modular production and the new division of labour ...

02-098930-Frigant:EUR-16(1)-098930-Frigant.qxd

10/16/2008

16(1): 11–25 10.1177/0969776408098930 Copyright © 2009 SAGE Publications Los Angeles, London, New Delhi and Singapore http://eur.sagepub.com

9:32 PM

Page 11

European Urban and Regiona l Studies

MODULAR PRODUCTION AND THE NEW DIVISION OF LABOUR WITHIN EUROPE: THE PERSPECTIVE OF FRENCH AUTOMOTIVE PARTS SUPPLIERS

★ Vincent Frigant and Jean-Bernard Layan

ON LY

GERPISA and University of Bordeaux, France

Abstract

PR OO F

This article focuses on the emergence of a new international division of labour in the auto parts industry. Its first section examines the hypothesis that the shift to modular production offers a chance to modify value chain geography inasmuch as modularity causes new opportunities and constraints in geographic proximity terms. An analytical matrix is provided and applied to New Accession Countries, with special consideration being given to French suppliers’ circumstances due to the requirement that host country characteristics and company specificities be analysed

Introduction

For nearly a decade now, empirical analyses have highlighted the rapid internationalization of autoparts suppliers, ostensibly pursuing three different types of logic (Sadler, 1998; Chanaron, 1999; Frigant, forthcoming): follow sourcing, where suppliers match customers’ internationalization moves; diversification, where they try to expand their customer portfolio by moving into carmakers’ home countries; and delocalization, where they move into low-cost countries to cut production costs. The simultaneity of these trends explains how major suppliers have been able to weave globalized productive networks that operate on a worldwide basis, and why they have set up production and R&D units in an increasing number of countries.1 At the same time, this internationalization drive is hiding the general transformation of the supplier industry’s economic geography. Alongside ‘traditional‘ internationalization issues (like the search for lower production costs) other issues have arisen, relating to the break-up and recomposition of

simultaneously. The second section tests this matrix using statistical data and culminates in a case-study. It will be demonstrated that New Accession Countries are being integrated with the rest of the Continent, due to firms’ ongoing search for location-related advantages and because of a tightening/easing interaction that is associated with proximity constraints.

KEY WORDS ★ automotive, division of labour ★ modular production ★ New Accession Countries ★ suppliers

the clusters at work in this process (Carrillo, 2004; Vale, 2004; Pavlinek and Janak, 2007). To understand this trend, we believe that instead of studies mobilizing traditional international economics it would be better to identify the spatial scales that are relevant to the value chain’s functioning. The need for this kind of approach becomes all the more apparent when we focus on the role played by emerging spaces, countries whose status in suppliers’ eyes can only be understood in the light of the central spaces where the suppliers originate, and in terms of their relationships with domestic customers (Humphrey, 2000; Carrillo and Contreras, 2007; Domanski et al., 2006; Chanaron, 1999; Sadler, 1998). The present article is an attempt to understand the role that the twelve New Accession Countries (NACs) fulfil for French autoparts suppliers (see Note 4 for the detail of countries studied). We wish to explore the new international division of labour that French suppliers implement between East and West in a context of modularization of car industry.


9:32 PM

EUROPEAN URBAN AND REGIONAL STUDIES

Modularization’s impact on the division of labour

Since the 1990s, modularization has been a key concern in the automobile industry. As a process intended to enhance the management of systemic complexity, modularity seems imbued with several virtues: reduction in design delays and costs; mass customization; extended division of labour; and externalization (Sako, 2003; Takeishi and Fujimoto, 2003; Volpato, 2004). This article will focus on this latter point.

Suppliers’ rising power: an opportunity to modify the automobile industry’s geography Since the mid-1980s, carmakers have being pursuing strategies based on increased outsourcing – an objective they have progressively attained, given current estimations that suppliers are responsible for European Urban and Regional Studies 2009 16(1)

Page 12

16(1)

75 percent of an automobile’s production price. This orientation is the product of a common strategy aimed at reducing the number of First Tier Suppliers (FTSs). The idea has been to structure the supply chain around a narrow hierarchy comprised of a few highly specialized suppliers in order to: (a) cut transaction costs by diminishing the number of bilateral relationships in need of management resources; (b) reduce design delays and costs by calling on suppliers capable of making the heavy investments required; (c) improve the quality of the components being delivered via Design-forManufacturing; and (d) cut production costs through a repercussion of the economies of scale and scope achieved by suppliers working on behalf of several carmakers (in dynamic terms, stemming from the economies of learning they have achieved by specializing in specific components). This kind of organizational choice is cumulative in nature and conducive to an increase in suppliers’ size and to their specialization in certain components. By retroaction, this encourages them to go further down a path of externalization and narrow hierarchy. At the same time, a dynamic of this sort suffers from threshold effects resulting from components’ technical breakdown. A further narrowing of the suppliers’ pyramid intimates a recomposition of tasks around a tandem that might be represented as ‘large components/broader supplier missions’. An initial stage launched in the late 1980s consisted of no longer externalizing simple components but entire functions (like the braking or exhaust systems). This helped to concentrate supply chains, but to further tighten the pyramid, the industry had to embark upon a new stage. Modular production fulfilled this role from the late 1990s onwards. It is true that compared to its theoretical principles (Ulrich, 1995) the move towards modularity was relatively disappointing. The automobile remains a profoundly systemic product that suffers from decomposability problems through and through (Sako, 2003). More than pure product modularity, the definition of modules that has tended to prevail centres on plurifunctional, physically compact blocs, like front-end modules, cockpits and rear-end modules. Defined thus, modules look more like macrocomponents (Volpato, 2004) than modules in the

ON LY

If we hypothesize that a new geography is currently deploying, we must assume that a new organization of interfirm relationships has emerged. After all, suppliers are by definition dependent on carmakers and we can only grasp their localization choices by understanding the relationships between these two actors. The hypothesis explored in this article is that autoparts suppliers’ localization behaviour has changed because transformations in the organization of vertical relationships have created new opportunities and constraints which justify such spatial reorganization. The idea here is that NACs’ role should be analysed as a peculiar implementation within this specific space of organizational transformations driven by the modularization of the car industry. The first section will detail our working hypothesis and show that modular production’s advance is an opportunity to transform firms’ geography. It offers a schema to explain the new forms of division of labour deploying as a result of modularization. The second section will assess this schema for NACs in an attempt to discern the role that countries in this part of the world fulfil in French suppliers’ eyes.

