Socially Responsible Investment vs. Vice Investing
A study on the differences between the financial performances of SRI versus controversial business involvement screened companies, in comparison with the Russell
3000 index.
Bachelor’s Thesis Economics and Finance January 2016
Youri de Jong Academic year 2015/2016 10273972 dr. T. Jochem (supervisor)
Abstract
In this paper I will do a study on the differences between the financial performances of SRI and controversial business involvement screened companies, in comparison with the Russell 3000 index. The ethically theoretical difference between these two kind of companies is that one is involved with socially responsible and the other one with socially irresponsible business activities. Further the companies are screened by KLD Social Ratings through ESG criteria: environmental, social and governance. In addition the added criteria is the controversial business involvement, this screen determines in this research the vice companies used in the regression.
As a result, the literature research and data analysis will lead to some concrete findings. The most important one is with the used sample, regression models and in the time frame of January 1st and December 31st 2012 there is a significant difference in abnormal returns between SRI and vice stocks in comparison to the Russell 3000 index. However there will be always an ethical dilemma which stock is better to invest in. It depends on the financial and non-‐financial preferences of an investor which one to choose.
Statement of Originality
This document is written by student Youri de Jong who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Content
2. Theoretical framework ... 6
2.1 Social Responsible Investment (SRI) ... 6
2.2 Vice Investing ... 8
2.3 Russell 3000 Index ... 10
2.4 Summary Chapter ... 10
3. Methodological framework ... 11
3.1 Data Selection ... 11
3.1.1. MSCI ESG KLD STATS -‐ Social Ratings ... 11
3.1.2 SRI ... 13
3.1.3 Vice ... 14
3.2 Capital Asset Pricing Model (CAPM) ... 16
3.3 The Fama-‐French Three-‐Factor Model ... 16
3.4 Carhart Four-‐Factor Model ... 17
3.5 BHRR & BHAR ... 17 3.6 Research Restrictions ... 18 3.6.1 Selection Process ... 18 3.6.2 Regression Model ... 18 4. Data results ... 19 4.1 Regression Analysis ... 19
4.2 BHRR & BHAR Analysis ... 20
5. Conclusion ... 21 References ... 23
1. Introduction
One of the new developments in the financial world is the rapid increase of socially responsible investments (SRIs), often named sustainable investments or ethical investments. SRIs entail an investment decision-‐making process that combines
environmental, ethical and social ideas. This process applies a set of investment checks to exclude assets based on corporate governance, ecological, ethical or social criteria. However a type of investment in stocks or companies that are involved with
controversial businesses, also know as vice investments.
One could say that the opposite of a SRI stock is a vice stock, that is involved with “vices” e.g. gambling, alcohol, tobacco, firearms and/or military sectors. Vice stocks are also considered by many to be sin stocks or socially irresponsible investments (Chong et al., 2006). In this paper a sample selection of returns of companies that are screened with controversial business involvements are used for the representation of the vice stocks. Further on in this paper I will sometimes use the abbreviation vice for the vice stocks.
Returning to the SRI stocks, since the 1960’s the modern view of social investing can be inferred by the political environment (Bauer et al., 2005) and seen as a result of the grown awareness of investors to the civil rights, environment and nuclear energy issues. Correspondingly, to meet the increasing demand for the application of ethical criteria in the investment process, specific mutual funds were designed with a corporate social responsibility (CSR) intention. In addition more and more companies are
introducing CSR related elements in their policies and strategies. Subsequently, the academic interest has risen as a result of the fast growth of the SRI industry (Renneboog et al., 2008). The increase in academic interest in the SRI industry and the fact that SRI is an actual subject nowadays makes this research relevant.
One of the most prominent datasets with quantitative measurements of SRI actions is the Kinder, Lydenberg, Domini Research & Analytics (KLD) Social Ratings (Mattingly and Berman, 2006). The dataset is the centrepiece of the data source of this research. In addition the indicator categories, which will be discussed later in this paper, social, governance, environment will be important as overarching screening themes for the SRI stocks used in the research. Furthermore indicator category for the vice stock used will be the overarching controversial business involvement screen. Further on I will describe the sample selection process of the screens.
This paper compares the returns of a ‘socially responsible stocks’ (SRI) and a ‘socially irresponsible stocks’ (vice) with the Russell 3000 index. Besides that, the purpose of this paper is to provide a relevant overview of the current situation of the literature on SRI and the counterpart the vice stocks, in order to show the main results and provide interesting questions for further research. The central question that arises, although the literature on SRI is still increasing rapidly, is whether there are significant differences in the financial performances between the SRI stocks and the vice stocks in comparison to the Russell 3000 index?
