Corps de l’article

The last ten years have witnessed a relentless and rapid rise in levels of foreign direct investment (FDI). Between 1990 and 1999 the stock of FDI nearly trebled in value to stand at $4,759 billion (UN 2000). The driving force behind this has been the wave of cross-border mergers and acquisitions, which have increased the size and geographical spread of many multinational companies (MNCs). MNCs are now widely seen as the “primary shapers” of the international economy (Dicken 1998) since they have the potential to shift investment and jobs across borders, influence aspects of government policy and draw on elements of a range of national business systems. In these ways, MNCs increase the internationally integrated nature of production and service provision.

The role of MNCs is particularly significant in countries such as Sweden, which are characterized by a high level of outward FDI. The process of the internationalization of Swedish capital has picked up pace over the last two decades, resulting in many large firms holding more assets outside than inside their country of origin. This has added importance in the case of Sweden because for most of the post-war period the economy was highly distinctive: strong economic growth was accompanied by low levels of inequality; the state played an active role in controlling economic activity; and employment relations revolved around a highly centralized system of pay bargaining. The combination of Swedish capital being highly internationalized and the distinctiveness of the domestic economy means that Swedish MNCs are a very interesting test case of the way in which MNCs are embedded in and shaped by their country of origin. To what extent and in what ways are Swedish MNCs influenced by the distinctiveness of the country of origin in the way they manage their international workforces? How has this country of origin effect evolved as FDI has increased? In particular, how has it been eroded by the adoption of practices that originated in other systems of employment relations?

These questions are important for a number of reasons. For management, there is growing evidence that MNCs are seeking greater integration in their international operations, including their approach to employment relations. A key issue for managers, therefore, is the form that this takes: the extent to which it is based on the practices operating in the distinctive Swedish system, or alternatively on practices characteristic of other systems. Employees and their representatives in the domestic operations of Swedish MNCs are also interested parties to this process. If the country of origin effect remains strong then there is little reason to expect employment practices to differ from those operating in other Swedish firms. On the other hand, where international integration is modelled around practices experienced by the firm in its foreign operations then domestic workforces will be subject to the introduction of new working practices. More generally, a surge in foreign direct investment by large firms may be a force for change in the nature of structures and patterns of employment relations in the domestic economy. Thus, the questions are of interest to policy makers, too.

We investigate these issues through examination of data gathered from a mini-case study of a large Swedish multinational. The literature on the transfer of practices within MNCs has largely concentrated on “forward diffusion”—the diffusion of practices from the home country to foreign subsidiaries—but has tended to ignore the diffusion of practices in the opposite direction, a process that has been termed “reverse diffusion” (Edwards 1998). This can take two forms: first, “strict” reverse diffusion in which a practice is diffused from a foreign to a domestic plant with the other operations in other countries unaffected; and, second, reverse diffusion as a subset of a wider process of “flow” diffusion in which a practice originates in one of the foreign subsidiaries and is subsequently diffused to all the other plants, including those in the home country. We show how the case study firm has adopted practices experienced in its foreign operations and deployed these in other parts of the firm. This process has eroded, though not removed, the influence of the Swedish system on employment relations in the multinational, with the British and American systems appearing to exert a growing influence. We explain the findings with reference to managerial perceptions of the strengths and weaknesses of different “national business systems”: Whitley (1992: 6) defines these as “configurations of hierarchy-market relations that become institutionalized in different market economies in different ways as a result of variations in dominant institutions.”

In the following section, we consider the key features of, and changes in, the Swedish business system in general and its employment relations system in particular. The article then seeks to establish the room for manoeuvre that MNCs have in diffusing practices across borders. In the fourth section, we present the evidence from the mini-case study, which is then located in an analysis of comparable studies of MNCs in the discussion.

The Swedish business system and the challenge of internationalization

The system of employment relations that existed in Sweden for most of the post-war period, termed “centralized self-regulation” by Kjellberg (1998), originated from the 1938 Saltsjobaden Agreement. Against a background of the social-democratic government threatening greater legal intervention to solve the problem of high levels of industrial conflict, employers and unions reached agreement on a range of formal elements to the system, such as new grievance procedures and greater regulation of industrial disputes. A supplementary step was to implement a system of highly centralized, economy-wide pay bargaining. Perhaps more importantly, though, was the informal understanding that was fostered by the agreement, in which both employers and unions accepted the need to co-operate to promote economic growth. Both parties saw the opportunity to realize gains, but also acknowledged the need to make sacrifices as a part of this understanding. The newly strengthened central union confederation, the LO, was willing to accept “socially responsible” pay rises as the price for being able to pursue a “solidaristic” wages policy which aimed to minimize pay differentials. Individual employers, on the other hand, were prepared to accept the loss of control over pay-setting that centralized bargaining involved in return for the assurances of industrial peace and the guarantee of not being undercut by low-paying competitors.

