Miika Räisänen The Performance of Responsible Equity Funds During Market Crises Vaasa 2020 School of Accounting and Finance Master’s Thesis in Finance Master’s Degree Programme in Finance 1 UNIVERSITY OF VAASA School of Accounting and Finance Author: Miika Räisänen Title of the Thesis: The Performance of Responsible Equity Funds During Market Cri- ses Degree: Master of Science in Economics and Business Administration Programme: Master’s Degree Programme in Finance Supervisor: Sami Vähämaa Year: 2020 Sivumäärä: 78 ABSTRACT: The purpose of this thesis is to study the performance of responsible equity funds during periods market crises. Previous literature has found evidence of outperformance by stocks of respon- sible companies and responsible mutual funds during market crises. Additionally, this thesis pro- vides evidence of the performance of responsible equity funds and compares risk-adjusted re- turns of responsible investing to those of conventional equity mutual funds during a sample period of almost two decades. This study employs a dataset of 110 US-based socially responsible funds and 120 US-based con- ventional equity funds from January 2000 to October 2019. These mutual fund groups are used to construct two time-series of the returns of an equal-weighted portfolio of the SRI funds and conventional funds, respectively. The abnormal returns of SRI and non-SRI are measured using the capital asset pricing model, the Fama-French three-factor model and the Carhart four-factor model. In order to measure the performance of these portfolios during crisis periods, these asset pricing models are extended to include crisis and non-crisis period alphas. The results suggest that both responsible funds and conventional funds do not generate abnor- mal returns during the whole sample period. In addition, SRI significantly underperforms non- SRI during crises and their non-crisis period performance are similar. However, during January 2010 – October 2019 the SRI portfolio outperforms the conventional portfolio during crises and normal market conditions, although, the difference between the two is not significant. Responsible investing does not provide investors with downside protection during periods of market turbulence. On the contrary, SRI underperforms during market crises which contradicts previous research. Additionally, the performance of SRI does not significantly differ from that of non-SRI during normal market conditions. These results suggest that investors should not expect abnormal returns while investing responsibly. However, investors should not favor non-SRI nei- ther, since the risk-adjusted returns of SRI do not significantly differ from those of conventional mutual funds. KEYWORDS: Responsible investing, Mutual funds, Market crises, ESG, CSR 2 Contents 1 Introduction 6 2 Literature review 8 2.1 What is socially responsible investing? 8 2.2 Development of SRI 9 2.3 Current SRI markets 12 2.4 Performance of socially responsible investing 13 2.4.1 Sin stocks 14 2.4.2 Negative, positive and best-in-class screening 15 2.4.3 Individual ESG factors 17 2.4.4 Socially responsible mutual funds and indices 20 2.4.4 ESG events 22 2.4.5 Active Ownership 24 2.4.6 SRI performance during market crises 24 2.5 CSR and financial performance 26 2.5.1 Cost of capital 27 2.5.2 CSR and accounting performance 28 3 Theoretical framework of socially responsible investing 30 3.1 Corporate perspective 30 3.1.1 Shareholder, stakeholder and instrumental stakeholder 31 3.1.2 Risk mitigation 35 3.2 Investor perspective 35 4 Data and methodology 44 4.1 Data description 44 4.2 Methodology 48 4.2.1 CAPM 49 4.2.2 Fama-French 3-factor model 49 4.2.3 Carhart 4-factor model 50 4.2.4 Factor models with crisis and non-crisis alphas 51 3 5 Empirical analysis 52 5.1 Descriptive statistics 52 5.1.1 Whole sample period and two sub-periods 52 5.1.2 Bear markets, corrections and recoveries 56 5.2 Factor models 59 5.2.1 Single-factor results 59 5.2.2 Fama-French three-factor results 61 5.2.3 Carhart four-factor results 63 5.2.4 Risk-adjusted returns of two subsamples 64 6 Conclusions 68 References 71 4 Figures Figure 1. ESG-Sharpe ratio frontier, adapted from Pedersen et al. (2019). 40 Figure 2. ESG-CAPM, adapted from Pedersen et al. (2019). 40 Figure 3. Cash-flow transmission channel (Giese et al. 2019). 42 Figure 4. Idiosyncratic risk channel (Giese et al. 2019). 42 Figure 5. Valuation channel, (Giese et al. 2019). 43 Figure 6. Morningstar Style Box, adapted from Morningstar (2008). 46 Figure 7. Cumulative returns of SRI and non-SRI portfolios (Jan 2000 - Oct 2019). 53 Figure 8. Cumulative returns of SRI and non-SRI portfolios (Jan 2000 - Dec 2009). 54 Figure 9. Cumulative returns of SRI and non-SRI portfolios (Jan 2010 - Oct 2019). 54 Tables Table 1. Characteristics of SRI and non-SRI equity funds 45 Table 2. Summary of investment styles of SRI and non-SRI equity funds 47 Table 3. S&P 500 performance during market crises 48 Table 4. Monthly performance statistics 55 Table 5. Monthly performance statistics for crisis periods 57 Table 6. Monthly performance statistics for recovery periods 59 Table 7. SRI and non-SRI fund performance using a single-factor model 60 Table 8. SRI and non-SRI fund performance using the Fama-French three-factor model 62 Table 9. SRI and non-SRI performance using the Carhart four-factor model 64 Table 10. SRI and non-SRI fund performance during two subsamples 66 Abbreviations AUM – Assets under management CAPM – Capital asset pricing model CER – Corporate environmental responsibility CFP – Corporate financial performance CSP – Corporate social performance 5 CSR – Corporate social responsibility DCF – Discounted cash-flow EMH – Efficient market hypothesis ESG – Environmental, social & governance SRI – Socially responsible investing UN PRI – United Nations Principles for Responsible Investment ROA – Return on assets ROE – Return on equity 6 1 Introduction Responsible investors take matters of environmental, social and governance (ESG) into consideration when making investment decisions – whether choosing between individ- ual stocks or when selecting mutual funds. As the society is increasingly concerned about climate change and corporate social responsibility (CSR), responsible investing offers in- vestors a framework to mitigate the effects of, for example, lacking governance stand- ards or the transition risk of climate change on their portfolios and generate returns, that are at least similar to those of non-SRI. Socially responsible investing (SRI) has become mainstream in the recent years while there is much heterogeneity surrounding the definitions and terminology of SRI. There has been research on SRI since the early 1970s but there is not a consensus on the per- formance of SRI. Opponents of SRI claim that due to a restricted investment set, socially responsible funds lack the diversification of conventional funds. This should lead to un- derperformance by SRI funds. Proponents of responsible investing argue that corporate social responsibility leads to better financial performance and thus to increased invest- ment performance. Others (incl. Nofsinger & Varma, 2014) claim that responsible invest- ments have insurance-like properties in that they perform better during market crises. However, due to definitional heterogeneity, research on SRI is difficult and many studies employ different methods to estimate investment returns. Small data sets and differing time frames are also a problem when comparing the results of these studies. Recent surge of supply of socially responsible investment products has increased the importance of research on SRI. Investment managers have their own definitions of SRI while claiming that these products have offered abnormal returns in the past. However, there is not a consensus on the outperformance of these products compared to conven- tional ones. The purpose of this research is to study the performance of socially respon- sible mutual funds during different phases of a market cycle – crisis and post-crisis peri- ods. Lins et al. (2017) find that investing in responsible companies during market crises 7 generates significant abnormal returns. Moreover, Nofsinger and Varma (2014) find that SRI mutual funds outperform their conventional peers during periods of bear markets. This thesis contributes to the existing literature by measuring the performance of SRI equity mutual funds during the two most recent bear markets in the US – the Dot-Com bubble of 2000 – 2002 and the financial crisis of 2008 – 2009, as well as during the seven market corrections of the 2000s. Do these mutual funds offer protection to their inves- tors or are investors better off investing in conventional mutual funds during extreme market conditions and do investors pay for the potential risk-mitigative property of SRI during normal market conditions? The research hypotheses of this thesis are the following. First, responsible investing overperforms non-SRI during market crises, and thus exhibits an insurance-like property. Second, SRI is hypothesized to underperform during non-crisis periods, during which in- vestors pay the premium for overperformance during crises. Third, the overall perfor- mance of responsible funds and their conventional peers do not differ significantly. The structure of this study is the following: the second chapter includes a literature re- view which defines the concepts related to the subject of this thesis and presents the literature on the development and performance of responsible investing. The third chap- ter presents the theoretical background of corporate social responsibility and responsi- ble investing from the corporate perspective as well as from the perspective of an inves- tor. The fourth chapter presents the data and the research methodology and chapters five and six are dedicated to the results and the conclusions of this thesis, respectively. 8 2 Literature review This section defines the concepts related to socially responsible investment as well as reviews previous literature on the development and performance of responsible portfo- lios and responsible mutual funds as well as the screening techniques employed by so- cially responsible investors. The previous literature has tried to answer whether inves- tors in SRI have been able to achieve their simultaneous goal of wealth-maximization and social responsibility. The field is subjected to heterogeneity on many different levels, as no consensus has been formed on the definition, terminology or even the perfor- mance of SRI. 2.1 What is socially responsible investing? Socially responsible investment funds have seen an increasing interest in them in the last few years. In 2018 the assets under management by socially responsible investors is es- timated at over USD 20 trillion (Forbes, 2018). Although responsible investing has been studied for over four decades, definitions and terminology of SRI have not been settled on. According to Sparkes and Cowton (2004) the most common terms in the field are socially responsible investing and ethical investing. Of these two, they argue that ethical investing is the older one, but it has increasingly been replaced by socially responsible investment. Sandberg et al. (2009) argue that the term ethical investing originates from the exclusion of certain companies on moral grounds by churches and that the term is was created in the UK. Over time, the US term socially responsible investing has come to replace ethical investment. According to Climent and Soriano (2011) SRI uses environ- mental, social and governance factors into investment process – including analysis, se- lection and choice of investment. The UN Principles for Responsible Investment (UN PRI) defines ESG integration as “the systematic and explicit inclusion of material ESG factors into investment analysis and investment decisions.” However, since the integration of ESG considerations to investment activities has remained inconsistent and difficult to 9 measure and compare, asset owners are facing difficulties in identifying ESG practices that truly add financial value. (Sloggett, 2016.) Sandberg et al. (2009) reviewed the website sections dedicated to SRI, online quarterly reports, investment policy documents and annual reports of 101 institutional investors which had signed the UN Principles for Responsible Investment. They state that socially responsible investing is the most common term used to express investments which inte- grate ESG factors in the investment process. The second most popular term is “ethical investing” whereas other existing terms are “social”, “responsible”, “natural” and “val- ues-based investing”. Cultural differences between regions and market conditions hin- der standardization – fund companies have incentives to develop their own definitions of SRI. Sandberg et al. (2009) argue that homogenization could be made possible by a top-down standardization movement. However, practitioners do not view the heteroge- neity of SRI as a problem in the way that academics do. Fabio and Jondeau (2019) argue that the use of ESG factors is caused by the increase in the availability of data. Sandberg et al. (2009) state that there is not a consensus on how these factors should be integrated nor on the definitions of this kind of investing. All in all, there are four levels heterogeneity in SRI: terminological, definitional, strategic and practical. There is some agreement on the definitional level as most SRI proponents agree on the definitions of the terms ethical and socially responsible investing. However, the debate is centered on how much weight is given to financial concerns compared to non-financial ones. (Sandberg et al., 2009.) 