Social responseble investment : a study on returns and adjusted risk of different categories of social responsible investment

SOCIAL RESPONSIBLE

INVESTMENT

A study on returns and adjusted risk of different

categories of social responsible investment

Frans de Heij

Abstract

As investors are more and more searching for sustainable investment opportunities and as the demand of sustainable stocks is increasing over the past years the question arises whether this has an significant impact on the returns. In this research sustainability is divided in four categories, social, corporate, ethical and environmental. Each of this categories is tested whether the returns and adjusted risk differs from the conventional portfolio. For this the Four-Factor Cahart Model is used. Stocks who have a high valuation on social and environmental aspect have a higher return and higher risk, while stocks that have a high score on corporate aspect underperform.

CONTENT

1. Introduction: ... 2

2. Theoretical Framework ... 3

2.1 Social Responsible Investment ... 3

2.2 Performance of SRI funds ... 3

Theory ... 4

Empirics ... 4

2.3 Capital assest pricing model ... 4

2.4 Four Factor Cahart Model ... 5

3. Method ... 6

3.1 Data categorization ... 6

3.2 Four factor Cahart model ... 7

4. Data Description ... 8

5. Results: ... 9

4.1 CAPM ... 9

4.2 Four-factor Cahart Model ... 10

5. Conclusion: ... 11

6. Discussion: ... 11

7. Literature: ... 12

1. INTRODUCTION:

As a response to the social and environmental challenges of global industrialization, there has been a growing interest among investors in Western Europe and USA to invest with environmental and social considerations in mind (Eurosif, 2008). Socially responsible investment (SRI) has therefore gained ground also among investment institutions. SRI is an investment process that integrates social, environmental, and ethical considerations into investment decision making (Renneboog, Horst & Zhang, 2008, p. 1723). From a marginal fraction associated with churches and charity organizations, SRI has become more mainstream, especially among institutional investors such as labour unions, pension funds and municipalities (Jansson & Biel, 2010). Based on data of the Social Invest Forum in 2000 almost 2 trillion dollars was under professional management in the United States in a socially responsible portfolio which rose to over 3 trillion in 2010.

For most providers of SRI, for example investments banks, SRI is a minor share of their investments. They implement SRI products with the aim to capture a niche market of SRI investors (Sparkes, 2011). However, for some investment banks all their capital is invested in SRI because they believe that SRI promotes sustainability and that SRI gives financial advantages in terms of risk reduction and long-term returns (Cumming & Johan, 2007)

SRI can be separated into four categories; ethical responsible, social responsible, environmental responsible and corporate responsible (Renneboog et al, 2008). In this research the effect of the four categories on the expected returns and sensitivity to the market risk is investigated. Previous research, such as Bauer et al., 2004, focussed on SRI as a whole package, while this research will divide SRI in the four categories mentioned above and see how the returns and adjusted risks of those categories behave. The research question is: ‘Do different SRI portfolio differ in adjusted risk and returns compared to NON-SRI portfolio?’ Two hypothesis will be tested.

HYPOTHESIS 1

H0: 𝜷𝟏𝒊 = 𝜷𝟏𝒋 ‘The portfolio do not differ in adjusted risk over the period of 1990-2013’ H1: 𝜷𝟏𝒊 ≠ 𝜷𝟏𝒋 ‘The portfolio differ in adjusted risk over the period of 1990-2013’

HYPOTHESIS 2

H0: 𝜶𝒊= 𝜶𝒋 ‘The portfolio do not differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

H1: 𝜶𝒊≠ 𝜶𝒋 ‘The portfolio do differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

Where 𝜷𝟏is the coefficient of the risk premium and α is the intercept in the four factor CAPM. I and J indicate different portfolio.

In order to test these hypothesis firstly the current state of literature will be listed and summarize, secondly a method for testing the hypothesis will be given, after that the results and conclusions will be listed.

