سرمایه گذاری مستقیم خارجی و سرریز از طریق تحرک کارگران

We analyze a model where a multinational firm can use a superior technology in a foreign subsidiary only after training a local worker. Technological spillovers from foreign direct investment arise when this worker is later hired by a local firm. Pecuniary spillovers arise when the foreign affiliate pays the trained worker a higher wage to prevent her from moving to a local competitor. We study conditions under which these spillovers occur. We also show that the multinational firm might find it optimal to export instead of investing abroad to avoid dissipation of its intangible assets or the payment of a higher wage to the trained worker.

The last two decades have witnessed an important change in the attitude of host countries towards multinational enterprises (MNEs). Most countries have removed their barriers to foreign direct investment (FDI) and have actively encouraged investment by foreign firms. Advocates of these policies claim that MNEs generate spillovers which benefit the host economy. Such spillovers may take several forms.1 First, there may exist backward and forward linkages between foreign affiliates and local firms (Lall, 1980 and Rodriguez-Clare, 1996). Second, foreign affiliates may increase local firms’ productivity through “demonstration effects”. For example, domestic competitors might successfully imitate technological innovations introduced by MNEs (Mansfield and Romeo, 1980 and Blömstrom, 1986). Third, spillovers arise when subsidiaries of foreign firms train local employees who later join local firms or set up their own companies, bringing with them all (or part of) the technological, marketing, and managerial knowledge that they have acquired. In this paper, we focus on this last form of spillovers, and we present a model in which technological spillovers arise due to the mobility of workers who have been trained by MNEs. Our main purpose is to study the conditions under which such spillovers occur. The fact that MNEs undertake substantial efforts in the education of local workers has been documented in many instances (e.g., ILO, 1981 and Lindsey, 1986), and empirical research seems to indicate that MNEs offer more training to technical workers and managers than do local firms (Chen, 1983 and Gerschenberg, 1987). In early stages, affiliates rely more intensively on expatriates, but subsequently they tend to replace them with (cheaper) local workers who have been properly trained in the meanwhile (UNLTC, 1993). However, evidence on spillovers due to workers’ mobility is scarce and far from conclusive.2 An early study by Behrman and Wallender (1976) shows that, while labor mobility is important in certain circumstances, it is minimal in others. Gerschenberg (1987) analyzes MNEs’ activity in Kenya. He concludes that mobility is lower for managers employed by MNEs than for those employed by local firms. In a study of the Taiwanese economy, Pack (1993) finds that labor mobility from MNEs to local firms is important and that often trained managers leave MNEs to run their own businesses. Aitken et al. (1997) study the effect of inward FDI on the wages of the local workforce in three different host countries. In Mexico and Venezuela, inward FDI increases the wages of the workers in MNE affiliates, but has no effect on the wages of local firms’ workers. In the US, inward FDI results in higher wages both in MNE affiliates and in local firms. Indirectly, this might show the existence of technological spillovers through labor mobility in the US, whereas in Mexico and Venezuela such labor mobility might be inhibited by either the higher wages paid by the MNEs or a larger technology gap. Our paper provides a formalization which is consistent with these findings. We build a model where a MNE trains a local worker to run its subsidiary. Later, the MNE and a local firm compete for the services of the trained worker. As a result, the MNE manages to keep the worker only if it offers better conditions than the local firm. Spillovers from foreign direct investment can take two forms. Technological spillovers arise when the trained worker is hired by the local firm. Pecuniary spillovers arise when the MNE pays the worker a higher wage to prevent her from moving to the local competitor. We find that the so-called “joint profit” effect (or “efficiency” effect) plays an important role in determining which type of spillovers arises. We show that technological spillovers do not occur if the joint profit of the MNE and the local firm is highest when the MNE can use the technology as a monopolist. This result is similar to that obtained in the literature on the persistence of monopolies (see Tirole, 1988). The empirical implication is that one should expect higher labor mobility and more spillovers (both technological and pecuniary) when the local firm can use the technology in activities that do not compete fiercely with the MNE. This occurs, for example, when the local firm can operate in markets for products which are unrelated or complementary to the MNE’s products. Unfortunately, we are aware of no empirical studies which try to link the existence of spillovers with the sectors of activity of multinational and local firms. We also find that a low level of “absorptive capability” by the local firm, which might be due to technological backwardness, reduces the potential for FDI generating spillovers. The empirical evidence seems to confirm that spillovers increase with the degree of absorptive capability of host country firms (see, for instance, Kokko, 1994 and Borensztein et al., 1998). Further, the mobility of the trained workers is higher the more general is the on-the-job training given by the MNE, which is consistent with the labor economics literature (e.g., Becker, 1964). Our addition to this literature is to show that it is not only the nature of the training (general versus specific), but also the degree of product market competition which affects labor mobility. Finally, the MNE might anticipate that investing abroad would lead either to technological spillovers or to higher wages and choose to export instead. Anecdotal evidence confirms that this may sometimes be the reason why MNEs export. An illustrative example is drawn from the history of the chemical sector (Kudo, 1993). After World War I, the leading German chemical company, IG Farben, decided to increase its activity in the growing Japanese market, whose chemical industry was still at an infant stage. IG Farben resorted to exporting and avoided FDI (and licensing) as much as possible in order to minimize the diffusion of technology to competitors. Other game theoretical models have analyzed spillovers to foreign markets, although from different perspectives. In Ethier and Markusen (1996), technological spillovers arise as a result of a double moral hazard problem. A foreign firm endowed with a superior technology might renege on an exclusive contract with a local licensee by transferring technology to other local firms, whereas the licensee might “cheat” by introducing a marginal improvement in the technology. Fosfuri and Motta (1999) and Siotis (1999) analyze the decision between exporting and FDI, but they simply assume that when two firms locate in the same region a proportion of their know-how spills over to each other. This “black box” type of spillovers is quite familiar in the R&D literature (e.g., d’Aspremont and Jacquemin, 1988). The rest of this paper is organized as follows. Section 2 presents the model, analyzes the equilibrium outcomes and discusses the results obtained. Section 3 concludes the paper.

Spillovers have often been treated as a “black box” mechanism, where their nature is left unspecified. This paper provides a specific mechanism through which technology might involuntarily move from a firm towards others located in the same country. Therefore, the paper offers a rationale to the empirical literature which has uncovered the importance of localized spillovers (e.g., Audretsch and Feldman, 1996). We have presented a model where technological spillovers from FDI might occur due to workers’ mobility. A MNE can transfer a superior technology to its foreign affiliate only after having trained a local worker. Once trained, this worker can later be hired by a local firm and technological spillovers might occur. Even when such spillovers do not take place, the host country’s welfare might improve because of the wage that the trained worker receives from the MNE to prevent her from moving to a local firm. We have also shown that, in some circumstances, the MNE might prefer to export rather than to invest, precisely to avoid diffusion of superior technology to the local rival and/or the payment of rents to the trained worker. Our model helps to identify the conditions under which a MNE retains the trained worker, and those under which she leaves to a local firm. The results are consistent with the industrial organization literature on persistence of monopolies. Technological spillovers arise (the monopoly ceases to exist) when the “joint-profit” effect does not hold, that is, when industry profits are higher if both firms can use the technology. This is more likely to happen when the local firm and the MNE do not compete fiercely in the product market (or sell in independent or vertically related markets), when on-the-job training is general rather than specific, and when the absorptive capability of the local firm is high. We have accordingly identified some empirical predictions to which our model gives rise.