INTRODUCTION
While the majority of studies on Corporate Social Performance (CSP) have dealt with firms from developed countries, in recent years there has been increased attention devoted to the CSP of firms from developing countries (Chapple & Moon, Reference Chapple and Moon2007, Reference Chapple and Moon2005; Visser, Reference Visser, Crane, Matten, McWilliams, Moon and Siegel2008; Gugler & Shi, Reference Gugler and Shi2009; Jamali & Mirshak, Reference Jamali and Mirshak2007; Baskin, Reference Baskin2006). Following Wood (Reference Wood1991: 693), we focus on developing country firms’ CSP defined as “a business organization’s configuration of principles of social responsibility, processes of social responsiveness, and policies, programs, and observable outcomes as they relate to the firm’s societal relationships.” Understanding the CSP of developing country firms is increasingly important for two reasons. First, developing country multinationals (DMNCs), which we define as firms headquartered in developing countries that operate foreign subsidiaries (Hitt, Hoskisson, & Kim, Reference Hitt, Hoskisson and Kim1997; Kotabe, Srinivasan, & Aulakh, Reference Kotabe, Srinivasan and Aulakh2002), now account for a large proportion of overseas foreign direct investment (OFDI) globally (UNCTAD, 2014; Meyer & Thaijongrak, Reference Meyer and Thaijongrak2013) and, as such, their CSP has increasingly widespread implications. Second, as they internationalize, DMNCs face reputational and legitimacy challenges (Merz, Peloza, & Kim, Reference Merz, Peloza and Chen2010; Zaheer, Reference Zaheer1995; Graafland, Reference Graafland2002; Gunningham, Kagan, & Thornton, Reference Gunningham, Kagan and Thornton2004) that they may try to address by improving their CSP.
In this article, we investigate why and how the extent and nature of DMNC multinationality, defined as “business activities that span across national boundaries” (Tseng, Tansuhaj, Hallagan, & McCullough, Reference Tseng, Tansuhaj, Hallagan and McCullough2007: 961), lead to changes in their CSP as compared to that of their counterparts operating only domestically. DMNCs venture abroad for several reasons. First, following the pattern established by multinationals from developed economies, DMNCs try to exploit competitive advantages they have developed at home (Cavusgil, Reference Cavusgil1980). These advantages might include technologies, brands or even CSP capabilities. However, previous research suggests that the competitive advantages DMNCs seek to exploit overseas are “non-traditional” including capabilities for: cost innovation (Zeng & Williamson, Reference Zeng and Williamson2007; Williamson & Zeng, Reference Williamson, Zeng, Ramamurti and Singh2009); efficiently unlocking latent demand in low-end segments (Prahalad, Reference Prahalad2006); optimizing products and processes for emerging markets (Ramamurti, Reference Ramamurti2012); dealing with weak institutions and infrastructure (Morck, Yeung, & Zhao, Reference Morck, Yeung and Zhao2008; Cuervo-Cazurra & Genc, Reference Cuervo-Cazurra and Genc2008) or optimizing their value chains globally in ways that allow their low-cost talent and resources to be leveraged effectively in emerging markets (Williamson, Ramamurti, Fleury, & Fleury, Reference Williamson, Ramamurti, Fleury and Fleury2013). The second main reason for DMNCs to internationalize is so-called “strategic asset seeking” (Rui & Yip, Reference Rui and Yip2008; Deng, Reference Deng2009; Makino, Lau, & Yeh, Reference Makino, Lau and Yeh2002). Internationalization can act as a springboard for DMNCs to acquire assets that may compensate for their competitive disadvantages, overcome late-comer disadvantage, counter-attack global rivals in their home market, by-pass trade barriers, alleviate domestic institutional constraints, secure preferential treatment from their home governments and access complementary assets to exploit their competitive advantages abroad (Luo & Tung, Reference Luo and Tung2007).
These rationales for DMNCs venturing abroad would suggest that any shift in their CSP relative to that of their purely domestic cousins is likely to be a consequence, rather than a cause, of their internationalization. When DMNCs do venture abroad, there is evidence that they face both reputation and legitimacy deficits in the eyes of the various international stakeholders in the foreign countries in which they operate (Merz et al., Reference Merz, Peloza and Chen2010; Petersen & Pedersen, Reference Petersen and Pedersen2002; Zaheer, Reference Zaheer1995; Gunningham et al., Reference Gunningham, Kagan and Thornton2004). These deficits result from both the fact that DMNCs are foreign (a potential issue for all multinationals) and that they originate from developing countries. In investigating how DMNCs address these deficits, we develop a number of hypotheses concerning the variation in CSP between DMNCs and their counterpart domestic-only firms, and the variation in CSP between DMNCs as a result of the extent of their multinationality and portfolio of host countries.
We test our hypotheses using a sample of 412 firms headquartered in five developing countries (Brazil, Russia, India, China and South Africa) between 2009 and 2012.
THEORETICAL DEVELOPMENT
In establishing themselves internationally, DMNCs often face two stakeholder-related issues: a legitimacy deficit and a reputational deficit (Merz et al., Reference Merz, Peloza and Chen2010; Petersen & Pedersen, Reference Petersen and Pedersen2002; Zaheer, Reference Zaheer1995; Graafland, Reference Graafland2002; Gunningham et al., Reference Gunningham, Kagan and Thornton2004), which they and their national counterparts do not face at home. We argue that DMNCs try to address both deficits by improving their CSP.
The Legitimacy Deficit
According to Suchman (Reference Suchman1995: 574), legitimacy refers to the “generalized perception or assumption that the actions of an entity are desirable, proper, or appropriate within some socially constructed system of norms, values, beliefs, and definitions.” Similarly, drawing on prior research, Deephouse and Carter (Reference Deephouse and Carter2005: 331) state that: “a central element of legitimacy, as currently understood, is meeting and adhering to the expectations of a social system’s norms, values, rules, and meanings.” Organizational legitimacy is a very valuable organizational resource that contributes to the long-term survival of firms (Ashforth & Gibbs, Reference Ashforth and Gibbs1990; Suchman, Reference Suchman1995). Suchman (Reference Suchman1995) identifies two forms of legitimacy, relevant for our purposes here: pragmatic legitimacy and moral legitimacy. Pragmatic legitimacy results from the self-interested behavior of stakeholders who will support a particular organization if they believe it is in their interest to do so. Moral legitimacy reflects stakeholders’ normative evaluation (positive or negative) of the organization’s activity, i.e. whether what the organization does is “the right thing to do” (Suchman, Reference Suchman1995: 579).
