Not in my homeland: Immigrant CEOs and the geography of corporate social irresponsibility

Abstract

Research Summary

This study examines how immigrant CEOs influence the geography of multinational enterprises' (MNEs) corporate social irresponsibility (CSI) incidents. Building on place attachment theory and social capital theory, we theorize that immigrant CEOs have strong psychological attachment to and social capital in their homeland countries, which could reduce the occurrence and media disclosure of their MNEs' CSI incidents in those countries. Moreover, this effect will be further enhanced if the CEO emigrates as an adult, if the focal firm has a higher sustainability rating, and if the CEO's homeland country has lower press freedom. A difference-in-differences analysis using a propensity score-matched sample of MNEs from the US S&P 500 during the 2007–2020 period supports our arguments.

Managerial Summary

Why do firms engage in different levels of corporate irresponsibility in different locations? Our study reveals a connection between the appointment of an immigrant CEO and the locations of multinational enterprises (MNEs') irresponsible incidents. Immigrant CEOs can not only reduce their MNEs' harmful acts in their homelands because of their strong psychological attachment but also leverage social relationships in their homelands to mitigate media coverage of their firms' negative incidents there. Such impacts of appointing immigrant CEOs are subject to CEOs' immigration as adults or children, their firms' sustainability ratings, and their homeland countries' press freedom.

1 INTRODUCTION

Corporate social irresponsibility (CSI) incidents, including environmental pollution, human rights violations, and governance fraud, have garnered increasing research attention due to their severe threats to social stakeholders' legitimate claims (Lange & Washburn, 2012; Strike et al., 2006) and their profound impact on firms' financial performance and reputational standing (Kölbel et al., 2017; Liu et al., 2022). CSI incidents exhibit significant geographical variation, occurring more frequently in certain locations while being less common in others. For example, our analysis of RepRisk data reveals that in 2020, 69.8% of worldwide CSI incidents were reported in five primary countries: the United States, China, the United Kingdom, Canada, and France. In recent years, multinational enterprises (MNEs) have increasingly taken their worldwide social behaviors seriously, recognizing that irresponsible actions in foreign countries can be attributed to negative intent by local stakeholders and lead to harsher penalties and more negative perceptions compared to domestic firms (Crilly et al., 2016). Therefore, understanding the geographical variation of CSI incidents—namely, answering the where question—is crucial for MNEs with global operations to assess and potentially mitigate the risks of involving in these incidents across different countries.

Notably, existing research has provided valuable insights into how macro-level factors, such as home- and host-country institutions as well as home-host country distance, shape the geography of MNEs' CSI incidents (Campbell et al., 2012; Spencer & Gomez, 2011; Surroca et al., 2013). However, focusing solely on macro-level factors leads to an incomplete understanding of the geography of MNEs' CSI incidents as microlevel factors are equally important. In particular, microlevel factors influence the internal dynamics and decision-making processes within MNEs (Crilly et al., 2008; Shea & Hawn, 2019), providing complementary insights into the mechanisms driving CSI incidents in different locations. Among various microlevel factors, scholars have long recognized the critical role of CEOs as central drivers of corporate social performance (Wernicke et al., 2022) because they can significantly shape firms' strategic decisions regarding environmental, social, and governance (ESG) initiatives (Bertrand et al., 2021; Tang et al., 2015). Despite the significant insights, we still know little about whether and how CEOs influence the geography of CSI incidents. Filling this gap, in this study, we focus on CEOs' immigrant background, a geography-related characteristic (i.e., reflecting CEOs' countries of origin), and theorize how it could influence the geography of MNEs' CSI incidents.

Immigration is an increasingly salient global trend and has an important impact on our societies. The international migrant population increased by 83% from 1990 to 2020, while the world population only increased by 29% over the same timeframe (United Nations, 2020). Today, immigrants play an indispensable role in contributing to the economy by leading and founding businesses. Indeed, 44.8% of Fortune 500 companies in 2023 were founded by immigrants or their children (American Immigration Council, 2023). In our sample of S&P 500 firms, we find that 28.6%1 of them were led by first-generation immigrant CEOs during the 2007–2020 period. Some well-known immigrant CEOs include Andrew Grove (CEO of Intel) from Hungary, Elon Musk (CEO of Tesla) from South Africa, Safra Catz (CEO of Oracle) from Israel, and Sundar Pichai (CEO of Google) from India (Boardroom Insiders, 2017).

Building on place attachment theory (Giuliani, 2003; Lewicka, 2011) and social capital theory (Nahapiet & Ghoshal, 1998; Portes, 1998), we argue that immigrant CEOs tend to reduce the occurrence and media disclosure of their firms' CSI incidents in their homeland countries—the countries where they were born and migrated from—because they have strong psychological attachment to and social capital in these countries. First, immigrant CEOs' strong attachment to their homelands could foster place-protective behaviors that mitigate their firms' engagement in CSI incidents in these countries. Second, immigrant CEOs tend to have strong social connections with local stakeholders in their homelands. Since CSI incidents are typically disclosed by the media (Hawn, 2021; Kölbel et al., 2017), immigrant CEOs could utilize their social capital to influence related parties (e.g., media and government) to reduce the public disclosure of their firms' CSI incidents in their homeland countries. Taken together, we argue that CEOs' immigrant background plays an important role in influencing the geography of their firms' CSI incidents, such that appointing an immigrant CEO could lead to a reduction in publicly reported CSI incidents in the CEO's homeland country.

Furthermore, we propose three moderators at the individual, firm, and country levels that could influence the main mechanisms we outline. First, we posit that the impact of immigrant CEOs on the geography of CSI incidents is more pronounced among adult immigrants—those who have reached adulthood before migrating to a new country—compared to non-adult immigrants. This is because adult immigrants who have resided in their homeland countries for an extended period have had the opportunity to fully develop their place-based identities and establish valuable social capital. Second, we hypothesize that the impact of immigrant CEOs on reducing their MNEs' CSI incidents in their homeland countries is stronger when the MNEs have higher sustainability ratings, which not only enhance immigrant CEOs' ability to engage in place-protective behaviors in their homelands but also strengthen their motivation to do so to protect their firms' sustainability reputations. Third, we posit that the impact of immigrant CEOs is weakened by the level of press freedom in their homeland countries. When press freedom is high, immigrant CEOs are likely less capable of influencing related parties to reduce media coverage of CSI because the media can freely and independently report on CSI incidents without external interference. Using difference-in-differences (DID) analysis and a unique propensity score-matched sample of MNEs from the US S&P 500, we find strong support for the above arguments.

Our study makes several important contributions to the strategy literature. First, we advance research on CSI by exploring the microlevel antecedents of the locations of MNEs' CSI incidents, thereby addressing calls for more attention to the microfoundations (e.g., the role of individuals, such as executive leaders) of firms' social and environmental behaviors (Crilly et al., 2008; Shea & Hawn, 2019). We identify CEOs' immigrant background as an important yet understudied microlevel factor that influences the geography of CSI incidents, enriching the existing literature on the geography of CSI incidents that has overwhelmingly focused on the role of macro-level institutional antecedents (Campbell, 2007; Keig et al., 2015; Sun et al., 2021).

Second, we contribute to research on the role of CEOs in shaping corporate social responsibility (CSR) by theorizing and testing the effect of CEOs' immigrant background on CSI. Different from previously studied CEO characteristics, such as education (Lewis et al., 2014), gender (Cronqvist & Yu, 2017), hubris (Tang et al., 2015), and political ideology (Chin et al., 2013), and their effects on a firm's overall CSI incidents, CEOs' immigrant background is a geography-related attribute that reflects where these individuals originally came from, thus allowing us to examine the nuanced influence of CEO characteristics on the geography of CSI incidents.

Third, we extend the emerging body of research on immigrant CEOs by introducing mechanisms based on place attachment and social capital. Existing studies on immigrant CEOs have limited their focus to outgroup identity in the destination countries where the CEOs migrated to (Mata & Alves, 2018) and the role of CSR in helping them overcome their liability of outsidership in these countries (Bertrand et al., 2021). Per the outgroup identity logic, immigrant CEOs may not need to devote additional effort to reducing CSI in their homelands because they do not suffer from outgroup identity there. However, our study shows that immigrant CEOs are often more committed to reducing CSI incidents in their homelands because they have strong psychological attachment to and social capital in these countries. Our study stresses the importance of considering alternative mechanisms besides outgroup identity to better explain immigrant CEOs' behaviors in their homeland countries.

2 THEORETICAL DEVELOPMENT

2.1 Research on the geography of CSI incidents

There has been growing scholarly attention on why MNEs exhibit varying levels of social performance across different countries (Goerzen et al., 2024). To date, existing studies have predominantly focused on macro-level factors that affect geographical variation in corporate social performance. These factors include characteristics of the home country, the host country, and inter-country relationships. For instance, growing stakeholder pressure in an MNE's home country can trigger the transfer of CSI practices from its headquarters to its overseas subsidiaries (Surroca et al., 2013). Also, host country characteristics, such as local peers' pressure, regulations, and socio-cultural norms, can influence MNEs' CSR behaviors (Gardberg & Fombrun, 2006; Husted, Montiel, and Christmann, 2016b). For example, widespread corruption in a host country can pressure and/or encourage an MNE's subsidiaries to engage in bribery locally (Spencer & Gomez, 2011). Moreover, the differences between home and host countries in terms of cultural, administrative, geographic, and economic distance significantly shape MNEs' engagement in CSR activities (Campbell et al., 2012).