2.2 Development of SRI According to Sloggett (2016) the first steps toward socially responsible investing were taken by faith-based organizations in the 1970s as they believed that their investing ac- tivities should reflect their values and that their investments could change the way com- panies practiced business. During the 1970s SRI was mostly viewed as a fringe activity. 10 Due to response to the demands of asset owners, the weight of SRI and ESG strategies has grown dramatically. (Sloggett, 2016.) Moskowitz (1972) was one of the first to consider whether social issues should be con- sidered in investment decisions. He stated that socially responsible investments do not have to be financially weak. He argued that the social awareness of SR companies will enable them to surpass their competitors. According to the author, some of the earliest adaptors of socially responsible investing were the two largest philanthropic institutions of the US – the Ford Foundation and the Rockefeller Foundation. These foundations stud- ied their investment portfolios and looked for investments which violated the social-im- provement guidelines. In addition, Yale and Cornell analyzed their investment portfolios through a social lens. At that time the first SRI funds were organized. These funds had mandates to favor companies which were socially responsible. For example, the Vantage Ten Ninety Fund was under obligation to invest at least 10 percent of its portfolio to firms which were involved in pollution control and combating the problems in inner cities and the war on hunger. (Moskowitz, 1972.) Church investment organizations were the first to use ethical investing but in the early 2000s more and more SRI funds have been offered to the public. The first SRI funds were formed in 1971 in the US and 1984 in the UK (Sparkes & Cowton, 2004.) The original model was to avoid companies which were deemed unethical and exclusionary screens are still the predominant methodology used by the SRI industry (Schepers & Sethi, 2003). A small number of funds have started to account positive factors in company analysis, such as employment of ethnic minorities, whereas some have concentrated their invest- ments in the environmental area, although, avoidance remains the dominant model in the industry (Sparkes & Cowton, 2004). In the recent years, socially responsible investment has shifted from margin to main- stream which has led to increasing interest in SRI by institutional investors (Sparkes & 11 Cowton, 2004). They also argue that in addition to significant growth in SRI, the industry has also matured by getting acceptance from large investment institutions such as pen- sion funds and insurance companies. Earlier, SRI was a niche that got interest only from few specialist retail investment funds. Sparkes and Cowton (2004) argue that corporate executives can no longer ignore ESG factors due to the fact institutional investors are the most important owners of publicly listed companies. Due to the recent shift towards SRI among large institutional investors, SRI funds are gaining bargaining leverage over cor- porate executives. Social issues will be emphasized in the corporate agenda and thus corporate decision makers must be aware of issues regarding corporate social responsi- bility. (Sparkes & Cowton, 2004.) However, according to Sandberg et al. (2009) there is a shift toward the term “responsible investment” amongst academics and professionals which reflects opposition toward overweighting the social issues compared to environ- mental and financial factors. In addition to the emphasis on SRI by institutional investors and corporate executives, the whole asset management industry is facing pressure from regulators to report on how they address SRI issues in some European countries (Ales- sandrini & Jondeau, 2019). According to Riedl and Smeets (2017) SRI equity holdings of investors are motivated by social signaling. Investors who talk more about their investments are more likely to hold socially responsible investments. However, the authors find that socially responsible in- vestors donate about 41 percent more to charity than conventional investors. This im- plies that SRI is not a substitute for charity donations. In addition to social preferences playing an important role in determining socially responsible investing, financial motives also affect whether investors invest in a socially responsible way. The results of Riedl and Smeets (2017) are based on data of individual investors. Although private investors may weigh financial performance less than individual investors, a survey by Amel-Zadeh and Serafeim (2018) finds that the primary motivation of institutional investors for using re- ported ESG information is the relevance to investment performance. 12 Riedl and Smeets (2017) add that investors accept higher management fees on SRI funds than on conventional funds. Most SR investors also expect that SRI funds will underper- form relative to conventional funds. Thus, some investors are willing to accept lower financial performance from mutual funds that match their social preferences. The au- thors also find that responsible investors have a longer investment horizon. Those inves- tors who hold their funds longer are more likely to be SR investors as well. Riedl and Smeets (2017) argue that as SRI continues to grow, social preferences may drive up prices of socially responsible companies and lower the prices of sin companies. 2.3 Current SRI markets According to Lang and Electris (2018) sustainable investments have grown 34% in two years, and at the start of 2018 they were at USD 30.7 trillion globally. The market share of responsible investing has grown in every market expect for Europe where assets held by professional SR investors grew by 11% from 2016 to 2018, but their total share of the overall market declined from 53% to 49%. Professionally managed responsible assets range from 18% in Japan to 63% in Australia and New Zealand. Japan has been the fastest growing region from 2016 to 2018 whereas the three largest regions were Europe, the United States and Japan. (Lang & Electris, 2018.) Currently, the largest SRI strategy globally is negative screening, followed by ESG integra- tion and shareholder activism. Negative screening employs the exclusion of certain in- vestments based on specific ESG criteria while ESG integration applies environmental, social and governance factors into financial analysis. (Lang & Electris, 2018.) In corporate engagement and shareholder action shareholders use power to influence corporate be- havior and decision-making. The three largest regions have different dominant strategies since the most of European SRI uses negative screening, the US market is dominated by ESG integration and shareholder activism is the preferred strategy in Japan. (Lang & Elec- tris, 2018.) 13 SRI has been a broadly practiced and accepted form of investing in Europe but there are signs that the market is maturing (Lang & Electris, 2018). The authors suggest that this drop may be due to stricter standards and definitions being implemented. In March 2019, the European Parliament adopted rules under its Sustainable Finance Action Plan to re- quire asset managers to standardize their ESG integration disclosure and to use common reporting standards. This is meant to prevent asset managers from overstating their ef- forts towards sustainable investing. Some European asset managers reported lower SR asset values for the 2018 Eurosif survey than in 2016 anticipating the stricter standards and definitions. (Lang & Electris, 2018.) In the US, SRI is continuing to expand. According to Lang and Electris (2018) the leading motivation for incorporating ESG criteria into investment process is client demand. Over half of the US money managers also explained their increasing interest in SRI as fulfilling a mission or values, pursuing social or environmental benefits and minimizing risks and fulfilling fiduciary duty. Investor demand for socially responsible assets is increasing and this change in investor values and views may have an impact for the expected perfor- mance of responsible investing. 2.4 Performance of socially responsible investing The key part of responsible investing is corporate social responsibility and at the heart of CSR is the debate of the stakeholder theory versus the stockholder theory. The litera- ture concerning ESG investing is a part of research on how corporate social responsibility effects corporate financial performance. This literature can be separated into four cate- gories. The first category studies whether investors can form portfolios based on ESG factors. Some researchers study the performance of socially responsible mutual funds relative to their conventional peers, while some form portfolios from individual stocks. The sec- ond strand of research addresses the stock market response to news related to ESG 14 factors by implementing event-study methodology. The third category studies the effect of sustainability and the cost of capital. The last strand studies the relationship between sustainability and operational performance which are usually measured by accounting- based performance measures, such as return on assets (ROA) and return on equity (ROE). 