There are two factors that might negatively influence the legitimacy of DMNCs when they expand internationally: a liability of foreignness and hailing from a developing country. The liability of foreignness (Zaheer, Reference Zaheer1995; Petersen & Pedersen, Reference Petersen and Pedersen2002; Bhanji & Oxley, Reference Bhanji and Oxley2013; Hymer, Reference Hymer1976) refers to “all additional costs a firm operating in a market overseas incurs that a local firm would not incur” (Zaheer, Reference Zaheer1995: 343). This liability may result from a kind of economic nationalism that foreign firms often face. Stakeholders faced with a foreign firm entering their country might ask themselves “does this firm adhere to our norms, values and meanings?” Their responses will impact the degree to which they support the foreign firm by providing it with the resources or demand necessary for its survival, which may give rise to legitimacy deficits (Mezias, Reference Mezias2002; Campbell, Eden, & Miller, Reference Campbell, Eden and Miller2012).
DMNCs are likely to face more critical questioning of their legitimacy by foreign stakeholders who often hold negative perceptions of firms from developing countries. Stakeholders (from both developing and developed countries) might consider (other) developing-country firms to have lax environmental or labor standards and be subject to ownership or influence by non-transparent and possibly corrupt governments (Bilkey & Nes, Reference Bilkey and Nes1982; Jaffe & Nebenzahl, Reference Jaffe and Nebenzahl2006; Noorderhaven & Harzing, Reference Noorderhaven and Harzing2003). Of course, not all of these negative, often simplistic, stereotypes apply equally to all developing countries (Kotler, Haider, & Rein, Reference Kotler, Haider and Rein1993). For example, the issue of corrupt links to government might be more serious in firms from Russia and State Owned Enterprises (SOEs) from China than for DMNCs from India and South Africa. However, even if there are significant differences in the effect that a particular developing country of origin might have on its DMNCs, there is substantial evidence in the literature that, on average, foreign firms from developing countries are associated with more negative stakeholder perceptions (Chinen, Jun, & Hampton, Reference Chinen, Jun and Hampton2000; Loo & Davies, Reference Loo and Davies2006; Verlegh & Steenkamp, Reference Verlegh and Steenkamp1999; Luo & Tung, Reference Luo and Tung2007).
We can expect this joint influence of the liability of foreignness and developing country of origin on the legitimacy of DMNCs to operate through both the pragmatic and moral legitimacy channels that Suchman (Reference Suchman1995) identifies. First, stakeholders might challenge the pragmatic legitimacy of DMNCs simply because they are foreign. For example, if customers have doubts about whether a foreign firm shares their own values and norms, they might hesitate to buy its products. They might challenge the firm’s pragmatic legitimacy by asking themselves questions like: “should we expose ourselves to the products of a firm whose values and norms we are not sure of?” Alternatively, state bureaucrats and politicians might ask themselves if it is in their country’s interest to facilitate the entry of a foreign firm with uncertain values and norms. This challenge to the pragmatic legitimacy of foreign firms might be more severe for DMNCs because, coming from developing countries, they are often associated with low quality/cost products, fears that they might displace local production (Cordell, Reference Cordell1993; Aulakh, Rotate, & Teegen, Reference Aulakh, Rotate and Teegen2000) or even pose security threats (Lindsay Reference Lindsay2015).
Second, stakeholders might ask themselves if it is morally acceptable for them to support a firm with questionable, or at least unknown, moral values, thus challenging the foreign firm’s moral legitimacy. Again, such challenges to the moral legitimacy of foreign firms might be more intense in cases of DMNCs because of their association with home countries with poor track records in areas such as labor and environmental standards and corruption (Jaffe & Nebenzahl, Reference Jaffe and Nebenzahl2006; Noorderhaven & Harzing, Reference Noorderhaven and Harzing2003; Visser, Reference Visser, Crane, Matten, McWilliams, Moon and Siegel2008). These negative perceptions of DMNCs can persist despite stakeholders acknowledging significant differences between firms across different countries and regions (Chambers, Chapple, Moon, & Sullivan, Reference Chambers, Chapple, Moon and Sullivan2003; Brammer & Millington, Reference Brammer and Millington2005).
The Reputation Deficit
According to Fombrun (Reference Fombrun1996: 36), the reputation of a corporation is “the overall estimation in which a particular company is held by its various constituents.” Despite being one of the most important intangible assets a firm can possess, a growing body of research suggests that a good reputation can help a firm gain sustainable competitive advantage leading to sustained superior financial returns (Dierickx & Cool, Reference Dierickx and Cool1989; Fombrun & Shanley, Reference Fombrun and Shanley1990; Roberts & Dowling, Reference Roberts and Dowling2002; Barney, Reference Barney1991). A good corporate reputation can send a signal to consumers that a firm’s products and services are of high quality, and hence justify a premium price (Shapiro, Reference Shapiro1983). Corporate reputation can also help firms overcome crises by functioning as a “reservoir of goodwill” (Jones, Jones, & Little, Reference Jones, Jones and Little2000; Schnietz & Epstein, Reference Schnietz and Epstein2005).
In comparing the concepts of reputation and legitimacy, Deephouse and Carter (Reference Deephouse and Carter2005) find a number of similarities, including the fact that they are both stakeholder evaluations, they both share some of the same antecedents (i.e. size, philanthropy, and strategic alliances) and they both enable firms to acquire resources. On the other hand, drawing on prior research (Lawrence, Reference Lawrence1998; Ruef & Scott, Reference Ruef and Scott1998; Deephouse, Reference Deephouse2000; Heugens, Van Riel, & Van Den Bosch, Reference Heugens, Van Riel and Van Den Bosch2004), they maintain that a major difference is the fact that, while legitimacy is related to social acceptance and stems from stakeholder perceptions about whether a firm can be expected to adhere to, or comply with, social “norms, values, rules, and meanings” (Deephouse & Carter, Reference Deephouse and Carter2005: 331), reputation is related to a perception of excellence and favorable standing which results from stakeholders comparing organizations to determine their “relative standing” in a relevant dimension. As Lawrence (Reference Lawrence1998: 1122) puts it: reputation “differentiates” the “outstanding” firms from the ranks of the merely “qualified”, i.e. legitimate (cited in Deephouse & Carter, Reference Deephouse and Carter2005: 332). Thus, while determinations of legitimacy can be seen as resulting from the question: ‘is this organization acceptable to us?’ those of reputation can be seen as resulting from the question: ‘of these organizations, which one is better, in terms of [X]?’