Despite the significant insights into the macro-level antecedents, little research has been devoted to the microlevel factors influencing the geographic variation in corporate social performance. Indeed, as stated by Shea and Hawn (2019, p. 1609), “We still know very little about the microfoundations of CSR.” Given the significant impact of microfoundational factors on firm-level decisions (Felin et al., 2015), it is crucial to explore the microlevel factors, such as individual motivations and psychological elements, underlying corporate social performance (Aguinis & Glavas, 2012; Crilly et al., 2008) in addition to macro-level factors of the country environment. Thus, in this study, we focus on a specific microlevel factor—CEOs' immigrant background—and examine how it influences the geography of firms' CSI incidents.

2.2 The influence of CEOs on corporate social performance

Research on upper echelons theory suggests that CEOs often have substantial discretion in influencing firm strategies and behaviors (Cannella et al., 2009; Hambrick & Mason, 1984). Existing literature found that the characteristics and leadership styles of CEOs are closely related to the extent to which their firms engage in socially responsible or irresponsible activities. For instance, Chin et al. (2013) found that CEOs' political ideologies manifest in their firms' social responsibility profiles. Tang et al. (2015) found that CEO hubris, which is framed as extreme self-confidence and pride, is negatively related to a firm's socially responsible activities but positively related to its socially irresponsible activities. Prior studies also found that CEOs' education and tenure (Lewis et al., 2014) and belief in a fair market ideology (Hafenbradl & Waeger, 2017) can affect firms' commitment to social responsibility. Overall, CEOs play an important role in influencing corporate social performance.

In light of the growing trend of global migration and the rising prevalence of immigrant CEOs, increasing research attention has been devoted to studying the potential influence of CEOs' immigrant background on corporate social performance. Existing research has recognized that immigrant CEOs are often labeled “outsiders” because of their foreign-country origin and different cultural backgrounds (Mata & Alves, 2018; Thams & Rickley, 2024). For example, Citibank's Indian-born CEO, Vikram Pandit, was described as “a cultural outsider” by the US media (Hagan, 2009). According to social identity theory (Tajfel & Turner, 1986), individuals tend to categorize themselves and others into groups and evaluate ingroup members more favorably than outgroup members. Therefore, CEOs who suffer from outgroup identity often confront heightened vulnerability to negative perceptions and discrimination (Mata & Alves, 2018; Park & Westphal, 2013). Building on social identity theory, Bertrand et al. (2021) argue that immigrant CEOs, who are typically perceived as outgroup leaders by local stakeholders, are more likely to enhance corporate social performance to overcome their liability of outsidership and gain local legitimacy and trust.

Although prior studies have widely recognized the outgroup identity of immigrant CEOs in their destination countries, they overlooked that these CEOs may not be considered outsiders in their homeland countries. According to the outgroup identity logic, immigrant CEOs may not exert additional efforts to engage in high levels of CSR in their homelands because they do not need to overcome outgroup identity there. However, this reasoning contradicts actual observations as immigrant CEOs often work extra hard to protect their homelands. For example, since the Russia-Ukraine war began, Jan Koum, the Ukrainian-born WhatsApp founder, has financed many humanitarian programs to support Ukraine and its migrants (Palchynska, 2022). Similarly, Google CEO Sundar Pichai and Microsoft CEO Satya Nadella actively directed their firms to aid India—their homeland country—during the COVID-19 pandemic (O'Brien, 2021). These phenomena highlight the need for alternative theoretical mechanisms to complement the outgroup identity logic when explaining immigrant CEOs' behaviors in their homelands. In the following section, we introduce two such mechanisms based on psychological attachment and social capital.

2.3 Immigrant CEOs in homeland countries: Psychological attachment and social capital

Building on place attachment theory (Giuliani, 2003; Lewicka, 2011) and social capital theory (Nahapiet & Ghoshal, 1998; Portes, 1998), we argue that immigrant CEOs have stronger psychological attachment to and social capital in their homeland countries than CEOs who are not from these countries.

First, immigrant CEOs have strong attachment to their homeland countries. Developed by environmental psychologists, place attachment theory captures “the bonding that occurs between individuals and their meaningful environments” (Scannell & Gifford, 2010, p. 1). Places are physical settings where human activities and social and psychological processes occur (Giuliani, 2003; Low & Altman, 1992; Stokols & Shumaker, 1981), and attachment refers to the cognitive and emotional linkage of an individual to a particular setting or milieu (Low & Altman, 1992). Although there are various ways to define “place,” it is undeniable that “home” is a prototypical type of place (Easthope, 2004). Some scholars have even argued that a “human-meaningful place is divided between home and not-home” (Lewicka, 2011, p. 212). Furthermore, the literature has emphasized that the meaningfulness of home is especially strong for individuals who have left home. For example, when asked about his roots, Google CEO Sundar Pichai answered, “I'm an American citizen but India is deeply within me. So it's a big part of who I am” (Economic Times, 2021). Scholars have found that for immigrants and refugees, attachment is often reflected in high-intensity longing for the homelands they have left (Deutsch, 2005; Scannell & Gifford, 2010), which might translate into a strong desire to return (Riemer, 2000) or to help through investments (Gillespie et al., 1999) or financial remittances (Vaaler, 2011).

Second, immigrant CEOs often possess strong social capital in their homeland countries. Social capital is defined as the resources and benefits that individuals or groups can access through their social networks and relationships (Coleman, 1988). These resources include information, trust, support, cooperation, and other advantages that result from social interactions and connections. Specifically, social capital can be conceptualized into three dimensions: structural, relational, and cognitive (Nahapiet & Ghoshal, 1998; Tsai & Ghoshal, 1998). The structural dimension refers to a network of relationships that provide information and resource access, the relational dimension refers to assets like trust and personal bonds developed through relationships, and the cognitive dimension involves shared meanings and norms among parties. We argue that CEOs' immigrant background creates social capital in their homelands encompassing all three of these dimensions. First, immigrants frequently maintain strong ties with their home countries (Portes, 1998). These enduring connections represent structural social capital that enables immigrant CEOs to tap into extensive networks within their homelands. For example, Satya Nadella, the India-born Microsoft CEO, maintains his connections with the Indian government to actively promote the technological ecosystem in India (Das, 2023; Sunilkumar, 2023). Second, immigrant CEOs can help build trust and collaboration with home-country business leaders, institutions, and government bodies. They are able to do so because they share similar backgrounds and attributes that are beneficial for friendships and relational social capital (Ertug et al., 2022; Soda & Zaheer, 2012). For example, Indian Prime Minister Manmohan Singh claimed Indra Nooyi, the Indian-born CEO of PepsiCo, as “one of us” (Kumar, 2021). Third, immigrants often share cultural and linguistic ties with their homelands, providing a natural basis for rapport and understanding (Putnam, 2000). These commonalities foster smoother communication and shared interpretations, enhancing immigrant CEOs' cognitive social capital in their homelands. Consequently, immigrant CEOs are uniquely positioned to leverage these multifaceted social capital resources in their home countries.

3 HYPOTHESES

3.1 Immigrant CEOs and CSI incidents in their homeland countries

Building on place attachment theory (Giuliani, 2003; Lewicka, 2011) and social capital theory (Nahapiet & Ghoshal, 1998; Portes, 1998), we contend that the appointment of an immigrant CEO leads to a reduction in CSI incidents in the CEO's homeland country because the immigrant CEO's psychological attachment to and social capital in their homeland can reduce the occurrence and disclosure of CSI incidents in this country.

First, immigrant CEOs' psychological attachment to their homeland countries could arouse place-protective behaviors that prevent their MNEs from engaging in socially or environmentally destructive activities in these countries. Most CSI incidents, such as massive amounts of oil being spilled that impact the local community (Nossiter, 2010), workplace injuries and illnesses due to hostile working conditions (Shi et al., 2022), and unsafe products and/or supply chains (Mena et al., 2016), just to name a few, are place-threatening activities. These incidents reflect the extent to which a firm ignores and/or violates stakeholder interests (Freeman, 1984; Hillman & Keim, 2001).

