The Complexities of Relocation and the Diversity of Union Responses: Efficiency-oriented FDI in central Europe G Meardi, P Marginson, M Fichter, M Frybes, M Stanojević, A Tóth 1. Introduction Drawing on research findings from a study of the operations of German- and US-owned multinational companies in the automotive components sector in three central European countries, the paper has two main aims. The first is to examine the widely invoked, but loosely defined, claim that much foreign direct investment into central Europe in the manufacturing sector involves relocation from western Europe, and thereby to refine the conceptualisation of relocation. It is argued that in a sector where investment flows are primarily determined by efficiency considerations, the EU’s eastern enlargement has prompted the international reconfiguration of production, entailing complex multidirectional shifts in production across borders. Despite an apparent structural shift of manufacturing from the old to the new member states (Pilat et al. 2006) and some media alarm (e.g. ‘All roads in Europe are pointing East for carmakers’, Financial Times, 15/6/2006), relocation is a contingent, not a pre-determined, outcome. The second is to present and account for the range of responses from trade unions, in central Europe but also in the west, to this international reconfiguration of production. Appropriate trade union strategies, it is argued, can be focused at the local as well as at cross-national level, according to circumstances. The paper is organised as follows. A first section will outline the different forms of contingency that affect the nature of relocations and, subsequently, trade union responses. The following ones will examine these factors in detail through sector-level information on the automotive parts sub-sector and through company case studies: three sections will discuss, respectively, sub-sector-level, company-and plant-level, and actor-level contingencies. The conclusion will consider implications for the Europeanisation of industrial relations in multinational companies. 2. The contingency of relocations a) The complexity of relocation decisions Despite widespread fears of, and attention to, the potential employment consequences of relocation amongst the EU-15 the available evidence suggests that relocation in its strictest sense, when productive activity is directly transferred eastwards from a location in the old member states to one in the new, is relatively rare. Of the job losses arising from company restructurings across Europe documented by the European Monitoring Centre on Change, around 5 per cent are attributed to relocation over the period from 2003 up until the end of 2005. These involved over 70,000 redundancies in some 200 companies (Pedersini, 2005). Even if the EMCC, given its methodology based on press reports, probably greatly underestimates the volume of restructuring cases, the proportion of relocations is unlikely to be overestimated, as these are exactly the cases that attract most media interest. An assessment prepared by the European Commission (2006: 65) cites findings on the employment effects of relocation from selected studies relating to Germany, Austria and the Netherlands. Estimates of job loss between 1990 and 2001 directly attributable to relocation in these countries varied between 0.3 and 2% of total job losses. In addition to ‘direct’ relocation, there is also scope for ‘indirect’ relocation. This is where investment projects which might previously have been located in the EU-15 are instead located in one of the new member states. Because of the dynamic nature of investment decisions, and where to locate them, it is frequently less than clear cut whether an indirect relocation has indeed taken place; hence there are no reliable estimates of their possible magnitude. A further effect on employment comes from the threat of relocation, which frequently has the effect of inducing cost-reducing and/or efficiency-enhancing changes in existing operations in the EU-15 which themselves have employment consequences. Examples include the high profile cases of Siemens in Germany and Bosch in France, over the summer of 2004 (Pedersini, 2005). Again there are no reliable estimates of the scale of such ‘threat’ effects available. In order to understand both why actual relocations appear to be less widespread than feared, but also to uncover the circumstances under which relocations – direct and indirect – and the threat to relocate are more (or less) likely to be occur, it is productive to adopt a contingent approach. The approach is based around a number of distinctions, each of which has bearing on the scope for relocation, actual and threatened. The focus is relocation of activities across borders within multinational companies, and not with international outsourcing of production to independent suppliers in other countries (for elaboration of this distinction see Galgóczi et al, 2005). A first distinction is between market- and efficiency-oriented motivations for foreign direct investment (FDI), in which the potential for relocation largely arises under the second. An important part of inward investment into the new member states is ‘market seeking’ in nature, motivated by the opportunity of opening up new markets for products and services (European Commission, 2006: 64). Such investment by its nature is market expanding, and the largest share involves service activities which by their nature need to be produced close to the point of consumption. Scope for relocation is minimal and jobs may actually be created for some functions in western sites. In contrast, efficiencyoriented FDI is motivated by considerations of comparative labour costs and/or labour quality and/or labour productivity, in which companies are looking to take advantage of superior unit labour cost conditions in the new member states as compared to those prevailing in the EU-15. Given significant differences in labour costs, the presence of workforces with established skills and qualifications, and improving labour productivity amongst the new member states (Marginson and Meardi, 2006), the resulting scope for relocation is potentially extensive. Second, a crucial distinction has been drawn between a first wave of efficiency seeking FDI into the new member states, which involved outward processing activity based almost exclusively on the search for cheap labour, and a second wave, in which the emphasis is on labour quality, flexibility and productivity as well as costs (Fichter, 2003; Radosevic et al., 2003; Rojec and Stanojevič, 2001). This structural shift in the nature of efficiency-oriented inward investment into central eastern Europe, which has been underpinned by declining unit labour costs as productivity has risen faster than wages, has been widely documented (e.g Graziani, 2002; Ladó, 2001; Tholen and Hemmer, 2005). The Czech Republic and Hungary in particular have increasingly specialised in 2 medium-high technology manufacturing: the share of high and medium-high technology in total gross value added in these countries rose from 6% in 1994 to 10% in 2002, against an unchanged average of 8% in the EU (Pilat et al. 2006). Whereas the first wave essentially involved the transfer of labour-intensive, low skilled operations, the second wave of efficiency seeking investment involves the reconfiguration of production networks in Europe. This results in a more complex