Local, yet global: Implications of caste for MNEs and international business

Caste is an informal institution that influences socioeconomic action in many contexts. It is becoming increasingly evident that international business research, practice, and policy need to programmatically address caste. To facilitate this endeavor, we review the limited research in IB that has addressed caste, and theorize caste as a distinct informal institution by distinguishing it from other systems of stratification like race, class, and gender. In addition, we propose a parsimonious framework to highlight the implications of caste for Indian and non-Indian MNEs in their Indian and global operations. In doing this, we focus on implications with respect to the internal organization and inter-organizational relationships of MNEs, and consider how these implications might differ as based on the MNEs’ organizational forms. We then build on these implications to discuss how MNEs and other stakeholders of international business can address caste inequalities via policies related to human rights, anti-racism, and affirmative action. By bridging theory, practice, and policy, we pave the way for MNEs to address global inequalities that relate to caste.


INTRODUCTION
As emerging economies become engines of global economic growth, we need to better understand their institutions, and the implications of those institutions for MNEs and international business (Elango & Dhandapani, 2020;Peng et al., 2008). Even though emerging economies have seen developments in their formal institutions, informal institutions are difficult to change and affect MNE operations in important ways (Minbaeva et al., 2022;Waylen, 2014). Underscoring this, IB scholars have examined the effects of informal institutions like guanxi, blat, jietinho and jaan-pehchaan on MNE operations (Chua et al., 2009;Ledeneva, 1998;McCarthy et al., 2012) and highlighted the need to further study such institutions (Dau, Chacar, Lyles, & Li, 2018;Parboteeah & Cullen, 2003).

The online version of this article is available Open Access
Supplementary Information The online version contains supplementary material available at https://doi.org/10.1057/s42214-022-00146-9.
One such informal institution is the caste system (Chrispal et al., 2021). Although caste is commonly associated with India and Hinduism, it is prevalent, in varying degrees and forms, across South Asia and among other religions, including Christianity and Islam. In addition, caste is also found among the South Asian communities who migrated to developed economies (e.g., the U.K., the U.S., Canada, and Australia) (Metcalf & Rolfe, 2010;Shankar, 2018;Zwick-Maitreyi, Soundararajan, Dar, Bheel, & Balakrishnan, 2017) as well as to former colonies as plantation labor (e.g., Fiji, Malaysia, Mauritius, Jamaica, and Trinidad) (Yengde, 2022).
The relevance of caste for socioeconomic action is substantiated by organizational research, which indicates that caste has implications for a range of socioeconomic interactions, including CEO selection (Damaraju & Makhija, 2018), entrepreneurs' network formation intentions (Vissa, 2011), friendships among MBA students (Bhardwaj et al., 2021), venture capitalists' valuations (Claes & Vissa, 2020), and microfinance lending (Patel et al., 2020). Therefore, studying caste is important for IB scholarship given (i) the growing significance of caste to MNEs around the world (Zwick-Maitreyi et al., 2018), (ii) increasing attention by IB scholars to societal economic inequalities (Doh, 2019;Nachum, 2021;Rygh, 2019), and (iii) the practical relevance of caste to operate in South Asia and achieve inclusive growth (Alamgir et al., 2022). However, IB scholarship has yet to produce a stream of work with a clear focus on the implications of caste to MNE research and practice and IB policy.
In view of the above, we ask: what are the implications of caste to MNEs and IB policy? To answer this question, we first outline the importance of studying caste and take stock of the state of knowledge, by presenting a brief review of the limited IB research that has addressed caste. Next, we theorize caste as an informal institution by distinguishing it from class, race, and genderconstructs with which IB scholars are more familiar. We then draw on our review as well as broader organizational research that has considered caste, i.e., work outside of IB research, to outline the implications of caste to MNEs. We conclude with a discussion on how MNEs can engage with macro policies on human rights, anti-racism, and affirmative action and research opportunities therewith.

Importance of Studying Caste
As a unique construct that has often been overlooked, caste is important for IB scholarship to study for four reasons. First, recent incidents in the U.S. (e.g., California's case against Cisco in 2020 alleging caste discrimination, litigation by a former HCL employee that his termination was based on caste, the statement by 30 female engineers to Washington Post narrating caste bias in their workplaces, Equality Labs' 2018 study on caste discrimination in the Silicon Valley, and the demand of Alphabet's workers union to include caste as a protected category in the U.S.) underscore the growing significance of caste in MNEs. These incidents also point to the need for a more careful consideration of caste when designing policies against workplace discrimination and as part of efforts to foster equal opportunities -both at the organizational and societal levels. Second, studying caste aligns with the increasing attention paid by IB scholars to rising socioeconomic inequalities (Doh, 2019;Nachum, 2021;Rygh, 2019), the role of firms in reproducing inequalities (Amis et al., 2020;Brandl et al., 2021;van der Straaten et al., 2020), and the implications of such inequalities for MNEs Lupton et al., 2020). By studying caste, insights can be developed about how MNEs can design their policies and practices so that they do not unwittingly reproduce societal inequalities in their own operations. Beyond this, research on caste by IB scholars can also generate insights that are related more broadly to informal institutions and their implications for MNEs and IB.
Third, understanding the implications of caste to MNEs and IB is necessary given the role of India specifically and South Asia more broadly as a key driver of global economic growth and its increasing integration with the global economy. Research in neighboring disciplines has indicated that institutions of inequality, such as caste (e.g., Mosse, 2018) and race (e.g., Pager & Shepherd, 2008), hamper economic growth. Therefore, considering caste in designing IB and MNE policies can help achieve greater economic growth in South Asia, which includes Afghanistan, Bangladesh, Bhutan, India, the Maldives, Nepal, Pakistan, and Sri Lanka (Alamgir et al., 2022).
Finally, as an institution of inequality, caste touches the lives of many South Asians, and limits socioeconomic opportunities for a vast number of them (Bapuji & Chrispal, 2020). For example, at nearly 30% of the population in India, the Dalit (untouchable) and Adivasi (Indigenous) Peoples alone number nearly 400 million and face severe inequalities. In Pakistan, where the biraderi system is practiced, major divisions exist between the Zamindars (landowners), the high castes, and the Kammis (service providers), the low castes who are relegated to low-status and low-income jobs like cleaning (Zulfiqar & Prasad, 2022). The caste system has also been a prominent feature of everyday life in Sri Lanka, where three parallel caste systems exist, one among the Sinhalese, one among the Sri Lankan Tamils, and one among the Indian Tamils (Samarakoon, 2022). For instance, in the Sinhala caste system, there are 15 castes ranging from the Radala (aristocratic caste) and Govigama (landowner castes), to the Rodi (beggar caste) (Silva, Kotikabadde, & Abeywickrama, 2009). Indian Tamils in Sri Lanka are predominantly plantation workers, of which, nearly 83% are Dalits and are often subjected to bonded labor (Balasundaram, Chandrabose, & Sivapragasam, 2009). In Nepal, over 60 caste groups exist, and researchers categorize these into distinct groups, such as ''high castes, Newars, hill ethnic groups, Tamang and other Bhotia (ethnic groups of Tibeto-Burmese origin), Terai castes, and low caste -untouchables'' (Stash & Hannum, 2001: 364). In Bangladesh, dozens of castes exist among Hindus (in a way similar to those found in India) as well as Muslims, which categorize them into three groups: Sharif or Ashraf (i.e., noble born/foreign Muslims), Atraf (low-born), and Ajlaf or Arzal (lowest of all) (Chowdhury, 2009). In brief, across the regions/countries in which it has been observed, the caste system creates inequalities and severely affects those who are placed lower in the hierarchy by providing differential access to resources and opportunities, as well as social rights and privileges based on one's caste position (Bapuji & Chrispal, 2020).

Review Procedure
To examine the discussion of caste in IB research, we reviewed the literature by covering 13 journals from the date of their first issue until July 15, 2022. First, we selected four IB journals that are commonly used in reviews of IB research: Journal of International Business Studies (JIBS), Journal of World Business (JWB), Management International Review (MIR), and Global Strategy Journal (GSJ). In addition to these, we also included Asia Pacific Journal of Management (APJM), a journal for research focused on Asia, as well as Journal of International Business Policy (JIBP), which has emerged as an important outlet for research on IB policy. As general management journals too publish IB research, we also selected seven mainstream management journals: Academy of Management Journal (AMJ), Administrative Science Quarterly (ASQ), Journal of Management (JOM), Journal of Management Studies (JMS), Organization Science (OS), Organization Studies (OST), and Strategic Management Journal (SMJ). To identify IB papers in these mainstream management journals, for each paper that was retrieved, we examined whether the paper addressed an IB topic (e.g., crossnational business practices, MNEs, expatriates, national culture) or had an MNE context. While we applied this filter to papers from mainstream management journals, in our review, we automatically included any paper dealing with caste that was published in the IB journals listed above.
Second, we developed a number of search terms, including ''caste'' and other terms that can be used to refer to caste (community/communities, tribe/ tribal, and Biraderi/Baradari, which refers to the caste system among Muslim communities), broad caste groups/categories and various terms used to refer to them (Other Backward Castes (OBCs)), Scheduled Castes (SCs or communities existing outside of the caste system), untouchables (commonly used to refer to SCs), Dalits (recent label to refer to SCs), Scheduled Tribes (STs or Indigenous tribes of India), Adivasis (recent label to refer to STs), SC/ST (commonly used together to refer to those eligible for constitutionally mandated quotas), and Bahujan (another term to reference lower/marginalized castes, which means ''the majority''), specific caste groups noted in research for their prominence in society (Brahmin/Brahman i.e., scholars, priests; Vaishyas/Vysyas, i.e., traders, merchants, and trading communities), and other hierarchical groups (Ashraf or Ajlaf/Azlaf within Muslim communities).
Third, we conducted a topic search in Web of Science using the selected search terms and limiting our search to the 13 journals listed above. The topic search may miss online first papers as well as papers that have not used the search terms in their title, abstract or keywords. Therefore, we also conducted a full text search in the 13 journals by individually searching the webpages of each journal. As a result of this process, we retrieved 54 papers: 17 in APJM, 13 in JWB, 11 in MIR, 9 in JIBS, 1 paper each in AMJ, OST, GSJ and JIBP, and none in the other journals. In total, the search term ''caste'' on its own identified 51 papers, while ''community/communities'' identified two more papers, and ''tribe/tribal'' identified another paper, with the remaining search terms (i.e., those other than caste, tribe, and community) yielding no additional papers. In the next section, we discuss the insights we generated from our review of these 54 papers.

