THE FINANCIALPERFORMANCE OF U.S.SOCIALLY RESPONSIBLEMUTUAL FUNDSA Study Performed during Normal and COVID-19 Crisis Times

(1)

T H E F I N A N C I A L

P E R F O R M A N C E O F U . S . S O C I A L L Y R E S P O N S I B L E

M U T U A L F U N D S

A Study Performed during

Normal and COVID-19 Crisis Times

M A Y 1 7 T H 2 0 2 1

MSc in Economics and Business Administration Finance and Investments

(Master's Thesis)

Carlotta Martinuzzi (S133869) Mille H. Jørgensen (S101968)

Supervisor: Cristiana Parisi

Number of Normal Pages: 107,6

Number of Physical Pages: 114

Number of Charaters: 244,753

(2)

Abstract

This thesis investigates U.S. socially responsible mutual funds’ financial performance as compared to that of U.S. conventional mutual funds. The sample period under investigation goes from the 1^st of January 2010 to the 31^st of December 2020, however subperiods covering normal times as well as the COVID-19 crisis market crash and recovery periods have also been considered in the analysis. Generally, this thesis finds that conventional mutual funds outperform their sustainable counterparts during the entire sample period, normal times, and the COVID-19 crisis recovery period. Contrarily, U.S. socially responsible mutual funds outperform conventional mutual funds during the COVID-19 market crash.

Additionally, this thesis investigates the relationship between socially responsible mutual funds’ screening processes and their financial performance over the same sample period specified above. It is concluded that this constitutes a curvilinear relationship, however, also that this relationship does not persist when particular investment areas are considered independently.

(6)

1. Introduction

Given the world’s recent subjection to the COVID-19 global crisis, the market has seen a dramatic downturn. This has affected companies in all markets across the globe but might have had a less powerful impact on one particular category. Specifically, it has previously been speculated and shown that sustainable companies generally face lower downside risk during economic crisis times (Nofsinger & Varma, 2014). Furthermore, it has also been hypothesized and found that sustainable companies can be a lower-risk investment providing non-significantly different returns from conventional funds during non-recession times too (Hamilton et al., 1993;

Sauer, 1997; Statman, 2000; Bauer et al., 2005, 2006, 2007). However, consensus on this matter has yet to be reached.

Thus, this thesis seeks to contribute to the existing literature on this topic by investigating the financial performance of U.S. socially responsible mutual funds¹ as compared to a matched portfolio of U.S. conventional mutual funds. This analysis is conducted over a sample period going from the 1^st of January 2010 to and including the 31^st of December 2020. Different subperiods are furthermore investigated, in order to establish the patterns of financial returns in both normal and COVID-19 crisis periods. Specifically, a normal period was identified as going from the 1^st of January to and including the 7^th of February 2020. Furthermore, the COVID-19 crisis period was split into two separate subperiods, whereof one covers the market crash (10^th of February 2020 to and including the 13^th of March 2020) while the other the recovery period after this initial crash (16^th of March 2020 to and including the 31^st of December 2020).

For this part of the analysis a matched pair analysis approach has been implemented, thereby assigning three conventional mutual funds to each ethical mutual fund based on characteristics of size, age and fund type. The two types of funds were regressed over the different time periods following either the CAPM, the Fama French three-factor or five-factor models.

Given the panel data set, random and fixed effects estimation approaches have been implemented.

The Hausmann test was here performed in order to choose between said methods.

On this matter, this thesis concludes that conventional mutual funds outperform their socially responsible counterparts during the entire sample period and during normal times. With

1 ”A service where financial experts invest the money of many people in many different companies” (Cambridge Dictionary, 2021)

(7)

respect to the COVID-19 crisis periods, the socially responsible mutual funds firstly outperform the conventional funds during the market crash but later underperform these counterparts during the recovery period.

In addition, this thesis investigates the relationship between socially responsible mutual funds’ screening processes and their financial performance. Firstly, screening intensities are considered in isolation without taking into account the specific investment area to which they are applied. Secondly, each screening process is further investigated within four specific areas of investment, namely environment, social, governance and products. This is a highly under- researched area in the existing literature, but highly relevant given that these screening processes are unique to socially responsible mutual funds, and thus could partly explain differences in returns between these and conventional counterparts. This analysis has been carried out over the same sample period and subperiods as specified above.

Similar to the method previously stated, the fixed and random effects estimation methods were utilized for the regressions. For this part of the analysis, however, new control variables were added to the CAPM, Fama French three-factor and five-factor models together with dummy variables for each screening process investigated.

In relation to this analysis, it is concluded that the relationship between socially responsible mutual funds’ screening intensities and their financial performance is curvilinear. This means that more extreme screening intensities are generally found to lead to higher financial performance.

This U-shaped curve is found to be skewed during the entire sample period and during normal times, whereas it is found to be more closely aligned with a perfect curvilinear relationship during both crisis periods. Moreover, it is concluded that the preferred screening processes leading to higher financial performance for socially responsible mutual funds are different depending on the investment area considered.

(8)

2. Background

2.1 COVID-19 Crisis

In December 2019, in the city of Wuhan (China), the epicenter of the so-called COVID-19 disease was detected and soon turned into a worldwide pandemic. Beyond the numerous tragic deaths and the destructive consequences of this virus on the healthcare system, indirect repercussions on the economy and financial markets arose. In particular, the pandemic caused a global macro-economic shock, driving the worldwide economy into a recession of unpredictable size and duration, and triggered a complex sequence of events in the economy (Barua, 2020).

Specifically, the immediate consequences of the COVID-19 pandemic were the temporarily closure of manufacturing sites around the world as well as other business activities. In addition, transport routes connecting different regions were closed, resulting in border lockdowns. These consequences were a result of measures taken in an effort to stop the pandemic’s spread, however at the same time these also produced a de-globalization effect. Goods and people were not allowed to flow cross-nationally, thereby hurting many companies by disrupting their supply chains. This caused a sharp decline in the production and demand throughout the whole system (Nicola, et al., 2020).

Given these circumstances, the financial market response to the spread of COVID-19 was particularly negative with a large fall in share prices. The major financial indices such as the S&P 500 and the Dow Jones Industrial Average (DJIA) experienced substantial downturns.

Specifically, in March 2020, the latter registered a decline of 26% in four trading days (Mazur et al., 2021). This is what has been called COVID-19 market crash.