2.4.1 Sin stocks One subject of which researchers are unanimous is the outperformance exhibited by sin stocks. High expected returns of sin stocks suggest that the cost of capital of these com- panies is higher than those of responsible firms. Hong and Kacperczyk (2009) hypothe- size that due to societal norms some investors abstain from investing in so called sin stocks – publicly traded companies involved in producing such products as alcohol, to- bacco and gambling activities. They also find that sin stocks are less held by institutional investors and they are less followed by analysts. Merton’s (1987) work on neglected stocks and segmented markets explains why sin stocks should be cheaper than their comparables. Due to being neglected by institutional investors, the valuations of sin stocks should be depressed. Hong and Kacperczyk (2009) show that sin stocks tend to have lower valuations compared to comparables using price- to-book and price-to-earnings ratios. The valuation ratios of sin stocks are on average 15 to 20% lower than those of comparables. This undervaluation leads to sin stocks having higher expected returns than comparables. The authors find that a portfolio long sin stocks and short their comparables has a return of 0.26% per month after adjusting for a four-factor model consisting of the Fama-French three-factor model with the momen- tum factor. Also, using cross-sectional regressions controlling for firm characteristics, they find outperformance by sin stocks of 0.29% per month. Fabozzi et al. (2008) argue that the society sees firms making “bad products” as “bad firms”. Some view that “bad firms” are thus bad stocks, although their valuations should only be determined by their risk-reward relationship according to financial theory. They 15 find that between January 1970 and June 2007 the average sin stock produced an annual return of 19.02% while the average stock market had an average annual return of 7.87%. The sin portfolio outperformed the relevant market index in 35 of 37 years. Their sample consists of 267 stocks in alcohol, tobacco, biotech, adult services, gaming and weapons. 2.4.2 Negative, positive and best-in-class screening ESG investing is often implemented by using different screening techniques on the in- vestment performance. Negative screening, which is the most common method of SRI, excludes companies if they do business in harmful industries or are associated in non- ethical activities. For example, companies in involved with business areas such as to- bacco, alcohol, gambling and firearms are usually excluded. Positive screening includes companies, that have demonstrated corporate social responsibility, in the investment portfolio. Companies are rated based on a set of ESG criteria and based on the rating investors choose the highest-rated companies. Best-in-class is a subclass of positive screening. It includes companies which are socially responsible without sectorial exclu- sion. Thus, the portfolio is better diversified than portfolios formed using negative or positive screening because best-in-class portfolios are invested in a larger number of in- dustries. Trinks and Scholtens (2017) examine how the implementation of negative screens affects an investor’s portfolio and whether the choice of the screening technique matters for returns. Most previous studies investigate the combination of tobacco, alcohol and gam- ing stocks whereas Trinks and Scholtens (2017) analyze the returns of fourteen contro- versial issues for the period 1991 to 2012 in several international markets. The authors do not exclude complete industries as they select at the level of the individual firm. They find that controversial stocks have positive abnormal returns and that negative screening causes financial underperformance. The different value-weighted sin portfolios outper- form the market which refers to the Fama-French 4 factor benchmark by 91 to 104 basis 16 points. Trinks and Scholtens (2017) also argue that screening can reduce the universe of investment objects significantly. Auer (2016) studies the effect of negative screening on portfolio Sharpe ratios using 2004 to 2012 data of STOXX 600 and ESG ratings from Sustainalytics. Auer (2016) prefers neg- ative screens to positive screening as he finds that the exclusion of unrated stocks pro- vides significantly higher returns than a passive benchmark. The author also finds that additional exclusion based on environmental and social scores neither adds nor destroys portfolio value while the exclusion of poor governance rating firms increases the Sharpe ratio of the portfolio. Hoepner and Schopohl (2018) analyze the performance of stocks which have been ex- cluded from the portfolios of two leading Nordic investors, Norway’s Government Pen- sion Fund-Global and Sweden’s AP-funds. They find that the portfolios of excluded stocks underperform relative to the benchmark index of the funds. The authors conclude that exclusionary screens employed by asset managers do not compromise the returns of their fiduciaries – the exclusions neither benefit nor harm the portfolios. Although liter- ature states that exclusionary screening is an outdated method, large institutional inves- tors seem to prefer it (Hoepner & Schopohl, 2018). Kempf and Osthoff (2007) study the performance of stock portfolios using three different screening techniques employed by socially responsible investors. They form the portfo- lios by using KLD ratings which measure the social responsibility of a company. The com- panies are evaluated according to two categories: qualitative and exclusionary criteria. Using the qualitative criteria, they form portfolios via positive screening and best-in-class screening. Exclusionary screens are used for the negative screening. By buying stocks with high ESG ratings and selling stocks with low ESG ratings, investors can achieve high abnormal returns – up to 8.7% per year. Negative screening is not efficient whereas pos- itive screening and best-in-class screening offer the best returns. They also find that best- in-class earns the highest alpha. 17 Halbritter and Dorfleitner (2015) create ESG portfolios using ESG data of ASSET4 from 2002 to 2011, Bloomberg from 2005 to 2011, and KLD from 1990 to 2011 for the U.S. market. High - low portfolios constructed using ASSET4 data exhibit a positive alpha whereas most of the portfolios formed using Bloomberg data do not exhibit a statistically significant relationship between ESG scores and returns. Only high social score compa- nies generate high statistically significant alphas. The ESG ratings have no effect on ab- normal returns when using KLD ratings, either. This result holds for overall ESG scores as well as for individual ESG factors. Even best-in-class method could not generate abnor- mal returns. The authors argue that the performance results are dependent on the rating approach and the company sample, and thus data based different ESG rating services offer different results. In addition, the use of most recent data is crucial as socially re- sponsible investing has developed immensely during the last decade. The outperformance of sin stocks should deter investors from negative screening since it would exclude these stocks from their portfolios. Neither the financial theory nor em- pirical results support negative screening, but still it is the most used form of responsible investing. However, increasingly institutional investors are viewing that negative screen- ing is the weakest form of responsible investing performance-wise (Amel-Zadeh & Ser- afeim, 2018). Responsible investors could increase their performance by adopting, for example, another screening technique. 2.4.3 Individual ESG factors The previous studies investigated the effect of overall ESG scores on portfolios. The fol- lowing papers examine the performance of portfolios formed on individual ESG factors. According to Borgers, Derwall, Koedijk and ter Horst (2013) most asset managers justify their socially responsible investments based on the argument that the implementation of ESG factors into investment processes provides positive effects to the investment per- formance. However, the evidence on the effect of ESG scores on portfolio performance 18 is not conclusive. For example, during the 1990s and early 2000s the use of governance scores generated abnormal returns, but since then the effect seems to have disappeared. The evidence of the effect of use of environmental screens on investment performance is mixed. Derwall, Günster, Bauer and Koedijk (2005) construct and evaluate two equity portfolios based on Innovest Strategic Value Advisors’ corporate eco-efficiency scores. They find that the more eco-efficient portfolio generated higher average returns and that it could not be explained by differences in beta, investment style or industry character- istics. Though Derwall et al. (2005) find a positive relationship between corporate eco- efficiency and future returns, Kempf and Osthoff (2005) and Statman and Glushkov (2009) find no evidence of outperformance due to environmental screening using KLD ratings data. Different data sources have different results yet again, as in the study by Halbritter and Dorfleitner (2015). In addition to environmental screens, the use of social screens has been researched as well. Unlike the performance of environmental screening, social screens have been found to generate high returns from 1984 to 2011, especially using employee satisfaction as a factor. Edmans (2011) investigates the relationship between employee satisfaction and long-run stock returns. Firms with high employee satisfaction earned an annual 4- factor alpha of 3.5% from 1984 to 2009. Edmans (2012) has similar results with an ex- tended sample period. The results are robust to controls for firm characteristics, differ- ent weighting methodologies and the removal of outliers. Employee satisfaction is posi- tively correlated with stockholder returns. Also, it seems that the market does not fully value intangibles which implies that social screens can improve the performance of in- vestors. The results of Edmans (2011 & 2012) are consistent with Kempf and Osthoff (2005) and Statman and Glushkov (2009) who found that investing based on KLD scores of employee relations and community can lead to high returns. In addition to social screens, governance scores have also resulted in abnormal returns. Gompers, Ishii and Metrick (2003) construct a “Governance Index” based on 24 19 corporate governance provisions for a sample of 1,500 large U.S. firms from 1990 to 1999. This G-index proxies for the strength of shareholder rights. By taking a long-horizon ap- proach as opposed to event-study methodology, they form long-short portfolios based on the G-index. Firms with strong corporate governance have a low G-index while firms with weak shareholder rights have higher G-index. The portfolio is long 10% of the lowest G-index companies while short the highest G-index firms. The authors find that corpo- rate governance has a strong positive relationship with stock returns. The long-short portfolio generated an abnormal return of 8.5% per year. The relationship between corporate governance and equity prices that Gompers et al. (2003) found in the period of 1990 to 1999 seems to have disappeared during the 2000s. Bebchuk, Cohen and Wang (2013) argue that the disappearance of the relationship is due to the markets learning about the correlation and reflecting the differences between good governance and poor governance companies into their prices. Although the G-in- dex was not able produce abnormal returns during the period of 2000 to 2008, good corporate governance was still reflected in the higher valuation, profitability and growth of these companies. Gu and Hackbarth (2013) find that the relationship between corporate governance and stock returns is positive and significant for transparent firms whereas the effect is small and insignificant for firms with opaque governance. Their data set covers 2,959 compa- nies during 1990 to 2006, which overlaps the data set of Bebchuk et al. (2013). Gu and Hackbarth (2013) argue that governance and transparency complement each other. The view that transparent firms are more valuable takeover targets is supported by the re- sults of the authors. Acquirers are able to identify synergies and bid more effectively on transparent firms. It seems that investors were able to use ESG screens to generate higher future returns until 2004 but since then the positive effect of implementing ESG scores seems to have disappeared. Although investors should not expect outperformance using ESG ratings 20 and governance information, they should not disregard ESG factors in portfolio construc- tion as there is no evidence of SRI underperforming relative to conventional investments. 