We argue that because immigrant CEOs have psychological attachment to their homelands, they are likely to take various place-protective actions to avoid the occurrence of CSI incidents in these countries. On the one hand, immigrant CEOs may tend to closely monitor subsidiaries' operations and supply-chain activities in their homeland countries to preclude potential violations of stakeholder interests. Given that place-attached individuals are typically strongly motivated to preserve natural resources in their homelands (Cole et al., 2021; Lewicka, 2011), we argue that immigrant CEOs may tend to monitor and prevent their MNEs from building heavily polluting factories in their homeland countries. They may even close or move existing facilities away from their homelands if they cause severe pollution. On the other hand, immigrant CEOs could be committed to providing useful resources and practices to help prevent CSI incidents in their homeland countries. Prior research shows that MNEs' headquarters are crucial as resource providers in promoting foreign subsidiaries' social engagement in host countries (Crilly, 2011). Therefore, immigrant CEOs may enhance their MNEs' provision of resources, such as capital, technology, and expertise, to support place-protective behaviors in their homeland countries. For example, immigrant CEOs can influence their MNEs to invest in improving working conditions and protecting employee rights in their homelands to reduce the likelihood of irresponsible workplace accidents. As an anecdotal example, Safra Catz, the Israeli-born CEO of Oracle, claimed that support for Israel is always among the priorities of Oracle (i24News, 2023). In sum, immigrant CEOs' psychological attachment to their homeland countries often translates into place-protective behaviors through enhanced monitoring and resource provision, which help their MNEs prevent socially irresponsible activities from occurring in these countries.

Second, immigrant CEOs' social capital in their homeland countries enables them to leverage their networks and relationships to reduce the public disclosure of CSI incidents in these countries. Unlike CSR information, which is usually self-disclosed by firms and distributed in CSR reports, CSI information is typically created by external stakeholders and distributed in the media (Barnett, 2014; Hawn, 2021; Kölbel et al., 2017). In particular, the media tends to focus on CSI incidents because audiences generally perceive negative information as far more interesting than positive information (Rozin & Royzman, 2001).

We argue that immigrant CEOs could leverage their social networks and relationships in their homeland countries to mitigate media coverage of their firms' CSI incidents there. Prior research has recognized that the media often serves the interests of powerful elites and corporations and that it shapes public opinion in ways that support these interests (Herman & Chomsky, 2010). In fact, research has found that CEOs can utilize their personal relationships with journalists to shield unfavorable information about their firms from media disclosure (Westphal & Deephouse, 2011). Hence, journalists may refrain from reporting negative news about CEOs and their firms to avoid losing connections with them (Shani & Westphal, 2015). Accordingly, we argue that when a CSI incident occurs, immigrant CEOs could strategically leverage their social capital in their homeland countries to either influence the local media to remain silent about this adverse event or manage the media narrative to alleviate the perceived severity of its potential consequences. Moreover, because the government often serves as a strategic actor that influences media content (Gehlbach & Sonin, 2014), immigrant CEOs could leverage their personal connections with high-ranking government officials to exert pressure on the media, thus diminishing the disclosure of CSI incidents or mitigating the severity of the associated reporting.

In conclusion, immigrant CEOs' psychological attachment to and social capital in their homeland countries could help reduce the occurrence of CSI incidents and CSI incidents being disclosed to the public in these countries. As a result, we argue that appointing an immigrant CEO could reduce a firm's CSI incidents in the immigrant CEO's homeland country. Accordingly, we predict the following:

Hypothesis 1.Compared to MNEs led by nonimmigrant CEOs, MNEs that appoint an immigrant CEO will experience a reduction in CSI incidents disclosed in the immigrant CEO's homeland country.

3.2 The moderating effects of individual, firm, and country characteristics

Our theorizing for the CSI-reduction effect of immigrant CEOs proposed above is based on two underlying mechanisms: (1) immigrant CEOs' psychological attachment to their homelands stimulates place-protective behaviors that reduce the occurrence of CSI incidents and (2) immigrant CEOs' social capital in their homelands enables them to influence related parties to reduce the media disclosure of CSI incidents. We further posit that certain individual-, firm-, and country-level factors could enhance or weaken these two underlying mechanisms and thereby moderate the CSI-reduction effect of immigrant CEOs. Specifically, we examine immigrant CEOs' length of residence in their homelands, their firms' sustainability ratings, and their homeland countries' press freedom as moderating factors at the individual, firm, and country levels, respectively.

3.2.1 Adult immigrant

We argue that immigrant CEOs who spent a longer time in their homeland countries and immigrated as adults, rather than as children, tend to have stronger psychological attachment to and social capital in their countries of origin. First, length of residence is positively associated with place attachment as it enables individuals to incorporate more place-based memories, preferences, and meanings into their self-identity (Hay, 1998; Rollero & De Piccoli, 2010). For example, Google's CEO, Sundar Pichai, moved to the United States as an adult to attend graduate school. He has demonstrated a strong attachment to India by leading Google's efforts to invest in India and initiate projects like “Internet Saathi” to bring internet access to rural India. In contrast, Gap's former CEO, Sonia Syngal, emigrated from India with her family as a child, and there is limited public information regarding Gap's contributions to India during her tenure. Therefore, we argue that immigrant CEOs who lived longer in their homelands (i.e., immigrated as adults rather than as children) have stronger homeland attachment and are thus more likely to take protective behaviors against CSI incidents in their homeland countries.

Second, immigrant CEOs who have spent more time in their homeland countries likely built stronger social capital in those countries. Individuals who immigrated as adults may have formed robust networks in their homeland countries during their formative and early professional years. Moreover, since values and beliefs are often shaped during adolescence and emerging adulthood (Arnett, 2001), immigrants who emigrate after reaching adulthood likely possess values and beliefs similar to those who remain in their homelands. This similarity fosters trust and personal relationships in their homelands. Also, adult immigrants have rich cultural and linguistic familiarity, allowing for better communication and shared meanings with their compatriots. Therefore, we argue that adult immigrants have better social capital in their homeland countries to mitigate media coverage of their firms' CSI incidents there.

Overall, we argue that immigrant CEOs who immigrated out of their homeland countries as adults, rather than as children, tend to establish stronger attachment to and social capital in their homeland countries, thus augmenting their homeland-protective behaviors and leading to a greater reduction in CSI incidents in their home countries. In contrast, immigrant CEOs who left their homeland countries at a very young age (e.g., in early childhood) may lack the attachment and social capital necessary to foster these protective behaviors, making them less likely to take action against irresponsible behaviors in their homelands. Therefore, we propose the following:

Hypothesis 2.The extent to which the appointment of an immigrant CEO leads to a reduction in CSI incidents disclosed in the immigrant CEO's homeland country is stronger if the CEO immigrated as an adult rather than as a child.

3.2.2 Firm sustainability rating

A firm's sustainability rating reflects the extent to which the firm has achieved sustainable development in social and environmental dimensions (Bansal, 2005). We argue that firms' sustainability enhances the impact of immigrant CEOs on reducing homeland CSI incidents. In particular, we propose that superior sustainability performance in firms not only strengthens immigrant CEOs' ability to reduce the occurrence of CSI incidents in their homelands but also reinforces their motivation to utilize social capital in homelands to reduce the media disclosure of CSI incidents there to protect their firms' reputations as sustainability leaders.

First, when firms have higher sustainability ratings, immigrant CEOs are likely better able to engage in place-protecting behaviors to avert CSI incidents in their homeland countries. Research suggests that by leading firms with a strong commitment to sustainable development, CEOs can amass and leverage a wealth of established knowledge, experience, and resources to prevent irresponsible behaviors from occurring (Cai et al., 2020; Lenox & King, 2004). Therefore, when leading firms with better sustainability ratings, immigrant CEOs are likely better equipped with resources and capabilities to convert their psychological attachment into action to curtail CSI incidents in their homeland countries. Specifically, they could better monitor their firms' behaviors in their homeland countries and transfer valuable resources, knowledge, and best practices to deter irresponsible incidents in these countries.

Second, when firms have higher sustainability ratings, immigrant CEOs likely have stronger motivation to alleviate the media disclosure of their firms' CSI incidents. Firms with higher sustainability ratings usually have better reputations for being socially and environmentally sustainable, so they may experience more severe punishments for wrongdoings due to higher stakeholder expectations (Rhee & Haunschild, 2006). Indeed, research suggests that stakeholders often raise their expectations for the actions of high-reputation firms and hold these firms to higher standards (Mishina et al., 2010). Therefore, MNEs with strong sustainability performance may sustain more significant reputational damage from getting involved in CSI incidents than MNEs with weak sustainability performance. To avoid reputational damage, immigrant CEOs whose firms have strong sustainability reputations are strongly motivated to avoid involvement in CSI incidents and to utilize their social capital in managing the media in the face of CSI occurrences.

In summary, we argue that when MNEs have higher sustainability ratings, immigrant CEOs exhibit stronger competence in curbing irresponsible actions and demonstrate greater motivation to reduce the media disclosure of CSI incidents to protect their firms' reputations. In contrast, when MNEs have low sustainability ratings, immigrant CEOs may not have the ability or essential resources to protect their homelands and may lack the motivation to reduce CSI disclosure due to their firms' poor sustainability reputations. As a result, we argue that firms' sustainability ratings increase immigrant CEOs' ability and motivation to reduce CSI occurrences and disclosure in their homeland countries, thereby enhancing the effect of appointing an immigrant CEO on the reduction in CSI incidents in the CEO's homeland country. Based on this reasoning, we hypothesize the following:

Hypothesis 3.The extent to which the appointment of an immigrant CEO leads to a reduction in CSI incidents disclosed in the immigrant CEO's homeland country is stronger if the focal firm has a higher sustainability rating.

