ppt

Journal of World Business 45 (2010) 19–28

Explaining variation in reverse diffusion of HR practices: Evidence from the German and British subsidiaries of American multinationals

Tony Edwards a,*, Anne Tempel b,1

a The Management Department, Franklin Wilkins Building, 150 Stamford Street, London SE1 9NN, UK b Institute for Applied Work Science, Ruhr-University of Bochum, D-44780 Bochum, Germany

A R T I C L E I N F O

Keywords:

Multinational companies

HR practices

Reverse diffusion

A B S T R A C T

This paper examines the phenomenon of ‘reverse diffusion’ of HR practices in multinational companies,

defined as the diffusion of practices from the foreign operations of MNCs to the home country. Drawing

on research incorporating fieldwork in the British and German subsidiaries of US-based multinationals,

we identify sources of variation in the incidence of reverse diffusion both between firms and between

subsidiaries of the same firm.

Crown Copyright � 2009 Published by Elsevier Inc. All rights reserved.

Contents lists available at ScienceDirect

Journal of World Business

j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / j w b

There is a growing literature concerned with the ‘reverse diffusion’ (RD) of HR practices in multinational companies (MNCs). This is defined as the diffusion of practices from the foreign operations to the home country, contrasting with ‘forward diffusion’ in which the direction is from the home country to the foreign operations and ‘horizontal diffusion’ in which practices travel from one foreign operation to another. RD is potentially very significant. Being the source of new innovations that are subsequently transferred to the heart of the multinational, reverse diffusion is one mechanism through which foreign operating units may enhance their status and claim on resources (Ferner & Varul, 2000). For example, Kristensen and Zeitlin’s (2005) detailed account of a British MNC showed how the Danish subsidiary enjoyed an elevated role through acting as the source of a number of new initiatives that were diffused across the firm. For the multinational as a whole engaging in RD has the potential to enhance its competitive position (Holm & Deo Sharma, 2006). The benefits to the firm are strongly shaped by its ability to process knowledge effectively (Ambos, Ambos, & Schlegelmilch, 2006). Thus the phenomenon of RD matters because it affects relations within, and the performance of, MNCs.

The issue has relevance for our understanding of MNCs more generally. Early theories of MNCs assumed that foreign firms were at an inherent disadvantage when competing with indigenous ones that understood the local context better. In order to compensate for this, they had to possess a ‘firm-specific advantage’

* Corresponding author. Tel.: +44 20 7848 4092; fax: +44 20 7848 4092.

E-mail addresses: tony.edwards@kcl.ac.uk (T. Edwards),

anne.tempel@rub.de (A. Tempel). 1 Tel.: +49 0234 32 27 894.

1090-9516/$ – see front matter . Crown Copyright � 2009 Published by Elsevier Inc. A doi:10.1016/j.jwb.2009.04.001

that was generated at home and spread to other countries (Hymer, 1976). More recent work has recognized that as firms inter- nationalize they are able to tap into a more diverse body of expertise and leverage this knowledge across their sites. Dunning (1993), for example, argues that MNCs can strengthen the ‘resource base of the firm as a whole’ by drawing on their experience in different countries. Similarly, Doz, Santos, and Williamson (2001: 4) argue that MNCs’ competitive position is increasingly shaped by their ability to ‘improve their capacity to mobilize knowledge that languishes underexploited within their far-flung network of subsidiaries’ (see also Bartlett & Ghoshal, 1998). Our argument in this paper is that the capacity of MNCs to mobilize knowledge of HR practices in such a way that leads to RD is very variable. There are a number of potential barriers to firms engaging in RD that stem not only from their ability to establish appropriate firm-level characteristics and conduits through which it can occur but also from the way that the center manages the various national subsidiaries. In sum, as they internationalize, MNCs increase their scope to augment the expertise held within the firm, but realizing this is far from straightforward.

One distinction in the impact of RD is between ‘evolutionary’ and ‘transformative’ diffusion. The evolutionary form brings about only incremental changes to a firm’s pre-existing operations as it results in an optimal mix of practices within an existing ‘modus operandi’. In this case, the multinational retains a set of assumptions concerning the way it operates but fine-tunes this through learning from its foreign plants. One example is the way that Japanese MNCs identified improvements in work organization in their foreign subsidiaries that were incorporated into a corporate model developed in Japan (Edwards & Ferner, 2004). Transformative RD, on the other hand, moves the multinational to a new modus operandi. This occurs when MNCs draw on practices in the foreign subsidiaries that are quite novel to the firm’s

ll rights reserved.

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–2820

domestic sites, as has occurred with the take-up of ‘variable’ forms of remuneration in many German MNCs (Edwards & Ferner, 2004).

We address two aspects of RD. First, we consider why some MNCs are more likely than others to engage in it. On this issue we examine the structures of firms and the mechanisms through which they can transfer knowledge as sources of variation. Second, we explore the way that subsidiaries across countries are differentially placed to be the source of RD. This entails an investigation of the national context in which the subsidiaries are embedded, the resources that subsidiaries can draw on and the role that they play within the multinational.

Addressing these twin questions requires a particular research design. Most empirical research into RD using case studies draws on evidence from multiple firms in a single host country, enabling the analysis to focus on variation between MNCs. A few studies use data from multiple subsidiaries of a single firm, allowing the researchers to assess variation between subsidiaries. We are able to achieve both goals through a design that incorporates multiple subsidiaries in multiple firms, involving analysis of the German and British sites of U.S. MNCs. The resulting data are used in an exploratory way to enhance theory on this issue and, accordingly, we establish a series of specific questions to be explored in future research.

The paper proceeds as follows. The various strands of the literature on RD are reviewed in the next section, leading to the establishment of a number of implications for our study. The third section details the way in which data were gathered. The fourth section presents and analyses the findings and develops questions for future research. The conclusions of our analysis are drawn in the final section.

1. Key themes in the literature

1.1. Knowledge transmission perspective

A first strand, the ‘knowledge transmission’ perspective, sees variation between firms as the result of differences in mechanisms through which the transfer of knowledge about HR practices can occur. A common distinction is between ‘codifiable’ and ‘tacit’ knowledge (Adenfelt & Lagerstrom, 2006; Kogut & Zander, 2003; Minbaeva, 2007). There is some evidence that codifiable knowl- edge concerning HR practices is transferred through such mechanisms as manuals, databases and management audits (e.g., Ferner & Varul, 1999; Gupta & Govindarajan, 2000). Such ‘procedural’ mechanisms are likely to be ineffective for tacit knowledge, for which ‘people-based’ mechanisms are better placed to generate the ‘social interaction’ between actors that is important in facilitating ‘inter-unit resource exchange’ (Tsai & Ghoshal, 1998). The literature has identified such people-based mechanisms as multidisciplinary teams, cross-national groups and the repatriation of those who have been on international assignments as effective conduits for RD of tacit knowledge (Bossard & Peterson, 2005; Gourlay, 2006; Harzing, 2001; Martin & Salomon, 2003; Sanchez Vidal, Sanz Valle, & Barba Aragon, 2007). In sum, this stream of literature suggests that the prevalence, character and combination of the conduits for transferring knowledge constitute an important source of variation between MNCs in the incidence of RD. Indeed, it is pertinent to explore whether appropriate mechanisms of this sort are a precondition for RD to occur.