More generally, there were mutual benefits to a system that closed off the option of securing profitability through cutting labour costs, forcing unproductive firms to close or rationalize and promoting a process of “dynamic modernization” (Iversen 1998) which rewarded high-tech, efficient firms. The process of dynamic modernization was facilitated by two additional factors. First, the state developed an “active labour market” program designed both to ease the redeployment of labour from one sector to another and also to facilitate corporate strategies that relied on a highly skilled workforce. Second, Swedish firms enjoyed a stable system of finance that provided the platform for long-term investments, with ownership of large Swedish firms being concentrated among a small number of investment groups, banks and wealthy families, such as the Wallenberg empire which is a key long-term shareholder in all large Swedish firms except Volvo.[1] In sum, the formation of this system was the result of an “historic compromise” which, in the words of Coates (2000: 96), constituted a “quite unique class accommodation.”

One of the distinguishing features of Sweden in this period was that it was characterized by high levels of economic growth and low levels of inequality and unemployment. Large firms appeared not to be handicapped by high labour costs and, remarkably, given the size of the country, in most industrial sectors Sweden possessed at least one leading company in international markets: Volvo in automotive; Electrolux in consumer goods; Astra in pharmaceuticals; and so on. However, in the last two decades, economic performance has been much less favourable. Economic growth slowed to levels below the OECD average and unemployment, which had previously been below 3%, rose in the early 1990s to around 10% and stayed just below this level for most of the decade. For firms, productivity growth was unfavourable in the international context, undermining the competitive position of Swedish firms in world markets. While the performance of the economy created pressure for change in employment relations, the internal tensions within the model were already apparent.

One such tension concerned the corollary of the centralization of authority within the LO, namely the weakness of unions at workplace level. Responding to pressure from their members, the unions successfully lobbied the government for legislation designed to increase unions’ influence over workplace issues. Consequently, the principle of self-regulation was abandoned as legislation relating to a range of issues, including job security and co-determination, was introduced. The co-determination law of 1976 was of particular significance since it required employers to negotiate with unions over proposed changes to issues such as working hours and conditions of employment. Private sector employers were opposed to the law initially but, in 1982, reached agreement with the unions on the nature that co-determination should take, accepting “the need to involve union workplace organization in order to increase productivity” (Kjellberg 1998: 106). Arguably, this legal change reinforced and institutionalized a tradition of consulting over, and negotiating, changes in work organization.

A perhaps even more significant tension was the way the system of centralized bargaining, and the moderate wage increases that it produced, meant that the ability of highly profitable firms to pay high wages was not fully utilized. Thus the unions instigated moves to secure a greater share of profits through the creation of “wage earner funds,” which were an attempt, short-lived as it turned out, to channel profits into union-controlled accounts. In essence, this represented an attempt to renegotiate the historic compromise in a way that employers found unacceptable (Coates 2000).

Against this background, large Swedish firms became openly hostile to the system of employment relations. In particular, employers sought to free themselves from centralized pay bargaining, which many perceived as posing serious constraints on their ability to set pay levels in accordance with the performance of individual sites and to use pay to promote quality and flexibility (Kjellberg 1998). This was particularly the case for firms seeking to replace piecework systems with new forms of work organization based on teamwork (for example, see Berggren 1993 on Volvo). More generally, employers sought to erode the tripartist traditions of the model and also campaigned against the growth of the public sector and the high levels of taxation it entailed, arguing for the neo-liberal economic policies that had become vogue in much of Europe and North America.

At the same time, large Swedish firms stepped up their internationalization strategies. Foreign direct investment from Sweden has increased rapidly in the last two decades. In 1980, the stock of outward FDI from Sweden amounted to less than $4 billion; by 1999 this had increased to more than $104 billion. As a percentage of all FDI, this represented a trebling, from 0.7% to 2.2% (UN 2000). Many Swedish firms have a long international history because their small home market led them to expand abroad at an early stage in their development (Hedlund 1981). The recent surge in FDI means that many are now much more highly internationalized than ever before: firms like Volvo and Electrolux employ more than half of their workforces outside Sweden, while others such as Astra and Pharmacia have merged with foreign firms. Many large Swedish firms, therefore, became much less concentrated in their country of origin. Hyman (2000) has described the internationalization of “national capital” and the challenges this has posed for national systems of employment relations. As he put it, “The most dramatic instance, perhaps, was the case of Sweden: the major employers in effect ’joined’ the EU long before the country’s formal accession, and demolished the classic centralized ’Swedish model’ of industrial relations the better to pursue more company-specific and internationalized employment policies” (2000: 8).

Of course, MNCs are strongly influenced by the business system in the country in which they developed. This is the case even for the most internationalized MNCs which still tend to raise finance and conduct R&D primarily in their home base, and also fill senior managerial positions with home country nationals (e.g. Ruigrok and van Tulder 1995). Accordingly, there is a detectable country of origin effect in most MNCs. Stressing the firm roots that most MNCs have in their domestic business system, and the influence these roots exert, is not to suggest that the country of origin imposes a straitjacket on management, forcing all MNCs from a particular country to be alike. Factors such as the ownership patterns of firms vary from one company to another, while the influence of key individuals, particularly founding families, in the development of firms is a further force for heterogeneity within a country. Nonetheless, the nature of dominant institutions in the home business system informs the behaviour of MNCs to a greater or lesser degree, creating general tendencies by nationality in the way they manage their international workforces (for a review, see Ferner 1997).