3.2.3 Homeland-country press freedom

A country's press freedom reflects the extent to which local stakeholders can independently express and spread information and perspectives (Wang & Li, 2019). A free press makes it easier for stakeholders to access relevant information about firms and makes it more difficult for firms to conceal or suppress negative information (Oh et al., 2020). We argue that in homeland countries where press freedom is more strongly protected, the negative effect of immigrant CEOs on CSI incident disclosure becomes weaker because immigrant CEOs are less likely to leverage their social connections to influence media coverage about their firms' CSI incidents in these countries.

First, media outlets in countries with strong press freedom operate with editorial independence (El Ghoul et al., 2019), thereby allowing them to freely report CSI incidents without external interference or constraints. When journalists have the autonomy to determine their coverage based on journalistic standards and the public interest, they are not swayed by outside pressures from powerful corporations and the government, which safeguards the integrity of their reporting. Therefore, in homeland countries with high press freedom, immigrant CEOs' attempts to manipulate media content through their connections with related parties (e.g., the media and government) are likely to be met with resistance and are unlikely to reduce the media disclosure of their firms' irresponsible behaviors.

Moreover, countries with higher press freedom often have well-established legal protections and safeguards for whistleblowers, encouraging individuals with insider knowledge of corporate misconduct to come forward without fear of retaliation. These protections ensure that even if firms try to suppress negative information through immigrant CEOs' social networks, whistleblowers can still share their insights with journalists, making it difficult for firms to control the media narratives of their wrongdoings (Bertoni, 2022). Journalists can then investigate and report on these revelations independently, reinforcing the media's role as a watchdog and providing a check on corporate misconduct (United Nations, 2023).

Taken together, high press freedom in their homeland countries makes it difficult for immigrant CEOs to influence the media narratives surrounding their firms' wrongdoings in these countries. Conversely, in countries with low press freedom, immigrant CEOs may find it more feasible to leverage their social capital to pressure the media to suppress unfavorable information about their firms' irresponsible incidents due to a lack of independent journalistic practices and regulatory oversight. Therefore, immigrant CEOs' social capital in their homelands is less likely to influence media coverage and reduce the disclosure of CSI incidents when their homeland countries have higher press freedom. Accordingly, we propose the following:

Hypothesis 4.The extent to which the appointment of an immigrant CEO leads to a reduction in CSI incidents disclosed in the immigrant CEO's homeland country is weaker if the homeland country has higher press freedom.

4 METHODS

4.1 Data

We tested our hypotheses using a sample of MNEs from the S&P 500 headquartered in the United States between 2007 and 2020. We focused on US firms because the United States is the largest destination country for international migrants (McAuliffe & Triandafyllidou, 2021). We focused on S&P 500 firms because they are the largest MNEs listed on stock exchanges and have the most public information about their CEOs and firm behaviors. In this study, we define MNEs as firms that had at least one foreign subsidiary in the period covered in our dataset.

Because no database capturing CEOs' immigrant background exists, we manually collected data from a diverse set of sources. Specifically, we used the Marquis Who's Who and Notable Names database, as suggested in earlier research (Cronqvist & Yu, 2017). We also searched public sources, such as news outlets, media interviews, Wikipedia, and company websites to find CEOs' biographical information. As a result, we were able to compile a comprehensive dataset with the following information: (1) whether a CEO is a first-generation immigrant2 in the United States, (2) the country of origin (i.e., homeland country) of an immigrant CEO, and (3) the age at which an immigrant CEO moved to the United States.

We collected data on MNEs' CSI incidents in each country from the RepRisk database,3 which has been widely used in CSI research (Hawn, 2021; Kölbel et al., 2017; Li & Wu, 2020). The RepRisk database records the countries where firms' ESG-related risk incidents occurred from 2007 to now, allowing us to calculate the number of CSI incidents disclosed in each country. Moreover, we collected firm-level data from other sources, including the BoardEx, ASSET4, Compustat, RavenPack, and FactSet Revere databases. We further compiled country-level information from multiple sources, including the World Bank, the IMD World Competitiveness Yearbook, and so forth.

4.2 Propensity score matching and DID analysis

To establish a causal effect of the appointment of an immigrant CEO on the reduction in CSI incidents in the immigrant CEO's homeland country, we conducted propensity score matching and DID analysis. The treated firms are firms that appointed an immigrant CEO to replace a nonimmigrant CEO in a given year. The control pool includes firms with nonimmigrant CEOs throughout the sample period. To estimate propensity scores, we conducted a probit regression in which the dependent variable equals 1 if a firm appointed an immigrant CEO in a year and 0 otherwise. In the probit regression, we controlled for variables at the firm and firm-country levels that may affect the disclosure of CSI incidents. At the firm level, the matching variables include firm size (the natural logarithm of firm total assets), firm performance (return on assets), degree of internationalization (the ratio of foreign sales divided by total sales), percentage of foreign directors (the ratio of foreign directors on the board), and media sentiment (the average sentiment scores of all news about a firm, with a higher score indicating a more positive media tone toward this firm). These variables ensured that the treated and control firms are similar on some confounding factors that could influence both the appointment of an immigrant CEO (Greve et al., 2015) and the disclosure of CSI incidents. At the firm-country level, we controlled for the number of subsidiaries in the homeland country (the number of an MNE's foreign subsidiaries in the treated firm's homeland country) and the percentage of sales from the homeland country (the percentage of an MNE's sales revenue derived from the treated firm's homeland country)4 to ensure that the treated and control firms have comparable business activities and exposure to the treated firm's homeland country, which could also influence CSI incidents in this country. All matching variables were measured in the year prior to the appointment of an immigrant CEO.

After calculating the propensity scores based on probit regression, we conducted one-to-three matching to identify three control firms with the closest propensity scores to each treatment firm in the same homeland country, Standard Industrial Classification two-digit industry, and year. To ensure proper balance, we required the control firm's propensity score to differ from the treated firm's score by less than 1%. After requiring no missing values in the variables used in our analysis, our final sample included 220 control firms matched to 76 treated firms. The treated firms appointed 76 first-generation immigrant CEOs to succeed nonimmigrant CEOs during 2007–2020. These immigrant CEOs originally came from 21 countries. We report the distribution of these immigrant CEOs' homelands in Table 1.

TABLE 1. Distribution of immigrant CEOs' homelands.

No.	Homeland	Freq.	%	No.	Homeland	Freq.	%
1	Argentina	3	3.95	12	Malaysia	3	3.95
2	Australia	6	7.89	13	Mexico	2	2.63
3	Belgium	2	2.63	14	Netherlands	1	1.32
4	Brazil	2	2.63	15	Poland	1	1.32
5	Canada	6	7.89	16	Russia	1	1.32
6	Colombia	1	1.32	17	South Africa	2	2.63
7	France	6	7.89	18	Spain	1	1.32
8	Germany	7	9.21	19	Sweden	2	2.63
9	India	13	17.11	20	United Kingdom	11	14.47
10	Israel	1	1.32	21	Venezuela	1	1.32
11	Italy	4	5.26		Total	76	100

To illustrate the similarity between the treated and control firms, we present the descriptive statistics of the seven matching variables in Table 2. The first four columns in Table 2 show that the treated and control firms balance on all matching variables in terms of means, minimums, 25th percentiles, 50th percentiles, 75th percentiles, and maximums. We also report the p-value of the difference-in-means test (t test) and Kolmogorov–Smirnov test in the last two columns of Table 2, which shows that the null of equal means (with p-values ranging from .137 to .784) and the null of equal distributions (p-values from .080 to 1.000) cannot be rejected at a 95% confidence level, demonstrating that the treated and control firms are very similar in all matching variables.

TABLE 2. Summary statistics for the treated and matched control firms.

