Prior research indicates that religion influences corporate decision making (e.g., El Ghoul, Guedhami, Ni, Pittman, & Saadi, Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012; Hilary & Hui, Reference Hilary and Hui2009). The linkage between corporate decision making and religion that earlier studies have found suggests that company decisions that aim at benefitting a community may likewise be influenced by religion. In particular, we believe that we may gain a deeper understanding of the potential influences underlying corporate decision making if we look at the relation, if any, between religiosity and company decisions that affect local communities. Moreover, because studies of corporate social responsibility (CSR) conceptualize the community as one of the important stakeholders to which companies should be responsive (Carroll, Reference Carroll1999; Garriga & Melé, Reference Garriga and Melé2004; Crane & Matten, Reference Crane and Matten2004), determining whether religion influences managerial decisions about the community will also have important implications for CSR studies.
Many firms now make significant investments in community initiatives that range from providing funding to charities or volunteering at schools to entering into public-private arrangements to provide housing for the economically disadvantaged (Smith, Reference Smith, Baker and Nofsinger2012). Apple and Google, for example, match any monetary gifts their employees make to charities. Walmart responded to Hurricane Katrina’s devastation in 2005 by partnering with the city to provide storm-ravaged New Orleans residents with over $20 million in cash donations, 100 truckloads of free merchandise, and more than 100,000 free meals (Barbaro & Gillis, Reference Barbaro and Gillis2005). IBM’s Corporate Service Corps program sends teams of IBM employees with a range of skills into communities in developing countries to work with them to help solve the economic, technology, education, and healthcare-related problems they face (Hamm, Reference Hamm2009; Marquis & Kanter, 2009).
In this article, we will use the label “corporate community involvement” (CCI) to refer to these kinds of corporate community initiatives, (i.e., corporate initiatives that aim to provide benefits to, or alleviate the perceived needs of or harmful effects upon, a community). As the previous examples suggest, CCI initiatives can involve significant amounts of collaboration with external communities. CCI initiatives, in fact, are sometimes categorized according to the degree to which the company involves the targeted community in its decision making (Austin, Reference Austin2000; Bowen, Newenham-Kahindi, & Herremans, Reference Bowen, Newenham-Kahindi and Herremans2010; Hardy & Philips, 1998; Morsing & Schultz, Reference Morsing and Schultz2006; Pater & Van Lierop, Reference Pater and Van Lierop2006).
CCI is generally considered part of the larger domain of corporate social responsibility (CSR). In a series of articles, Carroll (Reference Carroll1979, Reference Carroll1991, Reference Carroll1998, Reference Carroll1999) developed a now widely accepted definition of CSR that allows us to position CCI in relation to CSR’s other components. Carroll argued that the construct of corporate social responsibility includes four kinds of responsibilities: economic, legal, ethical, and discretionary. In terms of Carroll’s four-part CSR framework, CCI initiatives are part of the discretionary responsibilities of the firm (i.e., responsibilities that society expects from the firm as a “good citizen” but that remain discretionary insofar as they are not part of the firm’s required profit-making economic activities, nor does the firm violate any laws or ethical requirements if it does not engage in them). Moreover, they typically aim to benefit an external community instead of internal stakeholders who are involved in the economic operations of the firm. Among the main kinds of benefits a company’s community initiatives may provide are cash (e.g., matching employee contributions to charities as Apple and Google do), emergency assistance (e.g., the assistance Walmart provided to the victims of Katrina), and noncash or pro bono services or expertise (e.g., the services provided by IBM’s Corporate Service Corps). We note that what we are here calling “corporate community involvement” may be referred to in the literature on CSR as corporate philanthropy, corporate charitable contributions, community engagement, or corporate citizenship (Rochlin & Christoffer, Reference Rochlin and Christoffer2000).
This characterization of CCI raises an important question: If a company’s investments in CCI are discretionary—i.e., not economically, legally, or ethically required—why do companies make such investments? One answer is often referred to as the “business case” for community initiatives (Porter & Kramer, Reference Porter and Kramer2002; Rochlin & Christoffer, Reference Rochlin and Christoffer2000). Briefly stated, the business case for community initiatives (and, more generally, for CSR) argues that companies invest in community initiatives because—perhaps over the long term—they produce economic benefits for the firm (Carroll & Shabana, Reference Carroll and Shabana2010). This explanation of CCI, of course, in effect removes CCI from Carroll’s “discretionary” category of CSR and makes it part of what Carroll calls the firm’s “economic” responsibilities (i.e., part of its profit-making responsibilities).
We acknowledge that economic considerations may be a factor driving investments in CCI. However, economic considerations are not the only factors that influence managerial decision making. In particular, a significant body of research (Dyreng, Mayew, & Williams, Reference Dyreng, Mayew and Williams2012; El Ghoul et al., Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012; Hilary & Hui, Reference Hilary and Hui2009; McGuire, Newton, Omer, & Sharp, Reference McGuire, Newton, Omer and Sharp2012; McGuire, Omer, & Sharp Reference McGuire, Omer and Sharp2011; Omer, Sharp, & Wang, 2016) has shown that managers’ decisions are also influenced by their religious commitments. In this article, therefore, we propose to examine whether the influences driving corporate investments in CCI programs are at least partly related to a noneconomic factor rarely studied in this context, namely the influence that religion can have on a company’s top management. Examining the relationship between religion and the decisions top managers make regarding CCI programs may not only provide us with valuable insights into the forces that drive managerial actions, but should also shed light on the role religion plays in a company’s CSR-related decisions.
The purpose of this article, therefore, is to look at whether a company’s community involvement initiatives are empirically associated with the religiosity of the community in which it is located.Footnote 1 Our study focuses on the US where the dominant religions are Christian; we therefore examine the relationship between Christian religiosity and CCI initiatives. We study this relationship by proposing what we call the “religious morality hypothesis,” which is based partially on a review of the teachings of the major US (Christian) religious denominations regarding the responsibilities that managers owe local communities. This review shows that the major US Christian denominations insist that businesses have an obligation to assist their local communities and therefore to invest in CCI initiatives. We then draw on two bodies of research to argue that these religious teachings should lead managers and owners to engage in CCI initiatives.
One stream of research supporting the claim that the teachings of religion will lead local managers to engage in CCI initiatives projects a “moral community” hypothesis. These studies argue that the moral teachings of a community’s religions will affect the behaviors of those believers and nonbelievers who reside in that community (Clark-Miller, Reference Clark-Miller2008; Corcoran, Pettinicchio, & Robbins, Reference Corcoran, Pettinicchio and Robbins2012; Omer et al., 2016; Pescosolido, Reference Pescosolido1990; Pescosolido & Georgianna, Reference Pescosolido and Georgianna1989; Regnerus, Reference Regnerus2003; Stark, Doyle, & Kent, Reference Stark, Doyle and Kent1980; Stark, Kent, & Doyle, Reference Stark, Kent and Doyle1982). Because US Christian denominations teach that managers and owners should assist their local communities, the moral community hypothesis leads us to expect that managers and owners of organizations will conform to those moral teachings by engaging in CCI initiatives.
A second body of research on which we draw is institutional theory. Institutional theorists have argued that organizations are subject to coercive, mimetic, and normative pressures exerted by the institutions that surround them (DiMaggio & Powell, Reference DiMaggio and Powell1983). In particular, organizations are responsive to the “formal and informal pressures exerted on organizations … by cultural expectations in the society within which organizations function” (DiMaggio & Powell, Reference DiMaggio and Powell1983: 150). We expect, therefore, that religious institutions will exert pressures on business organizations to assist their local communities and business organizations will align themselves with those normative expectations by engaging in CCI initiatives.
The research on both the moral community hypothesis and institutional theory, therefore, leads us to hypothesize that the higher the level of (Christian) religiosity of the region in which a firm’s headquarters is located (i.e., the region in which the firm’s top management is likely to reside), the more likely it is that the firm will invest in the community-friendly practices those religious institutions promote. Institutional theory, however, suggests that institutional differences lead to different organizational outcomes. We therefore also hypothesize that institutional differences among the religious denominations will lead to different kinds of CCI responses.
The results of our study, which is based on a large sample of companies operating in the US during the period 1991–2013, provide support for these hypotheses. First, we show that there is a robust empirical relationship between companies’ investments in community-friendly practices and the level of religiosity of their local communities. This association remains after we control for a variety of firm characteristics. In addition, our results also suggest that institutional differences that exist among the Christian denominations lead to surprisingly different organizational decisions related to CCI.
The remainder of the article is organized as follows. The next section briefly describes the literature on CCI and highlights what is known about corporate investments in CCI. Next, we discuss the literature on which our hypotheses are based. We then turn to discussions of our sample, our measurement of CCI practices and of religiosity, and our research design. The subsequent section presents our empirical results. The final sections discuss the limitations, significance and contributions, and recommendations for future studies.
