Taxing International Corporate Income Economic Principles and Feasibility.doc

Taxing International Corporate Income:Economic Principles and FeasibilityMichael P. DevereuxCentre for Business Taxation, Oxford UniversityMay 2008AbstractThis paper reviews economic principles for optimality of the taxation of international profit, from both a global and national perspective. It argues that for systems based on residence or source, nothing less than full harmonisation across countries can achieve global optimality. The conditions for national optimality are more difficult to identify, but are most likely to imply source-based taxation. However, source-based taxation requires an allocation of the profits of multinational companies to individual jurisdictions; this is not only very difficult in practice, but in some cases is without any conceptual foundation. To the extent that the international tax system cannot conform to basic economic principles, and is becoming increasingly infeasible, more fundamental reforms such as a destination-based tax require further examination.This paper was prepared for the Conference in Honour of Richard Musgrave, Sydney, June 2-4, 2008. It draws on a related paper, “Taxing Foreign Profit: Principles and Feasibility”, which was presented at the European Tax Policy Forum conference in London, April 21, 2008, on “International Tax and Economic Welfare”. Address for correspondence: Michael Devereux, Centre for Business Taxation, Said Business School, Park End Street, Oxford OX1 1HP, michael.devereuxsbs.ox.ac.uk1. IntroductionFor a number of reasons, the structure of taxes on international corporate profit is on the political agenda, especially in Europe. Various cases in the European Court of Justice have required significant changes to existing tax structures. Competition over corporate tax rates seems to have become more intense. And the European Commission is proposing a radical EU-wide reform to remove “tax obstacles” to doing in business in Europe. This paper has two main aims. First, it aims to review and extend the existing academic literature on the optimal structure of taxes on international corporate profit. This has been studied for nearly half a century; yet its importance is growing continually as companies and their activities become increasingly international. Second, it aims to review, from a broad economic perspective, the feasibility of source-based taxation of corporate profit. In reviewing the academic literature, the paper begins with the aim of maximising global income and analyses tax structures which are consistent with the production efficiency concept of Diamond and Mirrlees (1971). Production is allocated efficiently if it is not possible to reallocate resources between activities in a way that would increase total output. The paper shows that in a real-world situation in which there are cross-border flows of portfolio and direct investment, and also international trade, then taxes on a source or a residence base would be distorting unless they were completely harmonised. By contrast, in the same situation, from a national perspective it is optimal to exempt outbound investment from tax. This is contrary to the standard prescription of economic theory, which argues that outbound investment should be taxed at the same rate as domestic investment, with foreign taxes being deducible from the residence country tax base. The reason for the difference stems from the substitutability of domestic and outbound investment. If £1 of outbound investment crowds out £1 of domestic investment, then the standard results hold. But if both forms of direct investment are financed at the margin by inbound portfolio investment, then the link between them is broken. In this case, it is not necessarily optimal to tax outbound investment at the same rate as domestic investment. The UK has recently proposed to move towards source-based taxation, exempting foreign source corporate dividends from UK tax. However, to counteract avoidance possibilities, it has also proposed worldwide taxation of passive income received by affiliates of UK companies. The need to propose worldwide taxation to combat avoidance casts doubt on the long-term feasibility of a source-based system of taxing international corporate profit. There are good reasons for such doubt, arising in part from the nature of international business: it is very difficult in practice to determine a particular allocation of profit amongst countries, and increasingly there is little conceptual justification. But if this is true, then the long-term future of the taxation of international corporate profit if there is one - must lie elsewhere. One possibility, which requires further examination, is to base tax on where consumers purchase the final good or service: a destination-based tax. 2. Economic principles for international taxationThe best-known concepts relating to the optimal taxation of international capital income date back to the 1960s, in contributions from Peggy Musgrave (Richman, 1963, Musgrave, 1969). Musgrave introduced the terms “capital export neutrality (CEN)” and “capital import neutrality (CIN)”, which are now in common use. CEN holds if any individual investor faces the same effective tax rate on her investments, wherever those investments are located. CIN holds if all investments undertaken in the same jurisdiction face the same effective tax rate. Suppose that, at the margin and in the absence of taxation, competition drives the marginal pre-tax rate of return on all investments in a jurisdiction to be equalised. Then in a simple framework (see, for example, Keen, 1993) CEN implies that (a) the international tax system will not distort the location decisions of any individual investor, (b) the pre-tax rate of return in all jurisdictions will be the same (production will be efficiently organised), but (c) investors in different jurisdictions may face different post-tax rates of return on their investment, and hence different incentives to save. CIN implies that (a) the location choices of investors may be affected by differences in effective tax rates, (b) marginal pre-tax rates of return will differ across jurisdictions (there will not be production efficiency), but (c) investors in different jurisdictions will face the same post-tax rate of return on each of their investments, and hence all face the same incentive to save. To analyses this in more detail, suppose that there are two countries, A and B, with one investor and one asset in each country. The rate of return on each asset depends on the amount invested in that asset: as more is invested, the marginal rate of return declines. Each investor can purchase shares in either asset. The effective tax rates may depend on the location of the investor and the location of the asset: for example, is the effective tax rate faced by the investor in A on the returns from the asset in B. We will return below to the question of how these effective tax rates are defined. In principle, each investor would want to invest in each asset up to the point at which the post-tax rate of return from each investment was the same: if the post-tax rate of return was not the same, then the investor could increase her overall return by switching from the investment with the lower rate of return to the investment with the higher rate of return. However at least in the absence of