The market structure which best describes the car industry is a highly rivalrous oligopoly characterised by workable competition. The industry experiences significant changes in firms’ market shares, both on a national and a regional level, the effective exit of producers and the entry of major, usually Japanese, car makers, first as importers and currently as new producers.
On occasions behaviour in the industry can approach that of a perfectly competitive market, but as product differentiation is a major feature of the industry, and the number of major competitors, although large, is under twenty autonomous companies worldwide, behaviour usually encompasses features of imperfectly competitive and oligopolistic markets. So the behaviour of firms includes the use of price competition but many forms of non-price competition as well. However, despite the latter the price variable cannot be ignored.
Clearly, competition in the car industry manifests itself in many ways and with different intensity. Firms do try to increase their market power be it through product differentiation or by actually trying to contain the full operation of market forces. This means that competition in the car industry cannot be divorced from the structure of the industry and the nature of long and short run supply. In other words the automotive industry attempts to attain the costs structure on the supply side of the supply and demand equation that will allow firms to operate profitably no matter how competitive and keen the market conditions and prices. The market is the working out of demand and supply in the automotive sector.
The supply conditions in the motor industry provide an interesting insight into cost behaviour in both the short and long run; that is, the way costs behave at different levels of output.
Although the total fixed costs of vehicle making are high in the short run, a firm with a standard, or average, level of output at near optimum scale, is able to spread them thinly (Table 1). One problem highlighted by Table 1 is how to treat labour costs. The economics textbooks traditionally regard labour as a variable factor of production in the short run. However, in the 1970s and early 1980s, because of the impact of employment protection legislation firms in the real world were loath to vary their workforces until long-run demand patterns became clear. Hence, as in Table 1, labour costs became regarded as ‘fixed’ in the modern motor industry. As the period since the 1980s has unfolded and as the necessity of improving efficiency to survive became more pressing, so the use of labour has become increasingly ‘variable’ once again.
As at least 66 per cent of unit average costs of a car are variable even if labour is treated as a fixed cost, rising to 76 per cent when labour is treated as a variable, it is obvious why car firms are loathe to engage in short-run price competition which involves cutting the price of existing car models. Even though car demand is price elastic it is unlikely that the increase in revenue would be enough to cover the increase in total production costs. So without reduced variable costs per car, or major production improvement, shifting the short-run (and hence the long run) cost curve downwards, short run price cuts will reduce profits unless the cars are sold out of stock. Indeed, short run price cuts are often designed to clear stock.
The actual manufacture of vehicles can be divided into a number of distinct operations each with its own optimum scale (Table 2). However, as well as technical economies, opportunities for lowering unit costs stem from optimising the sales network, advertising, finance, risk management and research and development (Table 3). These remain important, and there is no reason to suppose a significant reduction in their impact in future.
Where technical economies of scale are concerned and in terms of similar units, the most efficient car firm would, for instance, need to produce two million units using common body panels. In practice no such volumes for one model are achieved anywhere, but firms try to approach the optimum by using as many common panels as possible over a range of vehicles. Similarly, one type of engine is used over a variety of models. Another approach is for firms to collaborate and co-operate, so that different manufacturers use the same basic components. This turns an ‘internal’ into an ‘external’ economy of scale. Table 3 shows why so many different firms try to collaborate on marketing, and on research and development.
From the late 1980s the view appeared that technological change would reduce the optimum size of car firms. However, so far this merely means that new flexible production equipment in assembly plants facilitates the manufacture of a wide range of models on the same assembly line. This reduces the modelspecific, but not the overall, assembly optimum: indeed robots can actually increase the latter. In addition, new types of equipment, which can be used over a variety of models and over time, reduce the fixed cost per model. This reduces the numbers of a particular car that must be made to recoup tooling and development costs. Similarly, the use of computer aided design (CAD) might reduce the R&D costs per car. However, in all other areas of technical and non-technical activity the optimum has not yet been reduced by new technology. The most fruitful approach to reducing the optimum size lies elsewhere: a switch to what can be vertical disintegration through co-operation, collaboration and buying-out. This reinforces the long term historical phenomenon in the auto sector of outsourcing, but can also lead to short term and long term functional mergers.
Despite new technology, new flexible equipment and new operating systems, which can reduce the fixed costs of R&D or can put a variety of vehicles along a production line thereby reducing the carspecific optimum, large scale is still needed for optimal performance in many of the car making processes (Table 2). Thus while the ‘shape’ of the long-run average cost curve may change (say with lower cost penalties at sub-optimum scale), the minimum efficient scale is hardly altered.
The relationship between volume and costs shows that minimum efficient scale for an integrated car producer is around three million units a year. (Indeed the pursuit of risk bearing economies through globalisation and the increase in R&D demands, has if anything pushed this even higher towards four million.) Clearly if engines, castings, pressings, marketing and R&D were outsourced then the optimum size would be smaller. However despite speculation about vertical disintegration increasing and the emergence of further external scale economies very little has happened to alter the overall optimum despite increased outsourcing. (True, very small car makers and some heavy vehicle makers survive by outsourcing many of their ‘systems’, such as engines, power train requirements, and so on.)