2.2 Sustainable Investing

Given the growing importance of Socially Responsible Investing (SRI) as a financial instrument, it is worth giving the reader an overview of the market in which such investments are collocated. As disclosed by the US SIF Foundation’s 2020 Report on US Sustainable and Impact Investing Trends, $17.1 trillion out of $51.4 trillion in total US assets under management were invested according to sustainable strategies in 2020. This represents a growth of 42% as compared to the $12.0 trillion registered in 2018. Both Figure 1 and 2 display the size of sustainable investing,

(9)

with the former showing the development of this market between 1995 and 2020 and the latter displaying only numbers from 2020.

Figure 1. Sustainable Investing in the United States 1995-2020. Reprinted from US SIF, n.d., from https://www.ussif.org/fastfacts.

Figure 2. Size of Sustainable Investing Assets 2020.

As can be seen, the sustainable investment universe in the U.S. went from a value of $639 billion in 1995 to a value of $17.1 trillion at the start of 2020, thereby experiencing an increase of more than 25-fold. Moreover, a compound annual growth rate of 14% has been registered.

Among U.S. sustainable investors one can find individuals from all social classes as well as institutions such as foundations, religious institutions, pension funds, credit unions and

(10)

community development banks (The Forum for Sustainable and Responsible Investment, 2021).

These investors aim at maximizing returns while investing according to their personal values and beliefs centered around societal and environmental benefits. As seen, according to Figure 1, the most widely used strategy for investors interested in ethical practices is ESG incorporation.

Specifically, this means that Environmental, Social and Governance (ESG) criteria are applied to processes behind portfolio selection. Furthermore, according to responses given by money managers interviewed on the sustainable practices they implement, the preferred strategy was found to be ESG integration followed by a negative/exclusionary approach, which restricts certain investments within specific areas. The same conclusion was reached when institutional investors were questioned (The Forum for Sustainable and Responsible Investment, 2021). Given the increasingly positive attitude towards sustainable investing among investors, a growing body of literature have started to focus on the relationship between Environmental, Social and Governance (ESG) criteria and financial performance.

(11)

3. Literature Review

3.1 Definitions

While the definition of SRI varies greatly (Hamilton et al., 1993; Heinkel et al., 2001), there is a general agreement that these types of investments represent a way for individuals to combine “personal values, social considerations and economic factors into the investment decision.” (Michelson et al., 2004, p. 1). In other words, it is a strategy that seeks to maximize both financial and social returns. Similarly, Pivo (2005) defined SRI as the combination of “financial profitability of investments, social performance of enterprises, and ecological integrity” (Pivo, 2005, p. 6). Another acceptable definition is provided by Shkura (2019), who interpreted it as “an investment in tangible and intangible form focused on creating long-term value taking into account the impact on the environment, social sphere, quality control, and ethical obligations” (Shkura, 2019, p. 107).

There is furthermore a lack of consensus with respect to how this investment category should be called. According to Capelle-Blancard & Monjon (2012) “the terms social investing, socially responsible investing, ethical investing, socially aware investing, socially conscious investing, green investing, value-based investing, and mission-based or mission-related investing all refer to the same general process and are often used interchangeably” (Capelle‐Blancard &

Monjon, 2012, p. 189). Similarly, Hellsten & Mallin (2006), “use the terms “ethical investments”

and “socially responsible investments” interchangeably” (Hellsten & Mallin, 2006, p. 393). Thus, this thesis will take the same approach.

Although the terms “responsible” and “ethical” embrace a wide variety of investment possibilities, these types of investments are generally subject to a set of screens that limit their investment universe. As defined by Kinder & Domini (1997), a screen is “a criterion applied to a universe of potential investments that helps winnow the candidates” (Kinder & Domini, 1997, p.

12). There is commonly understood to be two approaches to applying these criteria, namely positive or negative screenings. The former is here an inclusionary approach while the latter is exclusionary (Berry & Junkus, 2013). Additionally, the best-in-class strategy has been discussed extensively in the literature. These three methods can be used separately or in conjunction with one another. Renneboog et al. (2008a), for example, found that 64% of the U.S. mutual funds are subject to more than 5 screens while it is just 18% that only use one in their investment strategy.

(12)

The negative screening is considered the oldest and most popular SRI approach (Berry &

Junkus, 2012; Renneboog et al., 2008a). Following some social, environmental, and ethical principles, this exclusionary strategy leads to the avoidance of certain firms or whole industries in the investment portfolio (Renneboorg, et al., 2008a). It is often companies involved in sinful business activities such as alcohol, tobacco, gambling, weapons (Berry & Junkus, 2012; Guenster, 2012) and nuclear power (Kiymaz, 2019; Nofsinger & Varma, 2014; von Wallis & Klein, 2014) that are not seen as suitable options for ethical investors. Inadequate labor conditions, poor environmental preservation, and the practice of animal testing (Berry & Junkus, 2012; Kiymaz, 2019; Renneboog et al., 2008a) are also often seen as violations of ethical behavior, meaning that firms engaged in such activities are also commonly excluded.

A second method used to ensure high ethical standards is the positive screening. Through this technique, firms with good CSR assessment and Environmental, Social and Governance (ESG) scores are selected (Renneboog et al., 2008a). The weight assigned to each stock within the portfolio is based on the level of social responsibility of the underlying company (Berry & Junkus, 2013). In particular, firms involved with the implementation of green technologies and the reduction of pollution are more likely to be selected based on environmental screenings.

Furthermore, diversity, employee satisfaction and human rights are often important standards that companies must live up to in order to meet social screening criteria. Finally, governance screens examine corporate dynamics such as the director’s compensation and board matters (Nofsinger &

Varma, 2014; Renneboog et al., 2008a).

Positive screenings are usually used in conjunction with the best-in-class strategy. The idea behind these two approaches is the same but the latter guarantees that there is a balance across industries in the investment portfolio (Kempf & Osthoff, 2007). In order to ensure this, for each industry, firms are classified according to their CSR level and included in the portfolio only if they are able to reach a minimum score (Renneboog et al., 2008a).

3.2 The History of Socially Responsible Investing

Responsible investing has ancient roots (Philips, 2011). The first group of known ethical investors appeared in the 16^th century in the U.S. under the name of Quakers. These individuals were part of a religious society founded by George Fox, who began to invest according to “beliefs in human equity and non-violence” (Bauer et al., 2005, p. 28). In particular, companies involved

(13)

in the production of unethical products started to be excluded from the investment universe.