2. THEORETICAL FRAMEWORK

Friedman (1970) argued that companies should only care for the interest of their shareholders, as profit was the main incentive for the shareholders to participate. Renneboog et al (2008) argue that besides making money shareholders benefit on other aspects if the company invests in SRI, in fact SRI shareholders are less concerned about negative returns than conventional investors. SRI shareholders expect companies to maximize welfare both socially and financially (Renneboog et al, 2011). Those social aspects exist of concern for ethics, environment, social relations and corporate governance (Galema et al., 2008). The question is whether a trade-off exists between the return and SRI.

2.1 SOCIAL RESPONSIBLE INVESTMENT

Schueth (2003) argues that investors who invest in SRI basically have two motivations, often complementary. One part feels the need to put their money to work in a manner that is closely aligned with their personal preferences. Also named ‘feel good’ investors. The other part feels a strong need to enhance the quality of life. This part is more focussed on what their money can do to catalyse positive change in society at large. They tend to be more interested in the “social change” strategies that are an integral part of the SRI field in the US. SRI is an investment process that integrates social, environmental, and ethical considerations into investment decision making (Renneboog, Horst & Zhang, 2008, p. 1723). Schueth (2003) defines SRI as the process of integrating personal values and societal concerns into investment decision-making.

SRI can be separated into four categories; ethical responsible, social responsible, environmental responsible and corporate responsible. Ethical responsible is about the field the company is active in, if these fields are related to for example weapons, animal testing, adultery or tobacco a company is not consider to be ethical. Social responsible is about the relations of the employee and the employers. If a company is social responsible it invests in its relation with the employers for example through higher wages and/or facilities. Environmental responsible is about reducing the ecological footprint of a company. A company is considered environmental responsible if it lowers the emissions and contributes to a more sustainable business cycle. Corporate responsible involves the contributions of a company to the society, for example providing education opportunities (Galema et al., 2008).

2.2 PERFORMANCE OF SRI FUNDS

Galema et al (2008) studied the contradiction between the empirical and theoretical research. They figured that the method used in the previous research was not sufficient. Two errors occur with the most common method are related to the use of regression models such as the Fama French models (1992) and the use of aggregate measures of SRI. When SRI and non-SRI firms have the same risk but higher book-to-market ratios due to the excess demand for SRI stocks, the Fama French model implies that as a result of this excess demand the trade-off between financial and SRI performance is related to the book-to-market ratio.

Aggregation is over different categories, all subject to SRI, which may lead to contradicting effects. To solve this error the categories have to be taken apart. For example, positive news on environment could have a positive effect on the return, while news related to good employee relations might have a negative impact (Scholtens & Zhou, 2008).

THEORY

In theory SRI has an excess demand, and that NON-SRI stocks have therefor a shortage of demand. These differences in demand for SRI and NON-SRI are either due to missing information (Merton, 1987),

environmental preferences (Heinkel et al., 2001) or due to externalities (Dam, 2008). An excess demand for SRI stocks would suggest that the returns on SRI stocks are lower than the returns of NON-SRI stocks. Portfolio theory indicates that SRI increase portfolio risk because the exclusions of stocks, sectors, and countries will lead to less opportunities for diversification (Langbein and Posner, 1980).

Schueth (2003) argues that SRI has matured to a point where any investment need can be met through portfolio design which integrates an investor’s personal values, institutional mission and/or social priorities.

EMPIRICS

However there is no empirical evidence that responsible stocks give a lower return (Renneboog et al., 2008). Galema et al (2008) found that SRI do have a significant lower return than non-SRI. Bauer et al (2004) found no evidence that conventional funds differ significant from SRI funds in the period of 1990-2001. Statman (2000) found that SRI outperform the market in the period of 1990-1998. Hamilton et al (1993) indicated that SRI funds have outperformed the market in the period 1981-1985, but underperformed in the period 1986-1989. These contrary findings might be due to different methods, terminology and portfolio. However SRI as a whole does not have a clear effect on the expected return. High scores on environmental aspects, in contrary to corporate governance aspects, have a negative impact on the book-to-market ratios. High scores on employment aspects do not have a significant impact on the expected returns (Hamilton et al, 1993). In line with Galema, Hong & Kacperczyk (2007) found that ‘sin’ stocks are under-priced and have a higher expected return than other stocks. ‘Sin’ stocks are stocks which are sometimes excluded from portfolio because of the negative ethical issues, for example stocks related to weapons.

