1
Master Thesis
Gaining Legitimacy through CSR:
The Effect of Internationalisation on the CSR Performance
of Emerging Market Multinationals
Author: Martika A. Küpper Student number: S3224430 Email: m.kuepper@student.rug.nl
Supervisor: dr. S. R. Gubbi Co-assessor: dr. H. U. Haq
MSc International Business and Management Faculty of Economics and Business
University of Groningen
Duisenberg Building, Nettelbosje 2, 9747 AE Groningen, The Netherlands P.O. Box 800, 9700 AV Groningen, The Netherlands
http://www.rug.nl/feb
2
Abstract
Prior research finds that many emerging market multinationals (EMNEs) increase the intensity of their corporate social responsibility (CSR) reporting as tool to gain legitimacy. However, less is known about the effect of internationalisation on their home country CSR performance. Building on neo-institutional theory, this paper investigates how internationalisation influences EMNEs overall CSR performance. We argue that with increasing internationalisation, EMNEs face greater pressure from their stakeholders to act socially responsible. Thus, to gain legitimacy and overcome liabilities of origin, EMNEs invest more in their CSR activities, and ultimately improve their overall CSR performance. In addition, we suggest that this effect is more pronounced when they expand into developed countries and with growing international market experience. To test our hypotheses, we conduct a longitudinal study covering 71 EMNEs from 13 different countries. We find that internationalisation positively influences EMNEs’ overall CSR performance when they focus on foreign consumer markets. In addition, our findings suggest that international market experience and expansion into developed countries provide learning opportunities that allow them to implement stronger CSR strategies at home.
Key Words: Emerging market multinationals, corporate social responsibility, institutional
theory, legitimacy, liability of origin
3
Acknowledgement
4 Table of Contents Abstract ... 2 Acknowledgement ... 3 List of Tables ... 5 List of Figures ... 5 List of Abbreviations ... 6 1. Introduction ... 7 2. Literature Review ... 9
2.1 Corporate social responsibility ... 9
2.2 CSR in emerging markets ... 11
2.3 Internationalisation, legitimacy and CSR ... 13
3. Methodology ... 18
3.1 Sample and data ... 18
3.2 Variables ... 19
3.3 Data Analysis ... 22
4. Results ... 23
4.1 Descriptive statistics and correlation ... 23
4.2 Regression results ... 26
4.3 Robustness test ... 29
5. Discussion ... 31
Contributions and implications ... 35
Limitations and future research ... 38
6. Conclusion ... 39
5 Appendix 4: Variance inflation factor ... II Appendix 5: Breusch-Pagan / Cook-Weisberg test for heteroskedasticity ... II Appendix 6: Wooldridge test for autocorrelation in panel data ... II Appendix 7: Robustness test 1 ... III Appendix 8: Robustness test 2 ... IV Appendix 9: Robustness test 3 ... V Appendix 10:Robustness test 4 ... VI Appendix 11: Robustness test 5 ... VII Appendix 12: Robustness test 6 ... VIII
List of Tables
Table 1: Sample description ... 19
Table 2: Description of variables used ... 21
Table 3:Descriptive statistics ... 23
Table 4: Correlation matrix ... 25
Table 5: Random effects regression on CSR performance ... 27
Table 6: Results Regression Analysis ... 29
6
List of Abbreviations
BG Business group
BOP Bottom of the pyramid
CSR Corporate social responsibility EMs Emerging markets
EMNEs Emerging market multinationals ESG Environmental, social, governance FDI Foreign direct investment
FE Fixed effects
FS/TS Foreign sales to total sales FA/TA Foreign assets to total assets IB International business LOF Liability of foreignness MNEs Multinational enterprises
NGOs Non-governmental organisations n.s. not significant
OLS Ordinary least square R&D Research and development
RE Random effects
ROA Return on assets
SOE State-owned enterprise TNCs Transnational corporations
UNCTAD United Nations Conference on Trade and Development VIF Variance inflation factor
7
1. Introduction
Recently, emerging markets (EMs) have undergone a process of liberalisation and rapid economic growth. This has, in turn, facilitated the internationalisation of many of their home country firms (Giuliani, Nieri, & Fiaschi, 2014). By 2025, it is estimated that emerging market multinationals (EMNEs) will account for 45% of Fortune 500 companies (KPMG, 2013). This growing importance of EMNEs in the global economy has also led to more scrutiny over these firms’ corporate social responsibility (CSR)1. Increasingly, factors such as environmental sustainability and social responsibility influence consumer purchasing decisions (Borin & Metcalf, 2010) and are becoming important investment factors (Scholtens & Sievanen, 2013). Thus, EMNEs face heightened pressure from both international and home country stakeholders to improve their CSR performance (Davies, 2000; Marano, Tashman, & Kostova, 2017).
Prior research argues that the weak institutions in EMs hinder the development of a strong CSR framework (Jamali & Mirshak, 2007, Nwabuzor, 2005). Moreover, economic and institutional factors provide little support for and exert limited pressure on firms to improve their CSR performance. Consequently, EMNEs are often less incentivised to invest in CSR activities and develop a well-established CSR strategy (Baughn, Bodie, & McIntosh, 2007; Miska, Witt, & Stahl, 2016; Nwabuzor, 2005). In this way, while many EMNEs engage in philanthropical CSR activities, such as charitable giving and donations, their CSR strategy remains less integrated into their overall business operations (Jamali & Mirshak, 2007).
However, when internationalising, EMNEs are exposed to a wider set of stakeholders and come under increased public scrutiny. This scrutiny ultimately puts greater pressure on firms to enhance their CSR performance and develop a well-established CSR strategy (Strike, Gao, & Bansal, 2006). At the same time, because of the institutional voids in EMs, many global stakeholders hold negative stereotypes towards EMNEs and question their credibility. Thus, when expanding abroad, EMNEs face liabilities of origin that impose additional challenges on them (Pant & Ramachandran, 2012). To gain legitimacy among stakeholders and meet different stakeholder demands, EMNEs increasingly engage in global legitimisation strategies and invest in CSR initiatives (Marano et al., 2017).
Despite the increasing importance of EMNEs in the global business field (Giuliani et al., 2014), few studies have analysed CSR in the context of EMNEs. Of the studies that did
1 In line with McWilliams and Siegel (2001), we define CSR as situations in which firms go beyond merely
8 analyse CSR in the context of EMNEs, most focus on the link between CSR and gaining legitimacy among global stakeholders (Fiaschi, Giuliani, & Nieri, 2017; Giuliani et al., 2014; Marano et al., 2017). For instance, both Marano et al. (2017) and Giuliani et al., (2014) find that with greater levels of internationalisation, EMNEs increasingly engage in CSR reporting to gain legitimacy among their global stakeholders. However, there remains limited understanding on whether and how internationalisation effects EMNEs’ CSR performance in their home country. In addition, we still lack understanding of how EMNEs’ experience in varied foreign markets affects their overall CSR performance. Thus, this study will extend prior research on CSR in EMNEs and investigate how internationalisation effects EMNEs’ overall CSR performance. Moreover, it seeks to answer how international market experience and developed country experience influences EMNEs’ CSR implementation at home. We apply neo-institutional theory and literature on legitimacy and CSR (Kostova & Zaheer, 1999; Marano et al., 2017; Pant & Ramachandran, 2012; Suchman, 1995) to answer:
How does internationalisation influence EMNEs’ overall CSR performance?