However, as MNCs become more geographically dispersed, they have increased scope to draw on elements of other business systems, adopting and diffusing practices perceived to be superior to those in place in the domestic operations. For Swedish MNCs, the combination of the apparent dissatisfaction of senior managers with elements of the domestic employment relations system, the small domestic market and the internationalization of capital has created considerable potential for this process, which in earlier work we have termed “reverse diffusion.” The process is evident in a number of recent studies of MNCs, to which we return in the final section. Before examining the case study evidence, we consider the freedom for manoeuvre that management in MNCs enjoy in attempting to diffuse practices across their international operations.

The diffusability of employment practices across borders

The character of the different business systems in which MNCs operate influences the scope which managers have to transfer practices across their international operations. The ease with which a practice can be transferred across borders is determined in part by the extent to which it is dependent on supportive and distinctive extra-firm structures. These extra-firm structures include the legal framework of employment relations, the nature of labour market institutions and the attitudinal and behavioural norms that characterize employment relations in a particular country. All employment practices, of course, are to some extent dependent on these legal, institutional and cultural supports, which we refer to here as “props.” The ease with which a practice can operate outside its original home environment—in other words, the extent to which a practice is “diffusable”—is determined in part by its dependence on these props.

One area in which practices may be highly dependent on such props is training. Some aspects of a firm’s approach to training are dependent on supportive institutions, a good example of which is the “dual” system of training in Germany. The role of the colleges and training bodies in administering, monitoring and certifying the system provides a crucial underpinning for firm-level practices. Consistent with this, Dickmann’s (1999) study of German multinationals showed that they have been constrained in their attempts to introduce German-style vocational training into their UK subsidiaries because the British economy lacks the “broader business institutions necessary to underpin particular practices” (Edwards and Ferner 2000: 7). Of course, some training practices, particularly those that are employer-led, are more diffusable in that they are less dependent on a set of supportive institutions (see Edwards 1998 for a discussion of a British MNC).

The nature and distinctiveness of the national business systems in which MNCs operate can also limit the diffusion of practices in a different sense. Managers at the HQ of a multinational may seek to operate a practice in a number of countries but might be prevented from doing so by the legal, institutional or cultural “constraints” of the country to which the practice is directed. Organizational actors in the recipient unit of a multinational may try to resist the introduction of a practice and may use their legal powers, rights provided by institutions, or appeals to the importance of local “custom and practice” in order to thwart the HQ’s plans. In this sense, some practices may not be “diffusable” because of the constraints posed by the nature of the host business system.

One area where such constraints are notable is in relation to work organization. There is clearly significant scope for MNCs to diffuse practices in this area; the literature on Japanese firms, for instance, shows that they have been innovators in lean production, taking practices such as cellular manufacturing and JIT to their foreign subsidiaries. However, the constraints posed by business systems do limit a multinational’s freedom of manoeuvre in this area. Changes to work organization have to be negotiated with employee representatives in most countries. Accordingly, Ortiz (1998) has shown how unions at Opel Spain were able to significantly influence the way that teamwork operated. As a result of union influence, team-working was introduced on a one-year experimental basis with workers only being a part of it if they so chose, while maintenance workers, who had been opposed to the initiative, were eventually excluded.

In many instances, these constraints are partial rather than absolute: that is, managers at the HQ may be able to diffuse a practice, but it may need to be altered so that it can be implemented in the new business system. As Edwards and Ferner (2000: 10) put it, a “practice may not operate in the same fashion in the recipient as in the donor unit but, rather, may undergo transmutation as actors in the recipient seek to adapt it to pre-existing models of behaviour, assumptions and power relations.” Thus the formal substance of a practice may be diffused but the operation of this practice may differ between countries.

One illustration of this process is the diffusion of Japanese management techniques. Broad (1994) has shown how the British managers in a UK transplant of a Japanese multinational resisted moves by the parent firm to shift responsibility for quality to teams of operators, preferring to retain the right to take decisions in this area themselves. Broad (1994: 58) argued that this reflects the “traditional obsession of British managers with prerogative and secrecy.” In this case, “Japanese” practices were diffused but took on a different form in the new environment. A further illustration of transmutation is in the area of corporate culture. In recent years a number of German MNCs have adopted aspects of the business cultures of the US and the UK, such as “mission” and “value” statements. These statements survive the transition partially in tact, in that the language remains the same. However, the significance of terms such as “shareholder value” in the German context is quite different where shareholders are commonly family members or banks which have held long-term stakes, rather than the uncommitted and distant shareholders which dominate Anglo-Saxon firms. Ferner and colleagues have termed this process “Anglo-Saxonization,” but argue that it has occurred “in the German manner” (Ferner and Quintanilla 1998; Ferner and Varul 2000).

Differences between national business systems, therefore, limit the diffusability of employment practices. This is partly because practices are dependent on the “props” present in the system in which they originate, and partly because their introduction to other countries is subject to the “constraints” posed by the recipient systems. We have also seen how practices can be modified, or “transmuted,” to fit the new environment. However, it is the differences between national business systems that also create the potential for cross-border diffusion in the first instance, as MNCs seek to gain a competitive advantage through transferring practices perceived as delivering improved performance across their operations. One possibility in Swedish MNCs is that senior managers have looked abroad for solutions to their perceived problems at home.