PR OO F

12

10/16/2008


10/16/2008

9:32 PM

Page 13

FRIGANT & LAYAN: MODULAR PRODUCTION AND THE NEW DIVISION

sites, opened 29 new sites, sold 26 locations and acquired 13 units. FTSs have been moving, and very quickly – but in pursuit of which logic? And towards what end?

The approach adopted: identification of proximity needs One major consequence of this reorganization is a considerable intensification in interfirm interactions (Lung, 2001). The modules’ very nature has forced FTSs to intensify their interactions with carmakers and with other FTSs to ensure compatibility. This is reminiscent of the lessons learnt from studies about Complex Products Systems: the modularization of strongly systemic products implies a reinforcement of interfirm knowledge exchanges (Prencipe et al., 2003; Hobday et al., 2005). It remains that knowledge exchanges are not the only ones to have intensified. The automobile is a mass production involving few assembly plants. Its supply chain is organized according to Just-In-Time principles. Delivering macro-components implies organizing a high-performance logistics chain that will be all the more complicated to set up due to the fact that modules are difficult to transport and feature many internal components. This too leads to leaner flows (but productive ones this time). This leaner flow system – in both cognitive and productive terms – raises proximity constraint questions about FTS location (Kirat and Lung, 1999; Rallet and Torre, 2005). Our hypothesis is that FTSs have put together a network of (production and research) entities in a way that respects the supply chain’s coherency. To understand the architecture of this network, we need to identify and associate the different forms of proximity that FTSs require for their various activities, starting with their interactions with carmakers. An approach of this sort seems appropriate due to the suppliers’ dependence on one (or several) carmaker(s). Theirs is a localized market, featuring the plant(s) to be delivered and the centre(s) being run by the carmaker on whose behalf the module is to be developed. Of course, the reality is even more complex, given that suppliers undertake several activities at once and have often achieved a high level of vertical integration. We need to look beyond

PR OO F

ON LY

theoretical sense of the term. However, they do constitute a powerful vector for externalization, even as they encourage a smaller number of suppliers (hence their concentration). The supplier industry consolidated considerably through internal growth – and even further through external growth, in the wake of the vast mergers and acquisitions trend that the sector experienced after 1995. The challenge for suppliers has been threefold: to increase their size and generate economies of scale; to diversify their customer portfolio; and to acquire complementary competencies, because modules are plurifunctional and mobilize dissimilar knowledge bases. In a context in which the market for modules is expected to grow rapidly,2 what we are now witnessing is a race towards modularity. There is a widespread belief, at least among Europeans and North American suppliers (Takeishi and Fujimoto, 2003), that they must become ‘first movers’ if they want to have any chance of forcing carmakers to adopt their own module definitions at a time when modular decomposition is not yet set in stone. At the industry’s current stage of restructuring, the main FTSs are the heart of the value chain, to the point that they are occasionally called 0.5 tier suppliers (Lung, 2001). Carmakers can no longer design certain modules by themselves (Autobusiness, 2004). This rising power has been accompanied by a modification in suppliers’ perimeters. If we consider size as an indicator, cumulative sales for the 30 large autoparts suppliers rose 73.2 percent between 1998 and 2005. Average sales were $6,381.9m in 1998 compared to $11,055.5m in 2005, when four FTSs each had revenues of more than $20b (Frigant, forthcoming). Alongside this, there has been a spectacular rise in suppliers’ production and research capacities. Becoming a ‘modular supplier’ implies completely designing modules as well as their constituent components. In the same vein, suppliers have had to build new production units, hence think about their localization. Moreover, the extension of their tasks (new missions) suggests that they have had to rethink their production sites’ spatial organization (site specialization, closures, unit linkages, etc.). This too has been an opportunity to change their economic geography. Taking one example, between 2001 and 2006, the French supplier Valeo closed 59

European Urban and Regional Studies 2009 16(1)

13


9:32 PM


Step 1: Breakdown of a modular supplier’s output

A modular supplier’s production process can be broken down into three phases. First, there is the module itself; that is, the macro-component whose final version will be delivered to the carmaker (seat, cockpit, etc.). At the other extreme, we find components that are used in the modules’ own manufacturing (i.e. seat padding). In between, and with relatively significant variations depending on the items in question, there might be a module preproduction stage. This could involve a preassembly of simple components (i.e. a seat’s metal frame). The idea here is to produce what we will call meso-components. These three stages of production differ according to two criteria: exclusivity and complexity. Macrocomponents are exclusive when dedicated to a specific vehicle model – and where they amount to a visible element of product differentiation. Components, however, are hidden elements (Baldwin and Clark, 2000). To generate economies of substitution (Garud and Kumaraswamy, 1995) and scale, a supplier will try to design reusable components on different modules. Mesocomponents constitute intermediary elements insofar as they involve designing basic architectures European Urban and Regional Studies 2009 16(1)

Page 14

16(1)

to be shared among several cars and even carmakers. These architectures will be specified at a later date as per carmakers’ specifications. For instance, the French supplier Faurecia has developed what it calls ‘platform components’. Faurecia is stepping up its ’component platform policy’, targeting maximum standardization for nonvisible vehicle parts. Product plans and market offerings are structured accordingly, and standards defined for both components and assembly processes. One example of this approach is a seat frame for the worldwide General Motors platform, and a common cockpit structure across different Ford Group models (Faurecia, 2006a: 20) These three phases also contrast in terms of complexity. Macro-components result from the assembly of multiple parts and components mobilizing different kinds of knowledge (e.g. with cockpits, this can involve electronics, electrics and plastics) whereas components tend to be elementary constituents mobilizing a lesser amount of differentiated knowledge. Meso-components occupy something of an intermediary position.

ON LY

activities based on a direct contact with carmakers and go further upstream to try and reconstitute the whole of the value chain. As exemplified by the large number of entities that FTSs possess, they have to organize also an intrafirm division of labour. We will apply an approach first proposed by Bordenave and Lung (1996), one that breaks the automobile production process down analytically to apprehend the technico-economic constraints weighing upon activities to deduce their ultimate localization. This approach must be adapted to our object of study, however: the modular-supplier production process. We will do this in two stages: by proposing a typology of the activities undertaken by modular suppliers; and by asking whether these activities must be located in proximity to a given carmaker. We will also distinguish between R&D and production since proximity constraints differ for each of these two aspects.