Matching variables	Group	Observations	Mean	Min.	25th percentile	50th percentile	75th percentile	Max.	p-Value (t test)	p-Value (KS test)
Firm size	Treated	76	9.848	5.130	9.031	9.766	10.946	13.601	.650	.582
	Control	220	9.763	5.150	8.930	9.683	10.651	13.722
Firm performance	Treated	76	0.145	−0.507	0.098	0.146	0.203	0.377	.415	.182
	Control	220	0.156	−0.092	0.092	0.144	0.226	0.571
Degree of internationalization	Treated	76	46.081	0.000	31.090	51.477	61.977	98.199	.199	.101
	Control	220	42.133	0.000	26.950	44.711	58.743	94.868
Percentage of foreign directors	Treated	76	23.855	0.000	12.000	22.000	37.000	77.000	.137	.080
	Control	220	20.363	0.000	10.000	20.000	30.000	70.000
Media sentiment	Treated	76	50.348	48.226	50.208	50.365	50.585	51.775	.715	.788
	Control	220	50.374	48.109	50.203	50.343	50.651	54.000
Number of subsidiaries in the homeland country	Treated	76	3.408	0.000	0.000	1.000	4.000	43.000	.784	1.000
	Control	220	3.700	0.000	0.000	1.000	3.000	79.000
Percentage of sales from the homeland country	Treated	76	2.221	0.000	0.281	1.252	2.378	38.922	.486	.266
	Control	220	1.845	0.000	0.000	0.827	2.026	32.484

4.3 Variables and measurements

4.3.1 Dependent variable

Our dependent variable—CSI incidents in the homeland country—is measured as the number of a focal firm's adverse ESG incidents publicly disclosed in the homeland country in a given year weighted by the severity of the incidents. For a treatment firm, the “homeland country” refers to its immigrant CEO's homeland country. For a control firm, the “homeland country” refers to its matched treatment firm's immigrant CEO's homeland country. RepRisk classifies the severity of incidents into three levels—high, medium, and low5—and reports the number of incidents at each level.6 In line with earlier research (Li & Wu, 2020; Wang & Li, 2019), we set the numeric values for the high-, medium-, and low-severity levels as 3, 2, and 1 and calculated the scores of CSI incidents in the homeland country as the sum of the number of CSI incidents in the homeland country at each severity level multiplied by its corresponding numeric value.7 The use of severity-weighted CSI incidents aligns with our theory because CEOs tend to pay greater attention to preventing incidents with more severe consequences. Therefore, more severe CSI incidents should carry more weight than less severe incidents.

4.3.2 Independent variable

We created a treatment variable and a period variable for our DID analysis. The treatment variable—treatment—is a dummy variable coded 1 for treated firms and 0 for control firms. The period variable—post-appointment period—is also a dummy variable coded 1 for 4 years after the appointment of an immigrant CEO and 0 for 4 years prior to the appointment. Our independent variable is measured as the interaction of treatment and post-appointment period.

4.3.3 Moderators

The first moderator—adult immigrant—is a dummy variable coded 1 if an immigrant CEO moved to the United States after the age of 18 and 0 otherwise. According to family law in the United States, a person is generally considered a legal adult at the age of 18. We searched for CEO biographies from a variety of sources, including company websites, company announcements, media interviews, Wikipedia, and other websites, and calculated the age when an immigrant CEO moved to the United States. Typically, an immigrant CEO is considered an adult immigrant if they moved to the United States after getting a bachelor's degree in a foreign country. In total, 76.3% of the immigrant CEOs in our sample moved to the United States as adults.

We measured the second moderator—firm sustainability rating—using a global ESG dataset from Thomson Reuters ASSET4, which collects data and scores firms on ESG dimensions. In line with earlier research (Ioannou & Serafeim, 2012), we measured firm sustainability rating as the average of the social and environmental sustainability scores developed by ASSET4 for a focal MNE in a given year. The social score reflects an MNE's social rating in employment quality, human rights, community, and product responsibility. The environmental score captures an MNE's environmental performance in terms of resource reduction, emission reduction, and product development. Both social and environmental scores are important indicators of CEOs' ability and motivation to handle sustainability issues.

We measured the third moderator—homeland-country press freedom—using the country-level press freedom index derived from annual data produced by Reporters without Borders. Specifically, a treated firm's homeland-country press freedom refers to the press freedom in its immigrant CEO's homeland country, and a control firm's homeland-country press freedom refers to the press freedom in its matched treated firm's immigrant CEO's homeland country. This index is compiled from data on abuses and acts of violence against journalists to reflect a country's press freedom level (Short et al., 2016; Wang & Li, 2019). We scaled this index by dividing it by 100 so the final measure ranges from 0 to 1, with a larger score indicating greater freedom of the press.

4.3.4 Control variables

We controlled for various factors at the CEO, firm, firm-country, and country levels that could affect a firm's CSI incidents in a country. At the CEO level, following earlier research (Bertrand et al., 2021; Tang et al., 2015), we controlled for CEO characteristics, including CEO age, CEO tenure (the number of years a CEO was in the role), and female CEO (coded 1 for female CEOs and 0 for male CEOs), all of which could influence a firm's corporate social performance. We further controlled for CEO duality (coded 1 if a CEO was also a chairman of the board and 0 if not) because a CEO's structural power might influence their discretion and, ultimately, their firm's strategic behaviors (Finkelstein, 1992). In line with prior research (Lewis et al., 2014), we also controlled for CEO MBA degree (coded 1 if a CEO had an MBA degree and 0 otherwise) because CEOs' educational background could influence their firms' social and environmental performance.

At the firm level, we controlled for board characteristics, including board independence (the ratio of independent directors to the total number of directors on a board) and the ratio of foreign directors, because these factors might influence a board's monitoring and advising functions, which may in turn influence CSI incidents (Lara et al., 2017; Miletkov et al., 2017). We also controlled for firm size, measured as the natural logarithm of a firm's annual total assets, as larger firms may have greater exposure to CSI risks (Li & Wu, 2020). Moreover, we included media sentiment, measured as the average sentiment scores of all news about a firm in a year using data obtained from RavenPack, with a higher value indicating a more positive tone toward a firm (Shi et al., 2018).

At the firm-country level, we controlled for a firm's CSI incidents in non-homeland countries to alleviate the concern that immigrant CEOs pay more attention to CSI issues than nonimmigrant CEOs in general. We also controlled for a firm's number of directors from the homeland country because such directors may have specific resources to help and monitor a firm when operating in the homeland country. Since a firm's business operations in a given country might influence its CSI incidents in this country, we controlled for a firm's number of subsidiaries in the homeland country to account for a firm's operations and experience in a focal country.

At the country level, since a country's institutional development can influence corporate social performance (Ioannou & Serafeim, 2012), we controlled for institutional quality using the World Bank's Worldwide Governance Indicators data, which has been widely used to measure the quality of countries' institutions (Abdi & Aulakh, 2012; Stephan et al., 2015). We also controlled for a country's economic indicators, including country size (log-transformed total population) and GDP growth (annual percentage), because market conditions can influence a firm's decision to engage in socially responsible behaviors (Bu & Wagner, 2016). Moreover, we controlled for a country's sustainable development using the sustainable development scores from the IMD World Competitiveness Index because a country's overall sustainable development may affect the occurrence of firms' socially irresponsible behaviors.

4.4 Estimation methods

We conducted a firm fixed effects DID analysis to estimate the impact of the treatment—appointing an immigrant CEO—on the change in CSI incidents in the post-appointment period in the CEO's homeland country while controlling for time-invariant firm heterogeneity. The following equation specifies our model:

$$y_{i,j,t}=b_0+b_1{\mathit{\text{treatment}}}_i\times \mathit{\text{post}}-{\mathit{\text{appointment period}}}_{i,t}+{\beta }_1\bullet X_{i,t-1}+{\beta }_2\bullet X_{j,t-1}+{\beta }_3\bullet X_{i,j,t-1}+a_i+a_t+{\epsilon }_{i,j,t},$$

where $y_{i,j,t}$ is the dependent variable (i.e., CSI incidents in the homeland country) for firm i in country j in year t; b₀ is the constant term; the coefficients b₁ corresponds to the predicted effects in Hypothesis 1; ${\beta }_1\bullet X_{i,t-1}$, ${\beta }_2\bullet X_{j,t-1}$, and ${\beta }_3\bullet X_{i,j,t-1}$ are the vectors of control variables at the firm, country, and firm-country levels, respectively, and their coefficients; and $a_i$, $a_t$, and ${\epsilon }_{i,j,t}$ represent the firm fixed effects, year fixed effects, and the error term, respectively. Because the specification includes firm- and year-fixed effects, it was not necessary to include the non-interacted treatment and post-appointment period dummy variables (Chen et al., 2021). We tested all our hypotheses using fixed-effects ordinary least squares regressions. We report the robust standard errors derived from the robust variance estimator (White, 1980), which produces consistent standard errors and yields asymptotically consistent estimates even when the errors are heteroscedastic.

5 RESULTS

Table 3 displays descriptive statistics and data sources for all the variables used in this study. Table 4 shows the Pearson correlation coefficients for these variables. Tables 5 and 6 report the results of the firm fixed effects DID analyses with subgroup analysis and three-way interactions, respectively.

TABLE 3. Summary statistics and data sources.

No.	Variables	Mean	SD	Min	Max	Data source
1	CSI incidents in the homeland country	0.157	0.753	0.000	10.000	RepRisk
2	Treatment	0.268	0.443	0.000	1.000	BoardEx and hand-collected data
3	Post-appointment period	0.415	0.493	0.000	1.000	BoardEx
4	Adult immigrant	0.777	0.416	0.000	1.000	Hand-collected data
5	Firm sustainability rating	0.667	0.270	0.075	0.983	ASSET4
6	Homeland-country press freedom	0.230	0.133	0.000	0.515	Reporters without Borders
7	CEO age	55.746	7.734	28.000	90.000	BoardEx
8	CEO tenure	4.553	4.936	0.000	51.000	BoardEx
9	Female CEO	0.100	0.300	0.000	1.000	BoardEx
10	CEO duality	0.403	0.491	0.000	1.000	BoardEx
11	CEO MBA degree	0.216	0.412	0.000	1.000	BoardEx
12	Board independence	0.927	0.259	0.000	1.000	BoardEx
13	Ratio of foreign directors	20.376	17.959	0.000	80.000	BoardEx
14	Firm size	9.734	1.399	3.939	13.929	Compustat
15	Media sentiment	50.330	0.462	46.990	55.000	RavenPack
16	CSI incidents in non-homeland countries	10.401	26.094	0.000	291.000	RepRisk
17	Number of foreign directors from the homeland country	0.139	0.411	0.000	3.000	BoardEx
18	Number of subsidiaries in the homeland country	3.534	11.268	0.000	294.000	Subsidiary data by WRDS
19	Institutional quality	0.910	0.919	−1.815	1.930	World Bank
20	Country size	18.099	1.406	15.347	21.035	World Bank
21	GDP growth	2.532	3.129	−10.149	25.176	World Bank
22	Country sustainable development	5.842	0.932	3.030	8.390	IMD World Competitiveness Index

• Note: N = 2179.