According to Renneboog et al. (2008a) modern roots of SRI can be traced back to 1960 when a series of social campaigns (feminism, civil rights, environmentalism, anti-racist movements) caused a shift in investor sentiment. From that point onwards, the predominant strategy based only on maximizing the financial returns shifted to a more ethical inclusionary approach.

In the second half of the 20^th century, investment strategies began to change due to a number of major events and catastrophes which led to a re-evaluation of what constitutes ethical behavior for many individuals. For example, following the outbreak of the Vietnam War, the first modern SRI mutual fund, called Pax World Fund, was founded in 1971 in the U.S. (Renneboog et al., 2008a; von Wallis & Klein, 2014). This fund was created by war opponents aiming at excluding investments in weapons. Furthermore, a few years later, the attention shifted towards the anti-racist movements born to fight the apartheid in South Africa. Responsible investors all over the world exercised pressure on firms operating in South Africa in order to divert their activities to other countries (Renneboog et al., 2008a; Sauer, 1997; von Wallis & Klein, 2014). Even mutual funds were under pressure to exclude these companies from their investment portfolios in order to meet clients’ demands (Renneboog et al., 2008a). In addition, there were two major disasters that increased investors’ concerns and awareness in relation to environmental issues facing the global economy. These were the explosion of the Chernobyl nuclear power plant in the current Ukraine and the accidental release of crude oil into the Alaska’s sea by the supertanker Exxon Valdez.

These both worked to contribute to an increase in demand of new green investment opportunities (Renneboog et al., 2008a; von Wallis & Klein, 2014). In this context, the research area of SRI was born with the aim of studying the changes in investors’ behavior.

Given the growing importance of acting in accordance with socially responsible standards, firms have experienced an increasing pressure from society, governments, and organizations to act ethically (Charlo et al., 2017; Doh et al., 2010; McWilliams & Siegel, 2000). Recent numbers for sustainable and responsible investments prove that investors are paying more attention to where they allocate their money. According to the The Forum for Sustainable and Responsible Investment (2021), $17.1 trillion were invested according to socially responsible practices by the end of 2019. This amounts to a 42% increase in the amount invested from the $12.0 trillion registered two years prior. Considering that the total US assets under professional management

(14)

amount to $51.4 trillion, sustainable investing strategies now account for one out of every three dollars invested.

As stated by Climent & Soriano (2011), investor demands within this universe are mainly focused on investment opportunities in “clean and green technology, alternative and renewable energy, responsible property and other environmentally driven businesses” (Climent & Soriano, 2011, p. 275). This can be explained by the great contemporary focus on climate change issues.

However, concerns are not only confined to environmental problems, but have been expanded to the social sphere as well. Socially responsible investors now also evaluate how corporations respond to the needs of their employees, minorities, women, and suppliers. Therefore, SRI has become strictly connected to the concept of Corporate Social Responsibility (CSR). CSR can here be defined as “situations where the firm goes beyond compliance and engages in actions that appear to further some social good, beyond the interests of the firm and that which is required by law” (McWilliams et al., 2006, p. 1). Given their growing interest in responsible investing, investors are attracted to investments in companies implementing CSR. Furthermore, it also happens that the shareholders are the ones moving the company towards more sustainable practices through their voting rights (Sparkes & Cowton, 2004).

To sum up, socially responsible investors are not only pursuing the goal of maximizing their financial returns but are also interested in holding assets in line with their ethical and social values (Renneboog et al., 2008b). This new investment strategy goes against the conventional wisdom of “don’t mix money and morality” (Goldreyer & Diltz, 1999, p. 23) and stands in contrast to the assumption that there should be a clear separation between the goals of firms and the goals of individuals (Friedman, 2007).

Indeed, economic theories suggests that the financial performance is compromised when including constraints in one’s investment portfolio. In particular, Modern Portfolio Theory (MPT) points out that excluding non-ethical assets limits diversification, resulting in higher investment risk (Chegut et al., 2011) and lower returns (Goldreyer & Diltz, 1999). Moreover, the classical theory of the firm implies that companies undertaking ethical investments incur higher costs, thus lowering their financial performance (Chegut et al., 2011).

Despite the fact that responsible assets might not benefit companies financially, however, corporate managers often include these investments in their strategies. One reason for this approach is explained by Borghesi et al., (2014). This study found that the CEOs’ personal

(15)

characteristics have an impact on their choice of engaging in CSR investments. For example, managers who appear more frequently in the media are more likely to invest in CSR because they see it as an opportunity to enhance their reputations. These results suggest that many CSR investments are mainly implemented to meet firm managers’ private needs. A second motivation for CEOs to undertake socially responsible investments can be the creation of value for shareholders. Borghesi et al., (2014) argue that gaining a good reputation for being socially responsible as well as providing benefits to employees may help firms attract new customers and investors while maintaining their existing relationships. In addition, CSR practices can reduce legal, political and tax risks. This is in line with the social theory of the firm, which suggests that incorporating socially responsible assets can benefit enterprises (Chegut et al., 2011) by increasing their efficiency and creating new markets (Kiymaz, 2019). This assumption is corroborated also by Goldreyer & Diltz (1999) who argued that “firms with social records will be more valuable”

(Goldreyer & Diltz, 1999, p. 23) because they present lower litigation risk and worker turnover, resulting in a reduction of operating costs. Given all these considerations, Hart & Zingales (2017) concluded that the goal of companies should be the maximization of shareholder welfare and not only the firm’s market value.

With respect to single investors, Derwall et al., (2011) divided individuals interested in socially responsible assets into two main categories: value-driven and profit-seeking investors.

The former mainly pursue ethical value rather than financial goals. Contrarily, the latter is more interested in generating superior returns (Wimmer, 2013). However, a study provided by Rosen et al., (1991) on 4,000 individual socially responsible investors revealed that even when these care about environmental and social issues, they are not willing to forgo returns. Other studies have also attempted to draw conclusions on the characteristics of the average responsible investor, however no consensus has yet been reached. On one side, Rosen et al., (1991) and Tippet & Leung (2001) argue that ethical individuals are younger, better educated and have an average income above that of conventional investors. On the other side, Lewis & Mackenzie (2000) and McLachlan & Gardner (2004), after surveying 1,000 and 109 socially responsible investors, respectively, conclude that these individuals can broadly be classified as middle-aged people with an average income. Furthermore, gender also seems to play an important role in the growing segment of SRI. In particular, it has been found that women are more inclined to invest on the basis of their ethical values (Tippet & Leung, 2001).

