DIVERSITY AND INCLUSION PROGRAMS IN MULTINATIONALS: OPPORTUNITY FOR REVERSE KNOWLEDGE TRANSFER

ABSTRACT This article aims to understand how multinational enterprises manage their diversity and inclusion (D&I) programs globally, identifying strategies, subsidiaries’ motivation to internalize practices, and addressing an issue little discussed in D&I literature to date: the reverse knowledge transfer (RKT). Through an exploratory multiple case analysis, this study contributes by integrating and advancing literatures of D&I, international business, and RKT, to the best of our knowledge not yet done. It highlights the importance of the headquarters’ strategy, the subsidiaries context as well as their relative size and power inside the organization, identifying main drivers for D&I and conditions for the RKT, culminating in a summarizing framework.

Although the issue has been on corporate agendas for quite some time now, its relevance and outcomes are far from being homogenous worldwide: while some countries are recognized for their egalitarian values and opportunities, others struggle with hierarchical social structures that project onto the work environment.MNEs play a pivotal role when it comes to this reality, not only because they must manage a diverse multinational workforce and cope with different legislation concerning D&I in every country they operate but also due to pressures to accelerate the pace of change in the world, exemplified by the United Nations' sustainable development goals (Newburry et al., 2022;Silva et al., 2020).Large, visible, and active MNEs in countries with different norms and standards than their home country face growing societal pressure to be accountable for the respective local social, environmental, and ethical issues (Kolk, 2016).
By adopting Bartlett andGhoshal's (1986, 1989) work as theoretical lenses, this article aims to understand how MNEs manage their D&I programs globally, identifying the overall strategy and transfer practices chosen by headquarters, subsidiaries' motivation to internalize such practices.Moreover, it addresses an issue little discussed in the D&I literature to date: the reverse knowledge transfer (RKT).Through an exploratory in-depth multiple case study, we dive into seven cases of global D&I programs involving headquarters, their Brazilian subsidiaries, and their transfer of practices.We analyze the approach chosen for globally managing the programs, the reasons behind it, along with home (headquarters), host, and industry contexts.As such, we contribute academically by advancing the literature of D&I in international business and RKT, to the best of our knowledge, not yet done.This paper empirically highlights and identify a) the importance of the context of countries hosting subsidiaries; b) the subsidiaries' relative size and power inside the organization; c) the main drivers for D&I; and d) conditions for the reverse transfer of practices to take place, offering a summarized framework of global D&I.Managerially, our findings can assist MNEs in designing their global D&I policies by providing considerations on successful D&I policy implementations, and by stressing the relevance of knowledge/practices generated at the subsidiary level, especially those located in developing and less egalitarian environments.This would also have an important social impact globally, given the relevance and reach of MNEs.

International business: a recurring dilemma
Global firms' main challenge regarding international human resource management is rather aligned with the great dilemma of MNEs in prioritizing either integration, coordination, and globalization advantages or differentiation, responsiveness, and localization advantages (Bartlett & Ghoshal, 1989).This dilemma can be understood as the institutional duality subsidiaries face to accommodate both home and host countries' institutional pressures (Cooke et al., 2019).For Bartlett and Ghoshal (1989) MNEs can opt for a multidomestic, global, international, or transnational strategy, depending on how much they value local responsiveness and global integration.Despite the common use of the terms global and multidomestic approach, these authors created the 'transnational' approach, but researchers in D&I rarely mention it.The transnational approach is appropriate when both local responsiveness and global compliance are needed.For Bartlett and Ghoshal (1989), the MNEs' role would be to coordinate the interdependent organizational units so that they would all share the same strategic vision and work towards a unique goal.
In another work, Bartlett and Ghoshal (1986) argue for assigning different roles to subsidiaries (strategic [regional] leaders, contributors, implementors, and 'black hole') depending on the strategic importance of their local environment and the competence of their local teams.With differentiated subsidiary roles and more dispersed responsibilities, headquarters could better explore underutilized human assets and capabilities, better detect opportunities, and develop more appropriate responses to threats.Subsidiaries with strategic leadership roles would be of great importance to headquarters, becoming a partner when making decisions and formulating policy.

