Reverse Knowledge Transfer in Multinational Companies: A Systematic Literature Review

The knowledge transfer between headquarters and subsidiaries is an important research topic in International Business (IB) studies (Alharbi & Singh, 2013). However, the mainstream literature has focused on knowledge transfers from parent companies to subsidiaries (Dunning, 2001; Johanson & Vahlne, 1977, 2009; Rugman, 2006; Vahlne & Johanson, 2014; Vernon, 1966, 1993), and not vice versa (from subsidiaries to parent company).

The literature on such bottom-up transfer, referred hereafter as reverse knowledge transfer (RKT), is still limited (Ambos, Ambos, & Schlegelmilch, 2006; Criscuolo, 2009; Hakanson & Nobel, 2001; J. Li, Strange, Ning, & Sutherland, 2016; Tseng, 2015), in spite of its growing importance to the knowledge generation (Frost & Zhou, 2005; Gupta & Govindarajan, 2000; J. Li et al., 2016; Tseng, 2015) of multinational enterprises (MNE) and its contribution to global competitive advantage (Chung, 2014; Eden, 2009; Frost & Zhou, 2005; Makela, Bjorkman, & Ehrnrooth, 2009; Tseng, 2015).

In addition to being limited in terms of the number of studies, the RKT literature is also limited in breadth, with papers usually falling short of analyzing the phenomenon as thoroughly and holistically (Pérez-Nordtvedt, Kedia, Datta, & Rashee, 2008) as possible, and often merely addressing a single aspect of the process. Thus, the literature has many gaps to be filled, and by failing to examine the company and the process as a whole, "the spread of innovation, technological, managerial and marketing advantages is not addressed" (Ietto-Gillies, 2005, p. 11).

Using a thorough systematic review of a 15-year period, from 2001 to 2016⁽¹⁾, in top tier journals, this present study finds interesting literature gaps to be filled and proposes a theoretical framework that conceptualizes the reverse knowledge transfer as a complex process with factors to be analyzed in detail such as the sender of knowledge (subsidiary), the receiver (parent), the direct and indirect mechanisms of knowledge transfer and the characteristics of the knowledge itself. In addition, the antecedents of the subsidiary and - most importantly - how the RKT benefitted the MNE are examined. In this work we intend to provide a detailed view of the phenomenon of reverse knowledge transfer, contributing to a better understanding of it and seeking to serve as a basis to assist corporate managers in global strategic planning and knowledge management and scholars in future academic research in the field.

To the extent that an understanding of how reverse knowledge transfer can benefit the parent company is interesting not only vis-à-vis internationalization but also vis-à-vis strategic management, we hope this paper to be all the more appealing. According to Autio (2005), most internationalization theories have tended to rely on either economic or organizational and behavioral theories; however, the internationalization context and its attendant theories have contributed little reciprocal insight to organizational and strategic theories. Autio therefore points to this as an important research gap and suggests empirical and theoretical academic works to be conducted "to better understand the sources and effects of the 'internationalization competitive advantage' as well as to articulate the practitioner implications of this potential effect" (Autio, 2005, p. 16).

This paper is organized as follows: second section provides a theoretical background on the importance of knowledge and knowledge management for IB and the subsidiary's respective role; third section explains the perceived research gap and how previous work has been framed. Fourth section explains the methodology used; fifth section discusses results and findings; and finally, sixth section concludes by listing some limitations and contributions of the research as well as possible avenues for future research.

Multinationals are, above all, in the relentless pursuit of international competitive advantages, meaning advantages over competitors based on some exclusive strategic resource (Barney, 1991). The Resource-based View (RBV) argues that for a company to have a competitive advantage its resources need to be VRIS: valuable, rare, difficult to imitate and difficult to substitute, as per Penrose (1959) and Barney (1991), among other authors.

