Isolating the Effect of Social Risk on MNEs’ CSR Reporting: A New Approach Based on China's Belt & Road Initiative

Abstract

This article highlights CSR disclosure as a strategic response of Chinese multinational enterprises (MNEs) to the social risk they face in host countries. Deviating from prior research that aims to directly measure social risk, we offer a new approach to isolate the effect of social risk by leveraging China's Belt & Road Initiative (BRI) as the research context, under which Chinese MNEs are largely protected from political risk in membership countries but are exposed to substantial social risk from local nongovernment stakeholders. Results from difference-in-differences analyses show that after the enactment of the BRI, Chinese MNEs investing in BRI countries significantly increases their likelihood of CSR disclosure than that of their counterparts investing in non-BRI countries. Further, such effects are more pronounced for state-owned MNEs and MNEs in natural resource industries. This research enriches the international business literature on the relationship between political risk and social risk, and that between corporate political actions and corporate social responsibility.

摘要

本文认为，企业社会责任披露是中国跨国公司一种应对东道国社会风险的战略手段。与以往直接衡量社会风险的研究不同，本文以中国‘一带一路’倡议为研究背景，提出了鉴别社会风险影响的新方法。虽然中国跨国公司在‘一带一路’沿线国家的投资使它们免受来自东道国政府的政治风险，但却仍面临着来自当地非政府利益相关者的社会风险。这一背景有助于我们鉴别社会风险，并研究其对跨国公司企业社会责任披露的影响。通过双重差分法分析，本研究发现：在‘一带一路’沿线国家投资的中国跨国公司比那些不在‘一带一路’沿线国家投资的中国跨国公司，更可能提高企业社会责任的披露；而且这种影响对国有跨国公司和自然资源行业的跨国公司更为明显。本文丰富了国际商务文献中关于政治风险与社会风险、企业政治行为与企业社会责任关系的研究

Introduction

Multinational enterprises (MNEs) are usually exposed to two types of nonmarket risk in their host countries, namely political risk which is imposed by the host-country government, and social risk which stems from local nongovernment stakeholders (Franks, Davis, Bebbington, Ali, Kemp, & Scurrah, 2014; Simon, 1984). The existence of such nonmarket risk in host countries reflects the general legitimacy challenges confronted by MNEs (Dowling & Pfeffer, 1975), or the unacceptance by the local government and nongovernment stakeholders. To overcome nonmarket risk, MNEs can undertake corporate political activities (CPA) and corporate socially responsible (CSR) activities to obtain legitimacy in host-country markets (den Hond, Rehbein, de Bakker, & Lankveld, 2014; Rodriguez, Siegel, Hillman, & Eden, 2006; Sun, Doh, Rajwani, & Siegel, 2021). While a large body of international business (IB) research has been devoted to understanding how MNEs respond to and cope with political risk, such as building relationships with the host-country government, connecting with local officials, or developing political capabilities (Li, Meyer, Zhang, & Ding, 2018; Luo, 2006; Rodriguez et al., 2006), the impact of social risk and MNEs’ responses has received limited scholarly attention. The insufficient concern about social risk in the existing IB literature is largely because socio-political risk has often been taken as a whole. Extant studies assume that political risk and social risk are highly correlated (see for instance, Ho, Oh, & Shapiro, 2023) and can be handled by MNEs’ similar nonmarket strategies, meaning that gaining legitimacy can reduce both political and social risks. Nevertheless, the two types of nonmarket risk may actually not go hand in hand, and gaining political legitimacy can sometimes come at the expense of social legitimacy. Therefore, it requires more attention from IB scholars to differentiate political and social risk and to examine the effect of social risk as well as MNEs’ reactions.

The few pioneering works on social risk in the IB field have analyzed the antecedents of social risk in host countries, as well as its influences on MNEs’ performance (Dorobantu, Henisz, & Nartey, 2017). For example, it has been shown that certain nongovernment stakeholders can launch significant activities against MNEs (Nartey, Henisz, & Dorobantu, 2018), and that such activities have a substantial adverse impact on MNEs’ valuation (Henisz, Dorobantu, & Nartey, 2014). Nevertheless, more research is needed to better understand how MNEs cope with and respond to social risk in host countries (Franks et al., 2014), especially for MNEs from emerging economies (EMNEs) like China because they usually lack internationalization experience (Lu, Liu, Wright, & Filatotchev, 2014) and are often considered to be irresponsible (Cuervo-Cazurra & Ramamurti, 2014). In this study, we aim to examine the effect of social risk in host countries on Chinese MNEs’ CSR disclosure. We argue that Chinese MNEs tend to use CSR disclosure as a strategic tool in response to social risk in host countries, because it is a global legitimate practice which conveys to host countries’ stakeholders about the focal firm's commitment to environmental and social responsibilities and sends signals that the firm's investment will align with global norms (Cormier, Magnan, & Van Velthoven, 2005; Ho et al., 2023; Marano, Tashman, & Kostova, 2017).

In order to empirically test our predictions, we leverage China's Belt & Road Initiative (BRI) as a unique research context. China's BRI has presented its significance in IB research and received growing attention from IB scholars, since it has both economic and political implications and exerts influence not only on Chinese MNEs but also on membership host countries (Lewin & Witt, 2022; Li, Van Assche, Li, & Qian, 2022; Witt, 2019). The BRI was officially enacted by the Chinese central government in 2015, aiming to facilitate international trade and investments and to advance economic, diplomatic, and political objectives (Blanchard & Flint, 2017; Lewin & Witt, 2022). With BRI cooperation agreements, the governments of membership countries (BRI countries) are obligated to create larger markets for Chinese MNEs and protect their investments in local markets. Therefore, political risk for Chinese MNEs investing in BRI countries gets alleviated, whereas social risk from local nongovernment stakeholders remains substantial (Buckley, 2020; Zhang, Alon, & Lattemann, 2018). Accordingly, the BRI context is ideal to explore the impact of social risk on Chinese MNEs’ nonmarket strategy.

