Harvard Law School - HUIT Sites Hosting

2 downloads 24078 Views 537KB Size Report
This article is dedicated to Orli Avi-Yonah for her unfailing inspiration and love, ..... corporate income tax on income from the sale of goods or provision of services ... commerce (a web page or even a server located in a jurisdiction) do not ...
Harvard Law School Public Law and Legal Theory Working Paper Series Working Paper No. 004 Spring 2000

GLOBALIZATION, TAX COMPETITION AND THE FISCAL CRISIS OF THE WELFARE STATE Reuven S. Avi-Yonah

A revised version of this working paper is forthcoming in 113 Harv. L. Rev. (May, 2000)

This paper can be downloaded without charge from the Social Science Research Network Electronic Paper Collection at: http://papers.ssrn.com/paper.taf?abstract_id=208748

Forthcoming, 113 Harv. L. Rev. (May, 2000) GLOBALIZATION, TAX COMPETITION AND THE FISCAL CRISIS OF THE WELFARE STATE Reuven S. Avi-Yonah1 I.

Introduction

The current age of globalization can be distinguished from the previous one (from 1870 to 1914) by the much higher mobility of capital than labor (in the previous age, before immigration restrictions, labor was at least as mobile as capital). This increased mobility is the result of technological changes (the ability to move funds electronically) and the relaxation of exchange controls. The mobility of capital is linked to tax competition, in which sovereign countries lower their tax rates on income earned by foreigners within their borders in order to attract both portfolio and direct investment. Tax competition, in turn, threatens to undermine the individual and corporate income taxes, which traditionally have been the main source of revenue (in terms of percentage of total revenue collected) for modern welfare states. The response of developed countries has been first, to shift the tax burden from (mobile) capital to (less mobile) labor, and second, when further increased taxation of labor becomes politically and economically difficult, to cut the social safety net. Thus, globalization and tax competition lead to a fiscal crisis for countries that wish to continue to provide social insurance to their citizens at the same time that demographic factors and the increased income inequality, job insecurity, and income volatility that result from globalization render such social insurance more necessary. The result is increasing pressure to limit globalization (e.g., by re-introducing exchange controls) which risks reducing world welfare. This article argues that if both globalization and social insurance are to be maintained, it is necessary to cut the intermediate link by limiting tax competition in a way that is congruent with maintaining the ability of democratic states to determine the desirable size of their government. From its beginnings late in the 19th century to the recent rise in payroll taxation, the welfare state has been financed primarily by progressive income taxation. The income tax 1

Assistant Professor of Law, Harvard Law School. I would like to thank the participants in workshops at Georgetown, Harvard, Michigan, Minnesota, NYU, Penn, UCLA, and Virginia law schools, as well as participants in the 1999 Harvard Seminar on Current Research in Taxation, the G-24 Technical Advisory Group, and the 1999-2000 MacArthur Workshop on Transnational Economic Security, for their helpful comments on previous versions of this article. Special thanks are due to Bill Andrews, Hugh Ault, Lucian Bebchuk, David Bradford, David Charny, Wouter van Ginneken, Michael Graetz, Joe Guttentag, Daniel Halperin, Jim Hines, Robert Howse, Howell Jackson, Louis Kaplow, David Kennedy, Duncan Kennedy, Robert Kuttner, Lance Lindblom, Jack Mintz, Dani Rodrik, Dan Shaviro, Reed Shuldiner, Joel Slemrod, Deborah Spar, Emil Sunley, Vito Tanzi, Louis Wells, Philip West, and Bernard Wolfman for their comments, to Stephanie Hunter for her meticulous and insightful research assistance, and to the Ford Foundation for its generous support. This article is dedicated to Orli Avi-Yonah for her unfailing inspiration and love, without which it would never have been written.

1

differs from other forms of taxation (such as consumption or social security taxes) because, in theory, it includes income from capital in the tax base even if it is saved and not consumed. Because the rich save more than the poor, a tax that includes income from capital in its base is more progressive (taxes the rich more heavily) than a tax that excludes income from capital (e.g., a consumption tax or a payroll tax). However, the ability to tax saved income from capital (i.e., income not vulnerable to consumption taxes) is impaired if the capital can be shifted overseas to jurisdictions where it escapes taxation. Two recent developments have dramatically augmented the ability of both individuals and corporations to earn income overseas free of income taxation: The effective end of withholding taxation by developed countries and the rise of production tax havens in developing countries. Since the United States abolished its withholding tax on interest paid to foreigners in 1984, no major capital importing country has been able to impose such a tax for fear of driving mobile capital elsewhere (or increasing the cost of capital for domestic borrowers, including the government itself). The result is that individuals can generally earn investment income free of host country taxation in any of the world’s major economies. Moreover, even developed countries find it exceedingly difficult to effectively collect the tax on the foreign income of their individual residents in the absence of withholding taxes imposed by host countries, because the investments can be made through tax havens with strong bank secrecy laws. Developing countries, with much weaker tax administrations, find this task almost impossible. Thus, crossborder investment income can largely be earned free of either host or home country taxation. When we switch our attention from passive to productive investment, a similar threat to the taxing capacity of both home and host jurisdictions emerges. In the last decade, competition for inbound investment has led an increasing number of countries (103 as of 1998) to offer tax holidays specifically geared to foreign corporate investors. Given the relative ease with which an integrated multinational can shift production facilities in response to tax rates, such “production tax havens” enable multinationals to derive most of their income abroad free of host country taxation. Moreover, most developed countries (including the U.S.) do not dare impose current taxation (or sometimes any taxation) on the foreign source business income of their resident multinationals for fear of reducing the competitiveness of those multinationals against multinationals of other countries. If they did, new multinationals could be set up as residents of jurisdictions that do not tax such foreign source income. Thus, business income can also be earned abroad largely free of either host or home country taxation. If income from capital can escape the income tax net, the tax becomes in effect a tax on labor. Several empirical studies have in fact suggested that in some developed jurisdictions the effective tax rate on income from capital approaches zero, and nominal tax rates on capital have tended to go down sharply since the early 1980s (when exchange controls were relaxed). As a result, countries that used to rely on the revenues from the income tax are forced to increase relatively regressive taxes. The two fastest growing taxes in OECD member countries in recent years have been consumption taxes (from 12% of total revenues in 1965 to 18% in 1995) and payroll taxes (from 18% to 25%), both of which are more regressive than the income tax. Over

2

the same period, the personal and corporate income taxes have not grown as a percentage of total revenues (the personal income tax accounted for 26% of total revenues in 1965 and 27% in 1995, while the figures for the corporate income tax are 9% and 8% respectively). The total tax revenue as a percentage of GDP in developed countries went up sharply during the same period (from an average of 28% in 1965 to almost 40% in 1994), and this increase is largely accounted for by the rise of consumption and payroll taxes. Moreover, there is evidence that as the degree of openness of an economy in OECD member countries increases, taxes on capital tend to go down while taxes on labor go up (the income tax is imposed on both capital and labor so that its stability may mask this trend). At some point, developed jurisdictions find themselves politically unable to raise income taxes on labor, consumption taxes, or payroll taxes any further. High rates of income tax on labor discourage work, high payroll taxes discourage job creation and contribute to unemployment, and high consumption taxes (e.g., on luxury goods) drive consumption overseas. If developed countries are unable to tax income from capital, and alternative taxes are not feasible, the only recourse is to cut the social safety net. But that net is needed more than ever because of both demographic factors and the increased income inequality, income volatility, and job insecurity than tend to result from globalization. Thus, globalization leads to an increased need for revenues at the same time that the increased mobility of capital that it brings limits the ability of governments to collect more revenues. This dilemma threatens to undercut the social consensus that underlies modern industrialized societies and threatens to create a backlash against globalization, despite its overall benefits. The previous age of globalization collapsed in the face of just such a backlash in the 1920s. To prevent this and to maintain the social compact underlying the modern welfare state, it is necessary (inter alia) to find a means to tax cross-border income from capital at rates that approximate the rates that are imposed domestically on labor income, i.e., to limit tax competition. However, any limits on tax competition must (as a normative and practical matter) be balanced against the desire of democratic countries to be able to set the size of their public sector as their citizens see fit. This article attempts to describe this problem and then develop possible solutions based on curbing tax competition in ways that are congruent with democracy and that balance the need of developed countries to maintain their social safety network with the need of developing countries to attract foreign capital. The article is divided into six parts. Part II describes in detail the ways in which international tax competition can lead to under-taxation of cross-border income from capital (compared with the rate of tax imposed on domestically invested capital in developed countries). Part II seeks to establish that both portfolio investment income and income from direct investment can be earned abroad largely free of tax, in the ways summarized above, and then presents the available data on the magnitude of the resulting revenue losses. Part III of the article analyzes the problem of tax competition from a global normative perspective. Tax competition involves three normative considerations: efficiency, equity, and

3

democracy. From an efficiency perspective, tax competition impairs the optimal global allocation of investments: When the effective tax rate abroad is lower than the effective rate at home, taxpayers will prefer international investments even when the return on domestic investments is higher before taking taxes into account, thus diminishing global welfare. From an equity perspective, if capital cannot be effectively taxed, the tax base tends to shift toward labor; and taxes on labor are more regressive than taxes on capital. Thus, tax competition impairs the ability of the income tax to redistribute wealth from the rich to the poor. These considerations weigh in favor of limiting tax competition. However, they should be balanced by respect for the ability of democratic countries to determine the desirable size of their public sector. Thus, Part III concludes by trying to define which forms of tax competition are and are not acceptable when the efficiency and equity arguments outlined above are balanced against democracy. Part IV of the article attempts to describe the costs and benefits of tax competition from the perspective of developed and developing countries. In the case of developed countries, tax competition makes it increasingly difficult to finance the social safety net that is part of the role of government in a modern welfare state. This is occurring at a time when the social safety net is coming under increasing pressure from both demographic factors and the side effects of globalization itself. In the case of developing countries, tax competition makes it difficult to maintain stable government revenues; and the loss of revenue is not adequately compensated by benefits flowing from increased foreign investments. Finally, Part IV attempts to analyze the division of tax revenues among countries that results from tax competition in terms of internation equity, arguing that the interests of developing countries should be preferred in choosing among different solutions to the tax competition problem. Part V of the article seeks to suggest possible solutions. It first describes and criticizes recent proposals to curb tax competition by the EU and the OECD, arguing that they represent useful beginnings. However, because those proposals are limited to member countries of the EU and OECD, they are insufficiently broad and therefore likely to be ineffective in the long run. Part V then seeks to develop alternative solutions, based on the imposition of withholding taxes on a coordinated basis by OECD members in which portfolio investment and the consumption of goods and services take place, in a way that adequately distinguishes between acceptable and unacceptable forms of tax competition. Finally, Part VI concludes by asking whether such a solution can be applied solely by the OECD or whether the tax competition problem should be addressed by a broader multilateral body (a “world tax organization”). It argues that while OECD-based solutions are feasible in the short run, in the longer run broader multilateral solutions, in which developing countries take an active role, are both preferable and feasible. Moreover, the World Trade Organization presents as the natural candidate for implementing such broader solutions.

4

II.