1.2. Corporate structure

A further, and much more modest, strand of the literature on RD identifies corporate structures as a source of variation between MNCs. One example of such a ‘structural’ approach concerning the

diffusion of technological innovations is Johnson and Medcof’s argument. They argue that the ‘network’ structure in which there are ‘active and flexible links between the HQ and the periphery’ (2007: 482) is the one most conducive to diffusion being initiated by the subsidiaries. An example in the HR field is Edwards’ (2000) analysis of UK MNCs where it is argued that the presence of certain features in MNCs provides fertile ground in which RD can take root. MNCs that have highly standardized systems of production or service provision, for instance, are those in which the technological context and occupational mix of staff at site level are broadly similar internationally, increasing the scope for MNCs to diffuse practices across sites. A second factor concerns the degree of ‘global spread’, with MNCs which are not concentrated in their original country being those in which the foreign operating units have greater weight within the firm and, therefore, are those in which RD is most likely to occur (Edwards, 2000). Thus we might expect that the presence or absence of these structures serves as one source of variation between MNCs in the incidence of RD. Overall, though, this line of analysis is under-developed in previous research and in particular the relationship between such structures and the knowledge transmission mechanisms discussed above are poorly understood.

1.3. National business systems

Consideration of these firm-level structures and mechanisms raises the issue of how subsidiaries of the same MNC are differentially placed to be the source of RD. Thus a third, and substantial, strand of the literature addresses this through the lens of the national business system in which subsidiaries are embedded. Much research on MNCs has studied how they struggle to overcome enduring national differences—the ‘psychic distance’ (Johanson & Vahlne, 1977) or ‘institutional distance’ (Kostova, 1999) that some refer to. Research into technological innovations in MNCs has shown that host country cultures and institutions are a significant influence on the ability of subsidiaries to initiate RD (Almeida & Phene, 2004; Ambos & Schlegelmilch, 2008; Andersson, Forsgren, & Holm, 2002; Cantwell, 1992; Frost, 2001; Frost, Birkinshaw, & Ensign, 2002). In the literature on RD of HR practices, national institutions also feature but the evidence is unclear as to which sort is most conducive to bringing about RD. For example, it can be argued that subsidiaries in countries where institutions promote clusters of firms which have stable relations with one another and collaborate to produce distinctive competencies are well placed to instigate RD (e.g., Kristensen & Zeitlin, 2005). In contrast, other evidence indicates that subsidiaries instigate RD more easily when they are located in countries in which market- based forms of economic co-ordination predominate (e.g., Ferner & Varul, 2000; Hayden & Edwards, 2001; Mtar, 2001; Tempel, 2001). It is evident that the influence of contrasting host country institutions in shaping the position of the subsidiary to engage in RD requires further research.

1.4. Dominance effects

One danger of focusing on national institutions is apparent in a further strand of the literature on MNCs which has implications for RD. Some observers argue that what matters is not so much the precise nature of a national system but rather how it is perceived by corporate managers, with these judgments being shaped by ‘dominance effects’ (Smith & Meiksins, 1995). Dominance effects arise from the evolving hierarchy of nations according to their economic performance, with those at the top of the hierarchy serving as models that firms across the world look to emulate. Popular systems and practices, such as ‘lean production’, achieve a ‘dominant’ status in part because of their

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–28 21

association with successful economies, and this conditions what senior managers look for in diffusing practices across borders. The evidence on RD of HR practices provides illustrations of how the success of the Japanese economy until the mid-1990s shaped what was the subject of RD (e.g., Edwards, 1998). As the performance of the Japanese economy has faltered, the influence of the American economy in particular and the Anglo-American economies in general, has become more apparent in studies of RD (e.g., Belanger, Berggren, Bjorkman, & Kohler, 1999; Ferner & Varul, 2000). The perceptions that result from dominance effects may be inaccurate and may conflate countries which are quite different into one category (as is often the case with the notion of ‘Anglo-Saxon’ business systems), but they matter nevertheless. Subsidiary level actors in countries that are perceived negatively by the HQ will struggle to be the source of RD unless they make innovations that are compatible with dominant models; in contrast, those in countries that are viewed favorably are much better placed. Thus perceptions of actors at the corporate HQ seem to be important in shaping the position of the various national operations in initiating RD.

1.5. Micro-political dimension

A fifth strand of the literature examines the micro-political dimension to RD. Whether RD occurs is not simply a matter of contact between sites; it also depends on the motivation of actors. For example, site level managers may be reluctant to share their distinctive expertise with other site managers with which they compete for new investment (Bjorkman & Barner- Rasmussen, 2004). Thus Edwards and Ferner (2004: 69) argue that ‘knowledge flows are paralleled by corresponding power and resource flows: the transfer of knowledge involved in RD is a political as much as a technical problem, in that it involves expending resources to overcome resistance’. A key role of the HQ, therefore, is to establish incentives for managers to share their expertise. This shapes their ‘motivational disposition’ (Gupta & Govindarajan, 2000) to engage in RD (see also Boojihawon, Dimitratos, & Young, 2007; Dobosz-Bourne, 2006). One way of doing so is stressing the benefits to donor sites in terms of safeguards for their future or for them to enjoy elevated status, such as a particular ‘mandate’. Another way is for individuals who have led the process of RD to be promoted or given the role of ‘internal consultant’. In this way, individual actors and sites can raise their profile by instigating the transfer of knowledge. It also falls to the HQ to set up penalties for those managers that seek to block RD, for example by benchmarking sites against one another and creating pressure on sites to emulate innovative ones (e.g., Mudambi & Navarro, 2004). The associated ‘positional strategies’ (Kristensen & Zeitlin, 2005) of managers at site level are crucial to whether RD occurs and these are shaped by the resources that they control. The profitability of the sites, managers’ possession of specialist expertise and the links that they have with local institutions can endow them with the power to bring about RD. A further challenge for researchers, then, is to shed light on the factors that shape the differential disposition and power of actors across sites to instigate RD.

1.6. Subsidiary role

A sixth strand of the literature on MNCs that has implications for RD concerns subsidiaries’ roles within companies that have strong inter-firm trading relationships. One line of research has stressed the way in which production processes that were previously nationally focused have become segmented across borders with the national operations performing a distinct function (e.g., Dicken, 2003). Some research has shown that many

MNCs are organized as ‘webs’ with sites having links with a number of others (Bartlett & Ghoshal, 1998; Birkinshaw, Hood, & Jonsson, 1998; Bjorkman, Fey, & Park, 2007; Boehe, 2007; Gupta & Govindarajan, 2000). Within such webs, sites may specialize in design and testing, others in assembly, others in providing customer support, and so on. The extent to which sites have a distinct specialty varies, with some taking on a unique role within the multinational and others taking on a role that may be similar to some of their counterparts but differ from others. This factor has not featured in the empirical literature on RD of HR practices but it has some implications for this issue; it is logical to expect that subsidiaries will be differentially placed in initiating RD according to their role within the multinational. Thus in explaining variation in the extent to which subsidiaries are the source of RD it is important to explore the role of such factors as the size of sites and their function within the firms, particularly whether they have been accorded a strategic role by the HQ or carved one out for themselves.