To isolate the effect of social risk, we adopt a difference-in-differences (DD) research design and use foreign direct investment (FDI) data of Chinese listed firms from the CSMAR database between 2011 and 2018. We classify Chinese MNEs investing only in BRI membership countries throughout the observation window as the treated group, and those investing only in non-BRI countries throughout the window as the control group. Our final sample is composed of 215 Chinese MNEs, among which 84 are treated firms while the other 131 are control firms. We show that after the enactment of the BRI, Chinese MNEs that only invested in BRI membership countries significantly increased the likelihood of CSR reporting and the contents disclosed in CSR reports as a response to the salient social risk in those countries when compared to Chinese MNEs investing only in non-BRI countries. Further, such effects are more pronounced for state-owned MNEs and MNEs in natural resource industries, as these firms usually confront greater social risk than their non-state-owned counterparts or those in other industries.

The current study aims to contribute to various streams of literature in the following ways. First, this study adds to the research on social risk in the IB context. Prior works often took socio-political risk as a whole (e.g., Ho et al., 2023), instead of distinguishing social risk from political risk. Moreover, extant studies have mainly examined the consequences of local nongovernment stakeholders’ actions on MNEs’ performance (see for instance, Dorobantu et al., 2017; Henisz et al., 2014), while neglecting how MNEs can cope with such risk (e.g., Hofman, Li, Sun, & Sun, 2019; Shapiro, Vecino, & Li, 2018). The current study therefore contributes to prior works first by differentiating social risk in host-country markets from political risk. Leveraging the unique research setting of China's BRI, we highlight that although FDI can be shielded against political risk through bilateral agreements between the home and host countries, it can still face substantial social risk imposed by local nongovernment stakeholders in the host country (Shapiro, Vecino et al., 2018). Besides, the current study also develops the exiting literature on social risk by revealing that MNEs can take actions, such as CSR disclosure, to mitigate social risk in host countries, thus improving the existing understanding about how MNEs respond to social risk in host-country markets (Campbell, Eden, & Miller, 2012; Marano et al., 2017; Scherer & Palazzo, 2007).

Second, prior research on EMNEs’ CSR usually attributes their legitimacy challenge to the institutional void in their home countries (e.g., Marano et al., 2017). The current study, however, sheds new light on the driving forces of EMNEs’ CSR engagement. We reveal that EMNEs’ political legitimacy resulting from official bilateral cooperation agreements may lead to increasing challenges of social legitimacy imposed by local nongovernment stakeholders. Such social risk in the host-country market thus provides a strong impetus for EMNEs’ CSR engagement to obtain social legitimacy.

Third, our study also advances the understanding of the relationship between political risk and social risk, and that between CPA and CSR. Extant studies have largely examined the effects of CPA or CSR on mitigating nonmarket risks in host countries separately (e.g., Albino-Pimentel, Dussauge, & Shaver, 2018; Marano et al., 2017). Some recent research has shifted scholarly attention to the relationship between CPA and CSR (see for instance, Li, Shapiro, Peng, & Ufimtseva, 2022; Sun et al., 2021). As to the role of CPA in managing political risk, prior studies have identified typical actions, like building connections with local governmental officials (Albino-Pimentel et al., 2018) or investing in host countries that have a high degree of political affinity with the home country (Li et al., 2018; Sun et al., 2021). The current research reveals that Chinese MNEs’ compliance with the Chinese government by investing in BRI membership countries, which can be seen as a way to leverage CPA to manage political risk, reduces their political risk in those countries, but the social risk resulted from local nongovernment stakeholders becomes more prominent or even increases. The findings therefore suggest that MNEs are supposed to distinguish the two types of nonmarket risk (i.e., political risk and social risk), and create different strategies, thus contributing to the ongoing discussion in recent IB literature about the complementarity between CPA and CSR (Li, Shapiro et al., 2022; Mellahi, Frynas, Sun, & Siegel, 2016; Sun et al., 2021).

Theoretical Background and Hypotheses Development

MNEs’ international investments are exposed to two types of nonmarket risk in the host country, namely, political risk and social risk (Simon, 1984). The two types of risk give rise to legitimacy challenges for MNEs (Kostova & Zaheer, 1999), which refer to the unacceptance and disapproval by the host-country government and nongovernment stakeholders, respectively (Dowling & Pfeffer, 1975; Rodriguez et al., 2006). Scholars have maintained that to succeed in the international market, MNEs need to acquire not only political or legal licenses from the host-country government, but also ‘social licenses’ from local nongovernment stakeholders, especially when they target natural resource or infrastructural industries (Demuijnck & Fasterling, 2016; Ho et al., 2023; Prno & Slocombe, 2012; Shapiro, Hobdari, & Oh, 2018).