Global D&I management and transfer
Specifically considering global D&I management, the multidomestic approach has been the mostly reported, given the primordial role played by cultural distance in this specific realm (Lauring, 2013).Adopting the same practices across subsidiaries might result in a lack of legitimacy and relevance within different societal contexts because every country has unique demographics in terms of minorities and cultural patterns of diversity (ÓBešić & Hirt, 2016).Studies consistently show that local context is of utmost importance when designing or transferring D&I practices, especially when headquarters and subsidiaries come from very different countries regarding development or location (developed/developing or north/south, western/eastern) (Newburry et al., 2022;Saqib & Khan, 2022).On the other hand, the global approach has been appointed as more sustainable in the long term, due to the advantage to leverage corporate resources, support, and experience.Egan and Bendick (2003) presented the case of the Royal Dutch/Shell Group, emphasizing that their global top-down approach may have taken longer to implement, but was the best choice, as they had a central core in all operations and headquarters' support.Recently, scholars have called for a compromise in terms of the global integration versus local responsiveness challenge, arguing that human resource practices should be adopted to fit both home and host country environments and emphasizing concepts such as hybridization (Yahiaoui, 2015) and synergy (Oppong, 2018).
In a reconciling effort, Sippola and Smale (2007) suggested the answer might be to develop a broader global concept of diversity philosophy that could be translated into more demographically and institutionally embedded policies and practices.On the same line, Umeh et al. (2022) found that D&I policies required a universal and common core alongside specificities to reflect the local context.Nishii and Özbilgin (2007)'s framework sees the best approach lying somewhere between the global and the multidomestic, defined individually according to the MNE's organizational structure, strategic objectives, and the different environments in which it operates.The "ultimate approach" should "facilitate knowledge creation, disseminate and organize initiatives across domestic diversity management offices, equip local units with the necessary skills and resources to move their diversity initiatives forward based on shared experiences in the branch network, and develop global strategies that transcend the narrower perspectives that emanate from domestic networks" (Nishii & Özbilgin, 2007(Nishii & Özbilgin, , p. 1885)).
There are also critical approaches to global D&I management.Syed and … Özbilgin (2009) propose a relational framework that considers the macro-national, meso-organizational and micro-individual levels of analyses to have a comprehensive understanding of the factors impacting the transfer of practices and avoid criticisms due to non-compatibility and cultural imperialism.On that same note, Fernando (2020) states that in MNEs, headquarters are usually perceived as "developed, modern, and superior" (p.2126) in comparison to subsidiaries, which end up adapting to the parents' models of diversity practice and becoming mere implementors of global practices.Santos (2020) analyzed how the MNE's country of origin may have affected D&I management on Brazilian subsidiaries, noting that the diversity issues they brought up were different (racial vs. cultural diversity) depending on the MNEs' origin and societal structure.The cases portrayed mainly a top-down transfer of programs, disregarding the local context.In addition, the author stated that many adopted practices were "just for show" and did not translate into real change and inclusiveness.
The efficient management of human resource practices is not entirely a function of the strategy chosen by headquarters but also depends on how the subsidiaries embrace such transfers (Ferner et al., 2012).The simple notion that subsidiaries are dependent and occupy lower positions in this hierarchy may justify a transfer and implementation of a set of practices, yet not their internalization, which happens when employees attribute meaning and value to those practices (Kostova & Roth, 2002).In matters where the subsidiaries and the headquarters' interests are aligned, the power capabilities of the subsidiary are likely to be supportive of the transfer of practices, even removing or circumventing possible host institutional obstacles (Ferner et al., 2012).63 (5) | 2023 | 1-22 | e2022-0007 eISSN 2178-938X