International Tax Competition and the Taxation of Capital a. Taxation of Savings: Portfolio Exemptions, Traditional Tax Havens, and the Global Tax

In 1984, the U.S. unilaterally abolished its withholding tax (of 30%) on foreign residents earning “portfolio interest” income from sources within the U.S.2 “Portfolio interest” was defined to include interest on U.S. government bonds, bonds issued by U.S. corporations (unless the bondholder held a 10% or more stake in the shares of the corporation), and interest on U.S. bank accounts and certificates of deposit.3 This “portfolio interest exemption” is available to any non-resident alien (i.e., any person who is not a U.S. resident for tax purposes) without requiring any certification of identity or proof that the interest income was subject to tax in the investor’s country of residence. The enactment of the portfolio interest exemption was the result of three factors, all of them rather fortuitous. First, as a result of the Reagan tax cuts of 1981, which dramatically lowered the U.S. effective tax rate,4 and the accompanying defense build-up, the U.S. government faced a significant budgetary deficit which could only be financed by borrowing abroad, especially from Japan (which had a large savings surplus). Second, the Japan-U.S. tax treaty imposed a 10% withholding tax on interest unlike most U.S. tax treaties with developed countries that have a zero withholding rate on interest.5 Third, the U.S. in 1984 decided to terminate its tax treaty with the Netherlands Antilles which had a zero withholding rate on interest and no limitation on benefits to Antilles residents.6 The Antilles treaty was used by U.S. corporations to borrow abroad without having a withholding tax imposed on the interest by channeling the loans through a Netherlands Antilles finance subsidiary. Thus, the enactment of the portfolio interest exemption was motivated by the desire of both the U.S. government and 2

Internal Revenue Code of 1986, as amended (“I.R.C.”), § 871(h). Interest on loans made by foreign banks was excluded from the exemption to protect the competitive position of U.S. banks. See I.R.C. § 881(c)(3)(A). Interest on bank accounts received by foreigners was exempt since the Tax Reform Act of 1976. See I.R.C. § 871(i). 4 See Hans-Werner Sinn, Why Taxes Matter: Reagan’s Accelerated Cost Recovery System and the U.S. Trade Deficit, 1 ECON. POL’Y 240, 246 (1985) (estimating that the adoption of ACRS resulted in a shift of between $.8 and $1.1 trillion of foreign investment into the U.S. or about 7% of world capital stock). 5 See Convention Between the United States of America and Japan for the Avoidance of Double Taxation and the Prevention of Fiscal Evasion with Respect to Taxes on Income, July 9, 1972, U.S.-Japan, art. XIII, 23 U.S.T. 967, 990. 6 The Netherlands Antilles were officially notified of the termination of the U.S.-Netherlands Antilles income tax treaties on June 29, 1987 to become effective January 1, 1988. However on July 10, 1987, the U.S. modified is termination to provide that Article VIII (exempting interest paid by U.S. persons to corporations and residents of the Netherlands Antilles from U.S. tax) continued in force. The Treasury reached an agreement with the Antilles on a protocol to take effect June 30, 1996 to the limited income tax treaty that terminated the U.S. tax exemption on interest paid to Antilles companies that were not U.S.-owned (i.e. not CFCs), U.S.-owned Antilles companies on non-eurobond debt, and U.S.-owned Antilles companies on eurobond debt issued after Oct. 15, 1984. See Andre Fogarasi et al., Current Status of U.S. Tax Treaties, 25 TAX M GMT INT ’L J. 523 (1996). 3

5

U.S. multinationals to borrow abroad without having to bear the cost of any withholding tax (which under the circumstances of 1984 was likely to be shifted to the borrower). Arguably, none of these reasons is valid today. The U.S. government is in budgetary surplus, and Japan is an unlikely source of funds (and if it were, given the current desire of Japan to attract foreign investors, the tax treaty could be renegotiated to reduce the withholding rate on interest).7 Moreover, under current conditions, both the U.S. government and U.S. corporations can probably afford to borrow abroad without bearing the cost of any withholding because of the size of the U.S. market and the widely held perception of U.S. bonds as relatively safe investments in turbulent economic times.8 However, the portfolio interest exemption is still with us.9 The result of enacting the portfolio interest exemption has been a classic race to the bottom: One after the other, all the major economies have abolished their withholding tax on interest for fear of losing mobile capital flows to the U.S.10 The following table shows current withholding rates in EU member countries and the U.S. on interest paid on banks accounts, securities (government and corporate bonds), and on dividends paid to foreign residents in the absence of a treaty:11 Country Belgium Denmark France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain

Bank accounts 0 0 0 0 10 0 10 0 0 15 0

Securities 10 0 0 0 10 0 10 0 0 15 0

Dividends 15 15 15 15 0 0 15 15 15 15 15

7

Interestingly, the reason this was not done in 1984 was probably the desire of Japanese investors to avoid being covered by the exchange of information article in the treaty which could have enabled the Japanese National Tax Administration to discover that they earned unreported foreign source income. 8 See J. Bradford Delong, Should We Fear Deflation?, 1 BROOKINGS A RTICLES IN ECONOMIC A CTIVITY 225 (1999); Clive Crook, The State of the World—Fragile but Euphoric, NAT ’L J., Jan. 16, 1999, at 86. 9 See Part V for an argument that the U.S. should abolish the exemption unilaterally given recent EU developments. 10 The race may have begun earlier than 1984 because the interest on U.S. bank deposits was tax-free before then and the Antilles treaty enabled U.S. borrowers to avoid withholding on their interest payments. But the symbolic significance of enacting the exemption gave the race a strong push forward as indicated by the resulting capital flight from Latin America (discussed below). 11 See VITO TANZI, TAXATION IN AN INTEGRATING W ORLD 130-31 (1995); see also Edward H. Gardner, Taxes on Capital Income: A Survey, in TAX HARMONIZATION IN THE EUROPEAN COMMUNITY 52, 60 (George Kopits ed., 1992) (providing similar but not identical figures).

6

United Kingdom United States

0 0

0 0

15 30

As the table indicates, most developed countries levy no withholding tax on interest paid to nonresidents on bank deposits, government and corporate bonds. Withholding taxes are levied on dividends despite the fact that dividends (unlike interest) are not deductible and therefore the underlying income has already been taxed once.12 However, the discrepancy between interest and dividends may in reality be less than the rates given above indicate for two reasons: First, a significant portion of the return on equity comes in the form of capital gains which are not subject to source-based taxation in any of the countries included in the table. Second, the withholding tax on dividends is generally easy to avoid for sophisticated investors. For example, one technique is to construct a “total return equity swap” in which the foreign investor receives payments equivalent to dividends from an investment banker in the source country who in turn hedges by holding the underlying stock and receiving the actual dividends. Most countries do not withhold on the dividend equivalent amounts, and the underlying dividends are free from withholding because paid to a domestic entity.13 This situation has led to calls for a “portfolio dividend exemption.”14 The standard economic advice to small, open economies is to avoid taxing capital income at source because the tax will be shifted forward to the borrowers and result in higher domestic interest rates.15 However, the countries in the table include large economies (the U.S., Germany, and the U.K.) in which the tax is not necessarily shifted forward. Rather, the principal reason for the lack of withholding taxes in most of the countries included in the table above is the fear that, if such taxes were imposed, capital would swiftly move to other locations that do not impose a withholding tax. Thus, the Ruding Committee, writing about the European Community, concluded in 1992 that “recent experience suggests that any attempt by the EC to impose withholding taxes on cross-border interest flows could result in a flight of financial capital to non-EC countries.”16

12

Moreover, even countries that integrate the corporate and shareholder taxes by giving shareholders a credit for corporate tax attributable to dividends (and most countries in the above table use this “imputation” method of integration) do not generally extend those credits to non-residents. 13 See Reuven Avi-Yonah & Linda Swartz, Virtual Taxation: Source-Based Taxation in the Age of Derivatives, 2 DERIVATIVES 247 (1997); Edmund S. Cohen, Individual International Tax Planning Employing Equity Derivatives, 4 DERIVATIVES 52 (1998). See Gregory May, Flying on Instruments: Synthetic Investment and the Avoidance of Withholding Tax, 412 PLI/TAX 1215 (1997) for other techniques relying on the portfolio interest exemption. 14 See Yaron Z. Reich, Taxing Foreign Investors’ Portfolio Investments: Developments and Discontinuities, 98 TAX NOTES TODAY 114-71 (June 15, 1998). 15 See, e.g., JACOB A. FRENKEL ET AL., INTERNATIONAL TAXATION IN AN INTEGRATED W ORLD 203-7, 214-16 (1991) (“We conclude therefore that no capital-income tax whatsoever is imposed … if capital flight to the rest of the world cannot be effectively stopped.”) For further discussion of this advice and its implications, see Part IV. 16 COMMISSION OF THE EUROPEAN COMMUNITIES, REPORT OF THE COMMITTEE OF INDEPENDENT EXPERTS ON COMPANY TAXATION 201 (1992) [hereinafter RUDING REPORT ].

7

The experience of Germany is a case in point: In 1988, Germany introduced a (relatively low) 10% withholding tax on interest on bank deposits but had to abolish it within a few months because of the magnitude of capital flight to Luxembourg (over $100 billion DM). In 1991, the German Federal Constitutional Court held that withholding taxes on wages but not on interest violated the constitutional right to equality; and the government, therefore, was obligated to reintroduce the withholding tax on interest but made it inapplicable to nonresidents.17 Non-residents may, however, be German residents investing through Luxembourg bank accounts and benefiting from the German tradition of bank secrecy vis-à-vis the government.18 The current situation resembles a multiple player assurance ("stag hunt") game: All developed countries would benefit from re-introducing the withholding tax on interest because they would gain revenue without fear that the capital would be shifted to another developed country. However, no country is willing to be the first one to cooperate by imposing a withholding tax unilaterally; and thus, they all defect (i.e., refrain from imposing the tax) to the detriment of all.19 In global terms, this outcome would make no difference if residence jurisdictions were able to tax their residents on foreign source interest (and dividend) income as required by a global personal income tax on all income “from whatever source derived.”20 However, as Joel Slemrod has written, “although it is not desirable to tax capital on a source basis, it is not administratively feasible to tax capital on a residence basis.”21 The problem is that residence country fiscal authorities generally have no means of knowing about the income that is earned by their residents abroad. Even in the case of sophisticated tax administrations like the IRS, tax compliance depends decisively on the presence of either withholding at source or information reporting. When neither is available, as in the case of foreign source income, compliance rates drop dramatically from over 90% to about 30%.22

See Leif Muten, International Experience of How Taxes Influence the Movement of Private Capital, 94 TAX NOTES INT ’L 49-17 (Mar. 14, 1994). 18 See Suzanne McGee, Tax Inquiry Turns Grayer in Luxembourg, W ALL ST . J., July 20, 1999, at C1; Tax Raids: Sleepless Nights in Frankfurt, Again, ECONOMIST , June 20, 1998, at 87. 19 The assurance game differs from the more familiar prisoners' dilemma in that the preferred outcome for all is mutual cooperation (rather than defection while the other parties cooperate). See M ANCUR OLSON, THE LOGIC OF COLLECTIVE A CTION: PUBLIC GOODS AND THE THEORY OF GROUPS (1971); M ICHAEL TAYLOR, THE POSSIBILITY OF COOPERATION (1987). 20 I.R.C. § 61. 21 Joel Slemrod, “Comment” in TANZI, supra note 11, at 144. 22 See General Accounting Office, Tax Administration: Approaches for Improving Independent Contractor Compliance, GAO/GGD-92-108 (July 23, 1992); I.R.S. Publication 1415, Table A-54 (July 1988). In the case of Germany, it is estimated that only about 10 to 20% of foreign source interest income is taxed; see Charles E. McLure, Jr., International Considerations in United States Tax Reform, in INFLUENCE OF TAX DIFFERENTIALS ON INTERNATIONAL COMPETITIVENESS 55 (1990) [hereinafter International Considerations]. 17

