest of their shareholders. According to Rennebooga et al. (2008) investors may derive utility other than financial from investing in SRI companies meeting high corporate social responsibility, and therefore care less about the financial performance

(7)

of their portfolios than the conventional or non-SRI investors. The so-called SRI in-vestors are less concerned about not meeting the performance of the benchmark than conventional investors (Rennebooga, Horst and Zhang, 2008). The SRI investors care more about the fact that companies focus on their environmental and social welfare instead of value maximization. Bollen (2007) argues that investors have a utility function which is not solely based on the standard optimization of risk and reward but they also have a set of personal and societal values which are taken into account. Rennebooga et al. (2008) conclude that SRI has experienced rapid growth around the world, reflecting the increasing awareness of investors to social, environmental, and governance issues. They also conclude in their paper that if investors prefer ‘aversion to unethical corporate behavior’ instead of value maximization, the SRI investors may require a lower rate of return from the firms with a high standard in social welfare. This is one of the reasons that a SRI or ESG index may underper-form their conventional non-SRI benchmark. According to Hamilton et al. ”socially responsible” investors favor certain companies over others according to criteria such as production of weapons or use of alternative energy sources. They found that socially responsible mutual funds do not earn statistically significant excess returns and that the performance of such mutual funds is not statistically different from the performance of conventional mutual funds (Hamilton, Jo & Statman, 1993). Bauer, Koedijk, and Otten (2005) found no evidence of significant differences in risk-adjusted returns between ethical funds for a period between 1990–2001. However, Kempf and Osthoff (2007) report significant positive risk-adjusted returns. Their research was based on the long-term performance of a portfolio consisting of SRI stocks from the KLD database. They invested 10% in the best scoring SRI stocks and shorted the 10% worst scoring SRI stocks during 1992-2004.

The methodology used to evaluate SRI performance has changed over time and there is no standard found for the best regression. The differences in the methodology used also show a difference in the conclusions made. Some of the first papers about the performance of a SRI portfolio use a single index model (Hamilton et al., 1993; Sauer, 1997). The more recent studies applied a multifactor model for their research. Including the size, book-to-market and momentum factors (Fama and French,1993; Carhart, 1997) to evaluate SRI performance (Bauer, Koedijk & Otten, 2005; Ren-nebooga, Horst and Zhang, 2008). In this research both the Fama-French three- and five-factor model is used, including small-minus-big, high-minus-low, robust-minus-weak and conservative-minus-aggressive. A five-factor model directed at capturing the size, value, profitability, and investment patterns in average stock returns performs better than the three-factor model of Fama and French (Fama and French, 2015). Fama and French provides evidence that the five-factor model explains between 71% and 94% of the variation in expected return for the portfolios formed.

(8)

3 Data

3.1 Sample

This study investigates the performance or 22 international Socially Responsible In-vesting indexes. These equity indexes have been constructed by MSCI (Morgan Stan-ley Capital International). The reason to choose for one supplier is the fact that it mitigates any unreliable differences in the ESG criteria that the supplier uses. The sample consists of 22 different ESG or SRI indexes covering 5 investment regions. For each region or country, the market proxy provided by Fama and French their database is used. Fama & French (2015) also used the one-month treasury bill rate as a proxy for the risk-free rate. From the MSCI index family, 22 indexes are included from five different profiles: the ESG Leaders, ESG Select, ESG Focus, ESG Universal, SRI standard and the well-known KLD 400 Social index. The KLD 400 Social was the first launched SRI index, constructed by KLD Research & Analytics. It has recently been acquired by MSCI and is now part of their index family.

The 22 SRI or ESG indexes cover different international investment areas. Eight indexes have a global investment area, of which four indexes exclude the United States. Six indexes focus on the European area, of which three indexes focus on European Monetary Union (EMU) countries only. Five indexes focus on the United States, one of them is the KLD 400 Social. The other three indexes cover the Asia-Pacific area. All indexes are so-called performance indexes and do cover all the investment returns, including the dividend payments of the companies. All indexes were converted to US dollars.

