رقابت در مالیات و عملکرد مورد نیاز برای سرمایه گذاری مستقیم خارجی

Tax incentives offered to attract firms engaged in foreign direct investment are often tied to performance requirements such as domestic content restrictions or adherence to environmental standards. The tax competition literature has repeatedly shown that competition between municipalities for mobile firms tends to drive taxes to low levels. One would expect a comparable result for burdensome performance requirements. Despite this, the evidence suggests that while taxes have indeed been driven down, performance requirements are as popular as ever. We explain this seeming conundrum by showing that in the presence of spillovers, binding performance requirements can act as a coordination device for firms. In equilibrium, municipalities choose performance requirements, which maximize joint surplus from investment. Competition between municipalities then transfers this surplus to firms via tax subsidies.

Of the many policy tools available to manipulate foreign direct investment (FDI), two of the most prominent are taxes and performance requirements (PRs) such as domestic content requirements. As Graham (2000) notes, both tax subsidies and PRs are frequently used by both developed and developing nations. Often these tax subsidies are directly contingent on the multinational enterprise (MNE) satisfying various PRs (MITI, 2000). Examples of such are numerous. For instance, several jurisdictions link incentives to employment levels. In Georgia, USA, when a firm creates 15 or more jobs, it is eligible for a payroll tax credit, provided that the average wage is at least 110% of the average in the lowest wage county.1 When the Korean firm Hynix invested in Eugene, Oregon, the firm was promised 3-year property tax exemptions with renewals conditional on it expanding employment by 10% each time (Rowan and Witt, 2003).2 Other incentives are conditional on the location of investment within a country. For example, firms investing in eastern Germany can have as much as 100% of their worker training paid for by the government (compared to 30% in the west).3 Elsewhere, these breaks are dependent upon export requirements, such as Slovakia's 100%, 5-year tax rebate for incoming MNEs that export at least 60% of their output. Elsewhere, incentives are given for certain undertakings. In response to concerns over pollution, Taipei China offers foreign firms investing in pollution abatement corporate income tax reductions of 5–20%.4 Similarly, almost every government offers special tax breaks only available to firms engaging in R&D.5 While tax competition between locations for FDI has been widely studied, we are unaware of any work which considers competition in PRs. As Wilson's (1999) survey highlights, tax competition models frequently predict a “race to the bottom” in which taxes are set lower than the host would like. This occurs as potential locations undercut one another in attempts to attract mobile investment, evidence of which is found by Devereux et al. (2002). On the surface, it might seem that if firms find PRs burdensome, then competition should reduce them just as it reduces taxation. This concern is especially prevalent among critics of globalization who fear an erosion of pollution and labor standards as jurisdictions compete for mobile firms.6 However, the bulk of the evidence suggests that no such erosion is occurring. In fact, Graham (2000) suggests that much the opposite is true. This poses a conundrum: Why are taxes bid away while PRs are not? We seek to resolve this issue by suggesting that while from the perspective of any individual firm PRs may not be viewed as beneficial, it may still be the case that from the joint perspective of all the firms and the municipality they may be highly desirable. One of the most highly touted benefits of FDI is that it creates spillovers. We interpret “spillovers” quite broadly to include not only the typical examples of effects between firms such as technology transfers and worker training (which benefits future employers), but also the effects between the firms and municipalities such as the benefits of employment, costs of pollution, and any other effects of FDI that MNEs do not internalize. Helleiner (1989) and Caves (1996) summarize a variety of evidence that points to the importance of spillovers. In a recent example Okamoto (1999) showed that US suppliers with ties to Japanese automakers have total factor productivity growth rates 8.9% higher than comparable firms without these links. Okamoto attributes part of this productivity gain to technological diffusion from the Japanese firms to their domestic counterparts. Javorcik (2004) finds similar productivity effects from foreign firms in Lithuania. When there are multiple firms the spillovers may interact and reinforce each other. Examples of this include agglomeration benefits, which are increasing in industry output, a host country labor market with an upward-sloping supply curve, and competitive effects in an oligopolistic market. Since firms do not internalize all of the effects of their actions, this can lead to individual firms making inefficient choices. PRs, which often specify local employment and training levels, research requirements, and percentages of local content, coordinate firms on actions that maximize total rents. The maximized rents generated by MNE activity are then extracted by the firms through competitively set tax subsidies. Thus, in equilibrium, generous tax packages will be offered at the same time as PRs are imposed. This is exactly the pattern described by Graham (2000), MITI (2000), and Moran (1998). The low tax rates are also indicative of the “race to the bottom” predicted by many models of tax competition for FDI. There are several models studying tax competition for discrete firms, including Black and Hoyt (1989), Haufler and Wooton (1999), and others. The theme running through these models is that governments offer subsidies to attract firms in a manner akin to second-price auctions, a process that transfers surplus from the host jurisdiction to the foreign firm. While to our knowledge there is no work examining competitively set domestic content requirements, the interaction between foreign firms operating under PRs and domestic firms has received considerable attention. Grossman (1981), Richardson (1991), Fung (1994), and Madan (1998) all contribute to this rich literature. They each study mechanisms by which the choices of MNEs can impact the profitability and performance of domestic firms. In this literature, the contribution of Lahiri and Ono (1998) is closest to our own. They examine how a small host country can use a combination of income taxes and domestic content requirements to determine the number of MNEs it attracts. FDI benefits the host through employment effects and via the endogenous price of an oligopolistic, non-traded good. They find that a tax on FDI is optimal only when the ratio of domestic to foreign input is sufficiently high. By considering a small host, they treat the firms’ outside options as exogenous, and do not consider inter-jurisdictional competition for firms. We develop a multiple jurisdiction framework in which the Lahiri and Ono model is a special case. Endogenous outside options lead to an equilibrium in which the PRs are set to maximize total surplus, yet much of this surplus is redistributed to firms via tax subsidies. Thus while we show that the Lahiri and Ono result holds for a partial equilibrium analysis, the gains to the host can evaporate in a more general setting. While there is a gap dealing with competition in domestic content restrictions, models of competition in other PRs exist. In particular, there is an active debate on the likelihood of a race to the bottom in environmental standards.7Leonard (1988), for example, predicts the emergence of “pollution havens” as jurisdictions relax environmental regulations in order to attract mobile firms. However, many empirical studies such as Eskeland and Harrison (2003) and Smarzynska and Wei (2001) find no evidence for pollution havens. Even when a negative relationship between FDI and pollution standards is found, the predicted effects are quite small. For example, Keller and Levinson (2002) predict that doubling a US state's environmental cost index would reduce its FDI by less than 10%. Our results, however, suggest that pollution havens should not arise in equilibrium. Instead, one should expect jurisdictions to maintain their desired environmental quality but that these jurisdictions should offer higher tax breaks for FDI. Illustrative evidence on this point is provided by Dean et al. (2002), who show that the average levy per ton of water pollution is more than twice as large in those Chinese provinces that offer some degree of FDI incentives as in those that do not. Furthermore, in their regressions they find that while the location decisions of MNEs from OECD countries do depend on the incentives offered, they do not depend on pollution levies. These are the patterns our model predicts. Thus, our model provides an explanation for the mystery of missing pollution havens. In a similar vein, our results present a new rationale for why competition for FDI does not appear to have had the predicted negative effect on labor standards. Note that the key to maintenance of such standards is the ability to tie them to tax subsidies. In this light, it is interesting to consider the attitude towards Trade Related Investment Measures (TRIMs) taken by the General Agreement on Tariffs and Trade (GATT) and the World Trade Organization (WTO). In the 1980s, disputes led the GATT to declare domestic content requirements as inconsistent with the group's rules. The Uruguay Round of the WTO led to a similar condemnation of other TRIMs. Both organizations, however, only banned TRIMs as conditions for the “entry or continued presence” of MNEs. This definition still permits the imposition of PRs as a condition for receiving investment incentives (see Graham, 2000, pp. 60–63 for details). This difference is consistent with our results as competition in PRs alone can lead to the erosion of standards whereas joint competition in PRs and taxes will not. Finally, we show that when considering joint determination of taxes and PRs, there exist heretofore-unrecognized benefits to federally-set PRs. Federally set PRs affect the surpluses generated in all potential host jurisdictions. Since tax competition for FDI is a second price auction, the surplus retained by the winning jurisdiction is the difference between the highest bid of a losing jurisdiction and the surplus generated in the winning location. Federally set PRs may be used to lower the surplus that would be generated in the second-highest bid jurisdiction driving it to lower its bid with the consequence that the winning jurisdiction retains a greater proportion of any surplus generated. Hence even federal PRs that generate inefficiencies (by distorting choices in the winning jurisdiction) can raise national welfare levels. Thus, we demonstrate that coordination in PRs can act as an imperfect substitute for coordination in taxes. We proceed as follows. In Section 2 we present the basic model, our main results, and extensions to the basic model. Section 3 discusses welfare properties of our equilibrium. Section 4 provides a specific application of our results to a model with R&D spillovers between firms. Section 5 concludes.