8

In the case of foreign source income, withholding taxes are not imposed for the reasons described above. As for information reporting, even though tax treaties contain an exchange of information procedure, it is vitally flawed in two respects: First, the lack of any uniform worldwide system of tax identification numbers means that most tax administrations are unable to match the information received from their treaty partners with domestic taxpayers. Second, there are no tax treaties with traditional tax havens, and it is sufficient to route the income through a tax haven to block the exchange of information.23 For example, if a Mexican national invests in a U.S. bank through a Cayman Islands corporation, the exchange of information article in the U.S.-Mexico tax treaty would not avail the Mexican authorities. The IRS has no way of knowing (given bank secrecy) that the portfolio interest that is paid to the Caymans is beneficially owned by a Mexican resident covered by the treaty. 24 The resulting state of affairs is that much of the income from portfolio investments overseas escapes income taxation by either source or residence countries. Latin American countries provide a prime example: It is estimated that following the enactment of the portfolio interest exemption about $300 billion fled from Latin American countries to bank accounts and other forms of portfolio investment in the U.S.25 Most of these funds were channeled though tax haven corporations and therefore were not subject to taxation in the country of residence. For all developing countries, various estimates of the magnitude of capital flight in the 1980s average between $15 billion and $60 billion per year.26 Nor is the problem limited to developing countries: Much of the German portfolio interest exemption benefits German residents who maintain bank accounts in Luxembourg, and much of the U.S. portfolio interest exemption benefits Japanese investors who hold U.S. Treasuries and do not report the income in Japan. 27 23

“Traditional” tax havens are countries with a low rate of tax and bank secrecy laws that protect the identity of foreign investors therein. For an attempt at definition, see ORGANIZATION FOR ECONOMIC COOPERATION AND DEVELOPMENT , HARMFUL TAX COMPETITION : A N EMERGING GLOBAL ISSUE 55 (1998) [hereinafter OECD REPORT ]. They are to be distinguished from countries with generally high tax rates that provide specific tax exemptions to foreigners; such “production” and “headquarters” tax havens are discussed below. 24 See generally TANZI, supra note 11, at 78-89. The U.S. has negotiated exchange of information agreements with several Caribbean countries to attempt to overcome this problem, but one holdout is sufficient. 25 See Charles E. McLure, Jr., U.S. Tax Laws and Capital Flight from Latin America, 20 INTER-A MERICAN L. REV. 321 (1989) [hereinafter Capital Flight from Latin America]; CHANDER KANT , FOREIGN DIRECT INVESTMENT AND CAPITAL FLIGHT 1 (Princeton Studies in Int’l Fin. No. 80, 1996). McLure has written that “I must admit that, as Deputy Assistant Secretary of the Treasury at the time this action [the portfolio interest exemption] was taken, I did not fight early enough, long enough, or hard enough against its enactment,” mostly because of the effect on Latin American countries. International Considerations, supra note 22, at 9 & n. 12. It is interesting that these estimates were made despite the possibility of investing in some forms of U.S. holding (e.g., bank accounts or investment through the Netherlands Antilles) prior to 1984. Apparently the portfolio interest exemption made a significant difference. But in any case, the bottom line results hold regardless of when the practical abolition of withholding took place. 26 KANT , supra note 25, at 21. 27 See Germans Find Tax Evasion Good Sport, FINANCIAL TIMES (London) June 19, 1999, at 3; Mark Milner, Dresdner Tax Arrests, GUARDIAN (London), Jan. 4, 1996, at 15; Hiroshi Oda, Japan, in BUTTERWORTHS INTERNATIONAL GUIDE TO M ONEY LAUNDERING AND PRACTICE 109, 110 (Richard Parlour ed., 1995)

9

Even in the case of the U.S., it is questionable how much tax is actually collected on portfolio income earned by U.S. residents abroad other than through mutual funds. One estimate has put capital flight from the U.S. in 1980-82 as high as $250 billion.28 Thus, in the absence of withholding taxes or effective information exchange, income from foreign portfolio investments frequently escapes being taxed by any jurisdiction. This is particularly significant because the flows of portfolio capital across international borders have been growing recently much faster than either world gross domestic product or foreign direct investment.29 It is currently estimated that international capital flows amount to $1 trillion a day; and although this figure is much larger than income from capital, it gives a sense of the magnitudes at stake.30 This situation has led knowledgeable observers like Richard Bird to write that “the weakness of international taxation calls into question the viability of the income tax itself … If something is not done to rectify these problems soon, the future of the income tax is bleak.”31 Other authors have written articles like “Can Capital Income Taxes Survive in Open Economies?” and “Is There a Future for Capital Income Taxation?”32 In section (c) below, we will try to assess the possible revenue losses from portfolio investment abroad, and whether they really amount to a threat to the income tax. In Part V, we will consider possible solutions.

Michael P. Dooley, Comment on the Definition and Magnitude of Recent Capital Flight, in CAPITAL FLIGHT AND THIRD W ORLD DEBT 79 (Donald R. Lessard & John Williamson eds., 1987). 29 Foreign direct investment (FDI) out of the U.S. grew from $396 billion in 1980 to $1,024 billion in 1997, while portfolio investment out of the U.S. grew from $62 billion in 1980 to $1,446 billion in 1997. Similarly, FDI into the U.S. grew from $126 billion in 1980 to $752 billion in 1997, while foreign portfolio investment into the U.S. grew from $90 billion to $2,240 billion in the same period. See U.S. Department of Commerce, Survey of Current Business (July 1998). Because of these numbers, the literature that found capital to be immobile internationally is puzzling, see Martin Feldstein & Charles Horioka, Domestic Savings and International Capital Flows, 90 ECON. J. 314 (1980). For an explanation of capital immobility in terms of information asymmetry, see Roger H. Gordon & A. Lans Bovenberg, Why Is Capital So Immobile Internationally? Possible Explanations and Implications for Capital Income Taxation, 86 A MER. ECO. REV. 1057 (1996), and for an extension deriving implications for optimal tax policy see Assaf Razin et al., Tax Principles and Capital Inflows: Is It Efficient To Tax Nonresident Income? (1996) (unpublished NBER Working Article 5513). For a recent review of the evidence and arguments why Feldstein and Horioka may have been wrong in 1980 and are probably wrong today, see Michael P. Devereux, Investment, Saving and Taxation in an Open Economy, 12 OXFORD REV. OF ECON. POLICY 90 (1996). 30 See Peter D. Sutherland, Sharing the Bounty; Challenges of Globalization, 148 BANKER 16 (1998) (estimating that average daily foreign exchange transactions grew from $15 billion in 1973 to $1.2 trillion in 1995 and that international capital flows now exceed trade flows by 60 to 1); Erik R. Peterson, Looming Collision of Capitalism?, 17 W ASH. Q. 65 (1994) (estimating daily global capital movements increased to well over $1 trillion by 1992). 31 Richard Bird, Shaping a New International Tax Order, 42 BULL. FOR INT ’L FISCAL DOCUMENTATION 292, 303 (1988). 32 Roger H. Gordon, Can Capital Income Taxes Survive in Open Economies, 97 J. OF FINANCE 1159 (1992); Jack Mintz, Is There a Future for Capital Income Taxation? (1992) (OECD Working Paper 108); see generally Charles E. McLure Jr., Tax Policies for the 21st Century, Keynote Address to the 50th Congress of the International Fiscal Association, Geneva (1996) [hereinafter Tax Policies for the 21st Century]. 28

10

b. Taxation of Multinationals: Headquarters and Production Tax Havens, and Permanent Establishments

There are three types of jurisdictions that may, under currently accepted rules, impose a corporate income tax on income from the sale of goods or provision of services across national borders. The demand jurisdiction (where consumption of the goods or services takes place) may impose a tax if the goods or services are provided through a “permanent establishment” (i.e., some physical presence in the demand jurisdiction) or through a subsidiary to the extent of the income attributable thereto. The supply jurisdiction (where production of the goods or services takes place) may impose a tax on the income attributable to such production. These two jurisdictions have the right to tax based on source. In addition, a third jurisdiction has a residual right to tax income that is not taxed at source based on being the jurisdiction where the corporation is resident (i.e., is incorporated or managed and controlled). The argument of this section of the article can be summarized as follows: Out of these three jurisdictions, the one most likely to want to tax multinationals on income derived therein, the demand jurisdiction, does not under currently accepted rules have the right to tax such income in an increasing number of cases. The jurisdictions that do have such a right (the supply and corporate residence jurisdictions) are unlikely to want to do so because of tax competition. Thus, income earned by multinationals from cross-border transactions is likely, to a significant extent, to escape the income tax altogether. (i) Demand Jurisdictions and the Permanent Establishment Threshold. Out of the three jurisdictions that have the right to impose an income tax on multinational enterprises, the one most likely to have a significant corporate tax rate is the demand jurisdiction because large consumer markets typically have corporate income taxes imposed on foreign as well as domestic corporations. Under the current international tax regime (as embodied in the tax treaty network) in order for a seller to be taxable in a demand jurisdiction it is necessary for the seller to have a physical presence (a “permanent establishment” or, in U.S. terms, a “U.S. trade or business”) in the demand jurisdiction. Traditionally, multinationals typically had to establish such a physical presence (a branch or a subsidiary) in demand jurisdictions in order to gain significant market share or to avoid tariff barriers. However, the tariff barriers have all but disappeared, and the rise of electronic commerce makes it increasingly feasible to sell large quantities of goods or services into a demand jurisdiction without establishing a physical presence therein. Most authoritative discussions of taxation of electronic commerce conclude that the most common forms of such commerce (a web page or even a server located in a jurisdiction) do not amount to a permanent establishment and thus fall under the taxing threshold of the demand jurisdiction.33 Therefore, See Office of Tax Policy, Treasury Dep’t, Selected Tax Policy Implications of Global Electronic Commerce (1996), reprinted in Daily Tax Rep. (BNA), Nov. 22, 1996, at L-1. 33

11

while multinationals currently pay some tax in demand jurisdictions in which they have subsidiaries engaged in distribution, this tax base is rapidly diminishing in the face of technological developments and cannot be relied on in the long run if the permanent establishment threshold remains unchanged.34 (ii) Supply Jurisdictions and Production Tax Havens. Supply jurisdictions have traditionally imposed significant corporate income tax rates, no less than demand jurisdictions (in fact, most production used to take place in the same jurisdictions as most consumption, i.e., in the developed world). However, a crucial development in the last decade or so is the increasing importance of production tax havens, i.e., jurisdictions that grant tax holidays specifically to production facilities located therein by foreign corporations while continuing to have an income tax imposed on domestic corporations and individual residents. This last feature distinguishes such production tax havens from traditional offshore tax havens that have no corporate income tax (or sometimes any significant tax) at all. This distinction is crucial because it means that a foreign investor in a production tax haven can enjoy the benefits of government services financed by non-tax haven levels of taxation imposed on relatively immobile factors of production such as labor and land while itself paying little or no corporate income tax. Tax competition has led to the proliferation of production tax havens. There are currently at least 103 countries that offer special tax concessions to foreign corporations that set up production facilities or administrative facilities within their borders.35 They include developed countries such as Belgium, Ireland, and Israel, and developing countries such as Malaysia and India. The extent of the tax holiday varies, but generally the statutory tax rate is reduced to 10% or less for foreign corporations investing in specified types of facilities or areas within the jurisdiction while a higher tax rate (at least between 30% and 40%, which is the normal range for OECD members) is imposed on local corporations and a personal income tax at even higher rates, as well as a VAT, is imposed on resident individuals.36 Studies by economists have shown that such tax incentives are quite widespread and that investments in such countries are significant. For example, Altshuler and Newlon have studied U.S. multinationals’ tax returns for 1986 and found that out of 1,827 foreign subsidiaries (“CFCs”) of such multinationals, 659 were located in countries with a tax rate below 20%. These CFCs had assets of $51 billion and earnings of $5.2 billion in 1986 which represented For a fuller exposition of this theme, see Reuven S. Avi-Yonah, International Taxation of Electronic Commerce, 52 TAX L. REV. 507 (1997) [hereinafter International Taxation of Electronic Commerce]. For proposals to change the definition of permanent establishment (and U.S. trade or business) to enable demand jurisdictions to levy tax without the need for a physical presence, see Part V below. 35 See RAYMOND VERNON, IN THE HURRICANE'S EYE: THE TROUBLED PROSPECTS OF M ULTINATIONAL ENTERPRISES, Table 2-1 (1998); UNITED NATIONS, INCENTIVES AND FOREIGN DIRECT INVESTMENT , Table III.2 (1996). 36 See VERNON, supra note 35. 34

12

over 40% of the total assets and earnings of all the CFCs.37 Almost three quarters of U.S. multinationals studied had CFCs in both high and low tax jurisdictions.38 Similarly, Hines and Rice have documented the extent of U.S. investment in foreign tax havens for 1982.39 They show that while the havens had only 1.2% of world population and 3.0% of world GDP, their share of the foreign operations of U.S. MNEs was 26% of assets, 21.4% of equity, and 30.6% of net income (but only 4.3% of employment and 4.2% of property, plant, and equipment).40 Two specific examples may help to illustrate this phenomenon. Intel Corporation, a top ten multinational, has operations in more than thirty countries around the globe. The company states that “[a]n Intel chip developed at a design center in Oregon, might be manufactured at a wafer fabrication facility in Ireland, packaged and tested in Malaysia, and then sold to a customer in Australia. Another chip might be designed in Japan, fabricated in Israel, packaged and tested in Arizona, and sold in China.”41 Specifically, outside the U.S., Intel has major manufacturing facilities in Puerto Rico, China, Malaysia, the Philippines, Ireland, and Israel.42 Thus, outside the U.S., all of Intel’s manufacturing facilities are located in countries granting tax holidays.43 This phenomenon is not confined to new-fangled high tech corporations like Intel. For example, new assembly plants for General Motors cars built since 1990 are located in Gravatai, Brasil (Opel cars); Rosario, Argentina (Chevrolet Corsica); Silao, Mexico (Chevrolet Suburban, Silverado); Gliwice, Poland (Opel Astra); Rayong, Thailand (Opel Astra); Jakarta, Indonesia (Opel Blazer, Vectra, Astra); and Shanghai, China (Buicks).44 Except for Mexico, all of these locations are in production tax havens.45 The rise of electronic commerce is likely to make it easier to locate production in tax havens. The Intel example, involving traditional (i.e., non-digitizeable) goods, is a case in point: Intel’s ability to coordinate production across the globe and locate its production facilities in production tax havens like Malaysia is based on modern communication technology including See Rosanne Altshuler & T. Scott Newlon, The Effects of U.S. Tax Policy on the Income Repatriation Patterns of U.S. Multinational Corporations, in STUDIES IN INTERNATIONAL TAXATION 92, 93 (Alberto Giovannini et al. eds., 1993) [hereinafter Effects of U.S. Tax Policy]. 38 Id. 39 See James R. Hines & Eric M. Rice, Fiscal Paradise: Foreign Tax Havens and American Business, 109 Q. J. OF ECON. 149 (1994). 40 See id. at Table 1. The share of net income in the havens is much higher than that reported for continental Europe (14.8%), which has 10.7% of world population and 34.5% of world GDP. 41 See (visited Feb. 1, 1999). 42 See id. 43 Intel has sales facilities in high-tax demand jurisdictions such as Japan, Germany, and the U.K. But this does not mean that Germany can tax Intel on chips sold in Europe; it all depends on whether the income can be attributed to the office in Germany, which it would not be if the sale was negotiated and concluded elsewhere. If Intel could sell its goods in electronic commerce, the need for such permanent establishments would be reduced (in recent ads, Intel has boasted of selling over $1 billion worth of chips in electronic commerce). 44 See Robyn Meredith, The Brave New World of General Motors, N. Y. TIMES, Oct. 26, 1997, at sec. 3, 1. 45 The Mexican location benefits from the North American Free Trade Agreement. 37

13

corporate intranets.46 Where goods can be conveyed in digital form, such as software, it is even easier to locate the entire operation in tax havens: The software can be written anywhere (e.g., in India or Israel where Microsoft has major operations) and transmitted elsewhere over secure corporate intranets. Information services in general have no inherent source, and on-line pornography, which is still one of the most profitable types of electronic commerce, is run largely from Guyana for both regulatory and tax related reasons.47 Are these locations influenced by the tax regime in the host country? There have been several studies by economists that examined the importance of tax differentials in location decisions. Hines has recently summarized ten quantitative studies of U.S. direct investment abroad and ten quantitative studies of foreign direct investment in the U.S. He concludes that although taxes are not the only determinant of the location of investments, “[t]he answer that emerges in a variety of contexts and from a variety of approaches is that, in spite of all the other economic and political considerations that are clearly important, taxation exerts a significant effect on the magnitude and location of FDI.”48 Hines estimates that the studies are generally consistent with a -0.5 elasticity of investment with respect to tax rates at the currently prevailing rates.49 These statistical findings are corroborated by Wilson who interviewed managers from nine firms regarding sixty-eight location decisions and concluded that “tax considerations largely dictate location decisions for business activities … such as administrative and distribution centers.”50 As for production locations, taxes are an important consideration in the location decision although they rarely dominate the decision process.51 The economic studies have also demonstrated the ability of multinationals to shift profit to low-tax jurisdictions through transfer pricing or otherwise. For example, Harris and his 46

Intel’s 1998 10-K states that if air traffic between the U.S. and its overseas locations were interrupted this would have a material adverse effect on its operations. This problem is not raised for goods that can be digitized. 47 See David Tillinghast, Comments on Professor Peroni’s Article on Reform of the U.S. International Income Tax Rules, 51 UNIV. M IAMI L. REV. 1013, 1014 (1997). 48 James R. Hines, Tax Policy and the Activities of Multinational Corporations, in FISCAL POLICY: LESSONS FROM ECONOMIC RESEARCH 414-415 (Alan J. Auerbach ed., 1997) [hereinafter Tax Policy]; see also James R. Hines, Lessons from Behavioral Responses to International Taxation, 52 Nat’l Tax J. 305 (1999). Earlier studies on this topic depended on time series evidence which is more doubtful than recent evidence from cross-sectional studies. See Deborah Swenson, The Impact of U.S. Tax Reform on Foreign Direct Investment in the United States, 54 J. PUB. ECON. 243 (1994); Timothy Scott Newlon, Tax Policy and the Multinational Firm’s Financial Policy and Investment Decisions (Ph.D. dissertation, Princeton University, 1987); David G. Hartman, Tax Policy and Foreign Direct Investment in the United States, 37 NAT ’L TAX J. 475 (1984). 49 See Tax Policy, surpa note 48, at 415; see also Marcel P. Devereux & Rachel Griffith, Taxes and the Location of Production: Evidence From a Panel of U.S. Multinationals, 68 J. PUB. ECON. 335 (1998) (concluding, on the basis of a study of 1,632 U.S. multinationals and their decisions where to invest in Europe in 1980- 1994, that these decisions were significantly affected by average effective tax rates). 50 G. Peter Wilson, The Role of Taxes in Location and Sourcing Decisions, in STUDIES IN INTERNATIONAL TAXATION 195, supra note 37, at 229. 51 See id. In general, as argued in Part IV below, taxes are usually not the dominant factor affecting real investment decisions but are decisive at the margin in the choice between two locations that are otherwise equally attractive.

14

colleagues have concluded on the basis of a sample of 200 U.S. manufacturing firms for 19841988 that having a subsidiary in Ireland or in the four Asian “dragons” (Taiwan, Korea, Singapore, and Hong Kong), all jurisdictions offering a tax holiday for foreign investors, is associated with lower U.S. taxes as a fraction of U.S. assets or U.S. sales “in a way that is consistent with tax-motivated income shifting.”52 Hines surveys several studies on the financial behavior of multinationals, including the use of debt and transfer pricing, and concludes that “[t]he evidence indicates that the financial behavior of multinational corporations is quite sensitive to tax considerations, though not completely determined by them … In some instances it may be less valuable to test the standard hypothesis that taxes influence financial behavior than to test the alternative hypothesis that only taxes influence financial behavior.”53 Some recent studies based on actual data derived from U.S.-based multinational tend to bolster the view that (a) tax competition among countries has driven effective tax rates down, and (b) U.S. investment abroad has become more sensitive to tax rates. Altschuler, Grubert, and Newlon used data from U.S. Treasury corporate tax returns for 1984 and 1992 to assess both questions for manufaturing affiliates of U.S.-based multinationals.54 Average effective tax rates for CFCs in about sixty countries were calculated based on dividing total income taxes paid by their total earnings and profits (a measure defined by the Internal Revenue Code that approximates “book” income).55 They found that average effective tax rates in manufacturing fell by more than 15% between 1984 and 1992. The statutory tax rates likewise fell by 14%, indicating that rate reductions were not compensated by base broadening measures.56 The authors also found that the elasticity of investment to tax rate differentials in open economies increased from 1984 to 1992, indicating that the location of real manufacturing capital by manufacturing firms may have become more sensitive to tax rates in this period. This result held when countries with populations of less than 1 million, including most traditional tax havens, were excluded from the calculations. The authors conclude that their results are “consistent with increasing international mobility of capital and globalization of production.”57

