سیستم مالیات شرکت، شرکت های چند ملیتی و یکپارچگی اقتصادی

Multinational firms are known to shift profits and countries are known to compete over shifty profits. Two major principles for corporate taxation are Separate Accounting (SA) and Formula Apportionment (FA). These two principles have very different qualities when it comes to preventing profit shifting and preserving national tax autonomy. Most OECD countries use SA. In this paper we show that a reduction in trade barriers lowers equilibrium corporate taxes under SA, but leads to higher taxes under FA. From a welfare point of view, the choice of tax principle is shown to depend on the degree of economic integration.

The rise in FDI and multinational firm activity is one of the most pronounced trends in the world economy over the last two decades.3 This trend has worried policymakers and academics, since multinationals are known to shift profits to low tax countries and governments are prone to compete for shifty profits.4 In response to these problems, the European Commission has focused on “harmful” tax competition (as in the “Monti Package”), and has more recently published a study on corporate taxation. The latter study aims at finding a system for corporate taxation that prevents profit shifting, reduces compliance costs for firms, and preserves national tax autonomy (Commission of the European Communities, 2001). One of the main proposals emerging from the Commission's corporate tax study is a switch from the corporate tax system employed by most European countries—called Separate Accounting (SA)—to a system of Formula Apportionment (FA). Apportionment systems are already in use internally among states, provinces, and cantons in federal countries such as the United States, Canada, and Switzerland, where its introduction has been motivated by the need to disentangle the activities of state subsidiaries from the activities of multistate enterprises as a whole in order to secure a tax base in all states where the enterprise has ongoing activities. Under Separate Accounting (SA) taxable income of a corporation's activity in each jurisdiction is based on computing the value of transactions between related affiliates as if they had occurred by independent parties in the market place (so-called arm's length pricing). The obvious weakness of this system is that it can be difficult to obtain market parallels on which such prices can be established. In particular, there is substantial evidence that Multinational Corporations (MNCs) arise because they possess firm-specific assets that are intangible in nature and difficult to trade at arm's length (Markusen (1995)). In practice, multinationals therefore have significant discretion when setting their transfer prices. The competing alternative to SA, Formula Apportionment (FA), implies that the corporate group combines the income of each of its operatives into a single measure of taxable income. The group then uses a formula to apportion taxable income to each of the jurisdictions in which the group has activities.5 The advantage of this approach is that manipulation of income between affiliates by use of transfer prices does not have an impact on the single measure of income for the corporate group. Given the growing importance of multinationals worldwide and the attention by policymakers to the issue of company taxation, it is perhaps surprising that very little work has been done to compare Separate Accounting to Formula Apportionment. Our objective in this paper is to undertake such a comparison in a framework that also allows us to investigate the impact of economic integration on tax policy, the choice of corporate tax system, and welfare. Our paper relates to a small literature that has mainly addressed corporate tax competition in the presence of multinational firms and transfer pricing under SA.6Konan (1996) models strategic taxation policy of home and host governments under SA when a multinational enterprise sets transfer prices on globally joint inputs. She finds that an equilibrium home-tax solution is to tax foreign earned profits at a higher rate than domestically earned profits. In Elitzur and Mintz (1996), the transfer price takes on a dual role affecting both the amount of profits shifted and incentives for the subsidiary's managing partner. Using a framework of Separate Accounting governments compete over MNC profits and impose corporate income taxes subject to a rule that approximates what the government believes is the arm's length price. In the tax competition equilibrium tax rates are affected by home country production costs, agency costs, and the productivity of the subsidiary, and it is shown that tax harmonization is likely to reduce tax rates. Haufler and Schjelderup (2000) analyze the optimal taxation of multinational profits under SA when firms can shift profits between countries by transfer pricing. They consider a setting where countries compete over corporate profits by choosing both the tax rate and the tax base (depreciation allowances) simultaneously. They find that recent corporate tax reforms in the OECD where corporate tax rates have been reduced while the tax base has been broadened, are optimal responses to the increased presence of multinationals and transfer pricing.7 Studies that compare the welfare or revenue effects of a switch from SA to FA are scant.8 Slemrod and Shackelford (1998) examine financial reports from U.S. based multinationals for the period 1989–1993 to estimate the revenue implications of implementing a U.S. federal Formula Apportionment system. They find that a switch from SA to FA using an equal three-weighted, three-factor formula would have increased US tax liabilities by 38%. Nielsen et al. (1999) use a two-country setup to compare SA to FA. In their model each MNC consists of a parent firm in one country and a subsidiary in the other. Both the parent firm and its subsidiary produce an output using a public input and (plant-specific) capital, and the public input is acquired by the parent company and made available to the subsidiary at a (transfer) price. They find that if the pure profits of multinationals are either very low or very high, and at the same time the costs of engaging in transfer pricing are of intermediate size, a switch from SA to FA reduces tax revenue and welfare. Finally Anand and Sansing (2000) show that a harmonized apportionment rule can prevail as the cooperative solution to a game between states (as can a system under SA), but a state can increase its welfare by deviating from the cooperative solution. This incentive gives rise to a Prisoner's Dilemma type of problem under FA. We emphasize that none of the papers reviewed above focuses on economic integration (taken to imply a reduction in trade costs) and transfer pricing, nor on how the interaction between the two may affect tax competition and the choice of tax system. Our analysis is related to Nielsen et al. (1999) in the sense that we study the effect of competition over corporate profits in the presence of multinationals and transfer pricing. Different from their analysis (and previous studies) is that the transfer price applies to a traded commodity that can only be shipped to the subsidiary at a (trade) cost. This allows us to analyze the impact of economic integration. Furthermore, we also take into account the fact that the transfer price as well as being a tax saving device gives rise to strategic effects.9 The latter is in contrast to the traditional literature on transfer pricing where monopoly is most often assumed. Under oligopoly, it has been shown by Schjelderup and Sørgard (1997) for SA and by Nielsen et al. (2003) for FA that transfer prices trade off tax incentives against strategic incentives. The strategic role of the transfer price is similar to the role of export (or import) subsidies (taxes) in strategic trade policy models (see e.g. Brander, 1985), but with the difference that the transfer price can be used either as a profit shifting device or as a strategic trade instrument. The strategic effect of the transfer price is as follows: if affiliates of an MNC face oligopolistic competition, the MNC can gain by setting the transfer price of internationally traded goods at a central level and delegating decisions about quantities (or prices) to its local affiliates. Such a strategy is beneficial to the MNC as a whole if it triggers favorable responses by local competitors. For example, under Cournot competition, a low transfer price set by the headquarters, turns the importing affiliate into a low cost firm that behaves aggressively by selling a large quantity. Such aggressive behavior induces the local rival to behave softly by setting a low quantity.10 The soft response from the rival is beneficial to the MNC as a whole. Hence, delegation can achieve higher profits than would arise if all decisions were undertaken centrally. The implication is that the transfer price has a strategic value in addition to being an instrument for profit shifting. Furthermore, since it is the headquarters of the MNC that conducts trade policy, the chosen transfer price is both credible and consistent with international trade agreements. To sum up, this paper differs from previous studies in that it analyzes how economic integration affects equilibrium tax rates, transfer prices and national welfare under SA and FA. Another novelty of the analysis is that we allow transfer prices to take on a dual role in the sense that they are both tax saving and strategic devices in markets with oligopolistic competition. We show that the transfer price is relatively tax sensitive for a high degree of economic integration under SA, while the opposite is true under FA. Hence, the conventional wisdom in the tax competition literature that increased economic integration leads to lower tax rates is supported by our findings under SA. However, under FA where increased integration reduces the tax sensitivity of the transfer price, increased competition over shifty profits allows governments to levy higher tax rates. A basic message that emerges from our analysis is therefore that from a welfare point of view the choice of system for corporate taxation hinges on the level of economic integration.