MNEs’ responses to nonmarket risks in host-country markets have received substantive scholarly attention. Nevertheless, existing research on political and social risks is unparallel. Prior research has largely focused on the effects of political risk in host-country markets (e.g., Brewer, 1993; Kobrin, 1979) and MNEs’ responsive strategies (Luo, 2006). It has been well recognized that MNEs can engage in CPA and develop political capabilities to alleviate political risk (García-Canal & Guillén, 2008; Rodriguez et al., 2006). Some recent works have shifted scholarly attention in the literature to social risk in host countries (e.g., Franks et al., 2014), and documented various forms of social risk, such as violence against foreign investors by the public (Oh & Oetzel, 2017), protests of community members or nongovernment organizations (Rodriguez et al., 2006), boycotts and sanctions by local customers and suppliers (Klein, Smith, & John, 2004), and social stereotyping and discrimination by different stakeholders (Cui & Jiang, 2012). It has been revealed that social risk in host-country markets can have a substantial detrimental effect on MNEs’ performance (Henisz et al., 2014; Oh, Shapiro, Ho, & Shin, 2020). Nevertheless, the question of how MNEs can proactively respond to and cope with social risk in the host-country market has remained underexplored in the literature. Social risk and political risk may not be highly correlated, and the two types of nonmarket risk cannot be handled by similar nonmarket strategies. Gaining legitimacy politically does not necessarily enhance legitimacy socially. On some occasions, CPA which help to overcome political risk may even increase social risk (Sun, Mellahi, & Wright, 2012). Therefore, it requires more scholarly attention in the IB field on the impact of social risk in host-country market and MNEs’ strategic responses. To address the gap in the literature, the current study investigates the effect of social risk in host countries on Chinese MNEs’ CSR disclosure. We propose that Chinese MNEs tend to use CSR disclosure as a strategic tool in response to host-country social risk, so as to obtain social legitimacy there.

Social Risk and MNEs’ CSR Reporting

In recent decades, some countries have initiated political agreements with each other to promote and attract foreign direct investments (Albino-Pimentel et al., 2018). Such agreements are usually characterized by political motivation (Shapiro, Vecino et al., 2018). The signatory government is obligated to protect foreign investors from political risks, such as potential intervention and expropriation of assets or profits (Kerner, 2009; Stevens, Xie, & Peng, 2016). Nevertheless, signing such agreements can also be seen by local nongovernment stakeholders as giving away investment opportunities to foreign investors that are otherwise available to local investors, thus being perceived as interference with the host country's sovereignty (Neumayer & Spess, 2005). Such perceived threat may therefore lead to actions against MNEs by local nongovernment stakeholders. In other words, when investing in host countries that have signed political agreements with the home country, MNEs can be largely protected from political risk but still exposed to social risk due to the resistant actions of the local nongovernment stakeholders. In some scenarios, MNEs may even face heightened social risk, as local nongovernment stakeholders tend to suspect that these foreign investments have political aims and even some hidden agenda (Li, Newenham-Kahindi, Shapiro, & Chen, 2013; Shapiro, Vecino et al., 2018).

In the current research, Chinese MNEs investing in BRI membership countries are likely to confront particularly high social risk in the local market for several reasons. First, the BRI, which represents a mix of aid, loans, trade, investment, and investment incentives (Buckley, 2020), is deemed as a tool through which China can wield its economic power to advance its economic, diplomatic, geopolitical, and other strategic objectives (Blanchard & Flint, 2017; Lewin & Witt, 2022). Thus, Chinese FDIs in BRI membership countries are likely to be seen as politically motivated, which may raise local stakeholders’ concerns. Such political ambition and intricacies of the initiative may lead Chinese MNEs to encounter increasing difficulties in obtaining social legitimacy from the local nongovernment stakeholders (Li, Liu, & Qian, 2019).

Second, a number of Chinese firms have been found to underinvest in CSR (Du & Vieira, 2012), and some have even been reported to operate irresponsibly in foreign markets, raising local concerns about environmental pollution, human rights issues, and other misconduct (Armony & Strauss, 2012). Partially because of these reports, the BRI has received skepticism among local nongovernment stakeholders that Chinese investors aim to extract resources and cause debt traps in host countries (Arduino & Gong, 2018). These concerns and questions tend to raise hostile attitudes and resistant actions of local nongovernment stakeholders toward Chinese MNEs.

When confronting salient social risk, Chinese MNEs are motivated to pursue ‘social license’ in the host-country market. It is generally assumed in the literature that effective CSR actions will lead to social license (Ho et al., 2023). Accordingly, we propose that CSR disclosure can act as a strategic legitimating tool that enables Chinese MNEs to obtain local stakeholders’ acceptance and trust, thus coping with social risk in host-country markets. This is because CSR reports provide useful information to local nongovernment stakeholders, such as product quality and workplace safety, therefore reducing information asymmetries (Young & Marais, 2012) and creating a less-biased image of Chinese MNEs beyond their stereotype. In addition, CSR reports can signal Chinese MNEs’ alignment with global norms and help them obtain positive evaluations from host-country stakeholders. For instance, disclosure of their efforts to protect the local environment and to create public benefits can produce a socially responsible image in the local stakeholders’ minds. Based on the above arguments, we propose our main hypothesis below:

Hypothesis 1 (H1): The likelihood of CSR disclosure among Chinese MNEs investing in BRI countries increases more than the likelihood of CSR disclosure among their counterparts investing in non-BRI countries after the enactment of the Belt & Road Initiative.

Boundary Conditions

The increase in Chinese MNEs’ CSR disclosure after the BRI is likely to vary with the ownership type and business focus of Chinese MNEs, since different ownership or industry characteristics can lead to different levels of social risk confronting MNEs. Therefore, in what follows, we further examine the boundary conditions of Chinese MNEs’ ownership type and business focus for the main hypothesis above.

Moderating effect of state ownership

SOEs are owned by the government or its agencies, and thus have a strong political imprint (Buckley, Yu, Liu, Munjal, & Tao, 2016). Accordingly, we further propose that when state-owned MNEs invest in BRI membership countries, they may confront greater social risk in the local market (Li, Xia, & Lin, 2017; Meyer, Ding, Li, & Zhang, 2014; Shapiro, Vecino et al., 2018), and thus are more inclined to disclose their CSR activities than their non-state-owned counterparts.