Given this distinction between legitimacy and reputation, we argue that when DMNCs enter foreign markets, in contrast to their national counterparts that do not operate abroad, they can be expected to suffer both deficits. First, due to their inherent liability of foreignness (Hymer, Reference Hymer1976; Zaheer, Reference Zaheer1995), DMNCs will not have the established reputations of local, host-country firms. Most likely, DMNCs will be unfamiliar to host-country stakeholders (Campbell et al., Reference Campbell, Eden and Miller2012; Kostova & Zaheer, Reference Kostova and Zaheer1999), who could doubt their ability to perform as well as local firms. Second, because DMNCs are likely to face not only local firms, but also incumbent MNCs with generally stronger and established global reputations when they enter foreign markets, they will suffer a reputational deficit as unknown entities, relative to these established competitors (Nachum, Reference Nachum2010; Mathews, Reference Mathews2006). It can be argued that, generally speaking, MNCs have established global reputations with the stakeholders in a given country, even if they do not operate in that country. When Coca-Cola re-entered Myanmar in 2013, for example, the company found that it had an established reputation with many consumers and other potential stakeholders as a result of its global activities, even though it had not been officially present in the country for over 60 years (Smith, Reference Smith2013). This is something that is not the case with DMNCs that are relatively unknown before entering a foreign market.
The reputational deficit that DMNCs might suffer because they are unfamiliar with foreign target markets will be exacerbated if foreign stakeholders fill the gap in their knowledge with negative perceptions associated with DMNCs’ developing-country origins (Kostova & Zaheer, Reference Kostova and Zaheer1999; Campbell et al., Reference Campbell, Eden and Miller2012). As Luo and Tung (Reference Luo and Tung2007: 494) state, a developing country of origin could “tarnish organizational reputation and hinder shareholder confidence and relationship building with global stakeholders” so that DMNCs suffer from a negative halo effect (Jaffe & Nebenzahl, Reference Jaffe and Nebenzahl2006; Noorderhaven & Harzing, Reference Noorderhaven and Harzing2003). Stakeholders of host countries might question the quality of the products or services of DMNCs given the extensive evidence that a country’s economic development level plays an important role in the way the products of its firms are received internationally (Verlegh & Steenkamp, Reference Verlegh and Steenkamp1999; Chinen, Jun, & Hampton, Reference Chinen, Jun and Hampton2000; Loo & Davies, Reference Loo and Davies2006).
The Potential of CSP to Mitigate Legitimacy and Reputational Deficits
There are a number of reasons why DMNCs would try to make up for their legitimacy and reputation deficits by enhancing their CSP. First, there is a great deal of empirical evidence to suggest that good CSP has a positive impact on a firm’s stock of intangible assets, including corporate reputation (Fombrun, Reference Fombrun1996; Hall, Reference Hall1992) and legitimacy (Rao, Reference Rao1994; Suchman, Reference Suchman1995). Second, researchers have also identified the ways in which CSP affects the reputation and legitimacy of firms operating across different countries in particular (Gardberg & Fombrun, Reference Gardberg and Fombrun2006).
By enhancing its CSP, a firm can increase its legitimacy (Suchman, Reference Suchman1995; Deephouse & Carter, Reference Deephouse and Carter2005), both pragmatic and moral. A firm’s enhanced CSP can signal to its stakeholders that it has good values, norms and morals, prompting stakeholders to believe they can only benefit by supporting the firm (pragmatic legitimacy) and that they should support the firm (moral legitimacy). A number of researchers have confirmed this positive impact of improved levels of CSP on corporate legitimacy (Handelman & Arnold, Reference Handelman and Arnold1999; Panwar, Paul, Nybakk, Hansen, & Thompson, Reference Panwar, Paul, Nybakk, Hansen and Thompson2014; Du & Vieira, Jr, Reference Du and Vieira2012).
Given the legitimacy deficits that DMNCs are likely to face in host markets and the potential of CSP activities to address these deficits, we expect DMNCs to adjust their CSP activities with this aim in mind. As Gardberg and Fombrun (Reference Gardberg and Fombrun2006) have argued, firms operating across many countries can use positive CSP to increase their legitimacy. Gardberg and Fombrun (Reference Gardberg and Fombrun2006) use the term “corporate citizenship”, which they see as part of CSP. We maintain that their arguments apply to CSP as well. CSP activities can help a firm’s integration into a particular society (Etzioni, Reference Etzioni2000) by creating public trust in the business (Harris, Moriarity, & Wicks, Reference Harris, Moriarty and Wicks2014; Fukuyama, Reference Fukuyama2000). This sociocognitive integration can be of particular help to firms operating across different countries, enabling them to overcome legitimacy deficits arising from nationalism associated with their “liability of foreignness” (Zaheer, Reference Zaheer1995) and potential information asymmetries they might face (Siegel & Vitaliano, Reference Siegel and Vitaliano2007; Darby & Karni, Reference Darby and Karni1973; Nelson, Reference Nelson1970).
Moreover, in terms of a “license to operate”, the “unwritten permit that society bestows on companies” (Thompson & Zakaria, Reference Thompson and Zakaria2004: 126), a number of researchers have argued that firms entering foreign markets often improve their CSP as a way to gain sufficient legitimacy to prompt various foreign stakeholders to “allow” them to operate in their new environments (Thompson & Zakaria, Reference Thompson and Zakaria2004; Gunningham et al., Reference Gunningham, Kagan and Thornton2004).
CSP can also help firms address their reputational deficits. Fombrun (Reference Fombrun1996) contended that positive CSP might help firms build reputational capital, whereas Turban and Greening (Reference Turban and Greening1997) and Pfau, Haigh, Sims and Wigley (Reference Pfau, Haigh, Sims and Wigley2008) found that CSP improvements have positive effects on corporate reputation. Brown and Dacin (Reference Brown and Dacin1997) have found that consumers see firms with enhanced CSP in a favorable light and Walter (Reference Walter, Baybars-Hawks and Samast2013) identified the conditions under which CSP tends to have a positive effect on corporate reputation. Additionally, Gardberg and Fombrun (Reference Gardberg and Fombrun2006) argued that in the same way as R&D and advertising expenditures can be considered investments in strategic differentiation, CSP can also be an investment in differentiation through a positive impact on corporate reputation (McWilliams & Siegel, Reference McWilliams and Siegel2000).
In sum, we expect that by enhancing their CSP, DMNCs will try to make up for their legitimacy deficits by obtaining a ‘license to operate’ in host markets and for their reputational deficits by showing that they are as good as local firms or MNCs by investing in enhancing their CSP. Footnote 1 Hypothesis 1 follows:
Hypothesis 1: DMNCs will demonstrate significantly more enhanced levels of CSP than their counterpart domestic (only) firms.
