Beyond Appearances: The Risk-Reducing Effects of Responsible Investment Practices

Abstract

This article examines the theoretical motivations underlying the conflicting beliefs in support of and against responsible investment (RI) and presents unique quantitative evidence to illustrate how such conflicting logics produce a curvilinear (inverted U-shape) relationship between screening intensity and two measures of risk. First, I argue that, whereas limiting the investable universe by using RI screening criteria increases the risk specific to the portfolio, very high screening intensity can reduce this risk. This is due to the fact that information benefits enable fund managers to be more selective, allowing them to select less risky firms. Second, by drawing on behavioral studies, I argue that this same curvilinear relationship occurs when examining the flow of money coming in and out of a fund. That is, high RI screening makes ethical investors “stickier” and less likely to pull money out of a fund because they are attracted to its ethical properties. I test my hypotheses using a data set of all known European RI screening equity mutual funds. I generally find strong support for both hypotheses. This has an important implication for investors: For high screening intensity and meaningful RI practices, RI is associated with a significant risk reduction.

Keywords

idiosyncratic risk responsible investment RI screening sustainability

The world is changing faster than it ever has before, and environmental issues, political risk, human rights, these kinds of issues are now fundamental; these issues are becoming more and more material. And if you have people managing your money, do you want them to have a really good understanding of the mega trends that are facing the world? Or would you prefer that they have a very traditional backward-looking, myopic approach of fund managers that look at the small number of financial metrics and then try to predict whether a company is going to be a good buy or not?

—Excerpt from my personal interview with James Gifford, former Director of the United Nations’ Principles for Responsible Investment

Over recent years, the theme of sustainability has emerged within the context of the financial sector. In particular, asset management firms—professional financial firms responsible for pooling money from different sources and managing portfolios of global wealth—have been questioned as to whether and how they consider environmental, social, governance (ESG), and controversial business involvement (CBI) issues during their investment processes and how, by financially supporting and/or reprimanding firms based on their performance on such issues, such financial institutions can affect the development and well-being of future generations.¹ This has brought about a phenomenon known as responsible investment (RI), defined in this article as any form of integration of sustainability issues within traditional investment processes.² Corporate social responsibility (CSR) and its translation into RI at the level of financial markets can be described as a form of “creating economic value in a way that also creates value for society by addressing its needs and challenges” (Porter & Kramer, 2011, p. 64).

From a niche market having roots in ethical investment,³ RI has evolved and is now driven mainly by the need for better risk controls in a crisis-stricken environment. The phenomenon is growing at such a staggering pace that recent scholars have pointed toward its “mainstreaming” in particular geographies such as France (cf. Arjaliès, 2010; Crifo & Mottis, 2013; Louche & Lydenberg, 2006). In 2010, the European Social Investment Forum (Eurosif), an RI think tank, reported that in Europe, RI currently represents assets worth EUR 5 trillion, or more than 46% of the overall assets under management (AUM) in the region (Eurosif, 2010), making it the leading geography in RI.⁴

However, several academic studies in finance (cf. Juravle & Lewis, 2008; Renneboog, Ter Horst, & Zhang, 2008b, for recent literature reviews) as well as anecdotal evidence from practitioner communities⁵ point out that engaging in RI remains highly controversial due to the fact that it entails an important departure from accepted practices in asset management. Indeed, key financial theories assume that a fund manager should be able to “diversify away” any specific risk coming from an individual firm by including other non-perfectly correlated firms within the portfolio. By doing RI, a fund manager may have his or her “hands tied” and be unable to include all the assets that he or she would have usually included. Furthermore, literatures in the CSR domain highlight the difficulties of the commensurability of sustainability and the lack of robust sustainability information upon which to base investment decisions (Delmas & Blass, 2010; Delmas, Etzion, & Nairn-Birch, 2013).

Yet RI proponents posit that by having more information, RI fund managers are able to make better decisions and select more stable and predictable firms. In line with stakeholder theory (Donaldson & Preston, 1995; Freeman, 2010), because RI funds include firms that have high corporate social performance, such funds should perform better than non-RI funds. Furthermore, they argue that RI assets tend to be “stickier” in that their clients are more committed to such investments in the long term (cf. Renneboog, Ter Horst, & Zhang, 2008a).

Due to these competing logics, it appears that many studies on RI have been inconclusive and there is still uncertainty as to the significance of its relationship with financial performance (cf. Dowell, Hart, & Yeung, 2000; Margolis & Walsh, 2003; Orlitzky, 2011; Waddock & Graves, 1997). Several reasons for this, such as measurement errors, “dichotomizing” between RI and non-RI funds, and the lack of available data have been put forward in the literature. Yet it is important to point out that another explanation may lie in the fact that examining risk has usually not been at the core of the debate with most extant studies focusing on the performance debate. Thus, the purpose of this article is to provide a differentiated empirical contribution to the literature on investments by examining whether such differing views can be reconciled. In particular, this article presents unique quantitative evidence to illustrate how such conflicting logics produce a curvilinear (inverted U-shape) relationship between screening intensity and two measures of risk.

I first draw on recent work in CSR and business ethics to provide an examination of the theoretical motivations underlying these conflicting beliefs and evidence support for what I term “Substantive Responsible Investment,” wherein a high level of screening intensity is imposed upon a portfolio. First, I argue that while limiting the investable universe increases idiosyncratic (specific) risk due to portfolio diversification issues, engaging in a substantive form of RI can be risk-reducing due to the fact that informational benefits allow fund managers to be more selective, thus producing a curvilinear effect. Second, by drawing on behavioral studies, I argue that this same curvilinear relationship occurs when examining the flow of money coming in and out of a fund because substantive RI makes ethical investors “stickier.” I test my hypotheses using a data set of all known European RI screening equity funds. I generally find strong support for my hypotheses that RI is associated with a reduction in both the specific risk of individual stocks and the money flow volatility of a fund when screening intensity is high and support the case for deeper and more meaningful RI practices.

Theoretical Framework

The Traditional View: RIs as Risk-Increasing

The practice of asset management is a widely acceptable means of investing current wealth in anticipation of higher expected future returns. Investing money in the markets entails a substantial amount of time, research, and sophisticated tools to understand which sectors and companies are likely to perform best and have lower risks in the future: capabilities and resources which an individual or organization may not have. The idea behind asset management is that it is more effective and less risky to pool money—“assets”—from several individuals or organizations and to outsource the collective management of these assets to a specialized firm. This allows risks to be spread across a diversified portfolio of assets, which would otherwise be more expensive to do due to high transaction costs. Asset managers also monitor developments in the markets and are able to select interesting opportunities. Asset management practices are thus embedded in a clear and agreed-upon mission of profit maximization through risk-return optimization.

This mission subscribes to a “traditional view” of economic theory going back to expected utility theory, which argues that investors are rational and wealth maximizing. An efficient portfolio can only be created through specific combinations of risk and return. The investor would (or should) then want to select one of those portfolios that give rise to efficient combinations of the two. This theory—modern portfolio theory (Markowitz, 1952)—is arguably the most strongly manifested theoretical underpinning in the practice of asset management. Highly sophisticated tools are available for asset managers to use to create an efficient portfolio and the performance of a fund is judged on whether it is able to achieve a return above a previously identified benchmark (usually a market index such as the S&P 500 in the United States or the Morgan Stanley Capital International [MSCI] Index in Europe).

Because an asset manager is expected to be able to fully diversify, modern portfolio theory argues that only risks related to the market (“systematic risk”) matter as the risks associated with the volatility of an individual security (“idiosyncratic risk”) can be “diversified away” by including non-perfectly correlated assets within a portfolio. That is, the specific risk carried by any individual security can be offset by the specific risk carried by another. While it has been argued that an asset manager cannot perfectly construct a portfolio that includes the entire market due to several constraints such as transaction costs (Malkiel & Xu, 2006) or exogenous legal or regulatory constraints (Lee & Faff, 2009), a mutual fund can still eliminate most specific risk by including randomly selected stocks (J. Y. Campbell, Lettau, Malkiel, & Xu, 2001; Statman, 1987). An RI fund, however, is a key example of a fund having constraints that prevent it from selecting stocks at random. The traditional view thus argues that having more stringent screening criteria related to sustainability issues limits an RI fund’s ability to diversify, thereby forcing them to bear a higher degree of idiosyncratic risk. Furthermore, by increasing the intensity of screening, the mean-variance frontier should continue to shift toward lesser favorable risk-return trade-offs. Thus, the expectation is a positive relationship between an increase in screening and the idiosyncratic risk of a portfolio.

Sustainability-Related Information as a Source of Risk-Reduction

Today, this is an argument that I do with the firms: When we meet, I ask, Let’s not talk about what you are doing now, not what you did yesterday, because this I can read, I can find in the sustainable development report of the annual report, instead, let’s talk about these criteria, these risks which we’ve identified as the most important, about what you will do in 3 and 5 years. (Excerpt from my personal interview with Franca Perin, Head of SRI, Generali Investments)

Given the changing and uncertain market environment at present, proponents of RI are posing new arguments in support of the practice as potentially risk-reducing. For instance, the Principles for Responsible Investment (PRI), a U.N.-backed initiative driven by asset owners, has put forward the belief that due to the informational benefits of doing RI, asset managers are able to select more predictable and stable firms. However, the previous focus in the academic literature on the performance debate has eclipsed theorizing about this risk-reducing potential. Indeed, it has been pointed out that there is little to no research focusing on idiosyncratic business risk at a portfolio level (Lee & Faff, 2009).