(16)

Despite the controversial nature of this topic, the main debate in the current literature centers around the financial consequences of being an ethical investor. Given the growing importance of responsible investing, an increasing number of academics and practitioners have attempted to address the question of whether SRI can improve financial performance (Bauer et al., 2005). However, so far, no consensus has been reached in the academic literature. While some argue that SRI portfolios display better financial performance than conventional portfolios, others argue that the performance is generally lower than or on par with conventional counterparts.

As such, the following sections present all the existing evidence in favor of and against the outperformance of SRI both in normal times and during crisis periods. Furthermore, existing literature on the topic of socially responsible screening processes and their effect on financial performance is presented.

3.3 Performance of SRI versus Conventional Funds

Although a large segment of the academic literature has attempted to address the question of whether SRI can improve financial performance, results are mixed.

In their paper, Hamilton et al., (1993) developed three different hypotheses concerning the relative returns of socially responsible as compared to conventional investments. These hypotheses are based on the assumptions of either outperformance, underperformance or equal performance between ethical funds or indices and conventional counterparts. Studies supporting either of these three possible outcomes will be discussed in more detail, however, given the contribution of von Wallis & Klein (2014) an overall picture is that “of the 35 studies that compare SRI vehicle performance to conventional benchmarks, 15 conclude that SRI vehicles perform in the same way as their conventional benchmarks, 6 find underperformance, and 14 exhibit outperformance compared to their various benchmarks” (von Wallis & Klein, 2014, p. 74).

Research on this area has taken different directions both on the choice of investments for the analysis (Schröder, 2004) as well as on the methodological approaches implemented. With respect to the former, several investigators have analyzed the financial performance of portfolios of individual socially responsible stocks (Diltz, 1995a, 1995b; Guerard Jr, 1997), while other researchers have, instead, examined the difference in returns between the Domini 400 Social Index (DSI) and unscreened benchmarks, such as the S&P500 (DiBartolomeo & Kurtz, 1996; Sauer, 1997; Statman, 2000). The DSI consists of 400 socially responsible companies not involved in the

(17)

production of alcohol, tobacco or military weapons. Moreover, the index excludes firms engaging in nuclear power and gambling services (Sauer, 1997). Lastly, a few scholars have compared the performance of socially responsible mutual funds to market indices or conventional unscreened counterparts (Hamilton et al., 1993). With respect to the methodological approaches, several papers have applied the so-called matched pair analysis approach as developed by Mallin et al.

(1995) (see Kreander et al., 2005; Gregory et al., 1997; Statman, 2000). This approach compares the performance of ethical and conventional investments through matched pairs that are formed based on characteristics such as geographical location, investment universe, fund size and fund age (Schröder, 2004). It is also common to find studies that base their analysis on single or multi- factor models, such as the CAPM, the Fama French three-factor and/or the Carhart models (Hamilton et al., 1993), either in conjunction with or without the matched pair analysis approach.

In terms of investment performance measures, the most commonly applied are Jensen’s alpha (Jensen, 1968), the Sharpe ratio (Sharpe, 1994) and the Treynor ratio.

The general view emerging from these SRI studies, is that there is little evidence of a statistically significant difference in the performance of ethical and conventional funds.

In the following sections, the three hypotheses formulated by Hamilton et al. (1993) are presented in turn and studies supporting each of these are discussed in more detail.

3.3.1 SRI = Conventional

Under the first hypothesis, financial returns are speculated to not be affected by different types of investment strategies based on ethical values, meaning that characteristics of social responsibility are not priced in the market (Hamilton et al., 1993). In other words, according to this assumption, holding assets in line with their ethical values neither benefit nor damage individuals or companies. As stated by Hamilton et al. (1993), “this is the hypothesis that is closest in spirit to the standard framework of finance, where factors that are not proxies for risk do not affect expected returns” (Hamilton et al., 1993, p. 63).

This theory has been supported by Guerard Jr (1997) who analyzed 1300 unscreened and 950 socially screened equity stocks over the period 1987-1994. Results suggested that the difference in the average returns between these two equity universes was not statistically significant. Consistent with this hypothesis, Hamilton et al., (1993) has demonstrated that socially responsible mutual funds do not earn statistically significant excess returns and do not outperform

(18)

their conventional counterparts over the period between 1981 and 1990. Many other studies provide evidence of this neutral relationship between the adoption of SRI practices and excess returns. For example, Sauer (1997) compared the risk-adjusted performance of a well-diversified portfolio of socially responsible stocks – the Domini 400 Social Index (DSI) – with two unrestricted benchmark portfolios. Using measures such as Jensen’s alpha, the Sharpe ratio, the monthly average returns, and variability, he found no sign of underperformance of the DSI when compared to the unscreened portfolios. Therefore, the author concluded that individuals interested in investing according to their ethical values can choose socially responsible investments without sacrificing returns (Sauer, 1997). The DSI was also used in a study conducted by Statman (2000), who compared the returns of the screened portfolio to the S&P500 between May 1990 and September 1998. Raw and risk-adjusted returns were both higher for the Domini 400 Social Index, however, the difference was not statistically significant. From his findings, Statman (2000) inferred that “pooling investing power for something other than making money is no worse at making money than pooling it for money alone” (Statman, 2000, p. 38). The paper signed by DiBartolomeo & Kurtz (1996) provide further proof of Hamilton’s first hypothesis, in that no significant difference was found between SRI and conventional investment returns using the CAPM between 1990 and 1993.