First, SOEs have greater access to resources from the home government, but also assume more political obligations than their non-state-owned counterparts in foreign investment (Li et al., 2017). Because of that, Chinese state-owned MNEs may be considered to be extracting resources from BRI countries and even threatening the national security of host countries. Social groups in BRI membership countries tend to be highly concerned about any political motives behind these firms’ investment activities. As a result, these firms are more likely to receive objection or repulsion from local nongovernment stakeholders. Moreover, Chinese state-owned MNEs may also be criticized for the lack of transparency (Zhang et al., 2018), thus increasing their social risk in BRI membership countries. Second, because of the political imprint (Buckley et al., 2016), Chinese state-owned MNEs can be easily labeled with their stereotypes by local nongovernment stakeholders, which give rise to legitimacy challenges in the host country (Kostova & Zaheer, 1999). In particular, some Chinese MNEs have been accused of irresponsible conduct in their past investment (Gong, 2018). Therefore, to cope with such heightened social risk, Chinese state-owned MNEs can be more eager to pursue social legitimacy in the host countries.

Based on the above arguments, we posit that Chinese state-owned MNEs will face greater social risk in BRI countries, and thus they are more likely to use CSR disclosure as a strategy to obtain social legitimacy in host-country markets. We hereby propose the following moderating hypothesis:

Hypothesis 2 (H2): The increasing of the likelihood of CSR disclosure among Chinese MNEs investing in BRI countries relative to that of their counterparts investing in non-BRI countries after the enactment of the Belt & Road Initiative will be more pronounced for state-owned MNEs than non-state-owned MNEs.

Moderating effect of natural resources industry focus

Prior research suggests that natural resource industries tend to attract a lot of international investments undertaken under bilateral economic agreements (Broadman, 2007; Colen, Persyn, & Guariso, 2016; Wapmuk, 2012) and that Chinese MNEs’ outward investments often target natural resources in host countries (Li et al., 2013; Shapiro, Vecino et al., 2018). Accordingly, we further investigate the contingent effect of Chinese MNEs’ business focus and predict that Chinese MNEs with a natural resource focus in their investment in BRI membership countries will confront greater social risk in host countries.

First, natural resource industries are of great importance to host countries’ economic development and national security (Colen et al., 2016; Hilson, 2012). Any political motivation behind foreign investment in those sectors will likely raise great concerns among local nongovernment stakeholders (Shapiro, Hobdari et al., 2018). Since BRI is considered as a tool for China to achieve its strategic goals (Blanchard & Flint, 2017) and thus often questioned by local nongovernment stakeholders for its motives (Arduino & Gong, 2018), Chinese MNEs that invest in the natural resource sectors in BRI countries are more likely to be perceived as politically driven and have some hidden agenda. Thus, they may face greater social resistance from local nongovernment stakeholders and have greater difficulty in obtaining ‘social licenses’ there. Second, business activities in natural resource industries (such as oil exploitation and coal mining) usually have a larger environmental footprint and social influence (Hilson, 2012) and are usually in close proximity to local communities (Shapiro, Hobdari et al., 2018), thus raising greater concerns among local civilians in host countries (Ho et al., 2023). Thus, MNEs in natural resource sectors are more likely to evoke social conflicts (Shapiro, Hobdari et al., 2018). This is especially true for Chinese MNEs, some of which have been reported to downplay environmental and social responsibilities in their previous international investment (Gonzalez-Vicente, 2012).

For the abovementioned reasons, we predict that Chinese MNEs with a natural resource focus that invest in BRI membership countries tend to have a stronger motivation to disclose their CSR activities so as to cope with the heightened social risk from local nongovernment stakeholders. Some recent studies in the mining industry have shown that MNEs’ commitment to CSR brings them with social license in the host-country market (Ho et al., 2023), which is consistent with our prediction.

Hypothesis 3 (H3): The increasing of the likelihood of CSR disclosure among Chinese MNEs investing in BRI countries relative to that of their counterparts investing in non-BRI countries after the enactment of the Belt & Road Initiative will be more pronounced for MNEs in the natural resource industries than MNEs in other industries.

Methods

Research Design and Sample

Our study aims to offer a new approach to capture the effect of social risk on MNEs’ response strategies, which complements prior attempts to measure social risk and calibrate its impact on firm performance. Specifically, we exploit China's BRI, which was proposed by Chinese President Xi Jinping and underpinned by a desire to better integrate China into the world economy (Liu & Dunford, 2016) and thus can be considered exogenous to individual Chinese MNEs’ CSR disclosure behavior. Under the BRI cooperation agreements, investments of Chinese MNEs are largely protected from political risk by the local government in BRI membership countries, but they are still exposed to substantial social risk among local nongovernment stakeholders (Buckley, 2020; Zhang et al., 2018). Accordingly, by comparing Chinese MNEs that invest in the BRI countries with their counterparts that invest in non-BRI countries, we can isolate the effect of social risk in a relatively clean way that is not possible in other contexts.

Our primary data comes from the Foreign Direct Investments (FDI) Database in the China Stock Market and Accounting Research Database (CSMAR), which compiles Chinese listed firms’ international investments across all sectors and countries. We identify firms that invested only in BRI countries as our ‘treatment group’, and firms that invested only in non-BRI countries as our ‘control group’. Besides, to tease apart the host-country-specific social risk from the social risk incurred by the treatment of our interest (i.e., the BRI), we further limit the sample to firms operating in the same host country before and after the treatment. That is, the ‘treatment group’ firms always operate in a BRI country whereas the ‘control group’ firms always operate in a non-BRI country before and after the treatment. Since the first official document on the BRI was published in March 2015 by the Chinese government¹, we set 2015 as the kickoff year of the treatment. The observation window is from 2011 to 2018. We ended in 2018 because the US-China trade war started in 2019, which may affect Chinese MNEs’ international strategies. We hereby have a 4-year window before the treatment (i.e., 2011–2014), and a 4-year window after the treatment (i.e., 2015–2018). Our final sample is composed of 215 Chinese MNEs, among which 84 are ‘BRI firms’ and the rest 131 are ‘non-BRI firms’, with 1,269 observations in total.

