تجارت بین الملل با حالت درونی از رقابت در تعادل عمومی

This paper endogenizes the extent of intra-sectoral competition in a multi-sectoral general-equilibrium model of oligopoly and trade. Firms choose capacity followed by prices. If the benefits of capacity investment in a given sector are below a threshold level, the sector exhibits Bertrand behavior, otherwise it exhibits Cournot behavior. By endogenizing the threshold parameter in general equilibrium, we show how exogenous shocks such as globalization and technological change alter the mix of sectors between “more” and “less” competitive, or Bertrand and Cournot, and affect the relative wages of skilled and unskilled workers, even in a “North–North” model with identical countries. Highlights ► We endogenize the mode of competition in a general-equilibrium oligopoly model. ► Technology and factor prices affect if sectors exhibit Bertrand or Cournot behavior. ► Globalization alters the mix of sectors between “more” and “less” competitive. ► It also affects the relative wages of skilled and unskilled workers. ► All this holds even in a “North–North” model with identical countries.

One of the oldest themes in international economics is that larger or more open economies are likely to be more competitive. This notion has been formalized in a variety of ways. Partial-equilibrium models of oligopoly have shown that trade liberalization or increases in market size generate a competition effect which reduces output and profit margins of incumbent firms and may make it harder for them to sustain collusion in repeated interactions.2Krugman (1979) showed that competition effects can also arise in a general-equilibrium model with differentiated products, free entry and general additively-separable demands. However, most subsequent studies of trade in general equilibrium have used the Dixit–Stiglitz model of monopolistic competition with CES preferences, which implies that firms' price-cost margins, and hence the degree of competition in the economy, are independent of market size. Melitz (2003) introduces firm heterogeneity into such a framework, and shows that trade liberalization favors more efficient firms at the expense of less efficient ones. However, this is a selection effect rather than a competition effect, since in the Melitz model each individual firm always has the same mark-up. Melitz and Ottaviano (2008) show that this can be relaxed in a model with a quadratic demand system similar to the one we use in this paper. However, since they assume that preferences are quasi-linear, they do not model the impact on factor markets. Much remains to be done to understand the implications of allowing firm mark-ups and the degree of competition to be endogenous in a general-equilibrium model. In this paper we provide a new explanation of how exogenous shocks such as growth or trade liberalization can lead to changes in the degree of competitive behavior throughout the economy. We do this by embedding a model of firm behavior along the lines of Kreps and Scheinkman (1983) in a framework of general oligopolistic equilibrium presented in Neary, 2003a and Neary, 2007. In the model of Kreps and Scheinkman, as simplified and reduced to an equilibrium in pure strategies by Maggi (1996), firms producing differentiated products first invest in capacity and then set their output prices. Although firms always compete in a Bertrand manner in the second stage of the game, the outcome may or may not resemble that of a one-stage Bertrand game. It will do so if the cost savings from prior investment in capacity are below a threshold level.3 By contrast, if the cost savings exceed the threshold, then the outcome is “as if” the firms were playing a one-stage Cournot game. Since it is well-known that, other things equal, Bertrand behavior is more competitive than Cournot (implying higher output and lower mark-ups), this model implies that the nature of technology in a sector is an independent determinant of the extent of competition there.4 All previous applications of this approach have considered only a single sector in partial equilibrium.5 Moreover, they have assumed that the crucial threshold parameter is exogenous. By contrast, a major contribution of our paper is to show that it is endogenous in general equilibrium. As in previous work, the threshold parameter depends on a technological component, which varies across sectors. In addition, it depends on a cost component, which is linked to economy-wide factor prices. This is because investing in capacity installation is assumed to require a different factor mix from routine production. Specifically, we assume that investment uses skilled workers while production uses unskilled. (Our results are qualitatively unchanged as long as capacity installation uses skilled labor more intensively than production.) This assumption is supported by much of the empirical literature on technology, trade and wages, where the distinction between production and non-production workers is assumed to coincide with that between unskilled and skilled workers. See for example Berman et al., 1994, Hanson and Harrison, 1999, Feenstra, 2003 and Bernard et al., 2008. Its implications have also been explored in a number of other theoretical studies. 6 An immediate implication of this view of the technology of production is that shocks to an equilibrium, such as trade liberalization, affect factor prices and therefore alter the cost component of the threshold parameter. As a result, such shocks change the mix of sectors between “more” and “less” competitive, or, equivalently, between those exhibiting Bertrand and Cournot behavior. The model thus suggests a new mechanism whereby exogenous changes can affect the degree of competition in an economy. It also throws new light on the impact of trade liberalization and technological change on the relative wages of skilled and unskilled workers. To set the scene, we begin by considering the model in the absence of oligopolistic interaction. Section 2 examines the case of a closed economy where each of a continuum of sectors has only a single firm. We show how the level of investment in capacity is chosen and in Section 3 illustrate the determination of equilibrium. Section 4 extends this model to an integrated world economy with home and foreign firms active in each sector, and explains how the mix between “Bertrand” and “Cournot” sectors is determined. Section 5 considers the effects of shocks to the initial equilibrium. Finally, Section 6 compares the autarky and free-trade equilibria, and shows how opening up such a world to trade affects the degree of competition and the distribution of income, even though the two countries in our “North–North” model are identical.

This paper has presented a new model which integrates some key features of industrial organization and general-equilibrium trade theory, and highlights a new mechanism whereby relative wages and the nature of competition within sectors are affected by exogenous shocks. The model extends to general equilibrium the work of Maggi (1996), building on Kreps and Scheinkman (1983), which predicts that firms will exhibit Bertrand or Cournot behavior depending on the costs of investing in capacity, where capacity serves as a commitment device to sustain higher prices. Maggi looked at normative questions only. In particular, he showed that the Kreps–Scheinkman approach resolves the apparent conflict between Brander and Spencer, 1985 and Eaton and Grossman, 1986, who proved respectively that the optimal export subsidy is positive in Cournot competition and negative in Bertrand competition.31 Here we have focused instead on positive questions, including the effects of exogenous shocks such as changes in relative factor endowments and trade liberalization on the distribution of income and on the margin between more and less competitive sectors. Of course, trade liberalization leads directly to more competition when the number of firms in each sector rises. However, just how much more competition is induced depends on whether firms are able to sustain prices above the Bertrand level. Our model shows that this in turn depends both on technology and on factor prices, with the latter determined endogenously in general equilibrium. When sectors differ in their requirements of unskilled labor, and when goods are more differentiated within sectors so inter-firm competition is less intense, trade between similar economies raises the relative return to skilled labor, making it more difficult to sustain higher prices through investment in capacity and so leading to greater competition throughout the economy. By contrast, when sectors have relatively similar unskilled-labor requirements, and when competition between firms is relatively intense, opening up to trade lowers the skill premium thereby reducing the cost of investing in capacity and sustaining higher prices. The model also exhibits other novel features. It shows that the effects of exogenous shocks to factor endowments and technology differ greatly depending on the relative importance of changes at the intensive or extensive margin in their effects on the demand for unskilled labor. And, although preferences are non-homothetic and fixed costs play an important role, the fact that the fixed costs are endogenous implies that the economy exhibits constant returns to scale in the aggregate, in striking contrast to standard trade models with exogenous fixed costs. More work is needed to explore the robustness of these and other properties of the model to alternative specifications of the workings of factor markets and the ways in which technology and factor prices interact to affect the nature of competition between firms.