However, not all DMNCs experience legitimacy and reputational deficits in equal magnitude. We can expect the magnitude of the impact of these deficits to depend on the degree of the DMNC’s multinationality (Hitt et al., Reference Hitt, Hoskisson and Kim1997; Kotabe et al., Reference Kotabe, Srinivasan and Aulakh2002). In other words, the greater the multinationality of a DMNC, the greater the negative consequences of likely reputational and legitimacy deficits will be and the more the firm will try to ameliorate them by enhancing its CSP. This is for three reasons.
First, a firm is increasingly exposed to different attitudes and standards of what is right or wrong as its multinationality increases (Donaldson, Reference Donaldson1996) and is more likely to suffer from a moral legitimacy deficit arising because its actions are not perceived to be “the right thing to do” (Suchman, Reference Suchman1995), in a particular national context. We can, therefore, expect DMNCs to enhance their CSP to match the level required by the most demanding of all the national environments in which they operate. Thus, the firm’s CSP is enhanced as its multinationality increases because the likelihood of entering a country with legitimacy demands that are higher than those it had grown used to also rises as it extends its operations internationally.
Second, the greater the number of countries in which a DMNC operates, the greater its risk of suffering a negative “international corporate reputation side-effect” (Zyglidopoulos, Reference Zyglidopoulos2002: 146). Given the increased cultural diversity a firm encounters as a result of increasing the number of countries in which it operates, there is a greater risk that its activities in one country will create a negative spill-over on its reputation in other countries, possibly even harming its global reputation. In order to mitigate some of this risk, all multinationals, including DMNCs with their greater reputational deficits, might choose to maintain more enhanced CSP levels than local firms.
Third, the increasing importance of non-governmental organizations (NGOs) in shaping the CSP environment for MNCs (Doh & Guay, Reference Doh and Guay2006; Teegen, Doh, & Vachani, Reference Teegen, Doh and Vachani2004) means that the likelihood of a firm becoming the target of an NGO campaign increases as it expands internationally (Ettenson & Gabrielle Klein, Reference Ettenson and Gabrielle Klein2005). NGOs, as “organizational manifestations of broader social movements” (Teegen et al., Reference Teegen, Doh and Vachani2004: 463), tend to target larger, more visible organizations (Rehbein, Waddock, & Graves, Reference Rehbein, Waddock and Graves2004). This is either due to the limited resources of NGOs that constrain them to only being able to detect the wrongdoings of larger, more visible corporations or that NGOs try to build their identity by positioning themselves against larger, more visible corporations (Rowley & Moldoveanu, Reference Rowley and Moldoveanu2003). Either way, as DMNCs expand internationally and their visibility increases, we can expect that they will be subject to increasing levels of scrutiny from NGOs, thus increasing the likelihood of becoming campaign targets (Rehbein et al., Reference Rehbein, Waddock and Graves2004; Ettenson & Gabrielle Klein, Reference Ettenson and Gabrielle Klein2005). A good illustration of this point is the way that Jeff Ballinger chose Nike as a target for his famous campaign against the conditions of workers in developing countries. Footnote 2 In 2000, at the International Association for Business and Society (IABS) in Vermont, in a debate between Jeff Ballinger and representatives from the CSR department of Nike, Ballinger noted that, at the time (mid 1990s), Nike was no worse than other firms in the sports goods industry, but that he chose Nike because of its size and visibility. By choosing to focus on Nike, he said, he was able to influence more firms in the process (Wokutch, Reference Wokutch2001). In an effort to protect themselves from becoming NGO targets, we can therefore expect DMNCs to engage more intensively in enhancing their CSP, the more multinational and visible they become (as found by Zyglidopoulos, Georgiadis, Carroll, & Siegel, Reference Zyglidopoulos, Georgiadis, Carroll and Siegel2012). Hypothesis 2 follows:
Hypothesis 2: The greater the multinationality of DMNCs, the better their CSP will be.
As the level of multinationality of a DMNC increases, however, we can expect the incremental impact of entering an additional country on the level of CSP activities to reach a limit and then decline. This is for two reasons. First, once a DMNC reaches a high degree of multinationality, it is likely to have entered a market with relatively high legitimacy demands and will have already enhanced its CSP substantially to try to improve its legitimacy to levels necessary to satisfy stakeholders in this demanding market. It may still need to undertake additional local CSP activities in the new market it enters to bolster its legitimacy there, but it will not be necessary to enhance its overall corporate CSP activities. Second, as the degree of a DMNC’s multinationality increases, the international corporate reputation side-effects (Zyglidopoulos, Reference Zyglidopoulos2002) will begin to operate in the opposite direction to that outlined above, creating positive reputational spillovers. We can expect then that a DMNC’s legitimacy and reputation deficits on entry to decline as multinationality increases, leading to a decreased need for enhanced CSP activities. Hypothesis 3 follows:
Hypothesis 3: The incremental impact of greater multinationality on CSP will decline towards a limit.
These relationships between multinationality and CSP are also likely to be impacted by the level of development of the countries a DMNC chooses to enter. If a DMNC internationalizes primarily by entering other developing-country markets, then the magnitude of its legitimacy deficit is likely to be low because stakeholder expectations are likely to be congruent with those it has already addressed in its home country (Cuervo-Cazurra & Genc, Reference Cuervo-Cazurra and Genc2008). In contrast, if a DMNC enters developed-country markets, even if CSP expectations vary, it is likely to encounter much higher stakeholder expectations and possibly higher levels of prejudice against its country of origin, leading to greater legitimacy and reputation deficits (Gaur, Kumar, & Sarathy, Reference Gaur, Kumar and Sarathy2011). Therefore, when a DMNC enters developed-country markets, it will face more intense pressures to address the negative consequences of these deficits by means of enhancing its CSP activity in order to succeed there, as compared to entry into other developing markets. Hypothesis 4 follows:
Hypothesis 4: DMNCs that operate relatively more in developed countries (than in developing countries) will tend to have more enhanced CSP levels.
