| La Lettre du GERPISA | no 115 (july 1997) |
I have presented these ideas in order to stimulate a discussion and begin to form a network. Clearly, the ideas would be better if I took the trouble to read all of the contributions to the GERPISA meeting carefully. I could also benefit greatly from going back through the extensive GERPISA documentation from previous meetings and the Actes du GERPISA. But if I do this, I will not have time to prepare this note before November. So apologies for its failure to take sufficient advantage of the work of others.
My strong belief is that each of the elements of potential GERPISA networks have much to gain from co-operation. The challenge is to create mechanisms that enable and stimulate this co-operation.
The enormous advantage of GERPISA is that it brings together researchers from a number of different countries (industrialised and developing) concerned with not only the global auto industry, but also questions of economic development. GERPISA brings together both researchers from the "South" and "North", which opens up the possibility of considering the auto industry simultaneously from, on the one hand, the perspectives of national development and government policy, and on the other hand the global expansion of the auto industry itself. At the same time, the GERPISA network offers an unusually wide degree of coverage of developing countries and an equality of access and status to researchers from industrialised and developing countries. These are unusual and potentially power ful characteristics that put the GERPISA networks in a unique position.
Our challenge is to take advantage of this potential by developing networks of researchers that can draw on wide ranges of information and expertise. My strong belief is that each of the elements of potential GERPISA networks have much to gain from co-operation. The challenge is to create mechanisms that enable and stimulate this co-operation.
The "new spaces" are heterogeneous, and it does not make sense to put them altogether. During the GERPISA meeting, the presentations were separated into to groups: (i) the "independent" automobile spaces (Mercosur, India and Russia), and (ii) those spaces which are linked to developed regions, Portugal, Mexico and Eastern Europe. I think that this basic division made sense.
There are many ways of approaching the issue of the new spaces.I am particularly concerned with the development of the dynamic new markets that are being transformed by a combination of foreign direct investment (FDI) and trade liberalisation. The four key spaces in my view are Mercosur, India, China and the ASEAN region. These were the four areas emphasised by the Renault and Peugeot representatives at the meeting. The suggestions which follow are concerned with these regions.
The focus on these regions does not exclude either the formation of other groups (which is something to be expected and welcomed) or the establishment of links between groups. However, it is important to establish research questions and objectives which are coherent, and this note tries to show that the four chosen areas provide a coherent focus for research.
Achieving global coverage is not an objective in itself. If the role of researchers in GERPISA is merely to provide information ("description") for others to theorise, then there is some merit in achieving the broadest possible coverage. My own view is that within the context of the important general questions about industrial models, there are significant issues about the development of the auto industry in the new spaces which require the development of researchable, mid-level hypotheses. I set out some of these below. The study of these hypotheses requires a narrower focus.
I foresee a matrix structure for studies. Comparative studies of each of these themes will be undertaken in some or all of the four regions. Different people/groups will work on one or more of these themes.
MercosurIndiaChinaASEAN
The space for this work will not remain open for long. My own opinion (which does not have much foundation) is that there is an opportunity to make a significant contribution to the study of the auto industry, but that we must move quickly. The time-scale is not three or four years. A network could produce a first volume of country/theme/comparative studies within 12-18 months. Waiting any longer than this would be a mistake.
This means that much of the work will have to be done between the annual GERPISA meetings. Given the enormous distances between the researchers involved, much of the work will have to be done through electronic mail.
The demand for cars. The demand for cars has been altered substantially in the new spaces by (i) the revival of economic growth, (ii) the easing of supply-side constraints in India and China, and (iii) changes in trade policy. At present, growth in demand is rapid, and there is enormous potential for car use in all four regions. From the point of view of GERPISA, a number of questions need to be considered:
- How stable is the growth in the demand for cars? Comparative work on the growth of the previous "new spaces" such as Italy in the 1950s and Spain in the 1970s would be useful. At the same time, it is important to consider the possibility of growth constraints or economic crisis, as was stressed by one of the contributors to the debate in the last session of the June conference.
- What type of cars are required for the emerging markets? There has been increasing talk of cars aimed specifically at developing countries. The Fiat Palio and the Honda City are two examples. However, the initial impact of liberalisation in the Mercosur region was to bring model ranges in Argentina and Brazil closer to those offered by the makers in the industrialised countries. Older models were phased out rapidly. While the Volkswagen Gol is still in production, it is likely to be the last car designed specifically for the Brazilian (or Mercosur) market. At the same time, it is worth noting that motor manufacturers have often ignored the cheaper end of the market. The "popular car" strategy in Brazil was suggested by the government, and up until that point the major car manufacturers had shown declining interest in this segment. What, therefore, determines the model ranges being offered by manufacturers? In principle, there is a conflict between the needs of the new motoring classes for cheap cars and the higher profits to be made on larger vehicles.
