Recent analyses of trade under imperfect competition have focused on whether markets are integrated or segmented. Integration implies that producers set a single price (or quantity) at the world level and let arbitrageurs determine the distribution of sales to national markets. At the other extreme market segmentation, combined with the as- sumption of constant marginal costs, implies that the game played between firms in one country is unaffected by the game played by the same firms in other countries. Neither approach is entirely satisfactory.

In Discussion Paper No. 277, Research Fellow Anthony Venables analyses a model of imperfect competition between two firms located in different countries. Firms' capacity choices are taken on an integrated or worldwide basis, since a unit of capacity can be used to supply any country. Each firm, therefore, chooses its world capacity at the first stage of the game, with the cost of installing a new unit of capacity assumed to be constant and identical in both countries. Decisions concerning price or quantity produced are taken subsequently for each (possibly `segmented') market, taking capacity as given.
Venables analyses three different equilibria. In the first case, the second-stage game is also played under the assumption of integrated markets. One firm's capacity choice does not change the other firm's allocation of output between markets and thus has no strategic role. In the second case, firms treat national markets as segmented at the second stage and choose the quantity of sales in each market, taking as given the rival's sales in that market. In this case capacity choice does influence, at the second stage, the distribution of the other firm's sales between markets.
The most interesting model is that in which each firm treats national markets as segmented and competes on prices, taking as given its capacity and its rival's prices and capacity. The characterization of oligopoly as price competition subject to capacity constraints is intuitively appealing, Venables argues, and allows rich interaction between markets. A domestic firm will, for example, anticipate that if it cuts prices in the export market this will cause the foreign firm to switch from this market into the domestic firm's home market. This reduces the incentive to export and means that the equilibrium volume of intra-industry trade is lower in this case than in the case of segmented markets and quantity competition.
Venables also presents numerical simulations of the model to illustrate the differences between equilibria. The results suggest that although strategic behaviour induces firms to increase capacity, the magnitude of this effect is small. Venables concludes that representation of equilibrium as the outcome of a capacity game followed by price games yields policy conclusions which are quantitatively but not qualitatively different from those that hold in the case of segmented markets and quantity choice. There are, for example, national gains from the imposition of import tariffs and export subsidies, but these gains are relatively small in the model he examines.

International Capacity Choice and National Market Games
Anthony J Venables

Discussion Paper No. 277, October 1988 (IT)