Reverse knowledge transfer in MNEs
A topic of increasing interest in international business has been the reverse knowledge transfer (RKT), which happens when subsidiaries transfer any type of knowledge or practice to their parent companies, in an unorthodox flux (Li et al., 2016).MNEs, with a dispersed network of units, can and have increasingly benefited from knowledge generated in those locations to create overall global value (Anderson et al., 2016), as subsidiaries can contribute not only to headquarters but to other subsidiaries (Cuervo-Cazurra & Un, 2004).However, the actual transfer of knowledge or practice in such an unorthodox flux is not always easy.For the RKT to be fruitful, headquarters must (a) be aware of the knowledge or practice to be transferred (Ciabuschi et al., 2015); (b) value it (Pérez-Nordtvedt et al., 2015) or miss it (Nair et al., 2015); and (c) have the necessary absorptive capacity (Ambos et al., 2006;Minbaeva et al., 2014).Usually, "knowledge is valuable to headquarters when the subsidiary had access [to it] by virtue of being in a certain location, with certain players, or in a certain context of superior or different knowledge stock" (Kogut & Mello,p. 17).As such, RKT is more likely when the subsidiary is large (Bezerra et al., 2015) and located in a developed country (Ambos et al., 2006).
The RKT is facilitated by the presence of (a) direct transfer mechanisms; (b) binding mechanisms; (c) social interactions; (Ambos et al., 2006;Chung, 2014) (d) similar cultures and a shared vision (Nair et al., 2015); (e) specific organizational structures such as international teams or projects (Minbaeva et al., 2014); and (f) formal rules and procedures (Chung, 2014).As can be inferred, efficient human resource practices are much needed for successful knowledge/practice transfers, which is why they are often discussed in terms of transfer mechanisms in RKT studies.However, studies in which human resource practices are the object of study are rare.

METHOD
This paper aims to understand how multinational enterprises manage their D&I programs globally, identifying their strategies, transfer practices, and motivations.The unit of analysis is the MNE.
The multiple case study method of investigation was adopted since it is specifically suited for investigating in-depth, contemporary (Yin, 2003), and global phenomena (Ghauri, 2004), such as the global management process of D&I policies.Although the method does not allow for statistical generalizations, it can provide valuable insights that contribute to theory building (Thomas & Myers, 2015).Research reliability was ensured using protocols and systematic procedures throughout all research stages, such as closely following a research design (Yin, 2003), with interviews with semi-structured guidelines and protocols, transcribing interviews, and keeping a solid database with all information gathered, not to mention triangulating different sources of data.

Case selection
Cases were selected by theoretical sampling (Eisenhardt & Graebner, 2007), with companies chosen from the same industry/country for comparison.The selection criteria were threefold: (1) relevance to the theme under investigation to allow significant insights; (2) a large percentage of MNEs in the industry; and (3) ease of access.The Brazilian insurance industry met all criteria.First, in Brazil, women and non-whites account for most of the population, yet white men still occupy 90-95% of the highest-paying positions (Ethos, 2016).In the Brazilian insurance industry, women account for 55% of the workforce but only 26% of executive-level positions (Galiza & Monteiro, 2019).Aware of these deficiencies, and as a reflection of the institutionalization of the D&I theme in the industry, the insurance companies in Brazil have publicly endorsed D&I management with the founding of Idis -Instituto pela Diversidade e Inclusão no Setor de Seguros, an institute for D&I in the insurance sector.Idis aims to foster awareness by engaging all stakeholders, encouraging initiatives, and tracking the diversity scorecard.Second, insurance is an industry with a large share of MNE activity in Brazil.Indeed, 47% of corporate ownership originates in foreign countries 1 , which allows for the study of international practice transfers, as intended in this research.This set of conditions ensured that an adequate number of international companies participated in the study.Finally, our ease of access increased substantially when an annual global event to foster the development of an inclusive and diverse workplace in the insurance industry was held in Brazil in September 2019 with presentations from the MNEs, which we could attend (Dive In Festival).

Data collection
Data collection consisted of both primary data (interviews) and secondary data, dully triangulated to "search for convergence among multiple and different sources of information" (Creswell & Miller, 2000, p. 126) and enhance the study's internal validity (Creswell & Miller, 2000;Pauwels & Matthyssens, 2004).Table 1 summarizes the data sources.Table 2 gives further information on the interviews held.Companies were free to nominate anyone they thought could best talk about D&I practices, so the respondents' positions varied, providing diversity in data collection.The interviews conducted followed a semi-structured questionnaire based on the theoretical background to guide and elicit the conversation, as proposed by Rabionet (2011).The participants did not agree to their data being publicly shared, so supporting data is unavailable.Additional information and possible clarifications were further provided by email.Triangulation was applied between data sources and methods (interviewee vs. lecturer from the same company presenting in the Dive In Festival; primary versus secondary sources of information, different types of secondary sources etc.), investigators, and theories (using the abductive method of investigation allowing us going back and forth when analyzing the data) (Denzin, 1978).This allowed data sorting, common themes' emergence, and increasing validity (Creswell & Miller, 2000).It is worth noting that many secondary data sources had the companies' names and could not be cited in our references for confidentiality reasons.
Table 3 presents an overview of the subsidiaries' characteristics.