One area in which Swedish MNCs may have sought to do so is appraisal and reward. The narrow differences in earnings between and within occupational groups that resulted from the “solidaristic wages policy” of the LO was something about which management had become uneasy by the 1980s (Kjellberg 1998). As we argued above, many large Swedish firms expressed the desire to link pay more directly to firm-specific factors in general and to use pay to promote quality and flexibility in particular. Therefore, it is plausible that Swedish MNCs have looked to their foreign subsidiaries for new appraisal techniques and systems of variable reward. One example is varying pay so that it is linked to individual performance, which is measured against specified objectives. Such “performance-related pay” (PRP) schemes were pioneered in the US and, to a lesser extent, the UK in the 1970s and 1980s (Sparrow and Hiltrop 1994; Storey 1992). Another example is profit related pay and share option schemes that are linked to the achievement of specific organizational targets. Whether such performance management practices are diffused across Swedish MNCs, and if so, the way in which such practices operate in Sweden are questions of great importance.

A second area in which Swedish MNCs may have looked abroad for new practices is management development. Historically, senior managers in Swedish MNCs appear to have adopted an informal management style, allowing subsidiaries considerable autonomy. Hedlund (1981) has described this as a “mother-daughter” structure in which the heads of subsidiaries reported directly to the group president, with the subsidiaries rarely being integrated into the business divisions in the home country. The increasingly integrated nature of the international economy and the competitive pressures it poses has raised the issue of how the geographically dispersed units are to be integrated with each other. In looking for ways of controlling and coordinating their international operations one possibility is that senior managers have looked to Anglo-Saxon firms, which have long between characterized by a more structured and formalized management style. Many Anglo-Saxon MNCs have managed their operations through the development of a cadre of international managers with experience of different countries (Ferner and Varul 2000). The way in which these managers were trained, selected, appraised and remunerated is likely to have been of considerable interest for senior managers in Swedish MNCs. The extent to which practices relating to management development are based on those characteristics of Anglo-Saxon firms is, therefore, a second important area for investigation.

These two areas, performance management and management development, are ones that formed the focus of the case study. We consider the way in which the multinational had adopted practices in these areas that had originated in their foreign subsidiaries. In so doing, we are sensitive to the role of the “props” and “constraints” posed by aspects of national business systems in limiting and shaping the process of diffusion. It is to an analysis of the case study data that we now turn.

The changing role of the country of origin effect in practice

The case study firm, Swedco, is a major Swedish multinational, producing high-tech manufactured goods and employing tens of thousands of employees. It is around one hundred years old and has operated internationally for most of this time. During the last two decades the company has gone through a period of international expansion so that it is now highly internationalized, selling about 95% of its products abroad in over one hundred countries. Furthermore, more than half of Swedco’s workforce is employed outside the home base.

The primary research instrument in the case study were semi-structured interviews with key respondents. These were conducted with three managers and a union official in the corporate HQ in Stockholm and a manager in each of the Belgian and British subsidiaries. The interviews lasted between one and three hours and focused on the firm’s approach to performance management and management development. These data were supplemented with analysis of documents and secondary sources, such as shareholder listings and annual reports. It is acknowledged that this is a “mini case study,” involving a relatively small number of interviews. However, the data do have significant strengths: the interviews were lengthy and open-ended; the documents supplemented the interview data in important ways; and the firm possesses certain characteristics that mean we can see it as a “critical case” of how the country of origin effect has evolved in Swedish MNCs.

The characteristics which accord Swedco the status of being a critical case partly relate to its organizational features. In particular, the firm possesses all six of the corporate characteristics identified by Edwards (2000) as promoting “reverse diffusion”: it is structured around international divisions; has a low degree of diversification; production is integrated internationally; there is a high degree of global spread; growth has been in part through acquisition; and it is mature in the sense that it has operated at the international level for a long time. Based on a study of British MNCs, Edwards showed how these features of multinationals made it more likely that senior management would have the motivation and the scope to draw on practices developed in the foreign subsidiaries. Since the process of reverse diffusion challenges the influence of the country of origin, these corporate characteristics make Swedco a particularly interesting example of the way in which the country of origin effect is eroded over time.

The other reason why Swedco can be seen as a critical case is its nationality. Many elements of the Swedish business system create conditions in which we might expect the country of origin effect to have been eroded by influences from other countries. First, the size of the domestic economy has meant that many large Swedish firms are highly exposed to international markets and have extensive international operations. Second, as we have seen, employer dissatisfaction with some of the distinctive structures within Sweden, particularly the tradition of solidaristic, centralized pay bargaining, have created fertile grounds for new practices to be diffused from elsewhere. Third, traditions of management style within Swedish firms, such as the team-oriented nature of decision making, mean that Swedish MNCs are likely to be favourably disposed to outside influences. Thus national as well as organizational factors shape the incidence and process of reverse diffusion and the way in which the country of origin endures or is eroded.

The research revealed a distinctly Swedish element to management style, which was particularly evident in relation to the firm’s approach to supervision, the character of hierarchical relations between occupational groups and the nature of communication within the firm. In international context, Swedish workers operate with relatively little direct supervision; indeed, there is no direct translation in Swedish for the word “supervisor” (Anderson 1995: 72). Swedish firms appear to have been at the forefront in the use of semi-autonomous teams. For instance, Berggren (1993) argued that the development of team working at Volvo enhanced the autonomy of workers and increased their influence over the pace and nature of their work. The legal and institutional context in Sweden, particularly the system of co-determination which reinforces union-channel workplace participation, provides a supportive context for such practices. In Swedco, the manager of international assignments described the company’s attempts to spread a “democratic” approach throughout the multinational and saw this as contrasting with approaches found in other countries:

I want to let my guys loose. I don’t want to control them and stand behind their backs. This is typically Swedish, to be a coach. When something goes wrong I have to go in and correct it and so on, but we are all part of a big team. … I think some people, some of the English people over here, have a small problem understanding how we can be so democratic.