3.2 Data from Morgan Stanley Capital International

As mentioned, the indexes used are constructed by MSCI. The data used is directly gathered from their database for the most accurate representation. For this research, 22 different indexes are used from 5 different regions. The regions used are: Europe, United States, Global, Global excluding United Stated and Asia-Pacific excluding Japan. All the indexes focus on only large-cap or large- and mid-cap. The dataset that MSCI uses for constructing the indexes is the MSCI ESG KLD STATS dataset. This dataset is created by Kinder, Lydenberg, Domini Research & Analytics (KLD) in 1991. The screening criteria used for the dataset is the ESG criteria and those consist of environmental, social and governance (ESG). When Morgan Stanley Capital International acquired KLD Research & Analytics, it became the criteria norm for every ESG or SRI index they have created. MSCI is now the single largest SRI index creator. The MSCI ESG KLD STATS dataset uses 13 measures that belong to four key ESG performance indicator. The 13 measures consist of environmental, social,

(9)

governance and 12 controversial business involvements such as controversial weapon production (MSCI, 2018). MSCI provides multiple SRI index categories with different characteristics, below I’ll shortly discuss the characteristics of each category used:

ESG Leaders: The MSCI ESG Leaders Indexes target companies that have the highest environmental, social and governance (ESG) rated performance in each sector of the parent index. The ‘best-in-class’ companies are chosen for the index. According to MSCI their factsheets, the highest rated companies according to the MSCI ESG Rating make up a 50% market capitalization of their sector for the ESG Index. The MSCI ESG Rating gives a company a score based on their impact on the environmental, social and governance risks. They analyze all the companies that are within the indexes that MSCI constructs, they rate the companies on a scale from ‘AAA‘ to ‘CCC’ scale. The ESG Leaders indexes don’t exclude any companies based on their products or services offered. (MSCI, 2018).

MSCI ESG Select: The MSCI ESG Select Index is for companies that show positive environmental, social and governance (ESG) factors while having the same risk and return characteristics as their MSCI parent Index. The indexes are con-structed in the same way as the ESG Leaders index but additionally over-weights any company with high MSCI ESG ratings. In contrast to the Leaders indexes, the Select indexes exclude producers of alcohol, tobacco, gambling, firearms, military weapons and nuclear power, are not included (MSCI, 2018).

MSCI ESG Focus: The MSCI ESG Focus Indexes are also designed with the best-in-class approach but the companies get their index weight in a way that the index has exactly the same risk and return characteristics as their MSCI par-ent index. The ESG Focus indexes exclude producers of tobacco and controversial weapons.(MSCI, 2018).

MSCI ESG Universal: According to the MSCI fact sheet, the MSCI ESG Uni-versal Indexes aim to provide a very diversified investment universe while improving the ESG exposure. The index gives more exposure to and over-weights companies that demonstrate both a higher MSCI ESG Rating (based on the current MSCI ESG Rating) and a positive ESG trend (based on the current MSCI ESG Rating Trend). The MSCI ESG Universal Indexes exclude only companies that are in violation of human, labour and environmental rights and companies that produce controversial weapons (MSCI, 2018).

(10)

MSCI SRI Standard: The SRI Indexes are capitalization weighted and provide exposure to companies with highest Environmental, Social and Governance ratings and exclude all companies that produce products which have a negative social or environmental impact (MSCI, 2018).

MSCI KLD 400 Social: The MSCI KLD 400 Social Index is one of the first So-cially Responsible Investing (SRI) indexes and was launched in 1990 by KLD Research & Analytics. The MSCI KLD 400 Social Index provides exposure to companies with high Environmental, Social and Governance ratings and exclude all companies that produce products which have a negative social or environmental impact. The index is constructed with 400 companies selected from the MSCI USA IMI Index and selects the companies with the highest ESG ratings in each sector. It excludes companies that produce alcohol, tobacco, gambling, civilian firearms, military weapons, nuclear power, adult entertainment and genetically modified organisms. This index can be found in almost any academic literature about SRI performance (MSCI, 2018).

3.3 Data from Fama French

The data used for the regression is provided by Kenneth R. French his database (2018). This database provides the market returns, risk free rate of the one-month US treasury bill and data of the additional 2 and 4 factors SMB, HML, RMW and CMA. SMB (Small Minus Big) is the average return on the small stock portfolios minus the average return on the big stock portfolios. HML (High Minus Low) is the average return on the value portfolios minus the average return on the growth portfolios. RMW (Robust Minus Weak) is the average return on the robust operating profitability portfolios minus the average return on the weak operating profitability portfolios. CMA (Conservative Minus Aggressive) is the average return on the con-servative investment portfolios minus the average return on the aggressive investment portfolios (French,K.,2018).

(11)

4 Regression

To test the research question, two Fama-French models will be used. At first, this research performs a regression of the 22 indexes with the use of Fama and French three-factor model. Secondly, the same indexes will be regressed using the extended five-factor model. According to Fama and French, the five-factor model performs better than the three-factor model of Fama and French (Fama & French, 2015).