David Harris et al., Income Shifting in U.S. Multinational Corporations, in STUDIES IN INTERNATIONAL TAXATION 277, supra note 37, at 301. 53 Tax Policy, supra note 48, at 430-431. 54 See Rosanne Altshuler et al., Has U.S. Investment Abroad Become More Sensitive to Tax Rates?, (November 1997) (unpublished article presented at NBER International Tax Conference no. 6383) [hereinafter U.S. Investment Abroad]. 55 As the authors note, this measure is much better than theoretical calculations of either marginal effective tax rates or average effective tax rates because such calculations make a large number of assumptions and use only rough approximations of country tax regimes. In addition, the use of Treasury data is superior to using publicly available data because precise numbers are available for each controlled foreign corporation (CFC). 56 Similarly, Lucy Chennells & Rachel Griffith, Taxing Profits in a Changing World (Institute for Fiscal Studies, 1997) report that average tax rates based on publicly available data for firms from ten developed countries show a decline from 40% in 1985 to 32.6% in 1994. Chennells and Griffith are skeptical about the existence of tax competition, but that may be because they focused only on a sample of ten developed countries and included only one country (Ireland) that can be classified as a production tax haven. 57 U.S. Investment Abroad, supra note 54, at 18. 52

15

A recent study by Harry Grubert focused directly on the questions whether there is evidence of tax competition and U.S. multinationals response thereto based on U.S. Treasury data for 1984 and 1992.58 He found that the mean average effective tax rate fell by almost 10% from 1984 to 1992, but there was no notable convergence in tax rates. Grubert interprets this as suggesting that there was no tax competition, but this does not follow from the data: It is possible that countries with high tax rates in 1984 reduced them because of competition with countries with low tax rates, but that the low tax countries responded with further decreases resulting in a lower average rate overall but no convergence.59 Significantly, Grubert found that tax reductions were greater in countries with a population of less than 15 million suggesting that small, open economies were more sensitive to capital movements (countries with less open economies cut their tax rates less). Using firm-level data, Grubert also found that the firms with low foreign tax rates in 1984 had the lowest rates in 1992 suggesting that those firms were highly mobile or more aggressive in their tax planning. Whether the tax competition phenomenon is negative or positive will be considered in Part III below. It is important, however, to note here that from the point of view of the investing multinational there is no added cost to investing in a low tax jurisdiction. This is in contrast with the standard Tieboutian analysis of tax competition in the state and local context in which individuals choose jurisdictions based on their preferred level of government services and bear the cost of reduced services if they choose the lower tax jurisdiction. 60 Here, the level of services provided by the host country is fixed before the tax holiday and is based on the revenues that were collected without regard to the tax holiday. Typically, the holiday jurisdiction would try to provide the same level of services with the tax holiday in place, replacing the revenue needed by increasing taxes on relatively immobile factors of production (i.e., land and labor). From the point of view of the multinational, having decided to make the investment somewhere, getting the tax holiday is a pure windfall since it can typically choose among several jurisdictions with similar levels of government services but different tax rates.61 (iii) Corporate Residence Jurisdictions and Headquarters Tax Havens.

See H. Grubert, Tax Planning by Companies and Tax Competition by Governments: Is There Evidence of Changes in Behavior?, (November 1997) (unpublished article presented at NBER International Tax Conference). 59 See id. Grubert also notes that there was no convergence in the European Economic Community. This contradicts the findings of the Ruding committee, see F. VANISTENDAEL & M ALCOLM GAMMIE, THE RUDING COMMITTEE REPORT ON EUROPEAN CORPORATE TAX HARMONIZATION (1992). 60 See Charles M. Tiebout, A Pure Theory of Local Expenditures, 64 J. POL. ECON. 416 (1956). 61 For example, Intel was able to get a $600 million tax concession from the Israeli government by threatening to locate a new facility in Ireland. This represented Israel’s entire development budget for the year. The government scientist objected to the award, arguing that the money was better spent in developing local hitech companies, but Intel’s prestige carried the day. See Evelyn Gordon, Reconsider Size of Aid Promised to Intel Harish, JERUSALEM POST , Dec. 14, 1995, at 8. The U.S. multinationals studied by Devereux and Griffith decided to invest in Europe regardless of taxes; only the choice of location within Europe was influenced by the effective average tax rate, see Devereux & Griffith, supra note 49. 58

16

The international tax regime allocates primary jurisdiction to tax active business income to the source jurisdiction. In situations where the source country refrains from exercising its right to tax (which it commonly does for the reasons described above), residual taxation is possible by the corporate residence jurisdiction. However, it is unlikely that corporate residence jurisdictions will actually impose a current tax on the foreign source active business income earned by multinationals whose headquarters are located in those jurisdictions (or whose parent corporation is incorporated therein). The reason is that most corporate residence jurisdictions either exempt foreign source active business income earned by their multinationals or permit their multinationals to avoid paying tax on the income earned by their foreign subsidiaries until actually repatriated to the parent in the form of dividends or otherwise (“deferral”). For example, under current U.S. rules, deferral applies to active business income earned by subsidiaries of U.S. parent corporations abroad.62 Significantly, deferral does not depend on whether or not the income was taxed abroad. Under deferral, a U.S. based multinational can avoid paying taxes indefinitely on its active foreign source income as long as it does not need to repatriate it.63 Under the well known Cary Brown formula, deferral is equivalent to exemption of the yield on the amount deferred.64 Thus, the income earned on foreign-source profits of U.S. multinationals becomes exempt from U.S. tax. 65 Another way of stating the advantage of deferral is that if deferral lasts long enough, the present value of any tax imposed when the income is eventually repatriated approaches zero. The obvious solution to this problem would be to repeal deferral, that is, to apply subpart F to all income rather than just to “subpart F” (i.e., generally passive) income as the Kennedy administration originally suggested in 1961. This solution has been repeatedly suggested by various administrations, incorporated in several recent legislative proposals, and See I.R.C. §§ 951-960 (“Subpart F”). In the 1980’s, U.S. multinationals like GM were forced to repatriate profits because their U.S. revenues were too low to sustain their dividend payouts to shareholders, but that may no longer be the case. But see James R. Hines, Altered States: Taxes and the Location of Foreign Direct Investment in America (1993) (unpublished NBER article). 64 See, e.g., Alvin C. Warren Jr., Would a Consumption Tax be Fairer Than an Income Tax, 89 YALE L.J. 1081 (1980). 65 See David G. Hartman, Tax Policy and Foreign Direct Investment, 26 J. PUB. ECON. 107 (1985). Hartman argues that the tax on repatriation is unavoidable and therefore deferral of foreign source earnings does not make a difference in terms of ultimate after-tax yield. He therefore concludes that FDI out of the reinvested earnings of foreign subsidiaries is unaffected by home country taxes. For example, if the parent’s tax rate is 30% and interest on earnings is 10%, 100 earned abroad and repatriated would be subject to tax of 30, leaving 70, and would grow to 77 in the following year; if the 100 were left abroad it would grow to 110 and be subject to tax of 33 when repatriated, leaving likewise 77. But Hartman assumes that there are no other domestic taxes except the tax on dividends. If, more realistically, there were to be a tax of 30% on the 7 of interest income in the first case, it would leave the taxpayer only 74.9, while in the second case the full 77 represent after tax income. This is the same as stating that the yield in the second case is exempt. 62 63

17

endorsed by academic commentators.66 But it has never gotten off the ground. In fact, the recent trend has been in the opposite direction: Code section 956A, which limited the ability of multinationals to avoid repatriation by imposing current taxation on excess passive assets held abroad, was repealed in 1996; and the extension of “check the box” to foreign entities means that U.S. multinationals can generally choose whether to have deferral or not based on the specific tax characteristics of the foreign entity and regardless of its foreign legal status as a corporation or otherwise.67 Why has deferral been so successfully maintained against all criticism, and why have other jurisdictions maintained an exemption for foreign source active business income? Consider what would happen were the U.S. to end deferral. In that case, it appears likely that parent corporations of new multinationals, especially in new industries like electronic commerce, will not be incorporated in the U.S.68 Instead, they will be incorporated in a country like Belgium which does not tax holding corporations (although it taxes other corporations quite heavily and even has a classical system like the U.S.). Gordon and MacKie-Mason have shown that a firm’s decision whether to incorporate is significantly affected by tax considerations.69 Similarly, the decision where to incorporate can be influenced by tax considerations.70 This can be shown in several ways. First, one can consider the evidence presented by Wilson based on interviews with corporate management that “tax considerations largely dictate location decisions for … administrative and distribution centers.”71 This is precisely the type of operation envisaged by Belgium: a corporate headquarters engaged in supervising worldwide manufacturing operations (located in production tax havens). Belgium is in effect marketing itself as a “corporate headquarters tax haven.” Like 66

For discussions of various proposals to abolish deferral see, e.g., Peter Merrill & Carol Dunahoo, Runaway Plant Legislation: Rhetoric and Reality, 74 TAX NOTES 221 (1996); Jane G. Gravelle, Foreign Tax Provisions of The American Jobs Act of 1996, 74 TAX NOTES 1165 (1996); Paul W. Oosterhuis & Rosanne M. Cutrone, The Cost of Deferral’s Repeal: If Done Properly, It Loses Billions, 58 TAX NOTES 765 (1993); Daniel J. Frisch, The Economics of International Tax Policy: Some Old and New Approaches, 90 TAX NOTES TODAY 19-7 (Jan. 24, 1990). 67 See, e.g., Robert J. Peroni, Back to the Future: A Path to Progressive Reform of the U.S. International Income Tax Rules, 51 UNIV. M IAMI L. REV. 975 (1997); Reuven S. Avi-Yonah, To End Deferral As We Know It: Simplification Potential of Check-The-Box, 74 TAX NOTES 219 (1996). The 1997 Tax Act also expanded deferral, e.g., by eliminating the CFC/PFIC overlap and by limiting the application of Subpart F to the securities and insurance industries. See also I.R.S. Notice 98-35, 1998-1 C.B. 300 (discussing the recent debate around and ultimate withdrawal of Notice 98-11 which attempted to limit the use of check the box to avoid Subpart F). Grubert, supra note 58, suggests that his data indicate that U.S. CFCs in Caribbean tax havens pay U.S. tax on less than 50% of their book income, which indicates that they are effectively avoiding the application of Subpart F (e.g., by the use of hybrids). 68 See, e.g., GARY C. HUFBAUER, U.S. TAXATION OF INTERNATIONAL INCOME: BLUEPRINT FOR REFORM 96-97 (1992). 69 See Roger H. Gordon & Jeffrey K. MacKie-Mason, Tax Distortions to the Choice of Organizational Form, 55 J. PUB. ECON. 279 (1994). 70 Average effective tax rates are significantly affected by choice of corporate domicile. See Julie H. Collins & Douglas A. Shackelford, Corporate Domicile and Average Effective Tax Rates: The Cases of Canada, Japan, the United Kingdom, and the United States, 2 INT ’L TAX AND PUB. FIN. 55 (1995). 71 Wilson, supra note 50, at 229.