Support for this hypothesis has not only been found the U.S. market but has similarly been observed in research on the Australian, UK and European markets. Cummings (2000), for instance, investigated the difference in ethical trust returns as compared to three local benchmarks in Australia. On a risk-adjusted basis, no significant divergence between the performance of the screened portfolios and the conventional ones was detected. The Australian market has also been studied by Bauer et al., (2006) between 1992 and 2003. In their paper, returns of 25 socially responsible mutual funds were compared to those of the Worldscope Australian Index. Equal performance was found over the period 1996 to 2003. A similar analysis has been carried out in Canada with the same results (Bauer et al., 2007). Furthermore, a comparable analytical approach was also used some years before by Bauer et al., (2005) in order to detect the presence of abnormal returns for socially responsible mutual funds in Germany, the UK, and the U.S. Using a sample of 103 ethical mutual funds over the period 1990 to 2001, no evidence of a statistically significant difference in returns between socially responsible and conventional funds was found. This also seemed to be consistent even after controlling for common factors such as size, book-to-market,

(19)

and momentum. In addition, ethical funds were found to be less exposed to market return variability than the conventional counterparts (Bauer et al., 2005). The German market of socially responsible funds has also been studied by Schröder (2004). In his paper, the returns of 16 German and Swiss funds and 30 U.S. funds were investigated against specific benchmarks. Most of the ethical investments’ indices exhibited a positive Jensen’s alpha, however, the statistics were found not to be significant. Again, socially responsible investments did not over- or under-perform conventional assets. In line with this evidence, using a matched pair analysis approach², Bello (2005) showed that ethical funds’ performance is not statistically different from the returns of conventional funds in “the characteristics of assets they hold, the degree of portfolio diversification, or investment performance” (Bello, 2005, p. 43). Kreander et al. (2005) provide additional documentation of no statistically significant difference in returns between 30 European ethical and 30 non-ethical funds between 1995 and 2001. In this study, the screened portfolios were matched with their conventional counterparts according to age, size, and investment universe. However, Gregory et al. (1997) documented that the use of a size factor in the inclusion of non-ethical funds did not control for a small cap bias. As far as the UK is concerned, Mill (2006) examined the performance of a unit trust that had switched from being a conventional fund to a fund focused on SRI. The comparison was made with three similar conventional funds whose investment style remained unchanged. Results suggested that the risk-adjusted returns of the fund remained unaffected even after the shift in investment objective. However, a higher variance for 4 years after the switch was observed by the fund. After this period of adjustment, the variability in returns returned to normal levels. This effect can be explained by fund managers’ “learning by doing”

effect³. Similar to previous studies, no significant difference was found between ethical and non- ethical funds in terms of financial performance.

To summarize, SRI has been observed in many studies to not provide either a financial disadvantage or benefit when compared to their conventional counterparts (Schröder, 2004). As theorized by Hamilton et al. (1993), this can be explained by the fact that ethical features are not priced in the market. Furthermore, it has been suggested by von Wallis & Klein (2014) that

“another reason why the performance of social and conventional funds is closely correlated might

2 Mallin et al. (1995)

3 Arrow (1971)

(20)

be that the specific holdings in the two different portfolios do not differ from their conventional counterparts as much as expected” (von Wallis & Klein, 2014, p. 76).

3.3.2 SRI < Conventional

Under Hamilton et al. (1993)’s second hypothesis, the expected returns of socially responsible funds are lower than those of conventional portfolios. If this assumption is proven to be true, one can conclude that the market is able to price characteristics of social responsibility.

The theory of underperformance of ethical funds as compared to traditional investments finds its roots in MPT, given its assumption that excluding non-ethical assets from the investment universe limits diversification. As a result, investors incur additional costs and thus experience lower risk-adjusted financial returns (Barnett & Salomon, 2006; Goldreyer & Diltz, 1999). The main argument in favor of this hypothesis is that managers investing in socially responsible assets have a confined number of funds to choose from that align with their preferred strategies. Limited allocation possibilities lead to extra costs related to monitoring, thus negatively affecting the performance of ethical funds (Rudd, 1981). According to Cowton (1998) it would not be surprising that ethical funds present lower returns as compared to traditional counterparts given that conventional investors can choose to hold the same portfolio of socially responsible investors , but not vice versa.

This negative relationship between SRI and financial performance has been supported by studies conducted by Michelson et al. (2004) and Tippet (2001), which were able to detect the so- called ethical penalty. However, little evidence of socially responsible funds’ underperformance as compared to conventional counterparts has been found in empirical research. In particular, only 6 studies were able support this assumption (von Wallis & Klein, 2014). For example, Mueller (1991) conducted an analysis on 10 socially responsible mutual funds over the period 1984 to 1988.

Results showed that risk-adjusted returns for ethical investments were on average 1.03% lower than those for the unscreened funds. A similar significant underperformance of socially responsible funds was found by Teper (1992) who compared the KLD 400 index with the S&P 500 between 1985 and 1989. The conventional S&P 500 has also been studied against the tobacco-free portfolio between 1986 and 1996. This analysis was performed by Kahn et al. (1997) and resulted in the conclusion of underperformance of the screened index. Following the work of Mallin et al. (1995), Gregory et al. (1997) performed a matched pair and cross-sectional analysis in order to study the

(21)

performance of socially responsible funds. This study provided evidence of a tendency for UK ethical unit trusts to underperform conventional counterparts. The authors concluded that this might be expected given the restricted investment universe of ethical unit trusts, lower portfolio diversification and “increased monitoring costs” (Gregory et al., 1997, p. 723). Furthermore, three of the major Australian socially responsible mutual funds have been shown to underperform their benchmarks over the period 1991 to 1998 (Tippet, 2001). According to the study, a possible explanation was the ratio between management fees and total income, which appeared to be substantial. Given previous research, Climent & Soriano (2011) attempted to address the question:

“Did it pay to be a green investor?” (Climent & Soriano, 2011, p. 285). Using a sample of U.S.

environmental mutual funds matched with conventional funds with similar characteristics during the period 1987 to 2009, they found a lower performance of ethical mutual funds as compared to that of the unscreened counterparts. It was concluded that it did not pay to be a green investor, “at least in US during 1987-2009” (Climent & Soriano, 2011, p. 285).

In summary, although the hypothesis of SRI underperformance is in line with MPT assumptions, little evidence supporting this hypothesis is present in existing literature.

3.3.3 SRI > Conventional

Finally, under Hamilton et al. (1993)’s third hypothesis, the expected returns of socially responsible funds are higher than those of conventional portfolios, meaning that “doing well while doing good” (Hamilton et al., 1993, p. 62) is possible. This outcome can be explained by a scenario in which many investors underestimate the probability that bad news is disclosed in relation to conventional companies. Such negative information would, on one hand, contribute to the underperformance of non-ethical investments and, on the other hand, lead to SRI outperformance (von Wallis & Klein, 2014). For example, the application of the different types of screenings may reduce incurred costs relating to corporate social crises or environmental disasters. “If financial markets tend to undervalue such costs, portfolios based on corporate governance, social or environmental criteria may outperform their benchmarks” (Renneboog et al., 2008b, p. 305).