Expanding on that came the knowledge based view (KBV), with several authors arguing "knowledge has emerged as the most strategically-significant resource of the firm" (Grant, 1996, p. 375; J. H. Li, Chang, Li, & Ma, 2014; Oliveira, 2007; Pérez-Nordtvedt, Mukherjee, & Kedia, 2015). Even though authors from other school of thought may not be so assertive about that, it is a consensus in the literature that after the shift from industrial to a post-industrial age, that knowledge is the most critical resource to firms (Nonaka & Takeuchi, 1995).

Accordingly, among the many benefits of internationalization for an organization, most scholars would agree that knowledge creation and sharing are important ones, so important, in fact, that noteworthy authors from several different schools in internationalization (Dunning, 2001; Johanson & Vahlne, 1977, 2009; Oviatt & MacDougall, 1994; Rugman, 2006; Vahlne & Johanson, 2014; Vernon, 1966, 1993) have frequently addressed the subject in their theories, paradigms and models. Although knowledge may not be explicitly central in some if not most discussions, it is important enough to garner attention, thus demonstrating its relevance and need for a better understanding of its mechanisms and flows in the context of multinational companies.

Figure 1 below summarizes the views of the main schools of internationalization as to where or with whom knowledge lies in the organization and its potential transferability, aiming not to contrast them but to show the importance of knowledge for internationalization theories in general.

Figure 1
Main Internationalization Theories and Their View on Knowledge and Knowledge Transferability
Source: prepared by authors. (*) date of first paper.

In this paper, knowledge is assumed to be a firm-specific strategic resource, one that is valuable, rare and difficult to imitate or substitute (Barney, 1991; Cuervo-Cazurra & Un, 2004; Loane & Bell, 2006), such as any kind of innovation, technology, management techniques, capabilities, marketing and production skills that can be transferred from one location to another.

As can be seen from Figure 1 above, most authors (Dunning, 2001; Johanson & Vahlne, 1977, 2009; Rugman, 2006; Vahlne & Johanson, 2014; Vernon, 1966, 1993) focused knowledge created at the parent company or headquarters level. It is reasonable to assume that this is because the main theories were conceived in periods preceding the current level of globalization and are rooted in the contexts of developed countries; therefore, they do not account for more recent phenomena such as emerging market multinationals or bi-directional information flows (Fleury & Fleury, 2007).

A very interesting, useful and more recent theory that may help address the above issue (EM MNE) is the Springboard Perspective (Luo & Tung, 2007; Petersen & Ivarsson, 2015). With EM MNEs being less likely to seek cost minimization strategies, as they are usually already located in low-cost countries, they usually follow a market-seeking or asset-seeking strategy. Strategic assets include technology, know-how, R&D facilities, human capital, brands, consumer bases, distribution channels, managerial expertise, and natural resources, all with the specific purpose of (a) strengthening their position at home, and (b) compensating any firm-level competitive disadvantages they may have. This perspective is thus of special interest to this study as it necessarily involves a RKT process. However, the springboard perspective does not specify any of the steps or actors in the process, stating only that "their resource commitment, especially investment size, is not necessarily a function of time, experience or learning" (Luo & Tung, 2007, p. 491). Regarding knowledge and knowledge transferability specifically, which is the focus of this research, the springboard perspective recognizes only the importance and progressive nature of learning.

Headquarters have been responsible for the bulk of a multinational's value creation and competitive advantages (Chandler, 1991; Ciabuschi, Dellestrand, & Nilsson, 2015) and "subsidiaries have traditionally been perceived as receivers of knowledge and theorized as inferior to HQ on several dimensions, including questioning the stock and value of knowledge they possess" (Michailova & Mustaffa, 2012, p. 389).

However, the subsidiaries are increasingly contributing to the company's value creation (Chung, 2014; Ciabuschi et al., 2015; Criscuolo, 2009; Eden, 2009; Frost & Zhou, 2005; Makela et al., 2009). Cuervo-Cazurra and Un (2004) for example have stated that strategic, or advantageous, international resources that are firm-specific can derive not only from a parent company advantage but also from a subsidiary advantage or a multinational advantage.