18

production tax havens, such headquarters tax havens are proliferating because of tax competition. Second, one can consider industries that are taxed exclusively on a residence basis because taxation at source is considered too difficult or too likely to lead to multiple taxation. Prominent examples are the shipping and commercial insurance industries which are not subject to source-based income taxation because there is no permanent establishment and production can be located in tax havens. One finds that in those cases the corporate residence is frequently also in a tax haven (Liberia and Panama are prominent examples for shipping, Bermuda for insurance) and therefore no income tax is levied (excise taxes are typically levied at source as a minimal but not adequate substitute). Third, there is evidence that even the current loose U.S. anti-deferral rules may have induced attempts to change corporate residence to avoid them. The U.S. has recently changed its tax rules to impose a tax on publicly-traded U.S. corporations that wish to change the residence of their corporate parent to another jurisdiction. 72 This was prompted by a wellpublicized case of a cosmetics distributor that was able to reincorporate tax-free in the Cayman Islands to avoid application of Subpart F to its future subsidiaries.73 Reincorporating in Panama or in the Caymans may be a step most U.S. multinationals would be unwilling to take. But establishing the corporate headquarters and formal corporate residence in Belgium, even when the entrepreneurs are U.S. residents, would seem to carry little business risk.74 Moreover, the shares of most multinationals are currently traded on several exchanges so that it would not appear to matter much which country the parent is incorporated in (since the shareholder protection laws of the trading jurisdiction will typically apply). While this may be unlikely for old, established U.S. corporations like General Motors or even Intel, 75 the major players in a new industry have yet to be incorporated, and future incorporation decisions may well be influenced by tax considerations.76 See Notice 94-46, 1994-1 C.B. 356. Similarly, in 1982 McDermott, Inc. changed its site of incorporation to Panama to avoid current U.S. tax of its foreign source income leading to the enactment of section 1248(i) (the “anti-inversion” provision), which was expanded in the 1997 tax act. See James R. Hines, Jr., The Flight Paths of Migratory Corporations, 6 J. A CCT ., A UDITING, & FIN. 447 (1991). 74 The anti-deferral rules do not apply to a closely held foreign corporation earning active business income. 75 However, even established corporations may be acquired by foreign multinationals, as the Daimler acquisition of Chrysler shows. If the majority of the shareholders in the new parent corporation are foreign, such a transaction escapes I.R.C. § 367, so that any future growth of the combined multinational can occur abroad free of the confines of Subpart F (Germany, like many other countries, has a much looser anti-deferral regime). 76 Reincorporation in another jurisdiction is common among the U.S. states, where both corporate and tax considerations have led many corporations to move to New Jersey and then to Delaware. See, PHILLIP I. BLUMBERG, THE M ULTINATIONAL CHALLENGE TO CORPORATION LAW 56-58 (1993) (discussing New Jersey's role in this evolution); Henry N. Butler, Nineteenth-Century Jurisdictional Competition in the Granting of Corporate Privileges, 14 J. LEGAL STUD . 129 (1985) (discussing the origins of jurisdictional competition for corporate sites). 72 73

19

The Treasury Discussion Paper on taxation of electronic commerce suggests that because of such considerations “a review of current residency definitions and taxation rules may be appropriate.”77 But it is hard to see what definition of corporate residence can be adopted that will avoid these problems. Looking at the corporation’s place of management and control seems difficult because the rise of corporate intranets means that there is no longer any need for corporate boards to meet in one physical location, and corporate management can be dispersed in many different countries (including headquarters tax havens) and communicate via e-mail or video conferencing on secure intranets. Looking at the residence of shareholders to establish corporate residence is likewise problematic because (as stated above) the shares of most multinationals now trade on several exchanges, and so there would be many corporate residences and no clear way to divide the tax base among them. 78 As long as residence jurisdictions compete with each other for the location of corporate headquarters, effective residence-based taxation seems unlikely. To sum up: Corporate income from cross-border transactions is increasingly unlikely to be taxed in the demand jurisdiction because of the absence of a permanent establishment. It is also unlikely to be taxed in the supply jurisdiction because production can take place in production tax havens. Finally, taxation by the corporate residence jurisdiction is likewise unlikely, at least on a current basis because of deferral and possibly not at all if corporate residence can be established in headquarters tax havens. Thus, unless current rules are changed, such income is increasingly likely to escape tax altogether. Whether this situation is problematic will be considered in Part III. c.

Revenue Losses

The preceding discussion has indicated that it is possible to avoid income taxation on cross-border portfolio and direct investment. But how much revenue is actually lost in this way? This turns out to be a much harder question to answer in the present state of the publicly available data. The recent OECD report on “Harmful Tax Competition: An Emerging Global Issue” gives as its main rationale for curbing such tax competition the risk that if nothing is done about it the revenue base of OECD member countries will erode.79 At the very beginning of the chapter containing its recommendations, the OECD Report states, “Governments cannot stand back while their tax bases are eroded through the actions of countries which offer taxpayers ways to 77

Office of Tax Policy, Department of the Treasury, Selected Tax Policy Implications of Global Electronic Commerce 7.1.5 (November 1996). Treasury article is posted on the Treasury Department's home page on the World Wide Web at http://www.ustreas.gov. It is also available from Tax Analysts as Doc 96-30614 (50 pages). 78 The possibility of ending corporate taxation altogether and taxing all income at the shareholder level is discussed and rejected on inter-nation equity grounds in Part IV(c) below. 79 OECD REPORT , supra note 23, at 8, 13-14. This Report is discussed in detail in Part V below.

20

exploit tax havens and preferential regimes to reduce the tax that would otherwise be payable to them.”80 Unfortunately, the OECD Report does not contain any numerical data to bolster its claim that tax bases are eroding as a result of harmful tax competition.81 Aggregate data on tax collections, reviewed in Part III(b) below, do not support the claim: There is no evidence that overall revenue from the personal or corporate income tax in OECD member countries has declined either as a percent of GDP or of total tax revenue from 1965 to 1995.82 However, these data do not distinguish between income from labor and capital, and it may be that a decline in the tax revenues from taxing capital is masked by a rise in revenues from taxing labor. This hypothesis would be consistent with the findings reported in Part III(b) below on changes in the overall tax mix. In this section, I will present the publicly available data on the revenue effects of tax competition. These data are quite sparse and incomplete but some of it is quite suggestive. Hopefully, the OECD will present fuller data in its future work on this issue. (i) Personal Income Tax Assessing the revenue losses from tax competition in the personal income tax area is difficult because it involves (illegal) tax evasion rather than (legal) tax avoidance.83 All developed countries and most developing countries include in taxable income foreign source passive investment income of their individual residents so that not reporting that income is illegal. However, as described in Part II, even developed countries find it difficult to actually collect the tax on such income in the absence of either withholding or information reporting.

Id. at 37. “The available data do not permit a detailed comparative analysis of the economic and revenue effects involving low-tax jurisdictions.” Id. at 17. The Report does note, however, that “the available data do suggest that the current use of tax havens is large, and that participation in such schemes is expanding at an exponential rate.” Id. 82 See Part III(b) below. Data for 1995-1999 do suggest a declining trend in corporate income tax rates in OECD and EU countries (from an average of 37.6% to 34.9% in the OECD and from 39% to 36% in the EU). It is less clear whether this translates to lower effective tax rates or lower revenues. See James R. Hines, Jr., Lessons From Behavioral Responses to International Taxation, 52 NAT ’L TAX J. 305-322 (June 1999). 83 In theory, it should be possible to obtain a good sense of the magnitude of revenue losses by summing up the withholding taxes that are foregone by OECD member countries through measures like the portfolio interest exemption. However, the most recent compilation of tax expenditure data by the OECD does not contain these data because most countries (including the U.S.) do not regard the reduction of withholding taxes as a departure from their normative tax base. See OECD, TAX EXPENDITURES: RECENT EXPERIENCES (1996). The only exception is Canada which estimates the tax loss from not withholding tax on interest on deposits at Cdn $325-430 million a year for the period 1993-2000, while the tax loss from not withholding tax on interest on long-term corporate debt is Cdn $460-550 million in the same period. Thus, Canada estimates that it loses about Cdn $1 billion (U.S. $650 million) a year from this source; although if it tried to collect it unilaterally, the investments would presumably move elsewhere. See Finance Canada Releases Government’s `98 Tax Expenditures Report, 98 TAX NOTES INT ’L 138-16 (July 20, 1998). 80 81

21

Tanzi and Shome have described the methods currently used to estimate the magnitude of tax evasion in general.84 The most commonly used method is the national accounts method which involves comparing the estimate of the base of a particular tax by the national accounts authorities and the base as reported to the tax authorities after making appropriate adjustments.85 Other methods include the sampling method (used by the IRS to estimate the so-called “tax gap” which was estimated at $195 billion in 1998),86 the budget survey method, a direct taxpayer survey, and indirect methods focusing on the underground economy. Indirect methods include the expenditure method which assumes that evaded income will show up as consumption, the physical input method which assumes that certain inputs (e.g., electricity) should relate to taxable outputs, and the monetary approach which associates evasion with currency holding. All of these methods have their problems, but they all reach consistently high estimates.87 For developed countries, the available estimates of the untaxed portion of the economy range from 6-20% of GDP in the U.S. (1980) to 20%-33% of GDP for Italy (1982).88 For developing countries, the estimates are as follows: Argentina (1989) Bolivia (1986) Brazil (1989) Chile (1986) Dominican Republic (1986) Ecuador (1986) Guatemala (1986) Honduras (1986) Mexico (1986) Peru (1986)

50% of GDP 44% of GNP 50% of GDP 14-32% of GNP 14-32% of GNP 30-32% of GNP 30-32% of GNP 30-32% of GNP 14-32% of GNP 35% of GNP89

However, these numbers include much domestic tax evasion that does not relate to tax competition. Kant has summarized the available data on capital flight using a variety of measures.90 These numbers are more relevant because it is reasonable to assume that most capital flight is invested in countries where it is not subject to taxation at source (given the wide availability of such investment opportunities, as described in Part II(a)). While Latin America 84

VITO TANZI & PARTHO SHOME, TAX EVASION: CAUSES, ESTIMATION M ETHODS, AND PENALTIES, A FOCUS ON LATIN A MERICA (1993). 85 See id. at 13. 86 See Amy Hamilton, The Tax Gap Game and Inklings of a Focus on Noncompliance, 79 TAX NOTES 933 (1998). These data are based on ten year old samples which may no longer be accurate. 87 See TANZI & SHOME, supra note 84, at 19. 88 See id. at 20-21 (Table 3). For Portugal, a 1975 study estimated that 30% of earned income and 40% of unearned income goes untaxed. 89 Id. 90 See KANT , supra note 25.