This risk-reduction belief, while novel in the investment field, has been studied at firm level. Several studies in the management field have shown that firms engaging in CSR practices may reduce the risk inherent in the firm’s operations as a result of external or internal factors that can affect a firm’s profitability (Jo & Na, 2012). Godfrey, Merrill, and Hansen (2009) illustrate how some types of CSR activities can create goodwill and provide “insurance-like” protection. By measuring the volatility in stock returns during specific negative events, the authors find that when a firm suffers a negative event, it is less likely to be financially penalized by its shareholders if it has had good CSR practices in the past. This is a similar finding to Bansal and Clelland (2004) who show that environmentally legitimate firms incur less unsystematic risk than illegitimate firms. Sharfman and Fernando (2008) suggest that improved environmental risk management can lead to a reduction of cost of equity capital, an outcome of reduced firm risk, because such firms can benefit from tax regimes and avoid penalization during environmental disasters and thus reduce the exposure they will face from current and future regulations. Similarly, Lee and Faff (2009) argue that the activities of leading sustainable firms are likely to have a downward influence of their idiosyncratic risk. The authors say that, for instance, this lower risk can happen because these firms have happier, more stable employees, lower fines, and good production levels. While other studies have found that firms may be punished to a greater degree because these events are unexpected (Konar & Cohen, 1997; Rhee & Haunschild, 2006) or that there is no real difference in the risk-adjusted performance of firms with high or low corporate social performance ratings (Humphrey, Lee, & Shen, 2012), there nevertheless appears to be some evidence that points toward a negative association between company-unique idiosyncratic risk and CSR (Bouslah, Kryzanowski, & M’Zali, 2013; Boutin-Dufresne & Savaria, 2004; Lee & Faff, 2009; Mishra & Modi, 2013).

Through screening, it is assumed that RI funds are able to have more information than conventional funds related to sustainability-related activities of firms, allowing RI fund managers to better understand which firms are more stable from a financial performance perspective. Through more in-depth and forward-looking information, RI fund managers are able to better understand the evolution of business models in terms of critical issues on a long-term horizon and make investment decisions based on a more comprehensive knowledge of investment risks. This goes against the traditional idea in finance that these specific risks are generally unpredictable. As such, as the investment universe becomes smaller, RI fund managers become more aware of the diversification challenge and construct portfolios with increasing selectivity.

However, if procuring useful information on sustainability issues were simple and straightforward enough, then it could be imagined that all fund managers would immediately start using the available sustainability information in the market, thus eliminating the “edge” of RI funds over conventional funds. However, due to the fact that sustainability reporting is voluntary (D. Campbell & Slack, 2008) and the nature of sustainability issues is difficult to measure, there is poor quality of sustainability information available from investee firms and specialist research providers fostering a high level of information asymmetry (Slager, 2014). Indeed, a recent study by Busch, Bauer, and Orlitzky (2015) highlights the importance of the trustworthiness of ESG data in order for a real shift toward sustainable business practices to occur. Unlike financial performance indicators, which over time have become well defined and standardized, to date there has been no convergence upon universally accepted environmental and social performance indicators, thus making such information questionable (Delmas et al., 2013). Furthermore, such information may also be proprietary and specifics are kept confidential (Delmas & Blass, 2010). Nevertheless, such methodologies remain useful for assessing such a complex concept (Delmas et al., 2013; Ferraro, Etzion, & Gehman, 2015). Therefore, to mitigate this difficulty, investors need to go above and beyond what is publicly available, engage with investee firms, and draw from multiple sources to make better and more selective decisions.

I thus posit that due to the fact that conflicting logics exist that necessarily pose certain trade-offs, RI practices can only be risk reducing when asset managers engage in “Substantive RI”: going beyond usual market practices to possess relevant and robust sustainability-related information. Such practices can include a dialogue with firms or the existence of a dedicated research team to provide valuable information for the fund manager. However, funds are only likely to engage in such practices if they are truly committed to RI. Fund managers engaging in Substantive RI thus differ from two types of asset managers: (a) traditional asset managers who rely solely on financial information to make investment decisions and (b) RI fund managers who only take public sustainability-related information (usually annual sustainability reports) “as is” without making further analyses to determine whether or not such information is timely, accurate, or relevant. During several discussions with asset managers, it emerged that those who were stronger advocates of RI tended to have more stringent screening practices to position themselves as different from those who used “soft” (easier to implement) categories. Thus, a higher level of screening intensity is a useful proxy to represent a subset of RI fund managers with a higher amount of selectivity and a deeper sustainability-related investment process, and such managers should therefore be able to gain a risk-reduction benefit due to a decrease in information asymmetry.

In sum, the relationship between RI screening intensity and idiosyncratic risk is by no means linear or straightforward. From the traditional view, increasing screens should minimize the possibilities for diversification. Thus, idiosyncratic risk will be lower for RI funds with a lower screening intensity as these funds, subject to the least amount of portfolio restrictions, are able to diversify firm-specific risks almost in line with non-RI funds. This risk increases as more and more screens are added as the fund begins to decrease possibilities of investing in other assets to diversify risks. However, it is from a certain level of screening (selectivity) wherein there is increased research and monitoring that sustainability-related information becomes useful and allows fund managers to select more stable firms, thus offsetting the costs of diversification.⁶

Thus, I posit the below hypothesis:

Hypothesis 1 (H1): The relationship between RI screening intensity and idiosyncratic risk is curvilinear. Idiosyncratic risk increases as screening intensity increases yet will begin to decrease from a point wherein RI practices become substantive.

Ethical Properties of Investments as a Source of Risk-Reduction

The fact that environmental, ethical, and responsible investment money is a bit more picky and has a bit more long-term thinking . . . then, it probably makes our shareholder base a bit “stickier,” in a way not to bail out and lack trust that the [asset manager] is doing a good thing. (Excerpt from my personal interview with Antti Savilaakso, RI Investment Manager, Nordea)

The second key argument in support of a rationale for engaging in RI is that clients of RI appear to be more “sticky” thus making the money that investors put into a fund (above and beyond what is gained or lost due to asset appreciation or depreciation)—money flow—less volatile and more predictable. This is evidenced by the fact that the RI movement continues to grow in spite of inconclusive evidence of a performance benefit. Indeed, practitioner studies have highlighted the fact that RI investments are “stickier” than non-RI investments during moments of crises. According to a report from the Social Investment Forum, which mentions a Lipper study, the first 9 months of the 2001 U.S. downturn saw a 94% drop in the dollars investors put into all mutual funds, compared with just a 54% drop for socially screened funds. Similarly, from the start of 2007 to the opening of 2010, a 3-year period when broad market indices such as the S&P 500 declined and the broader universe of professionally managed assets increased less than 1%, RI assets in the United States increased by more than 13% (USSIF, 2010).

The theoretical argument in support of this can be traced to the idea that ethical attributes produce high levels of stability that links to several studies in behavioral finance. Such studies argue that individuals make decisions based on cognitive limitations of their minds (cf. Simon, 1955) and through framing (Kahneman & Tversky, 1979, 1984; Statman & Caldwell, 1987). Individuals may willingly choose immaterial utility such as happiness or satisfaction gained from ethical considerations within their utility maximization (Gao & Schmidt, 2005; Qian, 2006). These behavioral studies form much of the research on why investors deviate from the value maximizing principle and incorporate other decision variables (such as extra-financial variables) into their investment decisions. For instance, Nilsson (2007) found that apart from financial return, social, environmental, and ethical issues are important determinants of investment while Lewis and Juravle (2009) found that investors are driven by a wide range of values. Other studies have found that having a negative ethical stance toward the stock market is a significant negative predictor of willingness to invest in stocks (Keller & Siegrist, 2006), that holding profit constant, people are willing to pay more for ethical shares (Hofmann, Hoelzl, & Kirchler, 2008) or accept lower financial returns for their investments in exchange for positive social returns (Glac, 2008), and that the strength of investors’ personal values is important in determining their investment choices (Pasewark & Riley, 2009).

Related to this, displaying commitment to sustainability is positive for the legitimacy of an RI fund if its clients value such ethical attributes. A sociological perspective of organizations posits that organizations tend to conform to norms to gain legitimacy in their field (DiMaggio & Powell, 1983; Lounsbury, 2001). With the onset of sustainability as a strategy that may increase firm legitimacy, organizations will engage in sustainability to be part of the norm and to not lose their reputation (Bansal & Clelland, 2004; Philippe & Durand, 2011). That is, organizations will engage in sustainability practices and do such things because they know that this is what their target investors want.

As evidenced by the signatories to the PRI, the RI movement in Europe is strongly driven by asset owners such as pension funds with a strong social contract and political agenda. As the financial industry started to become more transparent, these investors—oftentimes having the obligation to make public statements on their investment policies—have increasingly included questions on how sustainability issues are taken into consideration in the investment processes of asset managers when making requests for proposals (Gond & Piani, 2012). Because these clients are driving demand due to ethical considerations and have a reputational benefit to gain, it is expected that these types of investors will tend to support RI funds even when they are underperforming; that is, they will not easily sell their shares in funds whose sustainability strategy they are advocates of, making the money flow of such funds less volatile. In addition to this, the fact that RI is long-term oriented provides asset managers with a longer time frame on which to be measured; hence, clients take longer to pull out of bad performing funds. Such investor motivations illustrate why ESG information is taken into account and how such types of investors who seek to influence firms—consequential investors—and/or use sustainable investments as a mechanism of enhancing their own social identity—expressive investors—are more likely to pursue a long-term investment approach (Busch et al., 2015).

However, similar to the information asymmetry argument, it is only at high levels of screening (a substantive level of RI) that this can occur. A high level of screening is related to a high level of “ethicalness,” which translates into a higher ability to attract and retain investors with such considerations. Without this, investors will maintain preference over a well-diversified portfolio. In addition, investors may even penalize RI funds that are merely “greenwashing” or doing RI symbolically and not substantively. Thus, money flow volatility will be lower for funds with a lower screening intensity as these funds, being closer to traditional funds, will be favored by typical investors. Money flow volatility increases as more and more screens are added as the fund begins to decrease its possibilities of diversification without having a clear ethical identity. However, from a certain level of screening, investors become committed to the fund thus bringing stability to the inflow and outflow of money.

Thus, I put forward the below hypothesis:

Hypothesis 2 (H2): The relationship between RI screening intensity and money flow volatility is curvilinear. Money flow volatility increases as screening intensity increases yet will begin to decrease from a point wherein RI practices become substantive.