Regarding knowledge specifically, despite not considering it at first, in his later works Dunning (2001) acknowledged that some endogenous variables related to knowledge, such as technological or organizational innovations, management changes, new marketing techniques and others, were relevant in creating a company's competitive advantage and that companies were "increasingly growing internationally to create or gain access to resources or capabilities which complement their core competencies" (Dunning, 2001, p. 183). In other words, multinationals apparently did learn from international operations and consequently improved their competitive positioning.

Thus, the environment is changing to a scenario in which knowledge flow is abundant, speed and cost of communication are higher and lower respectively, managerial experience is increasingly becoming international, and alternative governance mechanisms are emerging (Autio, 2005), giving rise to new forms of ventures.

One such form of new venture, as previously mentioned, is the EM MNE: "international companies that originated from emerging markets and are engaged in outward FDI, where they exercise effective control and undertake value-adding activities in one or more foreign countries" (Luo & Tung, 2007, p. 482). EM MNEs expand rapidly and aggressively to overcome their latecomer disadvantages, which often lack of knowledge or expertise. In this sense, the subsidiary's knowledge is essential and strategic to overcoming the parent's deficiencies, which makes the KT process to parent company therefore essential and strategic too. A good example is the acquisition of Volvo Cars Corporation by Zheijiang Geely Holding in 2010 (Petersen & Ivarsson, 2015). The fairly small and unknown Chinese company wanted to use Volvo's technology and famous brand to benefit locally and abroad. The plan was to exploit not only Volvo's products, but also its technologies and R&D capabilities for development of Geely's own 2010 lineup (Petersen & Ivarsson, 2015).

Needless to say, for a parent company to benefit from a subsidiary's knowledge, headquarters must be aware of it (Ciabuschi et al., 2015; Kumar, 2013). If headquarters is aware of its own and its subsidiaries' knowledge, it can better formulate corporate strategy and allocate resources. If, on the other hand, it doesn't know exactly what kind of knowledge it possesses and what kind it lacks, "its ability to control and contribute to value creation process is unclear" (Ciabuschi et al., 2015, pp. 44-45).

Similarly, some authors argue that the RKT is a persuasion process, with the subsidiary having to convince the parent of the value of its knowledge (Pérez-Nordtvedt et al., 2015).

Transfer of knowledge, streaming from headquarters to the subsidiaries (hierarchical direction), is the traditional direction of flow, as framed in most theories (Dunning, 2001; Johanson & Vahlne, 1977, 2009; Rugman, 2006; Vahlne & Johanson, 2014; Vernon, 1966, 1993). For this reason, the opposite flow (from subsidiary to parent company) is called reverse knowledge transfer (RKT). Other authors also refer to the RKT concept as reverse technology transfer (Belderbos, Van Roy, & Duvivier, 2013; Criscuolo, 2009; Driffield, Love, & Menghinello, 2010), reverse knowledge flow (Maehler, Curado, Pedroso, & Pires, 2011), reverse transfer of practices (Chung, 2014) and akin.

We define a RKT as a process through which knowledge is transferred from a source that is affiliated (subsidiary) to the recipient (headquarters), and the recipient receives, assimilates and applies their knowledge in order to obtain a competitive advantage.

As demonstrated in the brief theoretical background developed, most studies have focused on the traditional flow, leaving a literature gap (Michailova & Mustaffa, 2012), to be filled regarding the study of RKT. Moreover, the existing studies in RKT seem to be overly focused on why these transfers occur and what are the enabling mechanisms (the how).

RKT is a growing phenomenon that deserves special attention and that poses many questions; thus this paper aims to provide a review of what has been written on the subject in top journals over the past 15 years in a quest for answers and to discern research gaps. A theoretical framework is also proposed for analyzing such a complex process. This endeavor can be achieved through a thorough systematic analysis, as described in detail in the Method section below.