2.3 CAPITAL ASSEST PRICING MODEL

The capital asset pricing model (CAPM) determines a theoretically required rate of return of an asset, if that asset is to be added to an already well-diversified portfolio, given that asset is non-diversifiable. The model takes the asset's sensitivity to market risk represented by the beta (𝛽1𝑖), as well as the expected return of the market and the expected return of a theoretical risk-free asset. CAPM shows the investor’s cost of equity capital is by beta 𝛽1𝑖.The CAPM is tested with a regression of historical data of the fund/stock on the market performance. The coefficient of the market performance captures the sensitivity to market risk (Mossin, 1966). The formula for CAPM is as below, where 𝑅𝑖𝑡 is the return of the stock, 𝑅𝑓𝑡 is the risk free rate, 𝛼𝑖 is the

intercept or Jensen’s Alpha and 𝑅𝑚𝑡 is the market return. Jensen’s Alpha measures the out- or

under-performance relative to the market proxy (Jensen, 1968). The beta, 𝜷𝟏𝒊 is a measure of the risk from exposure to general market movements as opposed to non-diversable factors. The market portfolio of all investable assets has a beta of 1. A beta higher than 1 indicates an investment with higher volatility than the market (Sharpe, 1970).

𝑹𝒊𝒕− 𝑹𝒇𝒕= 𝜶𝒊+ 𝜷𝟏𝒊�𝑹𝒎𝒕− 𝑹𝒇𝒕�

2.4 FOUR FACTOR CAHART MODEL

Fama and French (1992) argued that the CAPM could be improved by adding two variables, namely . Small market capitalization Minus Big (SMB) and High book-to-market ratio Minus Low (HML). The Book-to-market ratio is the book value of a firm divided by the market value of the firm. SMB and HML calculate the historic excess returns of small caps over big caps and of growth stocks over value stocks. These factors are formed with combinations of portfolios consisting of rated stocks and historical data (Fama & French, 1993). Explaining the expected return by the three variables is known as the three-factor model. In addition Carhart added in 1997 another variable, momentum. Momentum is the weighted average of the least performing firms minus the weighted average of the best performing firms. The four factor model is as below. Where SMB is Small Minus Big, HML is High Minus Low and MOM is momentum.

𝑹𝒊𝒕− 𝑹𝒇𝒕= 𝜶𝒊+ 𝜷𝟏𝒊�𝑹𝒎𝒕− 𝑹𝒇𝒕� + 𝜷𝟐𝒊𝑺𝑴𝑩𝒕+ 𝜷𝟑𝒊𝑯𝑴𝑳𝒕+ 𝜷𝟒𝒊 𝑴𝑶𝑴𝟒𝒊+ 𝜺𝒊𝒕

3. METHOD

In this chapter a brief explanation of where the data comes from will be given, as well as the categorization and will propose a technique to analyse this data.

3.1 DATA CATEGORIZATION

This paper will divide SRI into four categories in order to see whether these categories preform differently from each other and from NON-SRI stocks. The focus is on the US stock market in the period 1990-2014. The dataset for this research will contain the Dow Jones Industrial Average, consisting the daily data of 30 stocks and is collected from Datastream and the French (2014) website.

The categorization of the stocks will be based on screening as mentioned in table 1. Based on the definitions in table 1 each of the stocks has been examined and categorized. Information to do the screens was acquired by the following three rating agencies. CSR RepTrack© (2013), a company which rates large companies on their corporate responsibility based questionnaires. CDP global 500 (2013), a company which rates companies on their accomplishment on reducing their ecological footprint based on screening and testing samples.