The Fama and French three-factor model:

RSRI− rf = αi+ βi(RBM − rf ) + siSM B + hiHM L (1)

The Fama and French five-factor model:

RSRI− rf = αi+ βi(RBM − rf ) + siSM B + hiHM L + ciCM A + riRM W (2)

Firstly, the stock returns are calculated in excess of the risk-free rate to obtain excess returns of the SRI indexes. The market premium also has to be calculated, we obtain the premium by calculating the excess returns of the benchmarks with regards to the risk-free rate given by the K. French database. The regressions are performed using StataSE 15. StataSE performs a linear regression analysis with the use of the least squares method. This method fits a line through the set of observations and provides an overview with results how the dependent variable is affected by the independent variables, the three or five factors. The regressions for each SRI index are performed separately. The regressions are applied to 22 indexes for the monthly time period from the start date as given by table 1 till the 31th of January, 2018. The monthly excess return of the SRI index is the dependent variable, the independent variables are the risk adjusting factors of the Fama-French model.

The explaining variables of the three-factor model are the risk premium of the market, small minus big and high minus low. For the five-factor model, the ex-plaining variables are extended with the factors robust-minus-weak and conservative-minus-aggressive (Fama & French, 2015). The β provides the exposure to the market premium, s, h, c and r are the exposures to the other factors. The resulting intercept is the one this research is most interested in, the alpha (α), The alpha shows us the abnormal performance for the time period of the index. With the use of all the alpha’s we can answer this paper its research question.

(12)

5 Results

Table 1 shows an overview of the 22 indexes including the start date of the data, monthly mean excess returns, standard deviations and their Sharpe ratio. The table also provides the monthly mean excess returns, standard deviation and Sharpe ratio for the benchmark, abbreviated by BM. The Sharpe ratio measures the excess return divided by the total risk of the investment.

SharpeRatio = µi− rf σi

(3) Where µ = monthly mean return, rf = risk-free interest rate (U.S. 1-month treasury bill) and σ = standard deviation of the logarithmic returns. The mean, standard deviation and Sharpe ratio of the indexes are compared to their benchmark for the same time interval. The time interval is at maximum ten years. If the available data is less than ten years, the full available data history is used rounded to the month.

SRI Index Name Start Date Mean SRI Mean BM σ SRI σ BM Sharpe SRI Sharpe BM KLD 400 31/01/2008 0.00720 0.00780 0.04280 0.04530 0.16880 0.17180 USA ESG FOCUS 30/11/2010 0.01070 0.01140 0.03150 0.03270 0.33870 0.34750 USA ESG LEADERS 31/01/2008 0.00680 0.00780 0.04360 0.04530 0.15570 0.17180 USA ESG SELECT 31/01/2008 0.00700 0.00780 0.04340 0.04530 0.16140 0.17180 USA BROAD ESG LEADERS 31/01/2008 0.00710 0.00780 0.04500 0.04530 0.15690 0.17180 EU SRI STANDARD 31/01/2008 0.00310 0.00390 0.05780 0.05870 0.05320 0.06610 EU ESG LEADERS 31/01/2008 0.00180 0.00390 0.05660 0.05870 0.03250 0.06610 EMU ESG LEADERS 31/01/2008 0.00240 0.00390 0.06770 0.05870 0.03480 0.06610 EMU ESG UNIVERSAL 30/11/2009 0.00410 0.00720 0.05810 0.04830 0.07070 0.14990 EUROPE ESG UNIVERSAL 30/11/2009 0.00460 0.00720 0.04810 0.04830 0.09600 0.14990 EMU SRI STANDARD 31/01/2008 0.00280 0.00390 0.06690 0.05870 0.04150 0.06610 WORLD ESG FOCUS 31/01/2008 0.00450 0.00590 0.04690 0.04780 0.09700 0.12370 WORLD ESG LEADERS 31/01/2008 0.00440 0.00590 0.04650 0.04780 0.09450 0.12370 ACWI ESG LEADERS 31/01/2008 0.00470 0.00590 0.04670 0.04780 0.10140 0.12370 ACWI ESG UNIVERSAL 30/11/2009 0.00700 0.00940 0.03740 0.03750 0.18640 0.25150 EAFE SRI STANDARD 31/01/2008 0.00280 0.00400 0.05290 0.05260 0.05250 0.07570 WORLD EX USA ESG LEAD. 31/01/2008 0.00220 0.00400 0.05200 0.05260 0.04200 0.07570 WORLD EX USA SRI 31/01/2008 0.00270 0.00400 0.05210 0.05260 0.05270 0.07570 ACWI EX USA ESG LEADERS 31/01/2008 0.00270 0.00400 0.05230 0.05260 0.05230 0.07570 PACIFIC EX JAP. SRI 31/01/2008 0.00230 0.00520 0.06460 0.06340 0.03580 0.08200 PACIFIC EX JAP. ESG LEAD 31/01/2008 0.00250 0.00520 0.06370 0.06340 0.03990 0.08200 ASIA PAC. EX JAP. ESG LEAD 31/01/2008 0.00710 0.00520 0.05750 0.06340 0.12300 0.08200

Table 1: Mean monthly returns and Sharpe ratios

The table shows that in 21 out of the 22 cases the monthly mean excess return is higher for the benchmark. However, the higher mean return could indicate that there is a higher risk exposure, so the risk-adjusted returns should be compared (Sharpe, 1975). Comparing the Sharpe ratios provides the same results, in 21 out of the 22 comparisons the benchmark has a higher Sharpe ratio than the SRI or ESG index.