The remainder of the paper is organized as follows. Section 2, the theoretical framework, will provide an overview of the previous SRI research studies. Section 3, the methodological framework, gives a description about the data selection process and the regression research method used in this paper. Thereafter Section 4 reports the data analysis and main empirical findings. Finally Section 5 will give a relevant conclusion of this paper and the research question.
2. Theoretical framework
Chapter 2 will provide an overview of the previous SRI and vice investing research studies. In addition some background information about the Russell 3000 index will be discussed.
2.1 Social Responsible Investment (SRI)
There is a current discussion about the performance of socially responsible investing during the last thirty years. Past research suggests that the SRI findings unveil
insignificantly different results from conventional funds (Renneboog et al., 2011). Contemporary ethical investing is based on social cognisance and had rapidly increased over the past decades. Investors who invest in a SRI fund or stock have several social goals. These funds use screening criteria whereby specific sectors, e.g. tobacco, alcohol and military industries will be excluded. SRI funds will screen companies that meet the corporate governance, ecological, ethical or social criteria with positive results. The funds include only the companies that pass minimum threshold criteria in each industry in their portfolios.
The current increase of SRI indicates that investment world prefer to mix SRI’s comparable return with their fears on social responsibility (Jo et al., 2010). Social responsible investments had become part of the regular investment strategy, and this results in a number of conventional companies, that now offer SRI products to their clients.
In addition, it is a fact that mutual fund managers overall have other goals than fund investors. Investors want generally high-‐risk adjusted returns at low fees, while managers are interested in the money inflows and the corresponding management fees. In addition, most fund investors are almost never formally trained in portfolio analysis. This means that their investment decisions could be influenced by the media
considerations and the marketing activities of investment funds (Sirri and Tufano, 1998).
As a result excessive risk-‐taking by fund managers could be generated by the conflict between mutual fund managers and the funds’ investors, because fund
managers would like to generate strategies that strengthen their own revenues and this is not in the interest of the funds’ investors.
Furthermore most of the past SRI researches came up with overall the same results and conclusion. Bollen (2007) compares the flow-‐return relation of US SRI funds
to conventional US funds. The main finding was that US SRI fund flow is more sensitive to past positive returns, but less sensitive to past negative returns than conventional funds are. He assumed in his research that SRI funds are a homogenous group. Renneboog (2011) mentioned that the flow-‐return sensitivity of SRI funds in some situation is different in specific countries or regions. As a result of the awareness of SRI investors of the several types of social or ethical issues that can be involved by the value that differ across countries and cultures.
Renneboog (2011) has several main findings in his paper. First of all SRI investors from United Kingdom, United States, Europe, and Asia and the Pacific Rim region seem to care less about past negative returns than do investors in conventional funds. Secondly, he found that there is no relation between past average flows and future returns for both conventional funds and SRI funds. This means, that companies, which acquire more flows in the present, will neither outperform nor underperform in the future. While environmental screens negatively influence fund performance, there are several SRI characteristics that have a positive influence on future returns.
Furthermore he found that particular funds with lower fees or lower return, and smaller and younger funds do raise more inflows than do the older, bigger, riskier or more expensive funds. The conclusion of Renneboog (2011) his finding was that there is no relation between past average flows and future returns for conventional funds and SRI funds. High inflow funds neither outperform nor underperform in the future. This fact is in line with the efficient market hypothesis, which states that investors cannot predict fund returns, and with the fact that funds are confronted with decreasing returns to scale.
Beside, Bauer (2005) has used in his paper an international database of 103 ethical mutual fund and analyses ethical mutual fund performance and investment style. He applies the standard CAPM 1-‐factor model and the Carhart four-‐factor model, which controls for size, book-‐to-‐market and stock price momentum. In this paper I will also use the CAPM and Carhart four-‐factor model.
Bauer’s study has three main findings. His first finding was that there was not enough statistically significant evidence in difference in return between social
responsible and conventional mutual fund returns after controlling for main factors e.g. momentum, size and book-‐to-‐market. Second, socially responsible mutual funds have defined investment styles compared to conventional funds. E.g. social responsible funds
are normally less exposed to market return volatility compared to conventional funds. In addition, UK and German social responsible funds are extremely exposed to small caps. While social responsible funds in the US, on the other side, invest more in large caps compared to the conventional. Moreover, social responsible funds tend to be more orientated in growth, or less valued, so nothing in between.