By 2002, the only full line producer (such as, a manufacturer making a full range of volume-produced cars from minis to large executive cars) to exceed this level of production in all-Europe was VW Group (3.32m units). However, Ford (1.84m), Peugeot (PSA) (1.91m) and GM (1.96m) were adrift. Subsequent growth by Peugeot took them to 2.5 million in 2004 and planned expansion in the new EU states (Czech Republic, Slovakia) will take Peugeot to 3.1 million if everything else remained equal – for instance, no plant closures in the west. In the case of Ford and GM output fell over this period, but VW held its own. On a global scale, GM, Ford, Toyota, Nissan Renault and VW attain this scale of car production whilst Honda, Hyundai, and Peugeot Citroen are getting close with new net capacity.
Unit costs are reduced not only by operating at optimum scale but also by reducing input prices and non-scale cost reductions generally – this has been the aim of ‘lean production’. Consequently GM concentrates the manufacture of its small supermini in low-wage Spain, and by developing SEAT, VW did the same. Fiat uses Poland for its mini car sourcing. As the labour content does not increase proportionately with the size of the car, superminis are relatively labour intensive to make. It therefore makes economic sense to produce them in low-wage economies – provided the gains from the lower wages are not offset by lower labour productivity – even though the car plants in these countries have similar capital to labour ratios as other EU factories. This may be a comparative advantage for the new accession countries of 2004 to the EU but at present output there also consists of larger light medium and medium cars such as, GM Astra and Vectra; VW Golf and Bora. However, major new investments by PSA and Toyota will be geared to making small cars. The strength of the labour cost argument will diminish as further automation, especially in final assembly, reduces the overall labour content involved in car production and as wage costs rise in the low wage countries. Also, the total labour content at the final assembly stage is 10 per cent of total costs. In the manufacture of engines, systems and sophisticated components it is 5-7 per cent. Hence, the labour cost argument can be overdone. For instance, the process of labour cost increases is already occurring in the accession countries albeit from a low base. According to statistics published by Czech Automotive Industry Associations, since 1997 wages have increased by over 30 per cent in the Czech Republic: 1997, e424.97 per month; in 2000, e547.87 per month; in 2002, e556.72 per month; and, in 2003, e611.72 per month.
Until now lower labour costs have been a factor in attracting foreign investment. The view within the Hungarian auto sector is that the increased wage issue will become serious five years after accession. However, the Slovakians are more sanguine pointing out that wages and salaries are around one-eighth (Table 4) of the western auto industry level. The Czechs are one-sixth, and like the Slovaks, intend that any increase will be gradual and matched by improved productivity and efficiency.
As vertical disintegration, with car firms outsourcing significant activities like engine making, has not occurred in a major way the pursuit of economies of scale has been a major force behind the move to consolidation of companies into huge groups. At the same time the number of separate brands and models has not declined. These conflicting forces have been partly reconciled by the development of common platforms – body structure, suspension, powertrain – on which some product heterogeneity can be offered. The typical aim is 65 per cent of costs in common, and 35 per cent stemming from unique features. This allows consumer preferences to be met whilst controlling unit costs to a level that enables the market to generate profitable prices. Only in this way can car plants in small markets be efficient. Without such integration production in Poland, Slovakia, Hungary, the Czech Republic would not be viable.
The most significant development in the history of the global auto industry may not have been the developments brought together by Ford such as standardisation, inter-changeability and moving production lines but the all-steel car body and eventually the monocoque (without a separable chassis) by the Budd Corporation of the USA. The all steel car body was unequivocally an ‘improvement’ on previous practice: the quality was better, and if sufficient volumes could be reached the unit costs were lower. The key here was ‘sufficient volume’. This would thinly amortise tooling costs and give ultra competitive unit costs. This meant that only big volumes would suffice: to pay for the equipment; and to attain necessary volumes. From that time in the late 1920s the fate of the smaller firm in the mass market became sealed. The dire financial problems currently facing medium size Japanese firms such as Mitsubishi is but the latest manifestation of this. In addition big firms attempted to concentrate output on a limited number of models, retain them in production for a number of years, and attempt continuity of parts between models at a point in time and between successive models over time.
Attempts to break away from monocoque construction have been made but the norm is still the all metal, usually steel, unitary construction car body. However, platform strategies and greater production flexibility have reduced the cost penalty of offering variety in a market wanting more product differentiation. Adaptable platforms, space frames and so on are attempts to free production from the tyranny imposed by unitary construction. The attempt is to retain the cost structure of unitary construction whilst being able to produce a greater variety of products without losing control of unit costs in an increasingly competitive market.