We suggest that with increasing internationalisation, EMNEs become more embedded in the meta-institutional field, and thus, are more exposed to global legitimating actors and global norms (Kostova & Zaheer, 1999). Moreover, with greater convergence among CSR standards (Chapple & Moon, 2005; Hafsi & Farashahi, 2005), home country stakeholders, such as consumers, start placing greater importance on CSR (Davies, 2000). Thus, although the formal institutions at home provide little incentives to engage in CSR, internationalisation and the convergence of CSR standards increases the pressure of EMNEs to invest in CSR and improve their CSR performance at home. This leads to EMNEs voluntarily adopting global CSR practices that go beyond the law in their home countries (Chapple & Moon, 2005; Muller, 2006; Muller & Kolk, 2009). Therefore, we expect that internationalisation positively influences EMNEs’ home country CSR performance.
9 To test our hypotheses, we conduct a longitudinal study covering 71 EMNEs from 13 different countries between 2010 and 2017. We find support that internationalisation positively impacts EMNEs’ home country CSR performance. In addition, we also find that international market experience and expansion into developed countries enhances EMNE’s CSR performance at home.
Our study contributes to international business (IB) research on EMNEs, legitimacy strategies and CSR. Existing literature has shown that when internationalising, EMNEs must overcome challenges stemming from their liability of origin (Pant & Ramachandran, 2012). To do so, they often engage in CSR reporting (Marano et al., 2017). We add to this literature by shedding light on the effect of internationalisation on EMNEs’ home country CSR performance. In addition, our paper also adds to the literature on EMNEs’ expansion strategies (Luo & Tung, 2007; 2018) by showing that when internationalising into different countries, EMNEs can access knowledge and gain experience to improve their CSR performance. These learning effects are particularly pronounced when EMNEs expand into more developed countries. However, our findings suggest that to benefit from CSR experience and best practices acquired abroad, EMNEs must develop additional capabilities that help transfer and integrate the acquired experience into their overall operations (Luo & Tung, 2018).
Finally, we also advance research on institutional heterogeneity and multi-embeddedness (Marano et al., 2017; Marano & Kostova, 2016) by showing that the experience gained through embeddedness in various institutional contexts can be a valuable source of CSR knowledge and lead to the adaption of higher CSR standards.
The rest of the paper is structured as followed: Section two develops the theoretical background for our hypotheses and our conceptual model. Section three describes our methodology, including our sample, variables, and data analysis. Section four presents the results and the findings of our regression analysis. Section five discusses our findings, contributions, limitations and future research directions. Finally, section six concludes this paper by highlighting the most important findings of this paper.
2. Literature Review
2.1 Corporate social responsibility
10 pressure on firms to act in a socially responsible manner, there is still no consensus on what constitutes CSR and how it should be defined (Dahlsrud, 2008). Consequently, there are numerous definitions of CSR (McWilliams & Siegel, 2006). These encompass several different dimensions that include a large variety of corporate behaviour and responsibilities (Carroll, 1979; Wood, 1991). Although there is no universal applicable definition for CSR, an analysis of over 30 definitions by Dahlsrud (2008) found that most definitions are largely congruent and generally encompass five dimensions. These are: environmental, social, economic, stakeholder, and voluntariness. These dimensions are interrelated, and firms should aim at realising all of them simultaneously (Carroll, 1979). Furthermore, the broad range of these dimensions suggests that firms have responsibilities beyond their economic interest. Instead of passively complying with the law, they should actively engage in ethical behaviour and activities that further social good (Carroll, 1979; McWilliams & Siegel, 2001).
Another core idea of CSR is that firms have obligations towards a wide set of stakeholders, and thus need to consider their expectations and demands (Clarkson, 1995). However, different stakeholder groups may hold different expectations towards firms (Fombrun & Shanley, 1990), and companies must decide which of these to prioritise (Carroll, 1991). Given the different dimensions of CSR and the varying expectations of stakeholders, achieving strong CSR performance can be challenging. Firms that want to demonstrate a high level of social responsiveness must engage in a wide range of different activities encompassing environmental, social, and philanthropical aspects (McWilliams & Siegel, 2006).
Finally, research suggests that CSR is context specific and shaped by the institutional environment in which a company operates (Campbell, 2007; Gerson, 2007). The institutional context shapes firm strategy and decision making, influencing the actions and practices available to organisations (Robin & Reidenbach, 1987; Suchman, 1995), as well as firm’s CSR practices (Campbell, 2007). A country’s environmental context (which includes economic, social, political and technological conditions) can foster or impede the development of CSR. In addition, a country’s cultural context shapes the way CSR is perceived (Gerson, 2007). CSR can thus vary across countries (Halkos & Skouloudis, 2016) and practices that work in one country may not be suitable in another (Heslin & Ochoa, 2008).
11 CSR as situations in which firms go beyond merely complying to the law and engage in “social actions that appear to further some social good beyond the interest of the firm and what is required by law” (McWilliams & Siegel, 2001). It includes the prioritisation and management of different needs and expectations of diverse stakeholder groups that matter to the firm.
2.2 CSR in emerging markets
Emerging markets (EMs) take an intermediate position between developed and developing countries. Even though EMs differ in their pace of change and level of development, they share similarities that differentiate them from developed markets (Marquis & Raynard, 2015).
Generally, EMs are characterised by their low income, rapid industrialisation and economic liberalisation (Hoskisson, Eden, Lau, & Wright, 2000). Moreover, they often suffer from institutional voids, which refer to the underdevelopment or absence of an effective institutional system (Khanna & Palepu, 1997). The weak institutions in EMs not only hamper the efficient functioning of markets (Khanna & Palepu, 1997) but also have implications for CSR. Research suggests that a country’s macro-institutional environment shapes firms’ CSR behaviour and practices (Gerson, 2007; Jamali & Mirshak, 2007). Because EMs have less favourable conditions, their focal firms often lag behind firms from more developed countries, and CSR activities are disconnected from their overall business operations (Jamali & Mirshak, 2007; Welford, 2004). For instance, Welford (2004) found that companies in Asia implement less policies concerning CSR issues, such as fair wages, overtime and equal opportunities for employees, than their European counterparts. Instead of focusing on CSR activities enhancing the welfare of their employees, who are often merely seen as a production factor, they are forced to address more pressing issues, such as corruption and bribery. A study by Jamali and Mirshak (2007) similarly found that despite having the awareness of and intention to engage in CSR, many Lebanese firms lack a systematic and structured CSR approach. Although most firms engage in philanthropical CSR activities, such as charitable giving or donations, they do not see CSR as an integrative part of their overall business strategy. Moreover, they do not systematically monitor and assess the impacts of their CSR activities. Thus, their CSR strategy and its implementation remains “amateurish and sketchy” (Jamali & Mirshak, 2007). The same study also found that the turbulent external and political environment often diverts the companies’ attention from their CSR strategy as they must address more pressing issues concerned with their economic viability.
