دامپینگ به عنوان نشانه نوآوری

In the R&D-intensive industries, where technologies change rapidly, an innovative foreign firm may need to export greater than normal quantities to signal the level of the new technology it possesses. We find that such actions lead to sales below cost if the foreign firm has a relatively poor reputation for innovation, has a sufficiently high discount factor or possesses a new technology that significantly cuts its cost. We also show that antidumping reduces the costs of signaling, benefits the home firm, and may raise the profit to the foreign firm in the pre-duty period.

Over the last 25 years use of antidumping (AD) has spread from a handful of countries to more than 60 nations (Prusa, 2001). Recent evidence shows that AD users often target R&D-intensive sectors such as primary metals, chemicals, consumer electronics and mechanical engineering (Niels, 2000). In R&D-intensive industries, in which innovations occur frequently, firms are often unsure of the levels of technology currently used by the rivals. When the domestic firm cannot verify the foreign firm's production cost due to the innovation, the foreign firm may find it worthwhile to export larger than the normal quantities of output to signal its cost, even to the extent that the export price falls below its unit production cost. In this paper we explore this rationale for dumping. More specifically, we develop a model in which a home and a foreign firm compete in the home market over two periods. It is understood that the foreign firm has invested in cost-reducing R&D before the game, and knows the outcome of the investment while the home firm has only imprecise information as to what the rival's new cost is. The game begins with the two firms simultaneously choosing output levels. The home firm then uses the foreign firm's output to update its beliefs about the latter's cost. Given new beliefs, the two firms choose output levels once again, and the game ends. This game has a unique separating equilibrium outcome that satisfies the Cho-Kreps intuitive criterion (Cho and Kreps, 1987). Focusing on this equilibrium outcome, we find that the foreign firm dumps, i.e., exports below average production cost, if it has a sufficiently poor reputation as an innovator. Such might be the case if the foreign firm is a newcomer in the industry or originates in a developing or newly developed country where innovations rarely occur. Dumping also occurs if the foreign firm's discount factor is sufficiently high, or the new technology leads to a significant cost reduction in the sense to be made clear below. In the second half of this paper we apply the model to analyze the impact of antidumping (AD). In the theoretical AD literature it is customary to assume that an AD duty is set equal to the dumping margin. However, this tradition has been questioned by recent empirical studies. For example, the recent survey by Blonigen and Prusa (2001) points out that the legal definition of dumping is almost completely divorced from any economic notion of dumping and that AD has become simply a modern form of protection that improves the competitiveness of the petitioning firm against imports.1 The apparent gap between the theory and the practice of AD stems chiefly from the fact that AD rules have been amended repeatedly over the last quarter century to broaden the applicability of AD laws, especially in the U.S. In particular, the frequent use of “facts available” methods to construct or estimate the prices and costs of exports has made dumping margin determination completely arbitrary. As a result, the U.S. Department of Commerce, which is responsible for dumping margin determination, almost always rules that dumping has occurred, even against firms earning healthy profits from every sale of exports.2 From 1980 to 1992, for example, Commerce ruled that dumping had occurred in 93% of all cases (Irwin, 2002, pp. 114–115). Furthermore, the use of “facts available” methods nearly doubles the average US dumping margins, from around 35% to over 65% (Baldwin and Moore, 1991).3 What these empirical findings emphasize is the arbitrariness of AD margin determination in practice. In this paper we capture this fact by assuming that the actual dumping margin is drawn from a probability distribution function. It implies that the firms do not know the actual duty to be imposed. Since dumping margin determination is a probability distribution, the foreign firm will be saddled with an AD duty with a positive probability regardless of its cost, true or perceived. This assumption enables us to examine the effect of AD with minimal changes in our model. We only add the assumption that between the two periods the home firm files a petition and the home government investigates and announces an AD duty to be imposed in the second period. Then the second-period game is just a one-shot Cournot game under the tariff. More interesting is the possible effect of AD on the first-period or pre-duty game. We show that AD reduces the level of exports necessary for signaling so that first-period sales fall for the foreign firm and rise for the home firm. Reitzes (1993) and Kolev and Prusa (2002) obtain similar results but for different reasons. In their models the first-period game outcome is affected by AD as firms aim to influence future AD duties, but here it is affected as AD changes the cost of signaling. Now a review of the relevant literature is in order.4 The analysis of dumping under asymmetric information begins with Eaton and Mirman (1991). Hartigan (1994) presents a model similar to ours but with two important differences. First, he assumes price competition instead of quantity competition, and that the home firm exits if the foreign firm has low cost.5 Although his analysis cannot be faulted, predatory pricing, according to industrial organization economists, is rare in the real world.6 Absent predation, however, the foreign firm has an incentive to raise prices to soften the second-period price competition, so dumping never occurs, the point also noted by Eaton and Mirman (1991). More significantly, both Eaton and Mirman (1991) and Hartigan (1994) use the definition of dumping based on international price discrimination, while we focus on dumping defined as sales below average cost.7 There are two reasons for this shift in focus. Firstly, dumping as sales below cost has gained prominence over the last 25 years, especially in the United States. According to Horlick (1989, p. 136), for example, about 60% of all U.S. antidumping cases in the 1980s have been based on the allegation of sales below costs. In a more recent study, Clarida (1996) reports that cost-based allegations account for between one-half and two-thirds of all U.S. AD cases he examined. Secondly, trade economists have also shifted attention to sales below cost. The initial interest in this line of research was to show that selling below cost is consistent with rational behavior. For example, the pioneering work of Ethier (1982) showed that, if the industries have different adjustment costs across nations, sales below cost makes sense during shifting demands; see also Davies and McGuinness, 1984, Bernhardt, 1984 and Hillman and Katz, 1986 and Das (1992). Others have developed dynamic non-stochastic models of complete information to show that dumping occurs when exporters try to take advantage of learning curves (Gruenspecht, 1988), compete for scarce future quota rights (Yano, 1989 and Anderson, 1992) or enter the industry in order to discover their true costs (Clarida, 1993). This paper provides a rational for sales below cost under incomplete information. The remainder of this article is organized in 4 sections. The next section lays out the basic assumptions of the model. 3 and 4 present the model without and with AD, respectively. Section 5 concludes the analysis and suggests extensions.

