Imperfect Competition The theory of international trade has traditionally analysed trade in competitive markets, where participants are 'small', so prices and the market environment are beyond their influence. The policy relevance of such theory may be limited, since much international trade is conducted by large 'oligopolistic' multinational enterprises. The framework of imperfect competition permits firms to behave strategically, exercising their influence on the market. Extending this analysis to individual economies open to trade and the relations among them is a major step towards realism and applications to current policy issues.

CEPR held a joint workshop with the International Economics Study Group at the Centre on 10-11 September devoted to the policy implications of international trade with imperfectly competitive markets. Recent research has shown that there is scope within this framework for government intervention in trade. The main argument is one of 'rent extraction'. In imperfectly competitive markets, firms can earn supernormal profits or 'rents'. Domestic firms operating abroad benefit their home country, because the rents extracted increase national income. Policy intervention can improve national welfare if it enables home firms to gain a competitive advantage over foreign firms and so extract larger rents. Policy intervention in this situation usually takes the form of export subsidies.

Is 'targeted' export promotion desirable? Gene Grossman (Princeton and CEPR) discussed this issue in the first paper of the workshop, 'Targeted Export Promotion with Several Oligopolistic Industries', written with Avinash Dixit. Rent- extraction arguments are usually formulated in a partial equilibrium framework and ignore the possibility of domestic resource constraints. What difference would this make to the results? The increased rent earned as a result of export subsidies in one sector may be offset by lower rents earned in other sectors. Increased activity in the subsidized sector could bid up the price of domestic resources. This in turn may reduce the output and rent-extraction of the unsubsidized industry.

Grossman took as an example the US high-technology sector, in which several oligopolistic industries shared a common resource - scientists - which was specific to that sector and of which there was a limited (inelastic) supply. If each industry used the same technology, then a blanket subsidy of all the industries in the sector would lead to no change in their output or rent extraction - merely a rise in the wages of scientists! This indicated a need to explore differential subsidies between industries and 'targeting' certain industries in the sector. Where industries are identical, free trade with no subsidy is optimal; where industries differ, the government should target those industries where the increased rent-extraction (per scientist) is highest. Successful targeting, however, requires considerable information, even in the simple model presented. Grossman concluded that the potential benefits from strategic trade subsidies can be exaggerated when only one industry is considered in isolation.

The discussion made it clear that these results depended crucially on the assumption that the factor in limited supply was not employed outside the oligopolistic sector. Otherwise, the government could simply subsidize the whole rent-generating sector at the expense of the competitive sector of the economy. It was also suggested that in the case of US high technology a key factor had been the ability of American companies to 'steal' scientists from abroad: the importance of such labour mobility could reduce the effects of an inelastic domestic supply.

The presentation by John Sutton (LSE and CEPR) was drawn from a study 'Europe and Protectionism', undertaken by the Royal Institute of International Affairs for Department of Trade and Industry. Sutton saw two main propositions in the French memorandum to the EEC on European industrial policy. The first was an argument for R and D subsidies, either general or targeted on specific industries. Sutton noted that in practice most policies to encourage R and D applied to individual industries, and he explored some of the theoretical reasons for such 'targeting'.

Sutton argued that his theory of 'natural oligopolies', formulated jointly with Avner Shaked, provides a framework for understanding these policies. They start from a situation in which firms can employ product differentiation as well as pricing decisions as part of their strategy. The resulting equilibrium can involve a 'natural oligopoly', with only a few firms producing. The key point is that this equilibrium is asymmetric: a few firms produce, while the rest are losers who are inactive. Industrial policy matters here, since it can create situations where domestic firms are active at the expense of foreign competitors. The Airbus project may be an illustration. By using a mixture of subsidies and procurement policies, the Europeans were able to bring about the withdrawal of McDonnell Douglas.

The second argument in the French memorandum was the need to allow European firms to cooperate on large ventures, even at the expense of competition policy. Sutton considered the case for cooperative R and D ventures. Recent experience here has not been promising: European governments have not shown willingness to encourage their own firms to share information with other European firms. Furthermore, Sutton argued that the failed Grundig/Thomson-Brandt merger showed the French and German governments to have totally different priorities in balancing anti-trust policy against increased cooperation.

Participants queried whether strategic policy-making was theoretically sensible in the 'natural oligopoly' model. Technical problems might mean that no equilibrium need exist. Discussion also brought out the need for European countries to cooperate, if only in their policies to attract foreign direct investment. As the Nissan case showed, foreign firms can play off one country against another and thus charge a high price for bringing in new technologies and employment.