Greatplacetowork (2013), a global initiative to rate employees on their relation with their employers based on questionnaires. Some stocks have been added in multiple categories.

Table 1: Social Responsible Investment Screens. Source: Renneboog et al., 2011. Categories Type Screens Definitions

Ethical Negative Negative Negative Negative Negative Negative Negative Negative Positive Negative Tobacco Alcohol Gambling Weapons Pornography Animal testing Aborting Genetic engineering Healthcare Non-marital

Avoiding manufacturers of tobacco products Avoiding producers of alcoholic beverages

Avoiding casinos and suppliers of gambling equipment Avoiding firms producing weapons or firearms

Avoiding publishers of pornographic magazines or video tapes, or firms that provide adult-entertainment services

Avoiding firms that provide animal-testing services or involved in intensive farming of animals

Avoiding providers of abortion and manufacturers of abortion drugs or insurance companies that pay for elective abortions

Avoiding firms that develop genetically modified products Selecting firms whose products improve human health

Avoiding insurance companies that provide coverage to non-married couples Corporate

responsible Positive Positive Positive

Business practices Corporate governance Community

Selecting firms emphasizing product safety and quality

Selecting firms demonstrating best practices related to board independence, executive compensation, or other governance issues

Selecting firms with an active involvement in local communities Social Positive/ Negative Positive/ Negative Positive/ Negative Positive/ Negative Diversity Labour relations Human rights Foreign operations

Selecting firms pursuing active policies in employing minorities, women, gays/lesbians, and/or disabled persons; or avoiding firms discriminating on gender/race

Selecting firms that provide good workplace conditions, empowering employee and/or strong union relations; or avoiding firms with poor labour relations Selecting firms with policies to protect human rights; or avoiding firms with bas records on human rights issues

Selecting firms with human rights policies for foreign operations; or avoiding firms employing child labour overseas or operating in countries with oppressive regimes Environmental Positive/ Negative Positive Negative Environment Renewable energy Nuclear

Selecting firms with high environmental/ecological standards; or avoiding firms with low environmental standards

Selecting firms producing power from renewable energy Avoiding companies operating nuclear power plants

The SMB, HML, MOM, and Mkt factors will be gather from the French (2014) website. For the Risk-free rate 10 year US coupon are used.

3.2 FOUR FACTOR CAHART MODEL The Four-Factor Cahart model is:

𝑹𝒊𝒕− 𝑹𝒇𝒕= 𝜶𝒊+ 𝜷𝟏𝒊�𝑹𝒎𝒕− 𝑹𝒇𝒕� + 𝜷𝟐𝒊𝑺𝑴𝑩𝒕+ 𝜷𝟑𝒊𝑯𝑴𝑳𝒕+ 𝜷𝟒𝒊 𝑴𝑶𝑴𝟒𝒊+ 𝜺𝒊𝒕.

The model will be tested four times for each different category. The results of the different regression will be compared to see if either the 𝛼𝑖 or one of the 𝛽𝑖’s differs between the different portfolio. If 𝛼𝑖 differs between portfolio it indicates that the returns differ. If 𝛽1𝑖 differs across the portfolio it means that the sensitivity to the market risk is different. Differences in the other 𝛽𝑖’s indicates that the portfolio are differently related to the variables. As mentioned in the introduction this leads to the following hypothesis. Where 𝜷𝟏is the coefficient of the risk premium and α is the intercept in the Four-Factor Cahart model. I and J indicate different portfolio. Based on the differences between the coefficients and the intercept the hypothesis will be tested.