Review Findings
One of the first aspects that we noted in our review was the limited attention to caste in IB research. Only one paper (Chen et al., 2015) used the word ''caste'' in its title, abstract, or keywords, while another paper (Mani, 2021) used a related word, ''community/communities'' in its title, abstract, or keywords. Only 12 papers used the word caste (or community) five or more times and presented some discussion of caste, such as the effect of caste on HR practices, organizational culture, or business relations. A great majority of papers (41 papers) that mentioned the word ''caste'' offered little discussion about it and used it fewer than five times (the word is mentioned only once in 33 papers, twice in six papers, and three or four times in two papers). In other words, ''caste'' was not a major discussion point in the vast majority of the 54 papers we reviewed.
In some cases, caste was used passingly in the process of making a different or broader point. For example, caste was used in suggesting that dharma in Hinduism, or one's duties and responsibilities, was accommodated under caste and predicated on the social division of labor (Gupta et al., 2002). One paper noted that caste prejudice in India, where it was mentioned as a cultural element, was a challenge that an organization faced during internationalization (Garrido et al., 2021). In another instance, in exhorting IB scholars to examine how history matters, Jones and Khanna (2006) suggested that the relationship between FDI and long-run economic development in India was influenced by colonial history, i.e., how caste/community elites in different treaty ports (e.g., Bombay, Calcutta) interacted with the British. In short, a great majority of IB articles that considered caste have either mentioned it in passing or speculated about the likely implications of caste to MNEs, but did not offer further discussion to elaborate on such implications.
The studies that have examined caste to make a specific point and those that have elaborated on its implications were conducted predominantly in the context of India. We found that these papers provided important insights for international management along the lines of human resource practices, organizational behavior, social networks, and occupational roles. We summarize them in Table 1 and discuss each of them below.
Human resource practices Some scholars have suggested that caste influences human resource practices in organizations (Chatterjee & Pearson, 2000;Crino & White, 1980;McIntyre & Shull, 1979). Specifically, scholars suggested that some managers from caste-based societies might not be open to diversity and may recruit based on caste (Sparrow & Budhwar, 1997) or family relations, which poses a challenge for identifying, acquiring, and using talent (Cooke et al., 2014;Thams et al., 2013). Scholars have also noted that diversity is low in organizations in India because employees remain within small groups with whom they identify, based on caste and family considerations, and are distrustful of outgroup members, which leads to homogeneity within organizational workforce (Varma et al., 2006). Another study found that workers in a particular industrial belt were predominantly from the lower castes, and were subjected to abuse due to their caste position. However, the authors also highlighted that an organization that treated these workers respectfully was able to attract and retain talented workers (Saini & Budhwar, 2008).
In a study of the implications of religious diversity to MNE operations in India, Rao (2012) noted that companies reported that they did not use religion as a basis for any HR decisions. However, she reported that the U.S. headquarters of one MNE Human resource practices Sparrow and Budhwar (1997) Indian managers in the study were caste-conscious, not open to embrace diversity (despite being educated and from an elite class) and preferred recruiting from traditional labor pools. High power distance in India may be a result of caste system (besides family structure and colonial influence) Varma et al. (2006) As a reality of everyday life in India, caste has an effect within the workplace as well. Employees remain within small groups with whom they identify, and are distrustful of others, leading to caste homogeneity and low diversity in the workplace Saini and Budhwar (2008) Nearly all employees in the worker group in the two case study SMEs were from lower classes.
The caste system promotes submission of the low-caste workers to authority. One organization did not practice the 'abuse culture' prevalent in other companies against lower-caste workers, which helped it to attract and retain competent workers Rao (2012) Discussed caste in examining religion as a source of diversity. Inherited division of labor due to the caste system has discriminatory negative consequences, but also helped individuals from the vysya caste to specialize in business and succeed. The study noted that caste had an influence on employees, but its implications to MNEs was unclear. U.S. HQ asked for caste-wise breakdown of recruitment but stopped it on learning that it was not necessary by law and not an accepted practice in India Thams et al. (2013) Social exchanges have a dark side of in-group favoritism, which manifests in hiring and firing decisions based on caste considerations Cooke et al. (2014) Caste system is a barrier to talent management within Indian organizations Organizational behavior Harvey and Buckley (1997) Whether individuals who are socialized within the caste system are content when placed in cultures that do not have similarly rigid social systems Morris et al. (1998) Achievement orientation may be lower in India because there is limited social mobility and one's life outcomes are dependent on their ascribed characteristics i.e., caste Combs and Nadkarni (2005) Compared to American employees, Indian employees (i) prefer opportunity enhancement affirmative action programs to preferential treatment programs and (ii) have a less favorable view of beneficiaries of affirmative action Banai and Reisel (2007) Hierarchical arrangements like caste tend to perpetuate through informal means, despite their formal abolishment. In countries where power distance is accepted, employees are more likely to be averse to supportive leadership and amenable to autocratic leadership Mathew et al. (2012) Caste is an example of a sociological factor that influences the shaping of organizational culture within traditional Indian organizations Dheer et al. (2015) MNEs need to understand the cultural heterogeneity within India. One such subculture to understand is the practices and behaviors of Indigenous tribal groups Ananthram and Chan (2021) A study participant noted that some upper-caste individuals change their last names to lowercaste ones so that they can avail themselves of affirmative action quotas to access medical and engineering education for their children; this is a negative an example of jugaad, ''doing more with less.'' Business networks Prasad (1967) Shared ethno-cultural identity (i.e., caste) has an effect on business transactions and hiring in firms in Africa Kumar and Singh (1978) Interpersonal interactions of Indian managers are based on identification (i.e., categorization of people as 'own' or 'other') and obligation (i.e., a sense of personal indebtedness and responsibility to behave in socially expected ways); caste creates personal obligations with others (i.e., those who are not family or close friends), which in turn influences interactions, as well as how people interpret events and interactions Tsui-Auch (2005) One entrepreneur in Singapore was able to tap into international trading networks created by members of his caste Thams et al. (2013) Social exchanges -exchange of favors and rules governing them -in India are influenced by family and caste membership, similar to guanxi in China and on in Japan Verbeke and Kano (2013) Caste was identified as a group to illustrate in-group exchange of favor Chen et al. (2015) Accuracy of analyst forecasts was higher when they shared caste affiliation with CEOs of the companies they covered. However, this effect was stronger for pre-reform generation CEOs and weaker for post-reform generation CEOs had initially asked for a caste-wise breakdown of recruitment, assuming that recruiting from the lower castes was done in India similar to how minorities were recruited in the U.S., as part of affirmative action. However, they then ''dropped the idea because it was not necessary by law and not an accepted practice in the Indian private sector'' (Rao, 2012: 237). Although the participants and the author attribute this matter to religion rather than the caste system, this study also reported that a U.S. MNE had to switch to only vegetarian food in its cafeteria because their operations were located in South India and many of their employees were vegetarian. Similarly, it is mentioned in this study that another company arranged two separate caterers to assure the employees that meat was not cooked in the vegetarian kitchen. Although variation does exist across regions in India, vegetarian diet is predominantly practiced by a smaller proportion of the population placed higher in the caste hierarchy, e.g., Brahmins and Vysyas. 1

Organizational behavior
Caste has been considered an important force in organizations, which likely influences social identities and work values (Cullen & Parboteeah, 2010;Parboteeah et al., 2009). IB scholars suggested that the caste system has a strong influence in shaping organizational culture in traditional Indian organizations, and thus might pose difficulties for managing organizations in India (Mathew et al., 2012). For example, the achievement orientation of employees in India might be low, given limited social mobility and given that life outcomes are determined by ascribed characteristics, such as caste (Morris et al., 1998). It has also been suggested that a supportive leadership style might alienate workers who are used to authoritative leadership in countries with high power distance, due to institutions like caste (Banai & Reisel, 2007). Some scholars wondered whether candidates who have been raised and socialized within the caste system (who might therefore be culturally bound, according to the authors) would be content when they are placed in cultures that are different from it (Harvey & Buckley, 1997). In a similar vein, highlighting the multiplicity of cultures within an organization, some scholars have discussed the practices and experiences of Indigenous tribal groups as part of regional subcultures in India (Dheer et al., 2015). Researchers have also compared the attitudes of Indian and American employees towards affirmative action policies (Combs & Nadkarni, 2005), and found that, compared to American employees, Indian employees prefer opportunity enhancement programs to preferential treatment programs and also have a less favorable view of those who avail themselves of affirmative action. Connected to this point, some scholars reported views expressed by the participants that some upper-caste individuals claim to belong to lower castes to avail themselves of the quotas in educational institutions that are reserved for lower castes (Ananthram & Chan, 2021).

Article
Discussion on caste Mani (2021) 44% of directors and owners belong to Marwari, Gujarati, or Parsi communities, which together represent less than 7% of the Indian population. Further, families control 23% of shares of Indian firms, with families of Marwari, Gujarati and Parsi communities controlling 14% of shares. Firms with family/community control tend to be smaller, younger, and have lower market share, and there is a churn in the identity of largest firms over time, indicating an entrepreneurship effect rather than an entrenchment effect Occupational roles Bhatt and Miller (1984) Traditional society may not look at education, skill, or prescribed training as a determinant of work assigned, but rather base it on caste or family linkage Garg and Parikh (1986) Caste influences the roles adopted in family, caste, and community, and these roles also extend to organizations Becker-Ritterspach and Raaijman (2013) The caste system exacerbated professional distance between hierarchical employee groups, i.e., reflected in the unwillingness of employees to engage in manual work or cleaning Khanna (2015) Continuous disenfranchisement of lower-caste individuals prevents them from engaging in entrepreneurial activities

Business networks
Studies have identified caste as a useful theoretical lens in explaining the social ties (Patel et al., 2012) and networks that reinforce managerial, financial, trading, and organizational affiliations (Chen et al., 2015;Contractor et al., 2015;Cuervo-Cazurra, 2006;Nair et al., 2022;Puffer et al., 2013;Schmid & Mitterreiter, 2020;Tsui-Auch, 2005). Scholars have suggested that caste facilitates identification (i.e., categorization of in-group and out-group) and creates personal obligations, which then influence the interactions of managers (Kumar & Singh, 1978). Along these lines, multiple studies have investigated the influence of caste on relationships and reciprocity, including one that likened caste to guanxi, and explored the role that caste plays in the exchange of favors (Thams et al., 2013). Verbeke and Kano (2013) reiterate this point by noting that in-group exchange of favors occurs within caste groups. Further affirming the role that caste plays in identification and obligations, Tsui-Auch (2005) reported that caste membership helped an entrepreneur in Singapore to tap into the international trading networks of members of his caste. A similar observation was made about business transactions based on caste among Indian-owned business firms in Africa (Prasad, 1967). Some scholars suggested that the caste system helped individuals from the Vysya caste group to succeed in business and develop business groups that many MNEs ally with (Rao, 2012). Building on this, and other studies that underscore the effect of caste in facilitating business transactions, some scholars have examined the pattern of ownership and control based on caste or community identity and found a concentration of family or community control in Indian firms. Specifically, 44% of directors and owners belong to Marwari, Gujarati, or Parsi communities, 2 which together represent less than 6% of the Indian population (Mani, 2021).
Based on the premise that caste similarity facilitates information sharing and trust formation, scholars have also studied the accuracy of analyst forecasts based on CEO-analyst caste similarity (Chen et al., 2015). They found that the accuracy of analyst forecasts was higher when the analysts shared caste affiliation with CEOs of the companies they covered. However, this effect was stronger for pre-reform generation CEOs and weaker for postreform generation CEOs, suggesting that due to economic growth and urbanization, the effect of caste on social interactions may be waning. 3

Occupational roles
Scholars have also speculated that caste might influence the adoption of occupational roles in organizations (Garg & Parikh, 1986) and that work may be assigned based on caste rather than education, skill, or training (Bhatt & Miller, 1984). These practices and the resultant inequalities appear inconsistent with scriptural teachings that emphasized the equality of all living beings, i.e., they are part of the unity. On this same point, however, some scholars suggest that the very virtue of religious tolerance would also imply that such inequalities end up being tolerated (Ananthram & Chan, 2016). In addition, some scholars suggested that caste increases professional distance between hierarchical employee groups, resulting, for example, in the unwillingness of employees to engage in manual work or cleaning (Becker-Ritterspach & Raaijman, 2013). Further, Khanna (2015) provided an example of how the continuous disenfranchisement of lower-caste members creates barriers that prevent them from pursuing entrepreneurship.
In sum, our review shows that the research on caste in IB is limited but has generated interesting findings and observations about the influence of caste on human resource practices, organizational behavior, business networks, and occupational roles. Taken together, these studies have revealed some ways in which caste, as an informal institution, influences individuals, their attitudes, and behaviors, as well as their roles in society and organizations. Therefore, to consolidate this understanding and advance our understanding of caste as an informal institution, we elaborate on caste as a distinct informal institution next and distinguish it from race, class, and gender.