(13)

5.1 Results from the three-factor model

5.1.1 Results from European markets

Table 2 provides an overview of the results from the three-factor regression on the six European SRI indexes. The average adjusted R-squared is 0.9710. This means that the three-factor model explains 97,10% of the variation in the SRI index ex-cess return. The high adjusted R-squared is because of the significant independent variables: market excess return, SMB and HML. The table with results provides the information that all of the parameters are significant and three of the six alphas are significant lower than zero. None of the alphas are positive.

Index αi βi si hi Adj Rˆ2 SER

Europe Standard SRI -.0017184** (.0007358) 1.015807*** (.0148213) -.1224793*** (.0387805) -.1557169 *** (.0351736) 0.9813 0.00794 Europe ESG Leaders -.0026283***

(.0005679) .9823784*** (.0114398) -.1548111*** (.0299327) -.0820046*** (.0271487) 0.9884 0.00612 EMU ESG Leaders -.0008807

(.0015106) 1.05367*** (.0300388) -.2321807*** (.080132) .230868*** (.0722314) 0.9413 0.0164 EMU ESG Universal -.0024428**

(.0012176) 1.087614*** (.0287457) -.3907412*** (.0709407) .2310099*** (.0589978) 0.9608 0.01157 Europe ESG Universal -.0022306***

(.0004642) .9951688*** (.0109592) -.2069147*** (.0270461) -.0764698*** (.0224928) 0.9917 0.00441 EMU Standard SRI -.0017203

(.0012094) 1.090781*** (.0243617) -.2559184*** ( .0637435) .1279611** (.0578148) 0.9623 0.01304 Table 2: Results from European regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.1.2 Results from Global markets

Table 3 provides an overview of the results from the three-factor regression on world SRI indexes and table 4 provides the results of world SRI indexes that exclude the United States. The average of the adjusted R-squared is 0.9893, which means that 98,93% of the variation in the Global SRI indexes excess returns is explained by the factors of the three-factor regression model. Just as the European results, a lot of the variation is explained. However, the h parameter is not significant in any of the regressions.The HML factor has no explanatory power in the global regressions, this is in line with the research of Fama and French (2015) who suggest that the HML factor is redundant. All of the other parameters are significant and have explanatory power. The β’s are lower than one, meaning that our indexes are less volatile than the benchmark. The alphas are significant lower than zero in all of the indexes. The results show clear evidence that investing in socially responsible indexes, constructed on the Global markets, does cost investment returns when they are compared with

(14)

the benchmark.

WORLD ESG FOCUS -.0018341*** ( .0002781) .989115*** (.0059737) -.1820529*** (.019646) .0065364 (.0160654) 0.9959 0.00301 WORLD ESG LEADERS -.0019253***

(.0003479) .9785697*** (.0074733) -.1570571*** (.0245778) .0226259 (.0200984) 0.9935 0.00376 ACWI ESG LEADERS -.0017148***

(.0004027 ) .9855927*** (.0086511) -.0899762*** ( .0284513) -.0172115 (.0232659) 0.9914 0.00435 ACWI ESG UNIVERSAL -.0022982***

(.0003519) .9934417*** (.0091876) -.1833123*** (.0256783) -.0305676 (.0210532) 0.9920 0.00336 Table 3: Results from Global regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

EAFE SRI Standard -.0016309** (.0007173) .9943337*** (.0141254) -.1553246*** ( .0437071) -.0463327 (.0430296) 0.9787 0.00775 WORLD ex USA ESG LEADERS -.0021538***

(.0004511) .9785479 *** (.0088829) -.1690282*** (.0274858) .0003881 (.0270597) 0.9913 0.00487 WORLD ex USA SRI -.0017015***

(.0005652) .9881997*** ( .0111298) -.1133238*** (.0344381) -.0537801 (.0339043) 0.9864 0.00611 ACWI ex USA ESG LEADERS -.0015417**

(.0005871) .9926012*** (.0115618) -.0858182** (.0357747) -.0565897 (.0352201) 0.9854 0.00634

Table 4: Results from Global regression excluding United States (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.1.3 Results from United States markets

Table 5 provides an overview of the results from the three-factor regression on SRI indexes of the United States. The average adjusted R-squared is 0.9784, which means that 97,84% of the variation in the SRI index excess return is explained by the fac-tors excess market return, SMB, and HML. The results of the three-factor model regressions show that most of the parameters are statistically significant.The β’s are significant for the five indexes. The s parameter is statically significant for four in-dexes. For only one of the five indexes, the h parameter is statistically significant. Given the fact that four out of the five indexes have a significant negative alpha, the results provides the information that socially responsible investing may cost invest-ment returns for US indexes. Only the KLD 400 Social index has an alpha that is not significant lower than zero.