The method chosen for conducting the literature analysis was a systematic review - a reliable, replicable and scientific method for conducting literature reviews (Petticrew & Roberts, 2006). In the words of Petticrew and Roberts, a systematic review "explicitly aims to limit systematic error (bias), mainly by attempting to identify, appraise and synthesize all relevant studies (of whatever design) in order to answer a particular question (or set of questions)" (Petticrew & Roberts, 2006, p. 9) and hence is suitable for this thematic analysis.

Only peer reviewed articles from top academic journals were included in the review in order to ensure the quality of the work. Please find below the list of journals included in this systemic survey, which yielded an average impact factor of 1.84 and an average H index of 62.1. Any other top academic journal that is not listed below is because no relevant article was found as per out filters described further down this section.

Furthermore, a fifteen-year period, from 2001 to 2016⁽²⁾, was defined in order to have a long enough period to ensure a solid review but not so long that outdated material would be included.

Finally, only articles written in English were reviewed, in both EBSCO and PROQUEST databases. Keywords searched were reverse knowledge transfer OR reverse knowledge flow OR reverse technology transfer AND Internationalization AND subsidiaries. The same searches were performed with British spelling (i.e., internationalisation instead of internationalization) in order not to miss any potentially important article.

The searches yielded 154 articles, which were then manually filtered for repeated articles and scope. In this manual verification and selection for scope, a two-step process was followed. First, only title and abstract were read and a few articles were already discarded for not being related to Reverse Knowledge Transfer. In a second step, all articles were fully read in order to see how they approached the RKT phenomenon. Some were again not related to the RKT as the title and abstract might have led us to believe and were discarded. The remainders were all fully analyzed in the main section of the article.

As can be seen from Table 2 above, only 52 articles remained after quality, date, and duplication search/scope filters were applied. Some additional papers that did not directly come from our systematic search, but were referenced in those in our search may also be cited in the results section below due to their relevance to the theme. Discussion of the papers follows in the next section.

Table 1
Result of Systematic Literature Review Search by Journal

Table 2
Result of Systematic Literature Review Search

As already mentioned, with very few exceptions (Nair, Demirbag, & Mellahi, 2015; Pérez-Nordtvedt et al., 2008; Tseng, 2015), most studies tend not to analyze the RKT as a process, but to focus on a specific aspect of the process at a time, such as how or why the transfer occurs. However, the RKT phenomenon is a full process, and thus should be analyzed with all of its parts and contexts.

Gupta and Govindarajan (2000), Pérez-Nordtvedt, Kedia, Datta and Rashee (2008) and Tseng (2015), to cite a few, consider as factors influencing the degree of knowledge (a) flows from the sender (source unit), (b) to the receiver (target unit), (c) the transmission per se (formal and informal mechanisms) and (d) the knowledge characteristics (mainly complexity and tacitness). Figure 2 below summarizes a typical knowledge flow process. In an RKT process, the target unit will be the headquarters and the source unit would be the subsidiary holding the knowledge to be transferred.

Figure 2
MNC's Internal Knowledge Flow as a Transferring Process
Source: Prepared by the authors based on Gupta, A. K., & Govindarajan, V. (2000). Knowledge flows within multinational corporations. Strategic Management Journal, 21(4), 473-496. 10.1002/(SICI)1097-0266(200004)21:4<473::AID-SMJ84>3.0.CO;2-I; Pérez-Nordtvedt, L., Kedia, B. L., Datta, D. K., & Rashee, A. A. (2008). Effectiveness and efficiency of cross-border knowledge transfer: an empirical examination. Journal of Management Studies, 45(4), 714-744. 10.1111/j.1467-6486.2008.00767.x; Tseng, C. (2015). Determinants of MNC's knowledge inflows to subsidiaries: a perspective on internalization advantages. Management International Review, 55(1), 119-150. 10.1007/s11575-014-0230-7

But "the majority of studies on knowledge transfer focus on either the relationship between the source and the recipient, the recipient itself, the source itself or the type of knowledge being transferred. With few exceptions ..., research has failed to simultaneously examine all of these antecedents" (Pérez-Nordtvedt et al., 2008, p. 715), indicating a literature gap.