Data analysis and theorization
Data analysis followed an abductive approach (successively going back and forth from theory to empirical observation), as proposed by Dubois and Gadde (2002).Specifically, we have systematically and concurrently combined the theoretical background (initially D&I and international human resource management, then IB and RKT), empirical fieldwork (the seven cases), and case analysis to ensure the article's proximity to reality and suitability to the chosen theoretical lenses (Bartlett & Ghoshal, 1986, 1989).
Each case was first individually analyzed and color-coded with recurring subjects, which generated a matrix juxtaposing all cases and all subjects.This approach clarified the emergent patterns and enabled comparison and cross-case analysis.After conducting pattern-matching (Pauwels & Matthyssens, 2004;Yin, 2003) between the cases and literature, 1) the data were organized in tables (4 and 5) with classifications and the quotes that served as evidence of each construct and 2) further theorized/developed into figures 1 and 2. Following these procedures to establish a "clear chain of evidence" (Yin, 2003, p. 127), we expect to have ensured the study's validity.

D&I management approaches
Most analyzed subsidiaries are large (six out of seven) and have formal D&I management structures in Brazil (five out of seven).Also, all seven cases had a global philosophy on how to treat and respect their employees: Some were more focused on D&I issues than others, but all subsidiaries mentioned D&I as a general guideline.Despite these similarities, how the MNEs manage and transfer their D&I programs to the subsidiaries varies considerably, as per Table 4.  Three cases out of a sample of seven (1, 2, and 5) adopt a global approach, with D&I policies being decided by headquarters and locally implemented.These companies follow an 'implementer' role (Bartlett & Ghoshal, 1986).In Case 1, the Brazilian subsidiary is very small and has a limited number of employees, so it is unreasonable to have bottom-up, specifically designed D&I programs or practices.Also, because the MNE is globally recognized as one of the most active in fighting for equality in the insurance industry, the subsidiary acts as part of its global platform to raise awareness in the industry by hosting and participating in relevant events.Case 2, on the other hand, represents a subsidiary that started implementing D&I initiatives locally before the theme gained relevance to headquarters.Although they emphasized that the global approach adopted later was beneficial to them as the practices gained importance and structuring, the subsidiary sometimes feels like it is being held back by having to follow the same goals as every other unit, including those that had not even been discussing the matter beforehand.The subsidiary in Brazil could have developed into the role of 'strategic regional leader' for the MNE, especially considering its importance in terms of total employees and revenues (see Table 3), having more autonomy, and being more than just an implementer of standard practices.
We have a strong global network working on gender parity issues.We feel we were ready to advance a bit more and work on gender parity transversely.
Case 5 was given the opportunity to contribute more and adapt practices but did not feel the need to do so: global leaders asked local managers if they would like to choose their priorities to guide their corporate social and sustainable responsibility policies, which includes the management of diversity practices.Instead, they followed the same pillars of the global guidelines because they found them relevant and adequate.The local managers believe that "doing something together with the whole world is really engaging and powerful."Its relevance in terms of contribution to the MNE's global income and the fact that the Brazilian subsidiary is one of the only two based in South America (Table 3) could justify its role as a 'contributor' or strategic regional leader if they wanted to take a more active and participatory role in the management of D&I globally.
Case 3 is a classic example of a multidomestic approach.Its headquarters believe in the relevance of D&I practices and therefore provide a global structure with general guidelines on the issue, such as the global event they held on women's equality.Yet, they believe that the practices should be developed according to each subsidiary's context, so they encouraged the Brazilian unit to start discussing the matter internally, on their terms, and organized as they felt was best.
Finally, Cases 4, 6, and 7 were deemed part of a transnational D&I strategy, with headquarters setting a global structure including a broad set of values to engage and encourage subsidiaries in developing adapted initiatives toward a common goal.Although this transnational D&I strategy represents a compromise between the global and multidomestic approaches, it does not exactly fit the concepts of a "hybrid" (Yahiaoui, 2015) or "synergistic" (Oppong, 2018) approach.These subsidiaries are not just transforming or reinventing practices based on stakeholder strategies, host country institutions, or the relational context, like Case 2; they are actively participating in building those practices.Indeed, Case 4 reported that before a global strategy was in place for D&I management, they already had their own initiatives, as did other subsidiaries from other countries.When the headquarters decided to unify all of them, the theme gained importance, and the practices became much more robust and better structured, with "defined goals, key performance indicators, monitoring, regional and global reporting, etc."The interviewee was designated as the regional leader for D&I management and participated in the global committee that discusses, shares, and collectively decides the program's course whilst maintaining awareness of local distinctiveness.Moreover, Cases 4, 6, and 7 exhibit what Bartlett and Ghoshal (1986) have termed "strategic leader" behavior among subsidiaries, at least regarding D&I programs.Subsidiaries created value for headquarters, helping them define their global D&I, either as a co-creation or through the "export" or transfer of their practices to be incorporated into global norms.Headquarters have therefore benefitted from RKT and "hitchhiked" on subsidiaries' existing policies, adopting RKT practices, as can be seen in the transcript of case 6: There are opportunities to share our D&I practices.We make monthly calls; there is a quarterly newspaper showcasing the best practices, and an annual award.We were awarded "best place for women to work" [...] I have personally gone to headquarters to share some of our gender practices to share how the work environment can be more inclusive for women.
The cases of RKT came from large subsidiaries (Bezerra et al., 2015) located in environments more likely to generate knowledge on D&I (Kogut & Mello, 2017).Both Switzerland (headquarters Case 4) and Japan (headquarters Case 6) are small countries with much less diversity than Brazil, which probably faced far fewer D&I issues.Although Brazil cannot be considered a benchmark for gender parity, it ranks better than Japan in the Global Gender Gap Index from the World Economic Forum 4 .Considering that Japan ranks at 121st place out of 153 countries, it makes sense that they would be interested in learning from a subsidiary that earned the title of "best work environment for women" in their country.As for Case 7, the knowledge creation and transfer opportunity arose from not having a formal global initiative but rather one emerging at the subsidiary level.The local context, both industry and nationwide, was favorable to igniting this discussion and bringing it to the MNE, making the subsidiary act as a Strategic Leader in this matter despite its apparent minor relevance to the MNE (i.e., one of the ten subsidiaries located in South America, generating only 3% of total income, as per Table 3).The remainder of this paper expands this argument, exploring the drivers of the D&I initiatives and analyze the conditions for successfully transferring D&I management programs.