What competitive advantages do such multinational companies enjoy? Henrik Horn (Stockholm) addressed this question in 'Multinational Corporations and Control of Operations', written jointly with Wilfred Ethier (Pennsylvania). There are two conflicting answers to this question. On the production side, multinationals may possess technological advantages and benefit from economies of scale. On the organizational side, there are 'managerial' diseconomies, as the work of Oliver Williamson suggests: firms become more hierarchical as they expand; and as the number of management layers increases, there is a loss of managerial control. In the case of multinationals, Horn pointed out an additional managerial diseconomy, the 'interface' problem posed by international communication and coordination.

The large OPEC surpluses of the 1970s led to the proposal that the industrial countries should exploit their monopoly power in capital markets by taxing interest payments to OPEC. David Ulph (Bristol and CEPR) analysed this idea in 'International Trade in Resources and Capital under Various Market Structures,' written jointly with Alistair Ulph (Southampton and CEPR). Sweder van Wijnbergen had earlier considered such a tax but had assumed that the oil market was competitive and that OPEC could not act strategically. Ulph explored this proposal within the framework of a model in which OPEC exerts monopoly power in the oil markets by choosing the rate of extraction of oil and the level of oil prices, while industrial countries can exert their monopoly power in the capital markets through an interest tax. Ulph argued that van Wijnbergen's formula for the optimal interest tax could be misleading unless modified to take account of an imperfectly competitive oil market. The sign of this correction, however, was ambiguous and depended on the structure of OPEC pricing.

The discussion concentrated on some of the technical difficulties within the model. The model gave no definite answer as to how the 'rent' was divided between the industrial countries and OPEC. In the paper, the authors had merely considered the best choice of interest tax given a particular set of OPEC tariffs. In order to make the OPEC tariff endogenous, a far more complex model might be needed.

One of the primary motives for foreign direct investment is alleged to be 'tariff-jumping'. James Brander (UBC) and Barbara Spencer (Boston) analysed this process in 'Direct Foreign Investment with Endogenous Tariffs and Taxes'. The firm can choose to produce at home for export, but such exports are subject to tariffs. Direct investment abroad avoids such tariffs but involves tax payments to the host country. When the host country can be relied upon not to alter its taxes once the investment decision is made, the decision involves only a simple comparison based on existing taxes and tariffs. Brander and Spencer investigated the more complex case where a government's tax or tariff policy may be altered after the investment decision is made. They found that unemployment in the host country has a crucial effect on the outcome. In the full employment case, the host country would choose a tax on output and a tariff on imports which would make the firm indifferent between producing at home and exporting, or producing in the host country. There would be no advantage to tariff-jumping. In the case of unemployment, however, there is an incentive for the host to charge a lower tax on output to stimulate the firm's output and employment. With unemployment, 'tariff-jumping' becomes the equilibrium solution.

In the discussion, objections were raised to the nature of the output tax. John Sutton argued that in practice taxes on firms took the form of taxes on profits, not on output. Brander and Spencer needed an output tax in their model, because this gave the host government a trade-off between higher tax revenues and lower output and employment in the unemployment case. A profits tax would have no effect on employment in their model and would not offer such a trade-off. In reality there are other constraints, such as 'reputation effects', on how far a country can alter its taxation policies if it wishes to encourage future investment. Huw Dixon (Birkbeck) argued that the host government could overcome the problem of credibility by offering firms lump- sum inducements to invest which were paid prior to output being produced.

Tony Venables (Sussex and CEPR) presented the final paper of the workshop, 'Does Protection Reduce Productivity?'. This analysed the impact of protection on productivity when markets are oligopolistic. Firms used their commitments of fixed factors such as capital in order to influence the equilibrium in the product market. Because capital is used strategically, the resulting mix of inputs need not be the cost-minimizing ones. The nature of the distortion which results depends on precisely how firms view the strategic interaction in the market. In one case ('Cournot conjectures') there would be overcapitalization and hence high labour productivity; while in another case ('Bertrand conjectures') the outcome would be undercapitalization with low labour productivity.

Given this oligopolistic framework, how does protection affect labour productivity? Venables examined several possibilities. He found that in the Cournot case, quotas would reduce labour productivity and welfare, and that with Bertrand conjectures a tariff would raise both labour productivity and welfare. In general the effects depended very much on the nature of the conjectures held by firms.