reorganisation of activities across different locations than the direct relocations associated with the first wave of efficiencyoriented FDI. As both Galgóczi et al (2005) and Pedersini (2005) recognise, it becomes more difficult to assess the employment impact, either overall or at specific locations. Moreover, given the role of human capital in second-wave activities, employer commitment to the workforce (including job security and scope for social dialogue and compromise) tends to be stronger. Third, the nature of these international production networks and the role of new and acquired sites in the new member states within the network is likely to have a bearing on the scope for relocation. Considering the nature of these networks, three possibilities can be identifed. The first two flow from the possibilities that the twin processes of market integration and enlargement open up for multinational producers to simultaneously segment and coordinate their operations on a pan-European scale. As a result numbers of MNCs are radically adjusting their ‘value chains’, in which different activities are segmented across borders, depending on skills and cost considerations (Kaplinsky, 2001). Such segmentation can vary in the extent, if any, of the vertical integration of sites within the international production network that it involves. This can range, first, from instances where sites solely supply other sites that are final assemblers of products (which we call segmentation) to, second, those where each site is itself the source of distinct products for the market, and sites compete for product mandates from the multinational parent (which is best defined as segregation.) Edwards and Zhang (2006) draw attention to an alternative logic for organising international production networks.This third possibility involves standardisation of production at a series of different locations that are comparable to each other. This is ofclear relevance in the service sectors where FDI is motivated by market access. But they show that it is also relevant in those parts of manufacturing under two conditions. The first is where products might be perishable and/or transportation costs high and/or major customers require just-in-time delivery or after-sales service, conditions which are consistent with a market access logic for FDI. The second is potentially relevant to efficiency-oriented FDI, and arises where production is technologically difficult to separate across borders. An example is large machinery manufacturing, where segmentation is not pursued. Relocation could occur if unit labour cost considerations prompted the establishment of a parallel operation in the new member states to source part of the enlarged European market. In other words, the similarity of sites across borders can lead to direct competition, relocation, and possibly vertical restructuring if a segregation strategy is followed; to coexistence, but under conditions of ongoing review of product mandates, if a segregation strategy is preferred; and to the establishment of new capacity, with potential future relocation consequences, is a strategy of standardisation is pursued. A fourth consideration is the extent to which location decisions of companies are autonomous, or are driven by those of major customers. In some sectors, such as automotive components, the location decisions of powerful customers who demand just- 3 in-time delivery effectively require suppliers to follow them to new locations (UNICE, 2005). In other sectors, companies have greater discretion over the choice of which locations to supply their customers from. Fifth, the potential for relocation is far from being bounded by the enlarged borders of the EU; relocation has a global as well as intra-European dimension in which some production activities (not all low skilled) are moving to Asia, North Africa or extra-EU European countries. The extent to which there is scope for extra-EU relocation would seem likely to vary between and within sectors, according to factors such as technology, skill requirements and the need for proximity to major customers. b) Implications for union responses Potential union responses to relocation have been distinguished, in an actor-centred way, by Erne (2004) along the twin axes of democratic/technocratic and European/national. In drawing on this framework an analytical – as against normative – approach to actors’ strategies must, as urged by Ramsay (1997), be sensitive to the contingencies affecting both axes, such as product market, role of labour costs and vertical integration. On the European-national axis, the nature of the product market (including the incidence of labour and transport costs, and the degree of production standardisation, segmentation or segregation) may make ‘European’ answers appropriate whenever the product market more or less coincides with the European space. By contrast, where Asian competition is stringent and relocation beyond Europe a distinct possibility, union interest in Europeanlevel action will be undermined. Alternatively when local production systems have a degree of autonomy, the strategies of local ‘competitive solidarity’ (Streeck 2000) could be most appealing. On the democratic/technocratic axis, contrary to Golden (1997) the choice to mobilise against job losses does not necessarily follow from a shortage of information, a misunderstanding of the situation, or specific political conditions. In multinationals, information on inter-site comparison may be available (e.g. through the EWC), and there is potential for union cross-border information and therefore ‘informed’ democratic strategies (events at Renault and GM Europe contain elements of this). The range of available options depends on the seriousness of competitive pressures on labour costs stemming from ‘coercive’ employer comparisons if they are weak, local development strategies may be preferred to European solutions, while if they are extreme, there are unlikely to be any market-viable solutions, and therefore political intervention remains the only safety net. Only in mid-way situations of serious but not extreme threats, and when the product market is European, will European answers be appropriate and viable. In such situations there is potential for EWC intervention, starting from informationsharing in order to assess the real nature of the threats. In addition, integrated production is more conducive of union strategic action; so too is a degree of similarity of business activities across sites in different countries (Marginson et al, 2004). An important implication is that, as argued by Huzzard et al., 2004, the technocratic and democratic approaches can be combined, at both the local and the European level. Table 1 shows the effects of each contingency on union responses to relocations. The five contingent factors are not mutually independent, though. The first, in particular, affects the second, third and fourth. The second in turn affects the third and fifth. However, all 4 should be considered in a multi-level analysis (sector, sub-sector, company). Higher level constraints limit the choices available at lower level, but do not obliterate them, as the variety detected in section 4 will illustrate. 