سرمایه گذاری مستقیم خارجی و سیاست های کشور میزبان: یک دلیل منطقی برای استفاده از محدودیت های مالکیت

This paper examines host governments' motivation for restricting ownership shares of multinational firms (MNFs) in foreign direct investment (FDI) projects. An MNF with a productivity advantage is willing to invest in a host country. The host government wants to capture the MNF's surplus yet cannot observe it due to the MNF's private information about its firm-specific advantage. In contrast, a joint venture (JV) partner might observe this surplus depending on its ownership share. The host government can alleviate its informational constraints by using ownership restrictions to force a JV. This calls into question the wisdom of calls for ‘liberalizing’ FDI flows by the wholesale elimination of domestic JV requirements. We show that the optimal mechanism involves ownership restrictions that decrease as the size of the MNF's firm-specific advantage increases.

Foreign Direct Investment (FDI) has been one of the most widespread forms of international economic activity in recent years. In 1980, FDI stock abroad accounted for only 5% of world GDP whereas in 1998, this number almost tripled to 14% (OECD, 2001). Notwithstanding its many advantages (e.g., technology spillovers, international trade integration, etc.), restrictions on FDI are fairly common. The most obvious one is foreign ownership restriction. For instance, many emerging countries like India, Indonesia, Brazil, Mexico, Korea, Turkey, the Philippines, Thailand and most centrally planned economies impose restrictions on foreign equity participation.1 Likewise, many industrialized countries such as Finland, France, Norway, Sweden, Switzerland and even the U.S. and Canada have also developed some kind of indigenisation policy.2 The principal characteristic of FDI is the control over operations exercised by the investor firms (multinationals). Ownership of equity enables multinationals to exercise this control. According to the property rights theory (Grossman and Hart, 1986, Hart and Moore, 1990 and Hart, 1995), ownership rights affect ex-ante investment decisions through their influence on the distribution of ex-post surplus. Therefore, forcefully changing the ownership structure of a firm might distort ex-ante investment which can decrease the overall surplus of a project. Then, it is not clear why host governments insist on using equity restrictions when other potentially more efficient policy instruments (e.g., non-distortionary taxation and redistribution of rents) are available. In this study, we develop a model that explains the common use of ownership restrictions by host governments despite their inefficiencies.3 The purpose of the ownership restriction policy is not to limit access to foreign firms but to capture the rents from the MNF's activity.4 The model combines adverse selection with moral hazard. The novel feature of our approach is the watchdog role played by the domestic joint venture partner, which in turn induces truthful revelation by the MNF. As Caves (1982) noted, for FDI to occur, the MNF must possess a firm-specific advantage (e.g., expertise, managerial and organizational practices and trademarked brands) which it can exploit more profitably through internalization than through licensing and the host country must have a production advantage for the relevant market. Following these arguments, we assume that there is an advantageous production opportunity in the host country. There is a pool of local firms (LFs) each of which is capable of undertaking this project. There is also an MNF that can carry out the project via FDI. It is reasonable to assume that the MNF has a firm-specific advantage over the LFs such that it can create a higher surplus by making some effort that is both costly to exert and unobservable to others. Given this, it is more efficient for the MNF than LFs to produce in the host market.5 The host government wants to maximize its welfare which consists of the weighted average of its tax revenue and any possible profit of the local firm.6 There is an asymmetric information problem between the host government and the MNF. The host government cannot not observe the precise amount of the extra surplus generated by the MNF (since the tax sheltering activity of the MNF renders cash flows opaque for the host government taxation).7,8 The magnitude of this additional surplus depends on the effort level chosen by the MNF and the size of the firm-specific advantage the MNF has. If these were commonly known, the host government would simply let the MNF produce and tax the resulting profit. Under incomplete information, such a policy may result in ex-post inefficiency, since the host government's taxation can force the MNF to stay out of the host market. In this case, an alternative policy is to force the MNF to form a JV with a particular local firm chosen by the host government (ownership restriction). It is assumed that unlike the host government, the local JV partner might observe the resulting extra surplus depending on its degree of ownership (since it can monitor cash flows into the JV firm by using its ownership control rights).9 There are two opposing effects of using the ownership restriction policy. On the one hand, it causes the MNF to reduce its effort thereby leading the resulting surplus to decrease (hold-up problem). On the other hand, it helps the host government to ease its information problem. The host government takes this trade-off into account and determines its policy accordingly. In particular, the optimal mechanism dictates that more productive multinationals should be granted higher ownership shares but should face higher taxes. We show that the ownership restriction policy can mitigate the ex-post inefficiency resulting from asymmetric information. Our paper is related to several earlier studies. In this literature, one can distinguish two mainstream approaches. In the first one, there are models that analyze foreign direct investment in a symmetric information framework.10 On the contrary, models in the second approach use asymmetric information structure. Our paper belongs to this latter group. In this second group, Stoughton and Talmor (1994) use a mechanism design approach to model the game between the parent firm and its subsidiary. In their model, the host country would set the optimal combination of tax and indigenisation rates (which appear multiplicatively) by trading off its share in the cash flows with the output considerations of the multinational. Our model is different as we do not consider parent-subsidiary relations but focus on the direct relation between an MNF and a host government. We also do not consider a game in which there is a chance for the MNF to switch funds from one place to another due to the differential tax rates in different countries. The closest paper to ours is by Dasgupta and Sengupta (1995). They analyze the optimal regulation of MNFs by a host government interested in maximizing tax revenues when the MNF has private information about its benefits of controlling the firm. In their model, this control is independent of ownership shares. Given these assumptions, the authors determine the optimal ownership restrictions. Our paper is closely related to their paper but differs in the following. First, we extend their model by introducing moral hazard. Second, unlike their paper, the magnitude of the private benefit is closely related with the MNF's ownership shares. Third, there is an asymmetric information problem not only between the MNF and the host government, but also between the JV partners, with the latter being less severe. Finally, in our model the local JV partner plays a watchdog role and this makes truthful revelation by the MNF possible in the optimal mechanism. As a result, the motivation presented in this paper for using ownership restrictions is different than their paper. In addition to these papers, there are also papers that use asymmetric information framework to analyze the optimal regulation of multinationals without employing ownership restriction policy (Prusa, 1990, Gresik and Nelson, 1994, Bond and Gresik, 1996, Calzolari, 2001 and Konrad and Lommerud, 2001). The rest of the paper is organized as follows. In the next section, we describe the basic analytical framework. In Section 3, given the fact that some countries discontinue using ownership restrictions, we discuss possible reasons for the elimination of these restrictions. Section 4 concludes the analysis.