Our analysis is consistent with the position of Pérez-Nordtvedt et al. (2008), as can be seen in Figure 3⁽³⁾, indicating that few studies analyze RKT as a full process. However, we go a little further, we can incorporate two additional stages to the process: the pre-transfer, which pertains to the circumstances under which the knowledge was created on the subsidiary; and the post-transfer consequences to the parent company and the MNE as a whole.

Figure 3
MNC's Internal Knowledge Flow Complete Transferring Process
Source: Prepared by the authors.

The RKT itself: source, transmission factors, knowledge characteristics and recipient analysis

As previously mentioned, our study confirms that most studies on RKT can be separated into why RKT occur and how these transfer flows occur, leaving unanswered other types of questions (Becerra et al., 2008; Najafi-Tavani et al., 2012b). For a complete list on which paper focuses on which part of the process, please refer to Appendix.

"Why" questions: source and recipient analysis

When analyzing the why of an RKT, studies tend to focus on locational aspects of both sender (subsidiary) and receiver (parent) of the knowledge, supported more by the economic theories of internationalization. However, as discussed below, the main topics are not necessarily economic related which raises the question as to whether a different perspective might have been more suitable.

Main topics of research found in relation to the sender, or, in the case of RKT, the subsidiary (i.e., why knowledge from this specific subsidiary is valuable) included (a) the growing trend and benefits for the internationalization of firm R&D (Criscuolo, 2009; Criscuolo & Narula, 2007; D'Agostino & Santangelo, 2012; Di Minin & Zhang, 2010; Massini & Miozzo, 2012; Moncada-Paternò-Castello, Vivarelli, & Voigt, 2011; Sanna-Randaccio & Veugelers, 2007) and (b) discussions relating knowledge creation to the level of network embeddedness of the subsidiary (Hakanson & Nobel, 2001; Jiménez-Jiménez, Martínez-Costa, & Sanz-Valle, 2014; Najafi-Tavani et al., 2012b; Williams & Ecker, 2011).

Additionally, and related to those topics, spillover effects, that is, benefits accrued from proximity to companies of superior knowledge, were mentioned by other authors as a benefit and reason for having an international presence (Griffith, Harrison, & Van Reenen, 2006; Liu, Lu, & Choi, 2014). Taken as a whole, the above mentioned topics and authors refer, in different ways and through different concepts, to a parent company engaging in an RKT process to benefit from knowledge to which its subsidiary had access by virtue of being in a certain location, with certain players or in a certain context of superior or different knowledge stock, knowledge that the parent company would not otherwise have access to.

Finally, other interesting findings are that RKT (a) is more likely to occur when the subsidiary is located in an economically developed country (Ambos et al., 2006) and (b) tends to vary with size and age of the subsidiary (Bezerra, Borini, & Maclennan, 2015).

The second why question in the RKT depicted in Figure 4 above refers to why the parent company needs this kind of knowledge, or why is this knowledge valuable to this specific parent company. The salient topics found on this receiver side of the RKT process were related either to a catch-up strategy and latecomer perspective of emerging market companies or to a knowledge-based view as performance enhancers.