According to Moskowitz (1972), higher returns could also be explained by the fact that companies investing in social and environmental assets are associated with good managerial capacity.

Moreover, Schwartz (2003) argued that socially responsible mutual funds display better financial

(22)

performance because they are subject to more scrutiny as compared to their conventional counterparts.

Over the past decades, many scholars have attempted to provide supporting evidence for the idea that individuals interested in ethical funds do not have to forgo returns in order to invest in alignment with their values. In particular, 15 studies have been able to validate this hypothesis (von Wallis & Klein, 2014). Firstly, Luther & Matatko (1994) found weak evidence of higher risk- adjusted returns for UK ethical unit trusts as compared to the conventional counterparts. This result has furthermore been validated by Mallin et al. (1995) who conducted the same analysis over the period between 1986 and 1993 using financial measures such as Jensen’s alpha and the Sharpe and Treynor ratios. Both ethical and non-ethical unit trusts were, however, found to underperform the market. Moreover, Travers (1997) found significant abnormal returns for 23 selected socially responsible mutual funds from Europe, Australasia and Asia when compared to the MSCI EAFA (Morgan Stanley Capital International Europe, Australasia and Far East Index). Gil-Bazo et al.

(2010) applied a matched pair analysis approach and tested U.S. socially responsible mutual funds against conventional funds with similar characteristics over the period 1997 to 2005. The difference between the two categories of funds appeared to be statistically significant, however, the authors admitted that this result was “driven exclusively by SRI funds run by management companies specialized in SRI” (Gil-Bazo et al., 2010, p. 243). Several other researchers have examined the performance of social indices against benchmark portfolios. For example, in the paper signed by D’Antonio et al. (1997) 140 ethical companies out of the KLD 400 Index were selected for a test for abnormal returns. Although the authors concluded that socially responsible bonds outperform the Lehman Brothers Corporate Bond Index (LCB), this result has been imputed to the difference in credit risk between the ethical portfolio and the conventional benchmark. The outperformance of socially responsible investments has not only been supported when indices have been analyzed but also when stock portfolios have been considered. This was the case in a study performed by Derwall et al. (2005) who found significant abnormal returns when investigating the performance of socially responsible stock portfolios versus non-environmentally friendly portfolios between 1995 and 2003. A positive effect has been also detected when analyzing the relationship between corporate governance and financial performance. Both Gompers et al. (2003) and Tippet (2001) support the idea that excluding companies involved in unethical activities within this area from the investment universe leads to higher returns given the costs reduction. In the

(23)

paper of Shank et al. (2005), a sample of 11 ethical firms was compared with selected socially responsible mutual funds and a conventional benchmark. The results were different depending on the time period analyzed. On one hand, when horizons of 3 and 5 years were considered, no outperformance was detected in comparison to the market, both in terms of single firms as well as for mutual funds. On the other hand, when investigating a 10-year horizon, the socially responsible firms showed a statistically significant Jensen’s alpha, meaning that they were able to outperform the market. In line with this research, the study performed by Hill et al. (2007) showed that the outperformance of socially responsible companies as compared to conventional benchmarks, if registered, was significant only in the long-term. This paper analyzed ethical companies in Europe, Asia and the U.S. and for each of these regions selected an appropriate index as the benchmark reference (FTSE 300, Nikkei 225 and the S&P 500, respectively). According to this study, only the European firms displayed an outperformance which was statistically significant when compared to the benchmark in the short term.

Some researchers have attributed these abnormal returns to the capacity of firms to increase employee satisfaction (Edmans, 2011; Guenster, 2012). However, some suggest that the outperformance experienced as a result of this is unlikely to continue over the long term (Bebchuk

& Weisbach, 2010; Derwall et al., 2011; Guenster, 2012). This idea has been supported by Bebchuk & Weisbach (2010) who introduced the concept of the learning effect for corporate governance. This concept describes that as soon as individuals become conscious that investments in socially responsible assets lead to abnormal returns this outperformance will disappear. This theory has been supported by the study performed by Derwall et al., 2011, who provided evidence of decreasing abnormal returns for SRI in recent years.

To sum up, “Doing well while doing good” (Hamilton et al., 1993, p. 62) seems to be possible, but there is a question as to whether this can be obtained in the short-term.

3.4 Mutual Funds’ Performance in Crisis Periods

From the previous discussion on the existing literature of the performance of socially responsible investments as compared to conventional counterparts in normal times, it can be concluded that no consensus has been reached by researchers.

(24)

Turning to crisis periods, few academic sources investigate the effect of holding a portfolio containing ethical funds as compared to holding a traditional portfolio. This section will, however, be dedicated to the discussion of earlier studies conducted within this area of research.

The most complete investigation on the performance of U.S. socially responsible mutual funds during crisis and non-crisis times has been performed by Nofsinger & Varma (2014). In particular, their study aimed at exploring whether socially responsible funds limit downside risk for investors during economic downturns. The analysis was carried out using a sample of U.S.

ethical mutual funds matched with conventional funds with similar characteristics over the period 2000 to 2011. Thus, the technological bubble burst between 2000 and 2002 and the global financial crisis between 2007 and 2009 were considered as crisis periods. The model specifications implemented were the CAPM, the Fama French three-factor and the Fama French five-factor models. The crisis and non-crisis periods were analyzed separately, and thus different alphas were estimated for each period. Results of the investigation suggest that in normal times, conventional funds outperform socially responsible counterparts by 0.67-0.95% (depending on the model implemented) with alpha being significant at the 10% confidence level. Contrarily, in periods of economic downturn, the alphas of the U.S. ethical funds appear to be statistically significantly higher than those of the conventional funds (from 1.61% to 1.70%). Given these results, it was concluded that individuals have to sacrifice returns in non-crisis periods when investing in U.S.

socially responsible mutual funds, but that they are then rewarded with lower downside risk during economic recessions.

The question that arises is: “would investors be willing to give up some return in non-crisis market periods to gain some higher returns during crisis periods?” (Nofsinger & Varma, 2014, p.

181). According to the Prospect Theory proposed by Kahneman and Tversky (1979), investors appear to be “more negatively impacted by a loss than they are positively impacted by a gain of similar magnitude” (Nofsinger & Varma, 2014, p. 181). Thus, it is likely that investors would be willing to sacrifice some returns in normal times to reduce the downside risk during periods of distress. This phenomenon is also confirmed by the study performed by Glode (2011), who found evidence of an increase of demand for actively managed funds during economic recessions.