3. The automotive component sector in the enlarged EU Through a comparison of the motor, apparel, and maritime industries, Anner et al (2006) highlight how union answers to international competition are sector-dependent. Here we add the importance of sub-sector and geographic contingencies: the specific product and labour-market contingencies on the one side, and the regional dimension on the other. In this regard, the automotive component sector in the new EU member states differs from both the motor industry as a whole, and from the ‘global’ and the former EU15markets. As a whole, the automotive sector is considered as among the ‘winners of globalisation’, due to its high fixed- and human-capital intensity. Its process of internationalisation has allowed a specific form of trade union cross-border co-operation, focusing on firm-level European (and in some cases global) works councils, which have coexisted with the maintenance of national union strategies in each country in the form of pacts for employment and competitiveness (Zagelmeyer 2000). Fixed capital means that labourcost driven relocation are not that a viable short-term option, and high human capital involves space for employee ‘voice’ and participation, as well as the possibility of defending high wages through high quality and productivity. As an effect, national-level compromises have remained distinct in spite of converging global pressures (Katz and Darbishire 2000.) The opening of global markets and the increase of foreign direct investment in this sector has reduced neither employment nor relative wages in the triad of core producers (Japan, USA and Western Europe) between the late 1970s and the late 1990s (Spatz and Nunnenkamp 2004.) In 2003, the opinion-making German magazine der Spiegel could still title ‘Autoindustrie: die Job-Maschine’ (Spiegel, no 37, 8/9/2003), referring to the major planned investments of Daimler, BMW, Porsche and Volkswagen within Germany. Such a rosy picture, however, hides important sub-sector and regional differences. Notably, the parts and components sub-sector (whose importance is increasing due to the parallel process of outsourcing and value chain fragmentation) reacts differently to globalisation than vehicle production. The degree of labour intensity differs: in Germany, the ratio of workers to sale is 2.5 higher in the production of autoparts than in that of automobiles and engines (Spatz and Nunnenkamp 2004.) Weight and therefore the transportability of the product also differ: because they are more easily transported, parts are more exposed to foreign competition. The implication for trade unions in traditional car-making countries is that in the component sub-sector it is more difficult to keep both high wages and employment. Liberal-market economies such as the USA have witnessed an increase in intra-industry wage differentials, while a country with centralised sector-level wage bargaining such as Germany has avoided this but could not prevent a decline in employment in the components sub-sector (ibidem). This divergent intra-industry process must be seen in the context of internationalisation. While in 1978/79 the German automotive industry imported only 6% of its inputs from low-income countries, by 1997/98 the figure was 30%, and increasing (ibidem). The main origin of these parts (over a half of electrical parts, and a rapidly increasing share of 5 engines) is Central Europe. Internationalisation of automotive production chains does not involve far countries (such as the Asian ones, where foreign direct investment in this sector is market-seeking) but proximate, lower unit-labour cost countries within the context of regionalisation: NAFTA and the enlarged EU. In the new EU member states, the automotive sector is the most important industrial sector for inward FDI, second only to financial services (which is mostly marketseeking.) Within this sector however there has been a clear transition at the end of the 1990s, with the progress of EU integration, at the same time as a ‘second wave’ of investment replaced labour-intensive activities with more human-capital intensive ones (Rhys 2004). In the 1990s, investment was mostly in car assembly using imported parts, for market-access reasons (most Central European countries still had tariffs on car imports.) In the 2000s, car parts and components rapidly gained importance and overcame car assembly. This is particularly visible in Poland as well as in the Czech Republic, where the government agency CzechInvest claims that ‘nearly every car produced in Europe contains Czech-made components’ (as quoted in Sperling 2004: 190). EU access, by eliminating the last non-tariff trade barriers, has made cross-border production reorganisation extremely easy, and 90% of components – as against 60% of finished cars – produced in the region are re-exported. Regional integration involving areas with much lower labour costs changes the employment effects of internationalisation. While classic theories of trade assume mutual advantages through specialisation, this holds true only between similarly advanced countries. When countries differ strongly in their factor (labour and capital) endowments, the adjustment pressure can be much greater. The analysis by Spatz and Nunnenkamp (2004) finds that the negative correlation between share of imports in production and the relative wage of low-skilled workers in Germany is stronger for automotive parts and components than for automobile assembly. Not all automotive workers are among the ‘winners’ of globalisation, then. After EU enlargement, and only a year after its abovementioned confident analysis, the same German magazine had to headline ‘Bye-bye “made in Germany”’ (Der Spiegel no 44, 25/10/2004.) Yet the view – still frequently met among western experts and employee representatives – of Central Europe as late industrialising (or even less developed) countries (and therefore comparable to Mexico within NAFTA) is misplaced. Post-communist countries were heavily industrialised and still have a large pool of skilled workers, even if technical education has become less popular after 1989. Indeed, former Czechoslovakia, Hungary and Slovenia are, in terms of employment, more industrialised than Germany, and their market share in total OECD manufacturing market has doubled in 1995-2003 (Pilat et al. 2006). Even if in communist-times industry was lagging technologically, the shortage economy paradoxically required, to compensate for the frequent lack of appropriate tools or components, particularly flexible skills on the part of employees. This has been quickly discovered by foreign investors. Volkswagen, which originally expected the newly-acquired Skoda to be a low-segment platform, found that Czech employees were even better qualified than its German ones and that Czech-made cars quickly ‘cannibalised’ its more upmarket western brands (Sperling 2004), while Fiat upgraded its Polish factories to ‘models’ of innovation for the Italian ones (Meardi and Tóth 2006.) Spatz and Nunnenkamp (2004) expectation that readjustment would concentrate high human capital production in the West applies only to some niches. The low productivity 6 levels in aggregate of the new EU member states’ economies are due to the enduring technological and infrastructure gap, and to the enduring weight of agriculture in some countries, rather than to human capital. In the internationalised sub-sectors where investment in technology has taken place, productivity has reached western levels very quickly, leading to widening gaps in unit labour costs (Dyker 2004). Given the competitiveness of human capital in CE, relocations