- The determinants of auto policy. Although the world is supposedly moving towards freer trade, the development of the auto industry in all four of the regions is highly influenced and shaped by government policy. Each region has adopted a distinct set of policies towards the auto industry. Briefly, China maintains a highly directed auto policy which restricts the number of entrants to the local market and guarantees that the multinational car producers have minority stakes. The existence state-owned and recently private auto companies have a privileged position. In auto parts, entry is freer. In India, levels of protection remain high, but there is a free-for-all with regard to FDI. In Brazil, tariffs were reduced rapidly. When this policy ran into difficulties, the assembly sector regained protection while the nominal rate of protection for the parts sector is very low (calculations of the effective rate of protection in auto parts suggest highly negative rates). I am not sure about the situation in ASEAN, but I assume that given the large number of small plants scattered across the region (see table 3 of my GERPISA paper) there must be a strong policy influence. All this raises interesting questions:
- What are the political determinants of auto policy? This issue concerns the broader political forces determining policy and the link between auto policy and industry/trade/investment policies more generally. It also concerns the politics of different interests within the auto industry itself.
- What will be the impact of the Uruguay Round tariff reductions on the auto industry? China will not be affected to the same extent as the others. But current policies in all four regions would seem to be out of line with what is required under WTO rules. The likely impact on the auto industry country of the new regulations deserves close attention.
How sustainable are current auto industry policies in the four regions? We have already seen that the auto policy in Brazil was wrecked by a surge of auto imports in 1994-95 (Leite, da Silva et al., 1996). One of the most common reasons for policies aimed at promoting a local auto industry is the desire to neutralise the balance of trade impact of imported cars. This was behind the development of Brazilís auto policy in the 1950s. A combination of prohibitive tariffs on BU and CKD vehicles and incentives for local production made car ownership compatible with the balance of trade constraint. Now, the chosen policies appear to be less those of high tariffs and local content requirements, and more aimed at balancing imports and exports. A number of countries (such as Argentina and India) have policies which allow imports of parts or vehicles as long as they are balanced by exports of parts or vehicles. Leaving aside the question of how companies get around such restrictions, the question remains of how companies could balance imports and exports. For new entrants into Mercosur and India, production scales tend to be low (under 50,000 units per annum). Such scales are also the norm in ASEAN countries. What are the levels of local content for such units? If local content were low (say, one third), then a considerable proportion of local production would have to be exported to balance trade. Mercedes in India entered into a commitment to export half of its Indian CKD production, but it has been unable to do so. In India, even Maruti, which produces on a large scale and has 15 years of experience, still exports only 20% of the value of its imports. How can such policy regimes be made to work? What will happen when they do not?
Why are plant scales so different in Mercosur and ASEAN?
Will conformance with GATT rules lead to convergence?
Are there company or country differences in plant sizes?
What are the consequences of auto industry policy for national
development?
Part of the answer to this question will be considered under theme 2.
With the shift towards design at the suppliers, auto companies are looking increasingly for global supply networks to meet their needs. One contract will be awarded for a particular component, and the supplier should be in a position to supply this part to most (if not all) of the assemblerís factories around the world. Ford and GM negotiate one contract for brakes for each new model. The winner arranges supply not only to North America and Europe, but to Mercosur and India as well. Production, however, will not be centralised. The assemblers expect local producer to be available in each major market.
The impact of this shift on the structure of the world components industry is already apparent. First, there has been a considerable concentration of the industry, and mergers or takeovers between companies in different regions have become common. The scale required to be a successful components producer is rising rapidly. Second, the number of collaboration agreements is rising. Companies may retain independence, but they form networks to secure access to technology and worldwide production facilities.
The impact of these changes on companies in the new automobile regions has been
less studied, but just as dramatic. The case of Mercosur is particularly
striking:
Leading local manufacturers which had pretensions to compete in international
markets on equal terms have been forced to abandon this strategy. The largest
Brazilian auto components company has a turnover of "only" $750 million, which
is too small by world standards. The barriers to entry for technology leaders
in the components sector have increased enormously. The dreams of some
Brazilian firms of being leading independent producers can no longer be
realised. This led one of the industryís technological leader in Brazil, Metal
Leve, to sell its operations to a consortium of Brazilian and German firms.
The auto assemblers in Brazil are moving towards a policy of "follow sourcing". When parts were simple and were designed in-house, it was easy to find local producers with the necessary capabilities. The auto industry in Argentina and Brazil (and India?) created local supply chains consisting of both national and transnational companies. With the shift to global sourcing and design inputs, firms located in the new automobile spaces no longer compete directly for key orders. When Mercedes introduces a new truck in Brazil, many of the parts will already be in production in Europe, and Mercedes will look to Brazilian associates or subsidiaries of its European suppliers when allocating orders. The decision about which Brazilian company to choose will have been determined long before production in Brazil starts. At best, a Brazilian firm will form part of an international consortium bidding for a global contract. Therefore, even in the "domestic" or regional market, global links matter now in a way they did not a decade ago.