Related to this “democratic” approach is the team-oriented approach to relationships in which it is acceptable to “bypass the hierarchy” (Anderson 1995: 77; Hedlund 1981). Accordingly, one of the British managers in Swedco testified to the way this aspect of management style was present beyond the Swedish operations:

The company encourages a Nordic approach to openness. Swedes think nothing of jumping the hierarchy to put forward their ideas. They say to you “I didn’t come to you because I know that you had to go to him anyway.” Some British people find it irritating, but personally I like the style because it encourages people not to be intimidated by a formal chain of command.

The way in which Swedish managers strive for consensus and co-operation, what Anderson (1995: 76) terms the “quest for accord,” was also evident. Arguably, one instance of this tradition is the “historic compromise” of the 1930s described above, while the institutions created at this time were predicated on an assumption of compromise and concessions in industrial relations. Daun (1989, cited in Berglund and Lowstedt 1996) argues that Swedes are distinguished by open communication which increases the scope for a consensus to be reached (see also Lawrence and Edwards 2000). This resulted in an expectation that disagreements would be worked through in a calm, orderly manner. One of the Belgian managers saw this as contrasting with his own home country approach:

I sometimes find it very frustrating as a non-Swedish HR manager that I always have to be professional, diplomatic and avoid waves. I believe that there are a number of cases when you can’t do that. You cannot always agree or compromise. Sometimes you have to say no. In Belgium, we raise our voices, we explode sometimes. But Sweden says this is something you must not do.

The distinctiveness of the Swedish business system shows up in other characteristics of the firm. The stability of ownership enjoyed by Swedish firms has enabled senior management to preserve Swedco’s independence in an industry increasingly characterized by consolidation through mergers. This is despite 50% of the shares being held by individuals and institutions outside Sweden; the distinction between A and B shares means that the three largest Swedish shareholders possess only about 7% of the company but control almost 75% of the voting rights. Swedco itself has grown mainly through investment in new, “greenfield” sites and through a series of collaborative joint ventures and friendly acquisitions rather than through the hostile takeovers which are characteristic of large Anglo-Saxon MNCs. The Swedish business system has also shown itself to be conducive to competitive success in sectors requiring high levels of investment. Swedco appears to have benefited from this, operating in a high-tech industry with one quarter of its workforce employed in research and development.

There is evidence, therefore, of a distinctive Swedish element to management style that the company has attempted to spread throughout the company via expatriate assignments and international management training. This management style aims to encourage creativity, team working, openness of communication, the sharing of ideas and consensus in decision making across borders. Interestingly, when this distinctively Swedish style operates at the international level it favourably disposes the company to learning about systems and practices operating in other countries. The consequence is an apparent paradox: the country of origin effect favourably disposes the firm to the influences of other national business systems rather than closing off these influences. In this respect, the “Swedishness” of the firm encouraged the reverse diffusion of employment practices. Accordingly, there was evidence that practices in the areas of performance management and management development had been identified in the foreign subsidiaries and diffused across the firm.

Given the highly competitive nature of Swedco’s industry and the high skill levels of its workforce, a particular HR challenge is to recruit and retain skilled engineers and managers. One way in which the company has dealt with the challenge is by developing “flexible” or “variable” compensation systems. Such systems, which include bonuses, profit sharing and convertibles, are becoming more popular in Swedish firms (Berglund and Lowstedt 1996; EIRO 2000). In Swedco, one element of this has been the bonus systems that are linked to individual and company performance, which have been developed by an international policy-working group. One illustration is a debentures option program for 500 key staff globally, in which a bonus linked to base salary is paid if a pre-determined profitability target is met. In addition, for very senior managers, there is a “Short-term Incentive Plan” which focuses on the achievement of short-term goals, a practice which is characteristic of large Anglo-Saxon firms. Bonus systems are not confined to senior staff, however, as three years ago all employees obtained the right to subscribe to a convertible debentures scheme, something that about 40% of staff world-wide have taken up.

As well as this “bonus” scheme, Swedco has also developed an appraisal and performance-related pay (PRP) scheme that affects “base” pay. Many observers have argued that PRP schemes were most prevalent in the USA and have spread to Europe as the interest in “HRM” grew (e.g. Sparrow and Hiltrop 1994; Storey 1992). An individual performance appraisal system was introduced more than a decade ago, the purpose of which was to measure employees’ abilities against a “competence development plan.” Recently, these appraisals have assumed greater importance as they are now linked to an individual PRP scheme that affects all employees. The scheme comprises a framework developed at headquarters, which is then adapted locally to allow for the constraints posed by the business system in question. In fact, the company has been relatively unconstrained in implementing the scheme with even collective bargaining posing only a minor constraint. In the UK, the company’s negotiations with its one recognized union sets only a low annual pay rise, allowing management to vary the rest of the rise according to the performance of individual employees. In Belgium, the system of sectoral bargaining leaves some scope for PRP at organizational level because the sector level rise is very low. The company works council does not pose constraints to the operation of PRP either, as they do not deal with issues of pay determination. And, crucially, in Sweden too there have been relatively few constraints on the introduction of PRP. This was partly because of the union’s ready acceptance of PRP, which was described by a union official:

In Sweden the trade unions have been very positive about change and we have tried to develop with the company. The trade union no longer wants decision making power. We want to negotiate and consult, but we don’t want the power and the conflict.