PR OO F

14

10/16/2008

Step 2a – the duality of R&D site locations A given macro-component will be exclusive to a specific car model. Modules’ general architecture and interfacing rules must be designed in close collaboration with carmakers and the other FTSs manufacturing the complementary modules (Sako, 2003). Collective engineering teams are usually organized since, in spite of ICT advances, company staff members will need to interact physically to finalize the modules. Components and meso-components do not require the same types of interactions. In line with modularity’s ‘traditional’ properties, the buyer does not need to be familiar with the internal components’ exact specifications (Baldwin and Clark, 2000). The supplier tries to design generic components that can be introduced into modules that will be sold to several customers. There is a disconnect between carmakers and R&D activities relating to the modules’ detailed architecture (the components’ design and articulation). In short, R&D activities are dual in nature. On the one hand, interactions with carmakers have


10/16/2008

9:32 PM

Page 15


Step 2b –Three types of production entities

China ran into sourcing difficulties (Belzowski et al., 2006). This explains the preference for peripheral spaces. Mexico, for example, is a major location for the production of components on behalf of suppliers trying to service the North American market (Carrillo and Contreras, 2007). Meso-component production plants’ location is more ambiguous due to their intermediary position in terms of complexity and exclusivity. Production sites will be localized by arbitraging transport constraints and economies of scale. To surmount the former, it is tempting to increase the number of production sites and increase their proximity to the macro-component assembly plants. Yet suppliers also want to concentrate their sites to achieve economies of scale. The efficient solution seems to be an intermediary one involving a barycentric location at the centre of the modules’ final assembly site network. The number and location of units will vary from one host region to the next and depending on the type of modules being manufactured. The equation becomes all the more complicated if we consider that suppliers try to encourage user/producer interactions for the most complex meso-components (Lundvall, 1988). Certain sites will be located in the vicinity of R&D units specializing in this type of meso-component (cf. Carrincazeaux et al., 2001). This typology of FTS activity locations, starting with the need for proximity to customers and reversing back up the supply chain, offers an analytical matrix that should help us to better understand suppliers’ location strategies. To avoid excessive determinism, however, we need to take two precautionary measures before implementing this matrix. First, the matrix outlines orientations for suppliers’ localization choices. Depending on their past trajectory (notably the acquisitions they have made), suppliers will inherit a productive apparatus that has not necessarily been designed for their new purposes. This raises the question of sunk costs, which act as an obstacle to site relocation and specialization. Such spatio-organizational schemas should be viewed as ideal-types that are necessarily incomplete and in an evolutionary phase, if only because of customer portfolios’ instability. In addition, consideration should also be given to the home and host spaces. Two suppliers can view one and the same space in a different light. For example, Italy’s Piedmont region constitutes a central space

ON LY

increased with regard to modules’ specification and integration into overall vehicular architecture. On the other hand, there are fewer interactions nowadays defining a detailed architecture for the module and its internal components. Today’s FTSs are capable of ‘autonomous R&D’. They create R&D centres that can have a global vocation and specialize by module or functionality (Carrillo, 2004,Fourcade and Midler, 2005;). Moreover, they are relatively free to locate where they like. Centre engineers can be seconded to carmakers to finalise the module, working as part of a team project and/or in technical centres established in carmakers’ immediate vicinity.

PR OO F

At a productive level, the main issue is the circulation of component flows. Just-in-time (JIT) and In-line-sequence (ILS) logics now dominate the industry’s organization. Initial analyses of JIT geography may have affirmed the need for spatial proximity (Estall, 1985) but subsequent studies have shown that improvements in infrastructure and logistics, along with organizational learning, have led to relative easing in these constraints (Lung and Mair, 1993; Sadler, 1994). Different problems arise in a modular context. Given macro-components’ complexity (fragility, volume and weight) they can be costly to transport and will often be delivered directly onto production lines assembling different kinds of vehicle models. This means that deliveries will require a very delicate sequencing (exclusivity). Assembled modules suffer from a strong proximity constraint. The typical spatial organization that has arisen in recent years revolves around supplier parks where modules’ final assembly occurs (Frigant and Lung, 2002; Larsson, 2002). Components are not included in this kind of proximity requirement. Suppliers try to create specialized sites, seeking economies of scale and minimizing complexity or exclusivity. Since these components are easier to transport, there is less of a proximity constraint. Moving to a low-cost country may seem economically attractive because these kinds of activities are often labour-intensive and require a relatively unskilled workforce. But clearly, there are limits to feasible distances: American suppliers who delocalized activities of this sort to


15


16

10/16/2008

9:32 PM


Page 16

16(1)

Table 1 Does a specific activity have to be located near a particular carmaker?

Task Design

YES Technical centres or temporary proximity YES Supplier park

Meso-components (general definition of modules/ pre-assembly)

Barycentre (trade-off between economies of scale/transport constraints/proximity to R&D centres)

PR OO F

for a French supplier but an exotic one for a Japanese supplier. Having Fiat as a customer for a given module might mean each of these two suppliers making a totally different localization choice. In other words, our typology has to be specified for relatively specific ‘supplier-space’ tandems. It should also help us to answer the specific question we have asked: What role does a given space (like NACs) play for a specific party (like a French auto parts supplier)?

Specifying the matrix for French suppliers in NACs

Since our approach asks questions about suppliers’ relationships with carmakers, we should start out by specifying the positions they have acquired in NACs. Since the fall of the Berlin Wall, carmakers have been expanding their production capacities in NACs, converting Communist-era plants and creating greenfield ones (Lung, 2003). Today NACs produce nearly 3m vehicles, more than the Iberian Peninsula and France. Central and Eastern Europe’s (CEE) rise has not led to any significant fall in the capacities in Western Europe, where very few carmakers’ sites have in fact been closed down. However, it is an opportunity to redistribute the different types of production. NACs mainly specialize in the assembly of high-volume low-end vehicles and engines – although they also host the manufacturing of a number of high-end low-volume niche products (Layan, 2006). The Volkswagen Group has been a trailblazer in this movement. In 1991 it took over Skoda, whose European Urban and Regional Studies 2009 16(1)

Components

NO – Large R&D centres

ON LY

Production

Modules – Macro-components (specification/final assembly)