CASTE AS A DISTINCT INFORMAL INSTITUTION
Informal institutions are those socially produced and enacted rules that are mainly unwritten, but are nevertheless shared, communicated, and implemented outside of formal, codified procedures -''the actual rules that are being followed'' (Estrin & Prevezer, 2011: 44;Helmke & Levitsky, 2004). Intrinsic to informal institutions are customs, traditions, codes of conduct, shared mental models, societal norms, and culture (Williams & Vorley, 2015). Through teaching, oral traditions, and family and kin relationships, these informal institutions are passed from generation to generation (Tonoyan et al., 2010). Even though some scholars assume that informal institutions would wane over time through the institutionalization of formal Local, yet global rules, their permanence is predicated on the fact that they are not just remnants of traditions but are rather woven into the fibers of society (Radnitz, 2011;Waylen, 2014). In addition, informal institutions can be sustained through social networks 4 by way of ''blackbox exchanges'' like ''emotional currencies and reciprocal obligations'' between members of a society, such that these networks can strengthen informal institutions by mediating the changes enacted by formal institutions and adapting to them (Minbaeva et al., 2022: 13).
Many elements of the current caste system can be traced back to the Vedic period, i.e., 1500 BCE-500 BCE. Caste has remained a potent force, in part because it is reproduced by multiple societal norms and behaviors, even as legislative and political institutions have changed and often attempted to weaken, if not outright ban, it. This survival of the caste system, despite lacking support from formal institutions, and in spite of the formal systems denouncing it, underlines its strength as an informal institution (Chrispal et al., 2021). Informal institutions persist because they shape behaviors through rituals, norms, and practices and in turn are legitimized through everyday practices of individuals. In other words, the caste system does not manifest itself primarily in formalized rules or explicit norms, rather, it provides the repertoire for shared expectations among groups within the society via deeply ingrained notions of socioeconomic hierarchy and norms for social interactions (Bhatt et al., 2022).
Although IB research has not explicitly acknowledged or theorized caste as an informal institution as such, it has nevertheless noted the ways in which caste operates as an informal institution, in terms of governing relationships within organizations and society. For example, as discussed in the previous section, scholars have shown that caste has an influence on tie formation, even though this effect weakens when modern institutions take precedence in facilitating tie-formation (Chen et al., 2015). Other research has similarly shown that caste can also influence network formations (Tsui-Auch, 2005), social exchanges (Thams et al., 2013), and interpersonal interactions, i.e., caste enables obligatory behaviors such as acceptance, economic dependence, psychological support, and mutual protection (Kumar & Singh, 1978). These findings are in line with the view that informal institutions are sustained by social networks that serve as both pipes (through which resources, favors, and information about more or less acceptable behaviors flow) and prisms (that help participants make sense of what flows through the pipes) (Minbaeva et al., 2022).
With its implications for hierarchy, occupational roles, and assignment of value for labor, caste can seep into organizations and create a culture that is different from the formal, or informal, culture of those organizations. For example, as indicated in some of the studies we reviewed, caste can result in submission to authority even when that is not required by formal rules or by organizational hierarchy (Saini & Budhwar, 2008). Also, employees may adopt their caste role in organizations by extrapolating their roles in social settings to such organizations (Garg & Parikh, 1986). Further to this point, because the caste system attributes lower value to physical labor and delineates norms about professions and professional distances, it can also shape interactions among employees who perform various roles in organizations (Becker-Ritterspach & Raajiman, 2013). In addition, since caste also operates through beliefs and ideologies, such as through stereotypes and meritocracy, it can also create inequalities in the workplace in ways that are unintended by rules and formal culture (Bhatt & Miller, 1984;Combs & Nadkarni, 2005).
Managers can often face challenges in dealing with caste, because it operates as an informal institution that manifests in cultural practices and relationships (Jodhka & Newman, 2007). For example, some of the studies we reviewed (e.g., Rao, 2012;Thams et al., 2013;Tsui-Auch, 2005) highlighted the benefits of caste in the building of network formations and ties. Without accounting for the role played by caste in the formation of these networks and ties, scholars run the risk of legitimizing their benefits instead of scrutinizing them, i.e., seeing variations in social capital as they ''are'', rather than inquiring how they came about, and what externalities related to those consequences are borne by others and society at large. Recognizing caste and studying it may also be challenging because caste has also been examined under other forms and constructs, such as social class, power distance (Sparrow & Budhwar, 1997), or religious practices.
As the discussion above shows, the caste system is an informal institution whose presence can be difficult to observe, as it can manifest in subtle and informal ways, and in a manner that makes it appear as, or correlated with other, better-studied social stratification constructs. Therefore, to better understand caste, it is important to note its distinctiveness by briefly distinguishing it from other systems of stratification that are studied by or are familiar to IB scholars, such as social class, race, and gender (Acker, 2006;Pitesa & Pillutla, 2019;Shore et al., 2009). A comprehensive analysis of these systems of social stratification is beyond the scope of our study, given the complexities and nuances involved in each of them. Therefore, the following discussion is meant primarily to highlight the distinctiveness of caste as an informal institution, and serve as a starting point for further clarification and theorization.

Distinctiveness of Caste as a Construct
Given the origins of caste, its evolution over time, as well as the regional, linguistic, and sociocultural diversity in South Asia and among its people around the world, there are challenges in precisely and exhaustively defining and describing what caste is. However, it can be said that caste is an informal institution that places individuals in fixed patrilineal categories that are arranged in a graded social hierarchy and provides rules to govern the social and economic action of individuals. As an informal institution, caste and caste-like systems are found across religious populations (although not necessarily exclusively in such populations) in South Asia, including Buddhists, Christians, Hindus, Muslims, and Sikhs (Mines, 1972;Ram, 2017;Trivedi et al., 2016). We do not provide details about the origins of caste or its features, as they have been discussed in earlier works (e.g., Bapuji & Chrispal, 2020;Claes & Vissa, 2020;Vikas et al., 2015). 5 Instead, we focus on the categories that IB scholars who are unfamiliar with caste might encounter when they wish to explore caste in their studies. Further, although caste is prevalent in South Asia and has spread around the world with the immigration of its people, in this section and the next, we discuss it from the perspective of India because of India's prominence as a large economy. Moreover, the bulk of research on caste has occurred in the Indian context and thus, discussing it from the perspective of India allows us to maintain some level of precision in our description of the caste system and its implications to MNEs and IB policy.
For the purposes of public policy, such as affirmative action or for conducting a population census, Indian society is classified into four broad groups: (i) Others ( (Deshpande & Sharma, 2013); (ii) Other Backward Classes (OBCs), largely Sudra (peasant) castes who are deemed to have not achieved socioeconomic progress comparable to Others as discussed above. This group accounts for about 41% of the population (National Sample Survey Organization, 2005); (iii) Dalits (also known as Scheduled Castes or SCs, outcasts, untouchables, harijans, etc.), who constitute about 16.6% of the population (2011 Census); and (iv) Adivasis (also known as Scheduled Tribes or STs), who constitute about 8.6% of the population (2011 Census).
The similarities and differences of caste with social class and race can be understood in terms of the formation of categories and identifying markers. Individuals enter a caste or race category by birth, which remains unchanged during their lifetime. In contrast, even if individuals enter a social class by birth, some movement in social class is possible within an individual's lifespan, due to changes in wealth, income, occupation, and education. Further, caste and race categories remain stable across generations (due to endogamy) while class categories can change. In short, race and caste are assigned by birth and remain fixed for life and are stable across generations, while -by comparison -class is, in principle, dynamic and mutable.
An individual's racial category can be identified based on visible phenotypic traits such as skin color and facial features (Relethford, 2009). In contrast, the markers that identify an individual's caste category are difficult to discern for those who are unfamiliar with the caste system. These markers include the family name and region the individual is from, socioeconomic status of the family, dietary practices, and social practices, including religious ceremonies and rituals. Similarly, the markers that identify an individual as a member of a particular social class can be subtle (Schiller, 1971) and include factors such as: language, accent, cultural tastes, and behavioral style. In short, whereas race can commonly be identified by traits that are generally visible, the factors that can be used to identify class and caste are generally less visible (Acker, 2006;Vaghela, Jackson & Sengers, 2022).
Caste is similar to race and social class in some of its consequences, such as the possession of and access to resources. For example, due to a combination of historic reasons and informal barriers, one's caste influences their possession of and access to economic, social, cultural, and symbolic capitals, which result in differential outcomes with respect to health, education, and income (Bapuji & Chrispal, 2020). These consequences of social stratification that are based on caste are similar to the consequences based on stratification by race (Acker, 2006;Wiegman, 1995) and social class (Dahrendorf, 1959) -which are also driven primarily by informal, rather than formal, barriers in contemporary times. In short, caste, class, and race are similar to each other in how they advantage some groups and disadvantage others, and the mechanisms of such inequalities are primarily informal rather than formal.
Relative to caste, gender is more visible and can be generally inferred from the physical appearance of individuals. However, there are also similarities. Gender, beyond its biological confines, and caste, are both social institutions that are situated in ''historicity, agency and power'' (Martin, 2004(Martin, : 1260. The categories within gender presently are fewer than the thousands of categories present in caste. This difference notwithstanding, these two social institutions both categorize people into groups, whereby one of the categories becomes the standard, or the default, to which those in the other categories are expected to or want to measure up to. For instance, Simone de Beauvoir (1952) suggested that men are the lone fundamental against which women are assessed. Similarly, the lower castes are measured against the ritualistic purity of the upper castes, implying that the former are inferior to the latter. Due to these allocations of hierarchical rank, and the resulting dominance and subordination, power relations among gender and caste categories are legitimized (Pandey, 2013). When gender and caste intersect, these power relations become even more visible. For instance, Dalit women receive lower wage rates than their male and upper-caste female counterparts (Deshpande et al., 2018). Both caste and gender allocate different occupational and societal roles to individuals, although these roles evolve and transform, albeit possibly at different rates for both systems. Finally, similar to race, class, and caste, gender also results in inequalities. In short, gender and caste share similarities in how they identify one group as the standard (i.e., men and upper castes, despite considerable variation within that group on other parameters) and in how they assign status and roles accordingly, but they differ in terms of how group membership can be identified and the number of categories within the group.
As the discussion above shows, caste is an informal institution, but it is not a weak one, because it persists whether formal institutions are deficient or not. In addition, even though caste shares similarities with other systems of stratification, such as race, class, and gender, 7 it is distinct from them. Therefore, studying caste can generate insights that are relevant, and important, for MNEs, and also help in the effort to better understand informal institutions in general. Towards this end, we next propose a simple framework to organize our discussion of the implications of caste.

