The theory of trade policy towards imperfectly competitive industries has shown that it may be in the national interest to intervene in imperfect markets. Specifically, the literature has demonstrated three potential sources of welfare gains from policy intervention. First, the total profits of domestic firms can be increased through policies which induce more aggressive behaviour on the part of domestic firms towards foreign firms (thereby shifting profits from foreign to domestic producers) and less aggressive competition between domestic firms themselves. Second, policy can be designed to benefit consumers by reducing domestic prices and increasing consumer surplus. This can be achieved through policies which encourage domestic or foreign firms to sell more in the domestic market. Third, the range and volume of products available to domestic consumers can be widened, through policies which encourage entry of new firms.

 The precise nature of the policies needed to reap these potential benefits is, however, uncertain. For example, to shift export market profits from foreign to domestic firms, an export subsidy is called for if competition in the industry is in quantities (Cournot); but an export tax is appropriate if competition is in prices (Bertrand). Moreover, the different sources of potential gains may call for quite different policies. Thus, to pool profits, domestic firms should be encouraged to produce less; to achieve lower domestic prices, they should be encouraged to produce more. Similarly, to encourage new entry by domestic firms, domestic producers should be subsidized; to encourage new entry by foreign firms, domestic producers should be taxed.

Since we cannot reach general conclusions about the qualitative nature of optimum policies on the basis of economic theory alone, we need empirical studies of the industries concerned. Such research could establish broad, empirical regularities which might resolve some of the ambiguity. Alternatively, it could make it clear that no regularities exist, so that optimum policy towards a particular industry would require detailed knowledge of the nature of competition and the parameters of demand and cost functions in the industry.

This paper is a contribution to the small but growing body of such empirical studies. Previous studies include one by Dixit on optimum trade policy for the automobile industry, studies by Baldwin and Krugman on aircraft and semi-conductors, and studies by Smith and Venables on several UK and Continental industries. In our paper, we study two Norwegian industries - the ski industry (which is primarily a domestic industry with substantial import competition) and the Caribbean cruise shipping industry (which, from a Norwegian point of view, is a pure export industry). In the latter, profit-shifting and profit-pooling are the relevant policy objectives; in the former, consumer prices and product variety are the key issues.

In the Norwegian ski market, there are two domestic firms (one of which is the result of a 1986 merger of two producers). These two firms produce three brands of skis, which represent 55% of the market, while a number of foreign firms supply the rest of the market. Exports account for about 25% of Norwegian ski sales. We know little about the export markets; we do know, however, that the Norwegian share of most of these markets is very small. Arbitrarily, therefore, we have assumed a joint Norwegian market share of 10% divided among three brands of skis. The rest is assumed to be supplied by eight identical foreign firms. Thus, there are more active foreign firms abroad than in Norway, so there is a potential for new entry in the domestic Norwegian market by existing foreign producers. The industry is in a short-term equilibrium in which the largest domestic producer (the merged firm) experiences significantly higher costs and so is running at a loss. Exit and entry of new firms is therefore possible and even likely. The critical factor is whether the recently merged company will be able to cut fixed costs sufficiently to survive - if it does not, the likely outcome (in the absence of policy intervention) is for two of the domestic brands of skis to be replaced by new foreign brands.

We set up a numerical simulation model for the industry, in which the products of different firms are assumed to be less-than- perfect substitutes, and where the firms are assumed to use prices as strategic variables (so competition is of the Bertrand variety). The model is calibrated using actual 1985 data.