NO Component plants

capacities it expanded considerably in the Czech Republic (specifically at Mladá Boleslav) and in Slovakia (Bratislava). VW also built two Audi plants in Hungary in 1993. French groups have been more cautious, one example being PSA Peugeot–Citroën, which chose other emerging countries for its growth plans. It is true that in 1992 Renault took over the Slovenian joint venture Revoz, which it integrated into its network of West European assembly sites. It also acquired Dacia in 1999 – but this Romanian brand was kept on the edges of its Group strategy. Everything changed after 2000, when the PSA Group began (in 2005) collaborating with Toyota for the production of small vehicles in the Czech Republic (TPCA, in the town of Kolin) and built an assembly plant for its new Peugeot 207 in Slovakia (in 2006, at Trnava). Renault–Dacia production began taking off at the same time, thanks to the success of the Logan. Today, French carmakers’ capacities in NACs are estimated at nearly 800,000 units, or 10 percent of their European total. This recent expansion by French carmakers has considerably modified the situation for the French suppliers with whom they have often entertained longstanding and solid partnerships. According to our analytical matrix, two types of modular-supplier production should be established locally. First, the presence of carmakers implies that FTSs will build macro-component final assembly units in automobile plants’ immediate vicinity. Examples of supplier parks built towards this end include Peugeot–Citroën at Trnava, TPCA at Kolin and Renault–Dacia at Pitesti. Second, NACs do seem to offer useful locations for components production. Production-cost differentials remain very favourable to these countries, which are relatively accessible and feature a highly skilled workforce.3 The likelihood


10/16/2008

9:32 PM

Page 17


also true that this situation could evolve in the future, as illustrated by Renault’s choice to build in Romania a research centre dedicated to low-cost vehicles. At present, with module specification still being done in Western Europe, we can expect FTS units in charge of finalizing macro-components to remain located here. At the very least, we should expect to find customer-support units located as closely as possible to carmakers’ plants, so they can resolve any technical problems that might arise. Regarding R&D in meso-components and components, we have seen the autonomy of such R&D and can therefore predict that FTSs will decide to build these kinds of units in NACs due to the two strong advantages they offer: a historically recognized technical culture; and (global) production costs that remain low despite rapidly rising wages caused by the relative dearth of skilled labour. All in all, the analytical matrix suggests a complexification of FTS production and research networks on a European scale. To verify the accuracy of this analysis, we must first examine NAC trade with Europe and France. Analysis of changes in automotive parts and components volumes should give us an initial indication of which changes reflect the regional division of labour. This approach will not suffice, however, to answer our questions since it does not specify the nature of the elements being traded (macro or meso-components). We can supplement it by studying which productive systems are present in France and in NACs. The absence of data pertaining to the nature of the entities involved has forced us to adopt a monographic approach. Clearly, adopting a single perspective is insufficient, but it does constitute a launch pad enabling an initial evaluation of our matrix.

PR OO F

ON LY

that we will witness the creation of new entities in NACs does not mean, however, that we should expect these units to make a generalized move eastwards. NACs suffer from North African competition – a traditional zone of delocalization for French suppliers who appreciate its inexpensive workforce, tax-free economic zones and fluency in French. Turkey has also become a major space for the production of automobiles and supplies (Layan and Lung, 2007). In addition, certain components are made in plurifunctional units, i.e. ones where meso-components can also be manufactured. In these situations, suppliers must wait for the market to expand enough to justify relocation costs. Furthermore, productivity gains can be generated in units located in core countries (sometimes following wage concessions negotiated in the wake of a delocalization threat). All in all, the tendency to delocalize components plants should favour NACs. However, this is no more than a tendency. These kinds of plants have in no way disappeared from the core countries – and the practice of resorting to other peripheral countries remains in place. Regarding meso-components, the response depends on the volume of deliveries being made in NACs. If a supplier has many customers running local operations, it can hope to achieve sufficient economies of scale to cover its set-up costs. Otherwise – and given the relative quality of the East–West transport infrastructure – it is more efficient to maintain meso-component plants in the West and supply final module assembly units out of more central locations: North-east France or South Germany, for example, where plants can supply CEE as well as the rest of France and Germany. Local markets’ strong volatility offers a complementary argument (Lung, 2003), since this causes FTSs to proceed cautiously, due to the major volumes required to ensure plants’ profitability. All in all, if we assume an average degree of risk aversion, major relocation costs and relatively low transport costs, we can hypothesize that mesocomponents plants will tend to remain in the West. In terms of design activities, carmakers have maintained most of their research capabilities in Western Europe. Clearly, higher production has been accompanied by the decision to move engineering capacities (which are significant at VW). The fact remains, however, that vehicles’ design per se has been maintained in the central automobile spaces. It is

Empirical evidence: a statistical approach The rising number of supplier entities located in NACs Over the past 15 years, NACs have become a major pole for the supplier industry. We can estimate that from 2010 onwards, 15–20 percent of all automobile equipment made in Europe will be manufactured in CEE (SESSI, 2003). This mainly reflects Western FTSs’ massive European Urban and Regional Studies 2009 16(1)

17


9:32 PM


Page 18

16(1)

carmakers’ recent experience: fewer industrial workers, with the majority of white-collar workers only partially offsetting this loss. The significance of the moves to CEE and the relative decline in the number of production sites located in France suggest that the restructurings have been part of a division of labour process occurring on a Europe-wide scale. According to our analytical matrix, only some of these entities should be located in NACs; that is, most of the productive base should stay in France. Above all, only certain types of entities should prioritize moving eastwards to create interdependence between the East and West (thus generating more interbranch trading).

ON LY

establishment of production and even R&D operations in the NACs from the early 1990s onwards, with a noticeable acceleration since 2000. French firms are fully committed to this process. Based on data provided by FIEV (the French autoparts suppliers association), the Service des Etudes et des Statistiques Industrielles (SESSI) indicated in 2003 that between 1996 and 2000 CEE accounted for more than 30 percent of all French suppliers’ mergers and acquisitions, and for 31 percent of all the entities they set up abroad (but only 9.4% of their joint-ventures). From2001– 05, NAC accounted for nearly 54 percent of French suppliers’ greenfield plants established outside the country (but only 11% of companies purchased and 12.5% of all joint ventures, this being a mainly Asian speciality) (FIEV Data, processed by authors). This extension of capacities in CEE has recently been accompanied by alarmist discourses about alleged declines in the French supplier industry. Data used to substantiate this thesis include the sector’s falling 2006 revenues (-4%) and jobs (4.7%). Yet French industry has maintained its position as Number Two in Europe and not all indicators are convergent. Between 2001 and 2005, 10 new sites were opened in France (as opposed to about 20 in NACs). At the same time, two-thirds of suppliers’ brownfield investments of more than !1m took place in France (FIEV data). Mainly thanks to large companies, investments in the sector rose by 6 percent in 2006 (French Ministry of Industry data). Thus, more than some massive offshoring trend, what we are witnessing is French sites specializing in types of output that require greater skills, as well as a greater intensity of technical and financial capital. The more labour-intensive activities have moved to Eastern and Southern low-cost countries. This assertion is seemingly corroborated by other studies. According to statistics published by the European Restructuring Monitor (EU), the net loss of 24,000 jobs throughout the French automobile sector between 2001 and July 2007 resulted more from internal restructuring (68%) than from bankruptcies (16%) or offshoring (11%) (Database: [http://www.eurofound.europa.eu/emcc/erm/ind ex.htm]). At the same time, more than 10,000 jobs were created in France over this period by carmakers and suppliers. The supplier sector seems to have gone through an evolution comparable to