METHODS
Sample and Data Collection
We draw on the Thomson Reuters’ ASSET4 database to obtain corporate financial and social performance data. Thomson Reuters’ ASSET4 database is an established source of environmental, social and governance (ESG) information used for empirical research on CSP (Cheng et al., Reference Cheng, Ioannou and Serafeim2014; Ioannou & Serafeim, Reference Ioannou and Serafeim2012). ASSET4 relies on a team of 125 specially trained analysts to collect data on over 500 separate data points from multiple sources, including company reports, filings and websites, NGO websites, CSR reports, and established and reputable media outlets, for a sample of over 3,500 globally traded public firms. These data points roll up into 226 key performance indicators (KPIs) that form the basis for the rating process of firms’ three ESG performance pillars (Environmental Performance, Social Performance and Corporate Governance Performance). ASSET4 then transforms these data into objective and meaningful ratings through a weighting system that assigns weights to each KPI following several industry or regional considerations. To produce a firm’s rating for each pillar, ASSET4 first sums the products of each KPI and its weight for each pillar. These scores are then normalized, adjusted for skewness and fitted to a bell curve to derive ratings between 0 and 100.
We selected the whole universe of companies reported for five developing countries for the four year period between 2009 and 2012: Brazil (83 firms), Russia (32 firms), India (81 firms), China (83 firms) and South Africa (135 firms). We chose these particular developing countries as ASSET4 observes only very few firms from other developing countries. Furthermore, the data on these other countries are subject to missing values for variables pertinent to the current study. According to data from the World Bank’s World Indicators data for 2012, our sample countries represent 42% of global population and 15% of global Gross National Income in constant 2005 US$ (15.2% Gross Domestic Income in constant 2005 US$). Our initial sample consists of an unbalanced panel of 412 companies observed over the course of four years.
Measures and Estimation Methods
Dependent variable: We operationalize CSP using ASSET4’s social pillar. According to Thomson Reuters, the social pillar measures a company’s capacity to generate trust and loyalty amongst its workforce, customers and society through its use of management best practices. ASSET4 generates CSP scores by examining seven factors including employment quality, health and safety issues, training, diversity, human rights, community involvement and product responsibility. The measure is thus a reflection of the company’s reputation and the health of its ‘license to operate’, which are key factors determining its ability to generate long-term shareholder value. The variable takes values between 0 and 100, with higher values reflecting better, positive CSP levels.
Independent variables: We measured firms’ multinationality (‘Hosts’) by the number of countries in which firms had foreign subsidiaries in a given year based on information we collected from the sample firms’ annual reports and publicly available documents. Drawing on this information, we developed a binary variable (‘Foreign’) that denotes whether a firm is a DMNC (i.e. has at least one foreign subsidiary) or not for testing H1. We used ‘Hosts’ to test H2 and also created the squared term of ‘Hosts’ (‘Hosts_squared’) to examine non-linearity in the relationship between ‘Hosts’ and CSP (H3). Lastly, we classified the DMNCs’ host countries into developed and developing countries Footnote 3 to calculate DMNCs’ ratio of developed to developing host countries (‘Ratio’) for testing H4.
Control variables: To account for possible confounding factors, we control for a number of firm-level and country-level effects. At the firm level, we control for firm size (‘Size’) by including the number of employees of the firm as a deviation from its industry’s mean. Given that the existence of a possible relationship between CSP and financial performance has been debated for more than 60 years (Ullmann, Reference Ullmann1985; Waddock & Graves, Reference Waddock and Graves1997; Margolis & Walsh, Reference Margolis and Walsh2003; Orlitzky, Schmidt, & Rynes, Reference Orlitzky, Schmidt and Rynes2003; Bowman, Reference Bowman1975), we also control for financial performance using Return on Assets (‘ROA’).
We also control for industry membership (‘Industry’) with the expectation that firms that belong to the same industry will be more similar in terms of CSP than those in different industries. We used data from the Industry Classification Benchmark (ICB), maintained by FTSE International Limited, to classify sample firms by industry. ICB is a 4-digit classification framework designed to facilitate the comparison of companies across four levels of classification (Industry, Supersector, Sector and Subsector) and national boundaries. We chose the Supersector Level (2-digit) to categorize our sample firms. At the country level, we used data obtained from the World Bank’s World Development Indicators and the Heritage Foundation’s databases to account for a country’s economic, institutional and technological conditions and other demographic characteristics that delineate a firm’s external environment that might influence its propensity and capability to engage in social activities. We used gross domestic product per capita (‘GDP’) to account for a country’s economic conditions and the number of patent applications as a deviation from the sample countries’ mean to account for technological environment (‘Tech’).
Based on data obtained from the World Bank, we considered controlling for other factors that were likely to be important, such as the quality of the institutional environment (a construct that captures multiple institutional attributes including the control of corruption, government effectiveness and regulatory quality, inter alia); the access of local population to the internet; the size of the population; and the country’s degree of economic freedom. However, all these variables correlated strongly with GDP per capita (statistically significant correlation coefficients above 0.7), so we subsequently excluded them from the analysis. The high correlations between the sample countries’ measures are no surprise because the sub-sample of developing countries in our available dataset exhibits similar levels of development in different areas (Hoskisson, Wright, & Filatotchev, Reference Hoskisson, Wright, Filatotchev and Peng2013). This justifies our use of GDP per capita as a suitable, all-encompassing measure of development. Lastly, we create country dummies to indicate the country of origin and, due to the longitudinal nature of our dataset we use annual dummies to account for unobserved effects.
FINDINGS
In Tables 1 and 2, we report the descriptive statistics and bivariate correlations of the variables. In our sample, there are 219 DMNCs (out of 412 firms) that had at least one foreign subsidiary. The correlations matrix suggests that firms with a foreign presence exhibit enhanced CSP levels (higher scores).
Table 1: Summary statistics
+ absolute number
Table 2: Pairwise correlations matrix
* Statistically significant at 0.05
Analysis
Our data are clustered into three hierarchical levels: the firm level, the industry level, and the country level. Therefore, we model our data using a hierarchical linear model (HLM) in which we allow for dependence among the responses observed for units belonging to the same cluster (Rabe-Hesketh & Skrondal, Reference Rabe-Hesketh and Skrondal2012). By analyzing our data using HLM, we decompose the total variance in CSP responses into between-cluster variance and within-cluster variance. The HLM estimator treats the cluster variables (country, industry and firm clusters) as random variables rather than model parameters and gives estimates of cluster-specific (random) intercepts. According to Rabe-Hesketh and Skrondal (Reference Rabe-Hesketh and Skrondal2012), when the data exhibit hierarchical structures, the HLM estimator is more efficient than fixed-effects or random-effects estimators because it exploits both between- and within-cluster information. Running the likelihood-ratio test proposed by Wooldridge (Reference Wooldridge2010) also suggested the presence of between-cluster variance, supporting our decision to use HLM with our data.