(15)

Index αi βi si hi Adj Rˆ2 SER KLD 400 SOCIAL -.0006904 (.000718) .9566995*** (.0173304) -.1224432*** (.035068) .0175165 (.0303216) 0.9682 0.00767

USA ESG FOCUS -.0010056***

(.0003212) .9885381*** (.0097828) -.1624967*** (.0159993) -.0502038*** ( .0135112) 0.9926 0.00273

USA ESG LEADERS -.0012891**

( .0006081) .9709716*** (.0146773) -.0757237** (.0296993) .0103906 (.0256796) 0.9780 0.0065

USA ESG SELECT -.0011722*

(.0006692) .9744356*** (.0161534) -.1058369*** (.0326862) -.0366998 ( .0282622) 0.9731 0.00715

USABROAD ESG LEADERS -.0010871*

(.0005931) .977638*** (.0143145) .0340884 (.0289652) .0396129 ( .0250448) 0.9803 0.00634

Table 5: Results from United States regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.1.4 Results from Asia Pacific markets

Table 6 provides an overview of the results from the three-factor regression on Asia Pacific indexes. The average adjusted R-squared is 0.9168, which means that 91.68% of the variation in the SRI index excess return is explained by the three factors. The results of the five-factor model regressions show that almost all of the parameters are statistically significant. In contrast to the other markets regressions, the alphas are not significant lower than zero.

PACIFIC ex JAPAN SRI -.0027107 (.0019015) .969492*** (.0303102) -.3952009*** (.0704763) -.381153*** (.0804661) 0.9019 0.02033 PACIFIC ex JAPAN ESG LEADERS -.0023509

(.0018443) .9564576 *** (.0293985) -.3472959*** (.0683565) -.3736159*** (.0780458) 0.9050 0.01972 ASIA PACIFIC ex JAPAN ESG LEADERS -.0003332

(.0012845) .8989169*** (.0204752) -.141468*** (.0476082) -.0299057 (.0543565) 0.9435 0.01374

Table 6: Results from Asia Pacific regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.2 Results from the five-factor model

5.2.1 Results from European markets

Table 7 provides an overview of the results from the five-factor regression on the six European SRI indexes. The average adjusted R-squared is 0.9707, which means that 97,07% of the variation in the SRI index excess return is explained by the factors excess market return, SMB, HML, CMA and RMW. Interesting finding is that the adjusted R-squared of the extended model is a little lower than that of the three-factor model. However, a lot of the variation is explained by the model so the five-factor model is still a good fit for these regressions. The results of the five-factor model

(16)

regressions show that some parameters of the independent variables are statistically significant. As well as the β’s, the s parameter is statically significant for all European indexes. For three of the six indexes, the h parameter is statistically significant. The r and c parameters are both not statistically for any of the indexes, so this provides the information that investments and profitability has no explanatory power for any of the European indexes. Given the fact that four out of the six indexes have a significant negative alpha with both the three-factor and five-factor regression, we can conclude that socially responsible investing may cost returns in the European markets .

Index αi βi si hi ri ci Adj Rˆ2 SER

EU Standard SRI -0.0016397* (0.0008351) 1.023712*** (-0.0178141) -.1187629*** (.0438994) -.1977048*** (.0588546) -.0502528 (.0806158) .0595345 (.0649089) 0.9812 0.00797 EU ESG Leaders -.0025276*** (.0006514) .9806996*** (.0138962) -.1553269*** (.0342444) -.0747346 (.0459104) -.0037444 (.0628856) -.0022747 (.0506332) 0.9881 0.00622 EMU ESG Leaders -.0004377

(.0017015) 1.061931*** (.036273) -.2467743*** (.0905727) .1475711 (.1214264) -.1348167 (.1642539) .0626336 (.1340608) 0.9409 0.01646 EMU ESG Universal -0.0025298*