2.4.4 Socially responsible mutual funds and indices Another way to measure the performance of socially responsible investing is to research how SRI mutual funds and indices perform compared to their conventional peers, rather than studying how portfolios composed of individual responsible companies perform. Some research suggests that socially responsible funds achieve outperformance com- pared to conventional funds. However, these results are often not significant. One of the earliest studies concerning the performance of SR funds is by Hamilton et al. (1993). They have three hypotheses about the relative performance of socially respon- sible mutual funds and conventional funds. The first states that SR funds have risk-ad- justed expected returns that are equal to those of conventional portfolios. Thus, social responsibility of stocks is not priced by investors. The second one has it that the expected returns of SR funds are lower than the expected returns of conventional funds. In this case socially responsible investors have a positive impact on stock prices and thus low- ering the expected returns of socially responsible companies. The last hypothesis states that the expected returns of socially responsible mutual funds are higher than those of their conventional peers which is possible if many investors consistently misprice the news of corporate social responsibility. (Hamilton et al., 1993.) Climent and Soriano (2011) find that by using a matched-pair analysis, U.S. green funds performed poorly compared to conventional mutual funds during 1987 to 2009 (8.45% vs. 12.67%), whereas on a more recent period, 2001 to 2009, green funds were able achieve similar returns to conventional funds. Although green funds underperformed conventional funds, and thus supporting the second hypothesis of Hamilton et al. (1993), they were able to achieve better returns than SRI funds (8.45% vs. 7.19%). During the period, green funds were riskier than SRI funds and conventional funds when measuring 21 risk by volatility – green funds had an annual volatility of 17.56%, whereas the volatilities of SRI funds and conventional funds were 13.79% and 15.05%, respectively. They also found that green funds were more sensitive to market as they had higher betas, 0.99 while SRI funds had the lowest betas, 0.84. The lower performance of green funds in the first sub-period could be explained by their more restricted investment set or by poor investment selection process. Investors seem to be more willing to pay for green invest- ment products in the form of lower returns compared to conventional investments. (Climent & Soriano, 2011.) Statman (2000) finds that neither SRI funds nor conventional mutual funds of equal asset size were able to achieve better results than the S&P 500 Index. On average, SRI funds trailed the S&P 500 Index by -5.02 percentage points while conventional funds trailed by -7.45 percentage points. Although SR mutual funds were able to achieve better perfor- mance than conventional funds, the difference was not statistically significant. However, Statman (2000) finds that Domini Social Index, a capitalization-weighted index modelled on the S&P 500 Index, consisting of 400 stocks was able to achieve better raw returns and risk-adjusted returns than S&P 500. The results from the comparison between SR and conventional mutual funds support the first hypothesis of Hamilton et al. (1993), while the outperformance by the socially responsible index supports the claim that do- ing well by doing good is possible. Using a sample of 89 Australian SRI funds from 1986 to 2005, Jones et al. (2008) state that ethical mutual funds underperform their benchmarks. This underperformance is most considerable during 2000 to 2005. The risk-adjusted returns show that annual un- derperformance by SRI funds was 1.52% per annum. They conclude that investing in portfolios that are constrained by social, environmental and ethical criteria causes a fi- nancial sacrifice for investors. Renneboog et al. (2008) find similar results. The authors apply a multi-factor model to figure out whether investors pay the price when investing in SRI funds or do they get value for their money. Using a data of 440 active and dead equity mutual funds from continental Europe, the US, the UK, Canada and Asia-Pacific, 22 they conclude that SRI funds underperformed their benchmarks by -2.2% to -6.5%. Ex- cluding some exceptions (France, Japan, Sweden), the risk-adjusted returns of SRI funds were not statistically different from those of conventional funds. They find that fund re- turns tend to decrease with screening intensity on social and corporate governance cri- teria. Unlike in the case of conventional funds, the size of the fund does not decrease the performance of SRI funds. The literature suggests that there is no consensus on the effect of social responsibility on mutual fund performance. However, it should be noted that these studies employ different data sets, periods and even asset pricing models, so the comparison between these studies is hard. Thus, both proponents and opponents of socially responsible mu- tual funds can both find studies to back up their views on socially responsible investing. 2.4.4 ESG events Event studies are another method of studying the value of corporate social responsibility. If investors view ESG information as material, corporate efforts in CSR should be ob- served in event studies. Historically, investments in environmental technologies have been viewed as a cost to the firm, although, there is also support that improved corpo- rate environmental responsibility can enhance financial returns (Fisher-Vanden & Thor- burn, 2011). They state that engagement in voluntary environmentally responsible ac- tivities is a fast-growing trend in the corporate world. These activities include member- ship in public voluntary programs that encourage voluntary public disclosure of environ- mental performance measures and pollution reductions. Fisher-Vanden and Thorburn (2011) are motivated to examine whether better environmental performance is con- veyed into better financial performance. The authors study the abnormal stock returns using event study methodology surrounding announcements to join two voluntary envi- ronmental programs, the US Environmental Protection Agency’s Climate Leaders pro- gram and Ceres. The Climate Leaders program targets reductions in greenhouse gas emissions while Ceres involves more general environmental commitments. 23 Their sample consists of 117 announcements over 1993 to 2008. The authors find signif- icant losses of 1% in the market value of firms following announcements to join the Cli- mate Leaders program. For firms joining Ceres the abnormal returns are insignificant. Fisher-Vanden and Thorburn (2011) find that the stock price declines are larger for com- panies exhibiting poor corporate governance and that companies with weak corporate governance structures are more likely to join Climate Leaders. The authors argue that these companies join Climate Leaders despite the fact that it lowers shareholder value because they face more pressure from institutional investors or because the managers are not overseen properly by the owners allowing them to join these environmentally conscious programs. Krüger (2015) studies the shareholder value implications of 2,116 corporate events for the companies’ main stakeholders. He shows that investors have a significantly negative response to negative CSR news and that this reaction is especially notable for infor- mation concerning communities and the environment. Krüger (2015) argues that stock price declines following negative CSR news are consistent with the position that social responsibility is associated with significant costs. The median cost of negative CSR news is approximately USD 76 million. It is calculated as the product of the median sample market capitalization and the median 21-day cumulative abnormal return. Not only are negative CSR news met with negative abnormal returns, as the release of positive CSR news cause slight negative stock performance as well. As in the case of negative news, positive news are met with higher stock declines in the case of news concerning the environment and communities. Although positive news are not received well by inves- tors either, the abnormal returns are only slightly negative. Still, Krüger (2015) suggests that investors do not appreciate the implementation of CSR policies. It seems that investors have negative views on CSR as a whole. Positive news are met with stock declines the same as negative news, although positive news are met with slightly lower stock price declines. Investors view corporate social responsibility as a 24 expenses to a company. Negative CSR news often result in expenses such as fines and reparations whereas investors might not fully understand the benefits of incorporating social responsibility into a company’s actions. 2.4.5 Active Ownership Although most research on responsible investing assumes that investors are active in their stock selection process and passive thereafter as owners who do not try to engage with the management directly. However, some socially responsible investors act as ac- tive owners by trying to influence management behavior and by exercising their rights as owners. Dimson et al. (2015) argue that despite growing interest in active ownership, data limi- tations restrict the ability to answer which firms are the target of active engagement, and how these engagements are carried out. By using an extensive proprietary database consisting of CSR engagements with US public companies from 1999 to 2009 provided by a large institutional investor, they find that active ownership concerning ESG issues can be value enhancing. The sample of Dimson et al. (2015) consists of 382 successful and 1,770 unsuccessful engagement for 613 companies. Market reaction to ESG activism is positive as these engagements generate a cumulative size-adjusted abnormal returns of 2.3% over the year following the initial engagement. The results for successful engage- ments are even more promising, as they find that cumulative abnormal returns for them are 7.1%. 