are potentially more attractive in human capital intensive productions than in other – because the impact of labour costs is higher. The Czech Republic has quickly become the fourth most attractive country in the world for R&D projects in the automotive sectors (Czechinvest 2005), but Slovenia and Hungary follow close behind, with large companies such as Audi and General Electric as frontrunners. It is likely that extra-EU owned companies (especially American) are more prone to concentrate R&D in the new EU member states than west European ones, but in the medium-long term it remains that competition affects all employee groups in the sector. Such competition manifests itself in forms that evoke the image of ‘social dumping’. Large investors take their location decisions after tenders remaining ‘beauty tests’ giving the impression that labour costs and employer-friendly labour regulations are crucial. Local foreign investment agencies treat labour considerations as an important asset in open competition with the neighbours: the Slovaks stress low overall labour costs, and that the World Bank has awarded them the title of most flexible labour market in Europe (www.sario.sk); the Czechs highlight the low indirect labour costs, that ‘there is no history of large-scale strikes’, and that union membership has fallen from 2,292,000 to 610,000 (11% membership) between 1995 and 2005 (www.czechinvest.com); the Poles are proud of having the longest working hours in Europe (www.paiz.gov.pl); the Hungarians argue that they have the lowest earnings for highest productivity in the Central and Eastern European region (www.itdh.hu).1 The implications for European trade unions are twofold. First, given the lower impact of fixed and of transport costs, relocations are likely to be a greater threat for component factories than for assembly ones. Second, cross-border restructuring could be expected to lead to concentration of human-capital intensive production in western Europe, and of lower skilled production in the cheaper new EU member states. Keeping in mind Anner et al (2006)’s observation that in the automotive sector cross-border union co-operation takes place at firm level, these implications would expose employee representatives to a particular difficult situation . Not only would international competition be stronger, but those auto-component workers working in smaller firms (excluding ‘mega-suppliers’ such as Delphi) would lack the stronger firm-level representation channels (e.g. EWCs) of large car makers. Moreover, the geographic divide could overlap with professional divides making cross-border understanding more difficult. Fichter et al (2004), while studying German FDI in Hungary, had already noticed how international restructuring in some cases changed the nature of the German sites’ works councils: from blue-collar to white-collar dominated, leading to a changed union ethos and diminished mobilisation capacity. In manufacturing, it has not always been easy for trade unions to combine the representation of production workers, specialists, and white-collar employees. Indeed, in some countries these categories are organised in different trade unions, and there have been historical cases of direct opposition between the two (e.g. the Fiat strike in Italy in 1 All websites accessed on 24th July 2006. 7 1980 and the counter-mobilisation of clerical workers.) In an extreme scenario, western EWC white-collar reps would sit alongside eastern blue-collar counterparts: they would become less directly competitors, but also less able to understand each other on working conditions and employment prospects. The evidence from Central Europe to date – as the next section will show in detail – mitigates such a radical scenario, but does not provide a rosier one: competition is widespread even if not in just one direction or on one dimension. To summarise, the automotive component sector is particularly interesting for considering relocations for a number of reasons: it is regionalised rather than globalised (90% of production sold in the EU is produced in the EU); both labour skills and labour costs are important (lower capital intensity than car assembly), and human resources in the new EU member states are comparable to the western European, while much cheaper; foreign investment in this sector is purely efficiency-seeking (90% of production in Central Europe is re-exported), more so than final car assembly. This makes this subsector different from others, where investment is market-seeking (e.g. food, services) or only labour costs matter (e.g. apparel), and the enlarged Europe different from other regional blocs such as NAFTA, where human resources are less similar. It is a valuable test bed for the possibilities of negotiating and regulating international employee competition. 4. Case studies from the new frontiers of European autocomponent production As discussed in the previous sections, sector-level contingencies are integrated by company-level ones, relating to both product- and labour markets. This can be shown through the findings of research conducted from 2003 to 2005, involving a total of 12 plant case studies in the automotive component sector, four each in Hungary, Poland and Slovenia. In each country, the research design contemplated two German- and two USowned cases, but in Slovenia (where the number of investors is not large) two cases are actually hybrid, with a history of different investors from different regions. Interviews were conducted with management representatives (plant directors, production and human resource managers), employee representatives (either trade union or works council), trade union officials, local actors and, where possible, management representatives at the European headquarters and employee representatives on the European Works Councils. Of the five contingencies mentioned in section 2, the first (efficiency v marketorientation) is constant within the sector, and therefore is excluded from the analysis. The other four factors, together with the effects on relocation threats and industrial relations responses, all appear to be relevant, as summarised in Table 1. Plants are indicated by an acronym indicating home and host country (A: American, G: German, M: mixed, followed by H: Hungary, P: Poland, S: Slovenia.) Overall, across the case studies the examples of direct cost-driven relocations are rare: three among twelve companies, and of these one was from Poland to western Europe, for broader efficiency reasons than just labour costs. However, relocations as a threat and coercive comparisons are a daily reality in most plants. In the most integrated companies, it is usual for production teams to have open displays with their weekly production and 8 quality results compared to those of the other company plants.2 International competition is a direct experience for most employees, and not just an issue for the strategic level. The intensity of inter-site competition affects employees differently, though, depending on the role of labour costs in investment motivation. In none of the case studies labour costs are the only motivation: both employee representatives and, especially, managers stress the importance of skills, supply network reliability, public incentives and tradition (especially in the cases of brownfield investment). However, labour costs have played a primary role in three cases, and an important one – besides others – in another seven. In the former, relocation are an explicit issue in two cases (GH2 and, if on a minority of the production lines, AP2) and an indirect issue in the third (AS), where inter-site competition occurs in the form of contest for attracting future investment projects. This form appears to be overall more frequent, and although less urgently critical, it remains a constant consideration in industrial relations. Competition for the new investment exists also in the seven companies where labour costs are important, but beside other considerations such as quality and flexibility, while actual incidents of relocations are rare (and in one case, AP1, from the East to the West). At AH1 threats have materialised, first for blue collar workers and then also for white collar ones (finance, accounting, logistics and customer service.) In the other three cases relocations have occurred in the past (GH1, AH2) or are still currently threatened (GS). In the other two (AH2 and MS2) relocation threats are rather remote or very indirect. Finally, there are two cases (GP1 and MS1) where labour costs were a secondary consideration in comparison to quality, development and flexibility. As expected, here relocations are only remotely threatening. The second important issue is location autonomy. Most of the case studies come from rather autonomous producers of easily transportable components, and they have sufficient market power to have an own location strategy. Only for two of them location choice followed client choices (AP2 and GH2). As expected, relocations in these cases are decoupled from employment relations considerations: in one case they are not an issue, and in the other they are only within regional boundaries, in competition with neighbouring countries but not with western Europe. In both, however, later upgrading has increased the site autonomy. The situation can be obviously very different in the large number of smaller, often second- or third- tier suppliers, which are not covered by our research. The third important company-level contingency relates to the degree of vertical integration and standardisation. The process of outsourcing and fragmentation in the automotive industry has been particularly fast since the mid-1990s, and combined with the parallel process of internationalisation, causing strains on sector-level arrangements and allowing diversification of employment policies for different segments of the value chain, often located in different countries (Faust, Voskamp and Wittke 2004.) In case of vertical integration, there can be organisational integration or deep segmentation. This variable affects not only (in various way) relocation threats, but above all the industrial relations implications. Vertically and organisationally integrated or standardised companies relying on continuous transfer of semi-products, technology and 2 CE employees also frequently show their pride for their high position in such tables: for instance in Ger1 the Polish employees stressed how they outperformed their German colleagues on ‘Ordnung und Sauberheit’ (order and cleanness.) 9 information between different sites, indirectly encourage stricter standardisation of work organisation and a degree of industrial relations harmonisation. The continuous flow of information and contacts, and the requirement for highly standardised production imply that employment practices cannot be divergent. Research on the eastward extension of the EWC (Meardi 2004, Voß 2006) has shown that these are the cases where avoiding big cultural or employment conditions gaps between sites is in the employer interest, for the sake of joint projects and common understanding of productive issues. More generally, multinationals’ recent concern with business techniques such as program management, benchmarking, performance measurement may lead to more general mutual learning between stakeholders, and encourage ‘global strategies’ of ‘local players’ (Kristensen and Zeitlin 2005). The situation is different in plants working on specific products or semiproducts, often allocated to them on the basis of cost considerations, with little integration with other sites. Here, cross-border dialogue has less space to develop. The company with the strongest operational integration, GP1, is also the one with the strongest employee-side network of co-operation and information sharing. The establishment in the Polish location of industrial relations more similar to the German than to the Polish model is a welcome outcome for the corporation. Just as the company actively promotes the use of the German language, cross-border contacts between works councils and trade unions are seen as a tool to avoid conflicts and distortions of corporate projects. A broader socialisation process, including corporate culture, takes place, and even the EWC is instrumental in this goal. To a lesser extent this has occurred also in GH1 – and interestingly, only after the Hungarian subsidiary had been substantially upgraded. By contrast, plants such as GP2, AP1 and GS are entirely responsible for specific products, and while they are in competition with some other sites, cross-border co-operation is occasional, and mutual knowledge limited to little more than anecdotes, stereotypes. Although GP2 is covered by a long-established EWC, the 500-employee Polish plant in 2004 was still not represented in it, and the German works councillors did not see their integration as a relevant issue. A final factor affecting both the extent of relocation threats and the forms of employee responses is competition with extra-EU sites and – more indirectly – competing companies. Direct competition with China subsidiaries is most compelling in AH1, while GP2, GH2, AS feel the pressure from cheaper Eastern European locations: Romania, Ukraine, Croatia. The emergence of this problem is relatively new and combined with the already mentioned process of transition from ‘first’ to ‘second-wave’ FDI at the end of the 1990s. At GH2, for instance, relocation of part of the production to Romania in 2003 came as a shock and changed plant-level industrial relations that had been previously improving. In these cases, EU-level institutions such as the EWCs become largely redundant, because they do not correspond to the geography of the product market. While in the case of Romania, and to a lesser extent other European countries such as Turkey and Ukraine, an extension of the EWC is realistic, when competition is with Asia the difference between Europeanisation and globalisation of industrial relations (Müller, Platzer and Rüb 2004) becomes apparent, and Hungarian employee representatives at AH1 feel very distant from their western European counterparts. To conclude, the four company-level factors examined all exert some effects on the problem of relocations, determining both the nature of the threat to employees and – for the last two contingencies, i.e. vertical integration and extra-EU competition – the space 10 for employee responses. The latter will be examined in the next section, to assess if further, actor-level factors need to be considered. 