Thereby Bauer (2005) found when evaluating the fund performance using some important ethical indices that SRI indices act worse than standard indices in explaining ethical mutual fund returns.
In addition, there are three key arguments against integration of SRI funds. The arguments have a direct link to the relation, how SRI funds are empirically measured (Scholtens, 2011). First of all, SRI portfolios have increasingly costs and risk due to diminished diversification (Geczy et al., 2005; Renneboog et al., 2006; Cortez et al., 2008). Therefore there is a suspicion of increased monitoring costs from SRI-‐managers (Bauer et al., 2007). Third, SRI may lead to decreased returns, leading financial
managers to a breach of their fiduciary duty to provide the highest possible return with the lowest possible risk (Schröder, 2004; Bauer et al., 2005).
To analyse the impact of these aspects, SRI studies apply various methods of risk and return analysis, calculated mostly from modern portfolio theory. For example, models as capital asset pricing models (CAPMs), multi-‐index models, multi-‐factor models and arbitrage pricing theory (Scholtens, 2011). In the past 50 years the
empirical literature of the SRI studies has developed a lot, but most of them still relies on conventional portfolio evaluation (Elton et al., 2006).
Most of the SRI studies intention is to create estimates of the average returns of a population of socially responsible investment funds with and estimation errors and low biases (e.g. Bauer et al., 2005). It is crucial in SRI research that one must take into account for measurement error and misspecification (Kennedy, 2008). As a result SRI fund’s empirical average returns must be consistent and efficient (Greene, 2008).
2.2 Vice Investing
Besides past SRI research, there is on the other hand, past vice research. One of the studies was made and reported in 2001 by former chief investment officer of Credit Suisse, Tom Galvin (Waxler, 2004). Galvin named the industries that are included in vice, the “vice squad”, as earlier mentioned activities you’re not supposed to do and activities that are ethical not right, e.g. firearms, alcohol, tobacco and military. According
to Galvin’s research, in a part of the previous recessions specifically in the ‘82 and ‘90-‐ ‘91 recessions, in both the tobacco and alcohol industry outperformed the market. The vice fund performance is at his potential maximum in the early parts of a recession (Waxler, 2004). The vice sectors are classified as a safe bet in all markets, aside the ethical aspect of the vice. As a result Galvin conclude that the demand for vice related products stays steady and generally increase during economic volatility
conditions. Galvin records that the vice stocks significantly outperformed the market over the 18 months before he publishes his report in March of 2001. In addition a vice fund are easy for the public to understand and largely recession-‐proof.
Moreover past studies point to a higher performance of vice stocks (Lobe and Roithmeier, 2008; Hong et al., 2009). The companies in a vice fund or, in the later on discussed randomly selected, vice stocks sample are associated with unethical or immoral activities. According to Brush (2003) vice stocks perform better compared to SRI stocks. As a result of that vice stocks performance is better during bad rather than good times, because people might drink, smoke or gamble a little more (Jo et al., 2010). Though Brush conclusion is not general accepted in the investment world. A matching example is the Social Investment Forum (SIF) (2010) that suggests that approximate 65% of the SRI mutual funds outperform their benchmark in 2009.
Another innovative study by Merton (1987) concludes in his research about the characteristics of neglected stocks that the reason of the increase of the expected returns of the socks is directly related to higher litigation risk of the companies in the fund. In his paper he also described and illustrated why neglected stocks – e.g. tobacco companies – are underpriced and better performance than comparable companies.
According to Anderson (2008) investors use the vice sector investments as appropriate strategy during periods of recession, because the vice sectors tend to perform well regardless of economic fluctuations, and sometimes outperform during inconstant times. As mentioned in the US news in 2008, “Consumers don't give up necessities like toothpaste and laundry detergent in tough economic times. And they don't kick their habits, either” (Marquardt, 2008).
Merton, Hong and Kacperczyk (2009) found that vice stocks outperform the market because they have less analyst coverage in comparison to non-‐vice stocks with comparable characteristics and exhibit less institutional ownership. Generally banks, pension funds and insurance companies seem to avoid these vice companies due to
social norm pressures. Despite the fact that socially responsible investments funds do share this behaviour, hedge funds and conventional mutual funds do not share this behaviour, as they are natural arbitrageurs in the market, so buy also unethical stock if they are under-‐priced.
