Performance and investment style
of ethical mutual funds:
Evidence from eight European countries
Klaas-Jan Boessenkool*Rijksuniversiteit Groningen
Version: 6 August 2008
*University of Groningen, Faculty of Economics and Business Master’s thesis MScBa Finance
Author: K.J. Boessenkool, S1256637
Address: Allersmastraat 2a, 9716 AT, Groningen Phone: 06 30037833
Performance and investment style
of ethical mutual funds:
Evidence from eight European countries
Keywords: Socially Responsible Investments, ethical investing, mutual funds, investment, screens, style analysis, performance evaluation
JEL code: A13, G11, G12, G20, G23, M14
Abstract
1. Introduction
Over the past decade, Socially Responsible Investments (SRI) have experienced an enormous growth all over the world. In the US, one of the frontrunners in this movement, already 11% of all assets under control of professional management is invested in a socially responsible portfolio. (SIF, 2007). In Europe, the UK, Belgium and the Netherlands are the frontrunners in this upcoming movement. Despite the increase in investments, in several researches it is found that ethical mutual funds slightly underperform compared to the market and to conventional funds. The theoretical reasoning behind this finding is that ethical screening imposes constraints on the investment universe. This limits the possibilities of diversifying portfolios. In this research the performance and investment style of ethical mutual funds in eight European countries is analyzed in the period 2001-2008.
According to the Social Investment Forum (SIF), SRI can broadly be defined as “an investment process that considers the social and environmental consequences of investments, both positive and negative, within the context of rigorous financial analysis.”1One of the possibilities of SRI is investing in ethical mutual funds. Ethical mutual funds apply a set of social, environmental of ethical investment screens in order to select stocks for their portfolio. Ethical mutual funds can be bond funds, equity funds or balanced funds, the latter consisting of equities as well as bonds. This research exclusively focuses on equity funds.
In the last couple of decades, several researches have taken place that investigated risk and returns of ethical mutual funds. Most of these researches were done by researchers from the United States or the United Kingdom and exclusively focused their research on these two countries. The few researches that took into account other European mainly focused on the period before 2003. The typical models used in these researches are the Capital Asset Pricing Model (CAPM) and the Carhart (1997) four-factor model. The latter model estimates portfolio returns with a value-weighted market factor, but corrects for size, book-to-market value and momentum. The aim of this paper is to review previous research on the risk and returns of ethical mutual funds and extend it to different European countries and the most recent period. In order to find out whether the hypothesis of underperformance of ethical mutual funds also holds for other
markets and other timeframes, in total 116 ethical mutual funds from eight European countries are investigated in a more recent timeframe than ever done before. The ethical mutual funds under examination are based in Austria, Belgium, France, Germany, Italy, Switzerland, the Netherlands, and the United Kingdom. I will focus my research on the performance and investment style of equity funds from these countries from 2001 to 2008. In line with previous research, I will use the single factor CAPM model and the Carhart four-factor model to test my hypotheses.
The remainder of this paper will be organized as follows. In section two the theoretical backgrounds of socially responsible investments will be presented. Section three provides information on the data used in this research. In section four methodology and the model will be described. Section five presents the empirical results. In section six conclusions are drawn and recommendations for further research are given.
2. Theoretical Background
2.1 History of SRI
The origins of Socially Responsible Investing (SRI), or ethical investing, date back more than hundred years ago to several religious roots, like Judaism, Christianity and the Islam. In those days, the several religions set up a number of rules in order to prevent people from engaging in sinful behavior or taking interest in sinful institutions. However, SRI nowadays is not based on religious roots any more, though more on a growing social awareness of investors all around the world. Especially since the 1960’s, when several campaigns against war, racism and nuclear energy were carried, investors have become more concerned about the social and environmental consequences of their investments. Accordingly, the SRI industry experienced strong growth all over the world, especially in the United States and Europe. Ethical consumerism, stating that people were willing to pay a premium for (financial) products that are in line with their personal values, was an important factor behind this growth.2 Since the 1990’s this growth has been
explosive. In the US alone, 11% of all money under professional management is invested in a socially responsible portfolio.3
This research focuses on one part of SRI: ethical mutual equity funds. In order to construct an ethical mutual fund, investment screens are used. All of these funds apply several, but at least one positive or negative screen. A screen is used to filter the appropriate stocks out of the entire universe of stocks. A negative screen is typically applied on an initial assets pool, from which certain specific sectors (e.g. weapons, nuclear energy, tobacco, gambling, alcohol and pornography), are excluded. A positive screen is used to select companies that meet superior standards on issues like for example corporate governance, developing countries or environmental protection. This screen is often applied based on a best-in-class approach: firms within an industry are ranked based on certain criteria (e.g. ethical, social, environmental and corporate governance), and the leaders of this group are included in a portfolio.4
Social screens like environment protection, civil rights and labor relations have become very common. Since the year 2000, among others the Enron scandal and several other corporate scandals have stimulated the appearance of new screens based on governance, transparency and sustainability.
2.2 Theoretical Reasoning
In general there are three hypotheses about the performance of ethical mutual funds compared to non-ethical mutual funds.5 The first two hypotheses are on the risk-adjusted return (alpha) of ethical mutual funds, the last hypothesis is on risk exposure (beta) of ethical mutual funds.
The first hypothesis states that ethical mutual funds underperform compared to conventional mutual funds. According to this line of reasoning it is thought that making your investment decisions on the basis of ethical criteria worsens portfolio performance. The Efficient Market Hypothesis (EMH) posits that financial markets are ‘informationally’ efficient. Secondly, it states that portfolios returns are proportional to their risk and that an optimal portfolio is a well-diversified portfolio. Since ethical screening imposes constraints on the investment universe and
3 Social Investment Forum (2007) “Report on responsible investing trends in the US”, executive summary, p.2
4Renneboog, L., Horst, ter, J. and Zhang, C. (May 2007) “The price of ethics: evidence from socially Responsible Mutual Funds”, p. 7
consequently imposes limits on diversification, an ethical mutual fund will earn suboptimal returns. Besides this, the ethical screening process itself brings additional costs to the investors, which in turn worsens portfolio performance as well.
In contrast, the second hypothesis states that ethical mutual funds outperform conventional portfolios. According to this hypothesis, information on corporate governance and ethical, environmental and social performance may be underpriced by the stock markets. So the ethical screening processes generate value-relevant non-public information. Key assumption here is that portfolio managers of conventional funds do not use this value-relevant information. This is not in line with the EMH, stating that financial markets reflect all known information. There are two arguments that are in support of this hypothesis. First, sound environmental and social performance might be a signal of good managerial skill, which results in good financial performance. Secondly, with social and environmental screening one might prevent incurring high costs in social or environmental crises.
The third hypothesis states that ethical mutual funds have different exposure to risk and therefore different expected returns than their conventional peers. The use of screens for selecting stocks might have an impact on risk exposure of ethical mutual funds, in contrast to conventional funds, using no screens at all. For example, companies with sound corporate governance may have a lower book-to-market ratio and therefore lower risk exposure in a Carhart (1997) four-factor model.
2.3 Previous research
CAPM-Model
As far as socially responsible mutual funds are concerned, several researches have taken place in the past. Table 1 shortly summarizes the research methodologies of the most recent researches on socially responsible mutual funds.
funds in the US from 1990-1998. Using a matched pair analysis, he found that ethical mutual funds underperformed compared to both the Domini Social Index and the S&P500 index, but they performed slightly better, though not significantly, than their conventional peers. Concerning investment style, Luther, Matatko and Corner (1992), Luther and Matatko (1994) and Statman (2000) documented a small cap bias. This is confirmed by Gregory, Matatko and Luther (1997). More recently, Kreander et al. (2005) investigated the European market with a matched pair analysis for the period 1996-1998. Examining 30 funds from Belgium, Germany, the Netherlands, Norway, Sweden, Switzerland and the United Kingdom, no significant differences in return were found between ethical and non-ethical funds. Besides this, the management fee is found to be a significant explanatory variable for the Jensen measure.