Variables and Measurements

Dependent variables

Following prior related studies, we create two dependent variables. First, we adopt the variable CSR reporting (Marquis & Qian, 2014), which is a dummy variable equal to 1 if the focal MNE has issued the CSR report in year t + 1, and 0 otherwise. Second, we also create the variable, Number of CSR items (Luo, Wang, & Zhang, 2017), which is measured as the number of CSR items disclosed in the focal firm's CSR report (see Table 1 for a list of these items); the variable is coded as 0 when the focal firm did not issue the CSR report in year t + 1.

Table 1. Variables and measurements

Independent variables

Following prior research using the DD approach (e.g., Meyer, 1995), we create two dummy variables.² The first variable BRI firm equals 1 for Chinese MNEs that invested only in BRI countries throughout the observation window, and 0 for Chinese MNEs that invested only in non-BRI countries during the observation window. The second variable Post equals 0 for the before-treatment period (i.e., 2011–2014), and 1 for the after-treatment period (i.e., 2015–2018). The DD interaction term, BRI firm × Post, identifies the treatment effect of the BRI.

Moderating variables

We create two moderating variables. The first variable, SOE, is a dummy variable that equals 1 if the MNE is majority-owned or ultimately controlled by the government (at the central or local level), and 0 otherwise. The second variable, Resource sector, is a dummy variable that equals 1 if the MNE is in a natural resource industry, and 0 otherwise. To define natural resource industries, we refer to the official document ‘Guidelines on Promoting International Cooperation in Production Capacity and Equipment Manufacturing’, which is issued by the State Council of China.³ To capture the moderating effects, we generate three-way interaction terms by multiplying the two moderating variables with the DD interaction term, respectively.

Control variables

In our DD regression analyses, we control for a wide range of factors that may influence MNEs’ CSR disclosure, including factors at both the firm level and external environment level. Specifically, at the firm level, we first control for Firm size, which is measured as the logarithm of firm assets in year t, since larger firms are more likely to face greater public scrutiny over their social and environmental practices (Christmann & Taylor, 2001). We also control for Firm age, which is measured as the number of years since founding, in that younger firms more likely adopt new practices (Marquis & Qian, 2014). Besides, following prior work, we control for the variable Listed on Shenzhen, which is coded as 1 for firms listed on the Shenzhen Stock Exchange, and 0 for firms listed on the Shanghai Stock Exchange, because the stock exchange guidelines for firms to disclose CSR activities might be different (Luo et al., 2017). Next, we control for sample firms’ financial conditions. We include Firm profitability, which is measured as return on assets (ROA) in year t, and Firm leverage, which is measured as the ratio of debts to total assets in year t. We also control for R&D intensity, which is measured as the ratio of R&D expenditures to annual sales revenue, because prior works have revealed that innovative firms are more likely to develop and adopt CSR-related initiatives (Marano et al., 2017). Further, we control for Government subsidy, which is measured as the logarithm of government subsidy the focal firm receives in year t. In addition, we include variables related to firms’ internationalization, in that firms of higher levels of internationalization may be exposed to greater diffusion of new practices or other pressures (Marano et al., 2017; Marquis & Qian, 2014). Accordingly, we control for Export intensity, which is measured as the ratio of foreign sales to total sales revenue in year t, and Foreign shareholding, which is measured as the ratio of shares held by foreign owners to total shares in year t.

Following prior literature, we also control for some important external factors. At the industry level, we include Industry concentration, which is measured by the Herfindahl–Hirschman index, because competition may affect firms’ CSR activities (Campbell, 2007). Herfindahl–Hirschman index of an industry in year t is calculated per the following equation: $\sum\nolimits_{i = 1}^n {{( x_{it}/x_t) }^2}$, where x _it refers to the sales revenue of firm i in the focal industry in year t, and x _t refers to the sum of the sales revenue of all listed firms in the focal industry in year t. At the province or subnational region level, we include Key province, an indicator of whether the firm is located in one of the 14 key provinces (subnational regions) of the BRI, as designated by the Chinese government.⁴

In addition, we follow prior literature and add a series of control variables for host-country-specific characteristics. Specifically, we control for Institutional distance between China and the potential host country j, which is measured by the six indices of governance quality from the World Governance Index (WGI) of the World Bank and calculated using a Mahalanobis approach (van Hoorn & Maseland, 2016). We also control for Geographic distance, which is measured as the distance between the capital of the host country j and China's capital Beijing (in 10,000 km), using data from the CEPII database (Li et al., 2018).⁵ Besides, we create the variable Political affinity to control for the influence of diplomatic relations between host country j and China, which is measured as the correlation of the votes of China and country j at the United Nations General Assembly in year t, using data from Bailey, Strezhnev, and Voeten (2017). As well, we create the variable Institutions supporting collective actions to control for the influence of the objective nonmarket risk of the host country (Oh et al., 2020). It is measured as the sum of two standardized indices⁶, namely Reporters Without Borders’ World Press Freedom Index (WPFI)⁷, and World Economic Forum’s Global Competitiveness Index (GCI's) Judicial Independence (Oh et al., 2020). Finally, we include a full set of industry dummies to account for any fixed, unobservable industry heterogeneity, as well as year dummies to control for any macroeconomic conditions that may influence firms’ CSR disclosure. Table 1 provides a list of the variables and their measurement.