1.7. Linkages between factors

While these individual strands of the literature are useful in structuring expectations about variation in RD, the linkages between them are also important. The literature is only indicative on this. There is some discussion of the inter-relationships between the various factors that explain variation between subsidiaries; one way in which the different types of explanation are integrated is that the characteristics of the national system in which subsidiaries operate shapes the power of actors, demon- strating how the national systems and micro-political approaches are inter-dependent (e.g., Andersson, Forsgren, & Holm, 2007; Thory, 2008). Moreover, the power of subsidiaries is also conditioned by the role they have within the firm and, equally, the role they have within the firm is shaped by the resources that they develop locally (Birkinshaw et al., 1998), showing the linkages between the micro-political and subsidiary role explanations. There are also grounds for expecting to find linkages between the factors explaining variation between firms with those that explain variation between subsidiaries. For example, the nature of subsidiary roles is shaped by the structure of the company as a whole: if sites that are taking on strategic roles within the company are those most likely to initiate RD, then we might suppose that these are unlikely to be found in highly decentralized firms in which units are more or less self-sufficient, nor in highly centralized, hierarchical firms in which key strategic functions are kept in the home country; rather they are most likely to be found in networked organizational forms (e.g., Harzing & Noorderhaven, 2006). Another illustration is provided by Lazarova and Tarique who argue that there is an inter-relationship between ‘individual readiness to transfer knowledge and organizational receptivity to knowledge’ (2005: 361), demonstrating the linkages between the micro-political and the knowledge transmission strands of the literature. However, these inter-dependencies are only rarely explicitly developed and many potential linkages are not covered in the literature. We might anticipate, for example, that the participative, inclusive forms of networking in particular functions such as HR that constitute a mechanism through which RD can occur may be most common in network-based corporate structures, yet this is not developed.

This review has demonstrated sources of variation both between firms and between subsidiaries of the same firm. The literature is well developed in some of these areas (such as knowledge transmission channels and host country context) and less well developed in others (of which the ‘structural’ and ‘dominance effects’ explanations are examples), while previous research has only touched on the linkages between them. The

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–2822

paper aims to enhance our understanding of each strand and to further develop the inter-relationships between these factors.

2. Method

Most case study research on RD has been based either on data from subsidiary level only (e.g., Bjorkman & Barner- Rasmussen, 2004) or from the subsidiary level in one country and the HQ (e.g., Edwards, Almond, Clark, Colling, & Ferner, 2005; Ferner & Varul, 2000). Kristensen and Zeitlin (2005) and Belanger et al. (1999) stand out in incorporating fieldwork in several subsidiaries as well as the HQ into a single case study. The present study goes beyond even these in one important respect, namely that we analyze data from two subsidiaries and the HQ in a multiple case study design. We draw on two case study projects examining American MNCs. The UK project involved 14 firms between 2000 and 2003. The companion study was of 10 American MNCs in Germany and was conducted over a similar timeframe, using the same method of semi-structured interviews.

At an early stage in the research the teams collaborated in the production of an interview schedule. This established the key issues to be covered in each case study organization, which included the structure of the company, the reporting relationships between levels of the structure, the nature of the corporate culture and how this is communicated within the firm, and the nature of HR policy and practice in areas such as pay, performance management, training and industrial relations. This schedule served as a guide rather than a straightjacket for the interviews; in order to exploit the advantages of this method, a semi-structured approach was adopted with interviewers tailoring the content of the questions according to the individual’s position and also according to the issues that had arisen in previous interviews and those that arose during each interview. For the purposes of this paper this exploratory approach was crucial in addressing those sources of variation in RD for which there is not a well developed literature and in enhancing our understanding of the linkages between them. All of the interviews were recorded, fully transcribed and analyzed. Throughout the research the team members collaborated with each other, revising the interview schedules in the light of emerging findings. The choice of language in interviewing is a delicate issue, affecting how at ease is the respondent (Marschan-Piekkari & Reis, 2004). Accordingly, Ger- man respondents, whilst generally fluent in English, were interviewed in their native language with the resulting transcripts being translated.

The data source for this paper comprises three of these case studies, which were selected on the basis that, first, we were able to interview a number of key respondents in the corporate HQ in the U.S. and in the British and German operations and, second, that they are representative of the sectoral diversity of the wider projects—one is in services (Business Services), a second is in manufacturing (CPGco) and a third straddles manufacturing and services (ITco). In total, 55 interviews were carried out across these three firms. The cases enable us not only to compare subsidiaries in contrasting business systems but also to assess the extent to which their practices were incorporated into corporate policy, something that is necessary to assess the extent of RD. The findings from the wider projects are referred to at times, but the analysis in this paper goes beyond previous analysis in that it focuses solely on the issue of RD in the three companies for which we undertook research in all three countries, using these to enhance our theoretical understanding of RD.

A key methodological issue concerns the evidence that is necessary to justify an assessment that RD has occurred. One potential trap here would be to base this assessment on the claims made by a manager at subsidiary level. This would be dangerous

both because managers at this level may make claims that exaggerate their influence and also because they may genuinely be unaware that practices they shared with other sites were not in fact implemented. Thus the instances of RD in the fieldwork are only reported here where leaders at the HQ level could corroborate them. Thus triangulation, which is considered by Ghauri (2004: 115) to be ‘one of the defining features of a case study’, is central to our approach. Having established where RD had occurred, these multiple accounts were then used to assess both the mechanisms through which it took place and whether the impact was ‘evolutionary’ or ‘transformative’. This involved using the respon- dents’ views on the scale of change in corporate policy that RD caused and, crucially, whether the practice that was transferred slotted into an existing ‘modus operandi’ or shifted the multi- national to a new one.

In sum, the data mean we are well placed to address the paper’s goals. The fieldwork in two host countries and the home country is distinctive, the interview-based approach permitted extensive triangulation and the research in different countries was strongly comparable. While case study data of this sort are not well suited to the testing of some hypotheses, they lend themselves very well to the building of theory in general and the establishment of specific propositions for future research in particular. We use the data analysis to build a series of such propositions in the next section.

3. The case study findings

A theme of the data was of apparent openness to RD. Many respondents argued that their firm was open to learning from across its sites and that where managers in the subsidiaries could demonstrate the efficiency of a new practice then this could be adopted across the organization. This is in itself not convincing evidence of RD, of course, since that requires corroboration of practices originating in foreign subsidiaries and being diffused to the rest of the multinational. We start by charting such instances in our case studies, summarized in Table 1.