Carhart (1997) four-factor model
In the last couple of years, the majority of the studies applied multi-factor models like for example the Fama French (1993) three-factor model and the Carhart (1997) four-factor model. One of the first and most important in this respect is the paper by Bauer et al. (2005). In this paper the performance and investment style of ethical mutual funds from Germany, the UK and the US are examined in the period 1990-2001. A CAPM model as well as a Carhart-model is applied on a database of 103 funds, using Worldscope and Domini Social Index as market estimators and a local one month treasury-bill rate. The return difference between ethical and non-ethical funds is found to be negative for Germany and the US, but positive for the UK. However, none of these results are significant. It is found that German and UK ethical funds generally are more exposed to small caps, while US funds are more exposed to large caps. Furthermore, compared to conventional funds, ethical funds are more growth-oriented than value-oriented. Introducing time-variation in betas, using a Ferson & Schadt (1996) model, makes US domestic ethical funds’ underperformance and UK’s ethical funds outperformance significant. Finally, they also document a learning effect. Ethical mutual funds trailed conventional funds significantly from 1990-1997, but in the most recent period 1998-2001 they are able to catch up. The results of investigating the Australian market support these findings (Bauer et al., 2004).
overpriced compared to their fundamental price, which results in underperformance by ethical mutual funds. In line with their hypothesis, they find evidence of strong underperformance of ethical funds compared to their conventional counterparts in European and Asia-Pacific countries. Two reasons are put forward why investors might want to pay this premium. First, investors are willing to pay for the risk reduction, which exists because of ethical screening. Secondly, the behavioral bias “aversion to unethical corporate behavior” seems to play a role, since the premium is much higher than the risk reduction that comes in return. SIF (2003)6 states that, according to this reasoning, social investors are less likely to move investments from one fund to another. They are said to be “stickier” to their funds than conventional investors, despite the lower fund performance.
Finally, the paper by Scholtens (2004) exclusively focuses on the Dutch market for SRI funds, a market also included in this research. All Dutch SRI funds are investigated separately using a CAPM model and a Carhart (1997) four-factor model in the period 2001-2003. For conventional funds a somewhat higher return is documented, but this result is not significant. In line with previous research, Dutch SRI funds are found to be biased towards small caps. However, in contrast to Bauer et al. (2005), ethical funds are found to be tilted towards value stocks instead of growth stocks. The reason for this may be that the period under consideration was one of economic recession. Therefore fund managers were likely to increase their preference for value stocks.
This research will try to extend previous research on this subject. In this research, ethical mutual funds from eight European countries are examined in a more recent time frame than ever done before.
In line with Bauer et al. (2005) and Renneboog et al. (2007), for every country a portfolio will be constructed comprising all the SRI funds in that country. In line with Bauer et al. (2005), Scholtens (2004) and Renneboog et al. (May 2007), first a one-factor CAPM model is tested for each country. However, in contrast to these researches, the timeframe is more recent, being 2001-2008. Besides this, as a market benchmark seven different market indices are tested, being two world indices, a European index, a national index and three ethical indices. To my knowledge this has never been done before.
After this, a Carhart (1997) four-factor model will be tested for the same eight countries in the same time period. For estimating the market, one world, one European, one national and two ethical indices are used. MSCI indices are used to construct the size factor (SMB) and the
to-market-value factor (HML). This is in line with Scholtens (2004), but in contrast to Bauer et al. (2005) and Renneboog et al. (May 2007), who both used Worldscope information. In line with Scholtens (2004), momentum is constructed on a three-monthly basis using all the stocks that were in the main national index during the period 2001-2008. This is in contrast to Bauer et al. (2005) and Renneboog et al. (May 2007), who both use all stocks in the Worldscope database. Since the market for SRI is still growing and markets are changing in Europe, it will be interesting to see if findings in this paper support findings of the researches that examine other countries and earlier timeframes.
3. Data
3.1 Ethical mutual funds
A database is constructed that contains 116 socially responsible mutual funds from eight countries in Western Europe. These countries are Austria, Belgium, France, Germany, Italy, Switzerland, the Netherlands and the United Kingdom. The research is restricted to equity mutual funds, excluding bond, fixed-income, money-market and balanced mutual funds.
To determine the amount of socially responsible mutual funds for each country, the Morningstar SRI fund selector, as developed by Eurosif and Avanzi, is used.7This data source was also used in Bauer et al. (2005), but Renneboog et al. (May 2007) uses the S&P fund selector. The search results of the Morningstar selector are dated at 1-1-2007. The selected funds are considered to be ethical mutual funds during the entire sample period, from February 2001 to January 2008. The selection process leads to a final sample of in total 116 socially responsible mutual equity funds domiciled in eight European Countries. Table 2 presents all the ethical mutual funds from each country included in this research. Table 3 presents some key statistics on the ethical mutual fund market in the included countries. The size of the ethical mutual fund market is in absolute terms definitely largest in France and the UK. These countries have the largest number of funds and highest amount of total asset value. Median fund size is highest in Switzerland, followed by the UK and Belgium. Austria’s SRI funds have the highest initial expenses, although the countries don’t differ much from each other.
For each fund, monthly adjusted prices in Euro’s are taken from Datastream. Monthly adjusted prices take into account capital changes, like dividend payments and stock splits. The sample period is from February 2001 to January 2008. An earlier period is not advisable, since the amount of funds is too small and the available data is not representative. However, this dataset might suffer from a possible survivorship bias. This might lead to an overestimation of average performance. Bauer et al. (2005) prevents this problem by adding back dead funds. However, with the Morningstar fund selector this problem is inevitable, since the SRI fund selector only selects funds that are in the market at 1-1-2007. It is impossible to get the search results from earlier periods and to find out which funds were considered ethical funds before 1-1-2007.
For every ethical mutual fund the monthly adjusted price series have been turned into return series using the following formula:
Rit= 1 1
it it itP
P
P
, (1)where Rit is the return on stock i at time t en Pit is the price of stock j at time t.
After that, equally weighted portfolios are formed for each country, consisting of all the funds with available data at that time. Equally weighted portfolios are chosen, since every ethical mutual fund is considered to be just as important and influential as the other. The average portfolio return is calculated using the following formula:
Rpt=
N i it it R X 1 * , (2)where Rpt is the return on the portfolio at time t, Xi is the fraction of the funds invested in the ith asset at time t (that is one divided by number of stocks in the portfolio at time t), and Rit is the return on assets i at time t.
This procedure is followed for each of the eight countries. This leads to return series of eight equally-weighted portfolios, one for each country, consisting of all the ethical mutual funds in each country.
Unfortunately, in this research it is not possible to use log returns (continuously compounded returns), since log returns of individual stocks are not additive across a portfolio. The fundamental reason for this is that the log of a sum is not the same as the sum of a log, since taking a logarithm constitutes a non-linear transformation.8 For calculating portfolio returns based on continuously compounded returns it is necessary to first estimate the value of the portfolio at each time period en then determine the return. However in this research, the value of the
portfolios cannot be calculated properly across time, since for each country the number of ethical mutual funds in the portfolio fluctuates during the entire sample period. Some ethical mutual funds start for example in 2000, others in 2004. So in this case it is inevitable that first the individual funds’ returns are estimated. Subsequently, for each country the portfolio’s return is calculated adding these returns and dividing them by the number of the individual funds.
Table 4 describes the data used in this research. Looking at mean returns, it is expected that in general ethical funds in underperform the national indices and the global indices. However differences are only marginal. Standard deviations of ethical funds seem to be lower than those of the main national indices, which might hint at lower risk. Based on the Jarque-Bera statistic, all series, except the S&P1200, are found to be not normally distributed. However, because of the large sample size, violation of the normality assumption is not expected to give problems to using the ordinary least squares method (OLS). So this method is used in this research.