risk - it is by no means certain that both investors will hold both assets simultaneously: indeed it is unlikely that they will do so without some pattern in effective tax rates. That is because both investors share the same pre-tax rate of return on each asset; but if they face different effective tax rates, they will face different post-tax rates of return. In general, they will not both be able to equalise their post-tax rates of return. If they do not, then only one investor would hold both assets, and the other would specialise in the one generating the higher post-tax rate of return. CEN holds in both countries if and that : that is, each individual investor faces the same effective tax rate on the returns from both assets. In this special case, equalising post-tax rates of return for either investor will ensure that the pre-tax rates of return from the two assets will also be equal, implying that production will be efficiently organised. In this case as well, both investors can hold both assets though if their tax rates differ - - then their post-tax rates of return will also differ. By contrast, CIN holds in both countries if and : that is both investors face the same effective tax rate when investing in a single asset. In this case, and assuming that the effective tax rates on the two assets are different from each other, equalisation of post-tax rates of return will not generate equalisation of pre-tax rates of return. However, the post-tax rates of return faced by each investor will be the same.The distinction between these two notions of neutrality has led to some debate as to which is the more important (see, for example, Keen, 1993). The economic literature has generally favoured CEN, on the grounds that it generates production efficiency (discussed further below), though this has not always met with approval. Thus, for example, McLure (1992) has claimed that: “economists have generally favoured CEN because it maximises global welfare . but businessmen generally favor the level playing field provided by CIN”. However, through a number of contributions discussed below, the question of the optimal tax structure has now progressed well beyond a simple analysis of CEN and CIN. The remainder of this section reviews and develops broader principles. 2.1. Global optimisation and production efficiencyThe starting point for any analysis of optimal tax systems is the Diamond and Mirrlees (1971) framework which demonstrated that, within a single country, it is optimal to preserve production efficiency. This holds when it is not possible to increase total output by reallocating inputs to different uses; which implies that the marginal pre-tax rate of return is the same on all investments. If this does not hold, then total output could be increased by shifting capital inputs from a less productive use to a more productive use. Competition between investors would achieve this in the absence of distorting taxes. However, a number of caveats must be made before simply accepting that production efficiency is optimal. First, the Diamond-Mirrlees theorem relies on two critical assumptions: that there are no restrictions on the use of tax instruments available to the government, and that economic rent is fully taxed at 100% See Stiglitz and Dasgupta (1971). (or there is no economic rent). Keen and Piekola (1996) analyse optimal tax rates between co-operating countries when economic rents exist but cannot be taxed at a rate of 100%. In this case, the optimal tax system depends on similar factors to those identified by Horst (1980); namely the elasticity of the supply of savings and the elasticity of the demand for capital in each jurisdiction. Keen and Piekola also show that the optimal tax structure depends on the rate at which economic rents are taxed. Huizinga and Nielsen (1995, 1996) analyse optimal tax policy when economic rents are taxed at less than 100%, in the absence of cooperation amongst governments. A second caveat was introduced by Keen and Wildasin (2004). They point out that the Diamond-Mirrlees model does not directly apply in an international setting, since there is no longer a single government budget constraint, but each country has its own budget constraint. They analyse the case in which lump sum transfers between governments are ruled out, but where transfers can instead take place via trade taxes and subsidies. Under these circumstances, it may be the case that the optimal (Pareto-efficient) tax system does not generate production efficiency. However, as pointed out by Edwards (2005), if the aim is to generate a global optimum, it is not clear why governments should co-operate by adjusting their trade taxes, rather than agreeing to lump-sum transfers. In the latter case, we are effectively returned to the Diamond-Mirrlees setting of a single budget constraint. Although the global optimality of production efficiency is an important issue, we now leave these caveats to one side. In the following discussion, we instead focus on the implications for the design of international taxes on profit of a requirement for production efficiency. We then consider in Section 2.6 the rather different case of national optimality.2.2. Capital ownership neutralityTo begin with, we discuss the use of the term “capital ownership neutrality (CON)” by Desai and Hines (2003). This is similar, though not the same, as the use of the same term by Devereux (1990), as discussed below. Desai and Hines note that much foreign direct investment takes the form of acquisitions, rather than greenfield investment. This is consistent with there being differences in productivity according to the owner of an asset; the classic example is of a multinational firm being more productive than a domestic firm (there is a wealth of empirical evidence on this: see, for example, Criscuolo and Martin, 2004). Now suppose that investor A would have a more productive use of an asset than the current owner, investor B. In the absence of other factors, and taxes, there would be a improvement in world output (and potential gains to both A and B) if A purchased the asset from B. CON, as used by Desai and Hines, is a condition that the tax system does not distort the ownership of assets; in this case, the international tax system does not prevent A purchasing the asset from B. To analyse this in more detail, return to the simple framework above. However, now suppose that the rate of return on each asset depends on the identity of the investor: for example, for the asset in A, is the pre-tax rate of return earned by the investor residing in A, and is the pre-tax rate of return earned by the investor residing in B. To make things more concrete, suppose that, whatever the total size of the asset in A, the investor from A is more productive: that is, . Begin with investor A. In the absence of taxes, he will invest in asset A, which for a small investment suppose yields a return higher than asset B. He will continue to invest in asset A until the marginal return from asset A falls to the level of asset B, It is also possible that given his wealth, and lack of any further finance, the rate of return on asset A remains higher than the rate of return on asset B; in this case, investor A would invest only in asset A. However, we do not analyse this possibility. that is until the pre-tax rates of return are the same. The position of investor B depends e