LITERATURE REVIEW AND HYPOTHESES
Review of Literature on Corporate Community Involvement
As mentioned earlier, CCI is sometimes justified on the basis of the business case in support of such initiatives, (i.e., the claim that investments in CCI are associated with firms’ financial performance). Numerous studies have examined this claim but the results are mixed. A study by Patten (Reference Patten2008) found that CCI contributions to relief efforts had a positive effect on the firm’s value as measured by its stock price five days after the contribution was announced and firms making larger gifts experienced higher positive abnormal market returns. Khanna and Damon (Reference Khanna and Damon1999) showed that CCI contributions led to higher returns on capital and Galaskiewicz (Reference Galaskiewicz1997) found a strong positive association between CCI and return on sales, return on assets, and return on equity relative to industry averages.
Godfrey, Merrill, and Hansen (Reference Godfrey, Merrill and Hansen2009) found that firms’ CCI (in the form of philanthropy) provided shareholders with “insurance-like” protection against losses by reducing the decline in shareholder value that can occur when firms are found to be engaging in negative activities (e.g., environmental or product safety issues). Godfrey (Reference Godfrey2005) suggested earlier that CCI could reduce such shareholder losses by signaling that the firm considers others in its decisions, a signal that in turn mitigates the punitive sanctions that stakeholders tend to inflict on companies found to be involved in such negative events. On the other hand, a number of studies (Berman, Wicks, Kotha, & Jones, Reference Berman, Wicks, Kotha and Jones1999; Boatsman & Gupta, Reference Boatsman and Gupta1996; Seifert, Morris, & Bartkus, Reference Seifert, Morris and Bartkus2003) found little or no association between a company’s community contributions and its financial performance. Moreover, Wang, Choi, and Li (Reference Wang, Choi and Li2008) found that a company’s community contributions were positively associated with its financial performance up to a point, beyond which the positive association leveled off and eventually disappeared, giving the association an inverse curvilinear U-shape. A meta-analytic review of CCI studies by Orlitzky, Schmidt, and Rynes (Reference Orlitzky, Schmidt and Rynes2003) aggregated previous studies that examined the empirical relation between CCI and financial performance. Orlitzky and his colleagues concluded that the evidence provided by their aggregation of CCI studies showed that CCI is positively related to financial performance and the relationship is stronger than that between any other component of CSR and financial performance.
Several CCI studies found that investments in CCI are associated with a number of factors other than financial performance. Burke, Logsdon, Mitchell, Reiner, and Vogel (Reference Burke, Logsdon, Mitchell, Reiner and Vogel1986) found that CCI contributions are influenced by the age of the company. In addition, Useem (Reference Useem1988) showed they are influenced by the industry in which the company competes: the greater the visibility of the industry, the more likely the firms in that industry will engage in CCI initiatives. Studies by Bartkus, Morris, and Seifert (Reference Bartkus, Morris and Seifert2002), Buchholtz, Amason, and Rutherford (Reference Buchholtz, Amason and Rutherford1999), and Lerner and Fryxell (Reference Lerner and Fryxell1994) found that CCI is influenced by the CEO’s values. A number of other studies have shown that CCI is positively associated with firm size (Atkinson & Galaskiewicz, Reference Atkinson and Galaskiewicz1988; Buchholtz et al., Reference Buchholtz, Amason and Rutherford1999; McElroy & Siegfried, Reference McElroy and Siegfried1985; Seifert et al., Reference Seifert, Morris and Bartkus2003; Useem, Reference Useem1988; Wood & Jones, Reference Wood and Jones1995). Seifert, Morris, and Bartkus (2004: 145) suggest that the reason why large firms engage in more CCI activities is because “large firms have greater visibility which would attract greater public scrutiny and a higher standard of corporate citizenship.” Campbell and Slack (Reference Campbell and Slack2006) examined the suggestion that firm visibility is associated with CCI initiatives and found that high-visibility companies do engage in higher rates of CCI activities than low-visibility companies.
In spite of the large number of studies conducted on CCI and the range of associations those studies have addressed, there have been no studies to this point of the relationship between CCI and religion. Yet, as mentioned earlier, a large body of studies indicate that religion has a significant influence on corporate decision making (e.g., Cui, Jo, & Velasquez, Reference Cui, Jo, Na and Velasquez2015a, 2015b, 2016; Dyreng et al., Reference Dyreng, Mayew and Williams2012; El Ghoul et al., Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012; Hilary & Hui, Reference Hilary and Hui2009; Kumar, Page, & Spalt, Reference Kumar, Page and Spalt2011; McGuire et al., Reference McGuire, Omer and Sharp2011, Reference McGuire, Newton, Omer and Sharp2012; Omer et al., 2016). We attempt in this article to remedy this deficit by studying the empirical relationship, if any, between Christian religious views on community relations and corporate decisions to invest in CCI initiatives.
US Christian Religiosity and Community Relations
The religious traditions of the United States are extremely diverse. Nevertheless, the majority of the population (76 percent) identifies itself as Christian and only 3.9 percent of Americans identify as members of a non-Christian religion (Kosmin & Keysar, Reference Kosmin and Keysar2009; Newport, Reference Newport2011). Consequently, we will focus on how the teachings of the three main Christian denominational groups—Catholics, mainline Protestants, and evangelical Protestants—influence CCI initiatives. As we show below, the teachings of these three Christian denominations evidence a surprisingly high level of similarity given the heterogeneity they exhibit in other doctrinal matters (ILO, 2012). Zinbarg (Reference Zinbarg2001) and Melé (Reference Melé2003, Reference Melé2009, Reference Melé2011, Reference Melé2012) point out, for example, that all Christian denominations teach that each person possesses “human dignity,” a dignity that all men and women must respect. This view of human dignity, they argue, implies that the moral obligations of the manager go beyond the requirements of the law and must also include the obligation to engage in community development and avoid any negative economic impacts on the community.
Roman Catholicism
As Melé (Reference Melé2011) explains, Catholic moral views on business are based on a group of writings that are referred to as Catholic Social Teaching (CST), the greater part of which consist of a number of social encyclicals. These social encyclicals are pastoral letters that articulate the Church’s moral teachings on political, social, and economic issues and that Catholic popes began publishing during the late nineteenth century (Finn, Reference Finn2012; Thompson, Reference Thompson2015).
The moral views expressed in the Catholic social encyclicals are based on two fundamental claims (Thompson, Reference Thompson2015). The first is that each human being has an “inviolable dignity” based on the idea that each is created “in the image of God” (see, for example, Benedict XVI, 2009; Francis, 2015). The second foundational claim is that the dignity of each person implies that everyone—including, therefore, all business people—must work toward maintaining a community in which that dignity can flourish and develop (Melé, Reference Melé2012). As Pope Benedict XVI wrote in 2009, “business management cannot concern itself only with the interests of the proprietors [shareholders], but must also assume responsibility for all the other stakeholders who contribute to the life of the business: the workers, the customers, the suppliers, … [and] the community.”
Mainline Protestantism
The mainline Protestant churches do not have an institution like the Catholic papacy that has the power to publish authoritative statements on moral or social issues. Nevertheless, various groups of mainline Protestant churches have periodically come together to issue statements on current moral issues and they have recommended that their members adopt these views. A representative picture of Protestant moral beliefs can be put together from such statements, a picture that is very similar to that portrayed in CST. In the influential 2004 paper, Business as Mission, for example, the Lausanne Committee for World Evangelization, an international group made up of representatives from several mainline Protestant denominations, put forth some teachings that echoed the fundamental views of CST. The Lausanne Committee for World Evangelization (2004) wrote that “created in God’s image, humanity is also capable of creating” and consequently has “the responsibility to respect and care for each other.” The Lausanne statement urges businesses to contribute to “their community or social investment activities” and “to have a positive and lasting impact in the local community.”
Sufficient, Sustainable Livelihood for All, is another influential mainline Protestant document that addressed community issues and that was issued by a group of mainline Protestant churches. This document which was issued on August 20, 1999, by the Churchwide Assembly of the Evangelical Lutheran Church in America also claims that every person possesses a “human dignity” because each has been “created in God’s image.” It calls for “corporate governance that is accountable for the [negative] effects of a company’s practices on … communities” and praises businesses that invest in “sustainable community economic development that takes into account the overall health and welfare of people.” The largest mainline Protestant group is the United Methodist Church (UMC), which, in its 1972 General Conference, issued a resolution titled the “Economic Community.” Prefaced by repeated avowals that “human beings, created in the image of God, have an innate dignity,” the resolution asserted that “corporations (as the economic community) are responsible not only to their stockholders, but also to other stakeholders: their workers, suppliers, vendors, customers, the communities in which they do business, and for the earth, which supports them” (United Methodist Church, 2016: 136)
It is, perhaps, not surprising that the pronouncements of mainline Protestants regarding community assistance are similar to those of Roman Catholics. Like the Catholic Church, mainline Protestantism in the United States has a long history of commitment to social causes. Even before Pope Leo XIII (1891) wrote the first Catholic social encyclical on The Condition of Labor, the American Protestant clergyman Washington Gladden (Reference Gladden1877) had published his influential book The Christian Way: Whither It Leads and How to Go On, which urged his fellow Protestants to apply Christian values in their business lives. Twenty years later, Walter Rauschenbusch (Reference Rauschenbusch1922), a Baptist pastor, published the even more influential A Theology for the Social Gospel, criticizing the worst sins of capitalism and calling for a form of Christian socialism that supported labor unions, assistance for the poor, and social reforms (Sweeney, Reference Sweeney2005: 162–163).