(0.0013314) 1.084724*** (.0313079) -.3910782*** (.0759802) .3047696*** (.0995984) .0928635 (.1289972) -.0214559 (.1276607) 0.9603 0.01164 EU ESG Universal -.0021159*** (.0005069) .9945788*** (.0119196) -.2138639*** (.0289272) -.076907** (.0379192) -.0115462 (.0491119) .0103166 (.0486031) 0.9916 0.00443 EMU Standard SRI -.0021173

(.0013672) 1.111622*** (.0291659) -.2272112*** (.0718736) .0952148 (.0963588) .0189649 (.131987) .1537952 (.1062711) 0.9623 0.01305

Table 7: Results from European regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.2.2 Results from Global markets

Tables 8 and 9 provide an overview of the results from the regression on world SRI indexes with and without the United States. The total average of the adjusted R-squared is 0.9902, which means that 99,02% of the variation in the Global SRI indexes excess returns is explained by the factors of our regression model. This number is higher than that of the three factor model, so the five-factor model is better in explaining the variation for the global markets regression. Interesting finding is that the r parameter is in all of the eight cases statistically significant, so profitability has a high explanatory power. This is in contrast to the European regression where none of the r parameters was significant. Furthermore, all the β’s are significant. The s parameter is significant for six of the eight indexes, the h parameter is only significant for three indexes. In the regression for the global SRI indexes, the c parameter is significant for two out of the eight indexes, so investments have some explanatory power in global regressions. The alphas are significant lower than zero in all of the indexes, same as with the three-factor regressions.

(17)

WORLD ESG FOCUS -.0021311*** (.0003098) .9992708*** (.007653) -.1585682*** (.021768) .0314498 (.0209909) .0768047** (.0316528) .040671 (.025734) 0.996 0.00299 WORLD ESG LEADERS -.0025717***

(.0003711) .9995488*** (.0091667) -.1085268*** (.0260735) .0464384* (.0251428) .141377*** (.0379135) .0811867*** (.030824) 0.9941 0.00358 ACWI ESG LEADERS -.0025255***

(.0004163) 1.005079*** (.0102827) -.0294714 (.029248) .037415 (.0282039) .2090662*** (.0425295) .0401828 (.0345768) 0.9927 0.00402 ACWI ESG UNIVERSAL -.0028129***

(.0003436) 1.011091*** (.0093747) -.1311042*** (.0258983) .059444** (.0260056) .1939234*** (.0372839) -.0436628 (.0379906) 0.9937 0.00302

Table 8: Results from Global regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

EAFE SRI Standard -.0028662*** (.0007856) 1.035894*** (.0182699) -0.089254* (0.0459479) -.025948 (.0525941) .2111084*** (.0794327) .1691164*** (.059762) 0.9804 0.00744 WRLD ex US ESG LEAD. -.0027384*** (.0005036) .9877422*** (.0117112) -.1389163*** (.0294532) .0668029** (.0337135) .1593286*** (.0509175) .0070495 (.0383083) 0.9916 0.00477 WRLD ex US SRI -.0025132*** (.0006316) 1.008495*** (.0146882) -0.071149* (0.03694) -.0086072 (.0422833) .1724613*** (.0638603) .0612494 (.048046) 0.9869 0.00598 ACWI ex US ESG LEAD. -.0020669***

(.0006592) .9957023*** (.0153298) -.0599195 (.0385537) .0176371 (.0441304) .1591266** (.06665) -.0318341 (.0501449) 0.9859 0.00624

Table 9: Results from Global ex USA regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.2.3 Results from the United States markets

Table 11 provides an overview of the results from the regression on SRI indexes of the United States. The average adjusted R-squared is 0.9794, which means that 97,97% of the variation in the SRI index excess return is explained by the factors excess market return, SMB, HML, CMA and RMW. Just as with the global regressions, the adjusted R-squared is higher for the five-factor regression than for the three-factor regression. The results of the five-factor model regressions show that some parameters of the independent variables are statistically significant. As well as the β’s, the s parameter is statically significant for all indexes. For two of the five indexes, the h parameter is statistically significant. The r parameter is not statistically for any of the indexes, so this provides us with the information that investments has no explanatory power for any of the US indexes. Given the fact that in the three-factor and five-factor regressions, four out of the five indexes have a significant negative alpha, this research concludes that also in the US markets, socially responsible investing may cost returns. Agian, only the KLD 400 Social index has an alpha that is not significant lower than zero.