PR OO F

18

10/16/2008


Trade flows between the European Union and NACs

International trade in automotive goods has been marked recently by new actors’ arrival on the scene. Whereas in 1995 90 percent of world automobile component exports came from triad countries, today these only account for threequarters of the total, with exports from the Rest of the World having risen by a factor of ten. This exceptional growth is mainly due to NACs’ arrival on the international stage, and export volumes have risen by a factor of 30. NACs now have a market share of nearly 6 percent, a tenfold increase that has been marked (Table 2) by a clear acceleration since 2002. Given the exceptional export growth recorded by EU-15 countries (a 50% rise in 10 years) – one that enabled them to maintain their approximate 47 percent share of world trade – the complementarity between CEE and West European activities is obvious. NACs have experienced a parallel acceleration in exports over the past five years. The correlation of these two growth trends reflects one and the same process. CEE’s integration into the broader European automobile system was originally based on greater flows of finished and semi-finished automobile products coming from, and going to, both sides of the Continent. This acceleration of flows from the early 2000s onwards is a clear sign of a deeper Continental division of labour within the sector:


10/16/2008

9:32 PM

Page 19


19

Table 2 Main world exporters of automobile components (in $m and %) 1995 USA + Canada Japan EU (15) NACs (12)

29,851.00 19,322.88 51,349.59 1,440.31

2000 26.20% 16.96% 45.08% 1.26%

40,714.47 17,692.63 63,259.70 3,759.8

2005 28.29% 12.29% 43.96% 2.61%

42,003.22 26,009.04 108,200.01 13,014.95

18.32% 11.34% 47.18% 5.68%

Source: Authors’ calculations based on Chelem CEPII database.

Table 3 NACs: exports of automotive parts and components by destination (%) 1995

2000

2005

56.0 21.3 77.3

76.9 13.0 89.9

77.3 11.0 88.4


PR OO F

increased specialization of sites and growth in the number of production sites and in their size. Integration can also be perceived in flows’ destinations and origins. EU countries have built upon their already dominant positions as targets for CEE automotive exports (Table 3) and as sources of NAC imports (Table 4). Expansion has been remarkable, rising from less than $5b in 1995 to more than $15 b a decade later. What we are witnessing is a noteworthy preference for ‘regional sourcing’. This combined growth in exports and imports is characteristic of CEE’s integration into the European automobile system: the carmakers and major suppliers present in this region have broadened their zone of localization and redefined their specializations. The question is whether this general trend also applies to the French.

ON LY

EU (15) NACs (12) EU (27)

Table 4 NACs: imports of automotive parts and components by origin (%)

The rise in intrabranch trading between France and NACs

These trade balances seem to confirm NACs’ relative specialization (Table 6). Above all, they reveal the underlying dynamic that drives this specialization. If we examine France’s trade balance with NACs, the rise of the CEE supplier industry

EU (15) NACs (12) EU (27)

1995

2000

2005

77.6 12.4 90.0

82.8 8.5 91.3

79.5 12.0 91.5


Table 5 France/NACs trade in automotive parts and components ($m)

Exports Imports French surplus with NAC

1995

2000

2005

258 68 190

495 247 248

1,261 732 530


has had a clearly positive effect. The recent acceleration in the number of new sites that France has opened in NAC has led to further improvements in this trade balance (Table 5). This trend contrasts with overall changes occurring at a global or even regional level (Table 6). Our analytical matrix can be used to interpret these developments. Rising intrabranch cross-flows suggest the existence of localization-based intersite complementarity. In this sense, we seem to have confirmed the hypothesis that certain types of entities are being set up in CEE using goods produced in France, with these sites’ subsequent output then being destined for other French sites. The increase in the French surplus suggests (given the acceleration in number of operations established in CEE) that products made in CEE have a very European Urban and Regional Studies 2009 16(1)


20

10/16/2008

9:32 PM


Page 20

16(1)

Table 6 NACs: Automotive trade balances ($m) Parts and components

World EU (15)

Passenger cars

1995

2000

2005

1995

2000

2005

−1,038 −1,116

−1,986 −1,865

1,037 544

−1,759 −990

2,528 2,471

2,388 2,053


Slovenia Poland Romania Czech Rep.

2000

2001

+207 +15 +5 -24

+170 −27 +16 −55

ON LY

Table 7 France/NACs trade in automotive parts and components ($m) 2002

2003

2004

2005

+215 −30 +17 −34

+200 −63 +26 −55

+276 −115 +54 +0

+376 −73 +137 +7


PR OO F

different commercial value from those made in the West. This can be explained if we assume that the exports (from NACs) mainly involve components and that the exports (from France) relate to meso-components. To fine-tune this latter hypothesis, consider the countries where French carmakers have set up operations. French carmakers’ move into NACs has played a crucial role in changing France’s trade balances: exports (of components and meso-components) towards carmakers’ assembly plants and towards plants (for macro-components) built on suppliers’ parks have largely offset imports coming from delocalized sites. An element of confirmation could be found in the fact that France’s trade balances vary greatly from one NAC to another (depending on whether French carmakers are present). Note also the sudden changes in trade balances once the French have set up a new operation or kind of output (Table 7). The Novo Mesto plant’s integration into Renault’s West European production schema is the driver behind the permanent surplus in France’s trade of parts and components with Slovenia. This surplus has come with a major deficit in vehicle trading, with France absorbing a large share of all locally assembled Clios. Romania clearly shows the effects of the Logan’s September 2004 launch on the Clio II platform, hence its use of a large volume of parts from France and Spain. Similarly, the March 2005 establishment of a TPCA plant led to an immediate European Urban and Regional Studies 2009 16(1)

Table 8 Germany/NACs trade in automotive parts and components ($m)

Exports Imports Surplus/deficit with NACs

1995

2000

2005

749 574 174

2,565 1,768 798

5,160 5,985 −825


turnaround in France’s previous deficit with the Czech Republic. In Poland, however, France is still running a deficit – as if local French FTS entities were producing goods mainly destined to be re-imported by the West (the intra-NAC trade has been relatively limited so far, even if it is growing; see Tables 3 and 4). Germany offers some more food for thought (Table 8). As France is at present, Germany once experienced rising surpluses. However, the length and extent of German carmakers’ presence in NACs have ensured the profitability of their production entities in these countries, including units that produce meso-components. Central Europe can be portrayed as a ’complete space’ where all types of localizations (in the sense that our typology gives to this term) are possible here in the eyes of German suppliers, who seem to have moved on to a higher plane here. The recent reversal in the country’s trade balance with this region can be explained by the fact that many German or foreign suppliers are