5. Trade union responses While in aggregate terms the significance of relocation threats is more limited than in popular representations, in those situations where the threat - with its often dramatic social implications – is real the answers by trade unions are important for both the understanding of local outcomes and the possible further developments of union strategies as strategic actors. Removing from the analysis the three case studies where the relocation threat is currently only remote (AP1, AH2, MS), we are left with nine cases where the capacity for crossborder employee reactions and arrangements varies sharply. Four different scenarios are particularly clear. One extreme, if isolated, case, which may be labelled as ‘pro-active’ strategy, is detected in GP1: here there has been an early union engagement in internationalising companylevel industrial relations, through the European Works Council, a global works council, and a number of other international projects which translate into daily contacts through specific international teams on both works council and management sides. An important limitation of such strategy is that it does not cover the Chinese operations – but these are mostly market-seeking and therefore not in competition with European sites. Management, as already mentioned, has supported such efforts. The main strategy is an attempt at exporting the German model rather than defending it at home, and to enforce a spirit of ‘fair competition’ between sites. Through early consultation and information sharing, important investment in the NMS has not involved job cuts in the German sites – although these have come later under strong competitive pressure. According to the German works council, investment in the East has, through market expansion, createdthousands of new jobs for the German plants, although the net, direct employment effect is impossible to determine precisely. All investment decisions abroad have obtained preliminary approval from the German works council and have been discussed in the European Works Council. Such international procedures explicitly exclude, however, the discussion or negotiation of wages. Union engagement also took place in the case of the similar plant GH1, although in this case it was slower, and it was initially resisted by management. Company-specific factors that allowed consensual expansion of western solutions were missing, and therefore some preliminary ‘organising’ effort (as against a ‘social partnership’ approach) was required. Also, in this case a net negative employment effect for groups of German employees could not be avoided, and location decisions had not been subject to consultation. GH2 resembles this second scenario of co-operation through organising, even though in a more difficult market situation. An opposite, more defensive strategy from German unions is detected in GP2, another large German company, andprovides a third scenario. The Gesamtsbetriebsrat chair defends German arrangements rather than proposing their extension eastwards, arguing that ‘fifty years of union experience cannot be acquired in ten years’ (our interview.) A similar argument is used about production, to defend German employment: ‘in terms of productivity you can’t compare people who have 200 years of industrial experience with people who have nearly no industrial experience’ (ibidem). The German union has had 11 some success through local strategies: it has negotiated restructuring in several German sites and obtained (through a mix of resistance, mobilisation, expertise and political exchange involving local authorities) some social guarantees, but could not avoid continuous cost competition from Central Europe and cuts of hundreds of jobs directly due to relocations eastwards. The European Works Council is active but in a rather ‘reactive’ way, with no negotiation role. In 2004, it already covered some operations in central Europe (those more vertically integrated with the German ones), but not the Polish plant, of 500 employees, on the grounds that it was not in direct competition with German sites (although it is with Spanish, Italian, French and especially Turkish sister factories.) In spite of business similarities with the two previous case studies, the different degree of vertical integration and of geographic competition has meant that EUlevel cross-border employee co-operation has had much less space to develop. The company, in spite of its business being definitely global, has an adaptive approach in industrial relations, which in turn promotes local union strategies rather than coordination. The German views mentioned above are reflected by similar – and equally misinformed - views among Polish staff about their German counterparts. Defensive views in this case lead to a scenario of weak or missing cross-border links. AP2, a nonunionised factory where Polish workers deem trade unions to be ‘western luxuries’, can also be associated with this scenario. A fourth scenario is that of competition and diffidence. One of the case studies (AH1) includes experiences of tensions between western and eastern employee representatives, which were catalysed by the apparently technical issue of how to share seats in the enlarged EWC. Here, inter-site competition operates not only on employment and relocations, but also through ‘coercive comparisons’ on employment conditions, with for instance management trying to introduce ‘Polish practices’ in an Italian plant against local union resistance. Diffidence among western trade unions towards their eastern counterparts is even more apparent than at GP2, with questions raised as to whether ‘they are really unions’. Europeanisation through the EWC had worked well within western Europe, but is struggling to cope with the new geography of production, including the company decision to relocate their European headquarters from Spain to Hungary and the increasing direct competition from Chinese subsidiaries. We finally have to add the rather specific situation of the Slovenian plants (AS, GS and MS2), which does not fit with any of the above-mentioned scenarios. Here, trade unions are relatively strong, organisation is quite internationally integrated, but company-level cross-border links are still rather weak. The reason may be similar to the one that has discouraged Scandinavian unions from Europeanisation for some time (Andersen, 2005). Labour strength in Slovenia is rooted in the national industry and in national-level arrangements. Unlike in the other central European new EU member states, union density and influence is weaker in multinational companies than in domestic ones. The path towards Europeanisation may be then perceived as a risky undermining and departure from the traditional and safer national arrangements. The actor-level contingency which operates here is linked to national socio-political conditions. The findings are relevant for discussions on the Europeanisation of labour and of industrial relations. Our findings show how sector and company contingencies affect actors’ strategies on both axes suggested by Erne (2006): democratic/technocratic and European/national. Such contingencies explain the variety of union reactions even within 12 the same trade union in the same country (e.g. IG Metall), in spite of the apparently common institutional and ideological background (see GP1 and GP2). Unions in vertically integrated companies with product markets closely overlapping with EU boundaries will embrace Europeanisation as an appropriate response more readily than others. Also, direct inter-site cost competition and location autonomy will lend itself to an ‘organising’ approach (labelled as ‘democratic’ in Erne’s framework.) Other, more autonomous and development-oriented sites may tend to strategies of local ‘competitive solidarity’, including risks of concession bargaining. A problem with these situations is that such strategies tend to be self-enforcingly path-dependent and keep being followed even when their efficacy is declining, preventing the development of more courageous – but initially risky – lines, as seen in GP2. As to union strategies, the findings suggest that often much can still be done to increase cross-border information, and so far only integrated production networks have been conducive of union strategic action (e.g. GP1). Also, as argued by some recent comparative studies (Huzzard et al. 2004), technocratic and democratic approaches appear to be compatible, at both the local and the European level. Steps in this direction are detected at GH1, GH2 and AH2, where period of ‘organising’ and of ‘social partnership’ have alternated and have both been instrumental for union development (Meardi 2006). 