3.1. Evidence of reverse diffusion

The company in which the evidence of RD was most extensive was Business Services. The UK operations had pioneered a change to the performance management system consisting of a ‘general assessment tool’ for storing knowledge. This was an on-line database and replaced a series of forms; as such, it represented a relatively minor change to the system, constituting evolutionary RD. The UK operations were also responsible for a more significant innovation, relating to the handling of outsourcing projects in which Business Services carries out IT and accounting services for other companies. The way this was managed became the model for how outsourcing projects were handled across the firm. It was known as the ‘blueprint’ for managing outsourcing and led the firm to ‘treat and manage employees fundamentally differently’:

‘The (supermarket) project became essentially a case study for what to do and what not to do in managing the workforce or employees of an outsourcing arrangement and those dos and don’ts and methods and practices have been well documented and are now part of the foundation of how we look at managing employees in other outsourcing units’ (Global HR Director).

This resulted in staff that had been active in the supermarket project in Britain advising on how similar arrangements could be handled elsewhere. This is evidence of transformative RD since it

Table 1 The incidence of reverse diffusion in the case studies.

Firm Corporate characteristics Subsidiary characteristics Evidence of subsidiary initiatives Evidence of

reverse diffusion

Business

Services

Int’l standardization

high (MNC clients)

Large operations in both

countries, similar roles

of serving local clients

UK innovations in the management

of outsourced employees and

performance management taken

up globally

From UK Yes,

transformative

and evolutionary

Has operated as an

MNC for 40+ years

UK has grown more

rapidly and perceived to

be ‘five years ahead’

60% of ops outside US The UK has role of

‘virtual’ European HQ

in 3 divisions; Germany

has none

Leadership development program,

developed in Germany provided the

basis for corporate policy

From Germany Yes, evolutionary

Legacy of partnership

ITco Int’l standardization

high (global brands)

Large operations in both

countries with changing

roles

Appraisal system, involving setting

and assessing goals for each individual

on an annual basis that was developed

in the UK was incorporated into

company wide policy

From UK Yes, evolutionary

Has operated as an

MNC for 60+ years

Global cutbacks in 1990s,

but UK has expanded

significantly into the IT

outsourcing market

No initiatives from German subsidiary From Germany No

50% of ops outside US

High centralization Euro HQ in neither country

but UK staff occupy many

positions on it

CPGco Int’l standardization

high (global brands)

Large European operations

serving the regional market

UK site introduced ‘broadbanding’

involving a form of teamwork and

a fall in the number of pay

grades—pushed to the USA but not

taken up. Moves by UK operations

to share anti-harassment training

not taken up by HQ

From UK No

Has operated as an

MNC for 40+ years

Major prod’n and support

ops in UK but prod’n

relocated from Germany with

associated loss in status

No initiatives from German subsidiary From Germany No

50% of ops outside US Euro HQ in neither country

but UK has stronger contacts

to it

High centralization

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–28 23

constituted a significantly novel innovation that markedly changed pre-existing policy.

The instance of RD from Germany consisted of an innovation in leadership development. The context in Germany was of rapid growth and a large number of relatively new employees becoming partners. The leadership development program was designed to help those partners who were struggling with their new tasks of managing teams, dealing directly with clients and acquiring new business. After proving to be successful in Germany, it was developed into a global program which applies throughout the company. As one respondent described it:

‘In Germany they have piloted their own leadership develop- ment course which went down very well. . .(it) was successful. Those ideas were then developed and we now have a global leadership development program. . .but the germ of an idea started in Germany’ (Global HR Manager).

The German innovation went with the grain of the company’s approach to employee development, representing incremental changes to it. Thus this is evolutionary RD.

The research in Business Services threw up instances of RD from other European sites, including a form of 3608 feedback from France and a technique developed in Spain for handling repeatable work in discrete packages. In addition, the European region of the firm had been given the responsibility for developing a way of

assessing the software skills of employees which had been developed and rolled out across Europe with a view to being implemented globally. In sum, there were numerous instances of RD, including of the transformative variety.

There was also some evidence of RD in ITco. The one instance of the UK sites initiating RD involved a formula for identifying employee’s ‘Personal Business Commitments’ at the start of the appraisal process and benchmarks against which they were measured. This was incorporated into ITco’s international apprai- sal policy. Its impact was limited since it operated within a pre- established system of individual appraisal that was linked to a ‘forced distribution’. Thus it constituted a relatively minor adaptation to existing policy and, hence, it was evolutionary. Another initiative generated by the UK operations was in the area of job classifications where British managers had developed a new practice that had been ‘sent up the lines’ and had become policy in the European operations, though not across the firm so is not strictly RD. Our research into the German operations of ITco did not reveal any evidence of RD, nor of initiatives made by actors in Germany which had attracted interest from the HQ.

In the third case, CPGco, no clear evidence of RD could be found from either Britain or Germany. In the UK, this was not for lack of effort. British managers had pioneered practices in the area of developing supervisors and anti-harassment training but failed to attract interest from the HQ. A further innovation in the UK that made rather more headway, initially at least, related to

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–2824

‘broad-banding’. This involved the creation of teams of operators and a simplification of the payment structure within these teams. British managers pushed this practice to their U.S. counterparts, arguing that it resulted in greater flexibility and an enhanced team spirit but were unsuccessful. The German operations appeared not to have sought to instigate RD. Managers in Germany were seemingly more concerned with consolidating their position than influencing the nature of international policies. As one respondent put it: ‘We are really more like an administrative unit with little influence’ (HR Director, Germany).

As Table 1 shows, there is considerable variation between both firms and subsidiaries that needs explaining. Why is it that RD was most evident in Business Services, less so in ITco and not at all in CPGco? And how can we explain why the British subsidiaries were more frequently the source of RD than their German counterparts? We consider each question in turn.

3.2. Explaining variation between firms

3.2.1. Knowledge transmission perspective

The first source of firm-level variation is the conduits capable of transferring knowledge. Here, the evidence demonstrates that the most well developed mechanisms are in the firm with the most RD. Business Services had gone to considerable lengths to capture, codify and subsequently diffuse knowledge across borders. Such formalized structures were effective in diffusing codified knowl- edge on HR practices. One example was the ‘Knowledge Exchange’:

‘So all of the best practices, all of the learnings from key projects around the world – we have what’s called the ‘‘Knowledge Exchange’’ which is a place where all of that is stored and housed, content architecture, search engines, how to find the information, how to harvest it, how to categorize it in a way that’s easily accessible by our people’ (HR Manager, UK).

Another structure through which knowledge can be spread around the company is ‘Best Demonstrated Practice’, a formal record of practices that provided solutions to common problems. Perhaps most important was ‘People Enablement’, which com- prised IT tools capable of storing knowledge and making it available globally. This was described as ‘the creation, harvesting, organizing and purposeful distribution or sharing of everything we know how to do and everything we have ever done’ and as the ‘secret ingredient that makes everything else work’. This was a key mechanism through which the practices associated with the supermarket project were reverse diffused, through the knowledge being inputted onto databases and then becoming ‘embedded in the training the People Enablement function run, particularly in this case it formed part of the training that our business process outsourcing people do so that everybody knows about it’. This involved staff traveling across sites regularly.