Results

Descriptive Statistics

Table 2 presents the summary statistics and correlations of all variables. As can be seen, the dependent variable CSR reporting has a significantly positive correlation with the two moderating variables SOE and Resource sector (r = 0.284, p = 0.000; r = 0.160, p = 0.000, respectively). Likewise, the other dependent variable Number of CSR items is also positively correlated with the two moderating variables SOE and Resource sector (r = 0.286, p = 0.000; r = 0.168, p = 0.000, respectively). Table 2 also presents the variance inflation factor (VIF) values of all variables. As can be seen, all the VIF values are smaller than 5.00, far below the recommended threshold of 10.00, suggesting that multicollinearity is not a serious concern in our analysis (Ryan, 1997).

Table 2. Descriptive statistics and correlations

Notes: N = 1,269. All bold values are significant at the p < 0.05 level, two-tailed test.

Hypotheses Testing

Table 3 presents the DD regression results for the first dependent variable CSR reporting. Model 1 reports the main effect of BRI firm × Post. The positive and significant coefficient on BRI firm × Post (p = 0.007) supports Hypothesis 1 that the likelihood of CSR disclosure among Chinese MNEs investing in BRI countries increases more than that among their counterparts investing in non-BRI countries after the enactment of the BRI. To assess the economic significance, we follow Hoetker (2007) to calculate the marginal effect for each observation and take the average of the values. Our calculation of the marginal effect of BRI firm × Post indicates that investing in BRI countries increases the probability of Chinese MNEs’ CSR reporting by 12.91%. Moving on to Model 2, results show that the coefficient on the triple-DD term SOE × BRI firm × Post is positive and significant (p = 0.008), thus supporting H2 that the increasing likelihood of CSR disclosure among Chinese MNEs investing in BRI countries relative to that among their counterparts investing in non-BRI countries after the enactment of the BRI will be more pronounced for state-owned MNEs than non-state-owned MNEs. In Model 3, the coefficient on the triple-DD term Resource sector × BRI firm × Post is also positive and significant (p = 0.013), thus supporting H3's prediction that the increasing likelihood of CSR disclosure among Chinese MNEs investing in BRI countries relative to that among their counterparts investing in non-BRI countries after the enactment of the BRI will be more pronounced for MNEs in the natural resource industries than MNEs in other industries. We notice that after including the triple-DD term, the variables related to the two-way DD term BRI firm × Post become insignificant, which leads to a small incremental R-squared between Model 1 and Models 2–3. It suggests that the triple-DD term of Model 2 and Model 3 explained the models instead of the two-way DD term BRI firm × Post, indicating that the main effect is more pronounced for state-owned MNEs and for MNEs in the natural resource industries. Finally, Model 4 reports the results of the full model, again providing support for the two moderating hypotheses.

Table 3. Main results: DD regression results for dependent variable CSR reporting

Notes: Standard errors in parentheses. Exact p values in the table. Two-tailed test. This table uses the Probit model since CSR reporting is a dummy variable.

Table 4 reports the DD regression results with Number of CSR items as the other dependent variable. The results are highly consistent with those in Table 3. Model 1 shows that the coefficient on the DD term BRI firm ×Post is positive and significant (p = 0.002). Our calculation of the marginal effect of BRI firm × Post indicated that investing in BRI countries increases the number of CSR items disclosed by 0.423. In Models 2 and 3, the coefficients on the respective triple-DD term are both positive and significant (p = 0.000 and p = 0.002, respectively). Finally, Model 4, the full model, shows that SOE and Resource sector both have a positive and significant moderating effect. Taken together, the results in Table 4 provide further support to all the three hypotheses.

Table 4. Main results: DD regression results for dependent variable Number of CSR items

Notes: Standard errors in parentheses. Exact p values in the table. Two-tailed test. This table uses the Poisson model since number of CSR items is a count measure.

Figures 1–3 depict the effects by illustrating the differences in the number of firms that issue CSR reports. From Figure 1, we can see that the number of non-BRI firms issuing CSR reports in the post-BRI period increased by 10%, while the number of BRI firms issuing CSR reports in the post-BRI period increased by a higher proportion (i.e., 84.6%). And as Figures 2 and 3 show, the effect is more pronounced for state-owned MNEs and MNEs in natural resource industries.

Figure 1. Graphical presentation of the main average effect

Figure 2. Graphical presentation of the moderating effect of State ownership

Figure 3. Graphical presentation of the moderating effect of Natural resource industry

Robustness Tests and Supplementary Analyses

Analysis of firms’ CSR ratings as an alternative dependent variable (DV)

Prior research has shown that firms’ CSR disclosure can be consistent with their CSR performance (Marquis & Qian, 2014). Accordingly, we use firms’ CSR ratings as another dependent variable for a robustness test, because a high CSR rating signals the high quality of a firm's CSR practices and its conformity to societal norms and expectations. Following prior research, we use the data on Chinese listed firms’ CSR ratings from Hexun (e.g., Xiong, Lu, Skitmore, Chau, & Ye, 2016), an independent data vendor that evaluates Chinese listed firms’ CSR activities along five dimensions (i.e., environment, employees, suppliers-customers, shareholders, and society), based on their CSR reports and annual reports.

Table 5 reports the DD regression results with CSR rating as the dependent variable. Model 1 shows that the coefficient on the DD term BRI Firm × Post is positive and significant (p = 0.084), indicating that the increase of social risk in BRI host countries indeed motivates Chinese MNEs to improve their CSR performance, as evidenced by improved CSR ratings. Models 2 and 3 also show that SOE and Resource sector play a positive and significant moderating role (p = 0.023 and p = 0.016, respectively). Model 4 reports the results of the full model, which again provides strong support for our hypotheses.

Table 5. Robustness results: DD regression results for dependent variable CSR rating

Notes: Standard errors in parentheses. Exact p values in the table. Two-tailed test. This table uses the ordinary least squares (OLS) model since CSR rating is a continuous variable.