Several missing values for CSP, Industry and Size reduced the sample size to 387 firms (1313 observations). To determine whether or not the loss of sample firms was random, we tested for potential differences in the means of CSP, Size, FSTS, FATA and Hosts between firms omitted from the analysis and firms that were not. The analysis of means did not reveal any statistically significant differences across the two groups of firms, which suggests that our analysis does not suffer from a selection bias.
Table 2 shows very small correlation coefficients between the independent and control variables used in the analysis. The small correlations are also manifested in the small variance inflation factors (VIFs) we obtained when testing for the presence of serious collinearity. The VIFs’ mean value is 2.21 and the highest VIF is 4.30, substantially lower than the widely accepted threshold of 10 and very close to the very strict threshold of 2.5 suggested by Allison (Reference Allison2012). Therefore, multicollinearity is not a concern.
Results
Our main independent variables and controls enter the regression models with one-year lags. This allows us to make a—preliminary—causal interpretation of our findings, but it consumes one year’s observations. Table 3 shows the results of our analysis.
Table 3: Regressions results – dependent variable: corporate social performance
† p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
Base control year: 2010.
Our baseline, Model 0, includes all the firms (387) from developing countries in our sample and only the controls as independent variables. As expected, firm size is positively and significantly related to CSP. At the country level, higher levels of national development appear to be weakly conducive to improved firm CSP, whereas technological development appears to have a weak negative effect. Lastly, our year controls suggest that over time, the average CSP of sample firms has not improved compared to the base year (2010).
Model 1, which also includes all developing country firms, tests H1 that proposes that DMNCs have significantly more enhanced levels of CSP than domestic (only) firms by adding Foreign to the controls model. Foreign is statistically significant (p < 0.001; b = 13.86), suggesting that on a scale of 0-100, DMNCs have on average 13.86 units of CSP more than domestic-only firms, thus supporting H1. This translates to a predicted improvement in the CSP of DMNCs of 29% compared to that of domestic-only firms with a sample mean CSP of 47.6 units.
Model 2 tests H2, which proposes that DMNCs with a presence in relatively more foreign markets should experience improved levels of CSP. From this model and on, our sample size is reduced to 214 firms Footnote 4 and 515 observations, as we only include information on DMNCs. The effect of Hosts, a measure of the number of foreign markets in which a DMNC operates, is positive and statistically significant (p < 0.01; b = 0.32), supporting H2. If we consider the large magnitude of the coefficient of Foreign (Model 1), the small size of the effect of Hosts (Model 2) and that the preponderance of sample firms operate in relatively few foreign markets, we can infer that it is the initial forays abroad that provide the greatest encouragement to DMNCs to improve their CSP compared with later stages of international expansion. This is in accordance with the projections of H3, which we then further tested by including both the levels and squared term of Hosts in Model 3. The two variables were centered around their mean to avoid creating a multicollinearity issue (Belsley, Reference Belsley1984). In this model, the coefficient for Hosts remains positive and statistically significant (p < 0.01; b = 1.04), whereas the coefficient for its squared term is negative and statistically significant (p < 0.05; b = -0.02). Combining these two results suggests that Hosts is non-linearly related to CSP, which lends additional support to H3. The estimated relationship is illustrated in Graphs 1a (which shows how the predicted values for CSP in our sample firms vary as the number of host countries in which they operate increases) and Graph 1b, showing that the positive marginal effect of Hosts on CSP starts to decline from the second foreign market entry by a DMNC.
Graph 1: (a) and (b) The non-linear effect of Hosts on CSP.
Finally, Model 4 tests H4 by including Ratio in the model (the DMNC’s ratio of developed to developing host countries). The coefficient for Hosts remains positive and marginally statistically significant (p < 0.1; b = 0.23) and the coefficient for Ratio is also positive and statistically significant (p < 0.05; b = 3.12). These results lend support to H4, suggesting that both the number of host countries in which a DMNC operates and the balance between developed and developing host countries matter in determining its CSP. Our test for the interaction between these effects was statistically insignificant, suggesting that the two effects are independent.
Robustness Checks
We conducted numerous tests to examine the robustness of our analysis. First, to test the sensitivity of our results against alternative dimensions of a DMNC’s social activities, we re-estimated our models using Corporate Environmental Performance (CEP) as the dependent variable, measured using the environmental pillar obtained from Thomson Reuters’ ASSET4 database (reflecting a company’s impact on living and non-living natural systems and ecosystems). The results of this analysis are shown in Table 4. The results for CEP are qualitatively similar to those derived from the main analysis.
Table 4: Regressions results – dependent variable: corporate environmental performance
† p < 0.1, * p < 0.05, ** p < 0.01, *** p < 0.001
Base control year: 2010.
Our second robustness check sought to address the fact that our measure of multinationality, Foreign, required that a DMNC have at least one foreign subsidiary, but that firms without any foreign subsidiaries may still engage in international activities by exporting or through joint ventures. Clearly, it is possible that firms may interact with foreign stakeholders through these activities, just as DMNCs do and that these interactions might also present these firms with greater reputational and legitimacy deficits, compared to their domestic-only counterparts. To account for this possibility, we created a binary variable that classified firms engaged in other kinds of international activity (excluding those with at least one foreign subsidiary) as separate from purely domestic firms. We drew on ASSET4 data to measure Foreign Sales to Total Sales (FSTS), as well as firms’ Foreign Assets to Total Assets (FATA). We gave a value of 1 to firms that had zero foreign subsidiaries and reported positive FSTS and/or FATA greater than 5%. We then re-estimated Model 1 by adding this new variable as well as Foreign. Our results remained similar to those obtained in the main analysis while the new variable was positive, but statistically insignificant (p < 0.3; b = 1.91). This suggests that while developing country firms may interact with foreign stakeholders through foreign sales, only interactions that entail a physical foreign presence systematically induce firms to improve their CSP.
Our third robustness check was to test for Granger causality (Granger, Reference Granger1969), i.e. whether the relationships between multinationality attributes and CSP/CEP are strictly unidirectional. We conducted this additional test even though our independent variables entered our main model with one-year lags. This was implemented by separately regressing all our main independent variables (i.e. Foreign, Hosts, Ratio), FSTS and FATA on their lagged values, the lagged values of CSP/CEP and all control variables to test whether lagged CSP/CEP drives multinationality. In all models, lagged CSP/CEP was statistically insignificant, suggesting a strictly unidirectional relationship between multinationality and CSP/CEP.