The main idea behind this article is that in industries where technologies change rapidly and continuously firms may not have accurate information about the rivals' production costs. In such cases, a foreign firm that acquires a new technology may find it worthwhile to export more than the normal quantities to signal that it has low cost, even to the extent that the price falls below cost. To formalize such a scenario, we present a two-period game of incomplete information, in which the home and the foreign firm compete in the home market, and we identify the conditions under which sales below cost occurs. We then use the model to analyze the effect of AD protection. In doing so we take a novel approach, assuming that an AD duty is drawn from a probability distribution. The assumption is intended to reflect the arbitrariness of dumping margin determination in practice as emphasized in recent empirical studies. The (probabilistic) protectionist bias in duty determination implies that duty determination is influenced by the perceived efficiency of the foreign firm rather than by realized dumping margins, but as pointed out by a referee, this approach ignores possible strategic moves in order to influence duty determination; e.g., the home firm increasing first-period sales to lower the price to get a higher duty in the second period. Future research may explore this aspect. Our analysis also abstracts from the fact that, once the AD duty is set, the affected foreign firm goes through a review process periodically for the reassessment of the applicable AD duty (see Blonigen and Haynes, 2002 and Blonigen and Park, 2004). This review system makes it possible for the foreign firm to reduce the duty by raising the current-period price. In our model, once the duty t is drawn the second-period game is one of complete information. Each type therefore may try to cut back on exports to lower the AD duty. But it will do so only if it can increase profit beyond the basic Cournot profit of our model. Now, if the foreign firm is the low-cost type, the second-period price is always higher than the first-period price, because there is no need to signal in the second period. Thus, the review board is likely to lower the duty to reward the “good” behavior. By contrast, for the high-cost type the second-period price is always lower than the first-period price. Thus, the high-cost type will be less likely to get a reduction in the AD duty. The upshot is that the low-cost type may be able to get the duty reduced and raise the second-period profit while the high-cost type may not be able to. If so, Lemma 4 is strengthened, so the rest of our analysis may go through. Other extensions are possible. For example, it may be interesting to consider the option for the foreign firm to relocate in the domestic market to avoid an AD duty (see, e.g., Miyagiwa and Ohno, 1995, Blonigen and Ohno, 1998 and Blonigen, 2002), or allow for negotiated settlements leading to voluntary export restraints or price underpinning, common occurrences under AD threats (see e.g., Kolev and Prusa, 2002). Another extension concerns the relationship between AD and R&D investment. The effect of AD on R&D activities has just begun to receive attention; for example, Gao and Miyagiwa (2005) have used the Brander and Krugman (1983) model of reciprocal dumping to address the issue, while Konings and Vandenbussche (2004) have empirically analyzed the impact of AD on firm-level productivity growth. Investigating this issue requires an extension of the model to a three-stage game, in which the foreign firm's R&D decision-making becomes endogenous.