The weakness of these constraints reflects the changing nature of employment relations in Sweden. One important development has been the decentralization of bargaining, which has allowed employers greater flexibility to link pay to performance. Associated with this is the weakening of the LO’s “solidaristic wage policy” in recent years, resulting in a widening of pay differentials, particularly for white-collar workers (Kjellberg 1998). A further factor making it easier for Swedish firms to introduce PRP is the changing attitudes of younger workers who more readily accept the principle of greater wage differentials between individuals. Our respondents identified this as being particularly important in Swedco with its relatively young workforce; half of the employees are 35 or under. The findings provide an interesting contrast to Ferner and Varul’s (1998) study of German MNCs in which regulatory and institutional factors heavily constrained the adoption of an Anglo-Saxon style performance management system, requiring a significant modification to fit in with elements of the German business system.

A similar process of adopting Anglo-Saxon practices was evident in relation to management development. There has been a concerted attempt in recent years to develop a cadre of managers from across the multinational, with the subsidiaries given the role of identifying and nurturing non-Swedish employees as part of the international managerial workforce. As the Swedish HR manager explained:

The subsidiaries are putting forward more candidates now than they used to. They are improving. If you look at the number of candidates, the majority are still from Sweden, but you can see an increase in the numbers coming from abroad. By the time the next generation is in place, I think it will have turned around.

Recently, the firm has established a “Corporate Executive Team” that operates from four offices, only one of which is in Sweden while two are in the US and one in the UK. This coexists with a “Management Institute” that has the role of assessing and training managers from across the firm who have the potential to fill senior roles within the firm. This is part of the firm’s approach to succession planning, which involves the identification of future senior managers, known as “high potentials.” As the Swedish HR manager put it:

We have a group of 3 or 4 people from the management planning department who go around the world every year to every local company to make an evaluation of their management planning system. They ask, “what have you done? what kind of people have you identified? what plans do you have for these people?” So here in Sweden we know that the subsidiaries are doing it the right way.

The approach to assessing and training managers appears to reflect the influence of Anglo-Saxon multinationals. The identification of “high potentials” as part of an international cadre of managers is, according to Ferner and Varul (1998: 34), a common trait of British and American MNCs. In Swedco, the British subsidiaries appeared to have been particularly influential in the formation of the policy on management development. The manager of the Management Institute, a Swede, confirmed the important role of the UK subsidiary and UK universities in assisting her in policy development on training programs and management development:

When I am developing a training program for managers, I always include the UK. Firstly, it ensures that I get the language right but, secondly, there are a lot of good training and management development ideas in the UK that I would like to benefit from. I always bring someone in from the UK site onto the team. We are also developing links here in Sweden with the UK universities such as Cranfield and the LSE. They have good management training and methodology. We also talk to UK professors regarding what will happen in the future and what is going to be interesting.

In implementing this approach, Swedco appears to have undergone a similar process to that detected by Ferner and Varul (1998) amongst German MNCs. However, while the German firms appeared to modify the “Anglo-Saxon” approach to fit the German business system, Swedco appeared to be less constrained. For instance, in some of the German MNCs, there was a culture of secrecy surrounding the handling of “high potentials” who were not advised of their status in order to fit in with the egalitarian culture of the German firm. In contrast, within Swedco, no such modifications were made, with respondents in all three sites confirming that their high potentials were made aware of their status from the outset so as to motivate them to “develop themselves.”

How best can we understand the role of national business systems in the process of diffusion at Swedco? The evidence demonstrated that the practices in the two areas of employment relations studied, performance management and management development, were not highly dependent on a set of supportive “props” from the countries in which they originated. Moreover, the introduction of these practices into the domestic operations was not significantly constrained by the nature of the Swedish business system. Hence, the “diffusability” of the practices in question was high, and there was relatively little in the way of “transmutation.” Of course, the picture may well be very different for other areas of employment relations, such as work organization, employee representation and disciplinary and grievance procedures, where the props and constraints may limit the scope for diffusion to a greater extent.

In sum, while the influence of the Swedish business system over the case study firm was clearly visible, there is evidence that this has been eroded over the last decade or so. This can be assessed in terms of the perceptions of senior managers in the company of the strengths and weaknesses of the different business systems in which they operate. Our evidence suggested that senior managers in the firm perceived the USA and the UK as providing practices in the areas of performance management and management development that were seen as desirable. Practices from the two Anglo-Saxon countries were diffused across the firm’s international operations, including those in the home country. Indeed, Swedco is highly exposed to the Anglo-Saxon system: it makes substantial sales in the large American and British markets; it has 15% of its total employment in these two countries, including divisional head offices and R&D sites; a third of the firm’s shares are traded on the London and New York stock exchanges; and it has entered into a series of joint ventures with American firms. Thus the firm has undergone a process that has much in common with the process of Anglo-Saxonization of German firms referred to above, and it is a comparison with other studies of MNCs which forms part of the discussion.