Recently, more studies have investigated the effect of crisis periods on the performance of ethical and conventional funds in different markets around the world. Firstly, Nakai et al. (2016) examined Japanese socially responsible and conventional funds during the global financial crisis

(25)

in 2008. Results obtained using the Fama French three-factor model align with Nofsinger & Varma (2014)’s assumption that socially responsible funds outperform conventional funds during economic recessions. Contrarily, a study performed by Belghitar et al. (2017) found that UK ethical funds performed better only during pre- and post-financial crisis periods while they underperformed conventional funds during recession times. Moreover, Leite and Cortez (2015) found that the performance of ethical funds in France did not present abnormal returns in comparison to conventional funds during the technology bubble burst, the global financial crisis and the euro sovereign debt crisis. However, while socially responsible funds did not present statistically significantly different returns when compared to their traditional counterparts, they underperformed their peers during non-crisis periods. The financial performance of ethical funds during the global financial crisis and the technology bubble burst has also both been analyzed by Becchetti et al. (2015). Their paper investigated the performance of socially responsible funds and matched conventional funds during the period 1992 to 2012. In all the different geographical areas considered, abnormal returns were found in relation to ethical funds during the global financial crisis but not during the technology bubble burst. In the latter case, investors did not perceive the holdings of socially responsible investments to be a protection against the downside risk to which they were exposed to. Therefore, these findings partly stand in contrast to the Prospect Theory as proposed by Kahneman and Tversky (1979).

In summary, with respect to the existing literature on the performance of socially responsible funds as compared to conventional counterparts during crisis periods, no consensus has been reached by the different scholars.

3.5 Influence of Screening Processes on SRI Fund Performance

As outlined in previous sections, although many studies have been performed in relation to the topic of ethical and conventional funds difference in financial performance, most of these were not able to identify either a positive or negative relationship between financial returns and SRI. However, according to Barnett & Salomon (2006), these two opposite perspectives can be perceived as complementary. After many years of debate, their study seemed to solve the existing controversy by consolidating the opposing views in the literature. More specifically, it was suggested that a curvilinear relationship exists between social screenings and financial performance, “with the strongest financial returns to low and high levels of social responsibility,

(26)

and significantly lower financial returns to moderate levels of social responsibility” (Barnett &

Salomon, 2006, p. 35). In other words, they propose that the type and intensity of the screening applied affect the investment’s financial performance. Using a sample of 61 ethical funds during the period 1972 to 2000, Barnett & Salomon (2006) suggest that the number of social screens is inversely correlated with financial returns up to a certain level. On one hand, funds having a more restrictive approach eliminate underperforming companies, thereby boosting returns. On the other hand, funds with less constraints are able to improve performance by increasing diversification.

Contrarily, funds taking in-between approaches are not able to benefit either from the exclusion of firms with lower returns nor from decreasing the unsystematic risk, thus resulting in a lower financial performance.

The level of the returns is not only thought to be driven by the intensity of screening but also by the type of social screen employed. For example, while the benefits coming from community relations screenings are more than able to offset their costs, environmental and labor relations screenings appear to worsen financial performance (Barnett & Salomon, 2006). This perspective is furthermore supported by a study conducted by Diltz (1995a, 1995b) who analyzed daily stock prices between 1989 and 1991. Worse or better financial performance was, however, found in relation to different screenings than in the study by Barnett & Salomon (2006).

Specifically, from the results obtained, Diltz (1995a, 1995b) inferred that environmental and charitable screens together with the absence of military and nuclear involvement are able to increase funds’ performance. Contrarily, family benefits seemed to penalize financial returns. In conclusion, evidence suggests that social screenings “may either enhance or hinder portfolio performance” (Diltz, 1995b, p. 67).Moreover, using a sample of screened stock portfolios between 1992 and 2004, Kempf & Osthoff (2007) added new evidence to the existing literature. In their paper, the stocks were screened according to different criteria (positive, negative, and best-in-class methods) and the Carhart (1997) model was used to measure performance. The outcome of the study showed that the best-in-class approach was associated with a higher alpha.

A less complex position on the effect of screenings on SRI returns has been taken by Renneboog et al. (2008b). According to their paper, “all else equal, funds with one additional screen are associated with a 1% lower 4-factor-adjusted return per annum” (Renneboog et al., 2008b, p. 322).

(27)

A more comprehensive analysis has been presented in the paper signed by Nofsinger &

Varma (2014). The aim of their study was to explore how different combinations of screening processes affected the performance of ethical funds in normal as well as in crisis periods. Firstly, when alcohol, tobacco and gambling stocks were excluded from the investment universe, the funds underperformed the market in normal times. The difference in returns during crisis periods, however, was not statistically significant. Moreover, the deviation between screened funds and their conventional counterparts was not statistically significant. Results suggest that excluding stocks involved in the alcohol, tobacco and gambling industry did not reduce the downside risk faced by investors during the explored crisis periods. This conclusion was also reached when all product-related screens were applied to the sample. Secondly, when Environmental, Social and Governance (ESG) screens were employed, ESG funds seemed to perform worse than conventional counterparts during non-crisis periods. Contrarily, the screened investments outperformed traditional funds by 2.18% during periods of economic downturn. In both cases the coefficients were statistically significant at the 5% confidence level. Furthermore, when the different ESG categories were analyzed separately, all three classes performed better than conventional funds with governance being particularly positive. This was true both for crisis and non-crisis periods. Finally, the study attempted to address the question of whether negative or positive screening techniques would lead to higher financial performances. According to the alpha obtained from the analysis, funds using positive screens seemed to underperform conventional funds during normal times and outperform them during crisis periods. Both alphas appeared to be statistically significant at the 5% confidence level. Contrarily, funds applying negative screenings displayed alphas that were not all statistically significant. In conclusion, although investors of socially responsible funds were found to be better off during economic downturns, they were found to have to sacrifice returns during non-crisis periods. Furthermore, the abnormal returns found during crisis times were seemingly driven by the employment of positive screening techniques.

In summary while a large segment of the existing literature has investigated the performance of socially responsible funds as part of the same group, recent studies have outlined different classes of SRI and explored whether these perform differently. The results are mixed and thus, no consensus has been reached on this matter.