Lastly, we replaced ROA with return on equity (ROE) in the analysis. There were no major changes in the results. We also replaced the annual dummies with a time trend to predict the overall direction of CSP/CEP within the sample period. In all models for CSP, the time-trend was positive and statistically insignificant, whereas for CEP the time-trend was positive and statistically significant, consistent in both cases with the predictions associated with the annual dummies. The robustness of our empirical results to these sensitivity tests and alternative modeling gives us confidence in the empirical support we found for of all our hypotheses.
DISCUSSION AND CONCLUSION
In this article, we explored why and how the extent and nature of multinationality of DMNCs leads to changes in their CSP compared to the CSP of their domestic-only counterparts. We found that in trying to address the reputational and legitimacy deficits they face in foreign markets, DMNCs display enhanced CSP levels compared to domestic-only firms in developing markets. Our findings indicate that the CSP of DMNCs is positively related to their multinationality, the incremental impact of which declines towards a limit and that the more DMNCs operate in developed countries, the greater the positive impact on CSP will be (as compared to the impact of internationalizing into other developing countries). These results contribute to the International Business (IB) and CSP literatures in a number of ways.
With regard to the IB literature, our findings have a number of implications. First, we augment existing IB studies of the internationalization of DMNCs (Nachum, Reference Nachum2010; Williamson et al., Reference Williamson, Ramamurti, Fleury and Fleury2013; Zeng & Williamson, Reference Zeng and Williamson2007; Guillén & García-Canal, Reference Guillén and García-Canal2009; Deng, Reference Deng2009) by showing that a significant impact of increased multinationality is that it leads DMNCs to improve their CSP. Our empirical analysis supports this position that enhanced CSP is likely to be a consequence rather than a cause of DMNC internationalization. This position is in line with existing literature on the internationalization of DMNCs, which suggests that DMNCs expand abroad to either exploit the competitive advantages they have developed at home (Cavusgil, Reference Cavusgil1980; Williamson et al., Reference Williamson, Ramamurti, Fleury and Fleury2013; Prahalad, Reference Prahalad2006) or engage in “strategic asset seeking” (Rui & Yip, Reference Rui and Yip2008; Deng, Reference Deng2009).
Second, our hypotheses and results shed light on how the impact of internationalization on CSP might vary with the motivation of and strategies adopted by DMNCs. The classic model of internationalization postulates a strategy of incremental expansion and learning overseas, starting with similar markets and gradually moving to markets with less familiar characteristics, often termed the ‘Uppsala model’ (Johanson & Vahlne, Reference Johanson and Vahlne1977; Petersen, Pedersen, & Lyles, Reference Petersen, Pedersen and Lyles2008; Johanson & Vahlne, Reference Johanson and Vahlne2009). More recent evidence on the behavior of DMNCs suggests that they might also (or alternatively) follow a strategy of asset-seeking in which the primary goal of internationalization is to access distinctive foreign resources to complement assets amassed at home (Ramamurti & Singh, Reference Ramamurti and Singh2009; Meyer & Thaijongrak, Reference Meyer and Thaijongrak2013; Williamson et al., Reference Williamson, Ramamurti, Fleury and Fleury2013).
To the extent that DMNCs seek to minimize the challenges of learning about unfamiliar markets, we might expect them to focus on entering into other developing markets. In this case, it might be argued that internationalization would have little impact on DMNCs’ CSP because their CSP would often already be superior to domestic competitors in the host countries they enter. We find, however, that DMNCs’ internationalization does lead them to improve their CSP, even when they enter other developing markets, although less so than when they enter developed markets. One explanation for these results is that even when DMNCs enter other developing markets, they are likely to meet competition from multinationals from developed economies. Relative to these developed-country multinationals, they can be expected to face legitimacy and reputational deficits that we postulate they try to ameliorate by improving their CSP.
Our results also shed light on the impact that the entry of DMNCs in developed markets has on their CSP (Rui & Yip, Reference Rui and Yip2008; Deng, Reference Deng2012). The rationale for this behavior is to allow DMNCs to access strategic assets with qualities that allow them to catch up with the capabilities enjoyed by more experienced and better-resourced multinational competitors in global markets. Luo and Tung (Reference Luo and Tung2007) argue that this type of international expansion acts as a “springboard” upon which to acquire assets that may compensate for DMNCs’ competitive disadvantages, overcome late-comer disadvantage, counter-attack global rivals in their home market, by-pass trade barriers, secure preferential treatment for companies that go abroad offered by their home governments and access complementary assets to exploit their competitive advantages overseas. To the extent that they pursue this strategy, DMNCs will need to operate in developed markets from the earliest stages of their internationalization. Our results suggest that this will lead DMNCs to systematically increase their investments in CSP as they expand internationally more than is predicted by the Uppsala model.
Third, we contribute to the literature on the liability of foreignness (Zaheer, Reference Zaheer1995; Bhanji & Oxley, Reference Bhanji and Oxley2013; Gaur et al., Reference Gaur, Kumar and Sarathy2011) by exploring how DMNCs try to address this issue. The existing literature suggests that in going abroad, DMNCs might try to overcome their liability of foreignness by drawing on their home government for support (Gaur et al., Reference Gaur, Kumar and Sarathy2011), relying on their ownership-specific advantages or by increasing their isomorphism vis-à-vis local firms (Zaheer, Reference Zaheer1995). Our findings suggest that enhancements in CSP is a significant way in which DMNCs try to overcome their liability of foreignness, especially when they enter developed markets.
We also contribute to the CSP literature in three ways. First, we extend the existing work investigating the drivers of CSP in developing countries. Extant literature maintains that the CSP of developing country firms is driven either by the pressures local firms face as a result of their roles as suppliers of global MNCs or because MNCs investing in developing countries respond to pressures from their home (developed country) stakeholders to enhance their CSP world-wide (Gugler & Shi, Reference Gugler and Shi2009; Visser, Reference Visser, Crane, Matten, McWilliams, Moon and Siegel2008). We explain the role that internationalization plays in positively influencing the CSP of DMNCs and how this impact varies according the level of development of the countries DMNCs enter.
Second, we show that interactions with foreign stakeholders as a result of a foreign presence seem to be critical to stimulating CSP. Our findings suggest that developing country firms that sell abroad could also face reputational and legitimacy deficits that they may try to compensate for by making greater investments in CSP activities. However, even though the impact of a firm’s foreign activities (excluding the operation of foreign subsidiaries) on CSP was found to be positive, it was statistically insignificant. This result extends the single-country finding of Barin Cruz, Boehe and Ogasavara (Reference Barin Cruz, Boehe and Ogasavara2015) who found that the export intensity of Brazilian firms did not influence their engagement in CSP at the firm level of analysis. By testing this relationship using longitudinal data from five developing countries, our findings underscore the importance of a physical presence abroad in the form of foreign subsidiaries in stimulating DMNCs to significantly enhance their CSP.