10/16/2008

9:32 PM

Page 21


Even though the company is technically the PSA Peugeot–Citroën Group’s supplier subsidiary, it is managed autonomously and encouraged to develop its own customer portfolio. PSA’s proportion of Faurecia’s total sales has fallen over time, from 31.8 percent in 1998 to 25 percent in 2006. The company’s second and third largest customers are VW (22.6%) and Renault–Nissan (12.5%). In addition, Faurecia has considerably increased its productive and commercial internationalization. Initially very France-oriented, it is now mainly focused on non-French markets: in 2005, 69.2 percent of its sales, 79 percent of its manufacturing plants and 64.1 percent of its jobs were located outside the country (respectively, the figures were 56.3%, 67.1% and 41.7% in 1999). With regard to Europe, we note that the growth in Faurecia’s sales here has led to a remarkable extension in its production capacities. However, recent distortions in favour of NACs are also visible. Between 1999 and 2007, Faurecia increased the number of plants it owned by a factor of 1.87 in Western Europe versus 3.5 in CEE. Most of this growth occurred over the past few years, with 13 new plants being opened in NACs between 2002 and 2006. Note that the company had 4,669 employees in this part of the world in 2003, compared to 8,563 by year-end 2006. Including temporary employees, staff numbers here rose from 4,428 in 1998 to 9,934 in 2006. At the same time, staff numbers in Western Europe, after rising between 1998 and 2004 (44,969 employees) fell subsequently (to 40,299 in 2006). Between 2002 and 2005, six plants were closed in Western Europe and three others were sold; there were no such moves in NACs (see Figure 1). Similarly, in 2005 alone, five major downsizing plans were announced in the West, four in France and one in Spain. All in all, Faurecia has reoriented its productive apparatus towards CEE, reducing the West’s previous priority allocation of directly productive jobs that have now been transferred to the NACs (+35.1% industrial jobs between 2005 and 2006). This change does not mean that Faurecia is entirely abandoning its Western operations. France and Germany combined account for nearly 60 percent of the company’s European sites (Table 9). This preponderance still holds for the four types of entities we have distinguished – even if a number of disparities are apparent, depending on an entity’s function. Analysis of such entities should help us to test our matrix.

ON LY

selling to German carmakers which have set up operations here. German operations have proliferated due to slightly lesser transport constraints (spatial proximity) and improved transport infrastructure. South Poland offers a good example of how the proliferation of First (but also Second) Tier Supplier sites leads to the formation of veritable clusters benefiting from easy links to Germany and more broadly to the whole of Western Europe (Domanski et al., 2006). The growth in intrabranch flows and the direction of trade balances therefore seems to confirm the existence of a division of labour between CEE and Western Europe as well as an organization between these two types of spaces of a complementarity that can be explained in our analysis by the presence of entities featuring distinct functions. To further this analysis, we need to study the exact nature of such trade flows; to wit, whether they involve components, meso-components or macro-components. A monographic analysis may offer us a few elements of response.

PR OO F

Empirical evidence: Faurecia case-study

In revenue terms, Faurecia is France’s leading autoparts supplier, the Number Two in Europe and Number Ten in the world according to AutomotivesNews rankings. In 2006, the company had sales of !11.6b, employed more than 65,000 people, owned 181 production plants and 28 R&D centres worldwide, and maintained a presence in 28 countries. Its activities are split into six areas: seats, instrument panels, doors, acoustic packages, frontend modules, and exhaust systems. It has experienced strong growth since its creation, with sales rising by a factor of 3.8, and staff numbers by a factor of 1.4, between 1997 and 2006. This growth is mainly due to mergers and acquisitions. The company was born in 1998 out of a merger between three French suppliers (ECIA, ECTRA and B. Faure). In 1999, it acquired a US company called APAS (exhaust systems) and in 2000 took over a French company with a strong presence in Germany and Spain, Sommer Allibert (door panels, acoustic environment and instrument panels). This external growth is what enabled Faurecia to accelerate its sectoral and geographic diversification.


21


22

10/16/2008

9:32 PM


Page 22

16(1)

Table 9 Faurecia’s facilities in Europe (Number of units; January 2007)

Manufacturing plants Design & development Research & development Customer office

France

Germany

Other Western countries

NACs

37 5 7 3

27 5 3 15

39 2 0 9

21 0 1 0

R&D entities: from a refocusing in the West to an opening to the East?

ON LY

Source: Collected by authors from Faurecia website.

PR OO F

Faurecia has ostensibly chosen to maintain all its service activities in Western Europe. The absence of any customer offices in NAC signifies that relations between the supplier and its customers are shaped in the West. In turn, this indicates that local entities are relatively peripheral to the company system. This peripheral status is clear with regard to R&D activities. The company distinguishes between two types of entities: R&D units focusing research at the most upstream levels into modules’ architecture (and into their components); and design & development (D&D) units corresponding to engineering centres. In our analysis, D&D units should be located near carmakers’ design centres because these are the entities with which they interface. Table 9 confirms that Faurecia has situated such units in countries already possessing these kinds of automobile development centres: France and Germany (from where its three main customers come), as well as Spain and Sweden. Alongside this, Faurecia has R&D centres that tend to be specialized by main areas of competency or functionality. In Table 1, these centres correspond to what we call ‘autonomous research’ because they are not dedicated to any single carmaker. Almost all such sites are located in France and Germany. The company’s priority these last few years has been to reinforce their importance, enlarging and specializing such centres in an attempt to convert them into group centres of excellence. For example, in 2005, Faurecia developed and designed the French sites at Flers, Brières-les-Scellés and Magny-en-Vernois, converting them to world centres, respectively, in


seat mechanisms, complete seats and seat padding. In 2006, acoustic package and internal fittings divisional engineering activities were grouped in Cologne (Germany). The final site in this picture is the one located at Grójec (Poland). Opened in 1998, it is an engineering centre which collaborates with the company’s other R&D sites in areas like seat structures and mechanisms. Here Faurecia seems to be exploiting the autonomy it possesses to engage in this type of research. In terms of attractiveness, Poland benefits from a solid industrial tradition, in particular in mechanics, and can (still) offer relatively low labour costs. More dynamically, R&D capacities often emerge when a productive basis is being developed. Complex production requires a minimum of engineering capabilities located nearby. Once a critical mass of interactions has been created between production and engineering (Grójec combines two plants – the total site employed 1,323 persons by year-end 2006), it becomes legitimate to build real local R&D capacities to benefit from knowledge externalities and user/producer interactions.5

Production sites According to our analysis, Faurecia should locate two types of entities in NACs: component plants to benefit from low production costs; and macrocomponent plants supplying its main customers on an ILS or JIT basis. The reason is that Faurecia delivers several different modules, destined to be used on the different cars being assembled in the zone. This can be assessed using data that Faurecia has supplied on the nature of its units, classified into component and JIT plants.6


10/16/2008

9:32 PM

Page 23


Table 10 Faurecia manufacturing plants in Europe (January 2007) EU (15)

JIT Components Total

suppliers’ need to follow carmakers abroad, hence their need to engage in a process that revolves around an international extension of their production capacities whenever complementarity has a stronger effect than substitution.