TABLE 4. Correlation matrix.

No.	Variables	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20	21
1	CSI incidents in the homeland country
2	Treatment	0.04
3	Post-appointment period	0.00	−0.01
4	Adult immigrant	−0.02	−0.01	0.04
5	Firm sustainability rating	0.04	0.01	0.04	−0.06
6	Homeland-country press freedom	0.00	0.00	0.04	−0.12	0.05
7	CEO age	−0.04	0.02	0.02	−0.01	0.12	−0.01
8	CEO tenure	0.04	−0.12	−0.03	−0.03	−0.11	−0.01	0.46
9	Female CEO	−0.04	−0.09	−0.01	−0.01	0.07	−0.03	−0.41	−0.24
10	CEO duality	0.06	−0.18	−0.01	0.07	0.02	−0.09	0.27	0.20	−0.21
11	CEO MBA degree	−0.06	−0.08	−0.01	0.09	−0.02	0.04	−0.28	−0.15	0.42	−0.17
12	Board independence	0.00	−0.01	0.02	0.04	−0.08	−0.05	−0.14	−0.03	0.07	−0.05	0.03
13	Ratio of foreign directors	−0.01	0.12	0.11	−0.05	0.10	−0.02	−0.04	−0.13	0.00	−0.09	−0.10	0.10
14	Firm size	0.20	−0.01	0.08	−0.01	0.32	−0.01	0.17	0.05	0.05	0.08	−0.07	−0.19	0.04
15	Media sentiment	0.06	0.01	0.10	0.09	0.05	0.06	0.01	0.05	−0.01	0.01	0.05	0.05	−0.04	0.00
16	CSI incidents in non-homeland countries	0.61	0.06	0.06	−0.01	0.13	0.02	0.00	0.09	−0.05	0.07	−0.12	−0.06	0.02	0.41	0.07
17	Number of foreign directors from the homeland country	0.05	0.35	0.12	0.00	0.10	−0.08	−0.01	−0.06	−0.08	−0.08	−0.07	0.04	0.36	−0.02	0.01	0.05
18	Number of subsidiaries in the homeland country	0.02	−0.01	0.03	0.02	0.05	−0.12	0.04	−0.07	0.04	0.00	0.00	−0.06	0.04	0.10	−0.03	0.00	0.03
19	Institutional quality	0.02	0.00	0.06	0.06	−0.03	−0.81	0.03	0.02	0.02	0.04	0.00	0.02	0.01	0.04	−0.04	−0.02	0.11	0.16
20	Country size	0.05	0.01	−0.02	0.07	0.09	0.51	−0.05	−0.02	0.00	−0.02	−0.02	−0.04	−0.01	0.03	0.01	0.06	−0.01	−0.03	−0.50
21	GDP growth	0.00	0.01	−0.05	0.05	−0.02	0.28	−0.05	0.02	0.00	−0.03	−0.03	−0.02	0.02	−0.01	−0.03	−0.01	−0.03	−0.02	−0.26	0.37
22	Country sustainable development	−0.05	0.00	0.07	0.13	−0.02	−0.42	0.03	0.01	0.01	−0.01	0.07	−0.05	−0.03	−0.04	0.01	−0.07	0.04	0.07	0.62	−0.37	0.00

• Note: N = 2179. Correlation coefficients with absolute values greater than or equal to 0.05 are statistically significant at p < .05.

TABLE 5. Firm fixed effects DID analysis with subgroup analysis.

	Dependent variable (DV): CSI incidents in the homeland country
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6	Model 7
Variables	Full sample	Adult immigrant		Firm sustainability		Press freedom
		Adult	Child	High	Low	High	Low
Treatment × Post-appointment period	−0.138	−0.149	−0.009	−0.248	−0.071	−0.059	−0.207
	[0.044]	[0.044]	[0.953]	[0.025]	[0.375]	[0.589]	[0.008]
Wald test chi-squared value		6.52		4.55		4.86
p-Value		.011		.033		.028
CEO age	−0.006	−0.002	−0.017	−0.007	−0.009	0.005	−0.014
	[0.374]	[0.702]	[0.416]	[0.200]	[0.320]	[0.481]	[0.126]
CEO tenure	0.006	0.007	0.009	0.019	−0.004	0.008	0.002
	[0.317]	[0.223]	[0.652]	[0.069]	[0.524]	[0.340]	[0.753]
Female CEO	−0.269	−0.344	−0.468	−0.263	−0.713	0.113	−0.428
	[0.348]	[0.359]	[0.065]	[0.474]	[0.183]	[0.503]	[0.254]
CEO duality	0.091	0.041	0.169	0.075	0.136	0.096	0.119
	[0.133]	[0.521]	[0.386]	[0.395]	[0.153]	[0.359]	[0.127]
CEO MBA degree	−0.006	0.003	0.054	0.050	−0.048	0.067	−0.008
	[0.923]	[0.966]	[0.821]	[0.585]	[0.523]	[0.538]	[0.904]
Board independence	0.161	0.029	0.424	0.136	0.269	0.012	0.345
	[0.095]	[0.683]	[0.232]	[0.168]	[0.145]	[0.881]	[0.050]
Ratio of foreign directors	−0.004	−0.001	−0.015	−0.003	−0.005	−0.003	−0.002
	[0.099]	[0.673]	[0.039]	[0.142]	[0.137]	[0.169]	[0.382]
Firm size	0.064	0.075	−0.013	0.079	−0.011	0.043	0.032
	[0.199]	[0.079]	[0.937]	[0.216]	[0.850]	[0.449]	[0.568]
Media sentiment	0.002	−0.014	0.042	−0.050	0.032	−0.055	0.019
	[0.945]	[0.614]	[0.486]	[0.149]	[0.488]	[0.290]	[0.573]
CSI incidents in non-homeland countries	0.015	0.017	0.007	0.020	0.013	0.021	0.007
	[0.001]	[0.000]	[0.091]	[0.027]	[0.000]	[0.000]	[0.014]
Number of foreign directors from the homeland country	−0.013	−0.166	0.307	−0.119	0.351	0.150	−0.124
	[0.898]	[0.085]	[0.172]	[0.229]	[0.038]	[0.423]	[0.149]
Number of subsidiaries in the homeland country	−0.002	−0.000	−0.001	−0.006	−0.001	−0.009	−0.001
	[0.214]	[0.948]	[0.727]	[0.326]	[0.227]	[0.085]	[0.316]
Institutional quality	0.099	0.206	−0.185	0.091	0.353	−0.147	0.243
	[0.422]	[0.088]	[0.556]	[0.523]	[0.070]	[0.364]	[0.084]
Country size	0.293	1.017	2.082	2.395	−1.111	3.236	−0.957
	[0.822]	[0.432]	[0.619]	[0.169]	[0.495]	[0.264]	[0.454]
GDP growth	0.004	0.001	0.009	0.019	−0.007	0.009	−0.002
	[0.505]	[0.767]	[0.759]	[0.127]	[0.232]	[0.398]	[0.748]
Country sustainable development	−0.042	−0.009	−0.010	−0.117	0.016	−0.011	−0.050
	[0.287]	[0.819]	[0.923]	[0.068]	[0.782]	[0.868]	[0.281]
Constant	−5.641	−18.464	−38.088	−41.213	18.209	−58.814	15.755
	[0.808]	[0.430]	[0.607]	[0.197]	[0.519]	[0.273]	[0.464]
Firm fixed effect	YES	YES	YES	YES	YES	YES	YES
Year fixed effect	YES	YES	YES	YES	YES	YES	YES
Observations	2179	1694	485	1122	1057	997	1182
R-squared	.156	.228	.108	.183	.209	.246	.116

• Note: Two-tailed p-values based on robust standard errors are provided in brackets. Models 2–3 test the moderating effect of adult immigrant: Model 2 (3) is the subsample when the treated firms' immigrant CEOs were adult (child) immigrants. Models 4–5 and Models 6–7 test the moderating effect of firm sustainability rating and homeland-country press freedom, respectively. The high/low subgroups are based on the median values of firm sustainability rating and homeland-country press freedom.

TABLE 6. Firm fixed effects DID analysis with three-way interactions.