3.2 Benchmarks
Market indices
According to previous research, several national and international indices can be useful as a benchmark for the market. In this research the following indices are used.
First, as benchmarks for the global market are taken the MSCI World Index (covering 23 countries with developed markets and approximately 85% of market capitalization in those countries9) and the S&P 1200 Global Index (covering 29 countries and approximately 70% of market capitalization in those countries10). The MSCI World Index was also used in Luther, Matatko and Corner (1992), Schroder (2004), Kreander et al. (2005) and Bauer et al. (2005). To my knowledge, the S&P 1200 Global Index was not used before in this context; however the S&P 500 was used as a benchmark of the US market in Statman (2000), Schroder (2004) and Bauer et al. (2005).
Secondly, as a benchmark for the European market, the S&P 350 is taken. This index covers 17 European countries and around 70% of the region’s market capitalization. To my knowledge this index has not been used in previous researches.
Thirdly, the three additional SRI indices under examination are the Dow Jones Sustainability Index World (DJSI), the FTSE4Good World Index and the FTSE4Good Europe index. The DJSI
9 http://www.mscibarra.com
consists of more than 300 companies that represent the top 10% of leading SRI companies out of the 2500 companies in de Dow Jones World Index (excluded are among others industries in alcohol, gambling, tobacco and weapons). The FTSE4Good World Index consist of the 100 companies with the highest market capitalization from the FTSE All-World Developed Index, after filtering the latter based on the FTSE4Good Inclusion Criteria (excluded are among others industries in tobacco, weapons and nuclear power). According to the same method, the FTSE4Good Europe Index consists of 50 companies from the FTSE All-World Developed Europe Index11. These indices have all been used before. Scholtens (2004) concludes that FTSE4Good and DJSI are more powerful in explaining fund performance than conventional indices in the Netherlands. However, Bauer et al. (2005) finds that DJSI is less powerful in explaining ethical mutual fund performance than conventional indices in Germany, the US and the UK.
Finally, for every country its main national index is examined as a market benchmark comprising the most significant, highest market value stocks and most actively traded stocks. These are the ATX, comprising the 20 most important stocks traded in Austria, for Belgium the Bel20, comprising of 20 stocks, for France the CAC40, comprising of 40 stocks, for Germany the DAX, comprising of 30 stocks, for Italy the MIB, comprising of 40 stocks, for the Netherlands the AEX, comprising of 25 stocks, for Switzerland the SMI, comprising of 20 stocks and for the UK the FTSE100, comprising of 100 stocks. To my knowledge only Scholtens (2004) used a national index, being the AEX25, for investigating the Dutch SRI funds. He documented it to be less powerful in explaining fund performance than global indices and SRI indices.
Additional risk factors
For constructing additional risk factors size (SMB) and book-to-market value (HML), country-specific MSCI series are used.
A country specific SMB factor is constructed on a monthly basis for all eight countries in this research. Since the ethical mutual funds are examined for every country separately, also country specific indices are selected instead of one general European index. Since it was not possible to acquire information of book value and market capitalization on all stocks in a specific country, country specific MSCI Large Cap and MSCI Small Cap Indices are used. The MSCI Large Cap Indices cover around 70% of the free float-adjusted market capitalization in each market, while MSCI Small Cap Indices use all the companies with a full market capitalization in the range of
200-1500 million USD. Accordingly, total return indices are used, since they measure total market performance, including capital changes.
In order to construct HML for every country on a monthly basis, two estimators of book-to-market value are used. First, high book-to-book-to-market stocks are value stocks. As estimators, MSCI Value Indices are used, that measure the top 50% market capitalization of standard indices by ascending price to book value. Secondly, low book-to-market stocks are growth stocks. As estimators, MSCI Growth Indices are used, that measure the top 50% Market Capitalization of standard indices by descending price to book value. For HML, for the same reasons as SMB, total return indices are used.
In order to construct the momentum factor, all stocks are used that were in the national index of each of the eight countries during the period 2001-2008. Therefore I looked up all historical compositions of the national indices of each of the eight countries in the period 2001:1-2008:1. This comes down to for Austria (ATX20), Belgium (Bel20), France (CAC40), Germany (DAX30), Italy (MIB40), the Netherlands (AEX25), Switzerland (SMI20) and the UK (FTSE100). For each of the stocks that was in the national index during this period, monthly adjusted price are taken from Datastream. This measure corrects for capital changes, like for example dividend payments or stock splits.
The methods for constructing SMB, HML and momentum are in line with Scholtens (2004), but in contrast with Bauer (2005 and Renneboog et al. (2007), who use al the stocks in the Worldscope universe.
Risk-free rate
For the risk-free rate, the Euribor one month rate is used. Euribor (Euro Interbank Offered Rate) is the benchmark rate of the European money market since fifteen countries of the European Union officially adopted the Euro in 1999. Specifically it is the rate at which euro interbank deposits are offered by one primary bank to another.12Previous researches, like Bauer et al. (2005), Scholtens (2004) and Renneboog et al. (May 2007) use one month local Treasury bill rates. However, the European financial market has integrated a lot since 1999. Since many European countries are included in this research, and the research starts in 2001, for this research the European interbank one month rate is the most appropriate.
4. Methodology
4.1 Hypotheses
Based on theoretical foundations and previous empirical findings the following hypotheses are stated concerning the market, the SMB factor and the HML factor.
Concerning the market, European ethical funds are expected to underperform the market. Underperformance to the market, though not significant, was found by Statman (2000) for the US ethical funds. Scholtens (2004) found this for the Netherlands, Bauer et al. (2006) for the Australian Market, and Bauer et al. (2005) also found this result for Germany and the US. Renneboog et al. (May 2007) concluded insignificant market underperformance for Belgium, Germany, Italy, and Switzerland and significant market underperformance in France and the Netherlands. These countries are also included in this research. This leads to the following hypothesis:
H1: European ethical mutual funds don’t perform significantly different than the market
H1a: European ethical mutual funds perform significantly worse than the market In case of the size effect, it is expected that ethical mutual funds in all European countries are tilted towards firms with low market capitalization. For example, Bauer et al. (2005) found significant small cap bias for the German and UK market. Scholtens (2004) documented an insignificant small cap bias for Dutch SRI funds. Renneboog et al. (May 2007) found a significant small cap bias for Belgium, France, Germany, Netherlands and Switzerland and insignificant for Italy. These countries are included in this research too. The size effect is estimated by the SMB factor, which is constructed as the return on small stocks minus the return on large stocks. Since ethical mutual funds are expected to be tilted towards small firms, the estimator of SMB is expected to be positive. Formally:
H2: European ethical mutual funds are not significantly tilted towards firms with high or low market capitalization
Concerning the style effect, I expect that in all the European countries the ethical mutual funds are tilted towards growth stocks. Stocks with a low book-to-market ratio are said to be growth stocks. In contrast, stocks with a high book-to-market ratio are said to be value stocks. Bauer et al. (2005) found a significant growth stock bias for Germany and the UK. Renneboog et al. (2007) found a significant growth stock bias for Italy, and an insignificant growth stock bias for Switzerland. In contrast however, Scholtens (2004) documents an insignificant value stock bias for the Dutch SRI market. Renneboog et al. (2007) also documents insignificant value stock biases for Belgium, France, Germany and the Netherlands. The style effect is estimated by the HML-factor, which is constructed as the return on value stocks minus the return on growth stocks. Since ethical mutual funds are expected to be tilted towards growth stocks, the estimator of HML is expected to be negative. Formally:
H3: European ethical mutual funds are not significantly tilted towards growth or value stocks
H3a: European ethical mutual funds are significantly tilted towards growth stocks
4.2 Model
CAPM and Carhart four-factor Model
In this research an approach is used that is very well-known in the recent literature. Among others, it is also used in Bauer et al. (2005), Scholtens (2004), Derwall et al. (2005) and Renneboog (2007).