HYPOTHESIS 1

H0: 𝜷𝟏𝒊 = 𝜷𝟏𝒋 ‘The portfolio do not differ in adjusted risk over the period of 1990-2013’ H1: 𝜷𝟏𝒊 ≠ 𝜷𝟏𝒋 ‘The portfolio differ in adjusted risk over the period of 1990-2013’

HYPOTHESIS 2

H0: 𝜶𝒊= 𝜶𝒋 ‘The portfolio do not differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

H1: 𝜶𝒊≠ 𝜶𝒋 ‘The portfolio do differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

4. DATA DESCRIPTION

The data is collected and categorized as mentioned in chapter 3. In this chapter the data will be descripted and the characteristics will be given. In total over 180.000 observation where used for this research.

The category ethical is removed from the table as all the stocks were categorized as ethical. So four different portfolio remain, social, corporate responsible, environmental and conventional. The conventional category consist of the stocks who do not apply in any of the SRI categories. In table 4 the four categories are listed with their characteristics. Including scores of a two sided T-test for comparing means, between the partial portfolio and the complete portfolio. The t-test for comparing means is valid because of independency and the unequal variances of the portfolio.

Table 2: Characteristics of the categories

CATEGORY # OBSERVATIONS MEAN RETURN STANDARD DEVIATION MIN RETURN MAX RETURN T-TEST 𝐗𝟏 ���� − 𝐗���� ≠ 𝟎 𝟐 SOCIAL 22573 0.0007452 0.0247041 -0.24731 0.2646301 0.1424 CORPORATE 68849 0.0006172 0.0190364 -0.3138 0.2012 0.6325 ENVIRONMENTAL 35097 0.0007391 0.0215929 -0.299435 0.34755 0.1212 CONVENTIONAL 95440 0.0006225 0.0187909 -0.287356 0.2555532 0.5788

The conventional and corporate categories have the most observations, while social has the least. The mean return of social and environmental is the highest, while the mean return of the corporate category is the lowest, lower than the conventional. The highest return was made in the environmental category with 0.35%, the lowest was in corporate with -0.31%. The T-test scores are all insignificant.

5. RESULTS:

In this chapter first the results of the CAPM regression will be given. After that the results of the four factor Cahart model will be given. The results will be described and in the next chapter the results will lead to conclusions about the hypothesis.

4.1 CAPM

In table 3 the results of the CAPM regression are given. All the alphas are significant at the level of 1% while only the market for the conventional category is significant at 1%. The market for social is not significant at all, while the corporate is significant at 5% and the environmental at 10%. Social has the highest alpha, followed by environmental. The category corporate has a lower alpha than the conventional category. The conventional category has the highest coefficient for the market, and is the most related to the market. Social has the lowest, and insignificant, coefficient for market.

Table 3: CAPM results

CATEGORY ALPHA MARKET

SOCIAL -0.0105257*** 0.0001655

CORPORATE RESPONSIBLE -0.0114206*** 0.0002163**

ENVIRONMENTAL -0.0108079*** 0.0001787*

CONVENTIONAL -0.0112405*** 0.0003521***

* Significant at the 10% level ** Significant at the 5% level *** Significant at the 1% level

4.2 FOUR-FACTOR CAHART MODEL

In table 4 the results of the Four-Factor Cahart Model are given. All the four-factor alphas are significant at 1%. The coefficients for market are significant for corporate and conventional at 1% and for environmental at 10%, for social the market is not significant. SMB is significant for both the corporate and the conventional category at 5% and 1%. As for the conventional category all the variable are significant at the level of 1%. MOM and HML are both significant at 5% for the category social, and significant at 1% for the other three categories. The F scores are for all the categories significant at 1% except for environmental it was significant at 5%.

Table 4: Results Four factor Cahart Model

CATEGORY 4-FACTOR

ALPHA

MARKET SMB HML MOM F SCORE

SOCIAL -0.01052*** 0.002012 0.0007397** 0.0004308 -0.00059** 4.11** CORPORATE RESPONSIBLE -0.01142*** 0.002509*** 0.0005768*** 0.0002991** -0.00034*** 10.00*** ENVIRONMENTAL -0.01080*** 0.002081* 0.0007243*** 0.0002848 -0.00056*** 7.06*** CONVENTIONAL -0.01124*** 0.003926*** 0.0005078*** 0.0003923*** -0.00050*** 24.64***

* Significant at the 10% level ** Significant at the 5% level *** Significant at the 1% level

The four-factor alpha of environment and social are the highest, while the alpha of corporate is lower than the alpha of the conventional category. The category conventional has the highest coefficient, significant at 1%, for the market, while social has the lowest insignificant coefficient for the market.