22

accounts for the largest share of capital flight, it is an important phenomenon in virtually all regions of the developing world.91 Estimates of capital flight from all developing countries for the period 1980-1988 range from $15 billion to $60 billion per year.92 The total capital from Latin America alone accumulated abroad was estimated in 1992 at $300 billion. 93 For developed countries, estimates are rare, but Dooley estimated capital flight from the U.S. in 1980-82 at $250 billion. 94 It seems reasonable to assume that the income from such capital is untaxed by either the source country or the residence country (for the reasons discussed in Part II). In that case, it is possible to estimate the revenue loss of the Latin American countries on $300 billion of capital flight at about $15 billion per year (assuming a 10% average rate of return and a 50% individual tax rate). This is a significant figure, but it is probably far too low because it assumes that the capital investment represents after-tax income. If, more realistically, one assumes that the capital was never taxed, the lost revenue from $300 billion held abroad at a 50% rate would be an initial $150 billion plus an additional $15 billion per year.95 The figure for all developing countries may be as high as twice that amount. These are very high numbers for developing countries and suggest that they could raise very significant tax revenues if a way could be found to tax income of their residents that is invested abroad.96 Such revenue gains could be very important: The UN estimated that basic social services (such as universal primary education for boys and girls) could be assured for all developing countries for a mere $30-$40 billion a year.97 The corresponding numbers for developed countries (such as Germany) could also be quite large, although much smaller as a percentage of total tax revenues. Thus, tax competition seems to lead to significant tax base erosion of the personal income tax which is still a major (if not the largest) tax base for most developed and developing countries. (ii) Corporate Income Tax Estimating revenue losses from tax competition in the corporate income tax should be easier than in the personal income tax area because (as explained in Part II(b)) the revenue loss involves legal tax avoidance, rather than illegal tax evasion. Most developed countries exempt See id. at 21. See id. 93 See id. at 1. 94 See Dooley, supra note 28, at 79. 95 This assumes that it would be possible to tax the underlying capital which may be more difficult than to tax the income therefrom because it can be left in tax havens (as are profits from drugs and other criminal activities). However, if all that is left to traditional tax havens are illegal profits, one can imagine drastic actions being taken against them (such as shutting off telecommunications links). 96 This assumes that the residence country, rather than the source country, would receive the tax revenue. The solution envisaged in Part V below is intended to make residence-based taxation possible. 97 UNITED NATIONS, W ORLD SUMMIT FOR SOCIAL DEVELOPMENT (1995) [hereinafter W ORLD SUMMIT FOR SOCIAL DEVELOPMENT ]. 91 92

23

or defer taxing the active business income of foreign subsidiaries of “their” MNEs while developing countries grant explicit tax holidays to those subsidiaries. Nevertheless, there is little publicly available data on revenue losses either as a result of deferral or exemption by home countries or as a result of preferential tax regimes in host countries. In the case of home countries, OECD member countries are the home countries for about 85% of the world’s MNEs.98 However, the most recent data on tax expenditures published by the OECD reveal that few OECD member countries quantify the revenue loss from granting exemption or deferral to the foreign source active business income of their MNEs. The only countries to do so are the U.S. and France. The U.S. estimates the revenue loss from “deferral of income from controlled foreign corporations” at between $2.2 billion (in 1997) and an estimated $3.4 billion (in 2003).99 France estimated the potential revenue gain from taxing its MNEs on their worldwide or consolidated profit in 1994 at FF 18 billion (about $3 billion as of December 1998).100 These numbers are significant, but they represent only a small portion of all MNEs. However, their true significance for each home country can only be assessed if one compares these figures to other tax expenditures and to the overall budget. For the U.S., $3 billion is a small tax expenditure since the largest tax expenditures are $50 billion and more.101 It is also a small figure in terms of overall tax receipts from the corporate income tax ($189 billion in 1998).102 For France, on the other hand, 3 billion in 1994 represented the second largest tax expenditure in the budget.103 It also represented a larger share of total receipts from the corporate tax (119 billion FF in 1994).104 These figures may explain why France was one of the OECD member countries most interested in pushing the tax competition issue (it co-chaired the Special Sessions of Tax Competition that drafted the OECD Report which focuses entirely on the corporate tax issue).105

98

UNITED NATIONS, W ORLD INVESTMENT REPORT , (1996). Budget of the U.S. Government, FY 1999, Analytical Perspectives, Ch. 5: Tax Expenditures, W ORLDWIDE TAX DAILY (Feb. 6, 1998). 100 See OECD, TAX EXPENDITURES: RECENT EXPERIENCES 62 (1996) [hereinafter TAX EXPENDITURES]. 101 See Budget of the U.S. Government, supra note 99. The largest U.S. tax expenditures for FY 1999 are the exclusion of pensions ($72,375 billion), insurance proceeds ($14,200 billion), and the home mortgage interest deduction ($53,695 billion). 102 Kenneth J. Kies, Prepared Testimony of Managing Partner of the Washington National Tax Services PriceWaterhouseCoopers Before the House Committee on Ways and Means (Mar. 10, 1999). 103 The largest tax expenditure for 1994 was the “quotient familial” system of allowances for family size, which totaled FF 68 billion ($11 billion). Other significant tax expenditures were the exemption of income from investment growth bonds (FF 18 billion), the reduced rate for long-term capital gains (FF 14.5 billion), and the exemption of imputed income from housing (FF 13.7 billion). No other expenditures exceeded FF 10 billion. TAX EXPENDITURES, supra note 100, at 61-63. 104 See OECD, REVENUE STATISTICS OF OECD M EMBER COUNTRIES, 1965-1995, 111 (1996). 105 See OECD REPORT , supra note 23, at 7. The other co-chair was Japan; see below for a discussion why it was interested in this issue. 99

24

Information is sparse also for host countries that grant tax holidays or other preferential regimes targeted at foreigners. The OECD Report on tax competition gives no figures, except that total direct investment by G7 countries in traditional tax havens in the Caribbean and South Pacific increased more than five-fold in 1985-1994 to more than $200 billion. This represented a rate of increase well in excess of the growth of total outbound Foreign Direct Investment (FDI).106 If one assumes that the major motivation of such investment in traditional tax havens is avoiding taxation,107 and therefore that most of the income from the $200 billion escapes the net of existing anti-deferral regimes, this number would represent a significant loss of corporate tax revenue for the G7 countries.108 Two member countries of the OECD have preferential tax regimes targeted at foreigners: Ireland and Belgium. Ireland provides a reduced tax rate of about 10% to foreign MNEs that establish manufacturing facilities in selected areas (i.e., a “production tax haven”). Belgium grants near-complete tax relief to “centres de coordination” (i.e., “headquarters tax havens”). The estimated revenue loss for Ireland from its production tax haven in 1991-92 was 748.9 million Irish pounds (about $1 billion).109 The estimated revenue loss for Belgium from its headquarters tax haven in 1992 was 46 billion Belgian Francs (about $1.5 billion).110 These numbers are relatively small, but they represent only two out of the sixty-eight countries that grant preferential tax regimes to foreign MNEs. Moreover, the numbers are far from negligible for Ireland or Belgium: In Ireland, the tax expenditure on the production tax haven was the largest expenditure in the budget for 1991-92.111 In Belgium, the tax expenditure for the headquarters tax haven was the fifth largest expenditure in the 1992 budget.112 A problem with using tax expenditure figures that their accuracy may be doubtful. In particular, the U.S. figures (which are based on relatively sophisticated methods compared to other countries) may be understated because they do not take any behavioral shifts from changing the law into account. Thus, the $2-3 billion per year estimated revenue loss from deferral is not an accurate estimate of the revenue that would be lost were Subpart F repealed

See id. at 17. This seems a reasonable assumption because of the lack of other resources in those countries that would attract investors there. See Hines & Rice, supra note 39. 108 If the $200 billion generate income at 10% a year and the average corporate tax rate in G7 countries is 35% (both reasonable assumptions), the $200 billion in traditional tax havens represent a yearly revenue loss of $7 billion. Interestingly, Grubert reports based on U.S. Treasury data that even Subpart F, which is generally regarded as the toughest anti-deferral regime in the world, captures only 50% of the income of CFCs in Caribbean tax havens. See Grubert, supra note 58, at 55. 109 TAX EXPENDITURES, supra note 100, at 74. 110 See id. at 34. It is not clear, however, whether this revenue could have been collected absent that tax expenditure. See Part IV for a cost-benefit analysis of this issue. 111 The expenditure for married person’s allowance (745.8 million pounds) was almost as large, and no other expenditure was over 500 million pounds. See id. at 73-74. 112 The largest one was the exemption of dividends, at BF 112 billion. See id. at 33-34. 106 107

25

or severely curtailed (as some have suggested).113 In that case, U.S.-based MNEs would presumably shift revenues abroad, resulting in increased revenue losses.114 Nor is the $2-3 billion figure necessarily even a good estimate of the revenue that could be gained if Subpart F were expanded to cover all income because in that case (as suggested in Part II(b)) new MNEs may be incorporated outside the U.S., resulting in revenue losses.115 In addition, the U.S. revenue estimate has not changed (except for adjustments for inflation and other macroeconomic variables) since the 1980’s even though the Treasury data indicate a significant reduction in average effective foreign tax rates.116 But on the basis of the available data, it appears that the total revenue to be gained from repealing deferral is relatively small (although significant in relation to total U.S. corporate tax revenues from foreign source income).117 On the other hand, it should be recalled that the U.S. has the strictest anti-deferral regime of all OECD member countries and that its corporate tax rate is relatively low. If one takes a country like Japan, which has an ineffective anti-deferral regime and a high effective corporate tax rate, the revenue loss from deferral is likely to be much larger. The Japanese effective corporate tax rate has consistently been estimated at 40% or higher.118 This may explain why Japan was at the forefront of the OECD effort to combat harmful tax competition See HUFBAUER, supra note 68; National Foreign Trade Council, The NFTC Foreign Income Project: International Tax Policy for the 21st Century, Part One: A Reconsideration of Subpart F (1999) (Doc. 199911623 (150 original pages)) [hereinafter NFTC Project]. 114 The one-year exception from Subpart F for “active financing income” of banks and insurance companies was estimated to cost $1.4 billion, or about 50% of the entire loss from deferral. See Treasury Shares Dorgan's Concerns About CFC Extenders Provision, 1999 TAX NOTES TODAY 32-25 (1999). The foreign sales corporation regime (recently challenged by the WTO) costs $2 billion, see WTO Panel Issues Final Report Condemning U.S. FSC Regime, 1999 TAX NOTES TODAY 183-3 (1999). 115 Another issue that affects the accuracy of the estimate is that the revenue estimate assumes a continuation of the Subpart F method of taxing MNEs, i.e., a deemed dividend to the parent rather than a consolidated approach. As a result, CFCs with losses do not reduce the revenue estimate. In addition, repealing deferral implies treating all the interest expense of the MNE as fungible, as opposed to the current rule that only U.S. interest is allocated based on worldwide assets while foreign interest reduces foreign source income (and thus reduces the availability of foreign tax credits). If loss offsets and interest fungibility are both allowed, the $2 billion revenue estimate shrinks considerably. See Frisch, supra note 66; Donald J. Rousslang, Comment on Hufbauer and Derosa, Costs and Benefits of the Export Source Rule, 97 TAX NOTES INT ’L 129-27 (1997). 116 See Part III(c) above. 117 See id. According to Altshuler and Newlon, based on 1986 Treasury data, the U.S. collected only $1.585 billion from foreign source income of $47 billion of 340 U.S.-based multinationals (a 3.4% tax rate). Thus, an additional $2 billion (to use the lower estimate) would more than double tax collections from this source. The total foreign source income now is significantly higher because of the increase in outbound FDI. 118 See Chennells & Griffith, supra note 56, at 150 (estimating the Japanese effective tax rate in 1994 (the most recent year) at between 40.3% (ATR) and 52.9% (EMTR)); Julie Collins & Douglas Shackleford, Corporate Domicile and Average Effective Tax Rates, 2 INT ’L TAX AND PUB. FIN. 55, 61 (1995) (estimating the Japanese average effective tax rate for 1982-1991 at between 57% and 59%). In both cases these are the highest rates among the countries surveyed (Australia, Canada, France, Germany, Ireland, Italy, Japan, Spain, the U.K., and the U.S.). 113

26

and co-chaired the Special Sessions of the Committee on Fiscal Affairs that drafted the OECD Report on this subject.119 Thus, the limited available data on revenue losses from tax competition in the corporate income tax area suggest that these losses, while significant, are probably much smaller than the potential losses in the personal income tax area. On the other hand, the revenue losses may vary dramatically from country to country depending on their effective corporate tax rate and the stringency of their anti-deferral rules. Some countries, such as France and Japan, may suffer significant revenue losses from corporate tax competition, even in terms of their overall budget. In other countries, like the U.S., those losses may be relatively insignificant.