IMPLICATIONS OF CASTE FOR MNES
Given the limited IB research on caste, developing implications of caste for MNEs solely based on IB research might not be viable at this point in time. However, caste has been studied by management research as well. In addition, there is a significant body of work on caste in neighboring disciplines, such as economics and sociology (e.g., Mosse, 2018;Munshi, 2019). Finally, a substantial body of work in management as well as neighboring disciplines has investigated various phenomena in the context of other systems of stratification, such as race, gender, and social class. Scholars interested in studying caste can draw on all these streams of work or take inspiration from them as points of departure for their endeavors. Accordingly, in this section, we use both the IB research we reviewed as well as broader research on caste that is conducted in organizational contexts to develop implications of caste for research on MNEs. Drawing on that broader research on caste allows us to base some of our speculations on firmer ground as well as point to mechanisms (e.g., homophily, discrimination, imprinting, network closure) that might drive such implications.
In developing the implication of caste for MNEs, we rely on a long tradition in IB research that has documented that home country (or country of origin) and host country effects are important drivers of MNE behaviors and outcomes (e.g., Ferner et al., 2001;Lu et al., 2014;Mingo et al., 2018). For example, Milliman and colleagues have developed a typology to understand the fit of international HRM by contrasting the larger organizational or environmental context (domestic vs. foreign) with organizational level (corporate vs. subsidiary) (Milliman et al., 1991). Similarly, Collinson and Rugman (2008) developed a regional matrix by distinguishing between geographic scope Local, yet global and geographic reach of firm-specific advantages, while Ralston and colleagues contrasted national culture (Eastern and Western) with economic ideology, categorized as capitalism and socialism, to investigate managerial work values (Ralston et al., 1997). Also in this area, Buciuni and Pisano (2021) developed four innovation models by contrasting the level of geographic concentration of value chain operations and the control of lead firms over such operations.
Similar to these, and other IB studies that have examined an interplay of home and host country environments and MNEs characteristics as major dimensions to understand various phenomena, we focus on home and host country environments to outline the implications of caste for MNEs. This is helpful for our analytical purposes because we consider caste as an informal institution that is tied to the institutional environment of a country. Considering such broader dimensions is useful for early theorization of under researched phenomena to generate wide-ranging implications that can be subjected to further, more specific and nuanced, theorization and rigorous empirical testing. Accordingly, we use the geographic origin of MNEs and the geographic scope of their operations as the two dimensions we consider. We categorize these dimensions by contrasting India and the rest of the world. 8 Accordingly, we use a 2 9 2 matrix (Figure 1) to present implications of caste that result from the geographic origin of MNEs (i.e., India or rest of the world), on the y-axis, and the geographic scope of their operations (i.e., India or rest of the world), on the x-axis. This yields four cells to consider the implications of caste to: Indian operations of Indian MNEs (Cell 1); Indian operations of non-Indian MNEs (Cell 2); Global operations of Indian MNEs (Cell 3); Global operations of non-Indian MNEs (Cell 4).
We draw on our review of IB research on caste, as well as our understanding of broader organizational research on caste, to consider the internal organization of MNEs -particularly caste homogeneity and diversity across the organizational hierarchy; interorganizational relationships of MNEs -particularly industry value chain linkages and stakeholder relationships, and also the organizational form of MNEs -more specifically their structure. As the interaction of home and host country institutional environments varies based on the geographic scope of MNE operations, we expect the influence of caste to decrease in importance, in terms of the implications of caste in their strength and form, from Cell 1 (Indian operations of Indian MNEs) through to Cell 4 (global operations of non-Indian MNEs). 9 Indian Operations of Indian MNEs (Cell 1) We conjecture that the influence of caste is likely to be the widest-ranging and strongest in this cell, because caste is embedded in Indian society in multiple and subtle ways as our review and discussion of caste as an informal institution has shown. In addition, the staff in these MNEs is likely to be predominantly of Indian ethnicity and nationality and is unlikely to have much ethnic or national diversity i.e., particularly in terms of representation from non-caste contexts. Our discussion in this cell is about Indian operations of Indian-origin MNEs, and not on purely domestic Indian companies, for which the influence of caste might be even stronger due to their lack of exposure to international markets.
Internal organization. The influence of caste on internal organization might be seen through caste homogeneity and diversity across different hierarchical levels and occupations. Our review found that while IB scholars have briefly noted this point (e.g., Sparrow & Budhwar, 1997;Thams et al., 2013;Varma et al., 2006), other scholars have given it more attention. For example, researchers have found that management teams and boards of directors of publicly listed Indian firms have high caste homogeneity (Ajit et al., 2012;Bhagavatula, Bhalla, Goel, & Vissa, 2019;Bhattacharya et al., 2022;Dayanandan et al., 2019), with one study showing that nearly 70% of Indian corporate boards had a Blau index of caste diversity of zero (Ajit et al., 2012); and that across all listed Indian firms, 90% of board positions are occupied by two caste groups (Brahmins and Vysyas) who are estimated to form about 7% of the Indian population. Research found that caste homogeneity among upper echelons has a negative effect on firm value (Bhagavatula et al., 2019;Dayanandan et al., 2019), in a way that parallels patterns that have been observed with respect to homogeneity on other demographic attributes as well (Carter et al., 2003).
The studies we reviewed (e.g., Sparrow & Budhwar, 1997;Thams et al., 2013;Varma et al., 2006) noted that caste homogeneity may occur due to hiring based on caste affiliation. Broader Local, yet global organizational research has dwelt more on the mechanisms that might be behind this observation, and argued that caste homogeneity may be a result of homophily, as boards dominated by a particular caste tend to have a CEO from the same caste (Dayanandan et al., 2019). Bhagavatula et al. (2019) showed that this homogeneity is primarily a result of induced homophily (i.e., the total pool of directors consisting mainly of Others (i.e., castes that the Indian government recognizes as socioeconomically advanced), thereby resulting in the selection or appointment of more of them, if only based on this circumstance, and not necessarily as a result of any of the decision-makers choosing to appoint individuals of the same caste). However, there is also some evidence of choice homophily (where even after accounting for induced homophily, there are differences, which can be attributed to decision-makers making a choice to appoint those of the same caste as them). Similarly, Damaraju and Makhija (2018) argued that choice homophily (also known as taste-based preference, especially in labor market research on discrimination) operates in CEO appointments, particularly outsider CEO appointments, where, according to the authors, caste serves as a source of information.

Figure 1 Implication of caste for MNEs and international business
Local, yet global A similar pattern in hiring was observed by Soundararajan and colleagues in their study of the Tirupur textile industry, where the owner-managers recruited individuals from their caste, based on the expectation that these recruits would remain trustworthy and loyal to the owners (Soundararajan et al., 2018).
As different from the case for upper-level positions, where homophily has been argued to be a factor (perhaps because of the relatively small pool of candidates with higher credentials and the similarity in the profiles of the selectors and candidates), discrimination and stereotyping based on caste have been noted to play a role in the selection of candidates for lower-level positions, where the candidate pool is larger and the selectors tend to be those with higher credentials and experience. For example, applicants with Dalit (or lowest caste) surnames were 33% less likely to secure an interview, compared with other applicants who have similar credentials (Thorat & Attewell, 2007). Further, hiring managers reported using criteria such as ''family background'' for managerial jobs, which ended up advantaging candidates from upper castes (Jodhka & Newman, 2007). Similarly, for jobs such as airhostesses, Air India promoted the selection of candidates with certain ''physical and cultural capital'' that advantaged women from the upper castes or classes, which might have been a reflection of its clientele in the past, but which continued even after their clientele has become more diverse (Nair, 2004). In this way, Air India may have imprinted their recruitment and selection practices with a preference for certain qualities, which then resulted in homogenization of their recruits along caste or class lines. This kind of imprinting may have gained legitimacy and salience in MNEs given that their focus on global markets is likely to emphasize the possession of social, cultural, and symbolic capital, which are unevenly distributed along caste, as well as along some other demographic attributes (Bapuji & Chrispal, 2020).
The uneven distribution of various types of capital necessary for professional and managerial roles may also play a role in shaping caste composition within organizations. For example, multiple studies in the Indian IT industry have found that a large majority of employees are from the Others, predominantly Brahmins, which some authors note is ''not surprising, given their monopoly over higher education and formal sector employment, especially in South India'' (e.g., Fernandez, 2017;Upadhya, 2007Upadhya, : 1864Shanker, 2000). A similar caste composition, i.e., high percentage of Others, predominantly Brahmins, was also reported for the early accounting profession in India (Sian & Verma, 2021).

Inter-organizational relationships
The influence of caste on inter-organizational relationships may be visible in the differentiation within industry value chains and in relationships with stakeholders. For example, upper-caste individuals are found in high value positions and lower-caste individuals are found in low value positions in footwear (Knorringa, 1999), textile (Mitra, 2015;Soundararajan et al., 2018), and diamond value chains (Munshi, 2007). To elaborate using one of these studies, Munshi (2007) contrasted Kathiawadis (a community of lower-caste agricultural laborers and farmers, also known as Kanbi Patels), jobbers who polish rough stones into finished diamonds in Indian factories, with their predominantly upper-caste Palanpuri Jain or Marwari customers, who source rough diamonds mainly from Africa and export the polished stones globally. Munshi's analysis and discussion highlighted caste-based social capital within each stage of the value chain, which resulted in the differentiation observed because of the inability of businesses to trade across castes (Munshi, 2019). This differentiation may also be a result of the elite closure (i.e., higher-status individuals limiting their friendship ties to other high-status individuals), observed by Bhardwaj et al. (2021) in the networks of MBA students at business schools in India.
In addition to these patterns, recent research in management also points to the influence of caste on mergers and acquisitions, as well as on financing relationships. For example, Bhalla, Goel, Konduri, and Zemel (2019) found that firms are more likely to acquire target firms whose directors share a caste affiliation with their own directors, and also pay a higher premium in acquiring such target firms. Similarly, as discussed in our review findings, analyst forecasts were found to be more accurate when the analysts shared a caste affiliation with the CEO of the firm being analyzed (Chen et al., 2015). In contrast, pointing to a potential case of caste heterophily, higher-caste VCs were found to set higher valuations for companies founded by lower-caste entrepreneurs, particularly when the latter have elite education credentials (Claes & Vissa, 2020). In short, although not extensive, extant empirical evidence suggests that caste has Local, yet global an influence on interorganizational relationships and may serve as source of trust and information for those who share caste affiliation, which may also lead to closure in business networks.

Organizational form differences
The implications we discussed above, as well as the mechanisms that might lead to those implications, might vary based on organizational form. In the Indian context, the most relevant organizational forms are business group affiliated organizations, unaffiliated private firms, state-owned enterprises (SOEs), and foreign-origin multinational enterprises, the last of which we discuss separately in the next section (under Cell 2).
As discussed in our review findings, there is some evidence that family-controlled business groups act as carriers and repositories that blend modern management practices with particularistic behavioral patterns among top executives (Chen et al., 2015). In contrast, SOEs act as implementors of the nation-state's mandate, which formally makes particularistic, caste-based behavior illegal. In addition, SOEs are subject to affirmative action quotas, leading to higher caste diversity of their employee base and the presence of lower castes on boards (Dayanandan et al., 2019). Although the presence of lower castes on SOE boards is minimal (e.g., Dayanandan and colleagues reported that only three board members of the total 34,772 directors in publicly listed companies belonged to SC/ST and that these three were all on the boards of SOEs), these differences could have some effect on the implications of caste for state-owned MNEs.

Indian Operations of Non-Indian MNEs (Cell 2)
Even if MNEs that originate outside of India and the Indian subcontinent may not be exposed to caste in their countries of origin, caste may still influence their operations in India, as they adapt to local informal institutions. Further, Indian operations of these MNEs employ a proportionately large number of employees from India, who are likely to carry the influence of caste into the MNEs.

Internal organization
As Others are overrepresented in the labor market for executive talent, Indian subsidiaries of foreign MNEs may also exhibit caste homogeneity in their upper echelons due to induced homophily, although some evidence also points to choice homophily in CEO hiring (Damaraju & Makhija, 2018). The mechanisms driving such patterns are not empirically tested, but it is possible that imprinting is a plausible mechanism. For example, when a foreign MNE's Indian subsidiary was first created, its initial members of the top management team imprinted those positions (Burton & Beckman, 2007). Hence, successors might be drawn from the same or similar castes as the initial role occupants, due to factors such as profile similarity, and the influence of references and referrals, as well as social and professional networks.
Similar to our earlier discussion about the Indian operations of Indian MNEs, caste stereotypes may also influence hiring for non-Indian MNEs operating in India, as expressed by an HR manager of an international joint venture: ''The Jat group is arrogant. They do not listen to anyone. Ahirs are tame.
Brahmans are more learned and they speak well and Scheduled Castes are not vocal'' 10 (Jodhka & Newman, 2019: 4129). Such caste-based stereotypes have also been documented by studies set in preeconomic reform India (Sinha & Sinha, 1967). The extent to which these caste stereotypes operate in MNEs may vary based on geographic location, because caste-based heterogeneity in human capital varies across states in India (Asadullah & Yalonetzky, 2012;Deshpande, 2007). In addition, the ''dominant castes'', meaning those with locally significant political clout (Srinivas, 1957), such as Yadavs (OBCs) in Uttar Pradesh and Bihar, can gain higher access to employment opportunities in a particular region (Verma, 2010). In short, caste composition -and the implications of patterns in that composition -might vary based on location, depending on the relative influence of castes in a given location.
Inter-organizational relationships Based on reasons that are similar to those aforementioned, caste might also influence the interorganizational relationships of foreign MNEs through imprinting. For example, Nokia's manufacturing operations in India, which set up the first mobile phone related value chain, had an initial set of suppliers and vendors whose caste composition may have been influenced by the location choice of the foreign MNE (given that different castes are prevalent in different parts of industry value chains in different locations of India). The subsequent evolution of the value chain's caste composition might then be driven by imprinting processes. Similarly, the relationships that the local subsidiary develops to manage its operations, including managing possible resistance to such operations, Local, yet global might be influenced by the location. For example, Coca-Cola developed relationships with locally strong Yadavs (OBCs) to mitigate resistance to its operations in Uttar Pradesh (Varman & Al-Amoudi, 2016).