Drivers of D&I practices
Although the study did not quantitatively measure or compare the success of the internalization of practices among the cases, most showed signs of achievements in making the company culture more inclusive and raising employee awareness about improvements in their working environment.Many participants highlighted the engagement and support of the subsidiary's leadership as drivers of the initiatives' success.However, the employees' genuine interest was also deemed beneficial, meaning that the practices were being incentivized both top-down and bottom-up.This finding is consistent with (a) Ferner et al. (2012), who argue that transfers will be facilitated when subsidiaries' interests are concordant and homogenous amongst the many actors, and (b) Kostov and Roth (2002), who suggested that the host environment/context must be a priori amenable to such a practice for it to be fully implemented and internalized in the event of a transfer.
The host country environment-considering both the country and the industry-was also considered favorable to either developing or implementing D&I programs and practices.When asked about the motivation for moving this agenda forward, interviewees brought up aspects related to (a) keeping up with their counterparts and the insurance industry as a whole; (b) consumer trends and pressures; (c) the belief that a diverse work environment is financially beneficial for business; (d) willingness to build the company's brand as a good and inclusive workplace to attract talent, and contribute to society; (e) positive mindset vis-à-vis dealing with foreigners in a daily basis, amongst others, consistent with previous research (Jonsen et al., 2019;Silva et al., 2020).Table 5 depicts evidence of this context that seem to have played a part in facilitating not only the internalization of D&I programs, but also the RTK.