12 implement CSR practices and achieve substantial CSR performance (Ioannou, & Serafeim, 2012; Marquis, Toffel, & Zhou, 2016). In EMs, laws that address corporate behaviour are less developed, and the mechanisms that enforce and investigate firms’ compliance to such laws are similarly lacking (Nwabuzor, 2005). In addition, firms are subject to weaker industry self-regulations and face lower normative pressure from media and other institutions (Miska et al., 2016). Thus, the formal institutions put less pressure on organisations to develop an organisational culture that prioritises and fosters CSR values and provide little incentives to improve firms’ CSR performance (Tashman, Marano, & Kostova, 2018).
Additionally, previous research suggests that limited freedom, corruption and the absence of free press can impede the development of a strong CSR framework (Baughn et al., 2007; Nwabuzor, 2005). For instance, Nwabuzor (2005) found that the ability of stakeholders, such as customers, employees or NGOs, to access and disseminate information about corporate (ir)responsible behaviour is often limited. This, in turn, hampers the chance for public engagement and lowers public pressure on firms to act in a more socially responsible manner (Baughn et al., 2007).
Furthermore, scholars argue that the implementation of CSR policies is linked to a country’s economic development (Husted, 2005; Welford, 2004). According to Husted (2005), economic development drives the institutional and social capacity needed to foster environmental sustainability. Higher development of wealth per capita allows stakeholders, such as consumers and employees, to make greater demands for CSR. Consequently, firms in these environments have more incentives to improve their CSR practice (Ramasamy & Hung, 2004). However, EMs are generally characterised by low income (Hoskissonet al., 2000), with a large part of consumers located at the bottom of the pyramid (BOP). These consumers have less than $2 of daily spending income. Thus, firms targeting these consumers must work with several constraints, such as affordability, creating awareness, and ensuring access and availability of products (Prahalad, 2012). Consequently, firms must first fulfil the basics needs of the consumers before focusing on CSR strategies. In addition, because CSR strategies need to be targeted to the specific needs of the BOP consumers, it is also more difficult to implement standardised CSR practices (Singh, Bakshi, & Mishra, 2015).
13 practices (Baughn et al., 2007; Chapple & Moon, 2005; Muller & Kolk, 2009). We elaborate on how this may translate into greater CSR awareness and activities in the next section.
2.3 Internationalisation, legitimacy and CSR
Increasingly, firms from emerging markets have expanded their operations in both developed and developing countries (Giuliani et al., 2014; Luo & Tung, 2007, 2018). To access more efficient markets and eschew the institutional voids of their home countries, many firms from EMs pursue accelerated internationalisation strategies (Luo & Tung, 2007, 2018). By expanding abroad, they can leverage economies of scale and scope (Kogut, 1985), gain access to or acquire strategic resources and knowledge (Hitt, Hoskisson & Kim, 1997; Luo & Tung, 2018), and tap into new growth opportunities (Porter, 1990). Thus, internationalisation can enhance an EMNE’s competitive position and increase its overall competitive advantage at home and abroad (Luo & Tung, 2018; Nachum & Zaheer, 2005).
However, when entering foreign markets, EMNEs face challenges arising from their multi-embeddedness in different contexts and networks (Goshal & Bartlett, 1990). First, like all foreign firms, EMNEs face disadvantages stemming from their liability of foreignness (LOF). LOF arises from unfamiliarity with foreign contexts and discrimination of host country stakeholders (Zaheer, 1995). Second, EMNEs are exposed to and must consider pressures from a wide set of stakeholders, holding possibly contradicting expectations towards them (Detomasi, 2007; Meyer, Mudambi, & Narula, 2011). This exposure to a wider range of stakeholders and the embeddedness in different institutional contexts also has implications for their legitimacy. According to institutional theory, firms must adhere to the ‘rules of the game’ (North, 1990) which are set by the institutional context to gain legitimacy (Robin & Reidenbach, 1987). This legitimacy, defined as the perception that a firm’s actions are desirable and in line with the social system of norms, values and beliefs (Suchman, 1995), is necessary to ensure firm survival and future growth (DiMaggio & Powell, 1983; Suchman, 1995). Compared to domestic firms, establishing legitimacy is generally more challenging for multinational enterprises (MNEs). First, MNEs are embedded in institutional heterogenous environments and exposed to pressures from a wide set of stakeholders. Thus, they must contend with more varied institutional contexts (Kostova, Roth, & Dacin, 2008; Kostova & Roth, 2002). Additionally, information asymmetries make it more difficult for both the foreign firm and its stakeholder to understand and evaluate each other (Kostova & Zaheer, 1999).
14 underdeveloped institutions and institutional voids (Khanna & Palepu, 1997), many stakeholders, such as customers, governments and NGOs, hold negative stereotypes towards EMNEs and question their credibility. Thus, as EMNEs start internationalising, they must not only overcome LOF, but must also eliminate liabilities of origin to establish legitimacy among their stakeholders (Pant & Ramachandran, 2012). In addition, EMNEs often lack the capabilities and resources that many MNEs already established (Guillén & García-Canal, 2009). They often operate in low-end segments (Ghemawat & Hout, 2008), offering products that do not meet the quality and safety standards in more developed markets (Khanna, Palepu, & Andrews, 2012). Thus, to successfully expand into foreign markets, EMNEs first need to establish legitimacy and enhance their brand reputation abroad. For instance, before the Chinese refrigerator company Haier was able to successfully sell its products under its own brand-name in more developed countries, it had to upgrade its product standards to meet these countries’ safety and quality demands. In addition, it had to invest into enhancing its brand reputation and overcome liabilities of origin (Khanna et al., 2012).
To overcome these challenges and improve their reputation, many EMNEs engage in activities that help them meet the expectations of their stakeholders (Kostova et al., 2008; Marano et al., 2017). One way they do this is through the adoption of global CSR practices (Marano et al., 2017). Because an increasing amount of stakeholders expect firms to act in a more socially responsible way, investing in and improving their CSR performance can enhance legitimacy, and foster trust and credibility (Maignan & Ferrell, 2004; Peloza & Shang, 2011). In addition, as heightened global visibility increases the pressure on firms to act more socially responsible (Strike et al., 2006), increasing levels of internationalisation are likely to effect EMNEs’ CSR strategy and CSR performance.
Prior studies find that internationalisation has a positive effect on CSR (Attig, Boubakri, El Ghoul, & Guedhami, 2016; Brammer, Pavelin, & Porter, 2006; Kang, 2013). When entering new markets, firms are exposed to risk stemming from their unfamiliarity with the institutional context and from information asymmetries between the foreign firm and its host country stakeholders. This, is turn, increases their risk of violating host country regulations. To reduce this risk and signal commitment to the host country, many firms engage in CSR activities (Attig et al., Kacperczyk, 2009).
15 from global legitimating actors, such as NGOs (Marano et al., 2017; Marano & Tashman, 2012). This increased public scrutiny incentivises EMNEs to protect their global reputation by investing in CSR activities and abstain from CSR scandals (Marano et al., 2017; Strike et al., 2006; Tashman et al., 2018). By implementing a well-structured CSR strategy, EMNEs can disassociate themselves from the institutional voids in their home countries, and overcome the negative stereotypes that stakeholders hold towards them (Marano et al., 2017). In addition, they can promote favourable relationships with their stakeholders and enhance their reputation (Brammer, Pavelin, & Porter, 2009). This way they can reduce challenges stemming from their liability of origin (Marano et al., 2017).