Given the importance of scale economies, how is it that firms of different size survive? This issue is returned to in detail later but essentially it is a matter of ‘product differentiation’, although other factors such as state subsidies, protectionism and inefficiencies in big firms can help the smaller firms to survive. Most of the car firms or brands producing under 1,000,000 units are ‘specialists’ rather than ‘full-line’ producers, trying to charge a price premium sufficient to cover their higher unit costs. Those specialists, such as Alfa Romeo, producing cars less differentiated from the mass producers found it difficult to make profits even as part of a larger group, Fiat. Those with highly differentiated products such as BMW and Mercedes have prospered. However even these have become full line producers and in the case of Mercedes the Daimler Chrysler merger makes it a high volume producer. All firms seek to differentiate their products as success in this regard allows them to reduce the price elasticity of demand for their products, at a given price. Therefore, firms try to identify differentiated niches in the market by offering unique combinations of characteristics or attributes in their cars. This can aid new entry. After all it was the attribute ‘quality of build’ that did much to aid Japanese export growth.
The main independent variables determining car demand are prices and per capita disposable incomes. The estimated price elasticities for car demand as a whole vary from -0.1 to over -3.0, whilst income elasticities vary from 1.1 to 4.2. However, most demand models place the price elasticity co-efficient between -1 and -2. In addition, price elasticities for individual firms vary between -2.0 and -7.0.
Some of these studies use ‘quality-adjusted’ prices as the independent variable. That is, by adjusting for any extra ‘quality’ price is changed. The Japanese used competitive quality adjusted, but not low, list prices to enter most developed markets – especially so in Europe. The high income elasticity of demand for cars explains why with high economic growth in the years 1985-89 UK car demand reached record levels, why it fell back in 1990-1992, grew in 1993-8, and reached record levels in 2002-04.
Other independent variables affecting car demand include advertising expenditures, demographic considerations, the geographic nature of the market such as whether it is urban or rural, and credit terms changes. The latter can generate severe variations in short-term car demand. Population density has an inverse effect on car demand. Therefore, the UK and Europe with their dense populations will not reach US or Canadian car ownership per head figures even if everything else – such as, real incomes, car prices – were equal. There are various reasons for this: for instance in urban areas a good public transport network may exist, while in rural areas a car is almost a transport necessity.
The UK is typical of a developed car market. The low growth in annual car demand between 1964-1984, and the fall in demand in 1991-93 was the result of low per capita income growth rather than a saturated car market. This was made abundantly clear by the rise in car demand in 1985-89, the revived growth in 1993-8, and the boom in demand post-2002. In 1964 per capita car ownership in the UK was amongst the highest in Western Europe, but in the mid 1980s at 2.9 people per car it was one of the lowest. By 2003 it had reached the EU average of 2.0. There are over 30 million cars on UK roads of which 24 million represent a ‘used’ car – a car not new when bought. Of the remainder, two million were bought new by companies. Hence, despite the expansion of car ownership, the privately-purchased new car is still a ‘superior’ good, which although useful as an appliance is not strictly speaking a necessity.
New motorists usually become car owners via the used-car market, and because of this the, for instance, UK car stock is still increasing by around 500,000 per year. This growth in stock is common to all developed countries indicating that even there the market is not mature. The trading-in of one car to buy another means that most car demand is a ‘replacement’ process. It is because of this that car firms are concerned about the used car price for the models it makes as this will affect the cost of change for the buyers when selling their existing car and can affect the initial buying decision. This ‘residual value’ of cars is particularly important in the company car sector and in the luxury car sector. In the former case because of balance sheet and cash flow considerations and in the latter because the products are expensive and rapid price depreciation would make customers hesitant about purchasing a particular luxury car if they felt that its value would drop quickly and, thereby, increase the cost of replacement. When Toyota introduced its Lexus marque in Europe and the USA it deliberately limited the number of cars imported to avoid over-supply, and BMW is a master of this. The result has been strong demand in relation to supply which has maintained residual values and made people keen to buy the cars.
The motor industry is undergoing considerable technological change in both its methods of production and in the nature of the product. The latter involves reducing energy consumption by reducing weight, drag and improving the efficiency of engines, transmission and suspensions. The product and the process of manufacture are becoming more capital intensive. As a result the motor industry is no longer in the mature phase of the corporate life cycle, but is a new immature industry requiring huge inputs of physical and human capital and experiencing active inter-firm competition.
The motor industry has experienced continual structural change. The degree of market concentration is lower than that for production, illustrating the important role of imports. Excess capacity in Europe, taken together with the significance of scale economies, indicates that further structural changes are likely, especially through collaboration. In the mid 1980s the market clearing price was below the long run average costs of production of many firms. Hence, price competition reflected the freeness of the car market but, in contrast, the adjustment of capacity to demand was slow and laborious. The UK industry shed much capacity in trying to improve its efficiency. This has improved and although the recovery has had to contend with ultra competitive conditions, profitability returned in the late 1980s. Finally, the proper context for studying the motor industry is a West European one, if not a global one. The ‘home’ market is Western Europe, whilst a firm’s production is increasingly occurring on a European wide basis, and now worldwide.
With possible overcapacity, the advance of the Japanese, the appearance of new major centres of production like China and India, the effects of the single market and single currency and technological change affecting product and process the consumer will remain sovereign where the motor industry is concerned. The degree of competition will increase in the decade ahead and inefficient firms will have nowhere to hide. The rivalry in the motor industry oligopoly is set to increase.
Professor Garel Rhys OBE, is director at the Centre for Automotive Industry Research, Cardiff University Business School