Figure 4
MNC's Internal Knowledge Flow Complete Transferring Process
Source: Prepared by the authors

Many scholars studying EM enterprises see EM MNEs as increasingly engaging in springboarding and knowledge-seeking FDI to compensate for their deficiencies (Eden, 2009; J. Li et al., 2016; Luo & Tung, 2007), which may include liability of foreignness or third-world image (Nair et al., 2015), and, most importantly, competitive weaknesses compared to global leaders (Deng, 2013). Thus, RKT is an important ally in this catch-up strategy (Awate, Larsen, & Mudambi, 2015; J. Li et al., 2016), for it enables knowledge to flow from the subsidiary to the headquarters and to the rest of the corporation. This, in turn, enables the MNCs to obtain competitive advantages by combining local knowledge with their own technological and management capabilities (Chung, 2014; Tseng, 2015) and accrue the consequent benefits of it. Indeed, Sofka (2008, p. 786) claims the globalization of knowledge in the MNCs "is in fact a combination of investments into absorptive capacities, international experience and domestic scarcities".

Several authors appear to conflate the two why questions, posing the question in terms of why the knowledge created at one specific subsidiary is valuable to a specific parent company. Indeed, Belderbos, Lokshin and Sadowski (2015) do so, stating that "the roles of domestic and foreign R&D depend on the relative position of the home country with respect to the global technology frontier and the related relative opportunities for knowledge sourcing abroad" (p. 491). In other words, the value of RKT will be apparent when the parent company lags behind vis-à-vis certain knowledge and has the opportunity to acquire it from a subsidiary located in a host country that is more competitive than the home country (Nair et al., 2015).

'How' questions: transmission enhancers and restrainers

When analyzing how RKT progresses, most studies focus on the direct and indirect mechanisms that may invigorate or sabotage a successful transfer, therefore giving a more human resources' related aspect to the research.

In this regard, important topics discerned were the use of binding mechanisms or social interactions among employees to facilitate the transmission process. Studies focused on (a) general social interaction to promote overall communication and integration (Ambos et al., 2006; Chung, 2014; Criscuolo, 2009; Frost & Zhou, 2005; Noorderhaven & Harzing, 2009; Pérez-Nordtvedt et al., 2008; Verbeke, 2010; Zaragoza-Sáez & Claver-Cortés, 2011); or, more specifically (b), the use of expatriates (Bruning, Bebenroth, & Pascha, 2011; Ghauri & Park, 2012; Huang, Chiu, & Lu, 2013; Makela et al., 2009; Najafi-Tavani et al., 2012b; Vance, Andersen, Vaiman, & Gale, 2014; Zaragoza-Sáez & Claver-Cortés, 2011); (c) rotation jobs (Criscuolo & Narula, 2007; Hakanson & Nobel, 2001; Minbaeva, Pedersen, Björkman, & Fey, 2014); or (d) returnees to the home country (Liu et al., 2014).

Additionally, (a) having similar cultural aspects and a shared vision or logic (Betz, Oberweis, & Stephan, 2014; Colakoglu, 2012; Criscuolo & Narula, 2007; Ghauri & Park, 2012; Nair et al., 2015; Najafi-Tavani et al., 2012a; Öberg, 2013; Verbeke, 2010); (b) creating specific organizational structures such as project or joint R&D teams (Criscuolo & Narula, 2007; Lichtenthaler, 2010; Minbaeva et al., 2014); (c) having a more integrated business in general, with strategy and practices aligned (Hakanson & Nobel, 2001; J. H. Li et al., 2014; Najafi-Tavani et al., 2012b); and, finally, (d) having formal rules and the necessary information and communication tools well established (Chung, 2014; Criscuolo, 2009; Zaragoza-Sáez & Claver-Cortés, 2011) seem to improve the RKT process in the studies analyzed.

In addition to the above mechanisms, several papers highlighted the importance of the parent company valuing and being willing to receive the transfer of knowledge, a concept also referred to as having willingness to learn or learning intent (Belderbos et al., 2013; Pérez-Nordtvedt et al., 2008). Relatedly, some authors adopt the term absorptive capacity (Ambos et al., 2006; Criscuolo & Narula, 2007; J. Li et al., 2016; Sofka, 2008; Vance et al., 2014), which Frost and Zhou (2005) define as the ability to recognize the value of new knowledge, assimilate it, and apply it. In contrast, Minbaeva, Pedersen, Björkman and Fey (2014) define absorptive capacity as having the motivation, ability and opportunity to absorb knowledge.