Additionally, as EMNEs and EMs become more integrated and embedded in the global economy, CSR also becomes more important in EMs. Simply put, the greater convergence of global CSR practices creates additional pressures for EMNEs’ to improve CSR performance in their home country (Chapple & Moon, 2005; Davies, 2000; Muller & Kolk, 2009). As EMNEs already increase investments in their foreign subsidiaries to meet host country standards, there is an internal pressure to transfer CSR practices and standards to their home country operations (Chapple & Moon, 2005; Hafsi & Farashahi, 2005). In addition, it is more efficient for firms to standardise CSR principles and standards across all locations (Madsen, 2009). Based on these arguments, we suggest that internationalisation increases the pressure on EMNEs to adopt and integrate global CSR standards into their overall operations. Thus, we propose:
H1: Internationalisation positively influences EMNEs overall CSR performance.
Although many EMNEs increasingly engage in CSR practices commonly known in Western countries (Baskin, 2006; Muller & Kolk, 2009), the actual implementation of these practices is not always straightforward. To enhance CSR performance, firms must invest time and managerial capabilities as well as coordinate and manage their CSR strategy across their different locations (Bouquet & Deutsch, 2008). Thus, companies need to continuously make resource commitments to maintain their CSR performance (Muller, 2018). However, at the beginning of their international expansion, EMNEs may lack the resources and managerial capabilities to enhance their CSR performance. Instead of investing their resources into CSR, they must often focus their attention on more pressing issues, such as R&D investments or capability upgrading (Bouquet & Deutsch, 2008; Luo & Tung, 2007).
16 to a diversity of CSR templates and solutions (Roberston & Crittenden, 2003). This forces them to come up with more progressive CSR solutions in order to meet the diverging demands of different stakeholders (Marano & Kostova, 2016; Seo & Creed, 2002). This way, they become more successful in meeting diverging stakeholder demands and coping with their various rules and regulations (Kolk, 2010; Seo & Creed, 2002). In addition, having operations in several different countries allows EMNEs to build up international networks and form partnerships with already established MNEs. This allows them to gain access to knowledge (Dau, 2013), learn from best practices and incorporate global CSR standards into their own CSR strategy (Hafsi & Farashahi, 2005). Thus, with increasing international market experience, EMNEs become more efficient in deploying these practices (Dau, 2013; Kolk, 2010). Based on these arguments, we expect that with increasing international market experience, EMNEs develop a better knowledge to cope with a variety of regulations and rules and ultimately become more capable in integrating global CSR standards into their overall operations. Therefore, we propose:
H2: The greater EMNEs’ international market experience, the greater their overall CSR performance.
As EMNEs engage in rapid internationalisation strategies (Luo & Tung, 2007; 2018), they often face greater ambiguity and complexity. This in turn, increases their difficulty interpreting and learning from varied CSR practices (Jarzabkowski, Smets, Bednarek, Burke, & Spee, 2013) and transferring their acquired knowledge to their home country operations (Luo & Tung, 2007; 2018). In this way, to be successful, EMNEs must develop the capability to integrate the acquired knowledge and resources in their overall operations (Buckley, Munjal, Enderwick, & Forsans, 2016). Once EMNEs gain experience managing diverse environmental backgrounds, they might find it easier to transfer their gained knowledge to their home country operations (Luo & Tung, 2018). We thus suggest that as EMNEs are exposed to different foreign markets and increase their international experience, they become more capable in managing and integrating their acquired CSR knowledge into their home country operations. Therefore, we propose:
H3: EMNEs’ international market experience positively influences the relationship between internationalisation and EMNEs’ CSR performance.
17 developed countries that are characterised by a strong pollical environment and stringent CSR standards, EMNEs face greater external pressure to improve their CSR practices and comply with host country standards (Attig et al., 2016; Marano & Kostova, 2016). This, in turn, creates backward pressures, forcing EMNEs to improve their CSR standards in their home country (Chapple & Moon, 2005; Muller; 2006; Muller & Kolk, 2009). In addition, once EMNEs have adapted their CSR practices to meet the standards in developed countries, it also becomes more efficient to apply these standards to all countries. Thus, when having subsidiaries in countries with stringent CSR regulations, EMNEs are more pressured to increase CSR performance. At the same time, however, they also benefit more from the transfer of best practices into their home country (Madsen, 2009; Marano & Kostova, 2016; Muller, 2006).
In addition, when internationalising primarily into countries with strong speech and press freedom, EMNEs engagement in irresponsible behaviour is more likely to be detected and likely to damage their overall reputation. Thus, EMNEs are more incentivised to abstain from scandals and maintain a strong CSR performance (Fiashi et al., 2017; Tashman et al., 2018). Based on this, we argue that when exposed to greater external pressures in developed countries, EMNEs are forced adopt higher CSR standards. In the process, they benefit from increased experience in developed countries. Thus, we propose:
H4: The greater developed country experience, the greater EMNEs’ overall CSR performance.
18
3. Methodology
3.1 Sample and data
The sample consists of the largest firms listed on the UNCTAD’s top 100 non-financial TNCs from developing and transition economies (by foreign assets) between 2010 and 2017. This list is published annually by the World Investment Report and showcases the latest trends of global foreign direct investment (FDI). It provides information about a company’s sales, assets and number of employees of both domestic and foreign operations. This sample is appropriate for this research as it consist of EMNEs that have international experience in both emerging and developed countries. Because it covers 13 countries and five continents, it further allows us to examine EMNEs embedded in a range of different institutional contexts.
19
Table 1: Sample description
Country Country Code Number of firms
Brazil BR 5 China CN 16 India IN 9 Mexico MX 5 Malaysia MY 4 Russia RU 3 Singapore SG 8 South Africa ZA 4 South Korea KR 6 Taiwan TW 7 Thailand TH 1 Turkey TU 1
United Arab Emirates AE 2
Full Sample 71
3.2 Variables Dependent variable
To measure a firm’s overall CSR performance, we rely on the raw ESG scores derived from the Thomson Reuters Eikon database. These scores provide data on firms’ environmental, social and governance (ESG) performance. The ESG scores are measured at the firm level and are updated annually. They are a continuous measure between 0 and 100, with zero being the lowest and 100 the highest score. The ESG scores are calculated based on over 400 measures grouped into ten categories. The ten categories cover: resource use, emission, innovation, management, shareholders, CSR strategy, workforce, human rights, community and product responsibility. The scores are based on information derived from publicly reported data (Thomson Reuters, 2019).
Independent Variables
20 for internationalisation from the firms’ annual reports and the UNCTAD list of top 100 non-financial EMNEs.
International market experience is measured based on the total amount of countries in
which the company has operations. We derive this data from the company’s annual reports and their websites.
To measure developed country experience, we first distinguish between developing, transition and developed countries based on the United Nations World Economic Situation and
Prospects Report (WESP). Based on countries’ basic economic conditions, the WESP places
countries into three categories, namely developed, transition and developing (UN, 2019). We compute the proportion of subsidiaries located in developed countries compared to total countries. Information about the companies’ subsidiary portfolio is obtained from the companies’ annual reports and their websites.