Evangelical Protestantism
Yet the “Social Gospel movement” that Gladden and Rauschenbusch initiated and mainline Protestant churches endorsed was opposed—in a movement that came to be called “The Great Reversal”—by a sizable contingent of evangelical Protestants who argued that efforts devoted to social causes were a distraction from the “fundamentalist” Christian task of saving one’s soul and preaching the gospel, (i.e., of evangelizing [Kee, Albu, Lindberg, Frost, & Robert, Reference Kee, Albu, Lindberg, Frost and Robert1998; Moberg, Reference Moberg2007; Sutton, Reference Sutton2014]). Related to this was the widespread acceptance among evangelicals of “dispensational premillennialism,” the view that the second coming of Jesus is imminent and therefore it is urgent for Christians to devote all their efforts to ensuring their personal salvation and evangelizing the world before time runs out (Sutton, Reference Sutton2014).
Consequently, throughout most of the twentieth century, evangelical Protestants issued few public statements on business or economic issues. However, partly as a response to the 1973 Roe vs. Wade abortion decision of the Supreme Court, partly as a response to the “moral majority” movement of the 1980s, and partly as a result of the urgings of a young post-World War II generation, evangelicals began in the 1970s and 1980s to move away from a sole focus on personal faith and evangelization and toward advocating policies and practices aimed at improving the material lives of people (Roels, Reference Roels1997; Sutton, Reference Sutton2014). Indicative of this changing stance, in 1978, the Southern Baptist Convention—the largest evangelical denomination—issued a resolution that was entitled Declaration of Human Rights. The statement read: “We believe that they [human rights] spring from the Bible’s revelation that all persons are made in ‘the likeness of God.’” Consequently, the resolution declared, “Christians are obligated not only to provide in Jesus’ name cups of cold water for individuals whose human rights have been violated, but also to deal in a forthright and corrective way with social structures which abuse and violate human rights” (Southern Baptist Convention, 1978). In 2012, the Southern Baptist Convention adopted the resolution Affirming Human Needs Ministry and Community Involvement By Local Churches, which urged all church members to provide for “those in need” in the local community as “an expression of love, discipleship, and fidelity to our Lord as God gifts, directs, and enables us to be His hands and feet to a hurting and needy world” (Southern Baptist Convention, 2012)
The three Christian denominations, which comprise most of the American religious population, therefore all affirm the “dignity” of each person based on the conviction that every person is made “in the image of God.” They therefore all advocate that people, particularly Christian church members, respond to the needs of communities, calling on businesses to support all such responses and treat all their communities with fairness and compassion.
The Influence of Religion on Behavior
Following the work of Max Weber ([1904] 2009), scholars have examined how religion influences behavior. Numerous studies since then have looked at how the religious commitments of individuals affect their behaviors and attitudes. At this individual level of analysis, it has been shown that a person’s religion strongly influences that person’s political attitudes and his or her behaviors such as, for example, how the person votes (Fastnow, Grant, & Rudolph, Reference Fastnow, Grant and Rudolph1999; Jelen, Reference Jelen1998; Leege & Kellstedt, Reference Leege and Kellstedt1993). The religion of individuals also influences the extent to which they engage in or decide to abstain from premarital intercourse as well as the attitudes they hold about premarital childbirth (Jeynes, Reference Jeynes2003). Studies have shown that the likelihood a person will engage in tax fraud is inversely related to the level of that person’s religiosity (Grasmick, Kinsey, & Cochran, Reference Grasmick, Kinsey and Cochran1991; Petee, Milner, & Welch, Reference Petee, Milner and Welch1994; Stack & Kposowa, Reference Stack and Kposowa2006). More generally, a person’s religiosity is inversely associated with deviant behaviors (Agnew, Reference Agnew1998; Baier & Wright, Reference Baier and Wright2001). Juvenile delinquency has also been shown to be inversely related to the religiosity of the individual juvenile (Cochran & Akers, Reference Cochran and Akers1989). While a few studies have contested these associations (e.g., Hirschi & Stark, Reference Hirschi and Stark1969), more recent studies have confirmed them (Albrecht, Chadwick, & Alcorn, Reference Albrecht, Chadwick and Alcorn1977; Chadwick, Top, & McClendon, Reference Chadwick, Top and McClendon2010; Donahue & Benson, Reference Donahue and Benson1995; Sloane & Potvin, Reference Sloane and Potvin1986). In addition, several studies (Bloodgood, Turnley, & Mudrack, Reference Bloodgood, Turnley and Mudrack2007; Conroy & Emerson, Reference Conroy and Emerson2004; Kennedy & Lawton, Reference Kennedy and Lawton1998; Smith & Oakley, Reference Smith and Oakley1996; Terpstra, Rozell, & Robinson, Reference Terpstra, Rozell and Robinson1993) found that the religiosity of college students affects their personal attitudes toward ethics.
Previous studies suggest that an individual’s religion also influences his or her decisions in business (Agle & Van Buren, Reference Agle and Van Buren1999; Dyreng et al., Reference Dyreng, Mayew and Williams2012; El Ghoul et al., Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012; Emerson & Mckinney, Reference Emerson and Mckinney2010; Hilary & Hui, Reference Hilary and Hui2009; McGuire et al., Reference McGuire, Omer and Sharp2011, Reference McGuire, Newton, Omer and Sharp2012; Nash, Reference Nash1994; Omer et al., 2016). Emerson and Mckinney (2010: 8), for example, found that “business professionals who considered their religious faith to be highly important to them are significantly less accepting of ethically questionable behavior.” Agle and Van Buren (1999: 580) found a positive but “weak and inconsistent” link between the religious beliefs and practices of MBA students and their adherence to “a broad view of corporate social responsibility.” Chatjuthamard-Kitsabunnarat, Jiraporn, and Tong (2014: 1128) found that “religion motivates managers to treat other stakeholders and the society at large more favorably.” Thus, in business contexts as well, the religious commitments of individuals influence their personal behaviors and attitudes.
The Moral Communities Hypothesis
While the previously discussed studies establish that religion influences behavior, the level of analysis of these studies is at the individual level: they all look at how the individual’s personal religiosity affects that individual’s personal behaviors. More germane to our present study is a related body of research that looks at the influence of religion at the group level. In particular, studies on what has been called the “moral communities” hypothesis have looked at the extent to which the religiosity of a group affects the activities of the members of that group.
Although the term “moral community” is a twentieth century construct, research on how the religiosity of a group influences the behaviors of its members can be traced to the work of Emile Durkheim. In Suicide Durkheim (Reference Durkheim1897) pointed out that throughout Western Europe suicide rates were higher in regions where Protestants were more numerous, and lower in regions where Catholics were more numerous. The underlying reason for this distribution, he reasoned, was that although both Catholics and Protestants condemned suicide as immoral, the Catholic Church exerted stronger social control over its members than the Protestant churches and so achieved higher levels of conformity to its moral values. Contemporary research on the moral community hypothesis argues in a similar way that communities that have higher levels of religiosity will exert greater pressures on their members—both believers and nonbelievers—to conform to their moral values (Cochran & Akers, Reference Cochran and Akers1989; Sloane & Potvin, Reference Sloane and Potvin1986; Welch, Tittle, & Petee, Reference Welch, Tittle and Petee1991).
The moral community hypothesis has been supported by a large body of research (Clark-Miller, Reference Clark-Miller2008; Corcoran et al., Reference Corcoran, Pettinicchio and Robbins2012; Omer et al., 2016; Pescosolido, Reference Pescosolido1990; Pescosolido & Georgianna, Reference Pescosolido and Georgianna1989; Regnerus, Reference Regnerus2003; Stark et al., Reference Stark, Doyle and Kent1980, Reference Stark, Kent and Doyle1982). For example, Welch et al. (Reference Welch, Tittle and Petee1991) found that the higher the level of Catholic religiosity within a parish, the lower the likelihood that the members of that parish would engage in deviant behaviors. A study by Regnerus (Reference Regnerus2003) showed that the higher the level of religiosity within a county, the lower the likelihood that adolescents in that county engaged in theft and other forms of delinquency. Omer et al. (2016) showed that higher levels of religiosity within given metropolitan areas, are associated with a higher likelihood that accountants in those areas will issue more honest and conservative audit opinions. And two studies (Clark-Miller, Reference Clark-Miller2008; Corcoran et al., Reference Corcoran, Pettinicchio and Robbins2012), have found that the higher the level of religiosity of a nation as a whole, the lower the likelihood that citizens of that nation would see white-collar crime as acceptable.