First a single-factor CAPM model is applied. This model is based on the Jensen’s alpha and shows whether there are differences in the portfolio’s exposure to market risk. Formally:
Rit– Rft= αi + βi*(Rmt-Rft), (3)
However, in recent researches, like Bauer et al. (2005), Scholtens (2004) and Renneboog et al. (May 2007) this one-factor model is proven not to be the most efficient model in explaining portfolio returns. In order to confirm or, better, improve test results, an extension of this model is introduced, the Carhart 4-factor model. This model, introduced by Carhart (1997), also takes into account changes in style, size and momentum. This model is based on the Fama-French (1993) three-factor model, but includes also the momentum factor, according to Jegadeesh and Titman (1993). Besides using a market proxy, three additional control variables will be used to explain fund behavior; a capitalization-based factor, a book-to-market value factor and a momentum factor. Formally:
Rit– Rft = αi + β0i*(Rmt-Rft) + β1i*SMBt + β2i*HMLt + β3iMomt + εit, (4) where Rit is the return on a portfolio i at time t. Rft is the risk-free rate, measured by a one month interbank t-bill rate and Rmt is the relevant equity benchmark at time t. SMBt is the difference in return between a small cap portfolio and a large cap portfolio at time t. HMLt is the difference in return between a portfolio of high book-to-market stocks and one of low book-to-market stocks at time t. Momt is the difference in return between a portfolio of past three months winners and a portfolio of past three months losers at time t.
The additional risk factors
All three risk factors are constructed on a monthly basis for each of the eight countries in this research, since the ethical mutual funds are also analyzed on a national level. This way it is assumed to better take into account country specific aspects of the financial market. The additional risk factors SMB, HML and momentum are constructed in the following way.
Small-Minus-Big (SMB) is the difference in return between a small cap portfolio and a large cap portfolio. The estimator for the return on a small cap portfolio is the return on the MSCI Small Cap Index, the estimator for the return on a large cap portfolio is the return on the MSCI Large Cap Index. For each country the difference in return between a small cap portfolio and a large cap portfolio is calculated every month between February 2001 and January 2008 according to the formula: SMBt = MSCI Small Cap Indext– MSCI Large Cap Indext.
consisting of high book-to-market stocks and low book-to-market stocks is calculated every month between 2001 and 2008 according to the formula: HMLt= MSCI Value Indext– MSCI Growth Indext.
In order to construct the momentum factor, for each of the stocks in the national index the return over the last three months is calculated with the formula:
Rjt = 3 3 jt jt jt P P P , (5)
where Rjt is the return on stock j at time t en Pjt is the prices of stock j at time t.
Then, for every month the stocks are sorted from highest previous three-month return to lowest previous three-month return. Then two equally-weighted portfolios are created, one portfolio consists of the 30% highest return-stocks, the other consists of the 30% lowest return-stocks. Then for each portfolio the average equally-weighted return is calculated, using the formula: Rpt =
N i it it R X 1 * , (6)where Rpt is the return on the portfolio at time t, Xit is the fraction of the funds invested in the ith asset at time t (that is one divided by number of stocks in the portfolio at time t), and Rit is the return on assets i at time t. Momentum is then calculated by subtracting the average return of the portfolio of the lowest return-stocks from the average return of the portfolio consisting of the highest return-stocks. This procedure is repeated every month to keep this factor rolling.
5. Empirical Results
5.1 CAPM model
In table 5 the test results of the single-factor CAPM model are presented. From this table, several interesting conclusions can be drawn.
First, in general it is reported that the Dow Jones Social World Index (DJSI), the FTSE4Good Europe, the FTSE4Good World and the S&P350 give the best fit of the model, with adjusted r-squared results between 0,5274 and 0,7371. In all cases the S&P1200, MSCI World and the national index give the worst fit. In all cases, except the UK, the FTSE4Good indices give the best fit of the model. So overall, we can conclude that in these cases the ethical indices are more powerful in explaining fund performance than the standard non-ethical indices. However, the difference with the S&P350 is marginal.
Secondly, all models are tested for autocorrelation by means of a Durbin Watson statistic. All estimated values are between 1,7488 and 2,7406. The critical values are, with 84 observations and 1 explanatory variable (excluding the constant term) DL= 1,4750 and DU=1,5349. This gives boundaries between 1,4750 and 2,4651. Only in a few models, mostly FTSE4Good, DJSI and S&P350, the Durbin Watson statistic exceeds the upper limit by a small amount. This means there is very small evidence of negative autocorrelation in the residuals. But in most of the models, the estimators are expected to be best linear unbiased estimates (BLUE). In other words, the estimated values are expected to be efficient and true.
Hypothesis 1: Market underperformance
The estimators for beta show for all countries values between 0,0000 and 1,0000. This means that all ethical mutual fund portfolios are expected to move in the same direction as the market, but they are less risky than the market. All these estimates of beta are found to be significant on a 1% level.
So hypothesis one is rejected for Italy, it is concluded that ethical mutual funds perform significantly worse than the market. For all other countries hypothesis one is not rejected, it cannot be concluded that ethical mutual funds perform significantly different than the market. So based on the results for the CAPM model, we can conclude that ethical indices are more useful in explaining ethical mutual fund behavior than the standard indices. This is in line with research by Scholtens (2004), but contradicts the research of Bauer et al. (2005). Except for Switzerland, portfolio managers are not able to outperform the market indices and the ethical indices. So the investment style of ethical mutual funds, by restricting the investment universe by using ethical screens, indeed seems to come at a cost. However, except for Italy, the ethical mutual funds do not perform significantly worse than the market. The insignificant underperformance in seven of the eight countries is in line with previous researches by Statman (2000), Scholtens (2004), Bauer et al. (2005) and Renneboog et al. (2007). In order to find out more about the characteristics of the investment style of ethical mutual funds, the Carhart (1997) four-factor model is applied.
5.2 Multi-factor model
Table 6 summarizes the results of applying the Carhart (1997) four-factor model. Here the S&P1200 is omitted, since in the CAPM model, it continuously underperformed the other indices. Also the FTSE4Good Global is omitted, because the series gave exactly the same results as the FTSE4Good Europe.
Except for the UK, it is found that all adjusted R-squared values have increased compared to the CAPM model. If in both models the S&P1200 and the FTSE4Good Global are omitted, the average R-squared slightly increased from 0,6048 to 0,6428. This is an indication that the multi-factor model explains ethical mutual fund returns better than the single-multi-factor model. For all countries, an ethical index gives a better fit than the standard indices S&P350 and MSCI. However, for some countries this is the DJSI, and for some countries it’s the FTSE4Good Europe and the difference with the S&P350 is really marginal. The national index gives the poorest fit for all countries but the UK, where it gives the best fit.
of negative autocorrelation. However, this is such a small amount that this is not expected to influence the results significantly.
Hypothesis 1: Market underperformance
The test results for the estimator of beta is not different from these of the CAPM model. Except for Austria’s ATX20, all values are significant on a 1% level. In general, average beta is a little lower in the multi-factor model, compared to the single-factor model. This is due to the inclusion of three additional risk factors.
If one takes a closer look at the estimator excess market return, alpha, a strange phenomenon has appeared. The estimator of alpha is, besides Switzerland, also found to be positive for France and Germany. So in contrast to CAPM, not only Swiss ethical mutual funds are outperforming the standard and the ethical market indices, also French and German ethical mutual funds do. However, none of these results are significant on a 5% level. Austria, Belgium, Italy, the Netherlands and the UK show negative values for alpha. The negative estimators of alpha are, except for Austria’s ATX20, in none of the cases significant. So for all countries hypothesis one is not rejected, it cannot be concluded that European ethical funds perform significantly different than the market.