2.4.6 SRI performance during market crises According to Guiso et al. (2008) investors factor in the risk of being cheated when decid- ing whether to enter the stock market. They find evidence that lack of trust is an im- portant factor in explaining the limited participation in the stock market. Investment in 25 stocks requires an assessment of the risk-return trade-off and sufficient data to complete this assessment. However, another important factor is the trust that the data are reliable, and that the system is fair. The authors find that trust has a large positive effect on stock market participation, whereas lack of trust reduces the demand for stocks. (Guiso et al., 2008.) Their findings can also explain some of the SRI performance during bear markets and market crashes. Lins et al. (2017) study the value of corporate social responsibility during the 2008 – 2009 financial crisis by examining the performance of 1,673 nonfinancial firms with CSR data. The authors find that firms with high CSR ratings had significantly higher – between four and seven percentage points – stock returns during the crisis than those firms that had low CSR ratings. The authors argue that the effect of social capital in explaining stock returns is at least half as large as the effect of financial variables, such as cash holdings and leverage. In addition, they tested the performance of high-CSR firms during non- crisis periods. Their models show, however, that CSR affects returns only during the crisis period. Even during the recovery period after the crisis there was no difference in the stock returns between high- and low-CSR firms. It seems that investors value the social capital of high-CSR firms during a crisis of trust which translates into outperformance during crisis periods. Investment in a high-CSR firm can be viewed as a sort of an insur- ance policy that pays off when the economy suffers from severe lack of confidence. (Lins et al., 2017.) Nofsinger and Varma (2014) find similar results to Lins et al. (2017) in that socially re- sponsible investments add value by outperforming during periods of market crises. In crisis periods SRI funds outperform by 1.61 – 1.75%. Here, the selection of the factor model makes a difference. However, the authors find that this outperformance comes at a cost of underperformance during non-crisis periods whereas Lins et al. (2017) did not find any evidence of underperformance after the crisis period. In non-crisis periods re- sponsible funds underperform by 0.67 – 0.95 %. Nofsinger and Varma (2014) state that this asymmetric return pattern is especially found in ESG funds that use positive screen- ing whereas SRI funds focusing on sin stocks or funds that focus on religious principles 26 do not outperform in crisis periods. This makes the case for advocating positive screen- ing over exclusion even more compelling. Kim et al. (2014) offer another way of studying the relationship between corporate social responsibility and stock price crash risk. Crash is defined as the conditional skewness of return distribution and it captures asymmetry in risk. Higher standard of transparency and lower engagement in hoarding bad news by socially responsible firms is hypothe- sized to lead to lower crash price. By using a large sample of US public firms from 1995 to 2009 they find that CSR performance has a negative relationship with future crash risk after controlling for other predictors of stock price crash risk. CSR has the most powerful mitigating effect on firms with lower standards of corporate governance. Companies with strong CSR focus are less likely to conceal bad news, which leads to lower stock price crash risk. (Kim et al., 2014.) Trust is essential in financial markets and its benefits can be seen from the performance of ESG investing during market conditions when overall trust in the markets is low. In- vestors trust the data coming from socially responsible companies and thus they are not overreacting to bad news. Although the literature supports the view that socially respon- sible investments may provide some protection during market downturns, the literature is not unanimous on the outperformance during non-crisis periods. 2.5 CSR and financial performance Growing interest in ESG issues has increased the amount of research on the relationship between CSR and financial performance. One of the only areas of agreement in the ESG literature is about the effects of CSR on the cost of capital. Companies with higher ESG scores tend to have lower cost of equity capital, higher credit ratings and be able to bor- row with lower interest rates. Socially responsible companies are thus viewed as less risky by debt and equity investors which in turn implies that the expected returns of these companies are also lower. 27 2.5.1 Cost of capital Using a sample of 12,915 US firm-year observations from 1992 to 2007, El Ghoul et al. (2011) examine whether CSR affects the cost of equity capital. They argue that if socially responsible firms are viewed as less risky, then these firms should have lower cost of equity. Higher investor base should also lower the cost of equity capital for these firms. El Ghoul et al. (2011) find that after controlling for other firm-specific determinants as well as industry and year fixed-effects, firms with higher CSR scores have lower cost of equity capital. Thus, managers should pursue CSR-related activities as they lower financ- ing costs as well as benefit the society at large. By being good corporate citizens compa- nies can attract a larger investor base and further decrease their cost of equity. They conclude that sin stocks related to the tobacco and nuclear power industries exhibit sig- nificantly higher costs of equity capital. Thus, they confirm the findings of Hong and Kacperczyk (2009) that sin stocks have higher expected returns. Chava (2014) studies the effect of a firm’s environmental profile on its cost of equity and debt capital. The implied cost of capital, derived from analysts’ earnings estimates, is significantly higher on stocks excluded by environmental screens compared to non- screened companies. Not only does CSR influence cost of equity but cost of debt is af- fected as well as firms with environmental concerns are charged a significantly higher interest rates on their bank loans. These companies are also avoided by institutional in- vestors and banks, as Chava (2014) finds that their institutional ownership is lower and fewer banks are ready to loan them funds. Goss and Roberts (2011) study the effect of CSR on the cost of bank loans. Their sample consists of 3,996 loans to US firms. They have similar results to Chava (2014) in that companies with below average ESG records pay between 7 and 18 basis points more than socially responsible firms. In addition to Goss and Roberts (2011) and Chava (2014), 28 Sharfman and Fernando (2008) find that firms with better environmental risk manage- ment exhibit lower costs of capital. Investors view that lacking concerns toward CSR, especially environmental issues, can cause severe risks for a company. Corporate managements should consider involving ESG issues into their processes in order to lower the cost of capital. Companies which ignore corporate social responsibility face higher cost of capital and may be disregarded by in- stitutional investors. As responsible investing and sustainable banking grow, poor corpo- rate citizenship will be penalized which will further increase the cost of capital for these companies. 2.5.2 CSR and accounting performance While partaking in corporate social responsibility results in costs for the company, recent research suggests that ESG activities and accounting performance have a positive rela- tionship. The benefits of CSR outweigh the costs of it in the form of higher profitability. According to Kim and Statman (2012) proponents of corporate environmental responsi- bility (CER) claim that companies can improve their financial performance by increasing their investment in CER. Opponents of this view state that by reducing their investment in environmental responsibility, companies can have better financial performance. The third view on this topic is that CER investments by companies are increased when it im- proves financial performance, and these investments are reduced when a decrease leads to a higher level of financial performance. The results of Kim and Statman (2012) suggest that corporations adjust their investments in CER to maximize profits. Using KLD data from 1991 to 2000 they find that corporations which increased their level of CER and lowered it had subsequent increases in their profitability, as measured by return on as- sets, while companies which did not change their levels of CER had significantly lower profitability. 29 Barnett and Salomon (2012) hypothesize that the relationship between CSR and corpo- rate financial performance (CFP) is U-shaped. Moderate level of corporate social perfor- mance (CSP) is worse for financial performance than low or high level of CSP. By using an unbalanced panel of 1,214 firms and 4,730 firm-year observations from 1998 to 2006 and after controlling for a variety of firm, industry and year effects, the authors find sup- port for a U-shaped relationship. When a company’s overall KLD net score increases, its ROA and net income have an initial decline, and an increase thereafter. However, Barnett and Salomon (2012) continue that this relationship is not symmetrical, as those firms with the highest net KLD scores had significantly higher ROAs than those with the lowest responsibility scores. Jiao (2010) studies the effect of stakeholder welfare on firm valuation using KLD data. The final sample consists of 4,027 observations for 822 firms. Although the focus of the study is the relationship between Tobin’s Q and a constructed stakeholder welfare score, Jiao (2010) offers insight into the relationship between accounting measures, mainly profit margin and return on assets, and the stakeholder welfare score. Univariate tests reveal that ROA is higher for high stakeholder welfare group than low stakeholder wel- fare group and that this difference is highly statistically significant. In the case of profit margin, Jiao (2010) finds that overall stakeholder welfare score indicates a higher profit margin, based on the results from two-stage least squares regressions. The results are similar for individual responsibility scores such as community relations, environmental performance and employee relations, where employee relations has the highest level of significance. 30 3 Theoretical framework of socially responsible investing Proponents of socially responsible investing argue that doing well by doing good is pos- sible. Moskowitz (1972) was one of the first to argue that the risk-return relationship is not the only aspect that explains portfolio performance. Socially aware investors state that corporate social performance has a positive effect on corporate financial perfor- mance. Literature endorses this statement as socially responsible companies are found to have higher profitability and lower cost of capital than conventional companies. SRI should perform better than conventional investments if investors misprice the infor- mation concerning CSR. If investors overweigh socially responsible investments their ex- pected returns should be lower in the future. Opponents of SRI view that it restricts the investment universe and consequently leads to a lower level of diversification and performance. This view is justified by the outper- formance of sin stocks. Socially responsible investors exclude these stocks from their portfolios which drives the expected return of sin stocks higher. However, portfolios us- ing screening techniques such as best-in-class are diversified across industries which mit- igates the diversification argument. Best-in-class has often generated the highest alpha of different screening techniques (e.g. Kempf & Osthoff 2007) whereas the use of exclu- sionary screens has been found not to compromise the returns (Hoepner & Schopohl 2018). 3.1 Corporate perspective The corporate perspective on social responsibility studies the effect of CSR on companies from the viewpoint of the company. The neoclassical view of ESG engagements is that they are a cost to shareholders by requiring the consumption of the corporation’s re- sources. According to Friedman’s (1970) shareholder model the only social responsibility of a company is the maximization of shareholder value. Stakeholder view takes other stakeholders, such as employees, customers and suppliers, into consideration in 31 corporate decision-making as well. According to Goodpaster (1991) the term “stake- holder” has been invented in the early 1960s to signify that there are other parties in addition to stockholders who have a stake in the modern publicly listed corporation. In addition to these opposite views, instrumental stakeholder theory tries to combine these views in that shareholders’ value is maximized in the long term by recognizing the interests of all stakeholders. 