Driver Evidence
External (Market) "I have been in the insurance industry for many years, and I see that we have been fighting to make sure we become a more diverse environment, with equal opportunities career-wise.We are pushing the agenda through our institutions.We hold many events per year.We are also included in the global discussion and have made much progress."(Case 1) "Our decision to address D&I was very much in line with the market and to address our consumers' concerns.We needed to reflect better the markets where we operate."(Case 4) External (Legal requirements) "All companies face the quota law issue, so everyone started by hiring people with disabilities."(Case 2) Internal (Leadership) "I believe our diversity project has been yielding results because our regional and global CEOs sponsor it […] Human resource's work is generally not seen as a priority in corporate agenda, but I think we are moving forward because they see this work as strategical and relevant."(Case 4) Internal (Need felt) "We identified the need to work on diversity in general -regarding gender, race, etc. […] Also, as part of a multinational, we need to address cultural diversity because we talk to people from other countries in addition to dealing with Spain all the time."(Case 2) "It was the right thing to do, and made a lot of business sense […] Innovation only comes with different mindsets."(Case 4) "When we hire people with different mindsets and profiles, respecting the individual, we become a much more powerful company precisely because we have different voices.Listening to every voice makes people proud to belong, contribute, take center-stage, and bring innovation to the organization."(Case 6) Internal (Employees) "Last year, we created our diversity group to exchange ideas and engage our employees in the discussion of D&I.We held an internal campaign looking for volunteers and were surprised to hear back from more than one hundred people![…] The outcome was way better than expected."(Case 6) "We understood that talking about this topic in a more structured way was a potential driver for our company as an employer brand."(Case 7) Source: Elaborated by the authors.

DISCUSSION
It is common ground that the most relevant aspect to consider when managing global D&I programs is cultural distance.This occurs since mistakes are often made when headquarters try to impose practices on a host country that does not have the same diversity issues (Ferner et al., 2005) or the host country is not interested in changing the existing inequality structure (Alhejji et al., 2018).Therefore, researchers often advocate for the "best fit" approach (ÓNishii & Özbilgin, 2007), in which headquarters would support its subsidiaries, providing a unified philosophy and a structure to encourage them to develop their locally embedded practices.
This article argues that this promising way out actually fits the transnational approach of Bartlett and Ghoshal (1989), who observe that the headquarters coordinates all its units and encourages them to pursue their inherent strengths and eventually transfer their knowledge/ competence to other affiliates, acting as an interdependent organization.The case studies provided evidence that following a transnational approach for managing D&I issues can lead to RKT between subsidiaries (Cuervo-Cazurra & Un, 2004) and from subsidiary to headquarters, enhancing overall global value for the MNE (Ciabuschi et al., 2015).In taking advantage of RKT from subsidiaries with more challenging and therefore developed (in this regard) contexts, headquarters can accelerate the development of their D&I policies.Although this reality might be specific to the experience of the insurance industry in Brazil, it does offer some insights, which are summarized in Propositions 1 and 2: Proposition 1. MNEs following a transnational approach to D&I programs have higher chances of benefitting from RKT. Proposition 1a.MNEs with a strong global D&I program tend to benefit less from RKT.
Proposition 2. The transnational approach tends to be more successful in managing D&I programs.
Apart from investigating how MNEs manage and transfer their D&I programs amongst their subsidiaries, this study distinguishes decision-making aspects from headquarters (chosen strategy) and facilitating aspects from subsidiaries (host country and industry contexts).Figure 1 portrays those aspects and shows how a well-functioning relationship between headquarters and its subsidiaries can enable an exchange environment for implementing and internalizing D&I programs amongst its units.When it comes to globally managing D&I programs, headquarters must decide if they will follow a global, multidomestic, or transnational strategy (Bartlett & Ghoshal, 1989) based on the cultural distance between home and host countries.This choice could be aligned with the overall corporate strategy in terms of following or not following a global integration or a local responsiveness logic.However, being consistent is advised (Nishii & Özbilgin, 2007).
Similarly, the subsidiaries' context is also very relevant for determining the success of the transfer and internalization of practices (Fernando, 2020;Kostova & Roth, 2002).One should not overlook national-level factors such as societal, cultural, climatic, political, economic, and labor market (Bešić & Hirt, 2016) elements.This study considered Brazil's cultural and hierarchical reality, focusing on the insurance industry conditions that may have allowed for subsidiaries to express their keenness to implement and internalize D&I management practices, thus engendering successful transfers (Ferner et al., 2012) and incentivizing reverse transfers.The subsidiaries' relative size and power within the MNE's network defined their assigned/assumed roles as implementers, contributors, or strategic (regional) leaders, which were also proven key aspects when considering the stage of maturity, the importance, and the success of D&I practice exchanges.Subsidiaries assuming the role of strategic leaders felt they had the necessary support and autonomy to conduct their initiatives successfully.Implementers could have been more engaged but were not, either of their own accord or by the volition of the headquarters.• Context (country and insdustry) • Relative size