In the scramble for FDI, jurisdictions often compete by offering MNEs tax subsidies if they locate within their borders. Since firms find taxes burdensome, this can lead towards a race to the bottom in which large tax subsidies are offered. At the same time, however, jurisdictions link these subsidies to PRs such as domestic content restrictions. Since firms presumably find these restrictions burdensome as well, why these PRs are not competed away presents a mystery. We resolve this issue by demonstrating that PRs can play two important roles in the competition for FDI. First, locally set PRs act as a coordination device used by jurisdictions to maximize the gains from FDI, enhancing their ability to compete for FDI by offering more attractive tax subsidies. Because of this, such binding PRs are used as part of an equilibrium strategy even though a large part of the surplus may be extracted by firms through competitively set taxes. Additionally, in the absence of distributional concerns and inter-jurisdictional spillovers, such equilibria are efficient. Nevertheless, even when efficiency fails, the equilibrium use of binding PRs seems to be more likely than not. Second, the use of federally set PRs can act as a substitute for tax coordination. Because a federally set requirement affects the value of a firm to all jurisdictions, this can lower the bid, that is, increase the tax, of the losing jurisdictions. Thus, this second price manipulation can facilitate the retention of a larger proportion of the surpluses generated by FDI by the local jurisdictions. It is interesting to note that local PRs are designed to be purely surplus enhancing, while optimal national PRs can be designed to be surplus reducing as long as this carries a sufficiently large second price manipulation. There are two admitted limitations of our model. First, it is static. As found by King et al. (1993), when there are multiple periods and a firm can relocate, the firm may locate in an inefficient location because of the effect this has on bidding in the following round. As the authors note, this firm strategy does not work in a one-period model and therefore does not arise in the present context. However, it is not difficult to find situations in which PRs are used in a more complex dynamic setting even if the equilibrium is not fully efficient. Second, it assumes full information. It is perhaps more realistic to assume that firms have private information about the benefits they would generate in a given jurisdiction.20 In such a setting, PRs can assist governments in separating firms as well as coordinating them.21 In particular, with desirable and undesirable types of firms, using PRs to separate types can even be welfare improving. Thus, while we leave the details of such a model to future research, it is not difficult to imagine that even with private information binding PRs can exist as part of an equilibrium strategy. Despite these limitations, our results make two original contributions to the literature on competition for FDI. First, we have provided a rationale for why binding PRs continue to exist, and are not bid away in the competition to attract mobile capital. PRs can coordinate firms and increase the gains from FDI. This increase arises because PRs induce firms to internalize spillovers both between themselves and with the jurisdictions. Second, we have shown that federal PRs can be used as an indirect tax coordination device. By manipulating the tax packages that losing jurisdictions are able to offer federally set PRs allow winning jurisdictions to capture a greater share of the surplus generated by FDI.