	DV: CSI incidents in the homeland country
Variables	Model 1	Model 2	Model 3	Model 4
Treatment × Post-appointment period (Post)	−0.133	0.170	−0.146	−0.139
	[0.049]	[0.274]	[0.032]	[0.069]
Post × Adult immigrant		0.023
		[0.611]
Treatment × Post × Adult immigrant		−0.378
		[0.050]
Treatment × Firm sustainability rating			0.272
			[0.231]
Post × Firm sustainability rating			−0.169
			[0.184]
Treatment × Post × Firm sustainability rating			−0.566
			[0.021]
Treatment × Homeland-country press freedom				−0.187
				[0.817]
Post × Homeland-country press freedom				0.019
				[0.637]
Treatment × Post × Homeland-country press freedom				0.973
				[0.078]
Adult immigrant	−0.287	−0.315	−0.289	−0.290
	[0.083]	[0.068]	[0.078]	[0.093]
Firm sustainability rating	0.147	0.169	0.174	0.151
	[0.087]	[0.058]	[0.138]	[0.091]
Homeland-country press freedom	−0.010	−0.010	−0.009	−0.009
	[0.105]	[0.106]	[0.126]	[0.183]
Control variables	YES	YES	YES	YES
Firm fixed effect	YES	YES	YES	YES
Year fixed effect	YES	YES	YES	YES
Observations	2179	2179	2179	2179
R-squared	.160	.163	.166	.162

• Note: Two-tailed p-values based on robust standard errors are provided in brackets. Treatment × Adult immigrant is not included in Model 2 because it is time-invariant. All models include the same control variables as those listed in Table 5. Control variable coefficients are suppressed for brevity.

5.1 Results of hypotheses testing

Hypothesis 1 predicts that the appointment of an immigrant CEO leads to a reduction in CSI incidents in the immigrant CEO's homeland country. In support of Hypothesis 1, Model 1 in Table 5 shows that the coefficient estimate for the interaction between treatment and post-appointment period is negative and significant (β = −.138, p = .044). We plotted the average of CSI incidents in the homeland country for the treatment and control firms over a 9-year period centered around the appointment year. Figure 1 shows two lines that move in parallel in the pre-treatment years. Moreover, Figure 1 shows that the appointment of an immigrant CEO led to a 54.25% decrease in the number of CSI incidents in the 4 years following the appointment, whereas the control firms led by nonimmigrant CEOs experienced a 6.36% increase in the number of CSI incidents during the same post-appointment period. Figure 1 lends further support to Hypothesis 1 by showing that the appointment of an immigrant CEO led to a greater reduction in CSI incidents in the immigrant CEO's homeland country for the treatment firms than for the control firms.

FIGURE 1

CSI incidents in the homeland country around the appointment of an immigrant CEO. This figure depicts the levels of CSI incidents in the homeland country for the treatment and control firms over the 9-year period centered around the appointment year.

The validity of DID analysis requires meeting the parallel trends assumption. To assess whether this assumption was met in our study, we followed prior research (DesJardine et al., 2022; Kim & Kim, 2024) by creating eight dummy variables denoting whether a firm-year observation is 1, 2, 3, or 4 years before or after the appointment year. We then included the interactions of these eight dummies and the treatment variable and presented the corresponding estimated coefficients for these eight interactions in Figure 2, which shows that none of the interaction coefficients are statistically significant during the pre-appointment period, thereby supporting the DID assumption of parallel trends.

FIGURE 2

Dynamic treatment effects of the appointment of an immigrant CEO. This figure presents the coefficient estimates and 95% confidence intervals for the interaction terms between treatment and eight lead/lag indicators for time periods from t = −4 to t = 4 (t ≠ 0) in the years around the appointment of an immigrant CEO.

Hypotheses 2–4 propose the moderating effects of adult immigrant, firm sustainability rating, and homeland-country press freedom, respectively. We conducted subgroup analyses to test the moderating effects, and we report the results in Models 2–7 in Table 5. Specifically, Models 2–3 in Table 5 test Hypothesis 2, which posits that the extent to which the appointment of an immigrant CEO leads to a reduction in CSI incidents in the immigrant CEO's homeland country is stronger if the immigrant CEO is an adult immigrant. Corroborating Hypothesis 2, the coefficient estimate for the interaction between treatment and post-appointment period is negative and significant for the subsample covering the treated firms with immigrant CEOs who were adult immigrants and their matched control firms (β = −.149, p = .044), while the coefficient for the interaction term is negative but insignificant for the subsample covering the treated firms with immigrant CEOs who were not adult immigrants and their matched control firms (β = −.009, p = .953). Moreover, the absolute value of the negative coefficient in the subsample with adult immigrant CEOs is larger than that in the subsample with nonadult immigrant CEOs. A Wald test to compare the coefficient estimates for the interaction term across the two subsamples revealed that the difference is significant (Chi-squared value = 6.52, p = .011). Therefore, Hypothesis 2 is supported.

Models 4–5 in Table 5 test Hypothesis 3, which predicts that the extent to which the appointment of an immigrant CEO reduces CSI incidents in the immigrant CEO's homeland country is stronger when the focal firm has a higher sustainability rating. We divided the sample into two subsamples based on the median value of firm sustainability rating. As shown in Models 4–5 in Table 5, the coefficient estimate for the interaction between treatment and post-appointment period is negative and significant in the high firm sustainability rating subgroup (β = −.248, p = .025), whereas the coefficient for the interaction term is negative but insignificant in the low firm sustainability rating subgroup (β = −.071, p = .375). We further conducted a Wald test to compare the coefficient estimates for the interaction term across the two subsamples, which revealed that the absolute value of the coefficient in the high sustainability rating subsample is significantly larger than that in the low sustainability rating subgroup (Chi-squared value = 4.55, p = .033). Therefore, Hypothesis 3 is supported.

Models 6–7 in Table 5 test Hypothesis 4, which proposes that the extent to which the appointment of an immigrant CEO leads to a reduction in CSI incidents in the immigrant CEO's homeland country is weaker when the homeland country has higher press freedom. We divided the sample into two subsamples based on the median value of homeland-country press freedom. Models 6–7 in Table 5 show that the coefficient estimate for the interaction between treatment and post-appointment period is negative and significant in the low press freedom subgroup (β = −.207, p = .008), whereas the coefficient of the interaction term is negative but insignificant in the high press freedom subgroup (β = −.059, p = .589). A Wald test revealed that the absolute value of the coefficient in the high press freedom subsample is significantly lower than that in the low press freedom subgroup (Chi-squared value = 4.86, p = .028), thereby supporting Hypothesis 4.

In addition to subgroup analyses, we conducted three-way interactions to test the moderating hypotheses. As shown in Table 6, we found a negative and significant coefficient for the three-way interaction of treatment × post × adult immigrant (β = −.378, p = .050) and for the three-way interaction of treatment × post × firm sustainability rating (β = −.566, p = .021). We also found a positive and significant coefficient for the three-way interaction of treatment × post × homeland-country press freedom (β = .973, p = .078). These results are consistent with the results of the subgroup analyses and lend further support to the three moderating hypotheses.

5.2 Robustness checks and additional analyses

5.2.1 Running bivariate probit regression with partial observability

Although our theory argues that immigrant CEOs' homeland attachment reduces the occurrence of CSI incidents and that their social capital reduces the disclosure of these incidents, empirically, we could only capture CSI incidents that occurred and were disclosed by the media. We addressed this issue methodologically by using bivariate probit regressions with partial observability, following prior research (Shi et al., 2017; Wang, 2013). This method allowed us to model the occurrence of CSI and the disclosure of CSI simultaneously, thus mitigating the concern that some CSI behaviors that occurred were not disclosed. Bivariate probit regressions use two dependent variables: in our case, C_ij is a dummy variable that equals 1 if firm i tends to commit CSI incidents in homeland country j and equals 0 otherwise, and D_ij is a dummy variable that equals 1 if firm i is likely to be caught conditional on the occurrence of its CSI behaviors in homeland country j and equals 0 otherwise. Since we did not directly observe the realization of C_ij and D_ij, we observed Z_ij = C_ij × D_ij, where Z_ij equals 1 if firm i has committed CSI behaviors and has been disclosed and equals 0 if firm i has not committed CSI behaviors or has committed CSI behaviors but has not been disclosed. Following earlier research on bivariate probit regressions with partial observability (Feinstein, 1990; Poirier, 1980), we adopted different variables predicting C_ij and D_ij, respectively. We report the results in online Appendix A, which shows that the interaction between treatment and post-appointment period is negative and significant when predicting P(C), the likelihood of CSI commitment (β = −.538, p = .095), and P(D|C), the likelihood of CSI disclosure (conditional on commitment) (β = −.479, p = .049). These results support our arguments that the appointment of immigrant CEOs not only reduces the occurrence but also the disclosure of CSI incidents in their homeland countries.

5.2.2 Testing different time windows

We used alternative time windows—3 and 5 years prior to and after the appointment year (instead of 4 years, as used in the main analysis)—to test our hypotheses. The results using 3- and 5-year windows are largely consistent with our main results based on the 4-year window. We report the results in online Appendix B.