Discussion

It is clear that MNCs continue to be firmly embedded in, and strongly influenced by, their country of origin. In many cases, MNCs benefit from this because the depth of their contacts in the domestic business system enable them to exploit its strengths. Thus expertise concerning systems and practices with which the firm has familiarity in its home base can constitute a source of competitive advantage for international firms (Porter 1990). Indeed, Porter and colleagues argue that Swedish MNCs “are dependent upon Sweden as the dynamic diamond for their specific technologies and products” (Solvell, Zander and Porter 1993: 213). This embeddedness, therefore, results in a distinctive country of origin effect in the management of HRM. Yet the country of origin will not always provide systems and practices that senior managers in MNCs seek to deploy on an international basis. Where the domestic business system is perceived as being weak, the firm may look abroad for new practices. Its own international operations mean that it is partially embedded in other business systems, of course, giving the firm the mechanisms to draw on a different set of practices. In this way, “reverse diffusion” may occur in those areas where the home country does not provide an attractive model.

Solvell, Zander and Porter’s (1993) description of Swedish MNCs charts the way in which many have made limited moves to draw on the strengths of the business systems in which they have foreign subsidiaries, but also shows how the Swedish business system has shaped this process. As they put it, the “successful tapping of foreign diamonds was selective, and could not have worked without the support from a dynamic home base” (1993: 197). This diffusion of practices from the foreign subsidiaries shows up clearly in a detailed study of ABB, the Swedish-Swiss multinational, which revealed that while the firm was still influenced by its original roots this country of origin effect had been eroded markedly. Bélanger et al. (1999: 258–259) describe the firm as an “Americanized European-based company” that is “trying to implement American ideas.” They went on: “On the global level ABB’s culture is corporate, cosmopolitan, but as we keep emphasizing, Americanized.” The evidence from Swedco is consistent with this. It suggests that the company looked abroad for practices that enhanced flexibility in pay and appraisal and in developing managers. In these two areas, senior managers appeared to view “Anglo-Saxon” practices as having the potential to provide the sorts of practices that they were seeking. Of course, the process of adopting practices from foreign subsidiaries will not have been felt so strongly in all areas, particularly those where the constraints to diffusion are stronger, such as employee representation and work organization, which are strongly influenced by the legal underpinning of co-determination.

Moreover, we have seen how it was the very Swedishness of the firm, particularly its “democratic” management style and openness of communication, which exposed it to outside influences. This presented an interesting paradox; the distinct country of origin effect promoted the process of learning from other countries, which in turn eroded the country of origin effect. Many classifications of different types of MNCs assume that a strong country of origin effect is incompatible with the adoption of practices from overseas because the country of origin effect results in the diffusion of practices across borders flowing from the home base to foreign subsidiaries rather than vice versa. For instance, one category in Perlmutter’s (1969) typology is “ethnocentric” firms in which home country systems and practices are assumed to be dominant and home country practices are transferred to foreign subsidiaries. A further category, the “geocentric” firm, which attempts to build a common approach through an amalgamation of practices from a variety of national systems, can only emerge as the ethnocentric orientation is eroded. (Perlmutter also identified a further category, the polycentric firm, which adopted a decentralized approach). However, the case study demonstrated how the national influence from Sweden—the small domestic economy, the structure of the business system and the nature of managerial traditions—pre-disposed the company to learning from foreign systems. In this sense, the country of origin effect did not constitute an attempt to push Swedish systems and practices throughout the firm. Rather, the country of origin effect comprised a basic orientation that favourably disposed the firm to learning from foreign subsidiaries, promoting a geocentric element to management style in the form of the search for new practices. Thus in this case the country of origin effect promoted, rather than impeded, the process of adopting practices from abroad.

One way of making sense of the apparent paradox is to distinguish between different ways in which national business systems inform the behaviour of MNCs. One level is the basic orientation of the firm, the key values and assumptions that underpin management style; the other is the nature of particular practices operating in the firm. In Swedco, the basic orientation was strongly shaped by the firm’s domestic business and senior management made a concerted attempt to spread this across their international operations. This orientation was consistent with, and even promoted, the search for practices from foreign business systems. In this sense, reverse diffusion occurred “in the Swedish manner.”

Comparisons with evidence from other studies of MNCs are revealing in four respects. First, the nature of practices that MNCs look to diffuse from their foreign subsidiaries (the “substance” of reverse diffusion) is strongly influenced by senior managers’ perceptions of the inadequacies of the systems and practices of the country of origin as the basis for international operations. Thus the “Anglo-Saxonization” of German MNCs occurred in relation to organizational structures, control systems and management development, areas in which the German business system was perceived as being an inappropriate model on which to develop greater international integration (Ferner and Quintanilla 1998; Ferner and Varul 2000). A similar process is evident among many French MNCs (Mtar and Quintanilla 1997). Evidence from British MNCs, in contrast, suggests that where reverse diffusion occurs it is in the areas of work organization and training, both of which are areas where the British economy and system of employment relations has traditionally been very weak (Edwards 1998, 2000). Arguably, this is consistent with American MNCs in the automotive sector trying to revolutionize work organization along the lines of Japanese production. Hence, the substance of reverse diffusion varies according to managerial perceptions of the relative strengths of the country of origin.