On the other end of the spectrum, willingness to transfer knowledge has also been tackled by authors such as Huang, Chiu and Lu (2013), Najafi-Tavani et al. (2012a) and Criscuolo and Narula (2007). Indeed, it turns out that subsidiaries or executives may be unwilling to transfer their knowledge due to fear of losing autonomy, information monopoly (Criscuolo & Narula, 2007), and their relative power inside the organization (Alharbi & Singh, 2013).

Finally, a culture of sharing (Huang et al., 2013; Minbaeva et al., 2014) and trust building (Chung, 2014; Huang et al., 2013; J. H. Li et al., 2014; Najafi-Tavani et al., 2012b) is believed to be beneficial not only to one specific process of RKT, but to the process of knowledge creation of the MNE as a whole. Kumar (2013) also observes that managerial attention during the process is crucial to success.

'What' questions: what kind of knowledge is being transferred?

In line with the knowledge-based view (KBV) proponents, Pérez-Nordtvedt et al. (2008) claim knowledge value, rarity and non-substitutability are determinants of attractiveness for RKT.

Yet, despite attractiveness, another dimension that should be analyzed in an RKT process is the knowledge transferability; and it is on this last dimension that most authors analyzed in this paper have focused their efforts. According to Kotabe, Dunlap-Hinkler, Parente and Mishra (2007), knowledge transferability should be analyzed according to its tacitness, specificity and complexity.

Nair, Demirbag and Mellahi (2015) focus on the importance of analyzing knowledge complexity, claiming that it deals with "the comprehension of the knowledge and is also closely associated with the width that the knowledge spans (Grant, 1996). Hence, knowledge could prove to be more complex especially when it spans across multiple domains of expertise" (Nair et al., 2015, p. 284). The authors also add that the complexity of knowledge contributes to increasing a characteristic that several other authors (Chung, 2014; Liu et al., 2014; Pérez-Nordtvedt et al., 2015) attribute to knowledge: its stickiness, or its difficulty to be transferred.

But most authors analyze knowledge according to its tacitness (Choi, Cheng, Hilton, & Russell, 2005; Johnson, 2007), some focusing on differentiating the two poles: explicit and tacit knowledge (Becerra et al., 2008; Buckley et al., 2003; Jaw, Wang, & Chen, 2006). Tacit knowledge is defined as knowledge that cannot be codified or expressed explicitly and is rooted in everyday practices, routines, culture and values (Nonaka, 1994; Noorderhaven & Harzing, 2009). Explicit knowledge, on the other hand, is codifiable and expressed in data, manuals, and the like. However, even explicit sources of knowledge have a tacit component, as tacit knowledge assists in explaining explicit knowledge (Buckley et al., 2003; Jaw et al., 2006).

Tacit knowledge is considered much more difficult to transfer, since it is more difficult to teach (Hakanson & Nobel, 2001) and requires a high degree of trustworthiness (Becerra et al., 2008; Noorderhaven & Harzing, 2009). On the other hand, tacit knowledge is more difficult to imitate, thus rendering it more attractive (Pérez-Nordtvedt et al., 2008) and a source of competitive advantage to the firm (Becerra et al., 2008).

It is worth noting that even when the type of knowledge being transferred was not the focus of their research, several authors emphasized or mentioned the general tacit nature of knowledge and the consequent challenges of transferring it (Hakanson & Nobel, 2001; Huang et al., 2013; Najafi-Tavani et al., 2012a; Sofka, 2008; Zaragoza-Sáez & Claver-Cortés, 2011).