Data set and Method

Data set

I test my hypotheses on a data set of European RI equity mutual funds that employ screening practices. Such a screening process may be a very “black box” approach wherein financial analysts make use of a list of exclusions provided by external ratings agencies or internal sustainability analysts. It may also be more integrated wherein financial analysts work together with internal or external sustainability analysts during different phases of the investment process to determine inclusions or eliminations; in some such cases, the same person(s) are responsible for both sustainability and financial analyses. Funds that fall within this latter approach tend to perform more in-depth research and engage in dialogue with investee firms regarding sustainability issues. Screening practices are thus widely heterogeneous. Processes range from simple to highly complex and implementation can range from ad hoc to highly systematized. Given that 38% and 7% of all RI assets in Europe are negatively and positively screened, respectively (Eurosif, 2010), RI screening mutual funds represent an important sub-group of RI.⁷

I manually constructed the data set using a primary list of 529 RI Mutual Funds domiciled in Europe as identified by Eurosif. Of these funds, information on the screening criteria is available for 263 funds and is provided by Avanzi.⁸ To ensure homogeneity, I focus mainly on equity funds by eliminating funds that identified themselves by name as bond funds as well as those with more than 85% of their assets as non-equity.⁹ This focus on equity mutual funds is in line with previous studies (Lee, Humphrey, Benson, & Ahn, 2010; Renneboog et al., 2008a). Furthermore, I eliminated funds that had more than 80% of their assets invested outside of Europe.¹⁰ The data were then adjusted for outliers and errors relating to fund age, fund manager years, and net asset value (NAV). Finally, funds engaged in short-selling were eliminated. Ultimately, I reach a final unbalanced panel of 187 European RI screening mutual funds invested mainly in equities.

Historical data were provided directly to me by Morningstar and includes monthly fund returns, monthly NAVs, domicile, fund age, sectors, geographies, and securities splits, as well as Morningstar investment styles from December 2002, the earliest available date of complete coverage, to March 2012. I focus on the 9-year period of April 2003 to March 2012, separating these into three periods of April 2003 to March 2006 (“Period 1”), April 2006 to March 2009 (“Period 2”), and April 2009 to March 2012 (“Period 3”). Informally, Period 2 captures the U.S. mortgage crisis and Period 3 captures the European debt crisis. This brings the total number of observations for monthly fund returns to 20,304 (108 months × 187 funds). Finally, data for the Fama–French and Carhart factors for European equities are obtained from the online Kenneth French Data Library. Table 1 summarizes the sources and uses of data.

Table 1.

Sources and Uses of Data.

Data	Uses	Source
Initial list from Eurosif	529 funds	Taken from http://www.eurosif.org/
Historical data coverage from Morningstar	April 2003 to March 2012 (108 months)	Data provided to the researcher by Morningstar
Funds for which screening criteria from Avanzi available	263 funds	Data provided to the researcher by Avanzi
Final number after adjustments for outliers	187 funds
Historical data for Fama–French and Carhart models	April 2003 to March 2012 (108 months)	Taken from Kenneth French website: www.mba.tuck.dartmouth.edu/pages/faculty/ken.french/data_library.html
Historical data for Market return (MSCI Europe Index)	April 2003 to March 2012 (108 months)	Taken from MSCI Europe website: www.mscibarra.com/products/indices/international_equity_indices/gimi/stdindex/performance.html
Historical data for Risk Free rate	April 2003 to March 2012 (108 months)	Taken from Bundesbank website: http://www.bundesbank.de/Navigation/EN/Statistics/Time_series_databases/Macro_economic_time_series/its_details_value_node.html?tsId=BBK01.WZ9807&listId=www_s140_it03a/
Total monthly fund return observations (187 funds × 108 months)	20,304 fund month observations

Note. MSCI = Morgan Stanley Capital International.

Figure 1 illustrates the growth of RI funds based on the inception dates of the original sample of 529 funds. The figure shows that the RI boom happened in the late 1990s. As such, the final panel I take focuses on a more recent time period of high but rather stable growth and captures the period of the current financial crisis, which has not been previously examined in the literature. The data set is unique compared with previous studies in that it is focused on Europe,¹¹ the most relevant geography for RI. This data set is the most complete European data set to the best my knowledge.

Figure 1.

Number of European responsible investment mutual funds per year based on inception date.

RI Screening Intensity and Idiosyncratic Risk

Independent variables

To investigate the relationship between an RI fund’s selectivity and idiosyncratic risk, I use Screening Intensity¹² as the independent variable, which is the number of screens, either positive or negative, implemented by the fund (1 if the screen, either negative or positive, was implemented, 0 otherwise).¹³ The Avanzi database provides 24 screening criteria. These are divided into negative screening criteria where funds exclude investments in these areas (16 criteria) and positive screening criteria where funds are required to include investments in these areas (eight criteria). I then categorize the criteria (provided in Table 2) into four broader mutually exclusive areas, namely, ESG and CBI, with four, six, three, and 11 criteria, respectively. I follow in the same vein as previous studies (cf. Barnett & Salomon, 2006; Lee et al., 2010; Renneboog et al., 2008a) who use this measure and contend that doing so provides a more accurate and complex picture of the variation between RI funds as opposed to previous studies that examined the dichotomy between RI and non-RI funds. This study is an improvement over the aforementioned studies in the use of more screening criteria thereby allowing for more variation in screening intensity.¹⁴

Table 2.

Responsible Investment Screening Criteria Provided by Avanzi.

	Screening criteria	Type of screen^a	% of funds with screen
	Environmental (four screens)
1	Excludes excessive negative impact	Negative	33.21
2	Excludes products dangerous to health or environment	Negative	29.46
3	Includes innovative and beneficial products and services for the environment	Positive	21.43
4	Includes environmental protection	Positive	26.60
	Social (six screens)
5	Excludes human rights violations	Negative	51.43
6	Excludes labor rights violations	Negative	40.18
7	Excludes oppressive regimes	Negative	26.79
8	Includes human rights protection (measures to prevent and control human rights violations)	Positive	58.21
9	Includes promotion of economic and social development of local communities	Positive	49.29
10	Includes innovative and beneficial products and services for the quality of life (e.g., health care, social housing)	Positive	13.93
	Governance (three screens)
11	Includes good corporate governance	Positive	56.61
12	Includes responsible management of relations with customers (policies management and monitoring systems in terms of transparent communication with customers, product safety, fair competition, responsible advertisement, etc.)	Positive	52.00
13	Includes responsible management of employees (policies, management monitoring systems)	Positive	26.07
	Controversial business involvement (11 screens)
14	Excludes firearms	Negative	70.54
15	Excludes weapons and military contracting	Negative	72.68
16	Excludes nuclear energy	Negative	50.89
17	Excludes tobacco	Negative	70.00
18	Excludes gambling	Negative	45.89
19	Excludes pornography	Negative	46.43
20	Excludes alcohol	Negative	49.11
21	Excludes animal testing	Negative	28.93
22	Excludes factory farming	Negative	16.61
23	Excludes furs	Negative	10.71
24	Excludes genetically modified organisms	Negative	28.93

Eight positive screens; 16 negative screens.

The information on screening is the outcome of a long-term evaluation process and has not changed over time. In line with the notion that RI funds examine more long-term sustainability issues, it is useful to understand screening criteria as part of a long-term strategy of a fund rather than one that frequently changes. Given that the funds in the data set can be considered to be quite young with a mean age of 8 years, it is reasonable to assume that these screening strategies have not changed dramatically.

Dependent variables

To derive measures of risk, I first take measures of financial performance, in particular Risk-Adjusted Performance (RAP), Fama–French model alphas (FF Alpha; Fama & French, 1993), and Carhart model alphas (Car Alpha; Carhart, 1997). I also examine Sharpe (Sharpe) and Information ratios (InfoR) as additional analyses.

I construct the monthly RAP using the capital asset pricing model (CAPM) methodology (cf. Sharpe, 1964) in line with Barnett and Salomon (2006) wherein RAP is defined as the average monthly return, measured as the percentage change in a fund’s market value from the beginning to the end of a given month, adjusted by the fund’s specific beta. It is the fund’s return over and above what is expected based upon its beta. Specifically,

R A P_{i t} = (R_{i t} - R f_{t}) - β_{i} \times (R m_{t} - R f_{t}),

where R is the return on fund i in month t; Rf is the risk-free rate of return in month t, in this case, the historical monthly returns of the 6-month German treasury bond; Rm is the return on the market portfolio in month t, in this case, the historical monthly returns of the MSCI Europe Index; and β is the beta of fund i, in this case calculated as a regression from returns on the market index. Following Lee et al. (2010), I use a moving 3-year beta to address the significant time variation in beta estimates and to make it better aligned with the 3-year minimum investment horizon typically required for equity funds. I then compute annualized RAPs for three periods of 3 years.¹⁵

The Fama–French and Carhart alphas (a measure of abnormal return) are the intercept terms for fund i in month t from the following ordinary least squares regression equations:

R_{i t} - R f_{t} = F F A l p h a_{i t} + β 1 \times M k t_{t} + β 2 \times S M B_{t} + β 3 \times H M L_{t} + ε_{i t},

\begin{array}{l} R_{i t} - R f_{t} = C a r h a r t A l p h a_{i t} + β 1 \times M k t_{t} + β 2 \times S M B_{t} \\ + β 3 \times H M L_{t} + β 4 \times W M L_{t} + ε_{i t}, \end{array}

where R and Rf are as described above; Mkt is excess return on the market, SMB (“Small minus Big”) is the return on the mimicking size portfolio, HML (“High minus Low”) is the return on the mimicking book-to-market portfolio and WML (“Winners minus Losers”) is the return on the mimicking momentum factor. I compute annualized alphas for each of the 3-year periods.¹⁶

Idiosyncratic risk cannot be observed directly and a proxy needs to be used. I follow in the same vein as previous studies (Casavecchia & Hulley, 2010; Lee & Faff, 2009; Lee et al., 2010; Malkiel & Xu, 2006) and use the residual variance (the standard deviation of the residuals) of the estimated Fama–French (FFResa) and Carhart (CarResa) model residuals. The residual captures the deviation of the sample from the estimated (theoretical) function value and provides an observable estimate of the unobservable statistical errors. The mean square error (MSE) as reported by Stata is the variance computed from the sum of squares of the residuals, which adjusts for the influence of the end points in a regression function. As per Lee et al. (2010), I use the 3-year standardized residual variance. As a secondary measure, I compute the 3-year annualized standard deviation of RAP. As RAP is beta-adjusted (thus eliminating the market risk), the standard deviation provides a proxy measure of the volatility of excess returns.¹⁷

Control variables

The control variables used are similar to those of Barnett and Salomon (2006) and are typically employed in studying mutual funds. Fund age (FundAge) is the number of months since the inception of the fund, which addresses the learning effect of RI funds (Bauer, Koedijk, & Otten, 2005). Fund size (NAV) has also been found to affect fund performance (USSIF, 2010). Malkiel and Xu (2006) find that large firms are associated with lower idiosyncratic risk. Investments in equity are typically associated with higher levels of risk (Wermers, 2002). Given that firm risks are significantly affected by its industry association (Jo & Na, 2012), I control for sector differences as well as geographical differences. I also control for the fact that the residual variance may be a measure of the aggressiveness of fund strategies (Casavecchia & Hulley, 2010) and include the investment styles of Morningstar (Small growth, Small blend, Small Value, Mid-growth, Mid-blend, Mid-value, Large-growth, Large-blend, and Large-value). This addresses the need to disentangle the effect of sustainability performance of the investee firms from the fund manager’s performance (Lee & Faff, 2009). Finally, I include period fixed effects dummies to control for the time variation. Each period dummy is the difference in the conditional expected value of the dependent variable between the base year t = 1 and the year t = j.