6. Conclusion Instances of relocation are infrequent, even in a sector marked by considerable flows of efficiency seeking FDI into the EU’s new central European member states. Cases of indirect relocation (involving decisions on where to locate new investment projects)are more common. The threat of relocation is, however, a widespread phenomenon. A major reason accounting for the gap between threats and actual instances of relocation is that the threat to relocate has real effects, resulting in concessions at established sites which themselves diminish the potential benefits of an actual relocation for multinational producers. The nature of the threat and occurence of relocation emerges as a contingent one. The five contingencies examined proved influential: investment motivation (mostly sectorlevel), labour costs influence (sub-sector and company levels), location autonomy (company level), vertical integration (company level) and extra-EU competition (subsector and company levels.) In particular, the latter two affect the potential for crossborder union responses. As relocation threats are contingent and actual instances are mostly indirect, time is available during which unions can elaborate responses. The implication of our analysis is that there is no ‘one best response’, but rather a need for evaluating each situation. In particular, a ‘European’ response should not be seen as necessarily the best: depending on the circumstances it may be unviable, suboptimal or even inappropriate. The paper has illustrated the different situations through case study examples. Being plant-level studies from a non-representative sample in one sub-sector, they should be treated as indicators of a minimum degree of variation, rather than an exhaustive representation. The findings demonstrate that the scope for local actors’ ideologies and strategies is not obliterated, but must be located in the contingent constraints they face in order to avoid misplaced prescriptions. Notably, actors’ ideology (and history) still plays a role in 13 explaining the higher propensity for technocratic answers in the East than in the West: trade unionists in the new EU member states are generally better educated than their western counterparts, and much less prone to call a strike. Also, union strategy is detectable in some ‘offensive’ (as against reactive) use of the EWC (GP1), as well as in local strategies of upgrading (AH1). The importance of indirect competition has not led to open, structural competition between eastern and western sites, but to a more variable geography of situations where the competing sides may be shifting. Cross-border information sharing on the union side is still not common, but in those companies where it takes place, there are margins for ‘negotiated competition’ (Kädtler and Sperling 2002). East-west union co-operation within multinationals is still at a very early stage of development and faces serious structural constrains. Even in famous cases of crossborder mobilisation such as Renault (1997) and General Motors (2005, 2006) the participation of sites from Central Europe was limited by regulations (no right to solidarity strikes), implicit competition (the eastern sites were objectively benefiting, at least in the short term, from restructuring in the west), and insufficient interests (the Polish union leaders at Opel acknowledged that they would never win a strike ballot in solidarity with German colleagues.) However, open conflicts have been avoided, and mutual understanding and support have been growing. In the cases of competition, EWCs’ most common strategy has been that of avoiding forcing the situation through majority voting, in the hope that time will narrow the distance between the different sides (Carley and Hall 2006). In the meanwhile, the scenarios popular in the media of a radical ‘new international division of labour’ within Europe or of ‘wild competition’ and ‘social dumping’ are hardly borne out by the paper’s findings. The evidence is, so far, extremely varied, and at least in sectors where fixed costs are important a number of both local and international strategies are available to harmonise the pace of economic restructuring with that of social acceptance. References Andersen, S. K. (2005) EJIR – to add Anner, M., Greer, I., Hauptmeier, M., Lillie, N. and Winchester, N. 2006: The Industrial Determinants of Transnational Solidarity: Global Interunion Politics in Three Sectors. European Journal of Industrial Relations 12, 1, 7-27. Carley, M. and Hall, M. 2006: European Works Councils and Transnational Restructuring. Dublin: Report for the European Foundation for the Improvement of Living and Working Conditions. Czechinvest 2005: Automotive Industry in the Czech Republic. Prague. Dyker, D. 2004: Closing the productivity gap between eastern and western Europe: the role of foreign direct investment. Science and Public Policy 31, 4, 279-287. Edwards, T. and Zhanng, M. 2006: Multinationals and national systems of employment relations: innovators or adapters? PM has details somewhere … Erne, R. 2006: European trade-union strategies: Between technocratic efficiency and democratic legitimacy. In S. Smismans (ed) Civil Society and Legitimate European Governance. Cheltenham: Edward Elgar (forthcoming). 14 European Commission 2006: Enlargement, two years after: an economic evaluation. Occasional Paper No. 24 (May) Brussels: Directorate-General for Economic and Social Affairs, Commission of the European Communities. Faust, M. Voskamp, U. and Wittke, V. 2004: Globalization and the Future of National Systems: Exploring Patterns of Industrial Reorganization and Relocation in an Enlarged Europe. In M. Faust, U. Voskamp and V. Wittke (eds) European Industrial Restructuring in a Global Economy: Fragmentation and Relocation of Value Chains. Göttingen: SOFI, pp. 19-81. Fichter, M. 2003: Internationalisation of production: options and responses. AICGS/Daimler-Chrysler Working Paper Series (John Hopkins University). Fichter, M., Dörrenbächer, C., Neumann, L. and Tóth, A. 2004: Internationalization Of Production: Options And Responses. Evidence From German Enterprises In Hungary. Paper presented at the IRRA 53rd Meeting. San Diego, 3-5 January. Galgóczi, B., Keune, M. and Watt, A. 2005: Relocation: challenges for Trade Unions. Discussion Paper, DP 2005.01. Brussels: ETUI-REHS. Golden, M. 1997: Heroic Defeats: The Politics of Job Loss. Cambridge: Cambridge University Press. Graziani, G. 2002: Product Quality Upgrading