Alternative samples to identify the effect of BRI

The core assumption of our research design is that Chinese MNEs investing in the BRI membership countries mainly confront social risk instead of political risk. Accordingly, we limit the sample to Chinese MNEs operating only in the same single host country before and after the treatment in the main analysis, which enable us to separate the host-country-specific social risk from the effect of the BRI by controlling for a series of host-country-specific variables in the regressions. As a robustness test, we expanded our sample by including Chinese MNEs that invest in multiple countries. Likewise, we classifiy Chinese MNEs that invested only in BRI membership countries as our ‘treatment group’, and Chinese MNEs that invested only in non-BRI countries as our ‘control group’. In so doing, we collected 1,368 Chinese listed firms. Out of the 1,368 listed firms, 597 are in the treated group, and the rest 771 are in the control group. Tables 6 and 7 report the regression results for dependent variables CSR reporting and Number of CSR items, respectively. The results are highly consistent with those in Tables 3 and 4, thus providing further support to all three hypotheses.

Table 6. Robustness results: Expanding the sample to firms investing in more than one country (DV: CSR reporting)

Notes: Standard errors in parentheses. Exact p values in the table. Two-tailed test. This table uses the Probit model since CSR reporting is a dummy variable.

Table 7. Robustness results: Expanding the sample to firms investing in more than one country (DV: Number of CSR items)

Notes: Standard errors in parentheses. Exact p values in the table. Two-tailed test. This table uses the Poisson model since number of CSR items is a count measure.

Robustness test with propensity score matching approach

We also leverage the Propensity Score Matching (PSM) approach as a robustness test. Specifically, in the first step, we estimated the likelihood of a certain Chinese MNE investing in a BRI country, using BRI firm as the dependent variable and including the full set of control variables. In the second step, we reran the DD regressions based on the matched sample. Due to limited space, we reported the results with the PSM approach in the Appendix. Table A1 presents the matching process, while Table A2 presents the regression results based on the matched sample. As can be seen, the results are highly consistent with our main analyses, thus providing additional support to our theoretical predictions.

Parallel trend assumption tests

For the validity of the DD research design, the parallel trend assumption must hold. We therefore examined the parallel trend assumption in two ways. First, we used a regression approach. We created a linear time-trend variable t, which is coded as 1 to 4 for the years of 2011-2014, and then regress the dependent variables, CSR reporting and Number of CSR items, respectively, on BRI firm, t, and their interaction term BRI firm × t, as well as the full set of control variables. The results are reported in Table A3 in the Appendix. As can be seen, prior to the treatment, there were no significant differences in the likelihood of CSR reporting or the number of CSR items between the treated and control groups. Second, we also plot the parallel trend between the treated and control groups in CSR reporting before the treatment of our interest, which is presented in Figure A1 in the Appendix. As it shows, the treated and control groups followed a similar trend in terms of the likelihood of CSR reporting in the pretreatment period. Taken together, the parallel trend assumption holds in the current research, thus validating our DD research design.

Analysis in Logit models

We also ran Logit regressions for the dependent variable CSR reporting for robustness. The results are reported in Table A4 in the Appendix. As can be seen, the results are highly consistent with our main analyses, thus providing additional support to our theoretical predictions. Moreover, the odds ratio analyses for the Logit model show that the odds ratio of DID term BRI firm × Post is 2.4, which means investing in BRI countries multiply the odds of issuing CSR reporting by 2.4, or other words, the probability of BRI firms issuing CSR reports after investing in BRI countries is 70.59% (Hoetker, 2007).

Supplementary analyses

We further conducted separate tests on the disclosure of each CSR dimension to gain deeper insights. First, we created a series of dummy variables for each of the ten dimensions available in CSR reports. Each dummy variable equals 1 if the focal firm has released relevant content in the dimension in the CSR report in year t + 1, and 0 otherwise. We then ran Probit regressions. The results are reported in Table A5 in the Appendix. As can be seen, Chinese MNEs investing in the BRI membership countries are more likely to disclose their socially responsible activities in terms of protection of shareholders, employees, suppliers, customers, and the environment, as well as activities related to philanthropy, development of CSR institutions, and workplace safety, compared to their counterparts investing in non-BRI countries after the enactment of BRI.

Discussion

Leveraging China's BRI as a novel research setting (Buckley, 2020; Li, Van Assche, et al., 2022) and the DD research design, the current research shows that, after the enactment of China's BRI, Chinese MNEs investing in BRI countries significantly increase their likelihood of CSR disclosure, compared to that of their counterparts investing in non-BRI countries. Given that the BRI protects Chinese MNEs from political risk in membership countries to a large extent, but the social risk they face in those countries remains, we interpret our results as that Chinese MNEs adopt CSR disclosure as a strategic response to social risk in the host-country market. Further analyses show that state-owned MNEs and MNEs in natural resource sectors are more likely to disclose their CSR activities than their non-state-owned counterparts or those in other industries, consistent with the arguments that these firms usually confront higher social risk in the host countries. While we have identified the social risk MNEs face after investing in BRI countries as a significant force driving MNEs’ CSR disclosure, the impacts are relatively limited as manifested by the size of the Pseudo R-square in the regression models. The reason might be because we used corporate-level CSR information as the dependent variable, but some unobservable domestic factors might latently influence corporate-level CSR disclosure. Moreover, besides social risk, sometimes political risks might also explain part of the variance in MNEs’ CSR disclosure.