Third, our findings suggest that CSP may create a cumulative and transferable strategic asset. We find that once DMNCs improve their CSP to meet the standards of one host country, the benefits of further incremental CSP improvements appear to gradually diminish as they enter additional countries (as confirmed by our findings in respect of H3 and H4). This may be because once the DMNC has achieved a competitive level of CSP in a host country (possibly through corporate initiatives), that level of CSP could create spillovers that influence the perception of stakeholders elsewhere in the world. Alternatively it may be that the DMNC follows a strategy of standardizing their best practices across the world (rather than following a multi-domestic strategy), behavior observed by Dowell, Hart and Yeung (Reference Dowell, Hart and Yeung2000) in a study of environmental policies. In either case, a corollary of this finding is that if a DMNC chooses to enter developed countries with high expectations for CSP at an early stage in its internationalization, the incentive to make further improvements in CSP will be reduced as diminishing returns set in. These results are consistent with the idea that greater commitment to CSP creates a cumulative strategic asset that is transferable across countries a DMNC might subsequently enter. It is an asset that is especially valuable in light of the variance in expectations regarding CSP across different countries.
Our research has also implications for DMNC managers who are trying to expand the operations of their firms internationally. Interviews with senior management of DMNCs conducted by the authors in connection with other research indicate that DMNCs are fully cognizant of the reputational and legitimacy deficits they are likely to face when venturing abroad. The results we present in this article suggest that a significant proportion of DMNC managers believe that CSP initiatives can provide a means of addressing these deficits. At minimum, managers of DMNCs need to take this into account when deciding on the level of their own investments in CSP in order to avoid potentially losing competitiveness to rival DMNCs using CSP investment to this end.
Moreover, our findings suggest that the likely need for DMNCs to increase investment in CSP when they venture abroad is significantly greater when they expand into developed markets than when they expand into developing markets. On the other hand, our results also suggest that DMNCs will reach a level of CSP at which point diminishing returns to further investment set in. Narrowing our focus to the role of higher CSP in neutralizing the reputational and legitimacy deficits DMNCs face when entering foreign markets, our results imply that the management of DMNCs needs to assess the point at which the incremental impact of additional investment in CSP reaches its limit. Of course, there may be other reasons for DMNCs to continue to increase investment in CSP (i.e. CSP may provide benefits to their home countries by increasing differentiation (McWilliams & Siegel, Reference McWilliams and Siegel2000)).
These results also suggest that increased investment in CSP should not be seen purely in terms of its impact on legitimacy and reputation in a particular, isolated host market. Rather, because our findings suggest that improving CSP creates a transferable strategic asset, the potential returns on investment in CSP need to be evaluated more broadly, taking into account a greater ‘halo effect’ that may help DMNCs overcome legitimacy and reputational deficits in other markets they may enter.
In spite of these contributions, our work presented here has a number of limitations that future research could address. First, our study is limited by the fact that we only investigate firms from five specific developing countries. While we do not expect firms from these countries to be unrepresentative of the population of developing country firms, we should add the caveat that this sample is biased towards larger economies and countries that are more advanced than many less-developed ones. It may be the case that firms from smaller or less-developed economies tend to pursue somewhat different internationalization strategies that, in turn, create different incentives for these firms regarding CSP. Moreover, in spite of our best efforts to collect all available data, we also could not account for our sample firms’ type of international engagement. We could not, therefore, account for the type of subsidiary (i.e. sales, manufacturing), the kind of foreign entry involved (i.e. wholly owned subsidiary, joint venture), or the kind of strategy they followed (i.e. low cost, differentiation). Nor were we able to assemble data on the sequence in which each firm entered the different countries in which it operates. Future research could expand on our work by not only including firms from more developing countries, but also by conducting a more fine-grained investigation of the impact of the type of internationalization and strategy on changes in CSP.
Second, in this article, we do not explore the role of institutional distance, the similarity or dissimilarity between the institutions of home and host countries on the drivers of CSP (Kostova, Reference Kostova1999; Kostova & Zaheer, Reference Kostova and Zaheer1999). Neither do we investigate the impact of potentially conflicting CSP expectations and institutional demands across host countries. We did not have the data or the space to adequately explore these issues. We hope that future researchers will be able to build on our research here and explore these potentially important influences on the CSP of DMNCs.
Third, we do not address the role that different regulatory environments have on the CSP of DMNCs, a limitation to be addressed by future research. Even though some aspects of CSP are regulated, we do not expect this limitation to profoundly impact our findings because most aspects are not regulated, as indicated by the existence of substantial variation in the levels of CSP between different companies.
Fourth, while we investigate the social performance of DMNCs by examining their CSP and CEP (captured by the social and environmental performance pillars of ASSET4, respectively), we do not engage in a more fine-grained investigation of any of the constituent dimensions to both CSP and CEP, nor do we directly measure the reputation and legitimacy deficits that we discuss. ASSET4 captures a firm’s CSP by incorporating four such constituent dimensions (employment quality, health and safety, training and development, diversity, human rights, community and product responsibility), whereas the same database captures CEP by incorporating three such constituent dimensions (resource reduction, emission reduction and product innovation). Drawing on these dimensions, future research could attempt a more fine-grained investigation of the differences between the social performance of DMNCs and local-only developing country firms. Such an investigation could provide us with a better understanding of the priorities that DMNCs have compared to other firms and possibly even reveal the limitations of investing in different kinds of CSP. Moreover, by finding ways to directly measure the reputation and legitimacy deficits that DMNCs face in entering foreign markets, future research can evaluate the direct impact that CSP can have in reducing such deficits.
A final limitation of our research is that we lack the data that would allow us to distinguish between the CSP levels of DMNCs in their home versus host countries, nor do we directly compare the CSP of DMNCs with MNCs. However, given the interconnectedness of our globalized environment, we assume that if DMNCs improve their CSP anywhere in the world, this will have an impact on their reputation and legitimacy everywhere they operate (a reasonable assumption in light of our finding that multinationality has a diminishing positive effect on DMNCs’ CSP). Nonetheless, we might expect that CSP activities of DMNCs in their host markets might have a greater impact than CSP activities in their home markets. Moreover, directly comparing the CSP of DMNCs with that of MNCs could prove to be a potentially fruitful avenue for future research.
ACKNOWLEDGEMENTS
We would like to thank Bryan Husted and three anonymous reviewers for their valuable comments during the revision process.