The combination of the mining of databases and the face-to- face training points to the link between the procedural- and people-based conduits in bringing RD about. The People Enable- ment function ensured that in many new outsourcing projects ‘one of the experts from (the British supermarket project) put the deals together’. The balance between procedural- and people-based conduits seemed to be shifting to the former owing to a reduction in travel, limiting the extent to which people moved across borders, but the interaction between the two is clear.

ITco had established some channels capable of diffusing codifiable knowledge across sites. For example, ‘Best of Breed’ allowed subsidiaries to identify practices that could be spread across the firm, enabling them to initiate RD. However, these were not as high profile as their counterparts in Business Services. In the third case study, CPGco, the formal mechanisms designed to tap

into innovations across the multinational were less evident. While the corporate HQ had moved towards involving subsidiary actors in the development of international HR policies, these were largely through ad hoc working groups rather than the systematic, ongoing use of the ‘hardware’ of transfer that was such a feature of Business Services.

Overall, the analysis raises two propositions for future research. The case of Business Services suggested that the importance of procedural conduits of knowledge transfer is growing in relation to people-based conduits, while developments in ICT may have increased the ability of the former to handle even tacit knowledge, reducing the need for face-to-face contact. This leads to the first proposition:

1a. Procedural conduits of knowledge transfer are becoming more important for RD.

The data across the cases support the notion that the existence of appropriate channels for transferring knowledge are very important if RD is to occur, but is this a pre-condition? Arguably, RD could occur in the absence of such channels through direct contact between site actors, though such diffusion is likely to be limited in how widely spread it is across the firm. This leads to the second proposition:

1b. Where RD occurs through direct contact between sites as opposed to organizational conduits, its spread across the multi- national is likely to be more limited.

3.2.2. Corporate structures

The second source of variation between firms is to do with corporate structures. However, the variation between the case study firms can only partially be explained by the structures already identified in the literature. As Table 1 indicates, they are all internationally standardized in that the products or services have strong commonalities, ‘mature’ internationally and have a high degree of ‘global spread’. However, the evidence does point to a different structural factor not covered in the literature being crucial, namely the partnership structure that was part of the ‘administrative heritage’ of Business Services.

Where the HQ of MNCs exercises tight control, the scope for foreign sites to influence the development of global policies is restricted. Both ITco and CPGco had a legacy of strong centralized control. In the former, the influence of corporate management had markedly risen in the decade prior to our research. In the latter, centralized control had also increased:

‘The HR function used to be very independent and we had decision-making autonomy in HR issues. We used to report on employee numbers and that was about it. This has changed completely.. . . The directions now come from (the European HQ). That is, they don’t come from (the European HQ) but from (the corporate HQ) via (the European HQ)’ (HR Director, Germany).

In neither case was this so strong that it closed off all scope for subsidiary initiatives; indeed, the data indicates that the UK sites of both companies had looked to expand their role within the firms through supplying practices to other sites (albeit unsuccessfully). However, the centralist tradition appeared to have created a significant organizational group that may resist RD, a well-staffed corporate HR function. A corporate HR function may feel threatened by subsidiaries becoming more influential in policy- making and fear that the logic of this would be a reduction in HQ staff and status within the organization. There was evidence of this in CPGco where the privileged claim on resources of the corporate HR function was raised by one of our respondents.

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–28 25

A highly decentralized approach is also not conducive to RD since such MNCs may lack mechanisms that bring managers together across borders. Rather, it appears that a ‘network’ structure, in which authority is spread across the firm with the HQ playing a coordinating, as opposed to a controlling role, is more conducive to RD. This was the essence of the partnership legacy in Business Services, which persisted even after the company had become publicly listed. The ‘network’ structure afforded foreign operations a significant role in the policy formation process. As one respondent put it: ‘I actually think that our mentality to a certain extent is still a partnership . . .. We still like to consult and talk and get buy-in of the partners’ (UK HR manager). Another argued that ‘because our heritage is a partnership . . . other countries in the leadership and other groups round the world expect to be asked for their opinion, expect to have the opportunity to provide input and feedback’ (Senior HR manager, US). This was key in actors within the U.S. being favorably disposed towards modeling corporate policies on innovations made in Britain and Germany, differentiat- ing Business Services from the other firms.

This analysis raises two further questions that could be tested in future research:

2a. There is a curvilinear relationship between the degree of centralization of decision making and the extent of RD.

2b. Transmission channels that are effective in bringing about RD are most likely to be found in network-based structures.

3.3. Explaining variation between subsidiaries

In seeking to explain why the British subsidiaries were better placed to instigate RD compared with the German sites we structure our discussion around the subsidiary level factors previously discussed.

3.3.1. National business systems

The first of these concerns the role of the national systems in which subsidiaries are located. The evidence from the UK revealed instances of RD in a number of areas that arguably reflect key features of the British business system. One broad area where this was evident was appraisal, performance management systems and variable pay. RD was found in both Business Services and ITco in this area and this was consistent with a pattern of UK subsidiaries pushing innovations to other parent firms in the British project (Edwards, Collings, Quintanilla, & Tempel, 2006). The UK opera- tions have greater potential than their German counterparts to be the source of RD partly because the requirement to negotiate the introduction of, or changes to, variable pay with employee representatives is less widespread, and partly because some aspects of ‘variable’ pay mesh with a system of corporate governance in which managers pay heed to the primacy of ‘shareholder value’. A second area in which instances of RD from UK subsidiaries were detected was job classification, particularly in relation to teams. We have seen how there were moves in this direction in both ITco and CPGco. In addition, the development of teams of operators that function without occupational divides and elect their own ‘co-coordinator’ was an initiative developed in another firm in the wider project (Edwards et al., 2006). Again, the weakness of workplace employee representation in the UK means that British subsidiaries have greater potential than those based in Germany to make innovations in this area. A third instance of RD from the UK also bears the imprint of the British business system. In Business Services, as we have seen, the British operations pioneered a new way of handling employees transferred into the multinational through the outsourcing of IT, a phenomenon that has grown rapidly in the UK. As Grimshaw and Miozzo (2004: 13)

put it, ‘the UK was for much of the 1990s the leading country driving the European IT outsourcing market’. While this phenom- enon has increased in Germany in recent years, much of it is between firms with a strong ownership link, such as Siemens Business Services with Siemens; in contrast, in the UK outsourcing is more commonly between firms without ownership links. More generally, one of the key features of the British business system is its reliance on ‘arm’s-length’ trading relationships with co- ordination between firms governed by formal, written contracts rather than informal, collaborative networks. Given this, it is not surprising that it was the British subsidiary that developed expertise in this area.