Organizational form differences
The influence of caste may vary across the different forms that foreign MNEs take in India. First, relative to Western MNEs, Asian MNEs from high powerdistance contexts may see a stronger effect of casterelated imprinting or homophily within their operations. Second, foreign MNEs with a locally listed subsidiary and a long history (pre-reform era in India 11 ) may be more attuned to local institutional features, in ways that later entrants with wholly foreign ownership may lack (Chen et al., 2015). Third, foreign MNEs that initially entered India through joint ventures (e.g., Honda, Suzuki) may have had a different adaptation path compared to later entrants. Finally, mimetic isomorphism (Scott, 1995) may also be a factor, such that foreign MNEs that originate from a particular country (e.g., South Korea) might be influenced by each other's decisions regarding location choice, HR practices and corporate and business strategy, thus also exhibiting commonality with respect to the implications of caste. In short, differences in the geographic origin of an MNE, type of subsidiary, entry mode, and isomorphic pressures may yield different implications of caste for the operations of foreign MNEs in India.

Global Operations of Indian MNEs (Cell 3)
Drawing on prior IB research on home-country imprinting (Estrin et al., 2018;Shukla & Cantwell, 2018), we also expect influences of caste in the overseas operations of Indian MNEs. Compared to Cell 1 and Cell 2, we expect caste effects in this cell to be weaker (but still stronger than the effects in Cell 4 that we discuss in the next section), because caste as an informal institution may be weak or absent in those markets. In the global operations of Indian MNEs, caste effects may be witnessed predominantly through expatriate managers from India, through Indian-origin employees hired from the host countries, and via imprinting in organizational practices. As such, the effect of caste may be stronger in Indian MNEs that employ a larger proportion of individuals from India in their global operations or have closer interactions with India. However, as internationalization of Indian MNEs is a relatively recent phenomenon, scholarly as well as other research is scant on how caste influences global operations of Indian MNEs. By contrast, as we discuss in the next section (i.e., Cell 4), caste in non-Indian contexts and non-Indian MNEs has received some attention.
The role of caste might be visible in the diversity of upper echelons of overseas subsidiaries of Indian MNEs, due to induced or choice homophily (Damaraju & Makhija, 2018). The presence of discrimination and stereotypes that might influence hiring might also vary based on the host region. For example, recent immigrants from India to Western countries, who are highly educated and drawn from the post-economic reform generation, might not exhibit the preferences and biases of pre-reform generations (Chen et al., 2015), or might exhibit such preferences and biases via different channels. For example, while pre-reform generations might articulate their biases with direct reference to caste or equivalent terms (such as purity and cleanliness), post-reform generations might articulate them in more secular or universal terms (such as merit and cultural/personal choice). Providing an example of the implications of caste on internal organization, an engineer of HCL in U.S. filed a lawsuit against the company for unlawful discrimination, due to the caste rivalries of his and his supervisor from their home region in India (Murthy, 2020). The implication of caste in these MNEs might also depend on an MNE's core business because caste may have an effect (e.g., on recruitment) in some industries (e.g., BPO or call centers) but not in others (e.g., software), because discrimination based on caste may vary based on industry (e.g., research has noted the presence of hiring discrimination based on caste in the call center jobs, but not in the software jobs, Banerjee et al., 2009).
Additionally, caste orientations may differ across organizational forms. For example, caste influences on the internal organization and inter-organizational relationships might be different in MNEs that are affiliated with family-controlled business groups, because business ties are largely situated within the context of pre-existing social relationships in this case (Elango & Pattnaik, 2007). Caste influences might also be different in state-owned enterprises, because of their organizational heritage and history, as well as the impact of affirmative action quotas on SOEs. Such implications may also differ for entrepreneurial new ventures, which are often started by young founders who might be less sensitive to caste affiliations and more influenced, for example, by common university affiliations (cf. Chen et al., 2015).

Global Operations of Non-Indian MNEs (Cell 4)
We expect the implications of caste to be weakest in Cell 4, which examines the operations of non-Indian MNEs in markets outside of India. Our expectation is based on the fact that in the absence of caste as an informal institution in these contexts (or in contexts where it is present in a mild form), the implications of caste are likely to depend on the presence of Indian-origin individuals in the workforce. 12 Yet, investigation of such contexts may still reveal theoretically novel insights, because cases that fall in this cell possibly demonstrate the mechanisms through which caste might still manifest in even ''weak-form'' contexts. We also note that at times the caste influence in these settings may be subsumed by other effects in research to date -such as ethnic, racial or gender diversity. For instance, caste homogeneity within Indian employees may be overlooked if all of them are considered together as adding to ethnic or national diversity (as Indians) in the company; likewise, caste homogeneity may manifest as regional and linguistic effects and be considered as such, rather than as caste homogeneity itself (Claes & Vissa, 2020).
As mentioned, it is possible that the homogeneity of caste, and its related effects, might still be observed among Indian managers and employees even for cases that fall under this cell. For example, as an employee of Microsoft posted on its inhouse anonymous messaging app: ''Forget diversity. 40% of our team are Baniyas'' (belong to the Vaishya caste). 13 In studies of women in the U.S. tech industry, nearly all the participants from India were reported to be from the Others, and predominantly Brahmins (Bhatt, 2018;Twine, 2022). Even if such scenarios may not be representative of non-Indian MNEs more broadly, it might still be useful to examine caste homogeneity in teams with predominantly Indian-origin employees, to see whether such homogeneity -and any of its down-the-line implications for outcomes -are a result of caste identities becoming more salient through stereotype effects and identify situational cues of caste that might be present even in non-caste contexts (Hoff & Pandey, 2006. Caste identities might become salient in hiring, as an unintended consequence of the success of Indian-origin leaders in MNEs. As a large majority of individuals of Indian origin in the U.S. are from the upper-castes or those categorized as Others by the Indian government (a number reported to be around 90% in Kapur, 2010), it might not be surprising that several noted CEOs of Indian-origin in the U.S. are also from these same castes. For example, Indra Nooyi, ex-CEO of Pepsi, joined Pepsi because the company lacked someone of her ''ethnicity or international outlook who was female in senior management.'' She described her family as ''a good, conservative Brahmin family, deeply steeped in learning and education'' (Gelles, 2019). It is possible that stereotyping and related social cues (e.g., last name, vegetarianism, cultural practices, religiosity) that are linked to these successful individuals influence hiring and the resulting internal organization, as we discussed in Cell 1. In addition, given the prominence of Indian-origin individuals in executive and managerial positions in Silicon Valley, it is also possible that hiring is influenced by their access to social and professional networks of Indians (Inamdar & Alluri, 2021). In a study of intersectional gender inequalities in Silicon Valley, the preference that Indian-origin (and more broadly, Asian and White) individuals receive in hiring has been conceptualized as geek capital, a form of social and symbolic capital that provides candidates with ''privileged access to decisionmakers who can offer them entry-level and midlevel jobs'' and ''allows job candidates to be perceived as possessing 'merit-based' qualifications'' (Twine, 2022: 8).
While caste homogenization can have performance implications, as we discussed earlier (in Cell 1), it might also have a more immediate implication for employee diversity, which is a matter of primary importance, even more so for organizations in this cell. Caste can pose challenges for equity, diversity, and inclusion practices of MNEs, particularly in technology sectors that employ Indian-origin employees in large numbers. For example, Asian Americans (the category that includes Indians and other South Asians 14 ) constituted 6% of the U.S. population but made up 14% of all employees in high tech industries in 2014, as compared to 5.8% across all industries in the U.S. private sector. This number is even higher among professionals in high tech industries (19.5% compared to 9.7% for technicians) and among leading high-tech firms (41%, including 36% in executive and managerial positions) (U.S. Equal Employment Opportunity Commission, 2015).
The challenges posed by caste for equity, diversity and inclusion (due to the upper castes forming an overwhelming majority among Indian diaspora) are indicated by a few cases. For instance, Cisco Systems Inc. was sued in 2020 for discriminating against a Dalit employee, referred to as John Doe. It was alleged that upper-caste supervisors of John Doe revealed his caste identity, which resulted in discrimination (Attri, Joshi, & Bapuji, 2021). Specifically, one of his supervisors revealed that he availed affirmative action quota for his seat in IIT, thus invoking stereotypes about such individuals being less competent (Prasad et al., 2020). Subsequent events resulted in him losing out on pay and progression. This incident shows that, as we discussed in Cell 3, the reflection of caste-based biases may also be observed in the form of modern channels, as reflected in values like merit and competence. In another incident, an employee of NTT Data services, a Japanese multinational, used caste slurs against lower-caste individuals on social media, prompting the company to denounce such behavior and take (unspecified) action. Some studies that are conducted in the U.K. (Metcalf & Rolfe, 2010) and the U.S. (Zwick-Maitreyi et al., 2018) as well as testimonials collected by advocacy organizations in the U.S. (AKSCUSA, n.d.) point to unfair treatment of lower castes, use of caste slurs, and promotion of practices with caste undertones (e.g., rituals, vegetarianism, and cultural celebrations) in organizations. Incidents 15 such as these can have major implications for the reputation of the organizations [as seen in incidents that are related to discrimination on other demographic factors e.g., Denny's racial discrimination crisis in the 1990s (Kanso et al., 2010)], as well as the wellbeing of employees .
Within this cell, we can also direct our attention to interorganizational relationships, particularly to CSR initiatives that are used by organizations in other countries. For instance, a lawsuit was filed against a sect for exploiting the labor of lower-caste men to build a temple in New Jersey (Attri, Soundararajan, & Bapuji, 2022). This incident highlighted issues of modern slavery based on caste that MNEs should be aware of so that they can prevent similar phenomena from arising in their value chains (Ethical Trading Initiative, 2019). It also underlines the potential implications of firms' CSR relationships and related associations, e.g., Capital Financial group, a U.S. financial company, donated money to the sect that built the temple.
To summarize our discussion above about the four cells in Figure 1, IB research on caste, as well as broader organizational research on the topic, suggest that caste might have implications for MNEs as manifesting within a firm's internal organization and in its interorganizational relationships, which can differ based on the firm's organizational form. The mechanisms underlying these processes can range from homophily to institutional isomorphism to the development and use of stereotypes. In addition, the influence of caste varies based on the geographic origin and scope of the MNE, where our expectation is that the effects will be strongest in the operations of Indian MNEs in India and weakest in non-Indian MNEs' operations outside of India.

DISCUSSION
Our study (i) highlighted the scarcity of attention to caste in IB research, (ii) theorized caste as a distinct informal institution, as different from race, class, and gender, and (iii) provided an initial, and at times anecdotal and necessarily speculative -as limited/based on the scarcity of evidence and available research to date -outline of some of the implications of caste to MNEs. In doing so, we seek to motivate future empirical research that informs IB scholars to appreciate the relevance of caste and investigate it in their studies of MNEs. To spur thinking about such possibilities, below we discuss how studying caste from an IB perspective can generate insights about inequalities and informal institutions. We then present a more elaborate discussion on the policy implications and related further research opportunities.