Research on the moral community hypothesis has shown, then, that the religiosity of a geographic community—whether that community is defined as a county, a metropolitan area, a parish, or even an entire nation—influences the moral beliefs and behaviors of the members of that community (Greeley, McCready, Sullivan, & Fee, Reference Greeley, McCready, Sullivan and Fee1981; Leege & Welch, Reference Leege and Welch1989; Wald, Owen, & Hill, Reference Wald, Owen and Hill1988). The religion of a region seems to be such “a potent generator of conformity” (Welch et al., Reference Welch, Tittle and Petee1991: 159; see also Cochran & Akers, Reference Cochran and Akers1989; Sloane & Potvin, Reference Sloane and Potvin1986; Stark, Reference Stark1996), that it affects what people residing within that region believe and do (Ellison, Burr, & McCall, Reference Ellison, Burr and McCall1997; Greeley et al., Reference Greeley, McCready, Sullivan and Fee1981; Leege & Welch, Reference Leege and Welch1989; Wald et al., Reference Wald, Owen and Hill1988). The research on moral communities suggests, then, that the Christian religiosity of the regions within which companies are located will lead the management of those companies to engage in the kind of CCI initiatives that those Christian religions promote.
Institutional Theory
Institutional theory, like the moral community hypothesis, links group characteristics to the behaviors of agents embedded within the group. Institutional theory holds that institutions—relatively enduring rules, conventions, and practices of groups that influence the actions of the group’s members—are essential to explanations of organizational actions and structures. Although institutions have long been the object of study, contemporary interest in institutions as determinants of organizational actions and structures traces its origins to the seminal work of DiMaggio and Powell (1983), Friedland and Alford (Reference Friedland, Alford, Powell and Dimaggio1991), and Meyer and Rowan (Reference Meyer and Rowan1977). Meyer and Rowan (Reference Meyer and Rowan1977) argued that organizations exist in an institutionalized environment of professions, programs, and technologies that embody rules that serve as powerful myths. Organizations align themselves with these institutional rules and, by doing so, come to be seen as legitimate and worthy, strengthen their support and stability, increase their resources and survival capabilities, and secure their success. A few years later, DiMaggio and Powell (1983) argued that when organizations align themselves with their institutional environment, they tend to become similar or “isomorphic” with each other. DiMaggio and Powell identified three kinds of institutional pressures that lead to organizational isomorphism: coercive pressures exercised by other organizations and society’s culture, mimetic pressures when organizations cope with uncertainty by imitating other successful organizations, and normative pressures when organizations hire professionals with similar training, socialization, and professional norms.
Conversely, Friedland and Alford (1991) explored how institutional differences can impose conflicting pressures on organizations. Friedland and Alford argued that the main institutions of contemporary Western societies—capitalism, family, bureaucratic state, democracy, and Christian religion—each have a distinct “institutional logic.” Thornton and Ocasio (1999: 804) later provided a detailed definition of institutional logic as the “socially constructed, historical patterns of material practices, assumptions, values, beliefs, and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality.” Friedland and Alford argued that different institutions with their distinct institutional logic could affect organizations in different, even contradictory, ways by placing conflicting demands on them.
The work of DiMaggio and Powell (Reference DiMaggio and Powell1983), Friedland and Alford (Reference Friedland, Alford, Powell and Dimaggio1991), and Meyer and Rowan (Reference Meyer and Rowan1977) stimulated many studies based on the view that the behaviors of individuals and organizations are influenced by the surrounding context of institutions and their institutional logic, which can, depending on the institutional context, produce isomorphic as well as heterogeneous responses (for reviews of this work, see DiMaggio & Powell, Reference DiMaggio and Powell1991; Lawrence & Suddaby, Reference Lawrence, Suddaby, Clegg, Hardy, Lawrence and Nord2006; Schneiberg & Clemens, Reference Schneiberg and Clemens2006). Some of these studies found that institutions and their logic influence the CSR decisions of managers.
Matten and Moon (Reference Matten and Moon2008), for example, showed how institutional differences—specifically different political, financial, educational, labor, and cultural “systems”—between the United States and Europe can account for the differences in the level and kind of CSR decisions companies in these two regions make, whereas certain coercive, mimetic, and normative pressures can account for their similarities or isomorphism. Jamali and Neville (Reference Jamali and Neville2011) showed that a similar set of institutional factors (local political, financial, educational, labor, and cultural systems) shape the CSR decisions of companies operating in Lebanon. Holder-Webb and Cohen (Reference Holder-Webb and Cohen2012) argue that mimetic pressures and the coercive pressures of the US regulatory regime account for the isomorphism of the codes of ethics of firms operating in the United States. Avetisyan and Ferrary (Reference Avetisyan and Ferrary2013) show that institutional differences and similarities between France and the United States (in their respective national and international regulatory regimes, CSR rating groups, and other stakeholder groups) account for the differences and similarities in how each country’s firms engage in CSR. Campbell (Reference Campbell2007) argued that characteristics of the economic institutions (the “health” of a company’s surrounding economy, level of competition it faces, regulatory regime, degree of self-regulation to which a company is subject, presence of NGOs, etc.) surrounding companies influence their CSR responses.
The lesson of institutional theory, then, is that the institutions surrounding a company influence the corporate decisions it makes. We are therefore led to expect that what Friedland and Alford (Reference Friedland, Alford, Powell and Dimaggio1991) and Thornton (Reference Thornton2004) identify as one of the central institutions of Western societies—religion—will also influence corporate decisions. Institutional theory leads us to expect, that is, that the pressures exerted by a region’s local religious (Christian) institutions will influence the corporate CCI decisions of companies surrounded by those institutions.
In summary, if it is true that religious institutions influence corporate decision making because they inculcate their moral values throughout a “moral community,” as well as through the institutional pressures they generate, local religious views about the moral obligations of managers should influence the decisions local management makes about the community. In particular, because the dominant religions in the United States uniformly embrace the view that people in general and managers in particular have a moral obligation to deal fairly with local communities and promote their well being, and because, as we have argued, local levels of religiosity influence the decisions of local companies, we can hypothesize that companies whose top management is headquartered in geographical regions that have higher levels of Christian religiosity, will implement more CCI initiatives. We will refer to this as our “religious morality hypothesis”:
Hypothesis 1: Companies headquartered in regions with higher levels of Christian religiosity will engage in more CCI initiatives.
Institutional theory also leads us to expect, however, that differences among the various religious institutions influencing organizations—differences in their “institutional logic” including their “historical patterns of material practices, assumptions, values, beliefs, and rules” (Thornton & Ocasio, Reference Thornton and Ocasio1999: 804)—will lead to different organizational responses (Friedland & Alford, Reference Friedland, Alford, Powell and Dimaggio1991). We have seen that there are, in fact, significant historical differences between the beliefs and values of the Catholic and mainline Protestant denominations and those of evangelical religious denominations, particularly as related to their beliefs about the imminent return of Jesus and the significant value placed on evangelization and personal salvation. We hypothesize, therefore, that these differences in the institutional logic of US religious denominations will generate differences in the decisions corporate organizations make about their CCI initiatives. Specifically, this leads us to what we will call our “differentiated responses hypothesis”:
Hypothesis 2: The CCI decisions of firms will show differences related to the different institutional characteristics (history, beliefs, and values) of the religious denominations around them.
We can, perhaps, be more specific about the kind of differentiated responses we can expect to arise from the particular institutional differences that the three main Christian denominational groups—Catholic, mainline Protestant, and evangelical Protestant—exhibit.
There are identifiable historical differences in the way that the major denominations have shown their support for the view that businesses should assist and not harm their communities, differences that are particularly evident between the Catholic and mainline Protestant denominations and the evangelical denominations. Both the Catholic and mainline Protestant denominations have a long history of calling on their members to help relieve the needs of their communities. However, similar calls were not made by evangelical churches in the United States until fairly recently. Moreover, even today, many evangelicals continue to see such social involvement and concern as a less valuable peripheral aspect of their faith and continue to hold that they must remain focused on evangelization and personal salvation in light of the imminent return of Jesus (Moberg, Reference Moberg2007; Sutton, Reference Sutton2014). In fact, according to a 2010 Pew Research Center report, 58 percent of white evangelicals believe that Jesus will “probably” or “definitely” return by 2050 (Kohut et al., Reference Kohut, Keeter, Doherty, Dimock, Remez, Suls, Neidorf, Christian, Kiley, Tyson and Pushter2010) and therefore it is urgent to prepare for that return.
In sum, whereas both Catholics and mainline Protestants have a relatively long history of urging business people to assist their local communities and avoid activities that might harm them, evangelical Protestants, until recently, have not made similar demands on business people. Moreover, there remains some uncertainty among evangelicals today about whether the work of evangelization and one’s personal salvation should continue to take priority over all other matters, particularly in light of the imminent return of Jesus. We can hypothesize, therefore, that these institutional differences in “historical patterns … values [and] beliefs” (Thornton & Ocasio, Reference Thornton and Ocasio1999: 804), evident in the way that the three main Christian denominations have urged businesses to care for their communities, will result in differences in the influence these denominations exert on business people. Specifically, based on how differences, in their history, values, and beliefs have led mainline Protestants and Catholics to place greater emphasis on the obligation of businesses to provide material help to communities than evangelical Protestants have done, we can expect that Catholic and mainline Protestant religiosities are more likely to exert a positive influence on the CCI decisions of local companies than evangelical Protestant religiosity. This is the basis of what we call our “differentiated Christian denominational responses hypothesis”:
Hypothesis 2a: The Catholic and mainline Protestant religiosity of a region will have a greater positive influence on the CCI decisions of companies in that region, than the region’s evangelical Protestant religiosity.