The insignificant outperformance by ethical mutual funds in France confirms the results by Renneboog et al. (May 2007). The insignificant outperformance in Switzerland and Germany contradicts Bauer et al. (2005), who found insignificant underperformance in Germany and Renneboog et al. (May 2007), who found insignificant underperformance in Switzerland and Germany. The insignificant results of underperformance in the Netherlands confirms results by Scholtens (2004) and Renneboog et al. (2007), who even found a significant underperformance of Dutch ethical funds. The insignificant underperformance in Belgium and Italy confirms results by Renneboog et al. (May 2007)
So in short, results are mixed. Ethical mutual funds in France, Germany and Switzerland seem to outperform the market, while ethical mutual funds in Austria, Belgium, Italy, the Netherlands and the UK seem to underperform the market. However, none of the results is significant, so we cannot conclude that ethical mutual perform significantly different than the market.
costs in social or environmental crises. Key assumption is that conventional fund managers do not use this value relevant information. However, this line of reasoning is not in line with EMH, stating that financial markets reflect all value-relevant information.
The insignificant underperformance in five out of eight countries might suggest that the companies, meeting the ethical standards to be included ethical mutual funds, are overpriced in the stock market, and therefore earn a lower return. There are two potential explanations for this. First, ethical companies may be less risky than conventional firms, and therefore earn a lower return. They might be less risky, because they may have less risk to be engaged in corporate governance scandals or environmental disasters. A second reason for the potential overpricing of ethical companies might be the investors’ aversion to unethical corporate behavior. Because of this, investors may be content with lower returns on firms that comply to the rules of ethical and environmental behavior. The rising demand for shares of these ethical companies might lead to overpricing of these shares and underperformance to the market. However, this explanation is not in line with EMH, stating that financial markets are efficient.13
Hypothesis 2: Small Cap bias
Considering size factor SMB factor the following is documented. For Austria the estimator SMB is insignificantly positive, signaling an insignificant growth stock bias. For Belgium, France, Germany, Italy, the Netherlands, Switzerland and the UK, the estimator SMB is found to be negative, signaling a large cap bias. For Belgium, France, Germany and Italy this result is significant on a 10% significance level. For the Netherlands, Switzerland and the UK this is not the case.
So for the Belgium, France, Germany and Italy hypothesis two is rejected. This means ethical mutual funds from these countries are significantly tilted towards large or small cap stocks. However, opposed to what was intentionally thought, ethical mutual funds in these countries are significantly tilted towards large cap stocks. In the Netherlands, Switzerland and the UK, ethical hypothesis two cannot be rejected on a 10% significance level, meaning ethical mutual funds are not significantly tilted towards large or small stocks. However, an insignificant large stocks bias is documented. Also for Austria hypothesis two cannot be rejected. However for Austria an insignificant small cap bias is found.
So in Austria, ethical mutual funds are more exposed to small caps, while in the other seven countries ethical mutual funds are more exposed to large caps. This result is even significant for
Belgium, France, Germany and Italy. This contradicts the work by Scholtens (2004), who documented an insignificant small cap bias for the Netherlands. It also contradicts results by Bauer et al. (2005), who documented a significant small cap bias for ethical mutual funds from Germany and the UK. It also contradicts the results by Renneboog et al. (May 2007), who documented an insignificant small cap bias for Italy, and a significant small cap bias for Belgium, France, Germany, the Netherlands and the UK. The test results only confirm a finding by Bauer et al. (2005), who concluded that, compared to conventional mutual funds, US ethical mutual funds are more tilted towards large cap stocks. So in conclusion, this is a surprising result. To my knowledge, in European countries a large cap bias has never been found before. A reason for this may be that small ethical firms which performed good in the last decade, matured and developed into large firms in the last couple of years. Another explanation might be that the underperformance of ethical mutual funds in early stages, as found in researches having earlier periods under consideration, leaded to a growing focus on less risky large capitalized firms in the period under consideration here.
Hypothesis 3: Growth stock bias
Examining book-to-market value estimator HML, the results are also quite surprising. For Austria, Switzerland, Netherlands, Belgium, France and the UK the estimator is positive, signaling a value stock bias. For Austria, Switzerland and the Netherlands, this result is significant on a 5% level. For Belgium, France and the UK these results are not significant. For the other two countries, Germany and Italy, HML is found to be negative. Only for Germany this is significant on a 5% level. So for these two countries a growth stock bias is documented.
Thus, hypothesis three cannot be rejected for ethical mutual funds in Belgium, France, Italy and the UK. In these countries ethical mutual funds are not found to be tilted significantly towards growth/value stocks. For Germany, Austria, Switzerland and the Netherlands hypothesis two is rejected. However, for Austria, Switzerland and the Netherlands, ethical mutual funds are significantly tilted towards value stocks. In Germany ethical mutual funds are tilted towards growth stocks.
bias. The insignificant value stock bias in France and Belgium also confirms results by Renneboog et al. (2007). The insignificant value stock bias in the UK contradicts Bauer et al. (2005), who found a significant growth stock bias. Finally, the significant value stock bias in Switzerland contradicts Renneboog et al. (2007), who document an insignificant growth stock bias.
In conclusion, results are mixed. Some findings confirm previous research, while others contradict the results by others. However, that in six out of eight countries a value stock bias is documented is quite surprising, because in general, previous research predominantly documented growth stock biases. A reason for this might be the role of traditional value sectors like chemicals, energy and basic industries. Previously, these industries were predominantly excluded from ethical mutual funds, because of their higher environmental risk.14 However, the finding that in several countries ethical mutual funds tend to have a value focus, might lead to the conclusion that nowadays these firms also improved their ethical behavior and have become less risky, so that nowadays they can be included in ethical mutual funds. Secondly, it might also be the case that firms or sectors that previously were growth oriented have matured and have become more value focused.
Momentum
Finally the test results for the momentum variable are analyzed. For France, Germany, Italy, the Netherlands, and Switzerland the momentum factor is found to be insignificantly negative. For Austria momentum is insignificantly positive. For Belgium and the UK, results are mixed.
Though the results are all insignificant, comparing these results with previous research does not give any surprises. Negative values for ethical mutual funds in France, Germany and Italy confirm the work by Renneboog et al. (2007). The negative value for the Netherlands confirms results by Scholtens (2004), but contradicts Renneboog et al. (2007), who documented an insignificant positive momentum value. Opposed to this research, Bauer et al. (2005) also documented a positive and significant momentum factor in Germany and the UK. The predominantly negative values might lead to the conclusion that ethical mutual funds did not follow a very successful momentum strategy.
5.3 Analysis of the sign test
In order to find out whether the test results of the three hypotheses concerning alpha, SMB and HML differ significantly across countries, a sign test is conducted. To my knowledge this test has not been used before in a context like this.
In a sign test the sign of the difference between two estimated values is determined and then the number of negative (or positive) signs is counted. This number of signs is tested in a binominal distribution with a probability of success p=1/2 under the hypothesis:
H4: The estimator of this country is not significantly different than the estimator of the other country
H4a: The estimator of this country is significantly higher/lower than the alpha of the other country
In the case of the Carhart four-factor model, five different indices have been tested and the estimators are compared. Therefore, the significance values are based on the binomial distribution with n=5, p=1/2 and t as the number of negative/positive signs. When the significance value is under 0,0500 the null hypotheses of no significant difference in sign is rejected and it is concluded that the country has significantly lower/higher estimated values than the other country. The results of the sign test can be found in table 7. Based on these results we can conclude the following.