Control variables
21 state-owned enterprises (SOEs) also have unique governance structures that might affect their CSR performance (Cuervo-Cazurra, Newburry, & Park, 2016). In addition, SOEs may be less accountable to their stakeholders, and consequently, less inclined to invest in CSR practices (Chapple & Moon, 2005). Therefore, we control for state ownership using a dummy variable, which takes the value “1” if the company is state owned and “0” if not. Data for state ownership is derived from Orbis. Finally, we also control for industry and time fixed effects (Marano et al., 2017). We use the Standards Industry Classification four-digit level to identify the industries in our sample. Our sample comprises of eight industries: (1) agriculture, forestry and fishing (the reference group), (2) mining, (3) construction, (4) manufacturing, (5) transportation, communication, gas, electricity and utilities, (6) wholesale trade, (7) finance, insurance and retail, and (8) services.
Table 2: Description of variables used
Variable Measure Value Source
Overall CSR performance Firm’s CSR score Continuous Thomson-Reuters Eikon
Internationalisation (FS/TS)
Degree of firm internationalisation
Continuous Annual reports, UNCTAD top 100 non-financial EMNEs International market
experience
Number of countries in which firm has subsidiaries
Continuous Annual reports, company, websites Developed country experience Proportion of subsidiaries in developed to total countries
Continuous Annual reports Organisational Slack Ratio of current assets to
current liabilities
Continuous Orbis and Thomson Reuters Eikon Profitability Profit margin Continuous Orbis and Thomson
Reuters Eikon R&D intensity Ratio of R&D
expenditures to sales
Continuous Orbis Firm Size Natural log of total
number of employees
Continuous Orbis and annual reports
Business Group Indicator of membership in BG
0 or 1 Oxford Handbook of business groups (Colpan et al., 2010) State Owned Indicator of state
ownership
0 or 1 Orbis Industry Indicator of industry
(SIC)
22
3.3 Data Analysis
We test our data by applying a time-series, cross-section analysis. This type of analysis can control for unobservable, time-invariant and firm-specific effects (Halaby, 2004). The hypotheses are evaluated using StataSE 15.0. We have an unbalanced panel data set comprising of 71 companies covering the years 2010 to 2017.
To determine whether to use the pooled ordinary least square (OLS), the fixed effect or random effect model, we apply a number of specification tests. First, we ran the Lagrange Multiplier and the F-test to determine whether the pooled OLS method is more appropriate than the random effects or fixed effects model. Based on the outcomes of these tests, the pooled OLS model is not appropriate for our analysis (Appendix 1; 2). Next, we conduct the Hausman test to determine whether to use the random or fixed effect model. The results of the Hausman test suggest that the random effect model is most appropriate for our analysis (χ² =24.48, p > 0.0796).
Next, we check for multicollinearity using the variance inflation factor (VIF). All our variables have a VIF below the threshold of 10, suggesting that multicollinearity is not an issue in the analysis (O’Brien, 2007). We further test for heteroscedasticity using the Breusch-Pagan test. The results provide evidence for heteroscedasticity (χ² =5.01, p < 0.0252). In addition, we apply the Woolridge-test for autocorrelation. The outcome of this test provides evidence for autocorrelation (F = 17.015, p < 0.0001). To address the issues of heteroskedasticity and autocorrelation, we run our regression using clustered standard errors (Tashman et al., 2018).
Based on the prior tests, we run our model using the random effects regression analysis with clustered standard errors. This is a common method when using panel data that has many cross-section but only few time periods and when within-panel variation can be correlated with the error term (Tashman et al., 2018). Prior to the analysis, we standardised our independent variables and the interaction term to reduce the potential for multicollinearity between the main and the moderating effect (Aiken & West, 1991). In addition, we lagged our continuous variables by one period to reduce endogeneity between CSR performance and these variables and establish directionality (Attig et al., 2016; Tashman et al., 2018). The regression equation is as follows:
CSRit= αi + β1Intit + β2IntExpit + β3DCExpit + β4Int x IntExpit + β5OrgSlackit + β6Profit + β7RDit
+ β8Sizeit + β9BGit + β10SOEit + β11Industryit + φit + λt + εit,
23 profitability, RD is R&D intensity, Size is firms size measured as natural log of employees, BG is a dummy taking the value “1” if the firm is affiliated with a business group and “0” if not,
SOE a dummy taking the value “1” if the firm is state-owned and “0” if not, Industry is an
Industry dummy, t represents years, i represents firms, φi represents firm fixed effects, λt
represents time fixed effects and εit the corresponding error term.
4. Results
4.1 Descriptive statistics and correlation
Table 3 illustrates the descriptive statistics, including the number of observations, mean, standard deviation, maximum and minimum. The sample includes a total number of 71 EMNEs from 2010 to 2017. However, due to missing data, the number of observations varies across different variables (457 to 515-year observations).
Table 3:Descriptive statistics
Variable Obs. Mean St. Dev. Minimum Maximum DV 1. CSR performance 515 58.8995 17.18514 8.3 95.61303 IV 2. Internationalisationab 514 0.6081736 0. .273917 .01 1 3. International market experienceab 485 20.48247 14.28294 2 73 4. Developed country experienceab 483 0.2970335 0.1918871 0 0.7619 CV 5. Organisational slackb 515 1.793561 2.116378 0.303 32.347 6. Profitabilityb 496 15.04187 15.89708 -68.198 96.299 7. R&D intensityb 515 0.6617732 1.609901 0 10.145 8. Firm size (ln)b 457 10.6441 1.479499 2.397895 14.07015 9. Business Group 515 0.6543689 0.476036 0 1 10. State owned enterprise 515 0.1902913 0.3929125 0 1 11. Industry 515 4.671845 1.644198 1 8
24 Looking at the dependent variable overall CSR performance, the table shows that the firms have an average CSR score of 58.6, with the minimum score being 8.3 and the maximum score being 95.61. The standard deviation is 17.18, suggesting that the scores are quite dispersed. The first independent variable internationalisation was measured as the ratio of FS/TS. The results show that, on average, 61% of the firms’ sales stem from foreign customers. Internationalisation starts at 1% and reaches a maximum of 100%. The independent variable international market
experience was measured as number of countries in which the EMNEs have operations (incl.
their home country). The number of countries in which the EMNEs have operations ranges from 2 to 73. On average the firms have operations in 20 countries. Developed country
experience was measured as proportion of operations in developed countries compared to total
countries. The results show that, on average, 30% of the countries in which the EMNEs have subsidiaries are developed countries, with a minimum of 0% and a maximum of 76%. In addition, table 3 shows that 65% of our sample is affiliated with a BG and 19% is state-owned.
Table 4 represents the correlation matrix (only significant results are mentioned). In general, the pairwise correlations among our variables are relatively low and all below 0.7, which is the threshold for multicollinearity (Pallant, 2013). This further suggest that multicollinearity is not an issue in our analysis. Looking at the depend variable overall CSR performance, we observe a positive and significant correlation between international market
experience and overall CSR performance. The results suggest that international market
experience positively influences firm CSR performance. In addition, we also see a positive and significant relationship between developed country experience and CSR performance. This implies that a higher proportion of operations in developed countries positively influence CSR performance.