(18)

Index αi βi si hi ri ci Adj Rˆ2 SER KLD 400 SOCIAL -.0011878 (.0007205) .980173*** (.0182995) -.1130384*** (.0355601) -.0590149 (.0406623) .0296931 (.0493523) .206277*** (.0570321) 0.9705 0.00739 USA ESG FOCUS -.0010308***

(.000323) .9901383*** (.0100311) -.1567122*** (.018436) -.0482952** (.020642) .0275882 (.0257939) .0474941 (.0300063) 0.9926 0.00272 USA ESG LEADERS -.0017492***

(.0006197) .9881877*** (.0157373) -0.0558583* (0.0305811) -.0298905 (.0349689) .0682269 (.0424422) .1289592*** (.0490466) 0.9789 0.00636 USA ESG SELECT -.0014273**

(.0006957) .9866057*** (.0176675) -.0979425*** (.0343319) -0.0728171* (0.0392579) .0179886 (.0476478) .1118324** (.0550623) 0.9732 0.00714 USABROAD ESG LEADERS -.0015592***

(.0005948) .9952084*** (.0151053) .050536* (.0293529) -.0256105 (.0335645) .0458663 (.0407377) .1412302*** (.0470769) 0.9818 0.0061

Table 10: Results from United States regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.2.4 Results from the Asia Pacific markets

Table 11 provides an overview of the results from the regression on Asia Pacific indexes. The average adjusted R-squared is 0.9389, which means that 93,89% of the variation in the SRI index excess return is explained by the five factors. This is a lower number than in any of the other markets but again, higher than with the three-factor regression. The results of the five-factor model regressions show that almost all of the parameters are statistically significant. As well as the β’s, the r parameter is statically significant for all indexes. All the other parameters are statistically significant for two of the three indexes. In contrast to the three-factor model where none of the indexes had an alpha significant lower than zero, two out of the three alphas are lower than zero.

PAC. ex JAP. SRI -.0053088*** (.0016403) 1.055651*** (.0314104) -.175529** (.0696152) -.1960997** (.0842124) .4984054*** (.0912036) .2986076*** (.0952589) 0.9314 0.017 PAC. ex JAP. ESG LEAD. -.005087***

(.0015356) 1.045011*** (.0294047) -0.1090112* (0.06517) -.1800994** (.0788351) .5316878*** (.0853799) .3010468*** (.0891763) 0.9381 0.01591 AS.PAC. ex JAP. ESG LEAD. -.0007445

(.0012822) .8954878*** (.0245536) -.0736869 (.0544183) .1075988 (.065829) .2299213*** (.071294) -.0757862 (.0744641) 0.9471 0.01329

Table 11: Results from Asia Pacific regression (note: *,** and *** indicate 10%, 5% and 1% significance levels)

5.3 Limitations of the research

For this study there were some restrictions. The most important restriction is the data used for the research. Unfortunately, the SRI index sample was restricted to only MSCI indexes because of the data available. Furthermore, some indexes could not be used because of the time frame of ten years. In 3 cases the study has deviated from the 10 years period but only by a small margin. Therefore, the indexes where only five years of data was available haven’t been used. Overall, the 22 indexes used provides a fair estimation about the performance of SRI indexes.

(19)

6 Conclusion

Socially responsible investments (SRI) are no longer a niche segment of the inter-national equity markets. For investors who want their investments used for socially responsible objectives, it is interesting to know whether the stocks selected by the ESG criteria perform different than conventional stock investments. Most of the re-cent studies were focused on the performance of SRI funds instead of the more passive indexes. The problem with measuring fund performance is that the research has to deal with some uncontrollable variables such as managerial decisions, transaction costs or fund fees. Analyzing the SRI indexes provides a clearer view in the performance of the stocks selected by the ESG criteria. Morgan Stanley Capital International (MSCI) their ESG screening criteria is also used by many of the SRI fund managers. This means that, besides the investors ability to acquire an Exchange Traded Fund on the SRI indexes, this study is also relevant for the potential performance of any SRI mutual fund using the same selection criteria.

The regression results of this study can conclude that socially responsible investing in a SRI or ESG indexes may cost some investment return. The results of the three-factor regression show that 16 of the 22 indexes have an alpha significant lower than zero. The results of the five-factor regression show that in 18 of the 22 indexes studied, the alpha is significant lower than zero. None of the SRI or ESG indexes beat the conventional benchmark. Since the average adjusted R-squared of the five-factor regressions is slightly higher, the results of the five-five-factor model seem to be more precisely. This study indicates that Socially Responsible Investing Index performance is slightly lower than the benchmark during the ten-year period analyzed. This is in line with the optimal portfolio theory, since the SRI screening process reduces the investment universe which lead to a reduction in the risk-adjusted return.

Furthermore, the data results show that the betas are statistically significant in all the regressions. The only remarkable difference is that the SRI beta is lower than one in 18 of the 22 regressions when using the three-factor model. When using the five-factor model, 11 out of the 22 indexes have a beta lower than one. This would indicate that the SRI indexes have a lower volatility than the market. However, the Sharpe ratio is higher for the benchmark than for the SRI indexes in almost all cases. This means that the market has a more attractive risk-adjusted return than the SRI indexes.