The comparative results for performance indicator alpha are in table 7a. The relevant hypothesis here is whether the alpha is significantly lower than the alpha of the other country. Austria has significantly lower estimators of alpha compared to all other countries, followed by Belgium that, except for Austria also has significantly lower alphas than all other countries. France and Germany are not significantly different from each other. Also Italy, The Netherlands, and the UK are not significantly different from each other, but they have significantly lower values than France and Germany. This leads to the conclusion that ethical mutual funds perform significantly the worst in Austria compared to all other countries, followed by Belgium. All countries perform significantly worse than France and Germany.
Switzerland have significantly higher values than Belgium. All other results are not significantly different from each other. This leads to the conclusion that Austrian and UK ethical mutual funds are significantly more tilted towards firms with lower market capitalization than the other countries. Dutch and Swiss ethical mutual funds are also significantly more tilted towards smaller stocks than Belgian ethical mutual funds. The rest is not significantly different from each other. However, this is just a relative case. In absolute terms, only Austrian were really small cap focused, the other seven countries were large cap focused (see table 6).
The comparative results for book-to-market-value factor HML are in table 7c. The relevant hypothesis here is whether the HML is significantly lower than the HML of the other country. There’s a strongly significant order in HML values. Germany has significantly lower HML values than all other countries, followed by Italy that, except for Germany, also has significantly lower HML values than all other countries. Thirdly is France, which has, except for Germany and Italy, lower HML values than all other countries. All countries have significantly lower HML values than Switzerland. This leads to the conclusion that Germany is significantly more tilted towards growth stocks than all other countries, followed by Italy and then France. All countries appear to be significantly more growth oriented than Switzerland. These results are also relative. In absolute terms, only German and Italian ethical mutual funds are growth oriented, while the other six countries were value oriented.
In short, compared to other countries, the relative growth orientation in combination with a focus on high capitalized firms in Germany and France seems to pay off, since they perform relatively best. The small size focus in Austria does not seem to pay off, since they perform relatively worst. However, in general these findings are very relative inconclusive, so it is very difficult to derive meaningful conclusions.
6. Conclusions
After first reviewing the history and development of Socially Responsible Investing (SRI) in general, this paper analyses the performance and investment style of ethical mutual equity funds in eight European countries in the period 2001-2008. A database is constructed containing 116 ethical mutual funds from Austria, Belgium, France, Germany, Italy, the Netherlands, Switzerland and the UK.
cases. However, except for Italy, the differences are not significant. In line with previous researches, it is documented that generally ethical indices are more powerful in explaining ethical mutual fund performance than standard indices.
Secondly a Carhart (1997) four-factor model is applied, which controls for size, book-to-market value and momentum. This model improves the test results, and ethical indices are again more powerful in explaining ethical mutual fund performance. In France, Germany and Switzerland it is documented that ethical mutual funds outperform the market indices, but that there is still underperformance in Austria, Belgium, Italy, the Netherlands and the UK. However, none of the results is significant. Therefore we cannot conclude that ethical mutual funds perform significantly different than the market. In contrast to most previous research, it is concluded that in seven out of eight countries (not in Austria) ethical mutual funds are biased towards stocks with high market capitalization. This large cap bias is significant for Belgium, France, Germany and Italy. A reason for this may be that small ethical firms which performed good, matured and developed into large firms. Another explanation might be that the underperformance of ethical mutual funds in early stages, leaded to a growing focus on less risky large capitalized firms. Furthermore, in six out of eight countries a value stock bias is documented. For Austria, Switzerland and the Netherlands this is significant. This is in contrast with most previous researches that found a growth stock bias for ethical mutual funds. A reason for this might be that traditional value sectors have improved on their ethical and environmental behavior, so that they nowadays can be included in ethical mutual funds. Secondly, it might also be the case that firms or sectors that previously were growth oriented have matured and have become more value focused.
Finally, a sign test is applied in order to test the relative differences between the eight countries. Relative to other ethical funds in other countries, ethical mutual funds are found to perform significantly the worst in Austria and Belgium, and best in France and Germany. Relative to other countries Austrian and UK ethical mutual funds are the most focused on firms with small market capitalization. Ethical funds in Germany are mostly tilted towards growth firms, followed by Italy and then France. However, in general these results are very relative and inconclusive and no meaningful conclusions can be drawn.
References
Basso, A. and Funari, S. (2002) “Measuring the performance of ethical mutual funds: A DEA approach”, Working Paper, University of Venice, pp. 1-19
Bauer, R., Koedijk, K., en Otten, R. (2005) ”International evidence on ethical mutual fund performance and investment style”, Journal of Banking and Finance 29, pp. 1751-1767 Bauer, R., Otten, R. and Rad, A.T. (2006) “Ethical investing in Australia: Is there a financial
penalty?”, Pacific-Basin Finance Journal 14 (1), pp. 33-48
Brooks, C. (2002) Introductory econometrics for finance, first edition, Cambridge University
Press, Cambridge, United Kingdom
Carhart, M. (1997) “On persistence in mutual fund performance”, Journal of Finance 52, No. 1, pp. 56-82
Derwall, J. and Koedijk, K. (2005) “The performance of socially responsible bond funds”,
Working Paper, Erasmus University Rotterdam, pp. 1-29
Derwall, J., Guenster, N., Bauer, R. and Koedijk, K. (March 2005) “The eco-efficiency premium puzzle”, Financial Analysts Journal, pp. 51-63
Elton, E.J. and Gruber, M.J. (1995) Modern portfolio theory and investment analysis, fifth edition,
John Wiley & Sons Inc, New York, USA
Duurzaamgeldgids (2006), Issue September, Amersfoort, pp. 1-56
Gregory, A., Matatko, J. and Luther, R. (1997) “Ethical unit trust financial performance: small company effects and fund size effects”, Journal of Business Finance and Accounting 24, pp. 705-725
Guenster, N., Derwall, J., Bauer, R. and Koedijk, K. (2005) “The economic value of corporate eco-efficiency”, Working Paper, pp.1-39
Guerard, J. B. (1997) “Is there a cost to being socially responsible investing?”, The Journal of
Forecasting 16, pp. 475-490