Caste, Inequality, and Informal Institutions
With rising income inequalities and greater attention given to such inequalities, corporations face the threat of social movements, erosion of legitimacy, increased regulation, and decreased support for globalization -all of which have implications for MNEs . While corporations can directly contribute to income inequality in their role as wealth distributors, they also indirectly contribute to inequality through employment practices and via strategic actions that shape the institutional environment in their favour (Amis et al., 2020;. As such, MNEs have the opportunity to also use these practices to reduce inequality. Given the various institutional environments in which MNEs operate and the differential progress made on inequalities across these contexts, IB scholars are well placed to study the effect of institutions such as caste on inequalities. Local, yet global In studying caste as an institution of inequality, IB scholars can refer to and build on insights from the rich scholarship that has examined the implications of gender to HR practices and organizational culture, as well as the subtle ways (e.g., stereotypes that are related to personality attributes and managerial skills) in which gender inequalities arise in organizations (Koveshnikov et al., 2019). Studies of caste can also leverage and adapt the future research recommendations that researchers have put forth for gender, such as studying the dispersal of value-adding activities and conceptualizations of the ideal manager, as well as revisiting ethnocentrism and homosociality from a caste vantage point (Koveshnikov et al., 2019). In these and other ways, taking research on gender as a starting point might help IB scholars studying caste to accumulate insights on the implications of caste to MNE operations and decisions. Similarly, scholars may also find it useful to leverage applicable insights from research conducted on other stratification systems (e.g., class and race) to start expanding our understanding of the implications of caste.
Studying caste can also potentially generate new insights about some characteristics of informal institutions. As our review showed, IB scholars are aware of caste and have speculated on its implications, but the topic has not yet received close and systematic attention across studies, which raises important questions. For example, the challenges with identifying the implications of caste (e.g., as noted by Rao, 2012) and studying them might raise the broader issue of whether, when, and how informal institutions in general are difficult to observe, theorize, and empirically examine. This may be because informal institutions can manifest as social norms that are related to duties and responsibilities, influencing the social division of labor (Gupta et al., 2002), adoption of occupational roles in organizations (Garg & Parikh, 1986), preference for an authoritative leadership style (Banai & Reisel, 2007), and effort exerted and expectation of rewards (Morris et al., 1998). Such manifestations may be viewed and interpreted as individual attitudes and preferences, or as resulting from other factors, rather than as (also) being systematically linked to the effects of caste as an institution. Further, some of these social norms may also have religious roots, which may be taken to be primarily a cultural component or dimension (e.g., power distance), if not directly as religion itself (e.g., as discussed by Ananthram & Chan, 2021). Additionally, studies on caste can shed light on whether and how informal institutions of one type (e.g., caste system) interact with another (e.g., class or race) in MNE contexts, including how individuals raised within one system respond to another system (Harvey & Buckley, 1997). Finally, studies on caste may help researchers shed further light on how informal institutions can influence ethical decision-making and actions (Ananthram & Chan, 2016) as well as how they influence equality, diversity, and inclusion efforts in MNEs (Bhatt & Miller, 1984;Sparrow & Budhwar, 1997).
The persistence of informal institutions, even as formal institutions develop and change, is an important issue for IB scholars to examine, because this persistence has implications for MNEs (Minbaeva et al., 2022). For example, the persistence of caste can be studied by examining the characteristics of social network processes and structures that are influenced by caste, i.e., how they are built and maintained, and whether those networks are relatively closed or open, or whether they are affective or instrumental. Studies of caste can also examine the interplay between formal and informal institutions in terms of how the latter mediate and enable resistance to changes in formal institutions, e.g., through relabeling preferences and choices as based on the norms of meritocracy (Twine, 2022) or the recategorization of individuals in light of affirmative action based on gender (e.g., recategorizing women as those from different gender to those from the same caste, Bhattacharya et al., 2022). At the same time, a caste lens can help scholars uncover additional mechanisms (e.g., cultural practices and power relations, which we touch upon below) that can help understand the persistence of informal institutions and offer explanations that might complement mechanisms that are based on social networks (Minbaeva et al., 2022).
Broadly speaking, institutions are often associated with relations of power (Lawrence et al., 2013;McCarthy & Moon, 2018;Munir, 2015) and motivations that exist between actors to either ''create, transform, or disrupt institutions'' to serve their interests (Lawrence, 2008: 173). These power relations and motivations can also manifest in the behaviors of individuals in organizational contexts. Therefore, future studies can examine how informal institutions, like caste -whether in isolation, or in combination with other social institutions (e.g., gender, race, and class) -affect power relations and their influence on MNE operations (Festing et al., 2007;Oetzel & Doh, 2009). Studying caste as an informal institution can also open new avenues to study how informal Local, yet global institutions can materialize as, and influence, certain features, e.g., community ties, which are often attributed to other constructs, such as social capital. In other words, informal institutions such as caste could be potential antecedents of other well studied constructs such as community social capital.
As a complex and durable institution, studying caste can also offer useful insights on how organizations might reproduce inequalities within their boundaries, as well as on institutional frailties (i.e., vulnerabilities and weaknesses of formal institutions) in countries with a caste system (Verbeke & Calma, 2017). However, caste is not the only system of stratification that has been linked to inequality, and research on the implications of other such systems can also generate much needed insights, e.g., Japanese Feudalism -the persistence of family logics through the preservation of social capital (Bhappu, 2000), or Africa's spatial apartheid -coercion of marginalized peoples to remote locations to impede access to economic and social capital (Hirsch, 1991); and even modern formulations of status, such as firms hiring employees from Ivy League or elite schools (Rivera, 2015). In studying such inequalities, some scholars note that adopting a caste lens can generate useful insights on how structural inequalities get reproduced and reinforced within organizations and societies (Twine, 2022;Wilkerson, 2020). More generally, it is important for IB scholars to develop an evidence-based understanding of the implications of caste, and other systems of inequality, by studying them both independently and in conjunction with each other, in order to address the grand challenge of inequality and keep IB scholarship engaged with this important matter (Buckley et al., 2017;George et al., 2016).

Policy Implications of Caste in International Business
The policy implications of caste for MNEs are numerous, ranging from macro policies covering nations to MNE policies covering individual organizations, which are often influenced by the macro policies. Among the macro policies, we discuss those covering human rights, anti-racism, and affirmative action. We then suggest possible policies that organizations have developed, or can develop, in response to these macro policies. We also point to research opportunities and provide research questions to help researchers investigate the possible implications of such policies for MNEs. It is important that IB scholars pay attention to these policy implications so that the discourse on policy is influenced by scholarly research and evidence of such disinterested investigation rather than by other factors, such as the preferences of interest groups. We present a summary of the discussion in Table 2 and elaborate below.

Human rights policies
It has been argued that ''caste stands as a direct challenge to a fundamental underpinning of the human rights ideal -that all human beings are equal and deserve the same respect, dignity, and rights'' (Bob, 2007: 169). The UN's Committee on the Eradication of Racial Discrimination has noted that caste is a form of discrimination based on descent. As a result, UN member states are required to include caste discrimination cases in their reports of compliance with the International Convention on the Elimination of All Forms of Racial Discrimination. Several independent agencies (e.g., Centre for Human Rights & Global Justice at New York University) and advocacy organizations (e.g., International Dalit Solidarity Network) help the UN in its efforts to eradicate caste discrimination. At the macro policy level, these efforts aim to (i) recognize caste discrimination as a human rights issue, and (ii) eradicate caste-based forced labor. These macro policies translate into IB policies aimed at eliminating caste-based discrimination within MNEs and their global value chains (GVCs).
To prevent caste discrimination within their operations and thus support the respect, dignity and rights of all people, MNEs can include actions against caste-based discrimination within their workplace and employee policies. For instance, Dell's global social media policy asserts that it has zero tolerance for hate speech based on a protected classification and includes caste alongside gender, race, religion, and other such demographic categories (Dell, 2022a). Such policies can have a greater impact when they cover a wider set of areas, such as protection of customer data, equal opportunities, and harassment. Yet, Dell's code of conduct that more comprehensively outlines expectations from employees on all such areas includes gender, race, and religion, but not caste (Dell, 2022b). Given such developments, researchers can study how anti-caste-discrimination policy can extend across various areas within MNEs, in their products and services, and in the locations they operate. Further, as caste could be a taboo topic among practitioners, scholars can look at Local, yet global  Inclusion of caste-based discrimination in MNE policies, e.g., employee conduct policy, social media policy, and hate speech policy • How can caste discrimination policy be extended across various areas within MNEs, in their products and services, and in the locations they operate in? • How can implementation of such policies affect practices and behaviors within the organization? How can MNEs manage resistance to such policies? • Do MNEs face stigma for addressing caste? If so, would it be localized or global and how can it be managed? Policies to eradicate caste-based discrimination, e.g., training and education, performance management, and career progression • What are the various forms in which caste discrimination can occur within MNEs, and how does it affect employee lives, performance, and career progression? • What training and sensitization programs are effective against caste-discrimination for employees, managers, and executives? • How can MNEs create safe spaces for discussion of local inequalities like caste and manage them while maintaining their global reach and profitability? Eradication of forced labor based on caste (global, national)

Protocols to abolish forced labor based on caste (trafficking & slavery) within GVCs
• How do MNEs, wittingly or unwittingly, contribute to human rights offenses like trafficking and slavery against lower castes within GVCs? • How can MNEs recognize caste-based modern slavery and address it? Policies to reduce caste-based discrimination within GVCs • What policies/practices help MNEs better detect and address caste discrimination in GVCs, including that faced by the more vulnerable among the lower castes (e.g., marginalized within lower castes, women and sexual minorities)? How do organizations communicate and enforce such policy to suppliers in caste-affected countries? • How can MNEs mitigate the risk of caste-discrimination, e.g., via monitoring and regulation? What technological and partnership solutions are available and what determines their effectiveness?

Anti-racism
Inclusion of caste within antiracism frameworks (national) Policies to recognize caste as a protected category • How is racism related to discrimination based on caste? How can MNEs leverage their experience with racism globally to address caste discrimination? • How can racialization occur based on invisible characteristics within MNEs? What are the mechanisms through which such racialization is perpetuated (e.g., cultural practices, social interactions, and cultural/ethnic associations)? Inclusion of Indigeneity within anti-racism frameworks (national)

Policies to recognize Indigeneity
• How are racism and caste discrimination related to discrimination based on Indigeneity (i.e., STs or Adivasis)? What are their relative strengths/prevalence in MNEs? • What can organizations learn from indigenous knowledge and lived experiences within the areas of sustainability and grand challenges? And, to design policies to correct historic and ongoing marginalization?