The idiosyncratic risk equation is estimated as an unbalanced pooled cross-sectional regression:

\begin{matrix} I R i s k_{i t} = α_{i t} + β 1 \times S c r e e n i n g I n t e n s i t y_{i} + β 2 \times F u n d A g e_{i t} \\ + β 3 \times N A V_{i t} + β 4 \times % E q u i t y_{i t} + β 5 \times % E u r o p e_{i t} \\ + β 6 \times [S e c t o r V a r i a b l e s_{i t}] + [M S C a t_{i}] + [P e r i o d f i x e d e f f e c t s] + ε_{i t}, \end{matrix}

where IRisk is the 3-year annualized idiosyncratic risk measure for fund i in period t, Screening Intensity is the screening intensity for fund i measured as the number of screens, FundAge is the number of months since inception of fund i at the beginning of period t, NAV is the standardized average 3-year NAV in Euros of fund i, %Equity is the average 3-year percentage of investments in Equities, %Europe is the average 3-year percentage of investments in Europe, Sector Variables are the average 3-year percentage of investments in the financial services, health care, real estate, energy and utilities, information, and manufacturing sectors, and MSCat relates to the Morningstar investment styles. To test the hypothesis that the relationship between RI screening intensity and idiosyncratic risk is curvilinear, the square of Screening Intensity is introduced into the regressions.

RI Screening Intensity and Money Flow Volatility

To determine whether screening practices related to sustainability affect the volatility of money flow, I first calculate money flow using data from December 2002 to March 2013. This is the variation in percentage of the fund size due to the money inflow or outflow. I was provided with monthly fund size by Morningstar. I follow Qian (2006) and Renneboog et al. (2008a) and define the growth rate of the fund size beyond asset appreciation—Money Flow—as follows:¹⁸

M o n e y F l o w_{i, t =} \frac{(T N A_{i, t} - T N A_{i, t - 1} \times (1 + R_{i, t}))}{T N A_{i, t - 1}},

where TNA is the total net assets of fund i in month t and R is the return of fund i in month t. I use both unadjusted returns R and RAP, noting that investors consider excess returns rather than RAP (Del Guercio & Tkac, 2002; Ippolito, 1992). I then compute the standard deviation of money flow (SD Flow) for both R and RAP as the dependent variable for each of the three non-overlapping 3-year periods. Following Qian (2006), I control for past fund returns, fund fees, and investment style, which have been previously shown to affect money flow and estimate the below equation:

\begin{matrix} S D F l o w_{i t} = α_{i t} + β 1 \times {[S c r e e n i n g V a r i a b l e]}_{i} + β 2 \times [F u n d P a s t R e t u r n_{i, t}] \\ + β 3 \times F u n d f e e_{i} + β 4 \times [M S C a t_{i}] + ε_{i t} \end{matrix}

where the Screening Variable is the screening intensity, the total amount of negative criteria, or the totals of each criteria type. The Fund Past Return is composed of the average monthly cumulative return of fund i at the end of each period and of the square value of this return.¹⁹

Due to the fact that the relationship of the dependent variable and at least some of the explanatory variables is constant over time (i.e., my main independent variable Screening Intensity is not time-varying) I chose to use a pooled cross-sectional analysis that combines time-series regressions for several cross-sections. This is also useful to increase the number of observations and to solve the imbalance between the number of explanatory variables and the number of firms.

All models are run, checked, and corrected for normality and heteroscedasticity using Stata software. Following Lee et al. (2010), weighted least squares with robust standard errors was used to estimate the Fama–French and Carhart equations using the reciprocal of the residuals as the weights and ordinary least squares with robust standard errors was used for all other equations. Table 3 summarizes the list and operationalization of variables.

Table 3.

List of Variables.

Variable	Operationalization
Independent variables
Screening Intensity	Total number of screens (ESG and CBI) whether positive or negative
Envtot	Total number of environmental screens
Soctot	Total number of social screens
Govtot	Total number of governance screens
CBItot	Total number of controversial business involvement screens
Negative	Total number of negative screens
Measures of risk
FFResa	Standard deviation of the residuals of the Fama–French model annualized for each non-overlapping 3-year period
CarResa	Standard deviation of the residuals of the Carhart model annualized for each non-overlapping 3-year period
SDRa ^a	Standard deviation of Risk-adjusted performance annualized for each non-overlapping 3-year period
Beta	Regression of fund returns over the index MSCI Europe for each non-overlapping 3-year period
SDa^a	Standard deviation of unadjusted returns annualized for each non-overlapping 3-year period
Measure of money flow volatility
SD Flow	Standard deviation of the money inflow or outflow for each non-overlapping 3-year period
Measures of performance
RAP^b	Annualized risk-adjusted performance for each non-overlapping 3-year period
FF Alpha	Alpha from the Fama–French model annualized for each non-overlapping 3-year period
Car Alpha	Alpha from the Carhart model annualized for each non-overlapping 3-year period
Sharpe^c	Sharpe ratio
InfoR^d	Information ratio
Control variables
FundAge	Number of months since inception at the end of each non-overlapping 3-year period
NAV	Average of net asset value in euros at the end of the month for each non-overlapping 3-year period
% Equity	Average %equity at the end of the month for each non-overlapping 3-year period
% Europe	Average % of equity invested in Europe at the end of the month for each non-overlapping 3-year period
% FinServ	% of equity invested in financial services at the end of the month
% Health care	% of equity invested in health care at the end of the month
% RealEstate	% of equity invested in real estate at the end of the month
% EnergyUtilities	% of equity invested in energy and utilities at the end of the month
% Info	% of equity invested in technology and communication Services at the end of the month
% Mfg	% of equity invested in basic materials, consumer (cyclical), consumer (defensive), and industrials at the end of the month
MSCat	Morningstar investment styles

Note. ESG = environmental, social, governance; CBI = controversial business involvement; MSCI = Morgan Stanley Capital International; NAV = net asset value; RAP = risk-adjusted performance.

Takes SD for each year starting April and then annualizes for 3 years = SD3y × 3(1/2).

First take geometric product of monthly returns. Then, 3 years annualized return is [(RAP3yTotal)^(1/3)] – 1.

Sharpe ratio: Take Rfund – Rf, average then Sharpe = Average return / SD of excess return.

Information ratio: Take Rfund – RmMSCI, average then InfoR = Average return / SD of excess return.

Results

Descriptive Analysis

The funds are domiciled in 13 European countries with the majority of the funds coming from Luxembourg (20.86%), the United Kingdom (19.79%), and France (18.72%). The mean age of the funds is 7.9 years old, with the oldest fund established on April 1983 and the youngest on November 2008. The fund size, as represented by NAV, is highly fragmented ranging from 0.43 million Euros to 1.4 billion Euros. As expected, the majority of the funds’ investments are in Europe (75.14%) with the second largest geography being the United States. A majority of investments (83.21%) are in equities of which the largest percentage is in the manufacturing sector followed by financial services and the information sector. Morningstar has categorized 50.33% of the funds as large-blend funds. Table 4 presents the distribution of funds, Table 5 presents the correlation matrix, and Table 6 presents the descriptive statistics in detail. The results were tested for multicollinearity.²⁰ The funds implement on average 10 out of 24 screens with more than 70% of the funds excluding firearms, weapons and military contracting, and tobacco.

Table 4.

Distribution of Funds According to Domicile and Morningstar Category.

Domicile	%	Morningstar category	%
Austria	30.00	Small growth	0.00
Belgium	30.00	Small blend	0.00
France	18.72	Small value	0.00
Germany	5.88	Mid-growth	11.92
Ireland	0.53	Mid-blend	0.66
Italy	2.14	Mid-value	0.00
Luxembourg	20.86	Large-growth	16.56
Netherlands	2.67	Large-blend	50.33
Norway	0.53	Large-value	20.53
Spain	2.14
Sweden	12.83
Switzerland	3.21
United Kingdom	19.79

Table 5.

Correlation Matrix.