The substance of RD originating in Germany also reflected national effects. In Business Services the German subsidiary’s innovation related to the training of managers. Training and development in German firms is heavily shaped by the institu- tional context. This context can limit the ‘diffusability’ of practices. However, the innovation made by the German operations was not a direct product of the formal institutions in this area and so was relatively easy to diffuse. One interpretation of this is that RD occurred in an area in which the German business system provides a source of strength, taking the form of a disposition to developing employees, but not one in which practices were dependent on specific institutional supports. The wider study of American MNCs in Germany contained a case of RD of temporal flexibility, involving a flexible working time system, which had been subject to extensive negotiations with the works council, being incorporated into corporate policy (see Edwards et al., 2006). The original innovation was a product of the generally collaborative manage- ment-works council relations in the German site.

Thus we have shown how the character of the host country affects the substance of RD. The evidence also indicates that the greater compatibilities between key institutions in the British and American business systems help to explain why the British subsidiaries were more active in being the source of RD. In the same way as many corporate policies could be adopted with greater ease in Britain than in Germany because they went with the grain of local institutions, so there was more scope for UK practices to be adopted in domestic operations. It was easier for managers in the UK to ‘speak the language of the Americans’, as the works council chair in ITco in Germany put it, helping them to push their innovations into company policy. This ‘speaking the same language’ relates not just to the obvious linguistic dimension but also to similarities in corporate governance systems and the relative weakness of forms of employee representation. This leads to a further proposition:

3a. Both the distinctiveness of host country institutions and their compatibility with those in the home country shape whether a subsidiary is well placed to be the source of RD.

3.3.2. Dominance effects

While the evidence on the compatibility of national institutions is indicative, it is more conclusive in relation to the way that corporate managers perceived each country. It will be recalled that ‘dominance effects’ shape how managers in American MNCs perceive various countries. Arguably, this may draw U.S. MNCs to the UK since, rightly or wrongly, it is widely perceived to be performing more strongly and offer a more ‘flexible’ environment. Accordingly, a common message emanating from our data was the negative perception on the part of American managers of the German system, which was clearly communicated.

This negative perception centered on the issues of inflexibility, high labor costs and long decision-making time as shaped by collective bargaining and codetermination. This was felt particu- larly strongly in CPGco, where the negative perceptions of

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–2826

American managers had been exacerbated by works council resistance to ‘mandatory’ corporate policies, preventing corporate models on teamwork and performance-related-pay from being implemented in the German sites. Tellingly, HQ managers’ views primarily reflected these constraining aspects of the institutions in Germany, and largely ignored the more positive aspects of the system. One argued that

‘many American managers are of the opinion that you can’t locate production operations in Germany. In their view, everything takes too long here, legal regulations are so much stricter’ (HR Director, Germany).

In ITco senior managers exhibited negative perceptions of the German sites. As the chair of the works council put it, ‘for years we suffered from the stigma of having the highest wages and shortest working hours’. Respondents outside Germany stressed the legal constraints in introducing new practices into the country. For example, the UK HR Director argued that ‘there are some things from a legal perspective that we just cannot do in Germany’. The difficulties in having to negotiate changes to payment systems and the speed with which changes in employment levels can be made were examples. Another Briton indicated that German managers had used these constraints to block international HR policies: ‘I’m sure it’s convenient from time to time to argue that the works council would not agree’. In contrast, the British operations appeared to be held in a more positive light, illustrated by the fact that European HR posts were disproportionately staffed by Britons. Even in Business Services, where the German subsidiary had been the source of one case of RD, it seemed that they were perceived in a less favorable light than the British. One respondent complained that ‘it takes a lot before someone from Continental Europe or Asia is promoted to a leadership position’ (Finance manager, Germany).

Therefore, while the actual impact of institutions shapes the substance of RD from different subsidiaries, the way they are perceived by American managers is a key part of the explanation of variation between subsidiaries in the extent of RD. In essence, corporate management had a negative view of the German system, putting actors in these subsidiaries on the defensive in matters relating to international HR policies and thereby heavily con- straining their role as potential sources of RD. This presents a further issue for future research, namely are these perceptions of corporate managers specific to U.S. MNCs? It may be that what we are picking up here is a distinctively American set of attitudes and that in Swiss MNCs, say, German subsidiaries may be more favorably placed to initiate RD than their British counterparts. This leads to a further proposition:

4a. The perceptions of corporate managers that shape RD are not governed solely by global norms but rather are also shaped by a country of origin influence.

3.3.3. The micro-political dimension

Managers in the sites used a range of resources to exert influence on the wider firm and sought to engage in RD or block it in order to protect and advance their interests. The evidence from the UK demonstrates the key role of the strategizing of local managers. In ITco the downgrading of the role of national subsidiaries in favor of stronger business divisions and regions and British managers were well placed to exert influence outside the UK because they held a disproportionately high number of positions on the European axis of the company structure. The success of British managers in pushing innovations in the appraisal process should be seen in this light. In addition, the diffusion of the outsourcing innovations in Business Services from the UK reflected

a belief amongst British managers that the benefits of sharing this with other sites, especially the enhanced status within the company that it entailed, would outweigh any potential risks. A crucial factor in this respect was the very low likelihood of any disinvestment from the UK since they were highly profitable and had many key contacts with multinational clients.

The scope for managers in the German operations of Business Services to push their innovations to the rest of the firm was also significant, partly because German management had avoided collective bargaining and codetermination. This meant that workflow planning and working time, two key factors in the competitiveness of consulting firms, are shaped unilaterally by management without the restrictions of the Works Constitution Act and can be managed in line with global policies. This influenced its success in transferring practices in the area of leadership development. Of course, while the initiatives of subsidiary management were a key part of the process of RD, the agency they possess was exercised within constraints. Thus the strategiz- ing of local actors in ITco and CPGco was primarily concerned with redefining the perceptions that those at the HQ had of their operations. Subsidiary management in both cases had been seen by the HQ as orchestrating resistance to mandatory policies and were anxious to shed this image by conforming to corporate HR polices even where this involved some departure from local norms. This preoccupation with being seen to be ‘good corporate players’ was much more of a concern than attempting to promote themselves to HQ management as suppliers of new practices. This leads to a further proposition:

5a. The power of subsidiary managers to exert an influence on RD is shaped by the resources they deploy and this varies from country to country.

3.3.4. Subsidiary roles

This leads us to the final factor, the size of the sites and their function within the firms, particularly the ways in which subsidiary roles are differentiated. As the summary in Table 1 indicates, the UK subsidiaries had more influence within the wider company than those in Germany. This was true in ITco even though the UK and German operations fulfilled apparently similar roles, with the operations in both countries mainly serving customers in their respective national markets. Employment in both countries had increased and both were major bases for the firm in Europe. The position of the national subsidiaries differed, however, in that the UK operations had grown faster in the ‘IT services’ market. In Business Services there were five divisions and four of these had a European HQ. While these were ‘virtual’ in some senses – they did not have support staff, for example – there was a management group that had a geographical location and in three of the four cases this was London. This was ascribed partly to the ‘strong ties of the British subsidiary with the American operations’ (HR manager, Germany) and to the size and maturity of the UK operations. With 7000 staff in the UK as against 4000 in Germany, the UK was seen as being ‘about five years ahead of the German location in its development, client basis and headcount’ (Finance manager, Germany). The differential role of the subsidiaries was clearer still in CPGco. Significant changes in Germany arose from the selling off of production sites:

‘Things changed dramatically here when we lost the production operations. It was a question of image. We used to be an ‘‘A’’ subsidiary and suddenly we were degraded to a mere trading company. That was a real blow for us. The largest subsidiaries in Europe now are the UK and France with about 4,000 employees each. We now belong to the second grade subsidiaries in Europe’ (HR Director for Germany).