Regarding the control variables, we also see several significant correlations. First, R&D
intensity shows a significant and positive correlation with CSR performance. This suggests that
25
Table 4: Correlation matrix
*p < 0.05, a unstandardized data, bnon-lagged data
26
4.2 Regression results
Table 5 presents the results from the regression analysis using the random effect regression with clustered standard errors. Model 1 includes all control variables. The models 2 to 5 test the hypotheses, model 6 includes all variables combined, and model 7 is the full model also including the interaction effect.
The results of model 1, which tests the relationship between the control variables and overall CSR performance, show that profitability has a positive and significant effect on CSR (β=0.11, p<0.05). These findings are in line with previous research that suggests more profitable firms have more resources to invest in CSR performance (McWilliams & Siegel, 2000). In addition, R&D intensity has a negative and significant effect on CSR performance (β=-1.87, p<0.05). Contrary to the argument that more innovative firms are better at improving their CSR performance (McWilliams & Siegel, 2000), the results suggest that investments in R&D may draw away valuable resources and managerial attention needed for improving CSR performance (Bouquet & Deutsch, 2008). Lastly, firm size also has a positive and significant effect on CSR performance (β=2.17, p<0.1), suggesting that larger firms are more visible and thus, more pressured to improve their CSR performance (Strike et al., 2006). Organisational slack and state-ownership have a negative but insignificant effect and affiliation in a BG a positive but insignificant effect on firm’s overall CSR performance. These results of the control variables are consistent across all 7 models.
Next, model 2 tests hypothesis 1, which predicts that internationalisation positively influences EMNEs’ overall CSR performance. Internationalisation was measured as the ratio of FS/TS. Looking at the results, we find a positive and significant effect of internationalisation on CSR performance (β=2.93, p<0.05). Thus hypothesis 1 can be confirmed.
Hypothesis 2 is tested in model 3 and suggests that the greater EMNEs’ international market experience, the greater their overall CSR performance. The results show that international market experience has a positive and significant effect on CSR (β=3.16, p<0.01). Therefore, hypothesis 2 is confirmed.
27
Table 5: Random effects regression on CSR performance
DV CSR (1) (2) (3) (4) (5) (6) (7) IV Internationalisation (FS/TS) 2.93** (1.19) 2.44* (1.29) 2.23* (1.27) 2.50* (1.34) Int. Exp. 3.16*** (1.19) 2.21* (1.28) 0.84 (1.54) 0.60 (1.52) DC Exp. 5.00*** (1.33) 4.43*** (1.51) 4.36*** (1.51)
Interaction Internationalisation x Int. Exp. 0.68
(0.91) 0.70 (0.95) CV Organisational Slack -0.78 (1.24) -0.91 (1.23) -0.69 (1.24) -0.80 (0.25) -1.05 (1.19) -0.92 (1.19) -0.97 (1.21) Profitability 0.10** (0.04) 0.11*** (0.04) 0.10** (0.04) 0.11*** (0.04) 0.11*** (0.04) 0.11*** (0.04) 0.12*** (0.04) R&D Intensity -1.87** (0.80) -1.93** (0.74) -1.74** (0.84) -1.81** (0.80) -1.79* (0.95) -1.80** (0.90) -1.84** (0.90) Firm Size (ln) 2.17* (1.11) 2.13** (1.07) 2.03* (1.09) 2.03* (1.08) 2.47** (1.08) 2.37** (1.06) 2.32** (1.06) Business Group 3.71 (4.66) 3.95 (4.81) 1.72 (4.21) 1.89 (4.28) 3.28 (3.87) 3.05 (4.01) 3.19 (4.04) State Owned -3.80 (4.43) -0.86 (4.95) -4.10 (4.52) -1.86 (5.03) -2.25 (4.33) -0.16 (4.74) 0.10 (4.76) Industry and year fixed effects Included Included Included Included Included Included Included
Constant 25.59* (14.82) 24.68* (14.28) 28.71** (14.56) 27.60* (14.33) 19.83 (14.63) 20.49 (14.52) 20.84 (14.50) R2 0.1166 0.1147 0.1896 0.1814 0.2647 0.2631 0.2664 Wald χ2 78.02*** 110.30*** 114.12*** 129.94*** 104.32*** 122.40*** 124.76*** N 377 377 354 354 352 350 350 N Firms 71 71 68 68 68 68 68
*** p<0.01, ** p<0.05, * p<0.1, all continuous independent are standardized and lagged by one period, the control variables are lagged by one period
28 Model 5 tests hypothesis 4, which argues that greater developed country experience leads to increased overall CSR performance. The results show that developed country experience has a positive and significant effect on EMNEs’ CSR performance (β=5.00, p<0.01). Thus, hypothesis 4 is confirmed.
Model 6 analyses the effect of all independent and control variables on EMNEs’ overall CSR performance. When looking at the independent variables, we see that internationalisation (β=2.23, p<0.1) and developed country experience (β=4.43,p<0.01) have a positive and significant effect on CSR performance. International market experience has a positive but insignificant effect on CSR performance (β=0.84, p=n.s.). Regarding the control variables, the effect of profitability (β=0.11, p<0.01) and firm size (β=2.37, p<0.05) remain positive and significant, and the effect of R&D intensity on CSR remains negative and significant (β=-1.80, p<0.05). Affiliation with a BG is positive but insignificant, state ownership and organisational slack negative but insignificant.
Lastly, model 7 describes the full model, including the interaction effect of hypothesis 3. The results of model 7 are in line with our previous models. Both internationalisation (β =2.50, p<0.1) and developed country experience (β =4.36, p<0.01) remain positive and significant. However, international market experience, although still positive, no longer has a significant impact on CSR performance (β =0.6, p = n. s.). The interaction term between international market experience and internationalisation also remains positive but insignificant (β =0.70, p = n. s.). The control variables profitability and firm size have a positive and significant effect, R&D intensity a negative and significant effect.
29
Table 6: Results Regression Analysis
No. Hypothesis Results
1. Internationalisation positively influences EMNEs’ overall CSR
performance
Supported (β=2.93, p<0.05)
2. The greater EMNEs’ international market experience, the greater
their overall CSR performance.
Supported (β =3.16, p<0.01)
3. International market experience positively influences the relationship between internationalisation and EMNEs’ overall CSR performance.
Not supported (β=0.73,
p=n.s.) 4. The greater EMNEs’ developed country experience the greater their
overall CSR performance.
Supported (β=5.00, p<0.01)
4.3 Robustness test
To check whether the results of the regression analysis are robust or sensitive to the choice of measures, we carry out a series of robustness tests using alternative measures for some of our variables (Appendix 7 to 12).