Headquarters Subsidiaries
Source: Elaborated by the authors.
Based on Bartlett and Ghoshal's (1989) matrix of local responsiveness versus global integration, a transnational approach is more likely and adequate for cases where having D&I policies is highly important both to headquarters and the local context.Additionally, if the subsidiary has a historical/ positive stock of knowledge and relatively high power and/or size in the organization, there are higher chances of D&I RKT and/or D&I co-creation (evidence of Cases 4, 6, and 7) (Figure 2).If, however, having D&I policies is high in importance to headquarters but low in the local context, a global approach is more likely (Cases 1, 2, and 5), even if the subsidiary had an initial stock of knowledge (Case 2).Finally, if having D&I policies is not much valued at headquarters but is highly important to the local context, a multidomestic approach is probably the way to go (Case 3).In this case, D&I policies are created at the subsidiary level, and although possible, transference is less likely.The higher the stock of knowledge and the relative importance of the subsidiary (power and size), the higher the chances of a possible RKT or lateral transfer.The international approach does not apply to our research, as it would not refer to MNEs interested in D&I.
This article contributes to advancing the theme by bringing together concepts from D&I, international business, and RKT literature, to the best of our knowledge, not yet put together.Figure 2 graphically represents this intersection.It advances the discussion of the "best fit" approach (Nishii & Özbilgin, 2007) by bringing in the seminal work of Bartlett andGhoshal (1986, 1989), which is considered "the most influential set of organizational studies in the history of international management" (Rugman, 2002, p. 37), but surprisingly not much mentioned by D&I studies.The cases provided evidence that concepts such as the transnational approach and subsidiaries' strategic roles can be useful for advancing research on D&I international management and subsidiaries' influence in determining its success to constitute "a source of organizational identity and competitive advantages" (Kostova, 1999, p. 312).Furthermore, we corroborate that sometimes the subsidiaries can stimulate the emergence of new knowledge/ practices that are later shared with all MNEs' units, challenging the assumption that the traditional knowledge flow in D&I is the only available approach (Fernando, 2020).Also, by integrating international business, D&I, and RKT literature, this study moves beyond the traditional strategy and structure model of human resource discussions to a more processoriented and knowledge-based one (Dickman & Muller-Camen, 2006).Finally, literature on RKT usually dives into human resource literature when discussing transfer mechanisms such as socialization and binding mechanisms, but not as the object (knowledge or practice) of transference, which makes this paper original in its approach.Managerially, the study can assist MNEs in designing their global D&I policies by providing considerations on a successful D&I policy implementation; and by stressing the importance of considering knowledge/ practices generated at the level of subsidiaries, especially those located in developing and less egalitarian environments.This would also have an important social impact globally, given the relevance and reach of MNEs.
Limitations of the study are mostly related to method and sampling (i.e., seven cases from the same industry and country).The research sought only managerial generalizability and not statistical generalizability, focusing on the same industry/country for comparability reasons.Quantitative studies based on a larger and more diverse sample should be able to test and compare the hypotheses and framework raised in this research, leading to robust contributions to theory-building.It would also be promising to investigate further the effectiveness of adopting the transnational approach to managing D&I programs globally, differentiating between cases emerging from headquarters or subsidiaries.In addition, the relative importance of top-down versus bottom-up engagement with D&I practices in subsidiaries may be further studied.Finally, similar studies, including interviews with the subsidiary, headquarters' top management, and employees in general, could generate some additional contributions.

Figure 1 .
Figure 1.Framework for Managing D&I Programs Globally

Table 2 .
Description of interviews

Table 3 .
Description of subsidiaries in the sample *This figure includes income from Brazil and Colombia because the company disclosed only aggregate data.Source: Elaborated by the authors, based on data from company websites and annual reports.FGV EAESP | RAE | São Paulo | V. 63 (5) | 2023 | 1-22 | e2022-0007 eISSN 2178-938X