This shift in policy has been stated very clearly by Volkswagen in Brazil. The head of purchasing has stated publicly that VW has a policy of "follow sourcing". Local companies have to have the same technology as the German suppliers of the same part. Therefore, a Brazilian supplier should have a technology agreement or strategic alliance with a German auto part company (perhaps European company - this depends on European supply strategy). Failing this, VW would encouraged the German (European?) supplier to build a factory in Brazil. The final pressure on the Brazilian producers is the threat of imports. While most of the established assemblers import relatively little, they use a concept of "global sourcing" as a very explicit threat to local parts producers.
Similar processes are seen with new entrants into the Brazilian market. One parts company has won a contract with a Japanese new entrant because its European associate supplies the assemblerís plant in Europe.
This raises important research questions:
Will the formation of global alliances and centralised design lead to the collapse of R&D in the new automobile spaces? The "Metal Leve" strategy of acquiring independent technological capability seems to have failed. Will local producers be confined to adapting designs to local conditions (roads, fuel, climatic conditions)?
Does the development of new, developing country models (Honda City, etc.) open up possibility for supply networks not centred on the Triad production systems? Experience in Brazil with the Fiat Palio seems not to be encouraging, but I know little about this. How might government policy increase such a possibility?
What are the trade-offs between decentralised production to be close to markets/customers and centralised production for economies of scale? This point was raised in Laigleís analysis of the Valeo group at the conference. Paradoxically, just when tariff reductions make centralised production politically feasible, the shift in production systems towards JIT raises the question of decentralisation. This is an issue which is of considerable interest both at the inter-regional level and the regional level. The production system of the ASEAN countries, characterised by multiple small assembly plants producing similar vehicles, lends itself to more centralised component production. Alternatively, economic integration may lead to specialisation by plant and centralisation of assembly operations (as seems to be happening in Mercosur).
These changes also have implications for the industrialised countries. If, for example, Renault and Peugeot manage to double their non-European production (as the managers at the June meeting stated), then will the ability to supply worldwide become a more critical factor in supplier selection? Will this lead to further concentration in the parts industry?
In Brazil, in the early days of the auto industry, there were also forms of co-operation between the newly-implanted assemblers and the emerging local supplier base, for much the same reasons as in India (Addis, 1993). Similarly, there have been a number of studies which have emphasised the role played by the assemblers in upgrading suppliers and providing technical assistance (Posthuma, 1995). However, even the assemblers are aware that power relations in the Brazilian auto industry very much favour the assemblers. The head of purchasing at VW made this point publicly in 1995. The introduction of the new "Automotive Regime" (see the various papers on Brazil at the 1997 GERPISA conference) has further shifted the balance in favour of the assemblers by protecting the car market while simultaneously allowing imports of components at extremely low tariffs.
The Brazilian components producers (both locally-owned and transnational) regard the assemblers as exploitative. The ideas of "risk sharing" and "open books" are used in order to impose cost reductions on the suppliers. The assemblers frequently use the threat of imports to force price reductions. Contracts are definitely not long-term. Order sizes are not guaranteed.
Will the new Japanese entrants into the Brazilian market pursue a similar strategy to that of the European and American firms who dominate the market at the present time? Is it possible to see contrasts between Japanese and non-Japanese firms competing in the same markets (ASEAN, for example)? In Brazil, the assemblers have opted against partnership. What do they lose by this? If the key advantage of closer relations lies in design and development, does the centralisation of design mean that relationships can be more fluid at the periphery? The strategy of the Brazilian assemblers is somewhat surprising. There are also questions about how it is possible. If global sourcing is being operated, the threat of discontinuing supply might appear to be weak. Are the overseas suppliers other factories from the same group as the Brazilian company? If so, why do the suppliers allow this? Alternatively, the use of dual or multiple sourcing in Europe or North America might widen the choice available to the Brazilian assemblers.
On labour, I put forward some hypotheses from the work of Turner and Auer. I am sure that the papers presented at the meeting would provide further interesting hypotheses, but I have not had time to study them carefully. This note is just a first attempt to stimulate interest in the formation of the network of active study.
The Turner and Auer (1996: 235-36) propositions were:
I think that the fourth hypothesis can be discarded. The third hypothesis needs much more development, and the recent paper by Pil and McDuffie (1996) made available by Mr BÈcir to GERPISA members contains a number of useful observations about team working and "high involvement work practices" (although I am sure that there are others in past GERPISA papers). The prevalence of these practices and the way they are introduced are clearly an important questions for us. My feeling is that there will be enormous and significant differences in the use these practices in the four regions. The crucial question, of course, is what this means for both labour and management.