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–28 27

This ‘second grade’ status had severely curtailed the influence of German managers, limiting their ability to initiate RD.

This analysis raises two further issues for further research. First, the size, and particularly the growth, of each subsidiary seems to be linked to the prevalence of RD, leading to the following proposition:

6a. RD more commonly arises in subsidiaries that are relatively large and have recently grown.

The evidence provides some support for the expectation that subsidiary roles matter, but the impact may not be straightforward. That is, if the role is so specialized that the functions performed and technologies deployed in the sites are markedly different then the scope for diffusion to quite different sites is limited; equally, performing a function that is identical to other sites may also reduce it. In contrast, a role that is distinct in some ways but has some commonalities with other sites may be the most conducive to RD. A final proposition follows:

6b. There is a curvilinear relationship between the distinctiveness of subsidiary roles and RD.

The inter-linkages between these various factors have become even more apparent than was evident from a review of the literature. For example, we have argued that perceptions of host country contexts affect the way that subsidiaries are perceived and that this may vary by the nationality of the parent firm. In turn, subsidiary roles are shaped by the host country context and how this is perceived. Moreover, the resources that subsidiaries control affect, and are affected by, the roles that they take on. These inter-linkages should be a key element of further research.

4. Managerial relevance

A key implication for practitioners concerns corporate performance. As much of the literature argues, tapping into the diversity of practices within a multinational has the potential to enhance firm performance by strengthening the resource base of the firm. However, the logic of this analysis is that it is difficult for firms to realize RD: some are too centralized while others are too decentralized; some lack the mechanisms capable of transferring knowledge across borders; and the perceptions of senior managers can lead them to overlook pockets of expertise within the firm. Moreover, even where MNCs do engage in RD, if the process occurs in a way that is insensitive to practices being dependent on contextual supports then they may operate in a markedly different way to that which was envisaged. If managers are to use RD to beneficially augment the knowledge base of the firm then they must be sensitive to these factors.

A second implication for managers concerns relations within the firm. As we have argued, reverse diffusion should be seen as a political as well as a technical issue, a key part of the way in which managers strategize. Some managers engage in the process of RD partly to try to raise their status and claim on resources, or at least to establish themselves as good citizens within the company. In contrast, though, other subsidiary managers see RD as a threat since it entails sharing distinctive expertise with other units with which they compete and in these circumstances they may look to conceal their expertise by not engaging in it. Consequently, the strategies that subsidiary managers pursue in this respect not only affects the specific issue of whether RD occurs successfully but also shapes the power and influence of their respective subsidiaries more generally.

5. Conclusion

This paper has charted and explained variation in the extent and nature of RD between MNCs and subsidiaries. We have done this by exploring and developing various types of explanation, showing inter-relationships between the sources of variation and extending our understanding of the phenomena by identifying new ones. For example, we have shown how a partnership structure (or at least the legacy of one) is conducive to RD and how the way that national systems are perceived by corporate managers constrains the role of actors at subsidiary level. In short, we have enhanced our understanding of both inter-firm variation in the incidence of RD and inter-subsidiary variation concerning where it originates and generated a set of questions that can be addressed in future research.

The paper has limitations, of course. In particular, the six strands of explanation are not necessarily exhaustive. They were derived from the literature on RD in particular and on MNCs in general and these factors were used to structure the collection and analysis of the data. While the interviews were semi-structured and so provided some scope for unanticipated factors to emerge, this scope would have been greater with an unstructured approach. Moreover, the use of three case studies means that it is almost certainly the case that had other firms in other sectors been analyzed then different types of explanation might have emerged. In addition, this approach meant that empirical regularities across all types of MNCs could not be established. Despite these limitations, the research design has many strengths and is novel in important ways.

As the research questions in the previous section demon- strate, this is far from the final world on RD. A part of the research agenda for academics in the field of international HRM, therefore, is in the further development of the ideas advanced here. For example, can the propositions of the two types of curvilinear relationships – between RD and the extent of centralization of the multinational and between RD and the distinctiveness of the role of a particular subsidiary – be validated or further refined? Moreover, many of the research questions relate to the need to test the generalizability of the findings that emerged from the case study data and this points to the need for rigorous survey data.

References

Adenfelt, M., & Lagerstrom, K. (2006). Knowledge development and sharing in multi- national corporations: The case of a centre of excellence and a transnational team. International Business Review, 15: 381–400.

Almeida, P., & Phene, A. (2004). Subsidiaries and knowledge creation: The influence of the MNC and host country on innovation. Strategic Management Journal, 25: 847–864.

Ambos, T., Ambos, B., & Schlegelmilch, B. (2006). Learning from foreign subsidiaries: An empirical investigation of headquarters’ benefits from reverse knowledge trans- fers. International Business Review, 15: 294–312.

Ambos, B., & Schlegelmilch, B. (2008). Innovation in multinational firms: Does cultural fit enhance performance? Management International Review, 48(2): 189–206.

Andersson, U., Forsgren, M., & Holm, U. (2002). The strategic impact of external networks: Subsidiary performance and competence development in the multi- national corporation. Strategic Management Journal, 23: 979–996.

Andersson, U., Forsgren, M., & Holm, U. (2007). Balancing subsidiary influence in the federative MNC: A business network view. Journal of International Business Studies, 38: 802–818.

Bartlett, C., & Ghoshal, S. (1998). Managing Across Borders: The Transnational Solution (2nd edition). London: Hutchinson.

Belanger, J., Berggren, C., Bjorkman, T., & Kohler, C. (1999). Being Local Worldwide: ABB and the Challenge of Global Management. Ithaca: ILR Press.

Birkinshaw, J., Hood, N., & Jonsson, S. (1998). Building firm-specific advantages in multinational corporations: The role of subsidiary initiative. Strategic Management Journal, 19: 221–241.

Bjorkman, I., & Barner-Rasmussen, W. (2004). Managing knowledge diffusion in MNCs: The impact of headquarters control mechanisms. Journal of International Business Studies, 35(5): 443–455.

Bjorkman, I., Fey, C., & Park, H. (2007). Institutional theory and MNC subsidiary HRM practices: Evidence from a three-country study. Journal of International Business Studies, 38: 430–446.

T. Edwards, A. Tempel / Journal of World Business 45 (2010) 19–2828

Boehe, D. (2007). Product development in MNC subsidiaries: Local linkages and global interdependencies. Journal of International Management, 13: 488–512.