NACs

No.

%

No.

%

22 81

21% 79%

3 18

14% 86%

103

100%

21

100%

The rise of modular production and associated decomposition of production processes seems to be the harbinger of a new dynamic in the supplier industry’s international division of labour. Mainly driven by FTSs, in terms of proximity this has generated new opportunities and constraints. NACs are being integrated into French FTSs’ European productive spaces in line with a logic that combines a spatial division of labour with an imitation of those entities that already exist in the West. Given the significance of spatial constraints, the West should remain a production location, for certain activities at least. This conclusion draws upon a conceptual framework that tries to explain FTSs’ location choices via productive and organizational constraints. Firms’ geography stems from their exploitation of local advantages and from an interaction between the tightening and easing of proximity constraints. Clearly this analysis needs to be fine-tuned and supplemented. First at a theoretical level, the breakdown of production into three stages requires further extension, in particular with a specification that is determined by module types. In addition, we implicitly hypothesize that a product’s modular breakdown covers different stages in the production process and that it is ultimately possible to associate this with monofunctional entities. In turn, this helps us to establish a spatial division of labour. As heuristic as this hypothesis may seem, it still merits further exploration. Last, at the empirical level, the analysis functions by approximation due to a lack of data meshing with our typology. Our findings do appear promising, however, and we see this as an encouragement for the further extension of this research programme, specifically by carrying out new monographic analyses on FTSs.

ON LY

Source: Collected by authors from Faurecia website.

Conclusion

PR OO F

Although both types of entities can be found on both sides of Continental Europe, component plants are relatively more numerous in the NACs (Table 10). Faurecia seems to be pursuing two objectives: creating component plants in NAC; and setting up JIT plants in carmakers’ immediate vicinity (hence in Western Europe, where carmakers mostly operate). Reasoning in terms of factory creations should improve our understanding of the current relocation process (Table 11 and Figure 1). Faurecia has reinforced NACs’ role as component production sites, seeing as how out of ten components plants opened between 2002 and 2006, nine were located in NACs and only one in France. These plants are less often dedicated to a single carmaker and should therefore be analysed as part of a global sourcing cost-reduction strategy whose purpose is to benefit from the absolute advantages that CEE offers, and from the lesser proximity constraints that hobbles this kind of production. Faurecia’s aim is that over time 40 percent of its components should come from offshoring or foreign subcontracting (compared to 25% in 2005). At the same time, having captured new customers in the West, Faurecia now wants to set up JIT plants there as well, dedicating most of them to a single car model. This need to locate macro-component assembly sites in carmakers’ immediate vicinity also explains why sites that operate on a JIT and/or dedicated basis are being established in NACs in a way that amplifies the impression that what we are witnessing is no more than a partial delocalization trend. CEE’s attractiveness may stem in part from its low production costs, but we should not ignore


23


24

10/16/2008

9:32 PM


Page 24

16(1)

Table 11 Type of production plants being opened by Faurecia in Europe, 2002–06 (no. of units) Plant function

EU (15) NAC

Nature of production

Components

JIT

Undetermined

Dedicated

Non-dedicated

Undetermined

Total

1 9

11 4

1 0

11 4

1 8

1 1

13 13

ON LY

Sources: Collected by authors from Faurecia website and ‘Annual Report’ (different years).

here of 12 countries (Estonia, Lithuania, Latvia, Hungary, Bulgaria, Rumania, Cypress, Malta, Poland, Czech Republic, Slovakia and Slovenia). 5 The example of Delphi in Mexico seems analogous, with the US supplier having steadily increased its Ciudad Juarez R&D capacities over the years to the point that this has become a world R&D centre (Carrillo, 2004) 6 ‘They are usually located less than 30 kilometres from the assembly lines and are sometimes even located in the same industrial footprint as Faurecia’s customers.’ (Faurecia, 2006b: 9)

Opening component plant Closure

PR OO F

Opening JIT plant Opening undetermined plant S

Figure 1 The location of new Faurecia plants (2002–2006) or closed plants (2002–2005) in Europe Note: Faurecia opened another components plant in Romania after acquisition of a local supplier in 2006. It is not shown on the map as we could not find its exact coordinates. Source: Collected by authors from Faurecia website and ‘Annual Report’ (different years).

Notes

1 Each of the 30 leading FTSs owns on average 149 production plants and 25 R&D centres worldwide, while maintaining a presence in anywhere between 19 and 36 countries (Frigant, forthcoming). 2 In 2000, Autobusiness predicted that by 2010 the modules market would grow by 166%. Cockpits (+1031%), roofs (+1,149%) and doors (+636%) were supposed to experience the strongest growth. 3 In 2005, the cost of one man-hour in the automobile industry was €28.50 in France, €5.80 in Poland and 2.10 in Romania (Eurostats, Euler Hermes SFAC), 4 This analysis uses CEPII’s Chelem database, which surveys international product flows. NACs are composed


References

Autobusiness (2004) Modules and Systems. Stamford: AutoBusiness Ltd. Baldwin, C. and Clark, K. (2000) Design Rules: the Power of Modularity. Cambridge (MA.): MIT Press. Belzowski, B., Flynn, M., Edwards, M., Ban, L. and Martin, G. (2006) ‘Supply Chain Management: New Competitive Realities in the Automotive Value Chain’, International Journal of Automotive Technology and Management 6 (4): 387–405. Bordenave, G. and Lung, Y. (1996) ‘New Spatial Configurations in the European Automobile Industry’, European Urban and Regional Studies 3 (4): 305–21. Carrillo, J. (2004) ‘Transnational Strategies and Regional Development: the Cas of GM and Delphi in Mexico’, Industry and Innovation 11 (1/2): 127– 53. Carrillo, J. and Contreras, O. (2007) ‘The Historical Evolution of American Auto Firms in Mexico’, Proceedings of the 15th GERPISA International Colloquium (June), Paris. Carrincazeaux, C., Lung, Y. and Rallet, A. (2001) ‘Proximity and Localisation of Corporate R&D Activities’, Research Policy 30: 777–89. Chanaron, J-J. (ed.) (1999) International Division of Labor and Assembler–Supplier Relations. Evry: Actes du GERPISA. Domanski, B., Guzik, R. and Gwosdz, K. (2006) ‘The New Spatial Organisation of Automobile Industry in