Thus, SRI stocks or indexes are the most suitable for risk averse investors with specific ethical preferences. There will always be an ethical dilemma if it is better to invest in companies which have a high rating in environmental, social and governance criteria or in companies which may have a better performance in terms of risk-adjusted returns. Since the deviations of the alpha are small and as the research in previous studies show, SRI investments tend to move as the market and are a predictable

(20)

investment choice. However, this study its evidence from indexes suggest that socially responsible investing is sub-optimal in the analyzed timeframe. A suggestion for further research: compare more indexes for multiple timeframes.

(21)

7 Bibliography

Angel, J. and Rivoli, P. (1997). Does Ethical Investing Impose a Cost Upon the Firm? A Theoretical Examination. The Journal of Investing, 6 (4) 57-61

Aupperle, K., Carroll, A. and Hatfield, J. (1985). An Empirical Examination of the Relationship between Corporate Social Responsibility and Profitability. The Academy of Management Journal, 28(2), 446-463.

Bauer, B, Koedijk, K and Otten, R. (2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking & finance, 29 (7), 1751-1767.

Bauer, R., Otten, R. and Tourani Rad, A. (2006). Ethical investing in Australia: Is there a financial penalty? Pacific-Basin Finance Journal, 14 (1), 33–48.

Bauer, R., Derwall, J. and Otten, R. (2007). The ethical mutual funds perfor-mance debate: New evidence for Canada. Journal of Business Ethics, 70 (2), 111–124.

Bollen, N. (2007). Mutual fund attributes and investor behavior. Journal of Financial and Quantitative Analysis, 42(3), 683-708

Carhart, M.M. (1997). On Persistence in Mutual Fund Performance. The Journal of Finance, 52 (1), 57-82.

Fama, E. F., & French, K. R. (1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33(1), 3-56.

Fama, E., & French, K. (1995). Size and book-to-market factors in earnings and returns. Journal of Finance 50, 131-156.

Fama, E. F., & French, K. R. (1996). Multifactor Explanations of Asset Pricing Anomalies. Journal of Finance 51, 55-84.

Fama, E. F., & French, K. R. (2004). The capital asset pricing model: Theory and evidence. Journal of Economic Perspectives, 18(3), 25-46.

Fama, E. F., & French, K. R. (2006). The value premium and the CAPM. Journal of Finance, 61(5), 2163-2185.

(22)

Fama, E. F., & French, K. R. (2012). Size, value, and momentum in international stock returns. Journal of Financial Economics, 105(3), 457-472.

Fama, E.F., & French, K. (2015). A five-factor asset pricing model. Journal of Financial Economics, 116, 1-22.

Friedman, M. (1970). The Social Responsibility of Business is to Increase its Profits. Corporate Ethics and Corporate Governance, 173-178.

Galema, R., Plantinga, A. and Scholtens, B. (2008). The stocks at stake: Return and risk in socially responsible investment. Journal of Banking & Finance, 32 (12), 2646–2654.

Hamilton, S., Joe, H., and Statman, M., (1993). Doing well while doing good? The investment performance of socially responsible mutual funds. Financial Analysts Journal, 49 (6), 62–66.

Kempf, A. and Osthoff, P. (2007). The effect of socially responsible investing on portfolio performance. European Financial Management,13 ,908-922

Merton (1987). A simple model of capital market equilibrium with incomplete information. Journal of Finance, 42(3), 483-510

Mossin, J. (1966). Equilibrium in a capital asset market. Econometrica, 34 (4), 768-783.

Rennebooga, L., Horst, J. ter, and Zhang, C. (2008). Socially responsible invest-ments: Institutional aspects, performance, and investor behavior. Journal of Banking & Finance, 32 (9), 1723–1742.

Rennebooga, L., Horst, J. ter, and Zhang, C. (2011). Is ethical money financially smart? Nonfinancial attributes and money flows of socially responsible investment funds. Journal of Financial Intermediation, 20 (4),562–588.

Sauer, D. (1997). The impact of social-responsibility screens on investment per-formance: Evidence from the Domini 400 Social Index and Domini Equity Mutual Fund. Journal of Financial Economics, 6(2), 137-149

Sharpe, W.F. (1964). Capital Asset Prices: A theory of market equilibrium under conditions of risk. Journal of Finance, 19 (3), 425-442.

(23)

Statman, M. (2000). Socially responsible mutual funds. Financial Analysts Jour-nal , 56 (3), 30–39.

Performance of Socially Responsible Investing : evidence from indexes