5.2.3 Examining MNEs' CSI incidents in other host countries

Our theoretical arguments imply that immigrant CEOs tend to reduce CSI incidents in their homeland countries; therefore, this CSI-reduction effect should not exist in other host countries that are not their homelands. To test this argument, we selected the United Kingdom and Canada as two host countries because they are the top two host countries with the most CSI incidents reported in RepRisk; we then examined whether immigrant CEOs who are not from these two countries tend to reduce CSI incidents there. We adopted two new dependent variables—CSI incidents in the United Kingdom and CSI incidents in Canada—and dropped CEOs from these two countries. Specifically, when CSI incidents in the United Kingdom (Canada) is used as the dependent variable, the treated firms are firms that appointed immigrant CEOs who are not from the United Kingdom (Canada), and the control firms are firms with nonimmigrant CEOs. In line with our prediction, the interaction between treatment and post-appointment period is not significantly related to CSI incidents in the United Kingdom (β = −.086, p = .280) or Canada (β = −.058, p = .461), as shown in online Appendix C, corroborating our argument that immigrant CEOs are important in reducing MNEs' CSI incidents in their homeland countries but not in other host countries.

5.2.4 Comparing different types of CSI incidents

RepRisk classifies CSI incidents into four types: environmental issues, social issues, governance issues, and cross-cutting issues. We created four new dependent variables to measure the number of CSI incidents in an immigrant CEO's homeland country weighted by the severity level of each of these four types. The results in online Appendix D show that the interaction of treatment and post-appointment period has a negative and significant effect on homeland CSI incidents for environmental, social, and cross-cutting issues but not for governance issues. The lack of significant results for governance issues is likely attributable to the fact that our sample firms are the largest public firms in the United States and uniformly adhere to the same governance standards.

6 DISCUSSION AND CONCLUSION

6.1 Theoretical contributions

This article offers several theoretical contributions. First, our study contributes to the emerging research stream on geographical variations in CSI (Husted, Jamali, and Saffar, 2016a; Surroca et al., 2013). Despite the growing discussion regarding MNEs' CSI activities, we lack empirical research comparing CSI across different locations. While previous studies have mainly focused on the influence of macro-level institutional factors on the locations of CSI (Campbell, 2007; Keig et al., 2015; Sun et al., 2021), microlevel individual driving forces, such as CEO characteristics, remain largely underexplored (Crilly et al., 2008). Our study expands this line of research by providing a novel perspective drawn from the environmental psychology and social capital literature that links immigrant CEOs' place-oriented psychological attachment and social capital—which capture microlevel individual differences—to their firms' CSI in their homeland countries. It is worth noting that our focus on CEO-level driving factors is not meant to discount firm- and institutional-level factors but instead provides a complementary view to the existing literature. By doing so, our study answers the calls for a more nuanced understanding of the microfoundations of corporate social performance (Crilly et al., 2008; Shea & Hawn, 2019) and an exploration of the locations of CSI (Surroca et al., 2013).

Second, we contribute to research on how CEO characteristics influence corporate social performance by investigating CEOs' geography-related attributes (i.e., immigrant CEOs' homeland countries), thereby extending existing research on CEOs' non-geography-related characteristics (e.g., education, gender, hubris, political ideology, etc.) (Chin et al., 2013; Cronqvist & Yu, 2017; Lewis et al., 2014; Tang et al., 2015). Studying CEOs' geography-related attributes provides a more nuanced understanding of the impact of CEOs on the geography of CSI incidents rather than on overall CSI incidents, as studied in earlier research. Prior research on CEOs' geography-related attributes remains limited and has focused on the role of local CEOs in their hometowns, suggesting that CEOs are more socially responsible when leading firms headquartered in their hometowns due to hometown identity (Lai et al., 2020; Ren et al., 2023). However, it remains unclear whether CEOs who have left their homelands, settled in a new home, and are leading firms headquartered in foreign countries still maintain psychological bonds and identification with their home countries. This question is crucial since migration often reshapes CEOs' values and decision-making processes (Le & Kroll, 2017; Morgan et al., 2020). Our findings highlight that CEOs who migrated from their homelands continue to exhibit a sense of commitment and responsibility toward their homeland countries.

Finally, we contribute to the research on immigrant CEOs by introducing psychological attachment—and social capital-based mechanisms that influence firms' strategic behaviors. Prior research has adopted social identity theory and provided a rich discussion regarding how immigrant CEOs suffer from outsider identity in their destination countries (Mata & Alves, 2018) and thus need to build local legitimacy through socially responsible practices (Bertrand et al., 2021). Our study stresses that this discussion needs to be extended because the outsider identity logic falls short in explaining how immigrant CEOs behave in their homeland countries where they are not considered “outsiders.” To advance our understanding in this regard, our study adopts two fresh mechanisms based on psychological attachment and social capital to explain how immigrant CEOs affect CSI incidents in their homelands. As one of the first studies to examine the influence of immigrant CEOs beyond their destination countries, our study provides a more complete understanding of the impact of immigrant CEOs. While outgroup identity theory is crucial for explaining immigrant CEOs' behaviors in their destination countries, our mechanisms based on place attachment and social capital are particularly relevant for explaining their behaviors in their homeland countries.

6.2 Managerial and practical implications

Our study offers managerial and practical implications for both firms and policymakers. For firms, our findings provide fresh and valuable insights for CEO selection and monitoring. By understanding how CEOs' immigrant background affects CSI, firms can make more informed decisions to improve and monitor their reputations in different host countries. The impact of CEOs' immigrant background should be considered when evaluating CEO candidates in addition to other qualities like being entrepreneurial and innovative (Chaganti et al., 2008). Appointing an immigrant CEO is particularly effective in reducing CSI incidents in the CEO's homeland when the CEO has spent a longer time there, when their firm has better sustainability performance, and when their homeland country has lower press freedom. Moreover, a useful takeaway from our study is that appointing an immigrant CEO will not increase the CSI-related risk exposed to non-homeland countries, alleviating the concern that immigrant CEOs may transfer socially irresponsible activities from their homelands to other countries.

For policymakers, our paper uncovers the possibility that CEOs may leverage their social connections to influence media coverage and conceal irresponsible practices. This finding highlights the need for more regulatory attention to maintaining press freedom in countries to prevent firm executives' intervention in media disclosures. Policymakers can enforce laws to enhance press freedom and protect media transparency. Such enforcement involves ensuring a safe environment for journalists, whistleblowers, and anticorruption activists from threats that may arise as a result of their efforts to reveal corruption (United Nations, 2018). Moreover, our study provides evidence that CEOs' place attachment is an important mechanism that can curb firms' socially irresponsible behaviors. Accordingly, we suggest that in addition to formulating and implementing effective policies, local governments can create opportunities to enhance firm executives' emotional bonding and psychological attachment to their countries, which can act as an alternative mechanism to prevent CSI activities. Therefore, host governments should consider actively creating opportunities to make their countries a “meaningful place” for executives of MNEs.

6.3 Limitations and future research

This study has limitations that open avenues for future research. First, although we built on theories of place attachment and social capital to develop our theoretical arguments, the archival nature of our data did not allow us to completely unpack the micro-processes of the psychological and relational bonds between immigrant CEOs and their homeland countries. Future research can collect finer-grained data via interviews and surveys to provide more substantial evidence of the psychological and relational mechanisms underlying executives' place attachment and social capital.

Second, while our study focuses on immigrant CEOs' own efforts driven by homeland attachment and social capital mechanisms, it is possible that homeland subsidiaries and employees also contribute to reducing CSI incidents in their respective countries. For example, homeland employees may react to the appointment of an immigrant CEO from their home country by paying more attention to their actions and avoiding mistakes that could lead to CSI incidents. In other words, the CSI-reduction effect may result from the joint efforts of both immigrant CEOs and the homeland subsidiaries. However, our data is limited in teasing out the CEO level vs. the homeland subsidiary level mechanisms. We encourage future studies to examine whether homeland subsidiaries and employees react to the appointment of immigrant CEOs from their country, which has received little attention in strategy research.

Third, while we examined the impacts of immigrant CEOs on their firms' CSI incidents, other firm strategies may also be influenced by CEOs' immigrant background. Prior research shows that immigrants may choose to contribute to their homelands in various ways, such as making direct investments (Gillespie et al., 1999) and fostering entrepreneurial activities (Balachandran & Hernandez, 2021). We encourage future research to explore the impacts of immigrant CEOs on their firms' behaviors and even their peer firms' behaviors in their homeland countries.

ACKNOWLEDGMENTS

The authors greatly appreciate the excellent editorial guidance by Professor Olga Hawn and two anonymous reviewers. Juan Bu acknowledges the financial support provided by the Rodney Merritt Miller Faculty Fellowship; Stephanie Wang and Dan Li are grateful for the support from the Samuel & Pauline Glaubinger Professorship and the L. Leslie Waters Chairship.

Open Research

DATA AVAILABILITY STATEMENT

Research data are not shared.