3.1.1 Shareholder, stakeholder and instrumental stakeholder Smith (2003) argues that the shareholder and stakeholder theories are normative theo- ries of corporate social responsibility. Not only do they dictate what a corporation’s role should be in the society, but they can also be seen as normative theories of business ethics – executives and managers should act accordingly to the “right” theory. The main difference between the two theories is that the stakeholder theory insists that managers take into consideration the interests of all stakeholders even if it reduces profitability of a company. (Smith 2003.) Dodd (1932) was one of the first to claim that the modern corporation should be involved in social responsibility in addition to its profit-making function, and that corporate man- agers should take the interests of other stakeholders into consideration. According to Dodd (1932) there was a growing feeling among business leaders that business should voluntarily take care of its responsibilities to the community already in the first half of the 20th century. Managers should not wait for legal compulsion regarding social respon- sibility. Dodd (1932) argues that when business managers take the welfare of its custom- ers and employees into consideration, it will in the long run increase the company’s prof- its. Lack of security felt by workers was largely responsible for the under-consumption during the Great Depression (Dodd 1932). Friedman (1970) argues that in a capitalistic system a corporate executive is an employee of the owners of the business. The sole responsibility of the executives is to their 32 employers. The criterion of performance is straight-forward but Friedman (1970) states that basic rules of the society steer the executives. While acting as an agent rather than as a principal, executives should be deterred from social responsibilities because they are spending the money and time of their employers (Friedman, 1970) – corporate social responsibility is seen as being caused by agency conflicts and moral hazard and it can cause agency costs. Executives should not reduce pollution beyond the amount that is optimal for the corporation or required by law. Friedman (1970) argues that this case would entail using the someone else’s money to achieve a social objective while reducing stockholder returns, implying that there is a trade-off between CSR activities and finan- cial performance. In effect, the executive is thus imposing taxes and deciding where to use the proceeds. According to Friedman (1970) taxation and expenditure of tax income are functions of a government whereas while the executive takes part in corporate social responsibility, she is simultaneously deciding whom to tax and where to use the pro- ceeds. Thus, there is no system of checks and balances overseeing the process. Friedman (1970) states that then the executive becomes a public employee, even though she is still an employee of a private enterprise. Furthermore, Friedman (1970) brings up that corporate social responsibility is often used as a justification for actions rather than a reason for those actions. He makes an argu- ment that, for example, it may be in the long-run interests of a corporation that is a major employer in a community to improve the community and offer it amenities. This is done entirely of its own self-interest although it could be rationalized as social respon- sibility. Therefore, Friedman does not prohibit companies from social responsibility – but it should be a byproduct of the company’s actions, not the goal. Although CSR is often viewed as incompatible with the view of shareholder value maximization, Friedman does not forbid it – as long as it benefits the shareholders. One of the main opponents of a multi-fiduciary stakeholder approach, Goodpaster (1991), states that the relationship between the management and stakeholders is ethi- cally different from the relationship between the management and the stockholders. He 33 claims that the management has nonfiduciary duties to other stakeholders whereas the fiduciary duty of a corporation is to its shareholders alone. Goodpaster (1991) argues that this fact is not considered by the stakeholder theory. He emphasizes that although the relationship is different between stakeholders and managers, the other parties do not lack a morally significant relationship to the management. Goodpaster (1991 contin- ues that this only means that the relationship is not fiduciary in nature. According to Smith (2003) the shareholder theory is often misrepresented. Sometimes it is considered as urging managers and executives to make profit by any means. The theory is also criticized as weighting short-term term profit maximization more than the long-run benefits. Still, Friedman (1970) emphasizes that short-term profit maximization should be foregone if it impedes the long-term shareholder value maximization. The stockholder theory of Friedman (1970) is challenged by the stakeholder theory of Freeman (1984). The stakeholder theory offers an alternative to the narrow view of Friedman by stating that corporate success should be viewed more broadly through the lens of sustainable growth and the consideration of ESG issues. Smith (2003) states that according to the stakeholder theory managers have a dual duty to both the corporation’s shareholders as well as the stakeholders. However, there is ambiguity regarding which stakeholders are to be considered, but most interpretations refer to at least stockholders, customers, employees, suppliers and the local community. The theory states that man- agers are agents of all stakeholders, not only of stockholders. The objective of the man- agers is to maximize profit while ensuring the long-term ability of the corporation to remain a going concern. (Smith 2003.) According to Freeman et al. (2004) the main assumption of the stakeholder theory is that values are a necessary part of doing business. The theory asks the managers to be clear about what brings its core stakeholders together and how the managers want to do business. The authors argue that this shared sense of the value the stakeholders cre- ate brings them together and allows the company to generate excellent financial 34 performance as well as in terms of its purpose. By developing relationships and building communities, the managers are able to best deliver the value the firm promises. Con- trary to shareholder theory, profits are the result from the value creating process rather than the driver – doing well by doing good. (Freeman et. al 2004.) In addition, risk man- agement should incorporate environmental and social risks beside merely financial risks. Just like the stockholder theory, the stakeholder theory is sometimes misunderstood. Sometimes it is claimed that the theory does not emphasize focus on company profita- bility (Smith 2003). However, the author argues that the theory’s ultimate objective – the concern’s continued existence – is only achieved by balancing the interests of all stakeholders, whose interests are usually satisfied through profits (Smith 2003). By adopting the stakeholder view, corporations would be able to observe their impact on the society in addition to financial aspects. Instrumental stakeholder theory views the shareholder view and the stakeholder view as mutually inclusive. The theory has it that corporate social responsibility has positive effects on corporate financial performance through, for example, reputational effects and employee satisfaction. Turban and Greening (1997) hypothesize that an organiza- tion’s corporate social performance may attract potential applicants by signaling on working conditions under incomplete information. They suggest that companies can achieve competitive advantage by attracting a larger applicant pool. Peterson (2004) ex- amines how the perceived CSP of an employer affects its employees. The results show that the employer’s commitment to CSR is associated with higher organizational com- mitment – especially in the case of female employees. Lacey and Kennett-Hensel (2010) also find that CSR initiatives can build trust between a firm and its customers and thus lead to committed customer relationships. Not only does superior corporate social per- formance benefit the stakeholders of a company, but it can also lead to a competitive advantage to the firm resulting in better financial performance. 35 3.1.2 Risk mitigation Not only does CSR affect the financial performance of companies, but it can also be ex- amined from the viewpoint of risk mitigation. By improving corporate social perfor- mance companies can reduce their sensitivity to negative news, business cycles and eco- nomic shocks. Godfrey et al. (2009) hypothesize that when facing a negative event, the decline in shareholder value is smaller in the case of high CSP companies. This insurance- like property of CSR stems from moral capital. Companies which exhibit low levels of moral capital are punished by stakeholders when negative news emerge. By implement- ing event-study methodology, Godfrey et al. (2009) find that CSR participation has an insurance-like effect when a company faces negative firm-specific news. Oikonomou et al. (2012) hypothesize that CSP influences negatively market risk at the firm level. They find that there is a negative but insignificant relationship between CSP and systematic financial risk. However, when investigating the relationship during times of low volatility, high CSR companies exhibit lower levels of market risk, and during times of high volatility, socially irresponsible companies are characterized by higher levels of systematic risk. While the shareholder viewpoint sees CSR as an agency cost to shareholders, doing well by doing good, as suggested by the stakeholder theory, is possible through the positive effects of corporate social performance on corporate financial performance. When tak- ing the positive effects and risk mitigative properties of CSR into account, it seems that managers do a disservice to the company when they fail to incorporate social responsi- bility and stakeholder approach to their decision-making processes. 3.2 Investor perspective This section summarizes the theorical background of risk, return and valuation of re- sponsible investments and provides another way to view the effect of social responsibil- ity in finance, is through the perspective of an investor. The classical financial theory views that capital markets are efficient and that the portfolios of investors are well- 36 diversified regarding firm-specific risks. Thus, the exclusion of stocks with higher ex- pected returns due to low ESG performance should lead to lower returns for the investor. Socially responsible portfolios are then under-diversified due to an additional constraint in mean-variance portfolio optimization models. If corporate social responsibility influ- ences corporate financial performance or firm risk, ESG information should be incorpo- rated into investment decision-making – this information is considered in the risk-return relationship of an investment. This viewpoint takes the materiality of ESG information into consideration. The degree to which investors take these non-financial issues into consideration affects the performance of responsible investments. Investors should also consider the risk mitigative properties of responsible investing as some views suggest that responsible investment experience lower tail-risk. According to the efficient market hypothesis (EMH), stocks already reflect all available information (Fama 1970). Abnormal returns, or alpha generation, are not possible since all information is public, and thus available for everyone. According to this hypothesis, higher returns are only achieved when higher risks are taken. If high-ESG stocks have higher returns, it is due to them having higher systematic risk, and vice versa. There are three versions of the EMH, and they differ by their definition of “all available information.” The weak form hypothesis states that “all available information” refers to information that is uncovered by investigating market trading data such as historical stock prices and trading volumes. According to the weak-form, technical analysis is point- less. The semistrong form maintains that all public information, including fundamental information on management’s quality, earnings forecasts and dividend payments, is re- flected in stock prices. The third version of the EMH is the strong form. It states that all information, including insider information, is already