Theoretical Contributions

Our study makes some important contributions to various streams of literature. First, our study contributes to the research on social risk in the IB field. Prior works have largely looked into the adverse effects of nonmarket risk on MNEs financial outcomes (Dorobantu et al., 2017; Henisz et al., 2014), while lacking in-depth discussion about MNEs’ responses or reactions. Besides, extant studies often took socio-political risk as a whole (e.g., Ho et al., 2023), instead of distinguishing social risk from political risk. Our study maintains that political and social risks may not go hand in hand and reveals that Chinese MNEs can take CSR actions, such as CSR disclosure, as a nonmarket strategy to mitigate the social risk in host-country market, thus developing the existing literature on social risk. Furthermore, this study makes an empirical contribution by offering a new approach to identify the effect of social risk. Instead of tempting to directly measure social risk, our study manages to isolate the effect of social risk from that of political risk by leveraging China's BRI as the research context. Due to the BRI cooperation agreement, Chinese MNEs are largely protected from political risk by the local government but are still exposed to social risk from local nongovernment stakeholders who may accuse Chinese MNEs’ FDI of lacking transparency or having some hidden agenda (den Hond et al., 2014). In so doing, we can compare Chinese MNEs investing only in BRI membership countries throughout the observation window with those investing only in non-BRI countries, thereby isolating the effect of social risk, which is independent of political risk, on Chinese MNEs’ strategic response. This approach departs from but complements prior related studies that aim to directly measure social risk (e.g., Röell et al., 2022).

Second, our study contributes to CSR research by offering a new explanation of the driving forces of EMNEs’ CSR. Prior research on EMNEs’ CSR usually considers the influence of the institutional void in their home countries (e.g., Marano et al., 2017) or the role of institutional distance (Marano & Kostova, 2016), as the antecedents of EMNEs’ CSR engagement. Our study, instead, is concerned with the social risk and legitimacy challenges imposed by local nongovernment stakeholders, and it shows that such social risk can be a strong impetus for MNEs’ CSR activities, thus developing the existing literature on EMNEs’ CSR.

Third, the current research also contributes to the exciting discussion in recent IB literature about the relationship between political risk and social risk, and that between CPA and CSR. Most studies separately examine the effect of CPA and CSR on alleviating nonmarket risks in host countries (Albino-Pimentel et al., 2018; Marano et al., 2017), although some recent research started to discuss the relationship between CPA and CSR (Li, Shapiro, et al., 2022; Sun et al., 2021). The current research enriches this discussion by showing that Chinese MNEs’ compliance with the Chinese government by investing in BRI membership countries, which can be seen as one type of CPA, reduces their political risk in those countries, but the social risk resulted from local nongovernment stakeholders becomes prominent or even increases. Besides, existing research has considered that political risk and social risk can be handled by similar nonmarket strategies, and that gaining legitimacy can reduce both political and social risks (Campbell et al., 2012; Marano et al., 2017; Scherer & Palazzo, 2007). Our study, instead, conveys an important message that gaining legitimacy politically might come at the expense of social legitimacy. Although CPA helps MNEs gain political legitimacy, it might unexpectedly arouse social risk in host-country markets. In such scenarios, CSR can play a strategic role in obtaining social legitimacy (Sun et al., 2012). In so doing, our study reveals the complementarity between CPA and CSR and thus contributes to the broad literature on the relationship between CPA and CSR (Li, Shapiro, et al., 2022; Sun et al., 2021).

Practical Implications

Our findings also have practical implications for firms undertaking FDI in host countries. Although political cooperation agreements between the home and host countries can shield MNEs against political risk to a large extent, they do not protect MNEs from social risk incurred by local nongovernment stakeholders; in fact, social risk in such a scenario is even likely to be heightened, due to local nongovernment stakeholders’ concerns about any hidden political agenda. Our findings suggest that as a strategic response, MNEs can enhance their CSR practices, such as releasing CSR reports and disclosing more contents in reports, to make their social commitment public to the society and various stakeholders. In so doing, Chinese MNEs can avoid falling into a ‘bully trap’ while still leveraging the power of the home and host governments (Witt, 2019). In addition, inspired by the recent literature on political CSR (Maier, 2021; Scherer & Palazzo, 2011; Scherer, Rasche, Palazzo, & Spicer, 2016), our current research may also imply that MNEs can co-create or reshape the institutional environment in host-country markets where there exist institutional voids by developing political CSR strategies (Scherer et al., 2016), such as contributing to global regulation and providing public goods (Scherer & Palazzo, 2011). In this way, the nonmarket risks in host-country markets can be alleviated as the local institutional environment gets improved. For policymakers in the home country, our findings suggest that when designing and implementing an initiative to promote FDI, it is also important to have concrete policy instruments and measures in place to encourage their firms to engage in CSR activities more actively, in order to succeed in their international investment.

Limitations and Future Research Directions

The current research also comes along with several limitations, which may point out some fruitful directions for future research. First, the current research did not take into account possible interactions between political risk and social risk. In particular, the relationship and degree of trust between nongovernment stakeholders and the local government might affect the magnitude of social risk. Future studies can therefore examine how the relationship between the government and nongovernment stakeholders in the host country affects the social risk that MNEs may encounter. Second, data availability has limited our ability to examine MNEs’ CSR activities in more depth. Due to the lack of foreign subsidiary-level CSR information, the current research compromises to use corporate-level CSR information. Such aggregated data also lead to a relatively small size of the Pseudo R-square in our regression models, because social risk in discrete host countries may only explain part of the variance in MNEs’ corporate-level CSR activities and disclosure decisions. We encourage future research to collect finer-grained data about Chinese MNEs’ host-country-specific CSR activities and use novel methods (e.g., textual analysis of local newspapers) to study how Chinese MNEs may tailor specific CSR activities to different host countries and various types of nonmarket risk. As China's BRI receives increasing attention among IB scholars (e.g., Lewin & Witt, 2022; Li, Van Assche, et al., 2022), we hope that this article can be a catalyst to encourage future studies to examine more carefully the success of CSR efforts of Chinese MNEs in BRI membership countries.