First, we run our models using an alternative proxy for internationalisation (Appendix 7, model 1 and 2). We measure internationalisation using the ratio of foreign assets to total assets (FA/TA). Whereas the ratio of FS/TS represents a firm’s overall dependence on the foreign consumer market, the ratio of foreign assets captures their dependence on foreign resources (Sanders & Carpenter, 1998). As not all firms disclose geographic information about their assets, we have fewer observations when measuring internationalisation as ratio of FA/TA. Testing hypothesis 1, we find that internationalisation has a negative but insignificant effect on firms’ CSR performance. Testing hypothesis 3, we find internationalisation has a negative and significant effect on CSR performance (β=-1.93, p<0.1) and international market experience itself a positive effect on CSR (β =4.25, p<0.01). However, the interaction term of internationalisation and experience is negative but insignificant (β=-0.52, p=n.s.). Thus, when measuring internationalisation based on the ratio FA/TA, hypothesis 1 and 3 cannot be confirmed. One possible explanation for this is that investments in foreign assets may demand larger amounts of financial and managerial resources, leaving the firm with less resources to invest in CSR.
30 p=n.s.). One possible explanation is that EMNEs’ home country experience does not provide them with the necessary skills and capabilities to significantly improve their overall CSR performance.
Next, we run our regression analysis using an alternative proxy for CSR by aggregating the firms’ environmental, social and governance score (Appendix 8). Aggregation of different CSR dimensions is a common approach used in previous CSR literature (Attig et al., 2016). The results mainly corroborate the findings of our regression analysis, providing support to both hypothesis 1, 2 and 4. However, when including the interaction between international market experience and internationalisation (model 14), internationalisation is no longer significant (β=2.15, p=n.s.).
We also check the robustness of our results using an alternative measure for firm size and profitability (Appendix 9). As an alternative measure for firm size, we use the natural log of sales. As an alternative measure for profitability, we use return on assets (ROA). Data for sales and ROA is obtained from Orbis. The results corroborate the findings of our regression analysis, providing support for hypothesis 1, 2 and 4. However, when testing the moderation effect of international market experience on the relationship of internationalisation and CSR, we no longer find a significant test of internationalisation (β=1.94, p=n.s.).
Next, we log transform our dependent variable overall CSR performance to check if our results remain significant (Appendix 10). We find that our results are mainly corroborated providing support for hypothesis 1, 2 and 4. However, when including the interaction term of international market experience (model 23), experience is no longer significant. (β=0.03, p=n.s.).
31 Next, we checked whether our results differ across industry. To do this, we split our sample into either manufacturing or non-manufacturing industries (Appendix 12). The results corroborate hypothesis 2 and show that, for EMNEs in both manufacturing and non-manufacturing industries, international market experience significantly and positively influences their overall CSR performance. However, we find that internationalisation, although positive for both, only significantly improves CSR performance for EMNEs engaged in the manufacturing industry (β=3.70,p<0.01). This may be because firms engaging in the manufacturing sector are prone to greater scrutiny, as their operations often have more negative environmental externalities. Moreover, we also find that developed country experience is no longer significant for firms from the manufacturing sector (β=2.72, p=n.s.), but remains significant for firms in other industries (β=6.00, p<0.01). Perhaps, because many MNEs locate their manufacturing activities in EMs, CSR standards in the manufacturing industry are already quite similar to global standards, and EMNEs already pursue well-established CSR strategies. Thus, developed country experience no longer significantly improves their overall CSR performance.
To conclude, the robustness tests provide strong support for hypothesis 4 and marginal support for hypothesis 1 and 2. When measuring CSR in terms of assets and when including experience as an interaction term, we no longer find a significant and positive effect of internationalisation on CSR investments. Moreover, when measuring experience in terms of years since incorporation, we no longer find support for Hypothesis 2. In addition, the robustness tests suggest that the positive effect of internationalisation depends on the firm’s origin and industry sector.
5. Discussion
32 Prior research found that when internationalising, EMNEs must overcome challenges stemming from their liability of origin (Pant & Ramachandran, 2012). To do this, they often engage in global legitimising practices, such as CSR reporting (Marano et al., 2017; Tashman et al., 2018). However, there is still a limited scholarly understanding of how internationalisation effects EMNEs’ CSR performance in their home country. Therefore, our study contributes to IB and CSR literature by shedding light on the relationship between internationalisation and EMNEs’ overall CSR performance. Specifically, we extend research by exploring how EMNEs’ home country CSR performance is affected by different expansion strategies and by prior international market experience.
33 investments (Bouquet & Deutsch, 2008). Thus, especially at the beginning of internationalisation, EMNEs may focus more on other activities, such as R&D and capability upgrading. Thus, they have less attention and resources available to invest in CSR activities. Another possible explanation for the significant effect of internationalisation of sales, is that, when highly dependent on foreign consumer markets, EMNEs may focus more on improving their CSR reputation through CSR investments. An increasing number of consumers demand more socially responsible behaviour, and factors, such as sustainability have become more important drivers in purchasing decisions (Borin & Metcalf, 2010). Prior research found that companies that appear more socially responsible are viewed more favourably among consumers. Thus, investing in CSR initiatives can enhance brand reputation and credibility (Hur, Kim, & Woo, 2014). Consequently, when EMNEs rely heavily on foreign consumers markets, they may be exposed to greater pressure from consumers to invest in CSR. As irresponsible behaviour in one of the EMNE’s operations can translate to negative spillovers and impact its overall reputation (Surroca, Tribó, & Zahra, 2013), the increased consumer pressure for greater CSR initiatives could lead to improved CSR performance across all of the EMNE’s operations.
34 that exposure to different countries promotes CSR initiatives, especially when the companies originate in countries with lower institutional pressure for CSR. Although the EMNEs may already have several years’ experience conducting business at home, the low pressure for achieving strong CSR in their home countries often hinders their ability to gain the necessary capabilities to establish a well-planned CSR strategy (Ioannou & Serafein, 2012; Jamali & Mirshak, 2009). Therefore, experience in foreign countries has a more significant effect on EMNEs’ CSR performance than experience doing business at home.
Third, prior research suggests that the institutional environment of a country strongly influences a company’s CSR activities and performance (Attig et al., 2016; Ioannou & Serafein, 2012). However, the institutional context in EMs provides limited pressure on EMNEs to improve their CSR performance. Thus, both formal and informal contexts impede the development of strong CSR performance (Baughn et al., 2007; Nwabuzor, 2005). However, prior research found that exposure to countries with stringent CSR norms and a strong institutional environment positively influences MNEs’ CSR performance (Marano & Kostova, 2016). Thus, we argued that developed country experience positively impacts EMNEs’ CSR performance. Our findings are in line with prior research and provide evidence that the greater EMNEs’ developed country experience, the greater their overall CSR performance. Our results show that developed country experience significantly influences EMNEs’ overall CSR performance. This suggests that when expanding into more developed countries, EMNEs are exposed to more stringent CSR regulations, and thus face greater pressure to improve their CSR performance (Marano & Kostova, 2006). This, in turn, creates backwards pressures, and it becomes more efficient to deploy developed countries standards into their home country operations (Madsen, 2009; Muller, 2006). In addition, having subsidiaries in advanced economies also provides EMNEs with opportunities to learn from best practices and gain the expertise necessary to develop and implement CSR strategies (Hafsi & Faashahi, 2005).