Second, the erosion of the country of origin effect is related to the timing and stage of internationalization. German MNCs have been relatively late at internationalizing, having relied until recently on exporting rather than international production. Thus the process of Anglo-Saxonization has been a recent phenomenon. In Swedco, however, the process of adopting Anglo-Saxon style practices began a long time ago, reflecting the much earlier development of international production. This is not to imply that the erosion of the country of origin effect through reverse diffusion can simply be read off from the length of time a firm has operated internationally; many MNCs have operated at the international level for decades but have not sought to integrate their operations and, consequently, have not undergone the process of reverse diffusion. However, amongst those MNCs under pressure to achieve a degree of international integration, the “maturity” of their production or service provision process shapes the strength of the country of origin.

Third, the evidence in this article suggests the need for a reconsideration of the concept of reverse diffusion. In other work on reverse diffusion, the emphasis has been on diffusion within the boundaries of the multinational. For instance, in Engineering Products, practices that were developed and operated in the French, Spanish and American plants were identified as being suitable for other units and were subsequently diffused across the firm (Edwards 1998). Similarly, in many German MNCs, practices were identified in the firms’ American and British plants, which took on the role of a “vanguard” subsidiary (Ferner and Quintanilla 1998). In Swedco, however, the practices that were diffused from outside Sweden were not directly from their British or American subsidiaries but rather were from the contacts to which these had given rise. For instance, the firm’s approach to management development was influenced by the links that managers had established with UK business schools. Furthermore, Swedco’s linkages with the British and American business systems also enabled it to engage in a process of benchmarking against Anglo-Saxon firms. Hence, we distinguish between “internal” reverse diffusion in which practices operate in the foreign subsidiaries before being diffused to the home country, and “external” reverse diffusion in which practices are identified in the business systems in which a multinational has units.

Fourth, the analysis is revealing in showing the links between different sets of factors that promote reverse diffusion. As we noted above, Swedco possessed all of the organizational characteristics identified in earlier work as creating the conditions in which reverse diffusion is likely to occur. In this paper, the main focus has been on the way in which its incidence is shaped by the national influence—the size, structure and managerial traditions of the home country business system. In particular, we argued that many features of the Swedish business system promoted this form of diffusion: the small size of the domestic market exposes Swedish MNCs to foreign influences to a greater extent than in MNCs of other nationalities; some structural elements of the system provide an incentive for firms to look abroad for practices that can be deployed at home; and some Swedish managerial traditions dispose senior managers in MNCs to learning from other countries. These twin lines of analysis both have the potential to further our understanding of reverse diffusion, but should not be seen as mutually exclusive. For instance, the highly internationalized nature of many Swedish MNCs is clearly related to the small size of the domestic business system. Thus both organizational and national business system characteristics, and the interaction between them, shape the incidence of reverse diffusion.

Finally, to return to an issue raised at the outset, what are the implications of this analysis for managers, employees and the wider business systems in which they operate? In comparable multinationals, that is those which are highly internationalized and integrated across borders, this article suggests that senior managers are increasingly engaging in the search for new practices from other countries and that they are doing so through the creation of management structures capable of identifying and diffusing practices. The evidence presented here adds to a growing body of work that shows that in MNCs based in continental Europe this process results in managers being assessed, paid, trained and selected for promotion in ways that are characteristic of large Anglo-Saxon MNCs. These changes may present opportunities to some managers but will also present challenges.

The same is true for employees, too, of course. In particular, Anglo-Saxon style performance-related pay appears to have eroded the narrow distribution of earnings characteristic of Sweden and, more generally, highlights the differing levels of performance among individuals within a work group. The implications for employees in Sweden and elsewhere in Europe are serious given that even in Britain the evidence on PRP in practice demonstrates that it rarely has a motivating effect on a workforce. The diffusion of variable forms of pay also presents a challenge to unions; the evidence demonstrated that the role of collective bargaining over pay was being eroded. Arguably, the diffusion of such practices compounds the difficulties that many unions have in negotiating with multinationals, particularly the threat of relocation that management is able to use in order to gain acceptance to the introduction of new practices.

The consequences for the domestic business system are also of considerable potential importance. Specifically, the erosion of the country of origin effect in MNCs could be a force for change in the home country system as a whole. Whitley (1999: 20) has argued that “MNCs from distinctive and cohesive business systems with strong associated institutions governing economic activities may well become more complex and differentiated as a result of FDI—and so encourage novel forms of co-ordination to develop—but are unlikely to change their fundamental characteristics.” The findings in this article support this view, but only to a point. There is a growing body of evidence showing that MNCs from Germany, France and Sweden, which certainly possess distinctive and cohesive systems with strong institutions, do indeed encourage novel employment practices to develop. However, MNCs that are striving to achieve a high degree of integration in their processes of international production or service provision are also likely to spread these across their operations. Accordingly, the evidence demonstrates that these innovations have a significant impact on the home country operations of MNCs in some areas of employment relations. This erodes, but does not eradicate, their fundamental characteristics. In a country where large multinational firms are particularly influential in terms of employment and output such as Sweden, we might expect these innovations to have a significant impact on the wider business system.