5. CONCLUSION:

Concluding, based on this research and the data used in this research, stocks performing well on socially and environmental aspect give higher returns and have a lower correlation with the market. The categories conventional and corporate had the highest coefficients for the market and the highest correlation with the market. The systematic risk of the categories conventional and corporate is lower as the 𝜷𝟏𝒊 is closer to 1. The categories environment and social are more volatile to the market performance. H0 of hypothesis 1 is

rejected. The different portfolio differ in adjusted risk over the period of 1990-2013.

HYPOTHESIS 1

H0: 𝜷𝟏𝒊 = 𝜷𝟏𝒋 ‘The portfolio do not differ in adjusted risk over the period of 1990-2013’ H1: 𝜷𝟏𝒊 ≠ 𝜷𝟏𝒋 ‘The portfolio differ in adjusted risk over the period of 1990-2013’

The categories social and environmental had higher alphas in both the CAPM and the Four-Factor Cahart Model and performed better than the categories conventional and corporate responsible. The portfolio do differ in intercepts in the Four-Factor Cahart Model in the period 1990-2013. H0 of hypothesis 2 is rejected.

HYPOTHESIS 2

H0: 𝜶𝒊= 𝜶𝒋 ‘The portfolio do not differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

H1: 𝜶𝒊≠ 𝜶𝒋 ‘The portfolio do differ in intercepts in the Four-Factor Cahart Model in 1990-2013’

6. DISCUSSION:

While this research does find that stocks who perform well on environmental and social aspects have a higher return, it is questionable if this is not due to the method and limited data. The total observations for

environment, 35097, and social, 22537, are combined less than the total observations for conventional, 95440, or corporate, 68849. It is logical that the category with more observations and more stocks has a higher correlation to the market. Also as environment and social are small sample the chance that they outperform or underperform the market is higher based on basic statistics. Based on this research there is a difference in adjusted risk and return, but a larger scale research is required to make this claim more vast.

7. LITERATURE:

Bauer, R., Koedijk, K. & Otten, R. 2004. International evidence on ethical mutual fund performance and investment style. Journal of Banking & Finance 29, 1751-1767.

Dam, L., 2008. Corporate Social Responsibility and Financial Markets. PhD thesis. University of Groningen: Groningen.

Fama E., & French K.R., 1992. The cross-section of expected stock returns. Journal of Finance 47, 427-465.

Fama E., & French K.R., 1993. Common risk factors in the returns on stocks and bonds. Journal of Financial economics 33, 3-53.

French, Craig W. (2003). "The Treynor Capital Asset Pricing Model". Journal of Investment Management 1 (2): 60–72.

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.

Heinkel R., Krause A., Zechner J. 2001. The effect of green investment on corporate behaviour. Journal of Financial and Quantitative Analysis, 36, pp. 431–449.

Hong, H.G., Kacperczyk, M.T., 2007. The price of sin: the effects of social norms on markets, Sauder School of Business

Jensen, M., 1968. The performance of mutual funds in the period 1945-1964. Journal of Finance 23, 389-416.

Merton R.C., 1987. A simple model of capital market equilibrium with incomplete information. Journal of Finance, 42, pp. 483–510.

Renneboog L., ter Horst J., Zhang, C. 2008. Socially responsible investments: Institutional aspects, performance, and investor behaviour. Journal of Banking and Finance, 32, pp. 1723–1742

Scholtens B., Zhou, Y. 2008. Stakeholder relations and financial performance. Sustainable Development, 16, pp. 213–232

Sharpe, William (1970). Portfolio Theory and Capital Markets.