35 their overall operations is easier. However, although our results show a positive and significant effect of both internationalisation and experience on EMNEs’ CSR performance, they seem to effect CSR performance independently. We do not find significant support that experience strengthens the relationship between internationalisation and CSR performance. One possible explanation for this is that we measured experience as the number of countries in which EMNEs had subsidiaries. However, although embeddedness in different countries exposes EMNEs to a large variety of CSR templates (Robertson & Crittenden, 2003) and can enhance the creativity of CSR solutions (Seo & Creed, 2002), EMNEs may still have to develop additional capabilities to transfer their knowledge to subsidiaries in other countries and integrate it into their overall operations. Therefore, experience in diverse foreign operations may improve CSR performance independently but does not create the necessary capabilities needed to enhance the relationship between internationalisation and CSR performance. Thus, EMNEs may need to gain additional operational and managerial experience in order to better integrate their gained international experience.
Lastly, our findings suggest that the effect of internationalisation and developed country experience may be contingent on a company’s origin and its industry. Whereas internationalisation seems to positively influence Asian EMNEs’ CSR performance, this effect was not supported when considering EMNEs stemming from non-Asian countries. In addition, whereas the manufacturing industry benefitted from internationalisation, developed country experience appears to not significantly influence its CSR performance. One possible explanation for these differences is that certain industries and regions may be more prone to public scrutiny. For instance, companies engaged in manufacturing are more likely to have many negative environmental and social externalities (Handayani, Wahyudi, & Suharnomo, 2017). This might increase their need to engage in global CSR standards. In addition, some regions may attract more FDI from already established MNEs, and thus, already adopted higher CSR standards (Chapple & Moon, 2005). Therefore, the positive effect of internationalisation is less significant for firms stemming from certain regions. Future research could investigate such differences and provide a more detailed view on the different firm characteristics that influence CSR performance.
Contributions and implications
36 First, it contributes to research concerning EMNEs’ internationalisation strategies and these strategies’ implications for EMNE’s home country operations. Prior research found that the international context in EMs provides little incentives for firms to improve their CSR performance (Baughn et al., 2007; Ioannou & Serafein, 2012; Nwabuzor, 2005). However, when expanding abroad, EMNEs face pressures from the global institutional field to adopt global CSR standards, and enhance their CSR activities (Marano et al., 2017; Tashman et al., 2018). We add to this existing body of research by showing that internationalisation positively influences EMNEs’ CSR performance. This suggests that internationalisation provides EMNEs with greater opportunities for learning from best practices and global CSR standards, as well as creates backwards pressures to improve their CSR performance in their home country. In addition, our research provides evidence that expansion into developed countries and experience in different foreign markets can also strengthen CSR performance. We suggest that when expanding into developed markets, EMNEs face greater pressures to improve their CSR standards to meet host country standards (Marano & Kostova, 2016). Thus, they are more incentivised to invest in CSR activities and improve their overall CSR performance. We further argue that once these CSR standards are adapted to developed countries’ CSR standards, it is more efficient for EMNEs to transfer these higher standards to all operations (Madsen, 2009). In this way, EMNEs improve their CSR strategies both at home and abroad. We also find evidence that international market experience in different foreign contexts further improves EMNEs CSR practices, as EMNEs can learn from different contexts and create more creative solutions to meet the demand of varied stakeholders (Seo & Creed, 2002).
37 country stakeholders (Marano & Kostova, 2016), which in turn creates internal pressures to improve the CSR standards in their home country.
Our findings also provide some practical implications and recommendations for corporate managers and policy makers. First, our findings show that international market experience positively influences EMNEs’ CSR performance. By understanding the value of multi-embeddedness, corporate managers can actively embrace the diversity of CSR standards to find more progressive CSR solutions that are suitable for their overall operations. In addition, our results suggest that EMNEs’ must develop additional capabilities to transfer and integrate the acquired knowledge into their overall operations. Therefore, it is advisable for corporate managers to invest in additional capabilities in order to reap the full benefits of their experience and acquired knowledge (Luo & Tung, 2007; 2018).
Second, our paper shows that EMNEs can benefit from gaining experience in more developed countries. Managers can facilitate the experience they gain in developed countries and transfer and integrate the CSR standards of their developed country subsidiaries into their overall operations. This way, they can meet the heightened demands for more socially responsible behaviour of their home country stakeholders (Davies, 2000) and prevent negative spillover effects caused by engaging in irresponsible activities in countries with more lax regulations (Surroca et al., 2013). Thus, managers should aim at developing an integrated CSR strategy and transfer best practices and global CSR standards into their home country operations. In addition, it might be valuable to deploy mechanisms that can assess the success of their CSR strategies.
38
Limitations and future research
We should note several limitations to our study that can also serve as avenues for future research. First, our sample consists of the largest firms in EMs. In general, larger firms are more visible and thus face greater pressure from their stakeholders to act more socially responsible (Strike et al., 2006). In addition, large companies usually have more slack resources that they can invest into their CSR activities than small or medium enterprises (SMEs) (Bansal, 2005). Thus, it may be more important but also easier for EMNEs to improve their CSR performance. Consequently, this study is not representative for SMEs stemming from emerging markets. Future research could extend our study by focusing on SME firms and investigating whether and how their internationalisation strategies influence their CSR strategies. In addition, we only focused on multinationals stemming from emerging markets. As such, our arguments are specific to the institutional context of EMs. Thus, future research could also investigate the generalisability of our findings and enhance our knowledge by comparing multinationals stemming from emerging and developed countries. Moreover, there remains little research on CSR in developing countries. Developing countries are likely to be even less favourable for CSR implementation than emerging markets. Thus, future research could assess how internationalisation and pressure from stakeholders influences CSR practices in developing countries.
Second, we argue that EMNEs improve their CSR performance to gain legitimacy among their stakeholders. However, although we find evidence that internationalisation improves CSR performance, we do not investigate whether this improved performance actually influences EMNEs’ legitimacy among stakeholders. Thus, future research could enhance our understanding about the effects of CSR on EMNEs’ legitimacy by investigating changes in firm reputation or perceptions among stakeholders.
39 foreign sales nor examines whether sales are gained via trade or FDI. However, prior research suggests that pressures to improve CSR are more salient when firms are tied to the country through FDI instead of trade and when embedded in countries with more stringent CSR regulations (Marano & Kostova 2016). Therefore, future research could focus on a smaller sample of EMNEs, and conduct case studies to get a more in-depth insight into their operations. Fourth, to measure expansion into developed countries, we calculated the ratio of operations in developed to total countries. However, we do not account for the size, activities or connectedness of the subsidiaries. Yet, some subsidiaries may be more important to an EMNEs’ operations than others as they may be larger, fulfil crucial operational activities, or be well-connected to the firm’s main operations. Consequently, these subsidiaries are likely to receive greater attention and resources from EMNEs and meeting the demand of their local stakeholders may be more salient. Future research could focus on certain EMNEs and conduct in-depth case studies investigating how different subsidiaries from developed and emerging countries influence the EMNEs’ HQ CSR practices.
Fourth, our study only uses secondary data. Thus, our findings could be advanced by employing methodologies that use primary data. Future research could, for instance, conduct interviews with managers to get deeper insights into the processes that influence EMNEs’ CSR activities. In addition, interviews could shed more light on how managers are influenced by the heightened demand for greater social and environmental sustainability among EMNEs.
6. Conclusion
This paper provides new insights into how EMNEs’ internationalisation strategies affect their overall CSR performance. Its aim was to answer:
How does internationalisation influence EMNEs’ overall CSR performance?