10/16/2008

9:32 PM

Page 25


Lung, Y. (2003) ‘The Changing Geography of the European Automobile System’, Working papers of GRES 2003–10. Accessed on 29 Sep. 2008 at: [http://ideas.repec.org/p/grs/wpegrs/2003–10.html]. Lung, Y. and Mair, A. (1993) ‘Innovation institutionnelle, apprentissage organisationnel et contrainte de proximités: les enseignements de la géographie du juste-à-temps’, Revue d’Economie Régionale et Urbaine 3: 387–403. Pavlinek, P. and Janak, L. (2007) ‘Regional Restructuring of the Skoda Auto Supplier Network in the Czech Republic’, European Urban and Regional Studies 14 (2): 133–56 Prencipe A., Davies, A. and Hobday, M. (eds) (2003) The Business of Systems Integration. Oxford: Oxford University Press. Rallet, A. and Torre, A (2005) ‘Proximity and Localization’, Regional Studies 39 (1): 47–59. Sadler, D. (1994) ‘The Geographies of Just-in-time: Japanese Investment and the Automotive Components Industry in Western Europe’, Economic Geography 70 (1): 41–59. Sadler, D. (1998) ‘Changing Inter-firm Relations in the European Automotive Industry: Increased Dependance or Enhanced Autonomy for Components Producers?’, European Urban and Regional Studies 5 (4): 317–28. Sako, M. (2003) ‘Modularity and Outsourcing: the Nature of Co-evolution of Product Architecture and Organisation Architecture in the Global Automotive Industry’, in A. Prencipe, A. Davies and M. Hobday (eds) The Business of Systems Integration, pp. 229–53. Oxford: Oxford University Press. SESSI (2003) La chaîne de l’équipement automobile. Paris: Minefi. Takeishi, A. and Fujimoto, T. (2003) ‘Modularization in the Car Industry: Interlinked Multiple Hierarchies of Product, Production, and Suppliers Systems’, in A. Prencipe, A. Davies and M. Hobday (eds) The Business of Systems Integration, pp. 254–78. Oxford: Oxford University Press:. Ulrich, K. (1995) ‘The Role of Product Architecture in the Manufacturing Firm’. Research Policy 24: 419–40. Vale, M. (2004) ‘Innovation and Knowledge Driven by a Focal Corporation: the Case of the Autoeuropa Supply Chain’, European Urban and Regional Studies 11 (4): 124–40. Volpato, G. (2004) ‘The OEM–FTS Relationship in Automotive Industry’, International Journal of Automotive Technology and Management 4 (2/3): 166–97.

PR OO F

ON LY

Poland’, Proceedings of the 14th GERPISA International Colloquium (June), Paris. Estall, R. (1985) ‘Stock Control in Manufacturing: the Just-in-time System and its Locational Implications’, Area 17: 129–33. Faurecia (2006a) ‘Annual Report’. Accessed on 29 Sep. 2008 at: [www.faurecia.com/data/en/download/ annual_reports/2006/faurecia_ra_2006_ang/]. Faurecia (2006b) ‘Institutional Brochure’. Accessed on 29 Sep. 2008 at: [www.faurecia.com/data/en/download/ annual_reports/2006/faurecia_institutional_2006_en.pdf]. Fourcade, F. and Midler, C. (2005) ‘The Role of 1st Tier Supplier in Automobile Product Modularisation: the Search for a Coherent Strategy’, International Journal of Automotive Technology and Management 5 (2): 146–65. Frigant, V. (forthcoming) ‘Winners and Losers in the Auto Parts Industry: Trajectories Followed by the Main First Tier Suppliers over the Past Decade’, in M. Freyssenet (ed.) The Automobile Industry in the 21st Century. New Perspectives for a Changing Market, Houndmills and New York, NY: Palgrave MacMillan. Frigant, V. and Lung, Y. (2002) ‘Geographical Proximity and Supplying Relationships in Modular Production’, International Journal of Urban and Regional Research 26 (4): 742–55. Garud, R. and Kumaraswamy, A. (1995) ‘Technological and Organizational Designs to Achieve Economies of Substitution’, Strategic Management Journal 16: 93–109. Hobday, M., Davies, A. and Prencipe, A. (2005) ‘Systems Integration: a Core Capability of the Modern Corporation’, Industrial and Corporate Change 14 (6): 1109–43. Humphrey, J. (2000) ‘Assembler–Supplier Relations in the Auto Industry: Globalisation and National Development’, Competition & Change 4: 245–71. Kirat, T. and Lung, Y. (1999) ‘Innovation and Proximities: Territories as Loci of Collective Learning Processes’. European Urban and Regional Studies 6 (1): 27–38. Larsson, A. (2002) ‘Learning or Logistics? The Development and Regional Significance of Automotive Supplier-parks in Western Europe’, International Journal of Urban and Regional Research 26 (4): 767–84. Layan, J-B. (2006) ‘L’innovation péricentrale dans l’industrie automobile: une gestion territoriale du risque de résistance au changement’, Flux 63/64: 42–53. Layan, J-B. and Lung, Y. (2007) ‘Les nouvelles configurations de l’espace automobile méditerranéen’, Région & Developpement 25: 157–76. Lundvall, B.A. (1988) ‘Innovation as an Interactive Process: From User–Producer Interaction to National Systems of Innovation’, in G. Dosi, C. Freeman, R. Nelson, G. Silverberg, L. Soete (eds) Technical Change and Economic Theory, pp. 349–69. London: Frances Pinter. Lung, Y. (2001) ‘The Coordination of Competencies and Knowledge: a Critical Issue for Regional Automotive Systems’, International Journal of Automotive Technology and Management 1 (1): 108–27.

Correspondence to: Vincent Frigant, GREThA UMR CNRS, University of Bordeaux IV, Avenue L. Duguit, F-33608 Pessac France. [email: [email protected]] European Urban and Regional Studies 2009 16(1)

25

10/16/2008

9:32 PM

Page 26

PR OO F

ON LY