priced in the markets. (Fama 1970.) Active portfolio management, including fundamental analysis and trend following, is a wasted effort according to the proponents of the efficient market hypothesis. Instead, they advocate passive investing by following the market. On a risk-adjusted basis, active 37 mutual funds should not be able outperform the market, especially net of costs. Thus, it should be expected that the conventional and responsible mutual funds examined in this thesis should have statistically insignificant alphas. Modern portfolio theory asserts that in the equilibrium an asset’s expected excess return depends on its exposure to systematic risk, or beta (Lintner 1965). If investors view ESG information as material to assessment of systematic risk or future cash flows, then ESG information will be factored in asset prices immediately, if the markets are efficient (Fama 1970). However, Merton (1987) argues that financial markets may exhibit anom- alies due to incomplete information. If investors’ opinions on, for example, ESG infor- mation differ, Fama and French (2007) state that informed investors generate positive alpha, while misinformed have negative alphas. Informed investors also make asset prices more rational. If investors have a neoclassical view – that CSR is considered an agency cost – then ESG news should elicit a negative reaction. If investors believe that CSR engagements are beneficiary for corporate financial performance, then this kind of news should have a positive stock price reaction. Either way, ESG information expands the information set of investors, which then can be used in the investment decision-making process and in- formed investors can benefit from adapting to this new material information, thus pric- ing ESG information. However, if investors as a whole disregard information concerning social, environmental and governance news, then markets are not efficient, and inves- tors are subject to systematic risk that they are not considering. In this case, investors underestimate the effects of global issues, such as climate change, on systematic risk. Investors may also ignore the unsystematic risks of ESG issues. Instead of solely focusing on systematic risk, investors should take total risk of a stock into consideration. Total risk is based upon a stock’s systematic risk and firm-specific risk. In efficient markets investors assumed to optimize their portfolios using Markowitz’ (1952) expected returns - variance of returns rule. Using this rule investors can eliminate 38 firm-specific risk by diversifying their portfolios. Investors can reduce the total risk of their investments by holding a portfolio of different securities instead of just placing their money in one type of assets. A limited investment universe means limited diversification opportunities, which should ultimately lead to lower risk-adjusted returns. When this is implemented to negative screening, this additional constraint on the optimization should lead to worse performance. These value-based views undermine the effect of diversification by excluding stocks with higher expected returns, such as sin stocks (Hong & Kazperczyk, 2009). These non-responsible stocks are left out of the portfolio and thus the portfolio’s expected return is worse than that of a non-constrained one (Fama & French, 2007). In empirical studies, such as Kempf and Osthoff (2007), this can be seen as the underperformance of negative screens. However, when investigating SRI funds as a whole, they seem to achieve similar to returns to conventional funds – the perfor- mance of SRI funds cannot be told apart from that of their conventional peers. But nei- ther class of funds do outperform their benchmarks. Recent studies have started implementing only material ESG ratings when investigating the performance of responsible investing. Khan et al. (2015) compare the returns of two stock portfolios. The first consists of companies which rate high on ESG issues which are deemed material for its sector. The second is formed from companies which have low ESG ratings. The first portfolio significantly outperforms the second one, which indicates that investors should pay attention to material ESG issues. Material ESG factors can affect future cash flows or risks which are not reflected in current market prices. Informed in- vestors can thus generate abnormal returns and benefit from the higher risk-return re- lationship through mispricing of growth potential and future cash flows and risks by the market. Pedersen et al. (2019) view that there are two lines of literature concerning the perfor- mance of responsible investing. The first states that ESG decreases investment perfor- mance and it is supported by empirical evidence of outperformance by “sin stocks” (Hong & Kacperzcyk, 2009). The second line shows that stocks with good governance 39 (Gompers et al., 2003) as well as stocks with high employee satisfaction (Edmans, 2011; Edmans, 2012) show sustainable abnormal returns. Pedersen et al. (2019) state that if ESG predicts future firm profitability, then it should also forecast high expected returns if the market does not price ESG information correctly. However, this requires that ESG also predicts investor demand. The authors propose theory which explains two effects of ESG scores of stocks. First, ESG scores provide information on a firm’s fundamentals. The second part of the theory explains how ESG scores affect investor preferences. Pedersen et al. (2019) consider three types of investors based on their view of ESG scores. Investors of the first type, type-U, are unaware of ESG scores and their goal is to maxim- ize their mean-variance utility. The second group of investors, type-A, are ESG-aware who also have M-V preferences, but they implement ESG scores in their investment de- cision making – ESG affects their view on risk and expected return. The last type, type- M, are ESG-motivated. They use ESG information and prefer high ESG score stocks. Figure 1 shows an ESG-Sharpe ratio frontier (A) constructed by the authors by computing the highest attainable Sharpe ratio (SR) for each level of ESG. Type-A investors choose the “tangency portfolio using ESG information” – they maximize their SR. Point B char- acterizes this tangency portfolio and it maximizes Sharpe ratio. ESG-motivated type-M investors choose their portfolio right of this tangency portfolio from the ESG-efficient frontier, point C, or the area left of B. The ESG-unaware investors choose a portfolio be- low the ESG-SR frontier (point D), as they ignore the information provided by the ESG scores. Point E expresses individual assets. Pedersen et al. (2019) also provide expected returns given by an ESG-adjusted CAPM (Figure 2). When many investors are type-U, and assuming that high ESG scores predict high expected returns, the figure shows that high-ESG stocks earn high expected excess returns. The prices of profitable high-ESG stocks are not driven up by type-U investors, which explains the high expected returns. In an opposite situation the expected returns of ESG-motivated investors are low, if these investors form the majority. In the third case 40 ESG-aware investors bid up the prices of high-ESG stocks to a level which reflects their expected returns. Thus, the relationship between ESG scores and expected returns dis- appears. Figure 1. ESG-Sharpe ratio frontier, adapted from Pedersen et al. (2019). Figure 2. ESG-CAPM, adapted from Pedersen et al. (2019). 41 Giese et al. (2019) use discounted cash flow framework to help explain the effect of ESG factors on a company’s equity valuation by examining future cash flows, risk and cost of capital. DCF models calculate the fundamental value of a company as the present value of its future cash flows which are discounted at an appropriate cost of capital: 𝑃𝑉 = 𝛴 𝐶𝐹𝑡(1+𝑟)𝑡 , (1) where PV is the present value of the company, CFt is a cash flow at time t and r is the cost of capital. Gregory (2014) state that cost of equity depends on risk and that it is important to sep- arate firm-specific risk and systematic risk. They continue that systematic risk is often macro-economic in nature – economic growth rate shocks, interest rates, and oil price shocks all affect the majority of stocks. They contrast systematic risk with firm-specific risk which is particular to a company. This distinction is crucial for investors since inves- tors can eliminate firm-specific risk via diversification. The required rate of return for investors, or cost of capital, is then determined by systematic risk. (Gregory et al., 2014.) Giese et al. (2019) argue that firm-specific risk is considered in the future cash flows of a company whereas systematic risk is considered while computing the cost of capital in a DCF model – markets are not indifferent to firm-specific risk. Giese et al. (2019) distinguish three transmission channels which can explain how ESG characteristics influence companies. Two of the transmission channels transmit through idiosyncratic risk – the transmission of ESG into future cash flows and the transmission to firm-specific downside risk protection. The last remaining channel shows the effect of ESG on company valuation though systematic risk. Giese et al. (2019) summarize the first idiosyncratic transmission channel from the work of Gregory et al. (2014) as follows (Fig- ure 3). Companies with high ESG scores have a competitive advantage through efficient use of resources, better human capital development or better innovation management. These companies use their competitive advantage to achieve higher profitability. Last, 42 this higher profitability leads to higher dividends. Ceteris paribus, this leads to higher investment returns. Figure 3. Cash-flow transmission channel (Giese et al. 2019). According to Giese et al. (2019) the second transmission channel connects strong ESG characteristics with lower tail risk (Figure 4). Responsible companies have better risk management standards and they can reduce risk through CSR engagements (Jo & Na 2012). These companies are not as affected by negative news as low ESG-rated peers (Hong & Kacperczyk, 2009). These incidents can have seriously detrimental effects on company value. Thus, Giese et al. (2019) argue that less-frequent risk incidents should reduce tail risk of a company’s stock price. Figure 4. Idiosyncratic risk channel (Giese et al. 2019). The last transmission channel explains how high ESG performance leads to higher valu- ations (Figure 5). Using a CAPM framework investors can calculate a required rate of return for a stock: E(Ri) = rf + βi [E(Rm) – rf], (2) where E(Ri) is the expected return of a stock i, rf is the risk-free rate, βi is the market sensitivity of a stock and E(Rm) is the expected market return. Higher systematic risk, represented by market sensitivity of a stock, leads to higher rate of return required by investors and vice versa. As Gregory et al. (2014) argue that responsible companies show 43 lower systematic risk then investors require a lower rate of return. These companies have then a lower cost of capital which in a DCF model leads to a higher valuation (Giese et al., 2019). Figure 5. Valuation channel, (Giese et al. 2019). If high ESG performance is related to lower tail risk, investors may be willing to pay a premium for responsible companies during good times to outperform during market crashes. Moskowitz (2000) states that active mutual funds appear to outperform during recessions – when investors care about performance the most. Moreover, according to the prospect theory by Kahneman and Tversky (1979), investors are more affected by losses negatively than by a gain of the same size positively. Thus, investors prefer a port- folio with asymmetric performance, because they get a higher utility by outperforming during market drawdown than what they lose underperforming during a bull market. If responsible investing can mitigate the size of a drawdown of a portfolio, then according to the prospect theory investors can improve their utilities by investing responsibly. 