2.3 Russell 3000 Index
The Russell 3000 Index is a widespread equity market index similar to the S&P 500 (Coggin et al., 1993). In addition the index represents 98% of the possible investments U.S. equity market (Madhavan, 2003). Furthermore the index measures the financial performance of the 3000 largest companies in the U.S.A. and is based on their total market capitalization. Besides the Russell 3000 their exist also the Russell 1000 an 2000. These indexes represent respectively the 1000 and 2000 largest companies. In June 2002 the median of companies’ market capitalization in the Russell 3000 was $700 million. Finally each year the Frank Russell Company reranks every company by their market capitalization in order to determine the new index.
2.4 Summary Chapter
In this recapitulatory section e.g. a few main findings of the past SRI studies will be mentioned. First of all Renneboog (2011) concludes in one of his findings that there is no relation between past average flows and future returns for conventional funds and SRI funds. Secondly Bauer’s study has one important finding that there was not enough statistically significant evidence in difference in return between social responsible and conventional mutual fund returns after controlling for main factors e.g. momentum, size and book-‐to-‐market. In conclusion one may say that according to past SRI studies it is not possible to conclude that there is an empirical evidence to invest in SRI stocks instead of conventional stocks.
Furthermore the overall past vice research concludes that the vice sectors are classified as a safe bet in all markets, aside the ethical aspect of the vice. During economic volatility conditions the demand for vice related products stays steady. In addition a vice fund or stock are easy for the public to understand and largely recession-‐ proof.
Finally the Russell 3000 index will be the representative for the market, because the index gives realistic view of the overall investment opportunities. In the models and regression analysis that will be described in the next chapters the Russell 3000 index will be the benchmark.
3. Methodological framework
Chapter three will give first of all a short description about the data selection. In
addition I will show the sample selection process of the SRI and vice companies, which I used for the daily stock returns. Further I will do a regression analysis for the CAPM, Fama French Three-‐Factor model and Carhart Four-‐Factor model for the SRI and vice stock returns. The benchmark in the above named models will be the Russell 3000 index. Lastly the results of BHRR and BHAR of the SRI and vice stock returns will be shown and will give a slightly different dimension to the study.
3.1 Data Selection
In this research the main data sources are The Center for Research in Security Prices (CRSP), MSCI ESG KLD STATS -‐ Social Ratings and Fama and French factors (1993). For the Fama-‐French factors I used the daily frequency, and will handle initially with the risk-‐free rate (rf). The stock returns of the sample selection of the SRI and vice stocks are both from CRSP and also at a daily frequency, because of the fact that it will show in this way a more detailed view of the returns. As a result of the earlier mentioned daily frequencies, the Russell 3000 index stock returns are likewise at a daily frequency and sourced from Yahoo Finance.
The time frame of the research sample is January 1st 2003 till December 31th 2012. The start date of January 2003 is chosen, because of the fact that there was a limited historical data available for SRI and vice stocks returns before this date. The end of 2012 is the last date of the sample, because of the fact that the access to more actual data in CRSP at that specific date was limited. In addition a few years for and after the crisis are taking into account in the time frame of ten years.
3.1.1. MSCI ESG KLD STATS -‐ Social Ratings
The norm for quantitative measurements of SRI actions is the Kinder, Lydenberg, Domini Research & Analytics (KLD) Social Ratings dataset these days (Mattingly and Berman, 2006). Furthermore the MSCI ESG KLD STATS dataset was created by KLD Research & Analytics, Inc. (KLD) in 1991. The criteria used for the dataset are the ESG criteria and those imply environmental, social and governance (ESG) criteria. In addition Morgan Stanley Capital International (MSCI) acquired KLD in 2010 (wrds, 2015).
The MSCI ESG KLD STATS dataset uses 13 measures that belong to four key ESG performance indicators categorized as follows: social, governance, environmental, and
controversial business involvement (MSCI, 2015). In the table below one will find that the screens in all the indicator categories social, governance and environmental could be positive or negative. In addition the controversial business involvement indicators will always be negative. In order to clarify the type “positive’, this means for example that a company is involved with green buildings, renewable energy or human right initiatives. The opposite, “negative”, means for example human rights violations. This implicates human rights concerns and that would be labelled as negative.
Indicator Categories Type Involvement Screens
Social Positive/Negative Community
Diversity
Employee Relations Human Rights Product
Governance Positive/Negative Corporate Governance
Environment Positive/Negative Environment
Controversial Business Involvement Negative Alcohol
Firearms Gambling Military
Nuclear Power Tobacco Table 1: MSCI ESG KLD STATS Involvement Screens (MSCI ESG KLD STATS, 2015)
In the next two subsections of this chapter I will show one the sample selection process and the use of the indicators and involvement screens.
3.1.2 SRI
The companies that are used for the SRI stock returns in this research are selected by their total rate of positive involvements as a company. As earlier mentioned the MSCI ESG KLD STATS -‐ Social Ratings is used in this random sample selection process. In addition the notion random means that the companies are not selected by name. First of all the companies with the highest total rate in January 1st 2003 are included in the sample, which means that the highest total rate was a rate of 4 and the second highest total rate was a rate of 3.
The companies that are rated with a rate of 4 are: DuPont Company and PG&E Corporation. Firstly DuPont Company, multinational chemicals and health care company headquartered in Wilmington (United States), has scored a high rate. This because of the fact that they scored positive with the diversity involvement screen by board of directors -‐ gender, product involvement screen by R&D innovation and with the environmental involvement screen by pollution prevention and clean energy. Secondly PG&E
Corporation, an energy-‐based holding company headquartered in San Francisco (United States), has also scored a rate of 4. This company has scored on positive involvements screens for community by innovative giving, diversity by with CEO and board of directors – gender and environmental by clean energy. The total number of companies included in the sample is 23 companies; this means that the other 21 companies are rated with a total rate of 3.
The last step of the process were to take the average returns of the 23 companies for each day in the sample time frame of ten years; January 1st 2003 till December 31th 2012. The so-‐called KLD returns (𝑟!"#), which are randomly selected, will be used in the next sections of this research.
Company Name Ticker Total Market Value 2003 ($millions)
3M Company MMM 66673,4685
AGL Resources Inc. ATG -‐
Ault Incorporated AULT -‐
Avon Products, Inc. AVP 15880,262
Dow Chemical Company DOW 38554,0549
DuPont Company DD 45765,3628
Ecolab Inc. ECL 7045,5033
Fannie Mae FNM -‐
Freddie Mac FRE -‐
Gaiam, Inc. GAIA 86,8879
General Electric Company GE 311755,4576
Hanmi Financial Corporation HAFC -‐
Hewlett-‐Packard Company HPQ 67883,9979
Interface, Inc. IFSIA 280,0025
International Business Machines Corporation IBM 157047,0941
Johnson & Johnson JNJ 153325,4852
Lucent Technologies, Inc. LU 9005,04
PG&E Corporation PCG 10905,4179
Rohm and Haas Company ROH 9500,9676
Sigma Designs, Inc. SIGM 163,2466
Spanish Broadcasting System, Inc. SBSA 682,5112
WGL Holdings, Inc. WGL 1340,719
Xerox Corporation XRX 10955,5992
Table 2: SRI companies in the sample
Table 2 shows the total market value of the SRI companies in the sample. In addition the data source Wharton Research Data Services has limited access to give a complete view of the market values.
3.1.3 Vice
Then the vice companies, which are randomly selected by multiple steps. Firstly I have looked to the controversial business involvement indicators in 2003, the start of timeframe of the sample. Thereafter in the sample there were 8 companies with a total rate of 2 took in to account: Kraft Foods, Inc., Altria Group,Inc., Playboy Enterprises Inc., The Pantry Inc., UST Inc., Alliant, Techsystems Inc., General Dynamics Corporation and Olin Corporation. In addition all the companies, which are involved with firearms, are included in the sample: Jarden Corporation and Sturm Ruger and Company Inc., Further Alliant Techsystems Inc., General Dynamics Corporation and Olin Corporation are already included in the sample, because those companies are already rated with a total
of 2. The last step in de process is the adding of all companies, that are not yet included, with a rate of 1 for tobacco involvements.
The last step of the process is similar to the SRI companies, the average returns of the 25 companies will be used for each day in the sample time frame of ten years; January 1st 2003 till December 31th 2012. The so-‐called vice returns (𝑟
!"#$), will be used in the next sections of this research.
Company Name Ticker Involvement(s) Total Market
Value 2003 ($millions)
Kraft Foods, Inc. KFT Alcohol/Tobacco -‐
Altria Group, Inc. MO Alcohol/Tobacco 110867,8525
Playboy Enterprises, Inc. PLA Gambling/Tobacco -‐
The Pantry, Inc. PTRY Alcohol/Tobacco 218,0384
UST Inc. UST Alcohol/Tobacco -‐
Alliant Techsystems Inc. ATK Firearms/Military -‐
General Dynamics Corporation GD Firearms/Military 17894,1467
Olin Corporation OLN Firearms/Military 1183,8409
Carolina Group CG Tobacco -‐
CNA Financial Corporation CNA Tobacco -‐
DIMON, Inc. DMN Tobacco -‐
Diamond Offshore Drilling, Inc. DO Tobacco 2652,3942
Eastman Chemical Company EMN Tobacco 3053,4158
Loews Corporation LTR Tobacco -‐
M&F Worldwide Corporation MFW Tobacco 245,6503
R.J. Reynolds Tobacco Holdings, Inc. RJR Tobacco -‐
7-‐Eleven, Inc. SE Tobacco -‐
Star Scientific, Inc. STSI Tobacco -‐
Standard Commercial Corporation STW Tobacco 253,7826
CNA Surety Corporation SUR Tobacco -‐
Schweitzer-‐Mauduit International, Inc. SWM Tobacco 440,8631
Universal Corporation UVV Tobacco 1293,2165
Vector Group, Ltd. VGR Tobacco 636,8227
Jarden Corporation JAH Firearms 738,3714
Sturm Ruger and Company, Inc. RGR Firearms 305,9781
Table 3: Vice companies in the sample
Table 3 shows the total market value of the vice companies in the sample. In addition the data source Wharton Research Data Services has limited access to give a complete view of the market values.
3.2 Capital Asset Pricing Model (CAPM)
The capital asset pricing model (CAPM) is a centre point of modern financial economics. The model determines a theoretically required rate of return of an asset. Included the assumption that the asset given is non-‐diversifiable and that the asset is added to an already well-‐diversified portfolio. Harry Markowitz was responsible for the foundation of the modern portfolio management in 1952. The years after the foundation William Sharpe, John Lintner and Jan Mossin developed in several researches the CAPM (Bodie et al., 2011).
The model shows the asset’s sensitivity to the market risk by the beta ( 𝛽!), which indicates the beta of the returns of SRI (𝛽!"#) or vice (𝛽!"#$) companies. Further the average returns (𝑟!") of the SRI (𝑟!"#) or vice (𝑟!"#$) companies are risk-‐free as shown in the CAPM model below. In addition the model shows the excess return 𝑟!!"""! − 𝑟!" , this implies the return of the market, Russell 3000 𝑟!!"""! , minus the Fama-‐Fench risk-‐ free rate 𝑟!" . Lastly the error term (𝜀!") is also included in the model.
𝑟!"− 𝑟!" = 𝛼!"+ 𝛽! 𝑟!!"""!− 𝑟!" + 𝜀!"
Subsequently the CAPM is tested with a regression of historical data of the stock returns on the market performance, with the Russell 3000 as the benchmark. The last element of the model, which still has to be mentioned, is Jensen’s Alpha or the abnormal return (𝛼!"). Jensen’s Alpha measures the out-‐ or under-‐ performance relative to the market proxy (Jensen, 1968). In addition the beta ( 𝛽!) is a measure of the risk from exposure to general market. If the beta is higher than 1, then that indicates an investment with higher volatility than the market (Bodie et al., 2011).
3.3 The Fama-‐French Three-‐Factor Model
The Fama-‐French three-‐factor model is an extensive version of the CAPM model. This is an alternative approach to explain macroeconomics factors as useful causes of
systematic risk uses firm characteristics (Fama and French, 1996). The model below will be sometimes abbreviated to FF3 in this research:
𝑟!" − 𝑟!" = 𝛼!"+ 𝛽! 𝑟!!"""!− 𝑟!" + 𝛽!"#$ 𝑆𝑀𝐵! + 𝛽!"#$ 𝐻𝑀𝐿! + 𝜀!"
The extension in this multifactor model is the adding of two variables: Small Minus Big (SMB) and High Minus Low (HML). Firstly SMB is the return of a portfolio of small stocks in excess of the return on a portfolio of large stocks. Secondly HML, i.e., the return of a portfolio of stocks with a high book-‐to-‐market ratio in excess of the return on a portfolio of stocks with a low book-‐to-‐ market ratio (Bodie et al., 2011).
3.4 Carhart Four-‐Factor Model
Mark Carhart added the momentum effect to the Fama-‐French three-‐factor model. The adding of this fourth factor to the model was to evaluate mutual fund performance (Carhart M.M., 1997). Furthermore Carhart discovered that the fluctuations of the alpha of many mutual funds could be construed to their sensitivities to market momentum. This model has become a common four-‐factor model used to evaluate abnormal performance of a stock portfolio nowadays:
𝑟!"− 𝑟!" = 𝛼!" + 𝛽! 𝑟!!"""! − 𝑟!" + 𝛽!"#$ 𝑆𝑀𝐵! + 𝛽!"#$ 𝐻𝑀𝐿! + 𝛽!"#" 𝑀𝑂𝑀! + 𝜀!"
3.5 BHRR & BHAR
In this section of the chapter the financial performances will be discussed in a different way than in the previous sections. The buy-‐and-‐hold raw return (BHRR) and buy-‐and-‐ hold abnormal return will be shortly explained. First of all the BHRR, which is shown in the formula below, is the average return of a stock in a particular time frame ( 𝑟! ): 𝐵𝐻𝑅𝑅 = 1 + 𝑟! − 1 = 𝑟! ! !!!
N is related to the time in the time frame and (𝑟!) is related to the return of the stock.
Furthermore the BHAR; the abnormal return due an event is estimated by the difference between the stock’s actual return and his benchmark. In this research the benchmark used is the Russell 3000 index. Abnormal return, also known as firm-‐specific return, may be interpreted as the unexpected return that results from an event (Bodie et al., 2011). To obtain a more accurate result the BHAR is used for the time frame:
𝐵𝐻𝐴𝑅 = 1 + 𝐴𝑅! − 1 = !
!!!
𝑟!
𝑤ℎ𝑒𝑟𝑒 𝐴𝑅! = 𝑟!− 𝑟!!"""
In case of using the BHAR formula the market will be represented by the returns of the Russell 3000 index (𝑟!!"""). The stock return, e.g. a specific portfolio, will be deducted by the market return.
3.6 Research Restrictions
3.6.1 Selection Process
In this paper there were a few restrictions. The most important one is the restrictions in sample selection process. Unfortunately the sample of SRI and vice was restricted, because it is was not possible to have a larger observation number than already used. In addition there were limitations in the data available in the time frame of ten years. This means that only the overall highly rated firms, firearms and tobacco in 2003 are used for the vice sample. Beside that the SRI sample included only the companies rated with a total KLD score of 3 and 4.
3.6.2 Regression Model
In the regression analysis the ordinary least squares (OLS) regressions are used. If one will replace in future research the OLS with the generalized least squares (GLS)
regressions, than the correlation across residuals will take into account (Bodie et al., 2011). Moreover the used regression models do not take into account time-‐varying volatility. The econometric technique of Robert Engle, Nobel Prize winner, are dealing with the time-‐varying and also useful for future research.
Further Black (1993) describes that when researchers scan and rescan a database of security returns, e.g. Fama-‐French security returns, in their search to explanatory factors they may find past patterns that are partly due coincidence. This process is known as data-‐snooping. However, the Fama-‐French model have shown that the SMB and HML variables have predicted average returns in various time periods and in markets all over the world, thus softening potential effects of data snooping.
In conclusion the research recommendations for future research would be if one will taken into account the above mentioned restrictions, than the results will be even more accurate.
4. Data results
4.1 Regression Analysis
In this section the results of the regressions will be interpreted and described. The financial performance of the SRI and vice sample of companies will be described through the CAPM, Fama-‐French three-‐factor model and Carhart four-‐factor model. Table 4 shows the result of all the used regression models in this research.
(Jan 1st 2003 – Dec 31st 2012)
(1) (2) (3) (4) (5) (6)
CAPM CAPM Fama-‐French Fama-‐French Carhart Carhart
VARIABLES SRI VICE SRI VICE SRI VICE
Beta 0.98969*** 0.90509*** 0.93300*** 0.83494*** 0.90042*** 0.83073*** (0.01157) (0.00862) (0.01260) (0.00892) (0.01293) (0.00929) SMB 0.24874*** 0.32515*** 0.27563*** 0.32863*** (0.02588) (0.01832) (0.02566) (0.01844) HML 0.19213*** 0.22824*** 0.08617*** 0.21455*** (0.02772) (0.01962) (0.02976) (0.02138) MOM -‐0.15693*** -‐0.02027 (0.01757) (0.01262) Jensen’s Alpha 0.00012 0.00051*** 0.00008 0.00045*** 0.00009 0.00045*** (0.00015) (0.00012) (0.00015) (0.00011) (0.00015) (0.00011) Observations 2,517 2,517 2,517 2,517 2,517 2,517 Adj R-‐squared 0.74408 0.81419 0.75628 0.84027 0.76363 0.84032 Standard errors in parentheses
*** p<0.01, ** p<0.05, * p<0.1 Table 4: Results regression models
The above table shows in every column the used regression models and on the left-‐hand side the Excess return, SMB, HML, MOM, Jensen’s Alpha and additional information.
The total number of observation in this research is 2517, as mentioned earlier the total observation have been extracted from the time frame used of January 1st 2003 till December 31st 2012.
It is remarkable that in the CAPM, Fama-‐French three-‐factor model and Carhart four-‐factor model the positive Jensen’s alpha is not significant in the used time frame and sample. While Jensen’s alpha for vice is positive and statistically significant from 0 with a significance level of 1%, which implies abnormal returns in all the used
the two Fama-‐French factors SMB and HML in the model, than the alpha of both SRI and vice decreases a little bit, because of the increase in risk factors in the model. Further both alphas increases when adding the momentum variable. In addition in both ways the Fama-‐French three-‐factor and Carhart four-‐factor model’s alpha is significant, which implies abnormal returns in either way.
Further if one add risk factor variables in the model, e.g. SMB, HML and MOM. Than the stock returns could be better explained. That is the reason Jensen’s alpha decreases a little in the Fama-‐French three-‐factor and Carhart four-‐factor model. Table 4 reports that the excess return or beta coefficient of the SRI stocks is 0.98969 and statistically significant in the CAPM. In fact the SRI stock returns flows approximately with the market. The vice beta of 0.90509 and is significant. In case of adding the two Fama-‐French factors and the momentum into the model the SRI and vice stock beta will decrease a little, but not extremely. The effect on the market return decreases a little, because of the adding of the SMB, HML and MOM variables. In this case the excess return will be more explained through these variables. In addition in case of both models’ beta is significant.
The table finally shows that the adjusted R-‐squared. The adjusted R-‐squared compares the descriptive power of the regression model. In both SRI and vice stock the adjusted R-‐squared increased when adding a variable into the model. This implicates that the Carhart four-‐factor model has a higher descriptive power, than the CAPM.
4.2 BHRR & BHAR Analysis
The last section of the data analysis is about the buy-‐and-‐hold rate for raw and abnormal returns. In the table below one could find the results:
BHRR BHAR
Russell 3000 0.729 -‐
SRI 1.192 2.926
Vice 4.854 9.489
Table 4: Results BHHR & BHAR
The above table shows that the 𝐵𝐻𝑅𝑅!"# of 1.192 is closer to the 𝐵𝐻𝑅𝑅!"##$%%!""", than the 𝐵𝐻𝑅𝑅!"#$. One could conclude that the vice stocks are more attractive to invest for risk searching investors. If an investor is more risk averse, than the SRI stocks are more attractive for them. The BHAR results are confirming the previously mentioned
5. Conclusion
This paper contributes to the research of socially responsible investing and socially irresponsible investing. The paper is a study on the differences between the financial performances of SRI and controversial business involvement screened companies, in comparison with the Russell 3000 index. The companies are screened by KLD Social Ratings through ESG criteria: environmental, social and governance. In addition the added criteria is the controversial business involvement, this screen determines in this research the vice companies used in the regression.
According to Brush (2003) vice stocks perform better compared to SRI stocks. In addition Galvin concludes that the demand for vice related products stays steady and generally increase during economic volatility conditions (Waxler, 2004). However past research suggests that the SRI findings unveil insignificantly different results from conventional funds (Renneboog et al., 2011). In addition the current increase of SRI indicates that investment world prefer to mix SRI’s comparable return with their fears on social responsibility (Jo et al., 2010). This results in the fact that social responsible investments had become part of the regular investment strategy.
As a result of the regression analysis in this paper the Jensen’s alpha of the SRI stocks is in all regression models not statistically significant. However in case of the vice stocks the alpha is significant with a significant level of 1% in all models. In addition the alphas are positive and this implies abnormal returns. This could imply that investor should invest in vice stocks instead of SRI stocks, aside the ethical aspect of the vice. In addition past studies also point to a higher performance of vice stocks (Lobe and Roithmeier, 2008; Hong et al., 2009).
Furthermore the data results show that the betas in all regression models are statistically significant and tend to move in either SRI and vice case as the market. The only remarkable difference is that the vice beta is in comparison with the SRI beta lower. Finally one could concluded on base of the BHRR and BHAR rates that the vice stocks are more attractive to invest for risk searching investors and SRI stocks for risk averse investors with specific ethical preferences.
In conclusion with the used sample, regression models and in the time frame of January 1st and December 31st 2012 there is a significant difference in abnormal returns between SRI and vice stocks in comparison to the Russell 300 index. However there will be always an ethical dilemma which stock is better to invest in. It depends on the
financial and non-‐financial preferences of an investor which one to choose. As shown in the literature overview and the data analysis, SRI investments are predictable
investments, because they overall tend to move as the market. On the other hand vice investing is a safe investment in all markets, aside the ethical aspect of the vice, because these kinds of stocks are barely sensitive to market fluctuations.