This paper has demonstrated that the transfer price of multinationals is relatively tax sensitive for high degrees of economic integration under Separate Accounting. Separate Accounting is the corporate tax system used by most OECD countries. In contrast, the transfer price is not very tax sensitive for closely integrated countries under a Formula Apportionment tax system, which is used in the USA and Canada, and proposed by the recent EU Commission report on corporate taxation.18 These findings are mirrored in the welfare analysis, where we find that a system of Formula Apportionment (Separate Accounting) dominates for high (low) degrees of economic integration. Thus, the choice of corporate tax system depends crucially on the perceived degree of economic integration, and our findings give support to the view brought forward by many other economists that increased economic integration may call for a substantial reform of the corporate tax system.19 In our model we have made a number of simplifying assumptions, two of which we would like to discuss in more detail. The first relates to trade costs, where we have assumed that it is the foreign subsidiary that pays these expenses. An alternative formulation is to let the exporting plant pay the trade costs. Everything else being equal, the importing plant is more competitive (has lower costs) when it does not pay trade costs. This implies that the transfer price needs not be set as low as in the case when the importing plant pays the trade costs. The alternative modelling assumption thus amounts to a scaling of the transfer price that does not qualitatively affect the tax sensitivity of transfer prices under SA and FA, nor our welfare analysis.20 The second simplifying assumption refers to the use of tax revenues. Would our results change if we allowed tax revenues to be used for public good production? Our analysis shows that tax revenues differ under SA and FA and a reasonable conjecture is therefore that this difference would be reflected in differences in the provision of public goods under the two tax schemes. For public consumption goods one would not expect our results to change qualitatively, but if there is decreasing utility from consuming public goods, the relative benefit of one scheme to the other would be less pronounced. If instead tax revenue could be used to enhance the productivity of firms, one would expect, depending on the cross derivative between private and public input goods, that the preference for one tax scheme would increase. However, the main thrusts of our arguments should survive, but this is an obvious topic for future research.