There is little doubt that the general trend on FDI has been towards liberalization, especially in the past two decades. Many countries have reduced restrictions on FDI and adopted incentives to encourage FDI. Still, some restrictions remain in place even in countries that generally welcome FDI. The most obvious restriction on FDI is the foreign ownership restriction. In this study, we provide one possible explanation on the widespread use of this type of restriction. We show that, under incomplete information, ownership restrictions can be welfare-improving by reducing the information problem of the host government. Under the optimal mechanism, the host government is more lenient towards the MNF as its productivity (represented by the firm-specific advantage) is higher. This is the result of rent extraction-efficiency trade-off. For low values of the firm-specific advantage, the cost of restricting the MNF's ownership share (efficiency effect) is low and as a result, the ownership restriction is high. Yet, for high values of the firm-specific advantage, this cost increases. (Efficiency effect dominates the rent extraction effect.) Consequently, the ownership restriction is low. In FDI literature, it is often claimed that removing barriers would increase FDI flows. Notwithstanding, the information base and tools for analysis of FDI barriers are very limited. First of all, the measurement of restrictions is tricky. This is true even for relatively direct types of restrictions such as foreign ownership restrictions where different limits applying to different firms in a sector, to different types of foreign investors, and to individual and aggregate foreign investment. Furthermore, there is no international agreement on standardized reporting of policies on FDI with the partial exception of GATS schedules, which in itself is problematic. GATS commitments are based on a “positive list”25 approach and hence do not take into account sectors and restrictions which are unscheduled. In other words, a positive commitment in some sectors does not necessarily imply a restriction. A country may simply choose not to list a particular sector in its schedule to retain policy flexibility. Besides, it is hard to classify and rank various restrictions. When evaluating the effects of different types of restrictions, it is desirable to assign weights to them according to their significance, however, this procedure yields arbitrary judgements and errors. In addition, governments have altered their policies frequently in recent years and consequently it is not easy to obtain up-to-date information. To overcome this difficulty, surveys and questionnaires were used to obtain information but the reliability of them is open to question. It is possible that some countries are just more forthcoming than others in reporting their restrictions and thus they receive higher restrictive scores not because they are more restrictive but because they are more complete in their reporting. As a result, to develop a clearer picture of the implications of FDI barriers and the possible gains from further liberalization, more information is needed on the nature and extent of the existing barriers.