Multinational Learning under Asymmetric Information.

AuthorJain, Neelam
PositionResearch information

Neelam Jain [*]

Leonard J. Mirman [+]

The aim of this paper is to analyze the competition between a multinational and the incumbent firm in a foreign market under asymmetric information about demand and unobservable outputs. It is shown that the incumbent firm increases its production in the first period to signal to the multinational that the demand is low. The multinational reduces its output in the foreign market in order to signal-jam. In addition, the multinational increases its production in the other market. However, total production of the multinational is lower. Implications for research and development expenditure by the multinational are examined.

  1. Introduction

    Multinationals seem to be ubiquitous, increasing their presence in many markets. It has long been believed that multinational firms exist in order to exploit economies of scale in production, that is, declining marginal costs. However, if that were the case, these "efficient" firms would take over the entire market. Yet such behavior does not seem to occur. Instead, multinationals seem to be encountering increasing competitive pressures as foreign markets mature. It is the premise of this paper that multinationals, like all other firms, suffer from decreasing returns to scale, that is, increasing marginal costs.

    In this paper, we take a different view of multinational firms. Generally, there are political and regulatory reasons for market structures to be different in each country, and politics is often used by larger firms to enter new markets. However, once in a market, multinational firms can exploit their flexibility derived from supplying several markets. At the same time, multinational firms suffer from lack of knowledge about the fundamentals of local markets. Indeed, the local firms usually have better information about their own markets than an "outside" multinational. [1] We study the behavior of multinational firms focusing on the market structure in various markets and informational asymmetries that characterize the foreign markets. In particular, we examine how the multinational firm can use its flexibility over several markets in order to learn about foreign markets and manipulate the beliefs of the local firms.

    Very little theoretical work has been done to analyze the effect of this informational advantage of the local firms in a dynamic model, especially in the context of international economics (Brander 1995 surveys static models based on asymmetric information in the context of trade policy). The exceptions are Eaton and Mirman (1991) and Horstmann and Markusen (1996). The former analyzes a two-period model of learning. However, it does not deal with multinational firms, and more important, the informational assumptions made there are not suited to the multinational analysis. Horstmann and Markusen allow learning by multinationals to be instantaneous and do not allow any competition between the multinational and the local, informed firm.

    In order to study the effects of informational differences on firm behavior and the implications of the market structure in new markets, we examine duopolistic competition under asymmetric information where the uninformed firm is a multinational that operates in more than one market.

    Specifically, we consider a multinational that produces a good in its home market as a monopolist and in a foreign market as a duopolist. The home demand conditions are common knowledge except for the random shocks, while a foreign demand parameter is known only to the foreign firm. The two markets are segmented in that the prices are determined independently. However, the multinational's costs of production in the two countries are interdependent, reflecting the main feature of multinational firms, namely, the existence of a firm-specific input that can be utilized in production across locations (Caves 1982). One source of such economies and externality is the research and development (R&D) expenditure that contributes to the production both at home and in the foreign market. This element has been captured through the cost function of the multinational for convenience. [2] We also present a model explicitly showing the relationship between the cost function and the multinational's expenditure on R&D as a gu ideline for other applications of the model.

    We analyze a two-period model in order to study learning by the multinational. The multinational is assumed to have a prior on the unknown parameter of the demand function. It updates this prior by using the observed price and its own output choice. We assume that the firms observe only their own output. Thus, the updating by the multinational is based on conjectures about the foreign firm's first-period output. In addition, the inability to observe the other firm's output also implies that the foreign firm cannot infer the multinational's actual posterior.

    The dynamic aspect of the model, combined with the assumption that the first-period outputs are not observable, adds interesting and complex elements of information manipulation by the two firms to the static or myopic analysis. The foreign firm has the incentive to distort its first-period production to prevent the multinational from learning about the true state of demand and thus enjoy greater profits in the second period. Thus, the informed firm "signals." However, since the quantities produced are not observable, the foreign firm cannot signal as effectively as it would like. Instead, unobservable quantity has the effect of providing the multinational with an incentive to signal-jam by producing less. The equilibrium is a result of these opposing tendencies.

    Specifically, the main findings of this paper are that the foreign firm increases its first-period output to make the multinational believe that the demand is low, while the multinational reduces its output to make the incumbent believe that it believes that the demand is high. The multinational increases its output in other markets. However, total output of the multinational falls. We illustrate applications of the model by showing that the level of R&D expenditure undertaken by the multinational also changes as a result of learning. In particular, we show that the first-period R&D expenditure falls because of signal jamming. The increased output of the foreign firm in the first period can be viewed as predatory dumping (in the sense of Eaton and Mirman 1991) [3] since it is intended to discourage the multinational from producing more in the second period.

    Previous work on signal jamming has considered firms that compete in the same market as well as in segmented markets (Mirman and Urbano 1988; Eaton and Mirman 1991). This work differs from those in introducing uninformed multinationals and thus allowing us to discuss the implications for the form of multinational entry into foreign markets. This paper is most closely related to Mirman and Urbano (1988), Eaton and Mirman (1991), and Horstmann and Markusen (1996). Although seemingly very similar, the question asked in Eaton and Mirman is very different. While they show how a firm with private information in one, monopolistic market can enjoy a strategic advantage in other, duopolistic markets, this paper shows how a firm faced with an information disadvantage in a duopolistic market manipulates its output in all markets it operates in to enhance its learning. In terms of analysis, asymmetric information in the duopolistic market, coupled with the unobservability of outputs, leads to the possibility of informat ion manipulation not only by the informed firm (as in Eaton and Mirman) but also by the multinational. Now the informed firm must conjecture the multinational's beliefs about the unknown parameter. Moreover, Eaton and Mirman's dumping result (overproduction by the informed firm in the first period) is only one of the possibilities due to nonuniqueness of equilibrium. [4] Thus, it is possible that a counterintuitive equilibrium (the opposite of what they claim) can also be supported in their model. In contrast, in this paper we show that when markets are integrated, the nonuniqueness problem does not arise and that dumping occurs in equilibrium. Thus, the properties of equilibrium are fundamentally different, depending on which market is characterized by the informational asymmetry. Mirman and Urbano (1988) study integrated markets as we do. We use that model as a basis to study multinationals.

    This work contributes to the theory of multinational enterprises in the spirit of Horstmann and Markusen (1996). These authors examine the mode of entry of the multinational in the face of an information disadvantage. They compare foreign direct investment with hiring a local firm to serve as its agent. However, they do not allow any competition between the uninformed multinational and the agent who is more informed. Our paper fills that gap and provides a different approach to studying the mode of entry.

    This paper contributes to the existing literature on signaling, signal jamming, and experimentation (Grossman, Kihlstrom, and Mirman 1977; Milgrom and Roberts 1982; Matthews and Mirman 1983). In our model, the decisions are made simultaneously, and the output choices of the other firm are not observable. Together, these features lead to interesting differences in the results from these other models. First, the assumption of unobservable outputs leads the multinational to reduce its output in response to signaling by the incumbent. Thus, both firms distort their outputs in the first period to affect the flow of the information in their favor. Second, since the output choices are simultaneous, the value of the second-period game depends on the information available to the multinational, unlike the Matthews and Mirman model, where the entry decision decides the value of the second-period game. Finally, in our paper, although there is an opportunity for the multinational to experiment, that...

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT