ESG Based investing
“The effect of environment, social and governance screens on U.S. SRI mutual fund
performance.”
Abstract
This study examines the effect of Environmental, Social and Governance screens on SRI mutual fund performance in the United States. The performances of SRI mutual funds are examined through the CAPM, Fama and French three-factor and four-factor regression models in a period between 2015 and
2019 using monthly data. Concluding, this study finds a significant difference in the effect of ESG screens on SRI mutual fund performance. The environment screen has a significant positive effect on
SRI fund performance, whereas the social and governance screens have an insignificant negative effect.
Bess E. van den Brink (11575123)
University of Amsterdam - Amsterdam Business School
BSc Economics & Business Economics - Specialization Finance Number of credits: 6
Supervisor: Drs. P.V. Trietsch, MPhil. June 2020
Statement of Originality
This document is written by Bess van den Brink 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.
Table of content
1. Introduction ... 1
2. Literature review ... 2
2.1 – Socially responsible investing ... 2
2.2 – Costs and benefits of socially responsible investing ... 3
2.3 – ESG investment screen ... 3
2.4 – Evaluating SRI Fund performance ... 5
2.4.1 Data quality ... 5
2.4.2 Social responsibility verification ... 5
2.4.3 Survivorship bias ... 6
2.4.4 Benchmarks ... 6
2.4.5 Sensitivity and robustness analysis ... 6
2.5 – Empirical evidence ... 7
2.5.1 The effect of ESG ratings on SRI fund performance ... 7
2.5.2 The effect of environmental performance on SRI fund performance ... 7
2.5.3. The effect of social performance on SRI fund performance ... 8
2.5.4 The effect of governance performance on SRI fund performance ... 9
2.6 – Hypotheses ... 10
3. Methodology ... 11
3.1 CAPM ... 11
3.2 Fama and French factor models ... 12
3.3 Cross-sectional regression ... 13
4. Sample and Data ... 14
4.1 Sample ... 14
4.2 Data ... 15
5. Empirical results ... 17
5.1 SRI Fund performance ... 17
5.2 Screen performance ... 18
5.2.1 ESG score ... 18
5.2.2 Environment, social and governance scores ... 19
5.2.3 Scores over time ... 20
6. Conclusion ... 22
6.1 Summary and discussion ... 22
6.2 Limitations ... 23
6.3 Implications and recommendations ... 24
7. References ... 25
1. Introduction
In the past decades, the interest in socially responsible investing has increased. In the United States, the amount invested in socially responsible investment funds has grown from 2.71 trillion dollars in 2010 to 8.72 trillion dollars in 2016 (USSIF, 2016). These numbers present 11% and 21.6%
respectively of the total assets under management in the United States. The growth of SRI funds is caused by the increased demand from institutional investors, the pressure from society, the media and non-governmental organizations and changes in regulations (Olmedo, 2010). This denotes the
importance of understanding socially responsible investments.
Socially responsible investing chooses stocks not only based on return performance but also on certain criteria, called screens. The most common rating to screen socially responsible mutual funds (“SRI funds”) is the environment, social and governance (“ESG”) rating. Mainly, the type of screen can be categorized within the environment, social or governance screens. Most papers focused on the potential difference between SRI funds and conventional funds performances and neglected the screening process. Conversely, papers that did include the type of screen have not found a consensus between the type of screen and SRI fund performance. Derwall et al. (2005) found that a portfolio with relatively high environment scores outperforms the portfolio with lower environmental scores. This is in line with the results from previous research which tested whether high-ranked ESG companies are related to higher abnormal returns (Kempf and Osthoff, 2007). However, this result is disproved by a study from Halbritter and Dorfleitner (2015), who states that the abnormal returns by high ESG ratings are insignificant. Therefore, there is no consistent evidence on the effect of environmental, social and governance screens on SRI mutual fund performance. Furthermore, there are studies on the effect of ESG screen ratings on firm value, but not on fund performance (Friede et al, 2015). This leads to the following research question for this study: “Do the Environment, Social and Governance screens of ESG-ratings affect socially responsible investment funds’ risk-adjusted returns?”.
This study will contribute to the existing literature as it tries to help to find the consensus in the relation between the ESG screens and the SRI fund performance. To answer the research question, a sample of 24 funds is retrieved from the United States Forum for Responsible and Sustainable Investment. The sample includes U.S. SRI mutual equity funds which exist for more than five years since the dataset starts at 2015 and ends at 2019. Furthermore, a sample from the CRSP database will be assessed to calculate the ESG scores for each SRI fund.
In this paper, the Jensen’s alpha will be estimated using the CAPM, Fama and French three-factor and Fama and French four-factor models. The Jensen’s alpha is used to compare the risk-adjusted returns of the SRI funds with their benchmark, which is the weighted average of the NYSE, NASDAQ and
Thereafter, the average is taken from the sample to look at the overall performance of SRI funds. Finally, the effect of each screen of the ESG score on fund performance is evaluated. This will be done estimating the relation between the risk-adjusted return and the environment, social and governance scores of SRI funds through a cross-sectional regression model.
According to McWilliams and Siegel (2006), socially responsible investment leads to a better
reputation which increases the risk-adjusted return. Therefore, the expected result is that ESG screens have a positive effect on the risk-adjusted return. This indicates that a high ESG screen rated SRI fund outperforms SRI funds with lower ESG screen ratings. To examine whether this is correct the
remainder of this paper is as follows. First, the theoretical framework goes deeper into what socially responsible investment entails, what screens can be used, how fund performance is measured and how ESG ratings have affected financial performance in previous research. Followed by a description of the methodology. Furthermore, the fourth section presents data collection and sample selection. In the fifth section, the results of the analysis are presented. Finally, the conclusion, limitations and
suggestions for further research on this topic are presented.
2. Literature review
This chapter is about the definitions and empirical findings of the effect of environment, social and governance screens on SRI fund performance. Furthermore, it reviews and discusses the specific impact of each screen of the ESG rating and this will help in presenting the existing gap in research and conducting the hypothesis for this research.
2.1 – Socially responsible investing
There is not one single definition to define socially responsible investing (SRI). The origin of SRI consisted of the term ethical investment with moral screening. The first mutual fund with SRI implementations was the ‘Pioneer Fund’, founded in 1928 and consisted of employing screens (Renneboog et al., 2008). After that the number of SRI funds increased which is caused by the increased demand from institutional investors, the pressure from society, the media and non-governmental organizations and changes in regulations (Olmedo, 2010).
There are three widely used definitions of SRI. The first definition is from Renneboog et al. (2008), who define SRI as an investment strategy where investors apply different types of screens before investing (Renneboog et al., 2008). Secondly, according to the Social Investment Forum (2001) SRI is defined as an investment that takes into account the social and environmental influences of
investments, regardless of whether it is positive or negative. Furthermore, Syed (2017) adds to this definition that the exclusion of certain type of industries is based on certain moral or ethical grounds. Moreover, mutual funds without socially responsible components are considered as conventional funds (Chegut et al., 2011).
2.2 – Costs and benefits of socially responsible investing
SRI comes with different costs and benefits. The two main advantages of SRI are the reputation benefits and the drive of personal values and goals (Bollen 2006; Godfrey 2005; Wang, Choi and Li, 2008). The disadvantage of SRI is additional costs created by the increased reputation (Palmer, Oates and Portey 1995).
The advantages are explained as follows. The drive of personal values and goals are based on a study by Bollen (2006), who suggests that taking social responsibility is an important aspect of economic decision making. This indicates that people will accept a lower rate of return when they make socially responsible investment decisions (Renneboog et al., 2008). According to McWilliams and Siegel (2006), a positive reputation results in a positive economic value. The second argument is about a potential increase in reputation. A high reputation increases the commitment of other stakeholders (Godfrey 2005; Wang, Choi and Li, 2008). Which creates more willingness to provide resources and capital to a firm and it increases employee satisfaction which will lead to a higher willingness to work for the company for a longer period (Rindova and Fombrun, 1991). Rennenboog et al. (2008) add to this that an increase in social responsibility leads to retention of motivated workers.
However, SRI comes also with costs. The first cost is stated in the traditional neoclassical approach. This theory argues that besides a positive reputation affect the SRI also creates additional costs for a firm (Palmer, Oates and Portey 1995). In the short-term additional costs will lead to lower profits in a competitive market. However, in the long-term, this can affect the competitiveness negatively, which influences the cashflows. Friedman (1970) argues in the shareholder theory that the only social responsibility a company should have is the maximization of its shareholders’ value. Thus, a reduction in profits is not in line with the theory. Nevertheless, there is still no consensus found in the relation between SRI and financial performance. This is due to the advantages and disadvantages of SRI that offset each other. The benefits and costs can’t be linked to different screens since all screens can result in an increase in reputation and can lead to additional costs.
2.3 – ESG investment screen
There are multiple types of investment screens used in the screening strategy of SRI funds. The ESG screen is the most widely used. The ESG screen is built upon three categories., the environment screen, social screen and governance screen. The environment screen measures the impact of a company on the living and non-living natural systems. In many research papers, the focus lies on eco-efficiency and the contribution to climate change. The environment factor has been discussed
thoroughly form a corporate, but also from an investor perspective (Dunphy et al., 2000; Kinder, 2005). Furthermore, the social screen mainly focusses on the employee-satisfaction, also known as the
theory, which states that employee satisfaction can increase motivation and retention to the benefit of shareholders (Maslow 1943), leads to the possibility of undervaluation of employee satisfaction since the theory has no clear predictions whether it is desirable for a firm or not. The governance screen consists of a set of principles and rules defining the rights, expectations and responsibilities of all stakeholders of a company. A well-organized corporate governance system can contribute to aligned interest between stakeholders. There are two types of corporate governance, internal- and external governance. Internal governance focuses on the board and specifically on the size of the board, the independence of the board and the level of debt financing used by the board (Huijgevoort, 2017). Studies on external governance are more focused on corporate control and ownership blocking (Reddy et al., 2008).
The ESG screen is mostly implemented into an investment strategy by an engagement strategy or by either positive or negative screens. When SRI was still relatively unknown, funds mostly selected stocks or industries looking at negative screens. Negative screens are used when a fund excludes certain stocks or industries based on the specific nature of their business and the compliance with specific ESG criteria. It’s the oldest and most basic screening method used. Another way of implementing ESG into an investment policy is by using a positive screen. A positive screen only includes assets that comply with specific ESG standards (Renneboog et al., 2007). When the businesses are selected on the existence of positive or negative screening, the selection process continues by a quantitative and financial selection. However, the engagement strategy works the other way around. A fund decides whether it wants to increase or decrease the ESG rating and then with which portfolio they can achieve this. This research is focused on ESG integration of the SRI strategy (Renneboog et al., 2008).
The main goal of ESG screening is positively or at least not negatively influencing the environment, social and governance (Syed, 2017). Different analysts have stated in a survey that they implement ESG screening in their mainstream investment process (Bourghelle et al, 2009; Nofsinger &Varma, 2014). ESG rating agencies are not transparent and there is a lack of publicly available information, which makes it difficult to make a comparison between companies (Olmedo, 2010). The evaluation systems for ESG ratings are based on the adaption of the stakeholder model. Wherein, the needs and interests of stakeholders are taken into account when evaluating firms ESG performance (Fassin, 2009; Podnar and Jancic, 2006). However, the selection criteria are not the same for all ESG rating agencies. Some evaluation criteria are widely used. The most commonly used evaluation criterium is the environmental screen (Huijgevoort, 2017).
2.4 – Evaluating SRI Fund performance
In the research of Chegut et al. (2011), they state that the five most important factors that should be taken into account when measuring SRI fund performance are data quality, social responsibility verification, survivorship bias, robustness and benchmarking & sensitivity. This research forms the basis of the literature review on the evaluation of SRI fund performance. In the following subsections, the five factors will be discussed.
2.4.1 Data quality
The quality of data is determined by the chosen measure for fund performance. The possible
measurement instruments are the risk-adjusted return and the Sharpe ratio. The most commonly used risk-adjusted return measure is Jensen’s alpha (Barnett and Salomon, 2006). This measurement is calculated by using regression models and isolating the alpha. The second most used measurement is the Sharpe ratio. This measurement allows for a good comparison of excess returns relative to their risk. However, this measure is only relevant in relative nature (Renneboog et al., 2008). Since this paper focuses on direct comparison the main focus lies on the Jensen’s alpha.
The estimated returns in previous studies do not take into account the same drivers. Only half of the studies provide a specific description of the drivers of their returns. Therefore, dividend yield and fees have not been consistently included in previous research on SRI fund performance. Furthermore, load fees are not considered in many studies. This is partly because it is a complicated concept to include. Renneboog et al. (2008) approach the fees differently country-wise since they state that fees differ per country. Conversely, in the study of Gil-Bazo et al. (2010), it is specified that management fees and load fees do not significantly affect the performance of U.S. SRI funds. In conclusion, the Jensen’s alpha is a better measure since it takes into account more factors than only the excess return per unit of risk.
2.4.2 Social responsibility verification
Verification contains two possible procedures, verification by a third party or by independent verification by the author(s). The most universally used is the verification by a third-party source. A third-party source is also called an ESG-rating agency and places a flag into the dataset of a fund, which implies that this fund is a verified SRI fund. However, independent verification by the authors(s) happens differently. The author reviews information about the fund, reads reports and performs interviews with individual fund managers (Chegut et al., 2011). Renneboog et al. (2008) describe that the weight on both independent author verification and third-party verification should be utilized to verify an SRI fund. There is a current discussion about whether different third-party sources are agreeing upon the term socially responsible investing since they might have different views about the SRI concept (Scholtens, 2005).
2.4.3 Survivorship bias
Survivorship bias is an important subject of interest in research on SRI fund performance. Survivorship bias looks at the influence of the exclusion of “dead funds”. Not all available data incorporate ‘dead funds’ and nonetheless survivorship bias is not universally recognized. There are four ways to deal with survivorship bias. First, not taking into account the survivorship bias while suggesting that the bias is present in the research. Second, not treating survivorship bias at all. Third, stating that there is a bias consisted of the database. Fourth, determining the bias from objective SRI information and experience. Nowadays, several studies consider the survivorship bias as a limitation of their research or consider it in their treatment.
2.4.4 Benchmarks
Benchmarks are essential in the evaluation of SRI fund performance (Grinblatt and Titman, 1994). The three most widely used benchmarks for SRI fund performance are sustainable indices,
conventional indices and matched pair analyses. Several studies use multiple benchmarks to get a better indication of fund performance (Chegut et al., 2011). The problem with the use of conventional indices is that these indices also include the irresponsible companies which differ from funds that only invest in responsible companies (Luther and Luther et al., 1992; Matatko, 1994). In the case when SRI benchmarks are not available, the most widely used benchmark is the matched-paired analysis. The main advantage of using matched-pair analysis is that the match can be based on specific investment characteristics (Chegut et al., 2011). Besides these indices, there is evidence that the standard equity indices would be a more sufficient benchmark for measuring the SRI fund performance (Bauer et al., 2006, 2005). However, the rise of SRI benchmark indices started small but ended up in being one of the largest indices used for evaluating SRI fund performance.
2.4.5 Sensitivity and robustness analysis
A sensitivity and robustness analysis is used for distinguishing funds’ extent of SRI strategy
incorporation, composition and the influence of management on the strategy determination. The extent of SRI strategy incorporation is widely discussed since style, type and number of screens, influence the return of SRI funds. The composition of funds influences some important characteristic of funds (Chegut et al, 2011). Furthermore, the influence of management is based on management skills. The most important management skill that influences the funds’ performance is market timing (Bollen and Busse, 2001). SRI fund managers benefit from the skill market timing ability, which does not differ from conventional fund managers (Chegut et al, 2011).
2.5 – Empirical evidence
This section will present the empirical results found in previous research. First, the effect of ESG ratings on fund performance is discussed followed by the effect of the environment, social and government screens. Research about the effect of each screen separately on corporate financial performance has been researched, however, there is no research on the effect of all ESG screens on SRI fund performance. In past research, the central question was mostly whether ESG ratings affect the firm value or corporate financial performance. In this research, it’s not the firm value that is evaluated, but SRI fund performance. To estimate the possible effects on SRI fund performance, the existing relations of ESG ratings and corporate financial performance are taken into account and tested whether this is consistent with SRI fund performance. Research about the effect of each screen of ESG performance on corporate financial performance has been researched thoroughly.
2.5.1 The effect of ESG ratings on SRI fund performance
There are negative and positive relations found in research on the effect of ESG ratings on SRI fund performance. The ESG effect over time would expect to show a decreasing Jensen’s alpha and a reduction of correlations. This is due to the learning effect, learnings in the capital markets will increase due to increased awareness. However, the picture remains the same since the 1990s and various trend test fail to notice a change of the correlation factors due to the time-factor (Friede et al, 2015). Furthermore, Chetterjee et al (2018) performed research on the effect of ESG scores on mutual fund performance of 73 U.S. mutual funds. In conclusion, they stated that lower and mid rated funds perform better than high rated funds, both in risk-adjusted return and absolute returns. However, Kempf and Osthoff (2007) find that high-ranked ESG companies are related to high abnormal returns. This statement is strengthened by Statman and Gluskhov (2009) for the period from 1992 to 2007. However, multiple studies also found different results. In a study from Halbritter and Dorfleitner (2015) is found that the abnormal returns by high ESG ratings are insignificant. This is further researched by Auer and Schuhmacher (2016) who coincide with their conclusion. There’s still no consensus found on the relationship between ESG ratings and corporate financial performance. This is due to that the use of different screening strategies leads to different performance measures (Nofsinger
et al., 2014).
2.5.2 The effect of environmental performance on SRI fund performance
Studies state that there is a significant relation between environmental performance and SRI fund performance. However, the results do not coincide with each other. Friede et al (2015) find in all categories of the ESG rating a positive relation with the corporate financial performance. The
environmental studies show the most favourable relation. According to McGuire et al. (1988), this can be a result of multiple arguments. First, there is a trade-off between environmental and economic
performance which is in line with the neoclassical approach explained above. Second, the costs to enhance environmental performance are not substantial and are therefore able to increase other managerial benefits like an increase in productivity or reputation strengthening.
This is based on the theory of Porter and van der Linde (1995), who argue that environmental improvement does not have to be expensive for a company, but this is only possible when a newly introduced regulation is well managed. It is possible to off-set the costs for increasing the
environmental performance by a reduction of other costs and could even lead to an increase in
revenues (McGuire et al., 1998). Klassen and McLaughlin (1996) strengthen this statement by arguing that stocks react asymmetrically to environmental news. They find that a decrease in price when there is bad news about the environmental development of a company is larger than the increase in stock price due to exposure to environmental information. But besides these theories, the relation between environmental improvement and corporate financial performance is still widely criticized.
McWilliams et al. (1999) criticize the relation because they find that the studies focus on short-term corporate financial performance. Which leads to a limited timeframe and highly sensitive results due to design issues (McWilliams and Siegel, 1997). Derwall et al. (2005) show that the positive relation between environmental and corporate financial performance has strengthened over time. A portfolio with a relatively high environmental score outperforms the portfolio with lower environmental scores in the period 1997-2003. The outperformance is due to undervaluation of available environmental information and the possibility of environmental premium capturing missing risk factors in their asset pricing models.
2.5.3. The effect of social performance on SRI fund performance
In studies about the social screen, the focus lies on the effect of human resources management on corporate financial performance, which is the ability of a company to create trust a loyalty within its work environment. Huselid (1995) finds in his research that there is a consensus in this relation since a well-organized human resource department can have a significant and positive effect on corporate financial performance. Jackson and Schuler (1995) confirm this with their research by stating that it can create a competitive advantage for a company.
Moreover, there are three theories to explain the previously mentioned conclusions. First, the ‘value-in-diversity’ theory states that a more diversified workforce improve the business performance and it offers a direct investment return, with greater profits and earnings. Second, the ‘diversity-as-process-loss’ theory is rather sceptical on the potential benefits of diversity. It states that diversity can also lead to counterproductivity. Third, a paradox theory explains that greater diversity is related to better business performance, but also more group conflicts. This seems negative, but in contrast, it shows that more conflicts force groups to look further than the easy common solutions. Contrary,
homogeneous groups lead to fewer conflicts but also to higher group unity with less innovation and flexibility (Herring, 2009).
Edams (2011) adds to these theories that employee satisfaction is difficult to price and are seen as intangibles. These intangibles result in a more reliable positive relation in the long-run than in the short-run and this would mean that the learning effect, as mentioned above, is less likely. And therefore, the expected relation of social performance on SRI fund performance is less probable to be positive.
2.5.4 The effect of governance performance on SRI fund performance
In most researches, the focus is either on external of internal governance. The studies that focused on internal governance studies considered board independence, level of debt financing and size of the board. Evidence suggests that independent boards are more likely to take steps that are in favour of shareholders and that these steps taken, result in a positive effect on earnings per share (MacAvoy and Millstein, 1999). Furthermore, Yermack (1995) finds in his research that a small board work more efficient than large boards, which leads to a negative relation between the size of a board and corporate financial performance.
Empirical studies focusing on external governance show a wide variety of outcomes. A study on the effect of corporate control and block ownership does not provide a clear relationship with firm value (Reddy et al., 2008). Additionally, corporate governance is highly affected by the country in which a firm is located. Therefore, many researchers are focusing either on the variation of governance between companies in different countries or differences among countries (Huijgevoort, 2017). Daszyńska-Żygadło (2016) argue that corporate governance does not only differ among countries but also within sectors. In conclusion, a consensus is also hard to find for the sub-factor governance since it depends on the industry and country.
Table 1: overview empirical findings
Summarizes the empirical findings of studies on the ESG screens
Paper Subject Region Findings
Huijgevoort, 2017 ESG score Europe The governance score affects Tobin’s q positively with 0.08%. The environment and social scores decrease Tobin’s q with 0.02% and 0.01% respectively.
Derwall et al., 2011 Environmental performance
North America
Tobin’s q is larger for firms with deemed eco-efficiency relative to lower eco-eco-efficiency firms. There is a positive relation between eco-efficiency and financial performance.
Klassen and McLaughin, 1996 Environmental performance North America
Cumulative Abnormal Returns react asymmetrically to environmental news. Positive news increases CAR with 0.68% and Negative news decreases CAR with 0.82%.
Huselid, 1995 Social
performance
North America
Well-organized human resource department increases Returns on Assets significantly by 1.1%.
. Edams, 2011 Social performance North America and Europe
The risk-adjusted return from the Fama and French five-factor model is positively affected by 0.56% with a higher ESG rating in the U.S and Europe.
Daszyńska-Żygadło, 2016
Governance performance
Europe High governance performance is not necessarily related to higher financial performance. Some sectors have a positive relation on Tobin’s q while other industries don’t.
2.6 – Hypotheses
Based on the literature, this section derives the hypotheses. The hypotheses in this research paper intend to analyze the effect of different screens of the ESG score on risk-adjusted-performance. Some argue that SRI fund differs in terms of risk-adjusted return significantly from conventional funds. Therefore, it is tested whether SRI funds with different ESG scores differ from each other. The following hypothesis can be derived from this statement:
H1: The ESG rating has a positive effect on SRI fund performance.
This hypothesis is testing whether the total ESG score affects SRI fund performance. However, the ESG score is based on the Environment, Social and Governance screens. To test whether these individual scores have a different effect on SRI fund performance the following hypotheses are adopted:
H2: The Environment score of the ESG rating has a positive effect on SRI fund performance H3: The Social score of the ESG rating has a positive effect on SRI fund performance. H4: The Governance score of the ESG rating has a positive effect on SRI fund performance.
These hypotheses are relevant since many papers suggest that there is a significant relation between environmental, social and governmental performance and corporate financial performance. However, there is no clear answer whether the environment, social and governance scores from the ESG rating are the drivers behind this relation. The empirical evidence on environmental performance shows relative to the social and governance screen the most positive relations with financial performance.
Therefore, hypothesis two is expected to have the largest and significant result in relation with SRI fund performance relative to hypotheses three and four. Hypotheses three and four are both expected to end up with a significant and positive effect since in previous research positive relations are found with SRI fund performance. However, there are also relatively little negative relations found which indicates that there might a negative effect of the social and governance score on SRI fund
performance.
3. Methodology
This section will describe the methodology used in this study. The measure used in this paper is the cumulative abnormal returns of the SRI fund compared to their benchmark. To estimate this measure about the performance of US equity SRI fund, the average Jensen’s alpha will be calculated for each SRI fund. Jensen’s alpha is a risk-adjusted performance measure which represents the abnormal return of a fund over the theoretical expected return (Jensen, 1967). In other studies, Jensen’s alpha is currently called the Risk-adjusted performance (“RAP”), since it takes the particular fund risk into account. This measure indicates whether an SRI fund has a significant different return than expected by its riskiness. The CAPM, Fama and French three-factor and four-factor model will be used to estimate Jensen’s alpha. Followed by regressions to test whether the Jensen’s alpha is affected by ESG, environment, social and governance scores and whether this differs between the Jensen’s alpha obtained from the CAPM and Fama and French three- and four-factor model (Henke, 2016).
3.1 CAPM
The first index model that is used the Capital Asset Pricing Model (CAPM). This is a single index model with the following set-up:
Where:
rSRI = Return of SRI funds
rft = Monthly risk-free rate
rmt – rft = Excess market return
⍺i = Jensen’s alpha, intercept of the model
β1 = Slope coefficient excess market return rmt = Return of NYSE/NASDAQ/AMEX index
εit = Random error term.
In this regression model is rSRI the return of SRI funds and is rft the risk-free rate. The return of the
one-month U.S. Treasury bill is used as the risk-free rate. To calculate the excess-market return the risk-free rate, rf, t, is subtracted from the market return, rmt. The market return is calculated as the
value-weighted return on all stocks listed on the NASDAQ, NYSE and AMEX. The constant ⍺i is
Jensen’s alpha, which indicates whether a fund can outperform the market or not. The average alpha of the SRI funds will be tested for significance. Furthermore, β1 represents the slope coefficient, which is
also known as the exposure of the funds systematic risk to the market risk. The random error term is denoted as εit.
3.2 Fama and French factor models
The ongoing discussion on the suitability of the CAPM in this kind of research brings up the need to make use of a multifactor asset pricing model. Therefore, the Fama and French three-factor model and four-factor model are included in this study. The Fama and French three-factor model is as follows:
rSRI – rft = ⍺i + β1(rmt – rft) + β2SMB + β3HML + εit
Where:
SMB = Small Minus Big
HML = High Minus Low
β2 = Slope coefficient SMB β3 = Slope coefficient HML
rSRI, rft, rmt – rft, ⍺I, β1, rmt, and εit are defined as in the previous model.
In this multifactor model, the parameters are the same as for the CAPM model. However, this model includes additional parameters. Namely, the SMB variable represents the size return variable, which is the difference between the average return on the three smallest market capitalization portfolios and the three largest market capitalization portfolios at time t. The variable HML stands for High Minus Low and represent the value-growth return. This is calculated by taking the average return of the two smallest value and two largest value portfolios minus the average return of the two smallest growth and largest growth portfolios at time t. β2 and β3 indicate the slope coefficients of SMB and HML respectively.
The Fama and French four-factor is presented below:
rSRI – rft = ⍺i + β1(rmt – rft) + β2SMB + β3HML + β4MOM + εit
Where:
MOM = Momentum factor β4 = Slope coefficient MOM
rSRI, rft, rmt – rft, ⍺I, β1, β2, β3, rmt, SMB, HMLand εit are defined as in the previous model.
The Fama and French four-factor model includes the variable MOM to the model. MOM stands for momentum, which is the spread between high-performing past-year companies and low-performing past-year companies over the past twelve months. Where β4 indicates the slope coefficient of the momentum factor.
3.3 Cross-sectional regression
A cross-sectional regression is used to explain the fund performance of SRI funds with high ESG ratings. First, the model is tested with only the ESG score included. In model (1) till (3) the effect of the ESG score on the Jensen’s alphas obtained from CAPM and Fama and French three-factor and four-factor models are estimated. Followed by model (4) till model (6) which estimates the effect of the three screens of the ESG score on the Jensen’s alpha from the CAPM and Fama and French three-factor and four-three-factor models. The ESG score is excluded from the model testing the other scores since there is a very high correlation between the ESG score and the E, S and G scores since the ESG score is based on the individual E, S and G scores.
Table 2: overview of hypotheses with models
To test the first hypotheses model (1) till (3) are tested. To test hypotheses 2, 3 and 4, model (4) until model (6) are tested.
Hypotheses Model
H1: The ESG rating has a positive effect on SRI fund performance
1) ACAPM = ⍺i + β1ESG + εit
2) AFF3 = ⍺i + β1ESG + εit
3) AFF4 = ⍺i + β1ESG + εit
H2: The environmental screen of the ESG rating has a positive effect on SRI fund performance.
H3: The social screen of the ESG rating has a positive effect on SRI fund performance.
H4: The governance screen of the ESG rating has a positive effect on SRI fund performance.
4) ACAPM = ⍺i + β1*ENV + β2*SOC + β3*GOV + εit
5) AFF3 = ⍺i + β1*ENV + β2*SOC + β3*GOV + εit
where:
βi = coefficient of the following variables
ESG = the ESG score of the SRI fund ranging from 0-100
ENV = the Environment score of the SRI fund ranging from 0-100 SOC = the Social score of the SRI fund ranging from 0-100 GOV = the Governance score of the SRI fund ranging from 0-100 ACAPM = Jensen’s alpha obtained from CAPM model
AFF3 = Jensen’s alpha obtained from Fama and French three-factor model AFF4 = Jensen’s alpha obtained from Fama and French four-factor model
ACAPM, AFF3 and AFF4 represent the estimated Jensen’s alphas from the CAPM, Fama and French
three-factor and four-factor models of the 24 U.S. SRI funds over the period of 2015 till 2019. ESG indicates the total ESG score for each SRI fund and ENV, SOC and GOV represent the environment, social and governance score respectively. βi represents the coefficient of these variables. Furthermore, ⍺i presents the interceptions of the models.
As an addition to the previous models, the same regressions will be performed but only for one year at a time instead of over the period between 2015 and 2019. Since a correlation between the ESG screens could lead to a different alpha, the factors are tested for multicollinearity when the correlation is above 0.70. To test for multicollinearity the VIF function is used. If the function is below 10 there is no problem in multicollinearity.
In addition to the regression of the Jensen’s alpha, the Sharpe ratio for each fund is calculated. A graph is included to see if there is a relation between a higher alpha obtained from the CAPM model and the Sharpe ratio. The CAPM alpha used is because this model has the highest adjusted R2, and
thus best explains the model. The Sharpe ratio can be calculated through the following formula: 𝑆𝑝 = ("#$% – "()*#$%.
Where Sp is the Sharpe ratio of the portfolio, rSRI is the mean return of the SRI funds’ portfolio, rf is the
risk-free rate, which is again the 30-day T-bill rate and σSRI is the annual standard deviation of the SRI
funds’ portfolio.
4. Sample and Data
4.1 Sample
The sample is retrieved from the United States Forum for Sustainable and Responsible Investment (USSIF). USSIF indicates a comprehensive sample of all socially responsible screened mutual funds in the United States. The focus of this research is to study the performance ESG screened mutual equity funds. The sample is based on the list of US SRI mutual funds provided by USSIF on the 12th
of May 2020. The list of USSIF included equity, bond, balanced, and international global foreign fund. From this sample, only the equity funds are taken into account. Furthermore, all funds who existed less than 5 years are excluded since the performance evaluation will be studied for a period of 5 years. By selecting the equity funds and funds who were existing for more than 5 years the sample was reduced to an amount of 24 socially responsible screened funds and tested for their performance in the period December 31st, 2014 to December 31st, 2019.
4.2 Data
Data on the US equity mutual funds is collected from the Wharton Research and Data Services (WRDS) database. First, the data of the fund information is collected to obtain the funds’ holdings and the relative weight of the holdings in each fund. Since there was no access to databases that provided information on the environmental, social and governance scores of mutual funds, the scores of each fund is calculated through taking the weighted average of the scores of the top-10 holdings of each fund. To obtain the ESG, environment, social and governance ratings for each holding the WRDS Sustainalytics database is used. The WRDS Sustainalytics database uses ESG score as the
measurement on how well the SRI funds are managing environmental, social and governance (ESG) relatively to their comparable equals. The WRDS Sustainalytics database gives a score between 0-100, with 100 for the highest ESG score and 0 for the lowest score. However, this database is updated until 2018. Due to this limitation in retrieving ESG scores the latest available ratings retrieved are assumed to remain constant over time. This, however, can lead to a look-back bias since it is assumed that all ESG data is similar for the other years. This assumption biases the results when the scores change.
To calculate the monthly returns of the funds the Factset database is accessed to obtain the monthly prices of each fund. To calculate the monthly returns the following formula is adopted:
("#$"#$%) "#$%,
where Pt represents the price of the fund at time t, and Pt-1 represent the price of the fund at one period
before time t.
As mentioned in the literature review research on fund performance should take into account the possibility of survivorship bias. However, due to the lack of access to the survivorship-bias free database, this research is not able to identify non-surviving funds, and therefore acknowledge that this study might suffer from survivorship bias. Renneboog et al. (2007) concluded in their research that the number of funds which are leaving the market is low in the SRI segment. Therefore, survivorship bias is considered small in this research.
For the risk-free rate, data is retrieved from the WRDS CRSP database. The monthly return of a 30-day Treasury Bill is used as the risk-free rate. Furthermore, from the WRDS CRSP database is also the data of the benchmark retrieved. The benchmark is the weighted average index of the NASDAQ, NYSE and AMEX indices. The data for the market risk premium variable, the Fama and French factors, SMB and HML, and Carhart’s momentum factor are retrieved from the Kenneth R. French Data Library. For the SMB variable, the companies are ranked based on their market capitalization relative to the median market capitalization of companies listed on the NYSE. The HML factor looks at the difference between value and growth stocks. Growth stocks have a high book-to-market ratio, while value stocks have a low book-to-market ratio. They are divided on the 30th and 70th percentile.
Companies below the 30th percentile are considered value stocks, while the companies above the 70th
percentile are considered growth stocks. For determining the momentum factor the same reasoning is applied. The high performing firms perform above the 70th percentile, while the low performing firms
perform below the 30th percentile.
Table 3: descriptive statistics SRI funds
An overview of the excess return of the SRI funds in the period 2015 till 2019. The number of observations is the number of months included in the sample.
Observations Mean Standard
deviation
Minimum Maximum Skewness Kurtosis
SRI funds 61 0,003 0,041 -0,138 0,083 0,08 0,16
In Table 3, the descriptive statistics for the SRI funds’ excess returns are presented. The number of observations is the months of data included in the dataset. The mean indicates that on average the difference between the SRI fund return and the risk-free rate is 0,30%. The skewness value of 0,08 implies that the pattern of responses is considered a normal distribution. The kurtosis value of 0,16 shows that the distribution is not flat nor peaked.
Table 4: descriptive statistics ESG scores
Presents the ESG score and the environment, social and governance scores in 2018. Ranging from 0-100. The number of observations is the number of SRI funds included in the sample.
Observations Mean Standard Deviation Minimum Maximum
ESG score 24 63,808 4,002 55,221 70,545
Environment score 24 62,290 3,511 55,573 68,594
Social score 24 60,545 3,545 55,044 67,783
Governance score 24 69,298 7,944 54,014 80,395
Table 4 shows the descriptive statistics for the SRI funds’ environment, social and governance scores. Scores above 50 indicate good relative EGS, environment, social and governance performances (Refinitiv, 2020). The mean variable shows that the minimum of all scores included in this dataset is above 50, so all funds included in the sample are considered good performing SRI funds.
5. Empirical results
This section of the study will present the results. The first subsection presents the regression results concerning the Jensen’s alphas obtained through the CAPM, Fama and French three-factor and four-factor models. This section is followed by the presentation of possible correlations between the environment, social and governance screens and finally the Sharpe ratio.
5.1 SRI Fund performance
This subsection first presents the graphical moving average of risk-adjusted return compared to the benchmark. Followed by regression analyses of the CAPM, Fama and French three-factor and four-factor model to estimate the Jensen’s alpha.
Graph 1: moving average of excess returns
The following graph represents the average risk-adjusted return of SRI funds and the NASDAQ, NYSE and AMEX index. The returns are estimated using the CAPM model for five years period between 2015 and 2019.
The graph above shows that SRI funds seem not to outperform neither underperform the risk-adjusted return of the benchmark. Moreover, it appears that the SRI funds tend to move together with the market index over time. However, this result does not hold for every moment in time and, therefore, it is not sufficient to make only conclusions based on this graph. Hence, the regression models are estimated.
The first hypothesis, whether there is a positive relation between SRI fund performance and a higher ESG rating, is tested through the CAPM, Fama and French three-factor and four-factor models. Table 4 presents the results for three different regression models.
-. 1 5 -. 1 -. 0 5 0 .05 .1 R isk-a d ju st e d re tu rn (x1 0 0 % ) 1/1/2015 1/1/2016 1/1/2017 1/1/2018 1/1/2019 1/1/2020 Date
Table 5: regression results Jensen’s alpha
Overview of the regression estimates. The dependent variable is the average risk-adjusted return of the 24 SRI funds. The CAPM, Fama and French Three-factor and Four-factor models estimate the Jensen’s alphas. The
models present the result for the period between 2015 and 2019. Whereas the number of observations is the number of monthly returns included in this sample.
CAPM Three-factor Four-factor
Market risk premium 1,0268*** 1,0361*** 1,0383***
SMB -0,0005 -0,0005 HML -0,0008 -0,0008 MOM 0,0005 Alpha -0,0034* -0,0038** -0,0038** Observations 1.464 1.464 1.464 Adjusted R2 0,8778 0,8784 0,8662
*, **, and ***, indicate respectively the variables significance at 10%, 5% and 1% level
As can be seen from Table 5, Jensen’s alphas are negative and significant. The risk-adjusted returns for the sample range from -0,344% to -0,382%. The risk-adjusted return is the average of the alphas from the 24 SRI funds. This indicates that the SRI funds significantly underperform the benchmark at a 10% significance level for the three models, while only the Fama and French models significantly show underperformance at the 5% level. The four-factor model includes additional variables, but the adjusted R2 decreases. The decrease of the adjusted R2 indicates that the model does not explain the
variables better than the models that included fewer variables. In conclusion, the SRI funds underperform the market index significantly.
5.2 Screen performance
In this subsection, the results of the relationship between screen performance and SRI fund performance are presented. First, the ESG score relation with the risk-adjusted return is presented, followed by the effect of the environment, social and governance scores.
5.2.1 ESG score
Table 6 presents the effect of the ESG score on the risk-adjusted return of the SRI funds. The ESG score coefficient is positive and significant at a 5% level at the CAPM model. However, the ESG score coefficient is not significant in the other models. In the three models, the ESG score coefficient ranges from 0,034% to 0,027%. Also, the adjusted R2 is low which indicates that the model does not
explain the ESG score effect properly which can be due to omitted variable bias. Moreover, the adjusted R2 is highest at the CAPM model and lowest at the four-factor model. This means that the
effect of the ESG score on Jensen’s alpha is best explained in the model where the effect is estimated on the alpha of the CAPM model. Since all coefficients are positive, the effect of the ESG rating on SRI fund performance can be considered positive, however, only significant in the CAPM model.
Table 6: fund performance and ESG screen
Overview of the coefficient of the ESG score and the alpha related to the ESG score. The CAPM, Three-factor and Four-factor present the coefficients with as dependent variable the risk-adjusted return from the CAPM, Fama and French three-factor and four-factor model respectively. All models present the result for the period
between 2015 and 2019.
CAPM Three-factor Four-factor
ESG score 0,0003** 0,0003 0,0003
Alpha -0,0249 -0,0228* -0,0209*
Observations 1.464 1.464 1.464
Adjusted R2 0,0909 0,0670 0,0520
*, **, and ***, indicate respectively the variables significance at 10%, 5% and 1% level.
5.2.2 Environment, social and governance scores
Table 7 shows the effect of environment, social and governance scores on the risk-adjusted return obtained from the CAPM, Fama and French three- and four-factor models. In the first model, the environment score has a positive coefficient with significance at a 10% level. However, the social and governance scores have negative and insignificant coefficients. The three-factor and four-factor models shows the same coefficient signs, but all coefficients are insignificant.
Moreover, all coefficients are again insignificant. The coefficient of the environment scores ranges from 0,019% to 0,026%, with the highest and only significant coefficient in the CAPM model. Furthermore, the social score has negative coefficients ranging from -0,019% to -0,005%. The governance score coefficients are also negative but ranging from -0,442% to -0,282%. From this can be concluded that the environment score has a positive effect on the risk-adjusted performance of SRI funds and the social and governance scores negatively affect the risk-adjusted performance of SRI funds, with the governance score having the largest negative effect. Furthermore, the Adjusted R2 is
highest at the CAPM model whereas also the environment score has a significant coefficient. Which indicates that the CAPM model best explains the variation of the model. However, all adjusted R2s are
Table 7: fund performance and environment, social and governance scores
Overview of the coefficients of the environment, social and governance scores on the risk-adjusted returns. The CAPM, Three-factor and Four-factor present the coefficients with as dependent variable the risk-adjusted return from the CAPM, Fama and French three-factor and four-factor model respectively. All models present the result
for the period between 2015 and 2019.
CAPM Three-factor Four-factor
Environment score 0,0003* 0,0002 0,0002 Social score -0,0002 -0,0001 -0,0002 Governance score -0,0001 -0,0001 -0,0001 Alpha -0,0038 -0,0044 -0,0028 Observations 1.464 1.464 1.464 Adjusted R2 0,2138 0,1780 0,1665
*, **, and ***, indicate respectively the variables significance at 10%, 5% and 1% level.
5.2.3 Scores over time
Table 8 shows that the ESG score coefficient is in relative the most years positive. However, the ESG score coefficient is only significant in 2015. In conclusion, the ESG score only affects the
risk-adjusted return significantly in 2015 with 0,09%. In the other years, the coefficients are not significant. This is partly in line with the results from Table 6, which showed a significant positive coefficient for the ESG score. However, the coefficients in Table 8 are mostly insignificant. Furthermore, the models have very low adjusted R2 which could mean that there is omitted variable bias.
Table 8: regression screens controlling for years
Presents the effect of the ESG score for each subsequent year on the risk-adjusted return obtained through the CAPM model.
2015 2016 2017 2018 2019
ESG score 0,0009* -0,0001 0,0035 0,0033 0,0042
Adjusted R2 0,1143 0,0360 0,0536 0,0910 0,1304
*, **, and ***, indicate respectively the variables significance at 10%, 5% and 1% level.
Table 9 presented below shows that the effect of ESG screens differ per year and can’t be considered consistent. The governance score coefficient remains their negative sing. However, the environment and social scores coefficients tend to change over time. The environment score coefficients are the only coefficients that include significance. In the years 2018 and 2019 the environment score coefficient is significant and positive ranging from 0,03% until 0,04%. This is in line with the conclusion from Table 7, that the environment score is the only score that has a significant and positive effect. However, the social score coefficients are negative in three out of the five years and in the other two years the coefficients are positive. With the relative most negative coefficients this is also in line with the results from the coefficient over 5-years from Table 7. The governance score coefficient has only negative signs which is the same sign as found in Table 7.
In conclusion, the environment score is the only screen that can to affect the risk-adjusted return positively, which confirms the hypothesis about the environment score. However, the hypotheses on the effect of social and governance scores are rejected. Both scores do not show significant and
positive coefficients in the regression over 5-years and neither for the regression for the separate years.
Table 9: regression screens controlling for fiscal years
Presents the effect of the environment, social and governance scores for each subsequent year on the risk-adjusted return obtained through the CAPM model.
2015 2016 2017 2018 2019
Environment score 0,0007 -0,0003 -0,0001 0,0003* 0,0004 **
Social score -0,0005 0,0007 0,0006 -0,0002 -0,0004
Governance score -0,0001 -0,0003 -0,0001 -0,0003 -0,0001
Adjusted R2 0,0922 0,0202 0,0540 0,1216 0,2070
*, **, and ***, indicate respectively the variables significance at 10%, 5% and 1% level.
To test whether the variables are correlated the above showed table is presented. Table 10 shows that there is a high correlation between the Social score and the Environment score. The high correlation indicates that a higher environment score is related to a higher social score. Since the correlation is high it is tested for multicollinearity. The correlation between the social score and governance score is in the same direction, however, the correlation is small. Furthermore, the correlation between the environment score and governance score is -0,1559. The negative correlation indicates that a higher environment score is related to a lower governance score, and the other way around.
Table 10: correlation between the scores
Shows the correlation between the environment, social and governance scores.
Environment score Social score Governance score
Environment score 1.0000
Social score 0.7960 1.0000
Governance score -0,1559 0.0529 1.0000
Table 11 shows that the highest VIF score. The VIF implies multicollinearity when it takes a value above 10. The VIF score is 3,06, which means that the high correlation between the environment and social score does not result in a problem for the fit of the model and the interpretation of the results.
Table 11: test for multicollinearity
Present the VIF score of the environment, social and governance scores.
VIF
Environment score 3,06
Social score 2,99
In Appendix Table 1 the Sharpe ratios are presented for each fund. In graph 2, the graphical relation between the alphas from the CAPM model and the Sharpe ratio is shown. This graph shows that a higher alpha is related to a higher Sharpe ratio. This strengthens the reliability of the alphas found in the previous regression models.
Graph 2: the moving average of alpha and Sharpe ratio
Presents the relation between the alpha from the Fama and French four-factor model and the Sharpe ratio.
6. Conclusion
The final part of this paper provides an overview of the results found and how these findings can be related to current literature. This overview is followed by the limitations of this study. Finally, the implications of this work are presented together with suggestions for further research.
6.1 Summary and discussion
This paper investigates whether the ESG score and the individual Environment, Social and
Governance (E, S and G) scores affect the U.S. socially responsible mutual equity fund performance in the period between 2015 and 2019. This study aimed to fill the existing gap in the literature on the effect of ESG screens on SRI fund performance. This led to the main research question: “Do the environment, social and governance screens of ESG-ratings affect socially responsible investment funds’ risk-adjusted return?”. To answer the question several models are used to assess the different effects of E, S and G scores on SRI fund performance.
First, this study analyzed SRI fund performance by estimating the Jensen’s alpha. This study found that SRI funds have a negative alpha in all models used. A negative alpha indicates that they underperform the market index significantly. Therefore, these results are not consistent with the studies from Staman and Gluskov (2009), Halbritter and Dorfleitner (2015), Auer and Schuhmacker (2016), whom all find positive abnormal returns of SRI funds. However, positive abnormal returns are not significant. -. 5 0 .5 1 Sh a rp e R a ti o -.01 -.005 0 .005
Secondly, the relation between the ESG scores and SRI fund performance is tested. The ESG score coefficients are positive in all models. However, the ESG score coefficient is only significant in the model with the CAPM alpha as the dependent variable. This indicates that a higher ESG score leads to a higher SRI fund performance. This is consistent with a study from Kempf and Osthoff (2007) and a study from Statman and Gluskhov (2009) for the period from 1992 to 2007. However, it contradicts with a study from Cherrejee et al. (2018) who find a negative relation and with a study from Halbritter and Dorfleitner (2015), which state that the abnormal returns by high ESG ratings are insignificant.
When controlling for the different screens of the ESG score, the results differ per screen. The Environment score shows only a significant effect in the model with the CAPM alpha as the dependent variable but shows in all models a positive effect on the SRI fund performance. This matches the results from the papers from Derwall et al. (2011) and Klassen and McLaughlin (1996). However, the Social score shows a negative and insignificant effect on all models. This can be due to the learning effect mentioned by Edams (2011). This study suggests that employee satisfaction is hard to price and is therefore not always taken into account. Furthermore, the Governance score shows only negative insignificant effects. This can be due to the differences per industry and per country.
According to Daszyńska-Żygadło (2016), results can differ per industry and this is not taken into account in this research. The effect of the screens per year does not differ much from the results over five-years. The environment screen is still the only screen that can affect the risk-adjusted return positive and significant. In conclusion, ESG and environment scores have a positive effect on SRI fund performance. While social and governance scores do not affect the SRI fund performance s positive nor negative significantly.
6.2 Limitations
The results of this study should be considered with some limitations. First, the ESG data is only accessible for the year 2018 and therefore the screening data from 2018 is matched to historical returns. This results in a look-ahead bias, which indicates that data is used in a certain period that would have not been the same during that respective period. However, the possibility that the fund policies did not change is present since fund policies and strategies are often strict. Second, only the top-10 holdings of each SRI fund are taken into account, which neglects the effect of potential ESG scores of other holdings. Third, the sample consists only of 24 U.S. SRI funds, this small sample size could explain the insignificance in some models. Fourth, in current literature is mentioned that the results might differ per country while in this sample only U.S. funds are included and therefore the results are not generalizable for other countries. Finally, survivorship bias is a potential issue. The data used does not take into account the survivorship bias issue. Therefore, the results can be upward biased since bad-performing funds will not provide their information.
6.3 Implications and recommendations
In this study, the ESG scores from 2018 are obtained. A starting point for further research could be performing the research with ESG scores for each of the different years, which eliminates the first bias. Another suggestion for further research is to include all holdings per fund instead of only the top-10 holdings. Furthermore, a small amount of socially responsible investment funds is considered. To reduce the effect of insignificant results, more SRI funds could be considered. Additionally, control variables, like funds’ age, funds’ size, management fees, country and industry, could be included to increase the explanatory power of the models. It is also interesting to research whether the effect of ESG scores stop at a specific level and what the optimal scores could be. This could be interesting since a very high score could be too expensive to eliminate the potential benefits and since it is not clear whether the risk-adjusted return changes due to changes in ESG scores or the other way around. Moreover, one could investigate whether there is an effect of survivorship bias on ESG scores. Finally, these recommendations are outside the scope of this thesis and could not be integrated into this study. However, the results are still applicable since they provide new evidence on the effect of screening categories on SRI fund performance.
7. References
Bauer, R., Derwall, J., & Otten, R. (2006). The Ethical Mutual Fund Performance Debate: New Evidence from Canada. Journal of Business Ethics, 70(2), 111–124.
Bauer, R., Koedijk, K., & Otten, R. (2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking & Finance, 29(7), 1751–1767.
Bauer, R., Otten, R., & Rad, A. T. (2006). Ethical investing in Australia: Is there a financial penalty? Pacific-Basin Finance Journal, 14(1), 33–48.
Bollen, N. P. B., & Busse, J. A. (2001). On the timing ability of mutual fund managers. Journal of
Finance, 1075–1094. Retrieved from https://onlinelibrary.wiley.com/journal/15406261
Bourghelle, D., Jemel, H., & Louche, C. (2009). The integration of ESG information into investment processes: Toward an emerging collective belief? Corporate Responsibility, Market Valuation and
Measuring the Financial and Non- Financial Performance of the Firm.
Chatterjee, S., Das, N., Ruf, B., & Sander, A. (2018). Fund Characteristics and Performances of Socially Responsible Mutual Funds: Do ESG Ratings Play a Role? Journal of Accounting and
Finance, Forthcoming.
Chegut, A., Schenk, H., & Scholtens, B. (2011). Assessing SRI fund performance research: Best practices in empirical analysis. Sustainable Development, 19(2), 77–94.
Daszyńska-Żygadło, K. (2016). Sustainable Value Creation-performance of European Manufacturing Companies. European Financial Systems, 86–93.
Derwall, J., Guenster, N., Bauer, R., & Koedijk, K. (2005). The Eco-Efficiency Premium Puzzle. Financial Analysts Journal, 61(2), 51–63.
Edmans, A. (2011). Does the stock market fully value intangibles? Employee satisfaction and equity prices. Journal of Financial Economics, 101(3), 621–640.
Elton, E. J., Gruber, M. J., & Blake, C. R. (1996). The persistence of risk-adjusted mutual fund performance. Journal of business, 133-157.
Friede, G., Busch, T., & Bassen, A. (2015). ESG and financial performance: aggregated evidence from more than 2000 empirical studies. Journal of Sustainable Finance & Investment, 5(4), 210–233.
Friedman, M. (1970). The social responsibility of business is to increase its profits. New York Times
Magazine, September 13(32), 122–124.
Gil-Bazo, J., Ruiz-Verdú, P., & Santos, A. A. P. (2009). The Performance of Socially Responsible Mutual Funds: The Role of Fees and Management Companies. Journal of Business Ethics, 94(2), 243– 263.
Godfrey, P. C. (2005). The Relationship Between Corporate Philanthropy and Shareholder Wealth: A Risk Management Perspective. Academy of Management Review, 30(4), 777–798.
Godfrey, P. C., Merrill, C. B., & Hansen, J. M. (2009). The relationship between corporate social responsibility and shareholder value: an empirical test of the risk management hypothesis. Strategic
Management Journal, 30(4), 425–445.
Grinblatt, M., & Titman, S. (1994). A study of monthly mutual fund returns and performance evaluation techniques. Journal of Financial and Quantitative Analysis, 1994(September), 419–444.
Henke, H.-M. (2016). The effect of social screening on bond mutual fund performance. Journal of
Banking & Finance, 67, 69–84.
Herring, C. (2009). Does diversity pay?: Race, gender, and the business case for diversity. American
Sociological Review, 74(2), 208–224.
Herringer, A., Firer, C., & Viviers, S. (2009). Key challenges facing the socially responsible investment (SRI) sector in South Africa. Investment Analysts Journal, 38(70), 11–26.
Huijgevoort, J. (2017). The relationship between ESG-factors and the corporate financial
performance.
Huselid, M. A. (1995). The impact of human resource management practices on turnover,
productivity, and corporate financial performance. Academy of Management Journal, 38(3), 635–672.
Jackson, S. E., & Schuler, R. S. (1995). Understanding human resource management in the context of organizations and their environments. Annual Review of Psychology, 46(1), 237–264.
Kempf, A. and Osthoff, P. (2007) ‘The effect of socially responsible investing on portfolio performance’, European Financial Management, 13(5), 908–922.
Kinder, P. (2005). Socially responsible investing: An evolving concept in a changing world. KLD
Research & Analytics, Inc., 1–65.
Klassen, R. D., & McLaughlin, C. P. (1996). The Impact of Environmental Management on Firm Performance. Management Science, 42(8), 1199–1214.
López, M. V., Garcia, A., & Rodriguez, L. (2007). Sustainable Development and Corporate Performance: A Study Based on the Dow Jones Sustainability Index. Journal of Business
Ethics, 75(3), 285–300.
Luther, R., & Matako, J. (1994). The performance of ethical unit trusts: choosing an appropriate benchmark. Britisch Accounting Review, (26), 77–89.
Luther, R., Matatko, J., & Corner, D. (1992). The investment performance of UK ethical unit trusts. Auditing and Accountability Journal, 57–70.
MacAvoy, P. W., & Millstein, I. M. (1999). The active board of directors and its effect on the
performance of the large publicly traded corporation. Journal of Applied Corporate Finance, 11(4), 8– 20.
Maslow, A. H. (1943). A theory of human motivation. Psychological Review, 50(4), 370–396.
McGuire, J. B., Sundgren, A., & Schneeweis, T. (1988). Corporate Social responsibility and firm financial performance. Academy of Management Journal, 31(4), 854–872.
McWilliams, A, & Siegel, D. (1997). Event studies in management research: Theoretical and empirical issues. Academy of Management Journal, 40(3), 626–657.
McWilliams, Abagail, & Siegel, D. (2001). Corporate Social Responsibility: A theory of the Firm Perspective. Academy of Management Review, 26(1), 117–127.
McWilliams, Abagail, Siegel, D., & Teoh, S. H. (1999). Issues in the Use of the Event Study Methodology: A Critical Analysis of Corporate Social Responsibility Studies. Organizational