Variable	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18
1. SDRa	1
2. FFResa	.5469	1
3. CarResa	.5463	.9962	1
4. Screening Intensity	.1550	.1771	.1840	1
5. FundAge	.0491	.0324	.0231	−.0042	1
6. NAV	−.1245	−.1242	−.1216	−.0286	−.0433	1
7. % Equity	.1753	.7419	.7446	.0604	−.0636	−.0303	1
8. % Europe	.0321	−.2470	−.2642	−.2232	.0218	−.0386	−.1507	1
9. % FinServ	−.1739	−.3467	−.3566	−.0119	−.1084	−.0704	−.1765	.4069	1
10. % Healthcare	−.1135	−.1264	−.1142	.0526	.1370	−.0256	−.0004	−.1527	−.1090	1
11. % RealEstate	.2608	.2430	.2194	.0519	.2022	−.1119	.0739	.1532	.1384	−.2202	1
12. % EnergyUtilities	.1294	.0069	.0031	−.0541	−.0490	.1172	−.0355	−.0306	−.1863	−.2277	−.0439	1
13. % Info	−.1162	−.0058	−.0106	−.0553	−.1227	−.0832	−.0412	−.2117	−.0929	.0063	−.2213	−.1732	1
14. % Mfg	.1311	.3033	.3149	.0464	.0990	.0534	.1815	−.1488	−.6146	−.1757	−.0135	−.3243	−.3383	1
15. Mid-blend	−.1035	−.0289	−.0410	−.0665	−.0204	.0078	.0601	.0826	.0757	.0283	−.0382	.0245	−.0530	−.0560	1
16. Large-growth	−.0892	−.1891	−.1850	.0614	.0379	.3107	−.1392	.1231	.0513	−.1204	−.0488	.2312	−.1477	−.0552	−.0444	1
17. Large-blend	.1794	.1908	.1801	−.1420	.2120	−.3256	.1360	.2609	.0639	−.1141	.2149	−.1873	.0295	.0775	−.0902	−.4511	1
18. Large-value	−.1007	.0317	.0384	−.0011	−.2092	.1248	.0750	−.3414	−.1738	.1519	−.1376	.1375	.1253	−.0693	−.0526	−.2631	−.5341	1
Minimum	0.01	0.85	0.81	1.00	0.87	0.43	7.85	13.25	0.00	0.00	0.00	0.00	0.00	2.15
Maximum	3.72	9.90	9.54	21.00	316.83	1,408.79	99.82	100.00	42.35	96.90	19.90	51.56	37.97	83.50
M	1.02	4.72	4.68	9.80	95.00	94.27	83.37	75.09	17.95	9.96	1.30	12.65	16.94	41.20
SD	0.67	1.72	1.73	5.40	62.49	174.59	23.82	26.43	8.77	8.11	2.18	7.65	6.03	11.21

Note. NAV = net asset value.

Table 6.

Descriptive Statistics and Relationship to Responsible Investment Screening.

No. of screens	No. of funds	Fund age (months)		NAV (EUR)		% equity		% Europe		% non-manufacturing
No. of screens	No. of funds	Average	SD	Average	SD	Average	SD	Average	SD	Average	SD
1	13	115.19	69.73	140.29	196.47	78.40	22.96	81.39	23.33	58.84	10.89
2	2	101.73	44.52	45.00	27.21	92.27	4.21	80.46	17.72	53.39	6.76
3	7	61.42	34.52	45.92	47.78	79.00	29.66	73.75	27.87	61.59	7.60
4	19	90.68	49.07	86.34	125.26	86.01	22.05	92.06	19.78	60.62	11.28
5	11	92.42	89.93	68.28	82.93	92.92	14.51	61.11	30.89	60.48	17.43
6	10	92.39	59.45	146.83	235.79	83.20	27.55	63.92	24.23	56.34	16.15
7	14	121.97	63.04	29.45	31.24	90.44	18.74	80.21	29.03	55.24	10.04
8	5	73.07	33.70	113.62	117.49	86.78	19.00	77.69	26.78	56.59	15.65
9	14	92.71	55.66	36.02	49.68	77.76	29.15	78.90	22.58	61.58	9.04
10	9	109.13	58.30	37.23	41.90	76.78	25.31	74.21	21.55	63.20	7.12
11	7	65.40	36.73	63.54	48.63	83.30	23.11	60.07	20.43	60.93	5.06
12	13	99.88	61.40	260.66	451.48	81.08	28.19	84.27	23.31	56.94	9.99
13	5	64.58	34.67	39.38	34.65	61.99	35.63	79.52	24.64	64.39	5.62
14	14	131.85	93.52	55.14	53.98	84.96	20.29	78.33	24.95	55.47	8.39
15	13	88.16	61.25	90.61	201.87	88.42	18.58	75.69	21.34	59.51	7.46
16	6	74.18	37.69	227.98	160.82	92.11	8.93	70.94	29.51	58.12	9.95
17	10	61.70	33.09	178.38	200.71	89.15	17.86	54.87	23.24	59.61	8.49
18	2	58.98	36.53	25.54	21.69	45.44	26.57	41.39	9.95	59.93	6.11
19	10	93.58	44.79	58.39	92.88	74.29	29.22	80.16	29.14	63.78	8.98
20	2	199.83	39.53	1.91	0.49	93.72	6.12	47.33	19.14	31.65	12.57
21	1	58.77		83.36		94.75		36.64		32.92
All funds	187	94.89	62.47	94.21	174.41	83.21	24.07	75.14	26.43	58.83	11.22

Note. NAV = net asset value.

Substantive RI Practices and the Reduction of Idiosyncratic Risk

Table 7 presents the regression results for the relationship between Screening Intensity on three measures of idiosyncratic risk for the simple linear model (Model 1) and the model with the squared Screening Intensity variable (Model 2). My findings suggest strong support for a positive linear relationship between Screening Intensity and Idiosyncratic Risk at a 99% confidence level for both the Fama–French and Carhart model residual variances and at a 95% confidence level when the standard deviation of RAP is used as the idiosyncratic risk measure and stronger support for a curvilinear relationship (an inverted U-shape) at a 99% confidence level across all three measures of idiosyncratic risk. The increase in fit from the linear to the curvilinear models suggests that the latter presents a more robust explanation for the relationship, thereby supporting H1. This curvilinear relationship is illustrated in Figure 2. Funds with the least amount of screens have lower risk; the risk increases with each screen at an average of 0.0526, 0.0529, and 0.0383 for the Fama–French, Carhart, and RAP standard deviation models, respectively, until reaching a peak of 13 screens for the Fama–French and Carhart models and 12 screens for the RAP standard deviation model. The risk then begins to decrease by a slightly lesser average of 0.0354, 0.0351, and 0.0257, respectively. A fund that screens at the highest levels begins to have a similar risk as those that screen at the lowest level.²¹

Table 7.

Responsible Investment Screening Intensity and Idiosyncratic Risk.

	Fama–French residual variance		Carhart residual variance		SD of RAP (log)
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
SI	0.0239 (2.58)*	0.1141 (3.07)***	0.0242 (2.62)***	0.1145 (3.09)***	0.0149 (2.48)**	0.0798 (3.18)***
SI ²		−0.0044 (−2.63)***		−0.0044 (−2.66)***		−0.0031 (−2.77)***
FundAge	0.0000 (0.07)	0.0001 (0.22)	−0.0000 (−0.02)	0.0001 (0.13)	−0.0001 (−0.33)	−0.0001 (−0.17)
NAV	−0.1279 (−4.52)***	−0.1292 (−4.64)***	−0.1301 (−4.67)***	−0.1314 (−4.79)***	−0.0466 (−2.52)**	−0.0476 (−2.64)***
% Equity	0.0479 (24.25)***	0.0474 (24.24)***	0.0478 (24.2)***	0.0474 (24.22)***	0.0037 (2.7)***	0.0034 (2.53)**
% Europe	−0.0057 (−2.84)***	−0.0062 (−3.15)***	−0.0065 (−3.29)***	−0.0070 (−3.62)***	0.0001 (0.1)	−0.0001 (−0.12)
% FinServ	−0.0451 (−7.13)***	−0.0453 (−7.08)***	−0.0461 (−7.36)***	−0.0464 (−7.3)***	−0.0170 (−4.23)***	−0.0171 (−4.3)***
% Healthcare	−0.0498 (−3.82)***	−0.0474 (−3.64)***	−0.0478 (−3.67)***	−0.0454 (−3.49)***	−0.0085 (−1.18)	−0.0067 (−0.97)
% RealEstate	0.1620 (4.89)***	0.1465 (4.31)***	0.1395 (4.51)***	0.1239 (3.99)***	0.0727 (4.81)***	0.0615 (3.82)***
% EnergyUtilities	−0.0035 (−0.33)	−0.0028 (−0.26)	−0.0053 (−0.5)	−0.0045 (−0.43)	0.0025 (0.43)	0.0031 (0.53)
% Info	0.0084 (0.98)	0.0034 (0.4)	0.0038 (0.45)	−0.0010 (−0.12)	−0.0081 (−1.37)	−0.0117 (−1.96)*
Mid-blend	−0.4105 (−1.17)	−0.4299 (−1.22)	−0.6464 (−1.5)	−0.6658 (−1.54)	−2.1176 (−1.9)*	−2.1315 (−1.91)*
Large-growth	−0.2956 (−1.31)	−0.1802 (−0.81)	−0.2848 (−1.24)	−0.1692 (−0.75)	−0.2267 (−1.65)*	−0.1437 (−1.02)
Large-blend	−0.1611 (−0.72)	−0.1043 (−0.48)	−0.1766 (−0.78)	−0.1197 (−0.54)	−0.1181 (−0.89)	−0.0772 (−0.58)
Large-value	−0.1837 (−0.83)	−0.1172 (−0.54)	−0.1924 (−0.86)	−0.1259 (−0.57)	−0.2428 (−1.88)*	−0.1950 (−1.5)
Period 2	0.8072 (5.93)***	0.7782 (5.65)***	0.7785 (5.78)***	0.7494 (5.48)***	0.6363 (6.5)***	0.6155 (6.37)***
Period 3	0.3162 (2.07)**	0.2830 (1.86)*	0.2584 (1.74)*	0.2251 (1.52)	−0.0367 (−0.33)	−0.0606 (−0.56)
Constant	2.0590 (4.86)***	1.8424 (4.21)***	2.2537 (5.41)***	2.0367 (4.71)***	−0.1651 (−0.55)	−0.3211 (−1.07)
R ²	.7391	.7432	.7393	.7435	.3991	.4103
Observations	341	341	341	341	341	341

Note. $I d i o s y n c r a t i c R i s k_{i t} = α_{i t} + β 1 \times S c r e e n i n g I n t e n s i t y_{i} + β 2 \times F u n d A g e_{i t} + β 3 \times N A V_{i t} + β 4 \times % E q u i t y + β 5 \times % E u r o p e + β 6 \times [S e c t o r V a r i a b l e s_{i t}] + M S C a t + P e r i o d f i x e d e f f e c t s + ε_{i t},$ ,

where Idiosyncratic Risk is the 3-year annualized idiosyncratic risk measure for fund i in period t, SI is the screening intensity for fund i measured as the number of screens, SI² is the squared variable of Screening Intensity; NAV is the standardized average 3-year net asset value in Euros of fund i, Age is the number of months since inception of fund i at the beginning of period t, %Equity is the average 3-year percentage of investments in Equities, %Europe is the average 3-year percentage of investments in Europe, Sector variables are the average 3-year percentage of investments in the financial services, health care, real estate, energy and utilities, information, and manufacturing (dropped) sectors, MSCat related to the Morningstar investment styles. Numbers in parentheses are the values for the t test. RAP = risk-adjusted performance; SI = Screening Intensity; NAV = net asset value.

p < .10. **p < .05. ***p < .01.