Boojihawon, D., Dimitratos, P., & Young, S. (2007). Characteristics and influences of multinational subsidiary entrepreneurial culture: The case of the advertising sector. International Business Review, 16: 549–572.

Bossard, A., & Peterson, R. (2005). The repatriate experience as seen by American expatriates. Journal of World Business, 40: 9–28.

Cantwell, J. (1992). The internationalisation of technological activity and its implica- tions for competitiveness. In O. Granstrand, L. Kakanson, & S. Sjolander (Eds.), Technology management and international business: Internationalisation of R&D and technology (pp. 75–92). Chichester: Wiley.

Dicken, P. (2003). Global shift: Reshaping the global economic map in the 21st century. London: Sage.

Dobosz-Bourne, D. (2006). Reverse diffusion of quality: Evidence for general motors UK and Poland. Journal for East European Management Studies, 11(3): 244–266.

Doz, Y., Santos, J., & Williamson, P. (2001). From global to metanational: How companies win in the knowledge economy. Boston: Harvard Business School Press.

Dunning, J. (1993). Multinational enterprises and the global economy. Wokingham: Addison-Wesley.

Edwards, T. (1998). Multinationals, employment practices and the process of reverse diffusion. International Journal of Human Resource Management, 9(4): 696–709.

Edwards, T. (2000). Multinationals, international integration and employment practice in domestic plants. Industrial Relations Journal, 31(2): 115–129.

Edwards, T., Almond, P., Clark, I., Colling, T., & Ferner, A. (2005). Reverse diffusion in US multinationals: Barriers from the American business system. Journal of Manage- ment Studies, 42(6): 1261–1286.

Edwards, T., Collings, D., Quintanilla, J., & Tempel, A. (2006). Innovation and the transfer of organisational learning. In P. Almond & A. Ferner (Eds.), American multinationals in Europe: Managing employment relations across national borders (pp. 223–247). Oxford: Oxford University Press.

Edwards, T., & Ferner, A. (2004). Multinationals, reverse diffusion and national business systems. Management International Review, 44(1): 49–79.

Ferner, A., & Varul, M. (1999). The German way? German multinationals and the management of human resources in their UK subsidiaries London: Anglo-German Foundation.

Ferner, A., & Varul, M. (2000). Vanguard subsidiaries and the diffusion of new practices: A case study of German multinationals. British Journal of Industrial Relations, 38(1): 115–140.

Frost, T. (2001). The geographic sources of subsidiaries’ innovations. Strategic Manage- ment Journal, 22: 101–123.

Frost, T., Birkinshaw, J., & Ensign, P. (2002). Centers of excellence in multinational corporations. Strategic Management Journal, 23: 997–1018.

Ghauri, P. (2004). Designing and conducting case studies in international business research. In R. Marschan-Piekkari & C. Welch (Eds.), Handbook of qualitative research methods for international business (pp. 109–124). Cheltenham: Edward Elgar.

Gourlay, S. (2006). Knowledge management and international HRM. In T. Edwards & C. Rees (Eds.), International human resource management: Multinational companies, globalisation and national systems (pp. 151–171). Harlow: Pearson.

Grimshaw, D., & Miozzo, M. (2004). High-tech business services and innovation in Germany and the UK: The case of IT outsourcing. London: Anglo-German Foundation.

Gupta, A., & Govindarajan, V. (2000). Knowledge flows within multinational corpora- tions. Strategic Management Journal, 21(4): 473–496.

Harzing, A. (2001). Of bears, bumble-bees and spiders: The role of expatriates in controlling foreign subsidiaries. Journal of World Business, 36(4): 366–379.

Harzing, A., & Noorderhaven, N. (2006). Knowledge flows in MNCs: An empirical test and extension of Gupta and Govindarajan’s typology of subsidiary roles. Interna- tional Business Review, 15: 195–214.

Hayden, A., & Edwards, T. (2001). The erosion of the country of origin effect: A case study of a Swedish multinational company. Relations Industrielles/Industrial Rela- tions, 56(1): 116–140.

Holm, U., & Deo Sharma, D. (2006). Subsidiary marketing knowledge and strategic development of the multinational corporation. Journal of International Manage- ment, 12: 47–66.

Hymer, S. (1976). The international operations of national firms: A study of direct investment. Cambridge, MA: MIT Press.

Johanson, J., & Vahlne, J. (1977). The internationalization process of the firm: A model of knowledge development and increasing foreign market commitments. Journal of International Business Studies, 8(1): 23–32.

Johnson, W., & Medcof, J. (2007). Motivating proactive subsidiary innovation: Agent- based theory and socialisation models in global R&D. Journal of International Management, 13: 472–487.

Kogut, B., & Zander, U. (2003). A memoir and reflection: Knowledge and an evolu- tionary theory of the multinational firm 10 years later. Journal of International Business Studies, 34(6): 505–515.

Kostova, T. (1999). Transnational transfer of strategic organizational practices: A contextual perspective. Academy of Management Review, 24: 308–324.

Kristensen, P., & Zeitlin, J. (2005). Local players in global games: The strategic constitution of a multinational corporation. Oxford: Oxford University Press.

Lazarova, M., & Tarique, I. (2005). Knowledge transfer upon repatriation. Journal of World Business, 40: 361–373.

Marschan-Piekkari, R., & Reis, C. (2004). Language and languages in cross-cultural interviewing. In R. Marschan-Piekkari & C. Welch (Eds.), Handbook of qualitative research methods for international business (pp. 224–243). Cheltenham: Edward Elgar.

Martin, X., & Salomon, R. (2003). Knowledge diffusion capacity and its implications for the theory of the multinational corporation. Journal of International Business Studies, 34(4): 356–373.

Minbaeva, D. (2007). Knowledge transfer in multinational corporations. Management International Review, 47(4): 567–593.

Mtar, M. (2001). French Multinationals’ International Strategy. PhD Thesis, Warwick University.

Mudambi, R., & Navarro, P. (2004). Is knowledge power? Knowledge flows, subsidiary power and rent-seeking within MNCs. Journal of International Business Studies, 35(5): 385–406.

Sanchez Vidal, M., Sanz Valle, R., & Barba Aragon, M. (2007). The adjustment process of Spanish repatriates: A case study. International Journal of Human Resource Manage- ment, 18(8): 1396–1417.

Smith, C., & Meiksins, P. (1995). System, society and dominance effects in cross- national organisational analysis.. Work, Employment and Society, 9(2): 241–267.

Tempel, A. (2001). The cross-national diffusion of human resource management practices in German and British multinational companies. Mering: Hampp.

Thory, K. (2008). The internationalisation of HRM through reverse transfer: Two case studies of French multinationals in Scotland. Human Resource Management Journal, 18(1): 54–71.

Tsai, W., & Ghoshal, S. (1998). Social capital and value creation: The role of intrafirm networks. Academy of Management Journal, 41(4): 464–476.