44 4 Data and methodology The next sections of this thesis empirically investigate the performance of socially re- sponsible investing and the insurance-like features of these investments. Asset pricing literature has a long history of measuring mutual fund performance and thus this thesis follows earlier papers by implementing similar asset pricing models and other methods. Definitions for the data and methodology employed in this thesis are explained in the next subchapters. 4.1 Data description As this thesis compares the investment performance of socially responsible equity funds and that of their conventional peers, an original sample of 357 SRI funds and 500 con- ventional equity funds was collected from Thomson Reuters Eikon’s Datastream data- base. All of the funds were based in the US. The samples included price data from De- cember 1999 to October 2019 inception dates for these funds. SRI funds have distinct fund names which were used as search terms to conduct a list of responsible equity funds. These search terms were “responsible”, “environmental”, “ESG”, “ethical”, “so- cial”, “SRI”, and “sustainable”. After excluding duplicate funds and non-equity, such as fixed income, balanced and money-market mutual funds, the filtered sample includes 110 SRI funds and 120 non-SRI funds. The data was augmented with information on assets under management, expense ratios, turnover statistics and investment styles of these funds. This information was collected from the Morningstar mutual fund database. Table 1 presents these characteristics of the mutual funds used in this analysis. Responsible funds are younger than their conven- tional peers and have much less assets under management (AUM). The conventional funds include a few funds with several billions under management which skews the av- erage AUM. Although the difference is much smaller between the median AUMs, still the responsible funds are much smaller. The average expense ratio is a little higher for 45 responsible funds, but when measuring median expense ratios, responsible funds have slightly lower expense ratios. The same repeats with turnover ratios as responsible funds have higher average turnover ratios but lower median turnover ratios. This could be ex- plained by a higher amount of passively managed funds in the responsible fund sample. Table 1. Characteristics of SRI and non-SRI equity funds The fund database of Morningstar categorizes mutual funds into several investment styles. The nine investment-style classes form a nine-square grid by classifying a mutual fund by its target market capitalization and according to its emphasis on value or growth factors. This matrix is called Morningstar Style Box and it is depicted in Figure 6. On the vertical axis funds are defined into three size categories – small, mid-size and large. Morningstar (2008) defines large-cap stocks as the group that forms the top 70% of the capitalization of each geographic area. Mid-cap stocks are defined as the next 20% and small cap-stocks represent the last 10%. The portfolios of mutual funds define their des- ignation as large, mid or small-cap oriented funds. On the vertical axis funds are defined as growth, value or blend, which is composed of a mixture of growth and value stocks. Stocks are designated as growth or value by comparing them to other stocks of the same capitalization band and are then scored by their characteristics, which include both for- ward looking measures and historical-based accounting and valuation ratios. (Morn- ingstar, 2008.) No. Funds Mean age Mean AUM Mean expense ratio (%) Average turnover (%) SRI (1) 110 12.81 1081.42 1.06 51.97 Non-SRI (2) 120 19.44 16248.60 0.98 45.17 (1) - (2) -6.63 -15167.18 0.08 6.80 No. Funds Median age Median AUM Median expense ratio (%) Median turnover (%) SRI (1) 110 12.40 128.60 0.90 31.00 Non-SRI (2) 120 15.83 877.30 0.99 37.00 (1) - (2) -3.43 -748.70 -0.09 -6.00 46 Figure 6. Morningstar Style Box, adapted from Morningstar (2008). Table 2 compares the frequencies of different investment styles and target market capi- talizations of SRI and non-SRI mutual funds based on their designation by Morningstar Style Box. Responsible funds of the sample are not represented in the small-cap classes while 13 of the non-SRI funds are invested in small-cap stocks. The share of SRI funds invested in mid-cap stocks is also low compared to large-cap. Non-SRI funds are similarly more invested in mid-cap stocks than in small-cap stocks. However, both fund groups are mostly invested in large-cap stocks. SRI funds do not emphasize value stocks over growth stocks, since their investment style is mostly considered as blend. For non-SRI the distri- bution between value, blend and growth is much more balanced, except for large-cap stocks, where value stocks are underrepresented. Further analysis using multi-factor models also describes the underlying factor exposures of SRI and non-SRI funds. This analysis can be used to investigate the investment styles of SRI and non-SRI as entireties. Since this research studies the insurance-like properties of responsible investing, several periods of market crises, bear markets and corrections, are located. Table 3 represents the performance of S&P 500 index during the last two bear markets and all seven market corrections of the 2000s. Bear markets are defined as market drawdowns of at least 20% 47 and market corrections are defined as market decline of at least 10% from peak to through. Table 2. Summary of investment styles of SRI and non-SRI equity funds The first bear market of the sample, the Dot-Com bubble, took place at the beginning of the sample period. During the 30-month crisis period the index nearly halved. Five years later started the most recent bear market, the global financial crisis, during which S&P 500 declined by 57%. Both bear markets lasted for over a year and the recoveries from these two market crises took over four years. While bear markets are relatively uncom- mon, market corrections are typical market behavior and thus much more common. The first market correction took place right after the Dot-Com bubble and the most recent one between September 2018 and December 2018, during which the correction was almost designated as a bear market. On average, market corrections are much briefer than bear markets and the recoveries are similarly fast. Data on the excess market return, size, value and momentum factors is compiled from the Kenneth French Data Library to estimate the risk-adjusted returns for the two groups. The database was also used to acquire Treasury bill rates for monthly risk-free returns. SRI Non-SRI Difference Investment style Frequency Frequency Frequency Large Value 6 (9.0%) 10 (10.5%) -4 (-1.5%) Large Blend 38 (56.7%) 28 (29.5%) 10 (27.2%) Large Growth 14 (20.9%) 22 (23.2%) -8 (-2.3%) Mid Value 1 (1.5%) 8 (8.4%) -7 (-6.9%) Mid Blend 5 (7.5%) 6 (6.3%) -1 (1.2%) Mid Growth 3 (4.5%) 8 (8.4%) -5 (-3.9%) Small Value 0 (0.0%) 2 (2.1%) -2 (-2.1%) Small Blend 0 (0.0%) 5 (5.3%) -5 (-5.3%) Small Growth 0 (0.0%) 6 (6.3%) -6 (-6.3%) 48 Table 3. S&P 500 performance during market crises 4.2 Methodology The empirical analysis of the mutual funds consists of a descriptive part and of single- factor and multi-factor asset pricing models. To conduct the analyses two time-series are formed from the mutual fund data. The first consists the time-series returns of an equally weighted portfolio of the SRI funds and the second is formed of the conventional funds. Using these portfolios, average and median monthly returns and standard deviations of these returns are calculated and compared to a US benchmark – S&P 500 index – using also local risk-free interest rates. The descriptive part also calculates these metrics during two subsamples, 2000 to 2009 and 2010 to October 2019, and during the two bear mar- kets and recent corrections as well as during the recoveries of these bear markets. Investment alternatives should be compared by using risk-adjusted returns. This analysis uses three traditional asset pricing models to estimate risk-adjusted returns for both mu- tual fund groups. After estimating the single-factor model further controls are Beginning End Bear market / Correction Performance, % Length (months) Recovery (months) Mar 00 Oct 02 Bear market -49.2 % 30 52 Nov 02 Mar 03 Correction -14.7 % 3 2 Oct 07 Mar 09 Bear market -56.8 % 17 48 Apr 10 Jul 10 Correction -16.0 % 2 4 Apr 11 Oct 11 Correction -19.4 % 5 5 May 15 Aug 15 Correction -12.4 % 3 11 Nov 15 Feb 16 Correction -13.3 % 3 4 Jan 18 Feb 18 Correction -10.2 % 1 6 Sept 18 Dec 18 Correction -19.8 % 3 4 Average: Bear market -53.0 % 23.5 50 Correction -15.1 % 2.9 5.1 Median: Bear market -53.0 % 23.5 50.0 Correction -14.7 % 3.0 4.0 49 established by controlling for book-to-market, high-minus-low (HML), and size, small- minus-big (SMB), in addition to the market factor (MKT). The second multi-factor model controls for momentum (MOM) as well as HML and SMB. 4.2.1 CAPM Capital asset pricing model, CAPM, is often employed in measuring the performance of mutual funds, responsible or otherwise. The model derives the expected return of a se- curity. The single-factor model estimated here is of the following form: Rit – Rft = αi + βi (Rmt – Rft) + εit (3) where Rit is the return on portfolio i in month t, Rft is the one-month Treasury bill rate in month t, βi is the slope coefficient of the regression for portfolio i and Rmt – Rft is the excess return on the market, value-weighted return of all CRSP firms incorporated in the US and listed on the NYSE, AMEX, or NASDAQ, and ε is the error term. The intercept term, αi, of the model is often called Jensen’s alpha (Jensen, 1968). This is an indicator to whether a portfolio has underperformed or outperformed its benchmark. 4.2.2 Fama-French 3-factor model Fama and French (1993) present a three-factor model to extend the single-factor model. The model controls for the effect of investment style on fund performance. HML is the difference between the returns between value and growth stocks. Fama and French (1993) argue that high book-to-market (BM), outperform low book-to-market stocks. Thus, stocks with a higher exposure to the HML factor should have higher future returns. The third factor in the model controls for the small firm effect. SMB mimics the risk factor in returns related to size. It is estimated similarly to HML, but instead of estimating the 50 difference between high BM and low BM stocks, the monthly difference in question is between small cap and large cap stocks. Again, higher exposure to the SMB factor should lead to higher expected returns. The three-factor model is estimated as follows: Rit – Rft = αi + β1 (Rmt – Rft) + β2 SMB + β3 HML + εit, (4) where β1, β2 and β3 are slopes in the time-series regressions, Rmt – Rft is defined as above, SMB is the average return on the three small portfolios minus the average return on the three big portfolios, SMB = 1/3 (Small Value + Small Neutral + Small Growth) – 1/3 (Big Value + Big Neutral + Big Growth), and HML is the average return on the two value portfolios minus the average return on the two growth portfolios, HML = 1/2 (Small Value + Big Value) – 1/2 (Small Growth + Big Growth). 4.2.3 Carhart 4-factor model The third factor model is the Carhart four-factor model. Carhart (1997) extends the three-factor model with a momentum factor. Jegadeesh and Titman (1993) find that a strategy which buys recent winners and sells past losers generates significant positive returns over holding periods of three to 12 months. The model is estimated as follows: Rit – Rft = αi + β1 (Rmt – Rft) + β2 SMB + β3 HML + β4 MOM + εit, (5) where β1, β2, β3 and β4 are slopes in the time-series regressions, the three factors are defined as above and MOM is the average on the two high prior return portfolios mi