Historically, most researchers working in traditional theories of internationalization (Dunning, 2001; Johanson & Vahlne, 1977, 2009; Rugman, 2006, Vahlne & Johanson, 2014; Vernon, 1966, 1993) have focused on the knowledge flow from parent companies to subsidiaries, not giving as much attention to knowledge created at the subsidiary level, perceived as somewhat inferior to headquarters' (Michailova & Mustaffa, 2012).

But, as demonstrated in this thematic analysis, there is a growing trend to knowledge co-creation, and the responsibility of knowledge creation has for some time been shifting from headquarters to the corporation as a whole and to its subsidiaries (Chung, 2014; Eden, 2009; Frost & Zhou, 2005; Gupta & Govindarajan, 2000; Makela et al., 2009), thus giving rise to the phenomenon of reverse knowledge transfer. This phenomenon is particularly important to and present in EM-MNEs as latecomers with a catch-up strategy (Ambos et al., 2006; Awate et al., 2015; Eden, 2009; Nair et al., 2015). Even so, many studies still originate from or focus on MNEs from developed countries (Buckley et al., 2003; Nair et al., 2015), leaving a gap in the literature. Scholars from China and India have started to address this gap, but other emerging regions of the globe, such as Brazil and other countries in Latin America, have yet to do so (Nair et al., 2015, 2015; Oliveira, 2007, 2012).

It bears repeating that although the springboard perspective (Luo & Tung, 2007) appears to be a suitable theoretical base for studies in EM MNE, it has thus far seen little use in studies of RKT, even in those specifically focused on emerging markets.

Our main conclusion from this thematic analysis is that reverse knowledge transfer is a complex process, and as such should be analyzed in terms of all of its components and its broader context. Figure 4 outlines the theoretical framework.

It is reasonable to assume that RKT is not actually a linear process, but one that is socially complex as well as reciprocal (Choi et al., 2005), whereby the subsidiary and headquarters are constantly interacting and mutually exchanging knowledge and information, both tacit and explicit.

However, most authors studying the phenomenon have focused on specific aspects of the process, overlooking the wider view (Pérez-Nordtvedt et al., 2008). Many have focused on the why of RKT or on the how of better knowledge transfer to the detriment of analyzing the process as a whole, including the antecedents of knowledge creation and, above all, the outcome of RKT, which is the most important part and yet one of the less researched. After all, although KT may be yet another process, one must emphasize the positive effect that the knowledge has on the recipient (J. H. Li et al., 2014), otherwise the transfer is pointless. J. H. Li et al. summed it up succinctly: "Only when knowledge is understood, assimilated and used for creating value, can highly-efficient knowledge transfer occur" (J. H. Li et al., 2014, p. 280).

Finally, our study corroborates with Michailova and Mustaffa's (2012), in showing that most studies are based on quantitative methods (mainly surveys); thus, another avenue for future research could be using case studies and more qualitative methods to understand this complex and contemporary phenomenon (Yin, 1994).

We understand that, despite our efforts towards a systematic analysis, our filters (dates, journals, languages, sources and databases) prevented us from doing a complete and exhaustive literature review on the subject. However, by providing complete disclosure on our metrics, we have sought to produce work that is as reliable and replicable as possible.

Also, as stated in the conclusion section above, we acknowledge that our theoretical framework is a simplification of reality, with the RKT process being broken down into specific and divisible parts for illustration and analytic purposes only.

This paper aims to make both practical and academic contributions. For practitioners, we sought to render the reverse knowledge transfer process clearer by breaking it down into component parts and aspects, thereby providing a comprehensive understanding of its complexity. Additionally, we hope that the work can contribute to the preparation of executives within MNE vis-à-vis knowledge management and how it can affect global strategies. To aid current and future academic work, we have inventoried and synthesized all relevant papers published in the target 15-year period (2001-2016⁽⁴⁾), which we believe is a reliable reference for further studies to address gaps and possible avenues of research.