(28)

4. Hypotheses Development

Keeping the previously presented literature review in mind, this section will now discuss the focus area of this thesis and present the underlying hypotheses tested. Firstly, it has to be noted that this paper seeks to expand on the previously introduced study performed by Nofsinger &

Varma (2014). This can be regarded as the most comprehensive analysis as of present on the financial performance of U.S. socially responsible mutual funds as compared to U.S. conventional mutual funds during both normal and crisis times. This study by Nosfinger & Varma (2014) analyzed the financial performance of these two types of funds over the period 2000 to 2011. This thesis seeks to carry out a similar analysis, but over a different time period going from 2010 to 2020. Thus, in other words, this thesis seeks to contribute to the existing literature within this area by analyzing the financial performance of U.S. socially responsible mutual funds as compared to U.S. conventional mutual funds over a time period which has yet to be investigated. This period includes a subperiod which can be constituted as normal times as well as a crisis subperiod, given the COVID-19 crisis of 2020. In order to provide a more complex perspective on the financial performance of these funds during the COVID-19 subperiod, however, this was split into two further subperiods, such that the initial market crash and the following recovery period could be analyzed separately. As far as the authors of this thesis are aware, such detailed analysis of a crisis period has never previously been carried out in other studies. The first subperiod constituting normal times was identified as going from the 1^st of January 2010 to and including the 7^th of February 2020. The two COVID-19 subperiods were identified as going from the 10^th of February 2020 to and including the 13^th of March 2020 and from the 16^th of March 2020 to and including the 31^st of December 2020. This market crash and recovery period, respectively, were identified based on the financial upturns and downturns of the S&P500 and Dow Jones Industrial Average (DJIA) indeces. These indeces are often used as proxies for the U.S. market in different financial models, and thus these seemed appropriate for insights on the general movements of this market.

While consensus has yet to be reached with respect to the difference in financial performance between these two types of U.S. funds both in normal and crisis times, the most common finding in previous studies, with respect to the former period, is that there is no statistically significant difference (Hamilton et al., 1993; Guerard Jr, 1997; Sauer, 1997). Thus, this thesis hypothesizes that:

(29)

Hypothesis 1: The difference in the financial performance between U.S. socially

responsible and conventional mutual funds is not statistically significantly different from zero over the entire sample period.

Hypothesis 2: The difference in the financial performance between U.S. socially

responsible and conventional mutual funds is not statistically significantly different from zero during normal times.

With respect to the latter period, the most common finding in previous studies is that socially responsible mutual funds outperform conventional counterparts during crisis times (Nofsinger &

Varma, 2014). Thus, this thesis hypothesizes that:

Hypothesis 3: U.S. socially responsible mutual funds outperform conventional counterparts financially during the COVID-19 crisis market crash and recovery periods.

Additionally, this thesis seeks to contribute further to the investigation of socially responsible mutual funds’ financial performance by adding new evidence to a particularly under-researched area, namely socially responsible mutual funds’ screening processes and their effect on these funds’

financial performance. This is a relevant topic to explore, as selective screening processes are unique to socially responsible mutual funds and thus partly might explain differences between their financial returns as compared to those of conventional counterparts. This analysis will once again be performed both in normal and crisis times and over the same periods as previously specified. With respect to this analysis, this thesis considers two layers of complexity. The first layer simply looks at the screening processes applied by each fund without considering the investment area to which they are applied, while the second layer of complexity take these areas into consideration. With respect to the former, given conclusions reached in previous studies (Barnett & Salomon, 2006), this thesis hypothesizes that:

Hypothesis 4: The relationship between the screening processes of U.S. socially responsible mutual funds and their financial performance is curvilinear.

Thus, in other words, this thesis hypothesizes that the most extreme screening intensities in either direction (e.g., very restrictive/exclusionary or no restrictions at all) will lead to higher financial performance. The existence of such a curvilinear relationship would have certain implications for commonly applied financial theories, such as MPT, which sees diversification as

(30)

ultimately leading to higher financial returns and lower risk (Chegut et al., 2011; Goldreyer &

Diltz, 1999).

With respect to second layer, no previous studies have performed this kind of in-depth analysis of the different screening processes’ effectiveness within different investment areas.

Previous studies (Nofsinger &Varma, 2014; Barnett & Salomon, 2006), however, have found there to be a difference given different areas of investment, and thus, this thesis expects to reach the same conclusion. However, the exact pattern of how these will differ cannot be hypothesized about, since there is no available information to base this assumption on. Thus, this thesis hypothesizes that:

Hypothesis 5: Both in normal and COVID-19 crisis times, the most effective screening processes leading to higher financial performance of U.S. socially responsible mutual funds differ depending on the investment area within which they are applied.

(31)

5. Models and Estimation Methods

5.1 Asset Pricing Models

In this section of the paper, the asset pricing models used in the analysis will be outlined.

Firstly, the Capital Asset Pricing Model (CAPM) and its underlying assumptions will be explored.

Thereafter, this section will further outline two later established and alternative models. These are the Fama-French three-factor and five-factor models, which have been constructed under the assumption that the CAPM does not fully explain an asset or portfolio’s returns.

5.1.1 Capital Asset Pricing Model (CAPM)

The Capital Asset Pricing Model (CAPM) is a model developed by Sharpe (1964) and Lintner (1965). The model is widely used in application and provides a simple way to measure risk as well as the relationship between risk and the expected return of an asset or a portfolio of assets (Fama & French, 2004). The model builds on the work of Markowitz (1959) and his mean- variance model, and thus assumes that investors are risk averse and only care about the mean and variance of their one-period investment return. In other words, investors will always choose a so- called mean-variance-efficient portfolio, which minimizes the variance of the portfolio return, given the expected return and maximizes the expected return, given the portfolio variance. Other than this key assumption, Sharpe (1964) and Lintner (1965) added two further assumptions to the CAPM. Firstly, there must be complete agreement between investors with respect to the joint distribution of asset returns over the one investment period. Furthermore, this must be the actual true distribution of the returns. The second assumption is that it has to be possible to borrow and lend money at the risk-free rate and that this will not change depending on the investor or on the amount borrowed or lent (Fama & French, 2004). One can visually display the implications of these assumptions by constructing a chart which measures the standard deviation of the portfolio return on the horizontal axis and its expected return on the vertical axis (Fama & French, 2004).

This visual representation can be found in Figure 3.