We turn next to an empirical examination of how Christian religiosity impacts companies’ CCI practices.
DATA, MEASUREMENTS, AND RESEARCH DESIGN
Data and Measurements of CCI
The sample of companies we use is drawn from the Kinder, Lydenberg, and Domini (KLD) STATS database (now the MSCI ESG KLD STATS dataset), between the years 1991 and 2013. The KLD database, which was launched in 1989, has been used since then in numerous CSR studies because of its high level of reliability (Sharfman, Reference Sharfman1996). Between 1989 and 2001, the companies in the KLD data included those listed in the S&P 500. In 2001 the dataset was expanded to include all the companies in the Russell 1000 (the 1,000 biggest companies in the US), and in 2002 it was expanded again to include all the companies in the Russell 3000 (the 3,000 biggest companies in the US). The KLD data also include 100 other companies that are included because of their exemplary CSR performance; however, we took these 100 companies out of our final dataset because we believed they might bias our study.
In the years between 1991 and 2013, KLD scored the CSR performance of firms in seven categories: environment, employee relations, corporate governance, diversity, human rights, products, and community relations. Within every category, the database gives each firm a score of one or zero for each of many distinct “strengths” (favorable characteristics) it can have in that category, and similarly for each of the many “concerns” (unfavorable characteristics) it can have in that category. Taking all the categories into account, the database has a total of about 80 strengths” and concerns” scores for each firm each year.
Previous studies of CSR that have used the KLD data have generally proceeded by aggregating all of a firm’s scores in all of these categories to arrive at a single annual score for each firm. Since our study only addresses the CCI performance of firms, however, we use only the scores a firm has received in the category of community relations to derive an annual score for each firm (the community relations index scores explained below). In its community relations category, KLD gives each company a binary score (0 or 1) on eight possible strengths and four possible concerns. Specifically, KLD community relations strength ratings are initiatives that benefit communities, including charitable giving, innovative giving, providing housing support, providing education support, maintaining non-US charitable giving, volunteer programs, community engagement, and other strengths such as having a strong in-kind giving program or engaging in other notably positive community activities. KLD’s concern ratings in the community relations category are activities harmful to communities, including having notable investment controversies, significant negative economic impacts, major tax disputes, and other community “controversy” concerns that mobilize community opposition. The KLD community strengths, we note, are the kinds of community-friendly initiatives advocated by the US Christian denominations discussed earlier, whereas the KLD concerns include the kinds of inequitable treatment they condemn.
Using the KLD data we construct two community relations index scores as our dependent variables: The first is what we call the community relations net score (COM_NET), which is simply the sum of a company’s community strengths scores in a given year minus the sum of its community concerns scores in that year. The second community relations variable we construct is what we call the community relations index score (COM_IDX). This variable is also constructed from a company’s community strengths and concerns for each year, but is designed to take into account certain variations that can occur in the KLD ratings from one year to the next. As indicated above, the KLD ratings consist of a zero or one that is assigned to each of a company’s strengths and concerns each year. However, the number of KLD strengths and concerns vary across the years. Consequently, we construct the community relations index score (COM_IDX) for each firm in a given year by first taking the sum of its community strengths scores for that year and subtracting the sum of its community concerns scores for that year; we then normalize this result by dividing it by the total number of possible strengths and concerns KLD used that year in its community relations category. (We note that this construction relies on a similar CSR index-making procedure previously used by Cui et al., Reference Cui, Jo, Na and Velasquez2015a, Reference Cui, Jo and Velasquez2015b, Reference Cui, Jo and Velasquez2016, and Jo & Harjoto, Reference Jo and Harjoto2011, Reference Jo and Harjoto2012). Thus, for each firm i in year t, we let COM ijt denote an indicator variable of community relations strength j for firm i in year t; we let COM ikt denote an indicator variable of community relations concern k for firm i in year t; and let COM jt and COM kt denote the maximum number of community relations strengths and concerns, respectively, in year t for any firm. The community relations index score, COM it, for each firm i for firm-year observation t is then:
$$CO{M^{it}} = {{\mathop \sum \nolimits_j CO{M^{ijt}} - \mathop \sum \nolimits_k CO{M^{ikt}}} \over {CO{M^{jt}} + CO{M^{kt}}}}$$Our community relations index score (COM_IDX) for each company, then, is the difference between the sum of its KLD community relations strengths values minus the sum of its KLD community relations concerns values (numerator), divided by the number of KLD community relations strengths and concerns KLD used that year (denominator). The community relations net score (COM_NET) and the community relations index score (COM_IDX) are our primary dependent variables. We use two distinct community relations variables to provide two separate measures of a firm’s community relations, which allows us to be more confident in the robustness of our analysis.
We also examine the impact of local religiosity on the extent to which a company embraces community initiatives that benefit the community (KLD strengths) and avoids those that harm the community (KLD concerns). To do this we construct two variables, COM_STR and COM_CON. The COM_STR variable is simply the sum of the scores KLD gives a company in a given year for its community relations strengths, while COM_CON is the sum it gives a company in a given year for its community relations concerns.
Measurement of Christian Religiosity
Our main independent variable measuring Christian religiosity is constructed from data provided by the American Religion Data Archive (ARDA). The ARDA includes the “US church membership data file at the county level” which indicates how many adherents to Christianity reside in each county in the United States. We divide the number of adherents residing in a particular county by the total number of residents of that county, to get the percentage of adherents to Christianity that reside in that county. We use this percentage as a proxy measure of the level of Christian religiosity in that county and we call this the county’s Christian religiosity score (REL). Several previous studies have used this same procedure with the ARDA data to derive the same kind of proxy measure of the religiosity of a county including: Cui et al. (Reference Cui, Jo, Na and Velasquez2015a, Reference Cui, Jo and Velasquez2015b, Reference Cui, Jo and Velasquez2016), Dyreng et al. (Reference Dyreng, Mayew and Williams2012), El Ghoul et al. (Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012), Grullon, Kanatas, and Weston (Reference Grullon, Kanatas and Weston2010), Hilary and Hui (Reference Hilary and Hui2009), McGuire et al. (Reference McGuire, Omer and Sharp2011, Reference McGuire, Newton, Omer and Sharp2012), and Omer et al. (2016). The ARDA dataset, however, is based on a census that is undertaken only every ten years (1971, 1980, 1990, 2000, and 2010). Our community relations scores and certain other data we use as controls, however, are compiled every year. We therefore linearly interpolate and extrapolate our religiosity variable (REL) to get values for the years that fall between each ARDA census (specifically, between 1990 and 2000, and between 2000 and 2010) which then allows us to match our annual REL values with the annual values of our two community relations measures (COM_NET and COM_IDX), as well as match them with the annual values of certain other control variables we explain below. This interpolation and extrapolation procedure follows Cui et al. (Reference Cui, Jo, Na and Velasquez2015a, Reference Cui, Jo and Velasquez2015b, Reference Cui, Jo and Velasquez2016), Dyreng et al. (Reference Dyreng, Mayew and Williams2012), El Ghoul et al. (Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012), Grullon et al. (Reference Grullon, Kanatas and Weston2010), and Hilary and Hui (Reference Hilary and Hui2009). We believe that the religiosity variable (REL) we construct is a valid proxy measure of the level of Christian religiosity present in a US county and we use this variable to determine whether and how a county’s Christian religiosity (REL) influences the community relations scores (COM_NET and COM_IDX) of companies headquartered in that county.
The ARDA dataset also indicates how many people living in a given US county belong to mainline Protestant, Catholic, or evangelical churches. Using this data we can calculate the percentage of the county’s population that belongs to each of these three groups. We use these percentages as proxy measures of the level of each county’s mainline Protestant religiosity (MAIN), its Catholic religiosity (CATHO), and its evangelical religiosity (EVAN). This lets us separate the Christian religiosity of each county (REL), into its three main constituents (MAIN, CATHO, and EVAN), and thereby enables us to determine whether and how each of these three religious groups influence the community relations scores (COM_NET and COM_IDX) of companies headquartered in that county. Like our main Christian religiosity variable (REL), the three main components of our Christian religiosity variable (MAIN, CATHO, and EVAN) are linearly interpolated and extrapolated to get values for the years that fall between the 10-year ARDA censuses.
Control Variables
In addition to our main dependent and independent variables, we construct a number of control variables to deal with issues that could affect our results. The first issue is related to where a company’s headquarters may be located in a given county. A company’s headquarters could be located near the border of a county and then its managers may live in the adjacent county and may be influenced by the religiosity of the adjacent county and not by the religiosity of the county in which their headquarters is located. Thus, the religiosity that actually influences a company’s management may depend on where its headquarters happens to be located within the county (i.e., near its borders or near the center of the county), as well as the kind of transportation available in that county, its road density, etc. In other words, our results may be affected by geographical factors. We deal with this issue by creating a variable that can capture such geographical factors in the error terms. We do this by constructing a variable that is associated with the religiosity of the county that is nearest to the county where the company’s headquarters is actually located. To do this we use the “County Distance Data” that is provided by the National Bureau of Economic Research which calculates the distance between counties using great-circle distances determined with the use of the formula of Haversine.Footnote 2 This enables us to construct the variable REL_I which is the level of religiosity of the country that is nearest to a company’s headquarters. We can then use the variable REL_I to determine whether there is any association between our main community relations scores and the level of religiosity of the nearest county (REL_I). By doing this we can ascertain whether geographical factors affect our results.