Figure 2.

Curvilinear relationship between responsible investment screening intensity and idiosyncratic risk.

Type of RI Screens and Idiosyncratic Risk

In addition to the heterogeneity in the intensity of screening, RI funds vary greatly in the type of screens applied to their investments. Previous empirical work has found that some types of social responsibility are linked to higher financial performance than others and that the type of screening may enhance or erode performance (Barnett & Salomon, 2006; Delmas & Blass, 2010; Lee et al., 2010; Renneboog et al., 2008a). Furthermore, it has been said that negative screenings have been initially favored because it is often easier to agree on what constitutes a problem than to agree on what constitutes excellence (Delmas & Blass, 2010). Table 8 presents the regression results of the relationship between Criteria type and Idiosyncratic Risk. Model 1 introduces the total number of screens for each group of environmental (Envtot), social (Soctot), governance (Govtot), and CBI (CBItot) criteria while Model 2 includes the variable Negative, which is the total number of negative screens. In line with previous studies (Hong & Kacperczyk, 2009; Jo & Na, 2012; Langbein & Posner, 1980) I find strong evidence that a higher number of exclusions of CBI and that negative exclusions are both positively related to idiosyncratic risk.²²

Table 8.

Individual Effects of RI Screening and Idiosyncratic Risk.

	Fama–French residual variance		Carhart residual variance		SD of RAP (log)
	Model 1	Model 2	Model 1	Model 2	Model 1	Model 2
negative		0.0409 (3.45)***		0.0421 (3.57)***		0.0208 (2.68)***
envtot	0.0124 (0.23)		0.0194 (0.35)		−0.0051 (−0.15)
soctot	−0.0360 (−0.9)		−0.0340 (−0.84)		0.0126 (0.46)
govtot	−0.0115 (−0.17)		−0.0328 (−0.47)		−0.0039 (−0.09)
cbitot	0.0706 (3.25)***		0.0718 (3.26)***		0.0282 (2.19)**
FundAge	−0.0000 (−0.05)	−0.0000 (−0.1)	−0.0001 (−0.12)	−0.0001 (−0.2)	−0.0002 (−0.45)	−0.0002 (−0.48)
NAV	−0.0859 (−2.3)**	−0.1011 (−3.26)***	−0.0821 (−2.19)**	−0.1021 (−3.32)***	−0.0335 (−1.47)	−0.0356 (−1.76)*
% Equity	0.0485 (24.83)***	0.0483 (25.18)***	0.0486 (24.66)***	0.0482 (25.12)***	0.0039 (2.8)***	0.0040 (2.88)***
% Europe	−0.0059 (−2.94)***	−0.0055 (−2.79)***	−0.0068 (−3.4)***	−0.0063 (−3.23)***	0.0000 (0.05)	0.0001 (0.07)
% FinServ	−0.0423 (−6.46)***	−0.0433 (−6.97)***	−0.0428 (−6.62)***	−0.0443 (−7.22)***	−0.0159 (−3.87)***	−0.0160 (−0.4)***
% Healthcare	−0.0504 (−3.79)***	−0.0486 (−3.75)***	−0.0484 (−3.64)***	−0.0466 (−3.61)***	−0.0083 (−1.16)	−0.0078 (−1.1)
% RealEstate	0.1519 (4.48)***	0.1572 (4.66)***	0.1286 (4.02)***	0.1344 (4.31)***	0.0702 (4.6)***	0.0705 (4.62)***
% EnergyUtilities	0.0003 (0.04)	−0.0011 (−0.11)	−0.0007 (−0.07)	−0.0028 (−0.27)	0.0040 (0.68)	0.0036 (0.61)
% Info	0.0064 (0.69)	0.0083 (0.98)	0.0017 (0.18)	0.0038 (0.44)	−0.0089 (−1.45)	−0.0084 (−1.41)
Mid-blend	−0.1798 (−0.49)	−0.3562 (−1.03)	−0.4141 (−0.94)	−0.5881 (−1.38)	−2.0865 (−1.86)*	−2.1043 (−1.89)*
Large-growth	−0.2284 (−1.01)	−0.2473 (−1.1)	−0.2172 (−0.92)	−0.2347 (−1.03)	−0.1944 (−1.38)	−0.2048 (−1.5)
Large-blend	−0.1314 (−0.59)	−0.1165 (−0.52)	−0.1526 (−0.67)	−0.1284 (−0.57)	−0.1033 (−0.79)	−0.1090 (−0.83)
Large-value	−0.1885 (−0.87)	−0.1729 (−0.8)	−0.2059 (−0.92)	−0.1799 (−0.81)	−0.2387 (−1.88)*	−0.2456 (−1.93)*
Period 2	0.8052 (5.88)***	0.8099 (5.97)***	0.7777 (5.75)***	0.7811 (5.83)***	0.6392 (6.47)***	0.6389 (6.53)***
Period 3	0.3329 (2.17)**	0.3366 (2.23)**	0.2776 (1.85)*	0.2794 (1.9)*	−0.0259 (−0.23)	−0.0262 (−0.24)
Constant	1.8050 (4.36)***	1.7937 (4.37)***	1.9704 (4.85)***	1.9714 (4.87)***	−0.2593 (−0.86)	−0.2447 (−0.8)
R ²	.7450	.7430	.7462	.7437	.4018	.4010
Observations	341	341	341	341	341	341

where variables are similar to those in Table 5 and criteria variables are the total of environmental screens (envtot), social screens (soctot), governance screens (govtot), and controversial business involvement screens (cbitot) and negative is the total number of negative screens. Numbers in parentheses are the values for the t test. RAP = risk-adjusted performance; NAV = net asset value.

p < .10. **p < .05. ***p < .01.

Control Variables Results

In general, the significant results on Control variables are consistent with previous studies and are both interesting and sensible given the period under study of 2003 to the first quarter of 2012. As expected, a lot of the volatility is accounted for by unexplained factors that are captured by the crisis Period 2 (April 2007 to March 2009) and to a lesser extent, Period 3 (April 2009 to March 2012). In line with the majority of previous findings, I find consistent findings across all models that being a larger fund is negatively related to idiosyncratic risk (similar to Lee & Faff, 2009, and consistent with Malkiel & Xu, 2006), whereas being invested in equities is risk-increasing. Being invested in Europe during the period of investigation appears to have been risk reducing. As a developed capital market, there was less of a boom and bust in Europe as compared with Americas or Asia during this period. Being invested in the health care sector was also risk reducing, the defensive sectors having been the most resilient during the crisis that more severely affected the real estate sector. Surprisingly, however, being invested in financial services appears to have been risk-reducing. It is notable that the influence of screening intensity is significant even when controlling for investment strategies, as represented by the Morningstar investment styles. The investment strategies did not appear to be significantly related to idiosyncratic risk, with some evidence that having mid-blend, large-growth, and large-value strategies were risk reducing, however, only in the RAP standard deviation model.

Substantive RI Practices and the Reduction of Money Flow Volatility

Table 9 presents the regression results for the relationship between screening variables and money flow volatility for the simple linear model (Model 1) and the model with the squared Screening Intensity variable (Model 2) as well as two additional models: one that considers total negative exclusions (Model 3) and another that considers total criteria type (Model 3) for each of the dependent variables: standard deviation of the flow of unadjusted returns and standard deviation of the flow of risk-adjusted returns. My findings suggest support for a positive linear relationship between Screening Intensity and money flow volatility at a 99% confidence level and stronger support for a curvilinear relationship (an inverted U-shape) at a 95% confidence level. The increase in fit from the linear to the curvilinear models suggests that the latter presents a more robust explanation for the relationship, thereby supporting H2. This curvilinear relationship is illustrated in Figure 3. Funds with the least amount of screens have lower money flow volatility; the volatility increases with each screen until reaching a peak of between 13 to 15 screens. The volatility then begins to decrease as a fund screens at high levels.

Table 9.

RI Screening and Money Flow Volatility.

	SD of Rf flow				SD of RAP flow
	Model 1	Model 2	Model 3	Model 4	Model 1	Model 2	Model 3	Model 4
SI	0.0066 (2.7)***	0.0326 (2.87)***			0.0041 (3.87)***	0.0168 (2.68)***
SI ²		−0.0012 (−2.26)**				−0.0006 (−2.03)**
negative			0.0078 (2.44)**				0.0068 (4.37)***
envtot				−0.0070 (−0.49)				0.0079 (1.14)
soctot				−0.0156 (−1.22)				−0.0118 (−1.8)*
govtot				0.0443 (2.33)**				−0.0007 (−0.06)
cbitot				0.0162 (2.93)***				0.0112 (4.06)***
cumulative returns	−0.0626 (−0.54)	−0.0251 (−0.21)	−0.0738 (−0.63)	−0.0297 (−0.27)	0.0805 (3.85)***	0.0781 (3.82)***	0.0801 (3.83)***	0.0807 (3.99)***
cumulative returns²	0.1770 (1.69)*	0.1529 (1.48)	0.1811 (1.74)*	0.1659 (1.58)	0.0086 (1.3)	0.0089 (1.34)	0.0090 (1.37)	0.0106 (1.74)*
Fee	7.9245 (2.01)**	8.9034 (2.29)**	7.5449 (1.92)*	7.1617 (1.78)*	2.1568 (0.95)	2.4879 (1.09)	2.0897 (0.95)	1.2448 (0.54)
Mid-blend
Large-growth	0.0433 (0.58)	0.0707 (0.87)	0.0526 (0.7)	0.0437 (0.58)	−0.0494 (−1.28)	−0.0356 (−0.84)	−0.0369 (−0.94)	−0.0399 (−0.99)
Large-blend	0.1312 (1.92)*	0.1366 (1.93)*	0.1246 (1.76)*	0.1398 (2.05)**	−0.0075 (−0.21)	−0.0044 (−0.12)	−0.0057 (−0.16)	−0.0166 (−0.47)
Large-value	0.0646 (0.86)	0.0757 (0.96)	0.0638 (0.83)	0.0706 (0.92)	−0.0500 (−1.32)	−0.0434 (−1.09)	−0.0458 (−1.18)	−0.0528 (−1.4)
Period 2	0.2638 (3.26)***	0.2779 (3.41)***	0.2577 (3.13)**	0.2783 (3.46)***	0.0312 (1.21)	0.0303 (1.18)	0.0293 (1.14)	0.0278 (1.09)
Period 3	0.1217 (2.55)**	0.1248 (2.61)***	0.1197 (2.48)**	0.1243 (2.59)**	−0.0192 (−0.85)	−0.0197 (−0.86)	−0.0208 (−0.91)	−0.0222 (−0.97)
Constant	−0.0562 (−0.49)	−0.1821 (−1.29)	−0.0294 (−0.25)	−0.0886 (−0.71)	0.0777 (1.4)	0.0223 (0.34)	0.0723 (1.41)	0.1064 (1.65)
R ²	.1627	.1721	.1626	.1732	.1674	.1760	.1822	.1912
Observations	384	384	384	384	383	383	383	383