From discussion of the workshop papers several themes emerged: the conclusions of multi-stage game theory, or perfect- equilibrium models are quite sensitive to the order in which moves are made, so that policy conclusions must be treated cautiously; assumptions about entry to oligopolistic markets are also crucial, as many of the results about the desirability of protectionist policies in such markets turn on the existence of monopoly rents; and some equilibrium concepts are of much more practical relevance than others.


The final session was devoted to a round-table discussion of the issues for future research. Stephen Smith (Department of Trade and Industry) set out current trade policy issues to which the methods used in the workshop could be applied. Michihiro Ohyama (Keio University) examined current Japanese research in this area.
It is possible to construct a model of an open economy and calculate the prices which would clear all product and resource markets in the economy. Then one can introduce a policy change, such as a tariff reduction, recompute the new equilibrium prices and examine the impact of the policy change, taking into account its effects across markets in the economy. It was argued that such 'computable general equilibrium' (CGE) models of international trade had an important role in evaluating policy changes. They could predict the magnitude and not merely the direction of policy-induced changes. A recent CGE model of the Canadian economy constructed by Richard Harris had included many elements of imperfect competition, notably non-constant returns to scale. The model predicted welfare gains from (multilateral) free trade much greater than those derived from CGE models which assumed competitive markets. It was agreed there was a need for developing appropriate CGE models for the UK and the European Community which could be applied to the analysis of policy.

Gene Grossman discussed the NBER research on strategic behaviour and trade policy, which was designed to bring together industrial economists and those in government. Meetings were organized around case studies of particular industries, with the economists bringing to bear the models which seemed appropriate for that industry. Peter Neary, Director of the CEPR International Trade programme, argued that this had great potential in the UK, where so little case-study work was done by economists.

The workshop followed the annual IESG conference, held at Sussex University. This focused on the theory of imperfect competition in international trade. The first three papers at the conference were all concerned in different ways with extending recent analyses of international trade in differentiated products. In the first, 'Growth, Trade and Income Distributon under Increasing Returns,' Paul Krugman (MIT) set forth a unified framework for the analysis of growth and trade in a situation where some goods were produced with increasing returns to scale. He divided the effects of growth and trade between 'terms of trade' effects, involving changes in the distribution of income, and 'scale' effects due to increases in the size of markets.

Conventional trade theory has some difficulty accounting for the existence of intra-industry trade, as well as explaining why the volume of international trade tends to be larger between economies which are similar. Henryk Kierskowski (Graduate Institute of International Studies, Geneva), in a paper written with Rodney Falvey, explored how the unequal distribution of incomes and quality differences between goods might account for these phenomena. Elhanan Helpman (Tel Aviv), in 'Imperfect Competition and International Trade: Evidence from Fourteen Industrial Countries', examined how successfully recent models of international trade under imperfect competition could account for the observed patterns of intra-industry and international trade.

Trade theorists have also attempted to explain within the framework of their theory the existence of multinational enterprises James Markusen (Western Ontario) in a paper written with Ignatius Horstman, 'Strategic Investments and the Development of Multinationals', explored the role of tariffs, transport costs and 'sunk' costs in creating multinational enterprises. He argued that when tariffs, transport costs and 'firm-specific' costs were large relative to 'plant-specific' costs, a foreign firm could pre-empt the entry of home firms in smaller countries.

George Norman (Reading) in 'Intra-industry Foreign Direct Investment: its Rationale and Trade Effects' prepared jointly with John Dunning (Reading), drew parallels between the possible explanations of intra-industry trade and foreign direct investment.

Current theory focuses on three factors in accounting for multinational enterprises: ownership advantages, such as patents; locational factors, such as tariffs; and the advantages of internalizing international transactions. In 'A General Equilibrium Theory of Trade, Foreign Direct Investment and Multinational Corporations,' Wilfred Ethier (Pennsylvania) attempted to incorporate these considerations in a general equilibrium model giving rise to multinational enterprises.
Uncertainty played an important role in the internalization of transactions in Ethier's model.

In 'Trade in Differentiated Goods and the Shifting of Protection across Sectors,' David Greenaway (Buckingham) and Chris Milner (Loughborough) examined how one might analyse the 'burden' or incidence of protection in a model with a non-traded-goods sector.

Kelvin Lancaster (Columbia), in 'Multiproduct Defensive Monopoly in an Open Economy', discussed how a firm producing more than one product could set prices and choose its product specification so as to inhibit the entry of foreign and domestic firms.