Another issue that may affect our results is related to certain company characteristics that may influence the company’s community relations. Community relations are a sub-category of a company’s CSR practices, and previous studies have shown that a company’s CSR practices may be related to certain company characteristics including its size, its risk, its investment growth opportunities, its financial performance, its advertising, and its R&D (Clarkson, Li, Richardson, & Vasvari, Reference Clarkson, Li, Richardson and Vasvari2011; Jo & Harjoto, Reference Jo and Harjoto2011, Reference Jo and Harjoto2012; Servaes & Tamayo, Reference Servaes and Tamayo2013). Since community relations are a sub-category of CSR, such company characteristics may also affect the company’s community relations. We test whether such characteristics affect a company’s community relations by using several control variables taken from the Compustat database (from which we draw our financial variables), and from the Center for Research in Security Prices (CRSP) dataset (which provides information on company stock returns). These controls include: each company’s R&D expenses ratio (RNDR), its total advertising expenditures ratio (ADVR), its debt ratio (DEBTR), its one-year sales growth rate (SALEG), its capital expenses ratio (CAPEX), the log of the market value of its equity (LOG_MVE), its size as measured by the log of the value of its total assets (LOG_TA), its risk as measured by how volatile its stock returns are (DEVRET), and its investment growth opportunities as measured by the market to book value of its equity (MBR).Footnote 3 We summarize these control variables in Table 1.
Variable Definition and Data Sources
Construction of the Final Sample
Our final sample is created by linking together the community relations index (COM_NET and COM_IDX) that we constructed using the KLD dataset, the county in which each firm is located, the financial variables that we drew from the Compustat and CRSP datasets, and the religiosity indices (REL, CATHO, MAIN, and EVAN) we created using the ARDA data. We first match each company in the KLD dataset with its community relations score, its county, and its financial variables. We then relate each company to the religiosity measures of the county in which it is headquartered.Footnote 4 Finally, we account for any lags and variations in our community relations (COM_NET and COM_IDX), religiosity (REL, CATHO, MAIN, and EVAN), and control variables. The final aggregated sample contains 28,963 firm-year observations dating from 1991 to 2013. We note that the samples we actually use in our regression analyses differ in minor ways from the aggregated sample because the data that is available for the various variables differs across regression models.
Analysis
Because our aim is to determine how a region’s religiosity influences the CCI initiatives of local companies, we begin by regressing the company community relation scores (COM_NET and COM_IDX) on the religiosity of the county in which each company is headquartered, along with the control variables we described above. To accomplish this we start by running the baseline fixed effect regressions below in order to deal with the time-invariant, firm-fixed effects in the relation between a firm’s community relation scores (COM_NET and COM_IDX) and the level of religiosity (REL):
$$COM\_NET\left( {or\,COM\_IDX} \right)_{i,t} = \alpha _0 + \alpha _1 REL_{i,\,t} + \sum\nolimits_j^n {\alpha _j CONTROL\,VARIABLES_{i,\,t - 1} + u_i + \varepsilon _{i,\,t} } $$Then, in subsequent analyses, we substitute the MAINi,t, EVANi,t, and CATHOi,t variables for the RELi,t variable.
RESULTS
Descriptive Statistics and Bivariate Correlation
Table 2, exhibits the summary statistics for our main variables as well as for our control variables. The scores range between -2.0 and 5.0 for the firms’ community relations net scores (COM_NET) and between -0.17 and 0.42 for community relations index scores (COM_IDX), implying there are wide variations in the community-friendly practices of the companies in our sample. The average level of Christian religiosity (REL) is 0.52, which means that on average the percentage of Christian adherents in a county (i.e., the total adherents in a county divided by the county’s population) is 52. The averages of companies’ financial characteristics reported in Table 2 are comparable with samples in previous studies, such as Dhaliwal, Li, Tsang, and Yang (Reference Dhaliwal, Li, Tsang and Yang2011), Ioannou and Serafeim (Reference Ioannou and Serafeim2015), and Jo and Harjoto (Reference Jo and Harjoto2011, Reference Jo and Harjoto2012).
Summary Statistics
Note.This table displays descriptive statistics for variables from 1991 to 2013, with 28,963 firm-year observations. Mean, standard deviation, minimum, median, and maximum are reported. Variable definitions are provided in Table 1. We measure the degree of religiosity by the percentage of adherents per county (total adherents/total population).
In Table 3, we present the spearman correlation coefficients for our main variables. Notably, our bivariate correlation coefficient between our main dependent variable (COM_NET) and main independent variable of religiosity (REL) is .04 (it is significant at the .05 level).
Bivariate Correlations
Note. N = 28,963. Variable definitions are provided in Table 1. * p < .05.
Multivariate Regression Results
We use a linear regression model to carry out some multicollinearity diagnostics. The results show that every individual variable’s variance inflation factor (VIF) values is under 6 (the mean VIF is less than 3 in every model). This implies that our main results are not significantly affected by multicollinearity. We also conduct regressions to account for any fixed effects that may affect firms in our sample. These regressions also allow us to impose some time independent effects related to variables that may be correlated with the regressors.Footnote 5 We also include dummy variables to control for year- and industry-fixed effects for all models.
Table 4 presents the results from the baseline fixed effects regression of the level of the community relations net score (COM_NET) and community relations index score (COM_IDX) on the level of Christian religiosity (REL) in Panel A (and on the level of MAIN, EVAN, and CATHO in Panel B), as well as with REL_I and our other control variables. Our results indicate that the influence of religiosity (REL) on the community relations net score (COM_NET) and community relations index score (COM_IDX) is positive and statistically significant, at least at the .05 level. Since our results remain essentially the same whether we use the COM_NET or the COM_IDX measures, we can conclude that our analysis is robust.
Regression Analyses Predicting Firms’ Community Relations Initiatives
Note. Variable definitions are provided in Table 1. Robust p values in parentheses. * p < .05, ** p < .01, *** p < .001.
We also notice that although our results are both significant and positive, the effect size of regressing the community relations net score (COM_NET) on our measure of Christian religiosity (REL) is 0.06, indicating that one unit of our Christian religiosity net measure increases local community involvement by 6 percent, depending upon our choice of firm size control measure. In addition, a one standard deviation increase in our religiosity measure (REL) increases the community relations net score (COM_NET) by 1 percent.
The positive associations among the level of our Christian religiosity measure (REL) and the level of our community relations net score (COM_NET) and our community relations index score (COM_IDX) all remain whether we control for company size using the lagged variable of the log of total assets (LAG(LOG_TA)) or control for size using the lagged value of the log of the market value of equity (LAG(LOG_MVE)). In addition, the associations remain both when we include and when we exclude growth opportunities as measured by the lag of market-to-book value of equity (LAG(MBR)).Footnote 6
The coefficients on the REL_I, the level of religiosity variable of the county that is nearest to a company’s headquarters, however, are all insignificant, suggesting that geographical factors do not influence our results. Our results also show that companies that have greater capital expenditures (CAPEX) or that are larger—as measured, for example, by the market value of their equity (MVE)—have higher levels of community-friendly initiatives. This result implies that if a company is sufficiently large or if it has a sufficiently high level of resources, it is more likely to invest in community-friendly initiatives. On the other hand, companies that have a high debt ratio (DEBTR) tend to invest in fewer community-friendly initiatives, perhaps because their debt payments constrain their ability to invest in such initiatives.Footnote 7
Similar to the methods used by Cui et al. (Reference Cui, Jo, Na and Velasquez2015a, Reference Cui, Jo and Velasquez2015b, Reference Cui, Jo and Velasquez2016) and El Ghoul et al. (Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012), Table 4 Panel B presents the results from the baseline fixed effect regression of the level of our community relation measures (COM_NET and COM_IDX) on the level of our three subsidiary religiosity measures (MAIN, EVAN, and CATHO) with controls. The importance of these regressions is that they reveal the different ways in which mainline Protestant, Catholic, and evangelical Protestant religiosities affect the likelihood that local companies will engage in community-friendly initiatives. What is surprising is that although the coefficients on mainline Protestant religiosity (MAIN) and on Catholic religiosity (CATHO) show that both of these have a positive and significant (at the 1% level) influence on our community relations measures (COM_NET and COM_IDX), the coefficients on evangelical Protestant religiosity (EVAN), while statistically significant (also at the 1% level), are nevertheless negative. This implies that evangelical religiosity has a negative impact on the community-friendly choices of local firms while both mainline Protestant religiosity and Catholic religiosity have a positive impact. These differences support our “differentiated responses hypothesis” (Hypothesis 2) and our “differentiated Christian denominational responses hypothesis” (Hypothesis 2a).
Next, we examine whether companies in communities with higher levels of religiosity seem to have a greater involvement in the positive dimensions (the KLD “strengths”) of community issues (COM_STR) or seem to be more involved in resolving their negative (the KLD “concerns”) community issues (COM_CON) (see Appendix for more detail on the KLD “strengths” and “concerns”).