Note. $\begin{array}{l} S D F l o w_{i t} = α_{i t} + [β 1 \times S c r e e n i n g V a r i a b l e s_{i}] + β 2 \times C u m R e t_{i t} + β 3 \times C u m R e t 2_{i t} + β 4 \times F e e \\ + M S C a t + P e r i o d f i x e d e f f e c t s + ε_{i t}, \end{array}$ ,

where SD Flow is the standard deviation of the flow of fund i in period t, Screening Variables includes the screening intensity for fund i measured as the number of screens (as well as SI², the squared variable of Screening Intensity), negative, the number of negative exclusions for fund i, or totals for environmental, social, governance, and CBI criteria (envtot, soctot, govtot, cbitot, respectively); CumRet is the average cumulative returns for fund i at the end of period t (CumRet2 is the squared variable of CumRet); fee is the fees of fund i; MSCat relate to the Morningstar investment styles. Numbers in parentheses are the values for the t test. RAP = risk-adjusted performance; SI = Screening Intensity.

p < .10. **p < .05. ***p < .01.

Figure 3.

Curvilinear relationship between responsible investment screening intensity and money flow volatility.

I further find that an increase in the number of negative exclusions is positively related to money flow volatility at a 95% level for the standard deviation unadjusted returns model and at a 99% level for the RAP standard deviation model. I also find evidence that CBI exclusions were associated with high levels of money flow volatility at a 99% level for both dependent variables. Finally, I find that cumulative returns, fund fees, large-blend investment strategies, and unexplained factors related to Period 2 and Period 3 were positively related to money flow volatility.

Discussion and Conclusion

This article addresses the need for more comprehensive research on the motivations underlying RI, particularly by examining how a practice with conflicting logics has nevertheless been able to grow and become mainstream at such a staggering pace. By drawing on literatures in finance, CSR, and business ethics and through a quantitative approach using a unique and manually constructed data set, I provide evidence that the two conflicting views regarding the relationship between RI and financial performance can be reconciled when instead I examine measures of risk and nuance the level of screening intensity, which brings to the forefront an argument in favor of substantive RI as key to providing a risk mitigation effect.

I theorize and empirically test the validity of two hitherto anecdotally based beliefs underlying the mainstreaming of the practice, namely, that informational benefits lead to a decrease in idiosyncratic risk at substantive levels of RI and that “stickiness” from ethical considerations and normative legitimacy lead to lower levels of money flow volatility, also at substantive levels of RI. My findings strongly support the two main hypotheses of the article. First, I find empirical evidence of a curvilinear (inverted U-shaped) relationship between RI screening intensity and idiosyncratic risk, pointing toward an illustration of how new (sustainability-related) sources of risk are being integrated more and more in financial market models. These findings evidence the fact that contrary to traditional asset pricing models that propose that only systematic risk matters, idiosyncratic risk matters for RI funds. More importantly, I concur with Derwall, Guenster, Bauer, and Koedijk (2005) and Lee and Faff (2009) in proposing that financial markets factor in the economic consequences of sustainability into current share prices. My results provide a substantial addition in that they show that the integration of sustainability issues within financial models is by no means a “yes” or “no” question. Instead, sustainability-related issues as a source of risk depend largely on the depth of information that the fund manager possesses and this integration does not occur in all types of RI funds. Furthermore, I find that negative screening and screening for CBI is positively related with idiosyncratic risk, supporting the notion that reducing these defensive stocks reduces possibilities for diversification and addresses calls for a closer examination of screening strategies (Barnett & Salomon, 2006).

Second, I find evidence of client “stickiness” at substantive levels of RI, illustrated by the fact that the level of screening intensity may affect the volatility of money flow. Similar to my results on idiosyncratic risk, I find a significant curvilinear (inverted U-shaped) relationship between screening intensity and money flow volatility. These findings evidence the fact that a fund’s level of “ethicalness” provides high levels of asset stability. Similar to my results on idiosyncratic risk, I highlight that these results show that the integration of sustainability issues within financial models is not straightforward. Instead, sustainability as a source of risk (in this case, in terms of money flow volatility) depends largely on the ethical attributes of the fund that provides its legitimacy. Again and importantly, this does not occur in all types of RI funds.

These results are an initial support for a better understanding of the controversial yet rapidly growing practice of RI in general and RI screening strategies in particular. The findings stress how traditional models of finance remain dominant yet new ideas related to sustainability are gradually penetrating and manifesting their presence. By moving away from the performance debate and the dichotomy between RI and non-RI funds and instead, focusing on the more complex process of sustainability integration, these results contribute toward a deeper theoretical understanding of a fast-growing phenomenon and provide evidence that extant models in finance that have neglected sustainability issues are incomplete and that there is a need to consider sustainability-related sources of risk, importantly reconciling sustainability-related RI practices with traditional portfolio-based financial models.

My findings present an interesting case for asset managers. Unlike previous studies that have hypothesized that doing RI can only either be “good” or “bad,” I illustrate how engaging in RI can be meaningful but with certain conditions. It makes the investor realize that screening strategies are not simply a marketing tool, but rather one that affects risk. Thus, depending on the trade-offs an asset manager wants to make based on his risk appetite and level of “ethicalness,” he must position himself clearly either on the lower or higher end of the continuum. Perhaps he may want to be closer to a traditional investor yet still be doing a little of RI. Perhaps he may want to really bite the bullet and become truly sustainable. Both can be risk reducing. However, being unsure of an RI strategy and remaining in the middle is problematic. I thus illustrate that fund managers do not need to believe in traditional finance thought that RI makes “his hands tied.” Rather, engaging in RI remains attractive because of its unique risk profile, which, however, supports a case for a deep level of commitment, that is, the commitment of the fund manager to engage with its investee firms and have a better understanding of sustainability issues. Given that doing RI may produce information externalities, implications of this in practice point toward increased standardization of performance measures, the creation of models and tools, and increased specialization of sustainability roles through training and other knowledge-building practices. Such recommendations are in line with those of Busch et al. (2015) who posit that unless ESG measures are also related to changes in firms’ underlying economic fundamentals, ESG data are bound to produce more market noise and distort stock prices. There are also implications, perhaps, on the role of better marketing, communication, and transparency of RI funds to attract and retain “sticky” clients.

Finally, these findings address concerns for regulatory mechanisms to be put in place at a European level, which have heretofore been minimal and voluntary. Local government initiatives appear to have had more success in this regard. For instance, after the United Kingdom led the enforcement of the Pensions Act in 2000, requiring reporting from its pension funds, other member states such as Sweden (2000), France (2001), Germany (2002), Austria (2004), Belgium (2004), Norway (2004), Italy (2005), the Netherlands (2007), and Denmark (2008) followed suit. These legislations, however, have been mostly limited to government pension funds. Some countries have been able to implement RI through supportive legislation at a local level. For instance, the Belgian Parliament has prohibited the investment in companies producing anti-personnel mines, sub-munitions, and depleted uranium weapons since 2007. This was imitated recently in France in 2010 wherein the French Parliament enacted a law prohibiting any direct or indirect financial assistance to the production or trading of cluster munitions. In March 2011, the Italian Senate approved a motion similar to that of France. More popularly, Norway’s “Petroleum fund” has had ethical guidelines since 2004, which has prohibited investments in tobacco and arms production, among others. By structuring and regulating the process of RI through strong reforms and EU-level support, investors can take on a broader role in society by pushing corporations to move toward sustainability without necessarily harming their performance and yet potentially even decreasing their risk.

Going in this direction in terms of both investment strategy and regulation provides much-needed essence to a practice that has been criticized as of late as being too non-prescriptive and susceptible to greenwashing. Doing so can become the springboard for sustainability issues to have a truly substantial impact on pricing in the long term and can provide a new role for asset managers in constructing the future of finance, an industry which—still recovering from the backlash of the recent financial crisis—can be responsible and can drive profound societal change.

Footnotes

Acknowledgements

I would like to thank Diane-Laure Arjaliès, Michael del Mundo, Marco Giorgino, Jennifer Goodman, Tobias Gössling, Rocco Mosconi, and seminar participants at Medici 2011 (Florence), EBEN 2011 (Antwerp), and PRI-Mistra 2011 (Sigtuna), associate editor Magali Delmas, and the two anonymous reviewers for their invaluable comments on previous drafts of this article; Enguerran Petit and Rodrigo Soares Takasaki for excellent research assistance; Morningstar Italy and Vigeo (formerly Avanzi) Italy for providing access to data; Kenneth French for allowing open access to his database; and the asset managers who provided their time and insights to inform this research.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biography

Daniela Laurel-Fois (PhD, Politecnico di Milano) is assistant professor in the Department of Finance, Audit, and Control at IÉSEG School of Management (Catholic University of Lille) in Paris, France. Her research interests focus on the intersection between finance and sustainability issues.

References

Arjaliès

D.-L.

(2010). A social movement perspective on finance: How socially responsible investment mattered. Journal of Business Ethics, 92(Suppl. 1), 57-78.

Bansal

Clelland

(2004). Talking trash: Legitimacy, impression management, and unsystematic risk in the context of the natural environment. Academy of Management Journal, 47, 93-103.

Barnett

M. L.

Salomon

R. M.

(2006). Beyond dichotomy: The curvilinear relationship between social responsibility and financial performance. Strategic Management Journal, 27, 1101-1122.