Hamilton, S., Jo, H., and Statman, M. (1993) “Doing well while doing good? The investment performance of socially responsible mutual funds”, Financial Analysts Journal 49 (6), pp. 62-66
Jegadeesh, N. and Titman, S. (1993) “Returns to buying winners and selling losers: Implications for stock market efficiency”, Journal of Finance 48, p. 65-92
Kreander, N., Gray, G., Power, D.M. and Sinclair, C.D. (2005) ”Evaluating the performance of ethical and non-SRI funds: a matched pair analysis”, Journal of Business, Finance and
Lee, D.D. and Faff, R.W. (2006) “The corporate sustainability discount puzzle”, Working Paper, University of Queensland, pp. 1-32
Luther, R. and Matatko, J. (1994) “The performance of ethical unit trusts: choosing an appropriate benchmark”, British Accounting Review 26, pp. 77-89
Luther, R., Matatko, J. and Corner, D. (1992) “The investment performance of UK ethical unit trusts”, Accounting, Auditing & Accountability Journal Review 5, p. 57-70
Renneboog, L., Horst, ter, J. and Zhang, C. (May 2007) “The price of ethics: Evidence from socially responsible mutual funds”, Finance Working Paper, No. 168/2007, pp. 1-45 Renneboog, L., Horst, ter, J. and Zhang, C. (June 2007) “Socially responsible investments:
Methodology, risk exposure and performance”, Finance Working Paper, No. 175/2007, pp. 1-45
Scholtens, B. (2004) “Style and performance of Dutch social responsible investment funds”,
Research Report, RijksUniversiteit Groningen, pp.1-30
Schroder, M.(2004) “The performance of socially responsible investments: investment funds and indices”, Financial Markets and Portfolio Management 18(2), pp. 122-142
Social Investment Forum (SIF) (2001, 2003, 2007) “Report on responsible investing trends in the United States”, www.socialinvest.org
Social Investment Forum (2005) “Report on socially responsible investment trends in the United States”, a 10 year Review, Washington
Appendices
Table 1: Research methodologies of SRI studies
15Study Country Period No. Of Funds
Performance Measures
Market Indices Reference Group (non-SRI funds)
Luther, Matatko and Corner (1992)
UK 1984-1990 15 CAPM FT All Share Index
or MSCI World Index
No comparisons with non-SRI funds Luther and
Matatko (1994) UK 1984-1992 9 CAPM FT All Share Index or a Small Cap Index
No comparisons with non-SRI funds Hamilton, Joe and
Statman (1993) US 1981-1985 1986-1990 32 CAPM Value-weighted NYSE Index 320 non-SRI funds, randomly selected Mallin, Saadouni
& Briston (1995) UK 1986-1993 29 CAPM FT All Share Index 29 non-SRI funds, matched by fund size and age
Gregory, Matatko
and Luther (1997) UK 1986-1994 18 A two-factor model with two indices
FT All Share Index and Hoare Govett Small Cap Index
18 non-SRI funds, matched by fund size, age, investing area and fund type
Goldreyer, Ahmed and Diltz (1999)
US 1981-1997 49 CAPM Wilshire 5000
Equity Index (for equity funds)
180 non-SRI funds, matched by investment objective, fund size and market beta
Statman (2000) US 1990-1998 31 CAPM S&P 500 or DSI
400 Index 62 non-SRI funds, matched by fund size Greczy, Stambaugh and Levin (2003) US 1963-2001 35 CAPM, Fama-French (1992), Carhart (1997), Pastor and Stambaugh (2002) CRSP NYSE/AMEX/ NASQAQ Index 894 non-SRI funds, including dead funds
Schroder (2004) US, Germany,
Switzerland 1990-2002 46 A two-factor model with two indices, Timing: Treynor and Mazuy (1966), Conditional: Ferson and Schadt (1996), (Strong) Style Analysis
MSCI World Index and Salomon Smith Barney World Index (for international funds)
S&P 500 and Wilshire Small Cap 250 Index (for domestic US funds)
No comparisons with non-SRI funds
Kreander, Gray, Power and Sinclair (2005) Belgium, Germany, Netherlands, Norway, Sweden, Switzerland and UK 1996-1998
(weekly) 40 CAPM Timing: Hendriksson-Merton (1981)
MSCI World Index 40 non-SRI funds, matched by fund size, age, country and investment universe
Bauer, Koedijk
and Otten (2005) Germany, UK and US 1990-2001 103 CAPM Carhart (1997) Conditional: Ferson and Schadt (1996)
MSCI World Index or DJ Sustainability Global Index (for international funds) FT All Share Index or EIRIS ethical balance (for UK domestic funds) S&P 500 or DSI 400 (for US domestic funds) 4384 non-SRI funds (Germany 114, UK 396, US 3874), including dead funds Renneboog, Ter Horst and Zhang (2006)
17 countries around the world
1992-2003 410 CAPM Worldscope value-weighted Equity Index
649 non-SRI funds in the UK
Bauer, Otten and Tourani Rad (2006) Australia 1992-2003 25 CAPM Carhart (1997) Conditional: Ferson and Schadt (1996) Worldscope value-weighted Equity Index or Westpac Monash Eco Index
281 non-SRI funds including dead funds
Bauer, Derwall and Otten (2006) Canada 1994-2002 8 CAPM Carhart (1997) Conditional: Ferson and Schadt (1996) Worldscope value-weighted Equity Index or Jantze Social Index 267 non-SRI funds including dead funds
Barnett and
Salomon (2006) US 1995, 1997, 1999 (yearly) 67 Average return No benchmark index No comparisons with non-SRI funds Renneboog, Ter
Table 2: European ethical mutual funds included in the research, based on country
France United Kingdom
AG2R ACTIONS ISR AGICAM ISICAE ABERDEEN ETHICAL WORLD A INC AGF EURO ACTIONS (C) ASSURANCES GENERALES AEGON ETHICAL INC.AC.A AGF VALEURS DURABLES AGFASSET MANAGEMENT ECCLESIASTICALAMITY A ATOUT VALEURS DURABLES CREDIT AGRICOLE AM AXA ETHICAL ACC I AXA EURO VALEURS RESP.C AXA INVESTMENT PARIS AXA ETHICAL ACC R
BNP PARIBAS ETHEIS BNP PARIBAS AXA FRAMLINGTON HEALTH ACC BNP PARI.RETRAITE HRZ.I 100 BNP PAM BNP PARI. CIS SUSTAINABLE LEADERS TRUST CAAM ACTIONS DURABLES CREDIT AGRICOLE AM CIS UK FTSE4GOOD TRACKER INC. CAAM ACTIV. DUR. CR.AGRICOLE ASTMGMT.CR CS FELLOWSHIP R
CREDIT MUT ETHIQUE(C) CREDIT MUTUEL DIRECT LINE FTSE4 GOOD TRKR.STD DYNALION DEV.DURABLE (C)CLAM - CREDIT LYON AM F&C STEWARDSHIP GROWTH SC1 AC. ECUREUIL 1 2 3 FUTUR FCPCAISSES D'EPARGNE F&C STEWARDSHIP INCOME SC1 AC. EPARGNE ETHIQUE ACTIONS CREDIT COOPERATIF F&C STEWARDSHIP INTL SC1 AC. ETHIS VITALITE(C) PRADO EPARGNE FAMILY CHARITIES ETHICAL
ETOILE ENVIRONNEMENT CREDIT DU NORD FIRST STATE ASIA PAC SUSTAINABLTY A ETOILE PARTENAIRES CREDIT DU NORD HALIFAX ETHICAL C
EURO CAPITAL DURABLE (C)GAN - GROUPAMA HENDERSON INDUSTRIES OF THE FUTURE A EURO ETHIQUE INSTITUTIONCIPF HENDERSON GLOBAL CARE GROWTH Z ACC EURO ACTIVE INVESTORS PHITRUST ACTIVE INVRS. HENDERSON GLOBAL CARE INCOME Z ACC EU.GOUVERNANCE (C) INTEGRAL DEVELPMENT AM INSIGHT EVERGREEN RET. AC.
EUROSOCIETALE (C) INTEGRAL DEVELPMENT AM INSIGHT EUR.ETHI.RETAIL ACC FEDERAL ACTIONS ETHIQUESFEDERAL GESTION JUPITER ENVIRONMENTAL INCOME INC GENERATION ETHIQUE (C) BANQUE DE FINANCEMENT MARLBOROUGH ETHICAL A ACC HSBC AM MULTIMAN ETH MONHSBC AME FRANCE FCP NORWICH UK ETHICAL A HSBC DVPPT.DURABLE A HALBIS CAPITAL MAN. OLD MUT.ETHICAL A AC.
INSERTION-EMPLOIS(D) CDCIXIS SCOTTISH WIDOWS ENVIRONMENTAL PENS IXIS EURO 21 CDC ASST.MAN.EUROPE SCOTTISH WIDOWS ETHICAL
LBPAM ACTIONS DEV. DURABLE C LBPAM STANDARD LIFE ETHICAL 1
MACIF CROISSANCE DURABLEEUROPE D STANDARD LIFE UK ETHICAL RETAIL ACC
MAM ACT ENVIRONNEMENT(C)MEESCHAERT FCP NORWICH SUSTAINABLE FUTURE ABSOLUTE GROWTH 1 MAM ACT ETHIQUE (C) MEESCHAERT FCP NORWICH SUSTAINABLE FUTURE EUR.GROWTH 1 MG CROISSANCE DURABLE(D)EUROPE MACIF GESTION NORWICH SUSTAINABLE FUTURE GLOBAL GROWTH 1 ORSAY CROISSANCE RESPONSABLE ORSAY AM NORWICH SUSTAINABLE FUTURE UK GROWTH 1 SARASIN ER.MIDCAP EX DUREXPERTISE AM SWIP GLOBAL SRI A
SARASIN EU.EXPAN.DURABLEEXPERTISE AM SWIP PAN EUR.SRI EQ.E EUR SGAM INVEST EUROPE DEV. DURABLE SGAM SVM ALL EUROPESRI A
Austria Belgium
3 BANKEN NACHHALTIGKEITSFONDS ATHENA RESPONSIBLE WLD. EQUITY CAP
ESPA WWF STOCK UMWELT T DEXIA SUSTAINABLE EMU CLC.C CAP.