Affirmative action
Affirmative action within organizations (national) HR policy to access talent pools from lower castes, via quotas, targets and equal opportunities • What type of affirmative action (e.g., quotas, targets, preferential treatment, opportunity enhancement) are successful, and what contingent factors affect such success (e.g., elite school, high-class, women, type of profession)? How can resistance to affirmative action policies be managed? • What barriers do MNEs face to access caste-diverse talent pools for positions at various organizational levels? How can they overcome the same? How can MNEs leverage their experience with affirmative action in other countries to design and implement the same in caste affected countries? HR policy to diversify managerial and leadership talent • What challenges do employees hired under affirmative action policies face in organizations? Do these (i) vary based on type of affirmative action or demographic profile of the employee? (ii) inhibit their wellbeing and career progression? • How can MNEs limit homogenization of top management teams based on caste? What policies and practices can MNEs develop to increase access of lower castes to managerial and leadership positions? What factors (e.g., recategorization/ homogenization) prevent those from lower castes from being selected on boards?
Local, yet global Hari Bapuji et al.
whether MNEs face stigma for addressing caste, or whether their legitimacy would be threatened and if such a threat would be localized or global.
MNEs also need to design policies to eradicate caste-based discrimination through training and education. For instance, Amazon and Apple have begun to include caste in company manuals and trainings (Venkatraman, 2022). However, some organizations have been more apprehensive in doing so. For instance, following resistance from some employees that discussion on caste would target them as Hindus, Google cancelled a talk aimed at educating its employees about caste and caste-based discrimination (Tiku, 2022). These developments raise questions such as: Why are some MNEs wary while others embrace the need for caste education? What performance management and career progression policies can MNEs put in place to prevent incidents like those of the Cisco employee discussed in Cell 4? More broadly, how can MNEs balance their efforts to address both globally recognized inequalities (e.g., gender, ethnicity) and less recognized inequalities (e.g., caste)?
Going beyond the confines of MNEs, there is a growing push for the development of macro policy targeted at the eradication of forced labor and human rights abuses, including those that are based on caste, within global value chains (GVCs) of MNEs. For example, multilateral organizations like the International Labor Organization (ILO) have developed protocols to abolish forced/bonded labor (90% victims in South Asia belong to lower castes) within global value chains (Anti-Slavery International, 2021). Therefore, scholars can examine how MNEs can recognize caste-based modern slavery and address it.
Stakeholders have been developing guidelines and base codes to reduce caste-based discrimination within GVCs. For example, the OECD Watch, a multistakeholder initiative, supported the revision of OECD guidelines for MNEs to specifically address discrimination and human rights offenses against lower-caste individuals (Ingrams, Wilde-Ramsing, & Fleuren, 2021). In addition, the Ethical Trading Initiative has developed a Base Code Guidance to assess and mitigate caste-based human rights risks within GVCs (Ethical Trading Initiative, 2019). These initiatives appear to be bearing fruit as some MNEs have adopted caste as part of their policies for GVC monitoring. For example, the Supplier Code of Conduct of Loblaw Companies Ltd. includes caste, alongside others, as a protected category and notes that suppliers are expected to prevent discrimination and monitor human rights abuses, as well as implement ongoing improvement programs and audits (Loblaw, 2022). Similarly, Amazon has included caste in its supplier code of conduct. These examples indicate that lead firms can use their power to dictate the ''terms and conditions of GVC membership'' to uphold social standards among their suppliers through the use of codes of conduct and monitoring of work conditions (Pietrobelli et al., 2021).
Although such policies are being announced, it is not clear yet whether and how much they are making an impact. As discussed before, because caste is an informal institution, it can be difficult to detect caste-based discrimination and exclusion, especially for those who are unfamiliar with the ways in which it can manifest. Further, MNEs might often not be motivated enough to use their power for social justice issues, and ''either cajole with too small a carrot to incentivize compliance among their suppliers or use too small a stick when there is non-compliance, or a combination of both'' (Van Assche & Brandl, 2021: 138). Therefore, IB scholars can study the ways in which MNEs can be motivated further to use their power and what Local, yet global policies or practices aid MNEs to better detect and address caste discrimination in GVCs, including that faced by the more vulnerable among the lower castes, e.g., women and sexual minorities. Research can also look into how MNEs can mitigate the risk of caste-discrimination via monitoring and regulation. For example, some scholars suggest that AI and block chain technologies can be leveraged to monitor supply chains for caste discrimination (Tambe & Tambay, 2020).

Anti-racism frameworks
While human rights policies cover the rights of all individuals and acknowledge structural factors, such as caste, race, and gender, which affect human rights, there is an increased recognition that eliminating barriers to equality due to ethnicity and racialization need more concerted and focused efforts. Therefore, some countries (e.g., Australia, Canada, and the U.K.) are developing their own anti-racism policy frameworks. For example, the Australian National Anti-Racism Framework (Australian Human Rights Commission, 2021) suggests that institutional bodies and workplaces should reduce inequalities and improve opportunities and delineate policies that protect the human rights of people of all races and ethnic groups by building a strong legal framework. Such national anti-racism frameworks may need to consider the inclusion of caste (and similar institutions such as clan or religion) that might get subsumed under the broader identity of ethnicity, and thus end up being overlooked. If demographic categories like caste are overlooked and not included in antiracism frameworks, disadvantaged groups may not have a recourse when they face discrimination and unfair treatment that is similar to what they may have attempted to escape by immigrating to developed countries in the first place. A similar consideration applies to MNE policies that consider ethnicity or descent but not caste. When cases of caste-based discrimination in the workplace emerge, they may go unrecognized byand not fall under -the national laws that cover workplace discrimination, ''because caste does not line up with any of the categories protected by the law, such as race and national origin'' (Wiessner, 2020). Not being able to find the right ''box'' to fit caste in might result in caste-based discrimination go unnoticed and unchecked, causing productivity losses similar to those that are experienced as a result of other forms of discrimination. To reduce such lapses, MNEs may need to consider organizational policies that recognize caste as a protected category and train employees to be sensitive to caste-based discrimination and identify it where it exists. This is particularly important because changing migration patterns mean that more lower-caste individuals are likely to enter the global workforce, which has hitherto been primarily populated by the upper-caste individuals (Chakravorty et al., 2016;Kapur, 2010).
Consideration of caste (and similar institutions) under anti-racism frameworks, but also within MNE policies, may also reveal insights into the similarities and differences between discrimination based on race and caste, and how MNEs can leverage their experience in identifying and combating racism globally to also address caste discrimination. Such a consideration can provide insights on how racialization (Hochman, 2019) may occur based on characteristics that are less visible. Investigations in this direction can also pave the way for better understanding how racism may be normalized or perpetrated through cultural practices and social interactions (e.g., last names, ceremonies, and endogamy, as noted by Chrispal et al., 2021) and cultural or ethnic associations (Parameswaran & Sebastian, 2007;Pariyar, 2020). The generation of more evidence and knowledge on these matters can help MNEs make their products and services more inclusive as well. For example, Twitter and Facebook have recently included caste in the policies that govern hate speech on their platforms.
Including caste as part of anti-racism frameworks aims for the equality of all individuals covered by caste. However, Adivasi (Indigenous) Peoples (i.e., Scheduled Tribes) need special consideration, for which MNEs can look to the anti-racism frameworks of other countries. For example, Canada's Anti-Racism Strategy draws on the perspectives of the ''First Nations, Métis, and Inuit Peoples'' to empower them and seek their guidance on initiatives that can meet their needs. Such strategies in Canada (as well as in Australia and New Zealand) that seek to reconcile with and learn from Indigenous Peoples can prove useful in developing approaches for engagement with Adivasi (Indigenous) Peoples in South Asia. Such learning efforts are particularly pertinent in the context of sustainability issues, where indigenous knowledge can be a resource for adapting to climate change and sustainable living (Magni, 2017). Further, these approaches can help MNEs better understand the forms of misrecognition, oppression, and marginalization that Adivasis faced historically and during the colonial period Local, yet global (Ghurye, 1980). 16 In turn, MNEs can better understand the views and aspirations of the Indigenous Peoples in South Asia that are about economic development and use of natural resources. More broadly, as a larger research project, a comparative analysis of Indigenous Peoples in various countries can help us better understand their shared experiences, so that policies and measures can be designed to acknowledge and correct historic and ongoing marginalization they face.
Affirmative action policies Affirmative action policies are enacted and implemented in various countries to correct historic and ongoing oppression, as well as to remove educational and economic imbalances among groups. These took different forms in different countries, such as ''standardization'' (access to tertiary education for disadvantaged groups) in Sri Lanka, Federal Character Principle (equity and diversity in public services) in Nigeria, and bumiputera (translates as ''sons of the soil'') policies that enable affirmative action for the majority Malay community in Malaysia (Penny, 2021;Sowell, 2004: 2).
Building on early affirmative action, which sought to advance non-Brahmins during colonial rule, independent India adopted extensive affirmative action measures to mitigate caste-based inequalities (Weisskopf, 2004). Besides reserving seats in the legislature, affirmative action reserved 49.5% of seats in public universities and public employment for those individuals belonging to SC, ST, and OBC communities, whose share in the population is nearly 70% (Deshpande, 2006;Higham & Shah, 2013). In addition to these, affirmative action policies have been used to increase entrepreneurial capabilities and initiatives, for example, through preferential loan schemes (Jayal, 2015).
While quotas in India are constitutionally mandated, affirmative action in the U.S. arises from executive orders and Civil Rights Act, 1964. In the U.S., affirmative action takes the form of a preferential boost for minority groups rather than quotas, and is monitored by the Equal Employment Opportunities Commission (Weisskopf, 2004). Further, affirmative action quotas in India targeted only the public sector and thus, the private sector did not implement these and its stakeholders exerted no great pressure toward this end either. In contrast, affirmative action in the U.S. targeted both public and private sectors, and progress was monitored by the Equal Employment Opportunities Commission as well as other stakeholders.
It would be useful for scholars to study the relative effectiveness of these policies by comparing outcomes in South Asian countries with and without affirmative action, as well as by comparing affirmative action based on caste with those based on other demographic categories (e.g., for Indigenous Peoples in Australia, Canada, and New Zealand). Scholars can also study the unintended consequences (e.g., increased salience of identities and resulting conflict) of affirmative action. For example, in Sri Lanka, preferential admission policies in universities were instituted by the Sinhalese majority to curb the ethnic Tamil group's dominance in education. To do this, students were given a standardized score, which was a score measured in relation to other students' score in their own ethnic group (Sowell, 2004: 89). Such preferential policies, along with the politicization of group identities, e.g., as based on language, may have deepened conflict between the Tamils and Sinhalese (Sowell, 2004).
Similarly, policies in India have made notable progress in terms of enabling socioeconomic mobility for marginalized castes (Piketty, 2020;Weisskopf, 2004); yet, in the public sector, progress falls short of the mandated quotas (Yengde, 2019). More broadly, the limited presence of lower castes among the social, political, and economic elite is conspicuous (Jodhka & Naudet, 2019Yengde, 2019). For example, as we have mentioned, of total 34,772 directors, Dayanandan and colleagues identified only three directors from SC/ST groups (whose population share is nearly 30%) and 69 directors from the OBC groups (who represent 52% of the population, according to the Mandal Commission Report in 1983) (Dayanandan et al., 2019). As another example, at the Indian Institutes of Management (IIMs), which are required to implement affirmative action quotas in faculty positions, only 61 out of 667 faculty members were from the SC, ST, and OBC groups (i.e., 9.1% compared with the 49.5% quota) (Egalitarians, 2021).
The limited success of constitutionally mandated affirmative action quotas in India in shifting the composition of the elite, when contrasted with the voluntary affirmative action policies (albeit with judicial and legal force for equal opportunities) in the U.S., 17 is worthy of examination. Scholars have argued that the outcomes in the U.S. and India are both positive, but also display differences due to a variety of reasons, including that the coverage in the U.S. includes the entire economy vs. the limit to the public sector in India, differential attitudes towards affirmative action and its beneficiaries, Local, yet global differential wealth of the countries, and the share of population covered being lower in the U.S. than in India (Sowell, 2004;Weisskopf, 2004).
The discussion above about the range of affirmative action around the world, their effectiveness, and their unintended consequences, provides a rich background for MNEs to develop their policies because caste diversity is likely low in MNEs. These policies can be aimed at diversifying the talent pool, eliminating biases in hiring, developing an inclusive leadership, and diversifying suppliers. First, MNEs can examine whether their hiring pool and hiring sources (e.g., colleges, hiring agencies, referrals from current employees) reflect the level of caste diversity in society. If access to a diverse pool of candidates is a challenge, MNEs can explore the feasibility of hiring from government institutions that implement affirmative action policies and/or provide short-term internships to candidates from caste-disadvantaged backgrounds to better assess their suitability. Also, MNEs can access talent pools from lower tier cities where more lower castes live. According to the census, the population of SCs and STs living in the six biggest cities in India amounts to 11.25%, as compared to their nearly 30% share in the total population (Devulapalli, 2019).
Second, MNEs can address caste diversity by developing HR policies to better access caste diverse talent pools through the utilization of quotas, targets, and equal opportunities. For example, MNEs can evaluate their ideal candidate profiles to see if these profiles might systematically advantage/disadvantage some castes over others and also investigate if caste stereotypes operate in their hiring practices. Based on such evaluations, MNEs can deploy and emphasize selection mechanisms that do not favor caste-privileged candidates (and in parallel, which do not discriminate caste-disadvantaged candidates), based on markers of class/culture. Third, quotas within MNEs are also being used as a tool to diversify managerial and leadership talent. Europe, for example, used mandatory quotas for women on corporate boards, where 40% of nonexecutive director posts were allocated to women (Leszczyń ska, 2018). Similarly, India also used an affirmative action that mandated all boards of a certain company size to have at least one woman director. While this quota increased gender diversity on boards, caste and community diversity remained unaltered (Bhattacharya et al., 2022).
To gain a deeper understanding of the HR policies discussed above (i.e., policies to diversify the talent pool, reduce biases in hiring, and develop diverse leadership), scholars can delve into the challenges faced by employees who are hired under affirmative action policies; whether such challenges vary by the type of affirmative action or demographic profiles of the employee, or whether and how they inhibit these individuals' wellbeing and career progression. Research can also shed light on how MNEs can limit homogenization of top management teams based on caste; what policies and practices MNEs can develop to increase access of lower castes to these leadership positions; and, what factors prevent individuals from lower castes from being selected on boards.
Finally, complementing the HR policies, MNEs can develop procurement policies that engage with broader affirmative action and diversify their supply bases. For example, MNEs can work with intermediaries, such as NGOs or industry associations (e.g., Dalit Industries and Chamber of Commerce) to tap into the supplier pool from lower castes. Scholars can also examine how MNE procurement policies can engage with affirmative action policies of the government and what partnerships can help achieve this effectively.
Overall, the implications of caste for MNEs and IB policy are wide-ranging. covering operations within MNEs as well as the role of MNEs as orchestrators of GVCs. These implications are also influenced by macro policies that are related to human rights, antiracism, and affirmative action. In the area of human rights, MNE policies can cover the elimination of caste discrimination and forced labor. Engaging with national anti-racism frameworks, MNEs can develop policies to recognize caste and indigeneity as protected categories. Finally, in the area of affirmative action, MNEs can develop HR and procurement policies to provide equal opportunities to castedisadvantaged individuals. The impact of these policies may vary depending on the strength of the implications of caste for MNEs, i.e., by their geographic origin and scope of operations. We briefly touch upon this in the following section by returning to the 2x2 matrix that organized the implications of caste to MNEs.