in Central-Eastern European Countries and the Coming EU Enlargement. In S. Baldone, F. Sdogati and L. Tajoli (eds), EU Enlargement to the CEECs: Trade Competition, Delocalisation of Production, and Effects on the Economies of the Union. Milan: FrancoAngeli, 53-72. Huzzard, D., Gregory, D. and Scott, R. (eds) 2004: Strategic Unionism and Partnership: Boxing and Dancing? Oxford: Oxford University Press. Kädtler J. and Sperling H.J. 2002: After Globalisation and Financialisation: Logics of Bargaining in the German Automotive Industry. Competition & Change, 6, 2, 149-68. Kaplinsky, R. 2001: The Value of Value Chains: Spreading the Gains from Globalisation. IDS Working Paper 110. Institute of Development Studies, University of Sussex, Brighton. Katz, H.C. and O. Darbishire 2000: Converging Divergences: Worldwide Changes in Employment Systems. Cornell: Cornell University Press. Kristensen, P. and Zeitlin, J. 2005: Local Players in Global Games: The Strategic Constitution of a Multinational Corporation. Oxford: Oxford University Press. Ladó, M. (2001), ‘Hungary: FDI and Its Impact on Industrial Relations’ in G. Gradev (ed), CEE Countries in EU Companies’ Strategies of Industrial Restructuring and Relocation. Brussels: ETUI, 73-135. Marginson, P., Hall, M., Hoffman, A. and Müller, T. 2004: The Impact of European Works Councils on Management Decision Making in US- and UK-based Multinationals. British Journal of Industrial Relations 42, 2, 209-234. Marginson, P. and Meardi, G. 2006: EU Enlargement & the FDI Channel of Industrial Relations Transfer. Industrial Relations Journal 37, 2, 92-110. Meardi, G. 2004: Short Circuits in Multinational Plants. The Extension of European Works Councils to Poland, European Journal of Industrial Relations 10, 2, 161-78. Meardi, G. 2006: Multinationals in the New EU Member States and the Revitalisation of Trade Unions. Paper presented at the Conference on Trade Unions in Central eastern Europe and Russia. Hatfield, 27-28 March. 15 Meardi, G. and A. Tóth 2006: Who’s Hybridising What? Insights on Multinationals’ Employment Practices in Central Europe. In A. Ferner, X. Quintamilla and C. SánchezRundes (eds) Multinationals and the construction of transnational practices: Convergence and Diversity in the Global Economy. London: Palgrave, pp. 155-183. Müller, T., Platzer, H-W and Rüb, S. 2004: Globale Arbeitsbeziehungen in globalen Konzernen? Wiesbaden: VS Verlag. Pedersini, R. 2005: Relocation of production and industrial relations. European Industrial Relations Observatory online Ref: TN0511101S. Pilat, D., Cimper, A., Olsen, K. and Webb, C. 2006: The Changing Nature of Manufacturing in OECD Economies. STI Working Paper 9, OECD. Radosevic, S., U. Varblane and T. Mickiewicz 2003: Foreign direct investment and its effect on employment in central Europe, Transnational Corporations, 12, 1, 53-90. Ramsay, H. 1997: Solidarity at Last? International Trade Unionism Approaching the Millenium. Economic and Industrial Democracy, 18, 4, 503-37. Rhys, D.G 2004: The Motor Industry in an Enlarged EU. The World Economy 27, 6, 877900 Rojec, M. and M. Stanojević 2001: Slovenia: Factor-cost-seeking FDI and Manufacturing. In G. Gradev (ed), CEE Countries in EU Companies’ Strategies of Industrial Restructuring and Relocation. Brussels: ETUI, . Spatz, J. and Nunnenkamp, P. 2004: Globalization of the Automobile Industry – Traditional Locations under Pressure? In M. Faust, U. Voskamp and V. Wittke (eds) European Industrial Restructuring in a Global Economy: Fragmentation and Relocation of Value Chains. Göttingen: SOFI, 105-29. Sperling, H-J. 2004: Going East – a Volkswagen Version of Globalization. In M. Faust, U. Voskamp and V. Wittke (eds) European Industrial Restructuring in a Global Economy: Fragmentation and Relocation of Value Chains. Göttingen: SOFI, 181-200. Streeck, W. 2000: Competitive Solidarity: Rethinking the "European Social Model". In K. Hinrichs, H. Kitschelt and H. Wiesenthal (eds) Kontingenz und Krise: Institutionenpolitik in kapitalistischen und postsozialistischen Gesellschaften. Frankfurt am Main: Campus, 245-61. Tholen, J. and Hemmer, E. 2005: The Effects of Direct Investments by German Companies in Central and Eastern Europe: Scopes, Motives, Strategies, Jobs. IAW Forschungsbericht 8, April, IAW, University of Bremen. UNICE 2005: Relocation: Challenge and Opportunity. Brussels: UNICE. Voß, E. 2006: Overstreched: European Works Councils Experience in New EU Member States. Dublin: Report for the European Foundation for the Improvement of Living and Working Conditions. Zagelmeyer, S. 2000: Brothers in Arms in the European Car Wars: Management-Labour Pacts in the Context of Regime Competition. MPIfG Working Paper 00/2. 16 Table 1 Contingent factor FDI motivation (efficiency v market) Human resources consideration (skills v cost priority) International configuration of production (standardisation v segmentation/segregation) Location decision (autonomous v dictated by customers) Extension of product market (EU v global) Level Sector Implications for union responses to relocations Seriousness of threat Sector, sub-sector Potential for social compromise, dialogue Sub-sector, company Degree of competition and of international integration/socialization Level of response (company or sector/political) Suitability of European response Sub-sector Sector, sub-sector 17 Indirect Cross-border union responses Relocation threats Relocation incidents Exposure to extra-EU competition Integration Location autonomy Plant Primary competitive advantage Table 2 – Case study summary: company contingencies and EWC effects GP1 Poland Germanowned Quality Yes Very strong (standardisation) No GP2 Poland Germanowned Cost Quality Yes Weak (segregation) Yes: Turkey, Eastern Europe West-toEast Direct & indirect AP1 Poland US-owned AP2 Poland US-owned Cost Quality Flexibility Cost Yes Weak (segregation) Yes East-toWest Remote indirect Partial Medium (standardisation) Yes Limited West-toEast (on few lines) Indirect None, no union Scenario: national defence GH1 Hungary Germanowned GH2 Hungary Germanowned AH1 Hungary US-owned Cost Quality R&D Flexibility Cost Yes Very strong (standardisation) No Strong direct & indirect Partial Medium (standardisation) Yes: Eastern Europe Direct Strong Local organising Scenario: Euro-local organising Medium Scenario: Euro-local organising Cost Quality Flexibility Yes Very weak (segregation) Yes: China West-toEast Direct & indirect AH2 Hungary US-owned Cost Quality Yes Very weak (segregation) No West-toEast (one incident) Remote GS Slovenia Germanowned AS Slovenia US-owned MS1 Slovenia Domestic (formerly German) MS2 Slovenia UK-owned (formerly Austrian) Cost Quality Flexibility Yes Weak (standardisation) Yes Direct Weak Scenario: Euro-scepticism Cost Yes Medium (standardisation) Indirect Quality Yes Medium (standardisation) Yes: Eastern Europe No Weak Symbolic EWC Scenario: Euro-scepticism Weak Local strategies Scenario: indifference Cost Quality R&D Flexibility Yes Medium (segmentation) Yes Indirect Remote Very strong EWC + information networks ‘Expansion’ of homecountry compromises ‘Fair’ competition and ‘negotiated’ globalisation Scenario: pro-active None Plant excluded from EWC Defensive ‘competitive’ strategies in western sites Scenario: national defence None Scenario: indifference Medium Some conflicts, diffidence within the EWC Relations with US union Scenario: competition None Scenario: indifference None Scenario: Euro-scepticism 18 19