We use the model to simulate the effects of production subsidies to domestic firms. If policy is judged on the basis of domestic welfare (the sum of domestic consumer and producer surplus, less subsidy payments) we find the optimum subsidy to be 15%. If we assume that there is no entry or exit of foreign firms from the Norwegian market, the merged firm will maintain production of both brands of skis if the subsidy exceeds 10%, or if the subsidy is 10% and costs are reduced. If there is free entry and exit of foreign firms, there are four possible outcomes. If the subsidy does not exceed 5% and the cost reduction achieved by the merged firm is sufficiently large, the merged firm may stay in business and all existing foreign firms may stay in the Norwegian market. Second, if the merged firm achieves substantial cost reductions and the subsidy exceeds 5%, the merged firm may remain and one foreign firm disappear. Third, if the merged firm cannot cut costs and the subsidy is 5-10%, part of the merged firm may be closed and one new foreign firm come in. Finally, if there are no subsidies and no cost reductions, the merged firm may be closed down altogether and be replaced by three new foreign firms. In the presence of free entry and exit, the optimum subsidy is therefore 5%: a larger subsidy will force at least one foreign producer out of the domestic market and consumer surplus will fall, since Norwegian consumers will no longer have access to this brand of ski.

In the absence of information on entry conditions (or more generally, on the fixed costs of new firms which might enter the market), the government could decide the optimum subsidy on the basis of a maximin criterion, i.e. it could choose that subsidy which makes the minimum domestic surplus (the surplus associated with the most unfavourable entry conditions) as large as possible. Under the maximin criterion the optimum policy depends on the reduction in fixed costs achieved by the merged domestic firm. With no cost reduction, the subsidy is 15%; with a substantial cost reduction, it is 5%. This produces perverse cost incentives, however: the merged Norwegian firm is better off with a 15% subsidy and no cost reduction than with a 5% subsidy and substantial cost savings.

In the Caribbean cruise shipping market, the issue is future capacity. Three firms, two Norwegian and one US, together have 78% of the Miami market. The US firm has contracted three new medium-sized vessels; one of the Norwegian companies has contracted one large ship for delivery in 1988 and holds an option for another in 1991; the other Norwegian firm is contemplating ordering a "supership" carrying 5000 passengers. If all these vessels were built, carrying capacity would increase by 150% by 1991. Policy enters the problem because of high Norwegian labour costs: firms have argued that they will be unable to maintain their position in the market unless they are given permission to register their ships under cheaper flags.

Again we set up a fairly simple numerical model and use it to simulate market equilibria. Since the issue is capacity, the relevant market game involves strategic choices of quantity and is therefore of the Cournot type.

If the US contracts for the construction of new ships are firm and if these vessels cannot be used other than in the Caribbean cruise market, the game is one in which only the two Norwegian firms are active players (the US firm is already committed to its new vessels and so does not have a strategy to play). Policy has its principal effect on the profits of domestic firms. If the US contracts are firm, lower costs are not in the Norwegian interest: with lower costs, both Norwegian firms are encouraged to construct more vessels and their joint profits will fall. We show that the entire cost saving from using foreign crews in this case is passed on to foreign consumers - leaving the Norwegian firms with a joint net loss of $1 million.

This result hinges crucially on the lack of a response by the US company. If the American contracts are not firm (or if the new vessels could be used in other markets), lower manning costs can give Norway a gain over and beyond the cost saving itself. With no reduction in Norwegian labour costs we find that it is profitable for the US firm to put two of its three contracted vessels into the Miami market; a commitment to do so will therefore be credible. The effect will be to block the projected Norwegian supership.

With lower Norwegian costs, however, the US firm will no longer be able to keep one of the Norwegian firms from expanding; as a result, a commitment to increased capacity on the part of the American firm is not credible and the optimum policy for the US firm will be to cancel their contracts or to use their vessels for other purposes. The net effect is to shift profits from the US to the Norwegian firms.

Generally, our analysis clearly shows that the empirical regularities which could resolve theoretical ambiguities cannot be established. The optimum policies towards the two industries depend critically on cost and entry conditions; even the qualitative nature of the policy implications may be reversed when these conditions are changed.

The two examples also show that the information needed to design optimum policies is of a kind which the government is unlikely to have, but which the market agents may possess. This asymmetry of information may produce perverse cost incentives.

All told, therefore, we may be better off if we leave imperfectly competitive industries alone.