Impact of policy implications for different types of MNEs
To provide a little more insight, despite the limited research, we speculate here how policies we discussed above would affect the different MNEs in our matrix. Since the influence of caste is the strongest in Cell 1, if these policies are adopted, we believe, they would have the most impact on MNEs in this cell, i.e., Indian-origin MNEs operating in India. However, the adoption of such policies may be challenging, because of homogeneity along caste lines in internal organization and inter-organizational relationships, which are deeply embedded in the Indian institutional context. Therefore, because of the reluctance by Indian-origin MNEs operating in India to consider caste as a dimension of diversity, policies that aim to address caste may be subtly overlooked, or not implemented. The impact of caste-discrimination policies on non-Indian MNEs operating in India (Cell 2) may also be strong, provided they are adopted. As mentioned, due to imprinting mechanisms, non-Indian MNEs may overlook policy recommendations to maintain relations with particular castes, or ignore the plight of other castes, if they are able to increase ease of operations. The lack of awareness of caste on part of these non-Indian MNEs, may prevent them from actually implementing such policies, or they may choose to ignore any violation of them when they are implemented, to access particular resources and build relationships with particular castes who control key GVC activities.
Indian MNEs may be pressured to adopt human rights, anti-racism, and affirmative action policies in non-Indian contexts (Cell 3) to maintain legitimacy in particular institutional contexts and countries, especially when caste is acknowledged within policy at the macro-level and mirrored by other organizations. However, due to home-country imprinting, some caste-related practices within the organization may continue to persist. Policy interventions may see most impact on the global operations of non-Indian MNEs (Cell 4) because these organizations may have greater leverage in monitoring and dictating intra-organizational and inter-organizational relationships, but this may also be contingent on the caste composition of the workforce in the industry.
As discussed above, policy interventions related to caste may have a differential impact on MNEs as based on their geographic origin and scope of operations, which provides scholars with suitable empirical settings to research the effect of caste diversity, costs of caste homogeneity, and mechanisms to improve caste diversity. First, scholars can study whether caste diversity would: yield diverse perspectives, resulting in richer knowledge exchange across the organization (Becker-Ritterspach & Raaijman, 2013) or whether it might have unforeseen implications (e.g., frictions in teams hampering cooperation); help MNEs promote and assess human capital that is observed or measured, rather than inherited social and cultural capital; improve organizational ethics and performance (Porcena et al., 2020); help MNEs gain a more committed and loyal workforce by promoting the representation of disadvantaged castes (Saini & Budhwar, 2008) and derive performance advantages, similar to those experienced by firms that employ more women (Siegel et al., 2019) or whether there are also contrary effects (e.g., loss of reputation or stigmatization among the current talent pool, which in turn leads to performance losses); and, finally, whether MNEs can gain further contextual intelligence about the middle and lower segments and effectively tap into those large markets by developing products and services that meet the needs and aspirations of consumers in these segments (Dhanaraj & Khanna, 2011;Khanna, 2015).
Second, scholars can study the costs of lack of caste diversity, and of inequalities more broadly. For example, studies suggest that not addressing the racial wealth gap will not only have a detrimental effect on human development for Black communities, but will also inhibit the U.S. economy holistically via its dampening effects on consumption and investment, e.g., lack of economic activity, lower incomes, debt, lack of savings. Together, these are expected to cost the U.S. economy between $1 and 1.5 trillion annually between 2019 and 2028, and about 4-6% of the projected GDP in 2028 (Noel, Pinder, Stewart III, & Wright, 2019). MNEs and scholars can conduct similar studies on the broader economic effects of caste-based inequalities.
Finally, to identify or develop tools to increase caste diversity, scholars can study diversity initiatives that have targeted other demographic categories, particularly race, because it shares similarities with caste (as discussed earlier). Recently more stakeholders and MNEs are paying greater heed to racial discrimination. Ernst and Young decided to fully fund ethnically diverse college students across the U.S. to expose these students to an array of career and leadership opportunities (Iacone, 2021). These positive developments, i.e., increasing and clearer commitments to gender and racial equity, provide opportunities for IB scholars to study these and inform the development of similar initiatives for caste and other inequalities. As issues of equity and inequality gain central place in many conversations around the world, MNEs would face pressures from Local, yet global stakeholders to follow universal values and develop policies and initiatives to address inequalities and discrimination in all contexts in which they operate and based on attributes relevant in those contexts (Donaldson, 1996;Melé & Sánchez-Runde, 2013).
In conclusion, the implications of caste as an informal institution on economic and organizational outcomes in MNEs are likely to be distinct and deserve more research attention. By researching caste and its implications, IB scholars can generate novel insights on institutions and inequality, ranging from the characteristics of informal institutions to how they might lead to the reproduction of inequalities in organizations. Further, IB research on caste can also yield important implications for policies, across levels, which are related to human rights, affirmative action, engagement with Indigenous Peoples, and inclusion. These policies can help MNEs to address inequalities within their boundaries, in their GVCs, and more broadly in the societies they operate in. As IB scholarship encourages studying non-Western contexts, and given the growing importance of South Asia to the world economy, IB scholars are well placed to study the role of caste as an informal institution and generate new insights useful to IB scholarship as well as to broader organizational scholarship. certain number of positions (i.e., quotas of 15% for Dalits and 7.5% for Adivasis) in politics, education, and government jobs for individuals from these underprivileged castes, based on their proportion in the population at the time of independence (Hoff, 2016a(Hoff, , 2016bMunshi, 2019). In addition, the Indian government has created institutions with judicial power to further the constitutional ideals of justice, equality and liberty to the historically disadvantaged castes.
Building on the practice of providing quotas to the historically disadvantaged groups, the Indian government also provided a quota of 27% to Other Backward Classes (predominantly Sudras) in educational and government institutions in 1990 (Jaffrelot, 2000). The castes deemed ineligible for affirmative action came to be known as Other Castes (or OC), and predominantly included the three upper varnas (Brahmins, Kshatriyas, and Vysyas) as well as a number of other Sudra castes. Additionally, in 2019, the government provided for a further 10% quota in educational and government institutions to those not covered by previous quotas (i.e., Other Castes) and are economically disadvantaged. The constitutional validity of this measure is currently under judicial review.
The quota system and its use for research in various population surveys has resulted in Indian society being classified into four broad groups: (i) Others (Other castes, also known as forward castes or upper castes and predominantly include Brahmins, Kshatriyas, Vysyas, and some Sudra castes that have achieved socioeconomic progress), who are estimated to constitute about 31% of the population 21 (Deshpande, 2010); (ii) Other Backward Classes (OBCs), largely Sudra castes who constitute about 40% of the population; (iii) Dalits (also known as Scheduled Castes, outcasts, untouchables, harijans, SCs), who constitute about 20% of the population; (iv) Adivasis (also known as Scheduled Tribes or STs), who constitute about 9% of the population.
Although the government has outlawed caste discrimination and initiated affirmative action to advance historically marginalized groups, the caste system continues to shape the socioeconomic lives of many Indians. For example, 27% of the population have reported practicing some form of untouchability, with 52% of Brahmins (at the top of the hierarchy and furthest from Dalits) reporting the practice, followed by 33% of OBCs, who sit right above the Dalits in the caste hierarchy (Thorat & Joshi, 2015). Further, endogamy is widely prevalent, with nearly 95% of adults marrying within their own caste, and openness to marry outside the caste is low, even among those who have migrated to developed countries (Narzary & Ladusingh, 2019;Rajadesingan et al., 2019).
Researchers have studied caste discrimination in various aspects of social life, including housing (Thorat et al., 2015), receipt of aid during natural calamities and their recovery from them (Aldrich, 2010(Aldrich, , 2011, imposition of caste occupations during natural disasters , and status ascription and social relations in elite educational institutions (Pathania & Tierney, 2018). Newspaper reports provide more detailed accounts of discrimination, such as segregation of lower castes in classrooms and separate water sources for them, as well as assaults on them and killings for various violations, including wearing a gold chain, attending a community festival, using an uppercaste surname on Facebook, dressing in the attire of upper castes, riding a horse, eating in front of the upper castes, and adopting the wedding practices of the upper castes (e.g., wedding procession, riding a mare).
Overall, even though it remains difficult to comprehensively and accurately capture the caste system, it is fair to suggest that it is an institutional feature that determines one's position in a graded social hierarchy and thus their rights and privileges related to education, work, and social and family life.
ABOUT THE AUTHORS Hari Bapuji is a Professor in the Department of Management and Marketing, Faculty of Business and Economics, The University of Melbourne, Australia. His research interests lie at the intersection of business and society, with a particular focus on how economic inequality affects businesses and vice versa. His research has appeared in leading management journals and has also been noted for its impact on practice and policy. He serves as a coeditor of Business & Society and is a co-founder of Action to Improve Representation.
Snehanjali Chrispal is a Lecturer/Assistant Professor in Management at Deakin University, Australia. Her research interests lie in inequality and organizations, particularly focusing on human rights issues like caste and gender, using institutional and critical theory. She has published in journals Local, yet global