HYB.GLOBAL VALOR DEXIA SUSTAINABLE EUROPECLC.C CAP.
KEPLER ETHIK AKTIENFONDSA DEXIA SUSTAINABLE NORTH AMERICA CLC.C CAP KEPLER SUSTAINABILITY AKTIENFONDS T DEXIA SUSTAINABLE PAC. CLC.C CAP.
GUT.KAPITALANLAGE PRI. VAL.AKN.EUROPA PAM EQUITIES EUROPE ETHICAL C RAIFFEISEN ETHIK AKTIENFONDS A BEVEK 21 PARTICIPATIE-21
RAIFFEISEN ETHIK AKTIENFONDS T TRIODOS VALUES FD.INTL. EQTIES.CAP RAIFFEISEN KPL.GESELL. ETHIK AKTIEN A Germany
ALGSPK.ETHIKAKTIEN GERLING INVESTMENT KPL. SELECT21
SUPERIOR 4 ETHIK AKTIEN INVESCO KPL.UMWELT UND NACHHALTIGKEITS FONDS
KEPLER SUSTAINABILITY AKTIENFONDS A UNION INV.PRIVATFONDS KCD UN.NACHHALTIG AKN. MEAG MUNICH ERGO KPL. NACHHALTIGKEIT
Italy Netherlands
AUREO FINANZA ETICA AEGON OPTIMUM DUURZAAM WERD.AAN.CAP
UBI PRA AZIONARIO ETICO ASN AANDELENFONDS
CAAM AZIONI SR ASN MILIEU & WATERFONDS 2
DUCATO ETICO GEO A ING BK.DUZ.RENDEMENT FD.
GESTIELLE ETICO AZI. ORANGE SENSE FUND DS.
SANPAOLO AZ.INTL.ETICO POSTBANK DUURZAAM AANDELFONDS
Switzerland ROBECO DUURZAAM AANDFDS
PICTET ETHOS CH SWISS SUSTAINABLE EQ. E E SNS DUURZAAM AANELENFONDS
RAIFF.FUTURA GLB.STOCK TRIODOS MEERW.AANDFDS
RAIFF.FUTURA SWS.STOCK
SWISSCA FONDSL. SWISSCANTO CH EQ.GREEN I
Table 3: Key statistics of ethical mutual funds in European countries at 01-01-2007
Country
Number of
ethical funds Bonds Equity
Total net assets in SRI funds in million EUR
Median value of fund size in EUR
Average initial fee in EUR Austria 17 5 12 € 286,35 € 31,73 € 4,28 Belgium 15 7 8 € 490,59 € 86,83 € 2,08 France 41 5 36 € 3.337,14 € 64,29 € 2,58 Germany 8 3 5 € 102,83 € 23,03 € 3,63 Italy 13 7 6 € 70,19 € 23,03 € 1,83 Netherlands 11 2 9 € 845,61 € 59,81 € 0,42 Switzerland 6 2 4 € 701,03 € 170,96 € 0,00 UK 36 0 36 € 4.994,82 € 102,06 € 0,29 Total 147 31 116 € 10.828,55
Table 4: Summary statistics based on monthly returns of ethical mutual funds, national
indices and global indices 2001:2-2008:1*
Table 5: CAPM test results with different market indices for all countries from 2001:2 to 2008:1
Austria Alpha Beta Adj. R-squared Durbin Watson
DJSI -0.0035 0.8724*** 0.6940 2.5989 FTSE4GoodEurope -0.0049 0.8396*** 0.7106 2.5811 FTSE4GoodGlobal -0.0049 0.8396*** 0.7106 2.5811 MSCI World -0.0058 0.9107*** 0.6503 2.4296 S&P 1200 -0.0096** 0.8143*** 0.5588 2.1798 S&P 350 -0.0057* 0.8467*** 0.6976 2.5901 ATX 20 -0.0268*** 0.3347*** 0.1569 2.0572
Belgium Alpha Beta Adj. R-squared Durbin Watson
DJSI -0.0038 0.8158*** 0.6633 2.5731 FTSE4GoodEurope -0.0048 0.7980*** 0.7019 2.4880 FTSE4GoodGlobal -0.0048 0.7980*** 0.7019 2.4880 MSCI World -0.0065* 0.8320*** 0.5924 2.3863 S&P 1200 -0.0101** 0.7396*** 0.5028 2.1536 S&P 350 -0.0057* 0.8009*** 0.6822 2.4996 NR_BEL -0.0165*** 0.5163*** 0.3588 2.2510
France Alpha Beta Adj. R-squared Durbin Watson
DJSI -0.0007 0.8991*** 0.6554 2.5674 FTSE4GoodEurope -0.0009 0.9062*** 0.7371 2.5910 FTSE4GoodGlobal -0.0009 0.9062*** 0.7371 2.5910 MSCI World -0.0029 0.9456*** 0.6234 2.5113 S&P 1200 -0.0065 0.8583*** 0.5522 2.3071 S&P 350 -0.0019 0.9121*** 0.7207 2.6129 NR_CAC -0.0309*** -0.1350*** 0.2717 1.7488
Germany Alpha Beta Adj. R-squared Durbin Watson
DJSI -0.0028 0.9340*** 0.6910 2.7046 FTSE4GoodEurope -0.0042 0.9015*** 0.7117 2.6086 FTSE4GoodGlobal -0.0042 0.9015*** 0.7117 2.6086 MSCI World -0.0057 0.9583*** 0.6249 2.4277 S&P 1200 -0.0097** 0.8545*** 0.5339 2.1733 S&P 350 -0.0052 0.9071*** 0.6953 2.6271 NR_DAX -0.0160*** 0.6198*** 0.5667 2.4347
Italy Alpha Beta Adj. R-squared Durbin Watson
Netherlands Alpha Beta Adj. R-squared Durbin Watson DJSI -0.0023 0.8944*** 0.6720 2.6520 FTSE4GoodEurope -0.0037 0.8640*** 0.6931 2.6162 FTSE4GoodGlobal -0.0037 0.8640*** 0.6931 2.6162 MSCI World -0.0055 0.9048*** 0.5904 2.4465 S&P 1200 -0.0095** 0.8002*** 0.4958 2.2100 S&P 350 -0.0047 0.8672*** 0.6739 2.6298 NR_AEX -0.0095** 0.6583*** 0.5769 2.4454
Switzerland Alpha Beta Adj. R-squared Durbin Watson
DJSI 0.0017 0.9138*** 0.6112 2.6058 FTSE4GoodEurope 0.0006 0.8936*** 0.6465 2.5813 FTSE4GoodGlobal 0.0006 0.8936*** 0.6465 2.5813 MSCI World -0.0013 0.9325*** 0.5466 2.4401 S&P 1200 -0.0053 0.8287*** 0.4635 2.2449 S&P 350 -0.0004 0.8979*** 0.6300 2.6309 NR_SWI -0.0037 0.8008*** 0.5333 2.5818
UK Alpha Beta Adj. R-squared Durbin Watson