Table 5, Panel A, reports the results from the fixed-effects regressions for community relation strengths (COM_STR) and community relation concerns (COM_CON).Footnote 8 In Panel B, in place of our Christian religiosity measure (REL), we use the three main components of Christian religiosity: mainline Protestant religiosity (MAIN), evangelical Protestant religiosity (EVAN), and Catholic religiosity (CATHO). One coefficient on REL, shown in Panel A, is positive for community relation strengths (COM_STR) and both coefficients on REL are significantly negative for community relations concerns (COM_CON). These imply that Christian religiosity mildly increases local companies’ community strengths, whereas it significantly decreases companies’ community concerns, which again confirms our religious morality hypothesis (Hypothesis 1). However, the coefficients on MAIN, EVAN, and CATHO provided in Panel B, imply that the impact of religiosity on community strengths and community concerns is not the same for all three religious groups. Both the expected positive relation between Christian religiosity (REL) and community relations strengths (COM_STR) as well as the expected negative relation between Christian religiosity (REL) and community relations concerns (COM_CON) seem to derive mainly from the effect of Catholic religiosity (CATHO) and mainline Protestant religiosity (MAIN). On the other hand, the coefficients on evangelical religiosity (EVAN) are significantly negative for community relation strengths (COM_STR) and significantly positive for community relation concerns (COM_CON). Moreover, our community relation strengths (COM_STR) regressions explain approximately 28% of the variance in beneficial community relations practices, whereas the COM_CON models explain approximately 19 percent of the variance in harmful community relations practices. These results imply that companies headquartered in areas with higher levels of mainline Protestant religiosity and Catholic religiosity, are concerned with rectifying harmful company practices and implementing initiatives that are beneficial to local communities, whereas companies embedded in communities with higher levels of evangelical Protestant religiosity tend to be neither concerned with implementing initiatives that are beneficial to the community nor concerned with correcting firm practices that hurt the community. Thus, Panel B results are supportive of our differentiated responses hypothesis (Hypothesis 2) and our differentiated Christian denominational responses hypotheses (2a).Footnote 9 Our main empirical findings are summarized in Table 6.
Regression Analyses Predicting Community Relations Strengths and Concerns
Note.Variable definitions are provided in Table 1. Robust p values in parentheses. * p < .05, ** p < .01, *** p < .001.
Summary of Empirical Findings
DISCUSSION
As expected, the results of our study support the hypothesis that Christian morality in general has a positive effect on company CCI initiatives. Panel A of Table 4 shows that the level of Christian religiosity of a county is positively associated with the extent to which companies headquartered in that county engage in CCI, while Table 5 Panel A indicates that greater levels of Christian religiosity are positively associated with greater levels of company initiatives that benefit the community and negatively associated with company initiatives that are harmful to the community. These associations between religiosity and companies’ CCI initiatives remain even after we test them with a number of control variables.
Our results also imply, however, that the association between Christian religiosity and CCI initiatives is complex. As Table 4 Panel A suggests, although Christian religiosity influences firms’ community involvement, the amount firms will invest in community initiatives is partly dependent on their size, the financial resources they have available, and the amount of debt they carry. We can explain this latter result if we assume that companies that have to make high interest payments are financially restrained in a way that limits their ability to invest in community-friendly initiatives. Our results show, therefore, that although noneconomic religious factors affect the likelihood that a company will invest in community-friendly initiatives, how much a company invests in its community depends also on its economic resources.
However, a more interesting aspect of the complexity of the relationship between religiosity and community initiatives emerged when we disaggregated our main religiosity construct (REL) into its three main constituents: mainline Protestant religiosity (MAIN), Catholic religiosity (CATHO) and evangelical Protestant religiosity (EVAN). As Table 4 Panel B shows, these three constituents of Christian religiosity have heterogeneous influences on the CCI initiatives of companies (measured by COM_NET and COM_IDX). Whereas mainline Protestant religiosity and Catholic religiosity both have a positive impact on a company’s community relations decisions, evangelical Protestant religiosity generally has a negative impact. Moreover, Table 5 Panel B demonstrates that while mainline Protestant religiosity and Catholic religiosity both encourage beneficial community initiatives and discourage those that harm communities, evangelical Protestant religiosity discourages beneficial corporate community initiatives, and encourages harmful corporate community initiatives!
The finding that unlike mainline Protestant and Catholic religiosity, evangelical Protestant religiosity appears to discourage firms from engaging in community-friendly initiatives and/or fails to prevent firms from engaging in activities that harm communities, is a consequence, we suggest, of the significant institutional differences that, as we saw, exist among the three main groups of Christian denominations. Whereas both Catholic and mainline Protestant denominations have a long and consistent history of advocating the equitable treatment of communities, evangelical Protestant denominations do not share a similar history. Moreover, a large number of evangelical Protestants still believe today—as most believed during most of the twentieth century—that efforts to help local communities are distractions from the more valuable tasks of personal salvation and of evangelizing as much of the world as possible before the imminent Second Coming of Jesus (Sutton, Reference Sutton2014). The differences we see in the way that the different denominations influence company CCI initiatives, then, support our differentiated responses hypothesis (H2), and our differentiated Christian denominational responses hypothesis (H2a).
Our results may be questioned, however, because surely there is nothing about the logic of evangelical beliefs and values that should lead us to expect that evangelicals would actually encourage companies to harm communities or discourage companies from helping communities. However, two factors help explain our results. First, initiatives that benefit or help communities entail costs and multiple economic factors pressure companies to avoid costs. If evangelical denominations do nothing to encourage costly investments in initiatives that benefit communities, their inaction will lead companies to avoid such initiatives. Second, corporate operations often have external effects that harm local communities. If evangelical denominations do nothing to encourage companies to prevent or ameliorate such harmful external effects, their inaction will lead companies to do nothing to prevent or mitigate external effects that harm local communities. Evangelicals’ inaction on these two fronts is explained by the logic of evangelical institutions, specifically the belief that their efforts should be wholly devoted to the more urgent and valuable ends of ensuring their personal salvation and evangelizing the world before the imminent Second Coming of Jesus. We are not arguing that their institutional beliefs lead evangelicals to actively discourage beneficial activities or actively encourage harmful activities. Our view, rather, is that the logic of evangelical beliefs and values can lead to a lack of action about corporate matters that evangelicals see as less valuable and less urgent than salvation and evangelization, and this inaction leads to the results our study finds.
To summarize, whereas Christian religiosity in general has a positive influence on the CCI initiatives of local firms—which supports our religious morality hypothesis—the three major constituents of US Christian religiosity—mainline Protestant religiosity, Catholic religiosity, and evangelical Protestant religiosity—differ significantly in the type of influence they exert. Although mainline Protestant religiosity and Catholic religiosity both exert an influence that is largely positive, evangelical Protestant religiosity exerts an influence that is largely negative. This supports our differentiated responses hypothesis and our differentiated Christian denominational responses hypothesis.
Limitations
We should note that our study has some important limitations. First, the data drawn from the American Religion Data Archive (ARDA) is not produced on an annual basis, and this has forced us to use a linear interpolation method to derive values for the years that are missing. The fact that this unavoidable procedure has been used in several previous studies (e.g., Cui et al., Reference Cui, Jo, Na and Velasquez2015a, Reference Cui, Jo and Velasquez2015b, Reference Cui, Jo and Velasquez2016; Dyreng et al., Reference Dyreng, Mayew and Williams2012; El Ghoul et al., Reference El Ghoul, Guedhami, Ni, Pittman and Saadi2012; Grullon et al., Reference Grullon, Kanatas and Weston2010; Hilary & Hui, Reference Hilary and Hui2009) gives us some confidence that it is permissible to use this procedure, but we acknowledge that its use has the potential to introduce some interpolation bias. In order to check for the presence of such bias, we replicated our regressions using only data from the years for which ARDA provides actual survey data (1990, 2000, and 2010). The (unreported) results of these regressions necessarily rely on a much smaller sample, but they mirror the same significant and positive associations between community relations and religiosity that we found in our regressions using the interpolated data. This suggests that our use of the linear interpolation procedure does not introduce systematic noise into our study’s results.
Second, although KLD is one of the most highly respected and independent CSR ratings in the world and has been widely used by scholars in the fields of business ethics, religious studies, finance, economics, accounting, marketing, strategy, and management, the KLD data has some important problems that should be noted. First, KLD does not publish the criteria it uses to score each of its categories which introduces a certain measure of uncertainty about the precise meaning of those categories. Second, the KLD data has an unbalanced panel structure and has been criticized for having some construct validity issues (Chatterji, Levine, & Toffel, Reference Chatterji, Levine and Toffel2009). A third problem with the KLD data is that during its initial years it contained data that may have given rise to a selection bias. During the 1990s, the KLD dataset included the companies in the S&P 500, plus a number of companies drawn from the Domini 400 Social Index. The companies taken from the Domini 400 Social Index were selected because of their CSR performance, and the inclusion of such companies may have introduced a selection bias into the dataset. To deal with this potential selection bias, we included year-fixed effects in our fixed effects regressions (and also in our unreported dynamic GMM estimation). In spite of all these problems, however, Sharfman (Reference Sharfman1996) and others have argued that scholars should have confidence in the KLD data and that these data capture the core elements of CSR.