Bauer

Koedijk

Otten

(2005). International evidence on ethical mutual fund performance and investment style. Journal of Banking & Finance, 29, 1751-1767.

Bouslah

Kryzanowski

M’Zali

(2013). The impact of the dimensions of social performance on firm risk. Journal of Banking & Finance, 37, 1258-1273.

Boutin-Dufresne

Savaria

(2004). Corporate social responsibility and financial risk. Journal of Investing, 13(1), 57-66.

Brundtland Commission. (1987). Our common future: Report of the World Commission on Environment and Development (UN Documents: Gathering a Body of Global Agreements). New York, NY: United Nations.

Busch

Bauer

Orlitzky

(2015). Sustainable development and financial markets: Old paths and new avenues. Business & Society, 55, 303-329.

Campbell

Slack

(2008). Corporate “philanthropy strategy” and “strategic philanthropy”: Some insights from voluntary disclosures in annual reports. Business & Society, 47, 187-212.

10.

Campbell

J. Y.

Lettau

Malkiel

B. G.

(2001). Have individual stocks become more volatile? An empirical exploration of idiosyncratic risk. Journal of Finance, 56, 1-43.

11.

Carhart

M. M.

(1997). On persistence in mutual fund performance. Journal of Finance, 52, 57-82.

12.

Casavecchia

Hulley

(2010, December). The effect of idiosyncratic risk on mutual fund flows and management fees. Paper presented at the annual meeting of Financial Management Association, Seoul, Korea.

13.

Crifo

Mottis

(2013). Socially responsible investment in France. Business & Society, 10, 1-18.

14.

Del Guercio

Tkac

(2002). The determinants of the flow of funds of managed portfolios: Mutual funds versus pension funds. Journal of Financial and Quantitative Analysis, 37, 523-557.

15.

Delmas

M. A.

Blass

V. D.

(2010). Measuring corporate environmental performance: The trade-offs of sustainability ratings. Business Strategy and the Environment, 19, 245-260.

16.

Delmas

M. A.

Etzion

Nairn-Birch

(2013). Triangulating environmental performance: What do corporate social responsibility ratings really capture? Academy of Management Perspectives, 27, 255-267.

17.

Derwall

Guenster

Bauer

Koedijk

(2005). The eco-efficiency premium puzzle. Financial Analysts Journal, 61, 51-63.

18.

DiMaggio

P. J.

Powell

W. W.

(1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48, 147-160.

19.

Donaldson

Preston

L. E.

(1995). The stakeholder theory of the corporation: Concepts, evidence, and implications. Academy of Management Review, 20, 65-91.

20.

Dowell

Hart

Yeung

(2000). Do corporate global environmental standards create or destroy market value? Management Science, 46, 1059-1074.

21.

Eurosif. (2010). European Social Investment Forum: European SRI Study. Retrieved from http://www.eurosif.org/publication/european-sri-study-2010/

22.

Eurosif. (2012). European SRI Study. Retrieved from http://www.eurosif.org/publication/european-sri-study-2012/

23.

Eurosif. (2014). European SRI Study 2014. Retrieved from http://www.eurosif.org/publication/european-sri-study-2014/

24.

Fama

E. F.

French

K. R.

(1993). Common risk factors in the returns on stocks and bonds. Journal of Financial Economics, 33, 3-56.

25.

Ferraro

Etzion

Gehman

(2015). Tackling grand challenges pragmatically: Robust action revisited. Organization Studies, 36, 363-390.

26.

Freeman

R. E.

(2010). Stakeholder theory. Cambridge, UK: Cambridge University Press.

27.

Gao

Schmidt

(2005). Self is never neutral: Why economic agents behave irrationally. Journal of Behavioral Finance, 6, 27-37.

28.

Glac

(2008). Understanding socially responsible investing: The effect of decision frames and trade-off options. Journal of Business Ethics, 87(Suppl. 1), 41-55.

29.

Godfrey

P. C.

Merrill

C. B.

Hansen

J. M.

(2009). The relationship between corporate social responsibility and shareholder value: An empirical test of the risk management hypothesis. Strategic Management Journal, 30, 425-445.

30.

Gond

J.-P.

Piani

(2012). Enabling institutional investors’ collective action: The role of the principles for responsible investment initiative. Business & Society, 52, 64-104.

31.

Hofmann

Hoelzl

Kirchler

(2008). A comparison of models describing the impact of moral decision making on investment decisions. Journal of Business Ethics, 82, 171-187.

32.

Hong

Kacperczyk

(2009). The price of sin: The effects of social norms on markets. Journal of Financial Economics, 93, 15-36.

33.

Humphrey

J. E.

Lee

D. D.

Shen

(2012). Does it cost to be sustainable? Journal of Corporate Finance, 18, 626-639.

34.

Ippolito

(1992). Consumer reaction to measures of poor quality: Evidence from the mutual fund industry. Journal of Law & Economics, 35, 45-70.

35.

(2012). Does CSR reduce firm risk? Evidence from controversial industry sectors. Journal of Business Ethics, 35, 1-16.

36.

Juravle

Lewis

(2008). Identifying impediments to SRI in Europe: A review of the practitioner and academic literature. Business Ethics: A European Review, 17, 285-310.

37.

Kahneman

Tversky

(1979). Prospect theory: An analysis of decision under risk. Econometrica: Journal of the Econometric Society, 47, 263-291.

38.

Kahneman

Tversky

(1984). Choices, values, and frames. American Psychologist, 39, 341-350.

39.

Keller

Siegrist

(2006). Investing in stocks: The influence of financial risk attitude and values-related money and stock market attitudes. Journal of Economic Psychology, 27, 285-303.

40.

Konar

Cohen

M. A.

(1997). Information as regulation: The effect of community right to know laws on toxic emissions. Journal of Environmental Economics and Management, 32, 109-124.

41.

Langbein

J. H.

Posner

R. A.

(1980). Social investing and the law of trusts. Michigan Law Review, 79, 72-112.

42.

Lee

D. D.

Faff

R. W.

(2009). Corporate sustainability performance and idiosyncratic risk: A global perspective. Financial Review, 44, 213-237.

43.

Lee

D. D.

Humphrey

J. E.

Benson

K. L.

Ahn

J. Y. K.

(2010). Socially responsible investment fund performance: The impact of screening intensity. Accounting & Finance, 50, 351-370.

44.

Lewis

Juravle

(2009). Morals, markets and sustainable investments: A qualitative study of “champions.” Journal of Business Ethics, 93, 483-494.

45.

Louche

Lydenberg

(2006). Socially responsible investment: Differences between Europe and United States (Working paper No. 2006-22). Gent, Belgium: Vlerick Leuven Gent Management School.

46.

Lounsbury

(2001). Institutional sources of practice variation: Staffing college and university recycling programs. Administrative Science Quarterly, 46, 29-56.

47.

Malkiel

B. G.

(2006, March). Idiosyncratic risk and security returns. Richardson: University of Texas at Dallas.

48.

Margolis

J. D.

Walsh

J. P.

(2003). Misery loves companies: Rethinking social initiatives by business. Administrative Science Quarterly, 48, 268-305.

49.

Markowitz

H. M.

(1952). Portfolio selection. Journal of Finance, 7, 77-91.

50.

Mishra

Modi

S. B.

(2013). Positive and negative corporate social responsibility, financial leverage, and idiosyncratic risk. Journal of Business Ethics, 117, 431-448.

51.

Nilsson

(2007). Investment with a conscience: Examining the impact of pro-social attitudes and perceived financial performance on socially responsible investment behavior. Journal of Business Ethics, 83, 307-325.

52.

Orlitzky

(2011). Institutional logics in the study of organizations. Business Ethics Quarterly, 21, 409-444.

53.

Pasewark

W. R.

Riley

M. E.

(2009). It’s a matter of principle: The role of personal values in investment decisions. Journal of Business Ethics, 93, 237-253.

54.

Philippe

Durand

(2011). The impact of norm-conforming behaviors on firm reputation. Strategic Management Journal, 32, 969-993.

55.

Porter

M. E.

Kramer

M. R.

(2011, January-February). Creating shared value. Harvard Business Review, 89(1/2), 62-77.

56.

Qian

(2006). Whom can you trust? A study on mutual fund governance. Retrieved from http://ssrn.com/abstract=685543

57.

Renneboog

Ter Horst

Zhang

(2008a). The price of ethics and stakeholder governance: The performance of socially responsible mutual funds. Journal of Corporate Finance, 14, 302-322.

58.

Renneboog

Ter Horst

Zhang

(2008b). Socially responsible investments: Institutional aspects, performance, and investor behavior. Journal of Banking & Finance, 32, 1723-1742.

59.

Rhee

Haunschild

P. R.

(2006). The liability of good reputation: A study of product recalls in the U.S. automobile industry. Organization Science, 17, 101-117.

60.

Sharfman

M. P.

Fernando

C. S.

(2008). Environmental risk management and the cost of capital. Strategic Management Journal, 29, 569-592.

61.

Sharpe

W. F.

(1964). Capital asset prices: A theory of market equilibrium under conditions of risk. Journal of Finance, 19, 425-442.

62.

Simon

H. A.

(1955). A behavioral model of rational choice. The Quarterly Journal of Economics, 69, 99-118.

63.

Sirri

E. R.

Tufano

(1998). Costly search and mutual fund flows. Journal of Finance, 53, 1589-1622.

64.

Slager

(2014). SRI indices and responsible corporate behavior: A study of the FTSE4Good Index. Business & Society, 54, 386-405.

65.

Statman

(1987). How many stocks make a diversified portfolio? Journal of Financial and Quantitative Analysis, 22, 353-363.

66.

Statman

Caldwell

(1987). Applying behavioral finance to capital budgeting: Project terminations. Financial Management, 16, 7-15.

67.

US Social Investment Forum (USSIF). (2010). Report on socially responsible investing trends in the United States. Retrieved from http://www.ussif.org/files/Publications/10_Trends_Exec_Summary.pdf

68.

Waddock

S. A.

Graves

S. B.

(1997). The corporate social performance-financial performance link. Strategic Management Journal, 18, 303-319.

69.

Wermers

(2002). Mutual fund performance: An empirical decomposition into stock-picking talent, style, transactions costs, and expenses. Journal of Finance, 55, 1655-1703.