Significance and Contributions
Our study contributes to the literature in several distinct research domains. The first is related to the factors that influence corporate community involvement. The research on CCI is expansive and significant efforts have been made to understand the factors that influence corporate decisions to invest in CCI initiatives (Bowen et al., Reference Bowen, Newenham-Kahindi and Herremans2010). Previous studies have looked at how investments in CCI initiatives are influenced by the firm’s size (Seifert et al., Reference Seifert, Morris and Bartkus2003), industry (Useem, Reference Useem1988), and financial performance (Galaskiewicz, Reference Galaskiewicz1997; Wang et al., Reference Wang, Choi and Li2008), along with other economic factors. To the best of our knowledge, however, our study is the first to examine the influence of religion on CCI and the first to find that Christian religiosity has a largely positive effect on CCI investments.
Second, our results show that the moral teachings of Christian religious groups influence the behaviors of people who reside among those groups (Ellison et al., Reference Ellison, Burr and McCall1997; Regnerus, Reference Regnerus2003; Welch et al., Reference Welch, Tittle and Petee1991). This result adds to the growing psychological and sociological findings of how religion affects behavior. Previous research has shown that religiosity influences adult decisions about premarital sex (Barkan, Reference Barkan2006), adult decisions to follow the law (Grasmick et al., Reference Grasmick, Kinsey and Cochran1991), truth telling among college students (Perrin, Reference Perrin2000), the voting behavior of adults (Greeley & Hout, Reference Greeley and Hout2008; Hirschl, Booth, & Glenna, Reference Hirschl, Booth and Glenna2009; Manza & Brooks, Reference Manza and Brooks1997; Regnerus, Sikkink, & Smith,1999), everyday adult decision making (Schieman, Reference Schieman2011), and adolescent decision making (Baier & Wright, Reference Baier and Wright2001; Bearman & Bruckner, 2001; Brownfield & Sorenson, Reference Brownfield and Sorenson1991; Donahue & Benson, Reference Donahue and Benson1995). We are now able to add to these findings the conclusion that religious moral views can also influence corporate CSR decisions; in particular, they generally have a positive influence on corporate CCI decisions.
We note, however, that although Christian religiosity usually has a positive effect on CCI, it should not be assumed that religiosity has a positive effect on all aspects of CSR. Earlier studies of the influence of religiosity on certain aspects of CSR—particularly its influence on companies’ environmental decisions—have shown that religiosity can have a negative influence (Cui et al., Reference Cui, Jo and Velasquez2015b), while our study of CCI shows that evangelical religiosity has a negative influence on companies’ CCI decisions. Thus, religiosity seems to encourage some aspects of CSR and discourage others. Whether religiosity will encourage or discourage a specific element of CSR (such as companies’ stance on the environment, diversity, employee relations, etc.) depends very much on the particular teachings and “institutional logics” of the specific religions or religious denominations being studied.
Third, our study also contributes to the literature on institutional theory. Previous studies have examined how different kinds of institutions affect organizational behaviors and structures (DiMaggio & Powell, Reference DiMaggio and Powell1991; Hotho & Pedersen, Reference Hotho, Pedersen, Demirbag and Wood2012). Virtually no institutional studies, however, have looked at how religious institutions affect organizations (Tracey, Reference Tracey2012). Our study of how the institutions of Christianity affect corporate community decisions, therefore, is among the first to examine the influence of this important social institution on organizations.
Fourth, we think that our study also has some significant lessons for managers. Our results indicate that the religious groups in the region within which a company’s headquarters is located will influence the company’s community-friendly programs, as well as the extent to which the firm may engage in practices that can prove harmful to the community. For example, when a company is located among a group of evangelical Protestants, it is more likely to have a tendency to avoid most kinds of community-friendly programs while allowing activities that are harmful to local communities. On the other hand, if it is located in a region that is populated largely by Catholics or mainline Protestants, the company likely will tend to avoid activities that can harm the community and engage in activities that benefit the community. Because local communities are stakeholders that often have the power to help or hinder the company’s ability to achieve its goals, the company should identify the tendencies produced by the religious groups around it and it should deal with those tendencies in ways that will be more likely to secure the support of these key stakeholders. The manager who wants to ensure favorable community relations should therefore ascertain the nature of the religious groups that surround its headquarters, and use the results of our study to estimate the kinds of community relations those religious groups are likely to encourage, since these might not be consistent with the kinds of relations the manager desires and may lead to the loss of the community’s support.
Directions for Future Studies
We believe that our study suggests a number of topics that call for future scholarly research. First, although our study shows that the religiosity of the area surrounding managers affects their community decisions, the nature of our data does not allow us to determine whether the personal religiosity of company managers affects the community decisions they make. We believe that a study based on interview or survey data would shed light on this interesting issue. Second, because our study looks only at the influence of Christian religiosity on corporate community initiatives, we do not know how the particular teachings and institutional logics of non-Christian forms of religiosity influence corporate community initiatives or, more generally, CSR initiatives. How, for example, do the teachings and institutional structures of Buddhism, Confucianism, Hinduism, Islam, Judaism, and Taoism affect the CSR decisions managers make? We believe that the methods and theoretical approaches we developed in our study might fruitfully be extended to study the influence of non-Christian religions on managerial decision making.
Third, our study looks only at US religious denominations and their impact on the US corporate community. Several studies, however, have argued that the role of religion in American life is quite different from the one it has in other nations (e.g., Micklethwait & Wooldridge, Reference Micklethwait and Wooldridge2009), to say nothing of the different religions that are dominant in other nations. In addition, corporate communities outside the United States place different emphases on community relations. For example, EIRIS, VIGEO, and Sustainanalitics, three corporate rating agencies located in the European Union, stress community relations to a much greater extent and they incorporate several indicators that are not addressed by KLD. Consequently, a company that in the US may seem to be quite community friendly might not be seen as being particularly community friendly in Germany or France. In the absence of additional studies, then, we are unable to conclude that our results are applicable to the religions or corporate communities of other nations. Additional studies are needed to determine whether non-Christian religiosities have the sort of influences on foreign firms that we identify in this study.
Fourth, it will be fruitful if future studies examine how corporate community involvement and the local religiosity of the mainline Protestant, Catholic, and evangelical denominations impact corporate financial performance. Given the difference in relationships exhibited by different denominations discussed earlier, it would seem useful to examine interactive effects among these three denominations to understand whether there are any particular distributions that are more beneficial or damaging.
ACKNOWLEDGEMENTS
We greatly appreciate editor in chief Bruce Barry, associate editor Jerry Goodstein, and two anonymous referees for their many valuable comments and detailed guidance.
APPENDIX: DEFINITIONS OF THE CORPORATE COMMUNITY INVOLVEMENT STRENGTH AND CONCERN ITEMS
Community Strengths (COM_STR)
Generous Giving. The company has consistently given over 1.5% of trailing three-year net earnings before taxes (NEBT) to charity, or has otherwise been notably generous in its giving.
Innovative Giving. The company has a notably innovative giving program that supports nonprofit organizations, particularly those promoting self-sufficiency among the economically disadvantaged. Companies that permit nontraditional federated charitable giving drives in the workplace are often noted in this section as well.
Support for Housing. The company is a prominent participant in public/private partnerships that support housing initiatives for the economically disadvantaged (e.g., the National Equity Fund or the Enterprise Foundation).
Support for Education. The company has either been notably innovative in its support for primary or secondary school education, particularly for those programs that benefit the economically disadvantaged, or the company has prominently supported job-training programs for youth.
Non-US Charitable Giving. The company has made a substantial effort to make charitable contributions abroad, as well as in the US. To qualify, a company must make at least 20% of its giving, or have taken notably innovative initiatives in its giving program, outside the US.
Volunteer Programs. The company has an exceptionally strong volunteer program.
Community Engagement. The company has a notable community engagement program concerning involvement of local communities in areas where the firm has major operations.
Other Strength. The company has either an exceptionally strong in-kind giving program or engages in other notably positive community activities.
Community Concerns (COM_CON)
Investment Controversies. The company is a financial institution whose lending or investment practices have led to controversies, particularly ones related to the Community Reinvestment Act.
Community Impact. This indicator measures the severity of controversies related to a firm’s interactions with communities in which it does business. Factors affecting this evaluation include, but are not limited to, a history of involvement in land use and/or development-related legal cases, widespread or egregious community impacts due to company operations, and criticism by NGOs and/or other third-party observers.
Tax Disputes. The company has recently been involved in major tax disputes involving Federal, state, local, or non-US government authorities, or is involved in controversies over its tax obligations to the community.
Other Concern. The company is involved with a controversy that has mobilized community opposition, or is engaged in other noteworthy community controversies.
Note. This appendix lists the strength and concern dimensions in the KLD Community category based on KLD ratings definitions, User Guide & ESG Ratings Definition, 2013 and 2014 (reprinted with permission from MSCI). Dummy values are used to identify a company’s community status for each strength or concern.