119

See OECD REPORT , supra note 23, at 7.

27

III.

The Problem of Tax Competition from a Global Normative Perspective

In Part II, I have argued that increased capital mobility, when combined with tax competition, leads to the likelihood that cross-border flows of capital will be taxed more lightly than either domestically invested capital or domestic labor. In this Part, I will attempt to evaluate whether this situation is problematic from a global efficiency or equity perspective. I will then attempt to balance those arguments against the right of democratic countries to determine the size of their governments and to develop a distinction between harmful and acceptable forms of tax competition. a. Global Efficiency and Global Welfare. (i) Capital Export Neutrality and Capital Import Neutrality. The traditional argument in favor of imposing the same tax rate on income from capital whether it is invested at home or abroad has been made in the name of capital export neutrality (CEN), a concept first developed by Peggy Musgrave.120 CEN refers to the choice that an investor resident in a home country has between investing her savings domestically or in a foreign host country. CEN obtains when home and host country investments that earn the same pretax rates of return also yield the investor the same return after taxes.121 CEN is violated, for example, if both the home and host countries fail to tax the income from an investment in the host country (while an investment in the home country is taxed). In that case investors would prefer to invest in the host country rather than in the home country even if the pre-tax yield on the domestic investment is higher.122 The result is a deadweight loss from a global efficiency perspective because investments will not be allocated to their most productive (highest yielding) pretax uses. In the long run, as more capital flows to the host country investments, the (pre-tax) returns on those investments will be reduced, and pre-tax returns on home country investments will rise, until equilibrium is restored (after tax returns will be equalized). The deadweight loss will, however, remain the same because some less productive host country investments will be made at the expense of more productive home country investments (capital will be oversupplied in the host country and under-supplied in the home country).123 See PEGGY B. M USGRAVE , UNITED STATES TAXATION OF FOREIGN INVESTMENT INCOME: ISSUES AND A RGUMENTS (1969); JCT Reports on International Taxation, 1999 TAX NOTES TODAY 124-8 (JUNE 29, 1999). 121 See RICHARD E. CAVES, M ULTINATIONAL ENTERPRISE AND ECONOMIC A NALYSIS 227 (1982). 122 Assume for example that an investor faces a choice between a home country investment yielding 100 and a host country investment yielding 70. In a tax-free world, the investor would choose the home country investment. Now assume that the home country investment is taxed at 40% while the host country investment is untaxed (e.g., because the host country does not tax investment income earned by nonresidents and the home country is unable to enforce its tax jurisdiction on foreign-source income of its residents). In that case the investor faces a choice between a home country investment yielding 60 (100- 40 tax) and a host country investment yielding 70 (70- 0% tax). The investor would then choose the host country investment even though it yields the lower pretax return. 123 See Assaf Razin & Efraim Sadka, Optimal Incentives to Domestic Investment in the Presence of Capital Flight 91989) (unpublished IMF Working Paper) [hereinafter Optimal Incentives]. 120

28

This traditional argument has led most economists to support residence-based taxation of world-wide income as the optimal tax rule for international taxation because in the absence of host country taxes, residence (home country) based taxation maintains CEN. For example, Razin and Sadka have argued that adopting the residence principle leads to an efficient allocation of the world’s pool of investment capital: If a country adopts the residence principle, taxing at the same rate capital income from all sources, then the gross return accruing to an individual in that country must be the same, regardless of which country is the source of that return. Thus, the marginal product of capital in that country will be equal to the world return to capital. If all countries adopt the residence principle, then capital income taxation does not disturb the equality of the marginal product of capital across countries which is generated by a free movement of capital. … [If] residents of a country are not taxed on their income from foreign sources … [and] the tax rate is not the same in all countries, then the marginal product of capital is also not the same in all countries. In this case the international allocation of the world stock of capital is inefficient.124 The economic case against CEN is traditionally made in the name of capital import neutrality (CIN). CIN requires that equal before-tax returns at the margin to competing (domestic and foreign) suppliers of capital to a host country producer translate into equal aftertax earnings.125 CIN is violated, for example, if foreign investors into a host country are taxed on their investment income at their home country rate (as required by CEN) while the host country does not levy an income tax on investment income. In that case, domestic (host country) investors will have a different net return on their investment in the host country than foreign (home country) investors. The result is that intertemporal marginal rates of substitution (that is, the choice between present and future consumption) will not be the same between countries, and the international allocation of world savings will be distorted.126 Assaf Razin & Efraim Sadka, International Tax Competition and Gains from Tax Harmonization, 37 ECONOMICS LETTERS 69, 69-70 (1991) [hereinafter International Tax Competition]. If governments cannot effectively tax foreign source income, Razin and Sadka advocate capital controls, or if those are not feasible, subsidies to domestic investment. See Optimal Incentives, supra note 123. See also Assaf Razin & Efraim Sadka, Capital Market Integration: Issues of International Taxation, (1989) (The David Horowitz Institute for the Research of Developing Countries, Article No. 4/90) (“[O]ptimal policy requires an efficient allocation of capital between investment at home and abroad so that the marginal product of domestic capital must be equated to the world rate of interest.”) Razin and Sadka derive this result from an application of Diamond and Mirrlees’ concept of “production efficiency.” See Peter A. Diamond & James A. Mirrlees, Optimal Taxation and Public Production I: Production Efficiency, 61 A MER. ECON. REV. 8 (1971). 125 See CAVES, supra note 121, at 227. 126 See International Tax Competition, supra note 124, at 70. In terms of the example in note 127 above, the home country investor prefers the domestic investment (yielding 100 before tax) over the host country investment (yielding 70) because both will subject him to a tax at 40% (so that the after tax yield of the domestic investment is 60 and of the foreign investment is 42). In that case he will have a lower after tax yield (60) than a host country investor in the host country who faces a 0% tax rate (70). After-tax savings will thus be lower in the home country. 124

29

In a classic article, Thomas Horst showed that if tax rates vary among countries, it is impossible for CEN and CIN to obtain simultaneously.127 How, then, is one to choose between a policy that assures efficient allocation of world investment (CEN) or world savings (CIN)? Horst showed that the choice depends on the assumption made regarding the relative elasticity (i.e., responsiveness to taxation) of the supply and demand of capital. If the supply of capital is fixed, i.e., savings rates are not responsive to taxation, and the demand for capital is elastic, CEN is preferred because it assures an efficient allocation of the home country’s supply of capital. If the demand for capital is fixed while the supply is elastic, CIN is preferred because it assures the efficient provision of the host country’s demand for capital.128 Horst argued that the most plausible assumption in the absence of empirical evidence is that the elasticity of supply and demand for capital is the same and is somewhere between 0 (fixed) and 1 (perfect elasticity). Thus, a policy that falls between CEN and CIN is called for.129 Why, then, do most economists prefer CEN to CIN?130 The answer is that the empirical evidence suggests that the elasticity of demand for capital is considerably greater than the elasticity of supply, i.e., that saving vs. consumption decisions are less responsive to taxes than the choice of alternative investment vehicles. Thus, for example, a recent article by Austan Goolsbee on the efficiency of the investment tax credit has found that the demand elasticity of investment exceeds 1.131 Empirical studies examining aggregate data on savings flows tend to suggest that the elasticity of savings with respect to the rate of return is small while the elasticity

See Thomas Horst, A Note on the Optimal Taxation of International Investment Income, 94 Q. J. OF ECON. 793 (1980). 128 See id. at 796. 129 See id. at 797. Horst acknowledges that the U.S. policy meets his criteria but argues that the compromise between CEN and CIN embodied in Subpart F would only coincidentally be correct according to the criteria he suggests. See id. at n. 5. 130 In addition to Razin and Sadka, a preference for CEN over CIN on the assumption that foreign-source income can be effectively taxed can be seen, e.g., in Gordon, Can Capital Income Taxes Survive in Open Economies, supra note 32; Roger H. Gordon & Hal Varian, Taxation of Asset Income in the Presence of a World Securities Market, 26 J. INT ’L ECON. 205 (1989); Michael P. Devereux & Mark Pearson, European Tax Harmonization and Production Efficiency, 39 EUROPEAN ECON. REV. 1657, 1660 (1995) (“[A]n optimal tax structure would preserve production efficiency but not exchange efficiency [i.e., efficient allocation of savings]... CEN is a necessary condition for production efficiency”); James Levinsohn & Joel Slemrod, Taxes, Tariffs, and the Global Corporation, 51 J. PUBLIC ECON. 97 (1993) (arguing that “the optimality of levying taxes that do not discriminate by location is … preserved when strategic considerations are introduced,” although they favor reducing taxes on foreign source income if tax policy cannot discriminate between strategic and non-strategic sectors). 131 See Austan Goolsbee, Investment Tax Incentives, Prices, and the Supply of Capital Goods, 113 Q. J. OF ECON. 121 (1997). See also Frisch, supra note 66 (pointing out that assuming zero elasticity for the demand for capital is unrealistic because it requires not only that capital be used inelastically in production of goods, but also that consumers’ demands for goods with differing capital intensities are all inelastic). There are also indications that demand elasticity for capital may be increasing over time. See U.S. Investment Abroad, supra note 54, which reports an increase in the elasticity of investment to tax rate differentials from about 2 in 1984 to about 3 in 1992. 127

30

of aggregate invesment is probably higher.132 An OECD study in 1994 concluded that “there is no clear evidence that the level of taxation … does generally affect the level of household saving.”133 The basic reason seems to be that many individuals tend to have a fixed target for the amount of saving they need at retirement (e.g., $1 million). If so, a cut in the tax rate on savings would lead them to reduce, rather than increase, their saving rate (because a lower before-tax saving rate would enable them to reach their target).134 In addition to arguments based on CIN, other arguments for rejecting CEN as the optimal policy have been made based on “national neutrality.” The issue underlying this debate is the proper treatment of host country taxes. CEN implies that they should be credited by the home country, and moreover that the credit should be unlimited (i.e., taxes at rates that exceed the home country rate should be refundable), so as to leave the investor with equal after-tax rates of return on home and host country investments.135 However, some economists have argued that host country taxes should be treated as costs and a deduction rather than a credit given so as to maximize the national income of the home country (which includes both after-tax profits of the home firm and home country tax revenues).136 However, others have pointed out that such “national neutrality” policies are likely to lead to retaliation which ultimately would hurt the home country as well as violate CEN.137 Other economists have argued that a credit would result in the host country raising its tax rate to capture tax revenue, and the home country would then respond by raising its rate to gain some benefits from restricting capital flows, resulting in too little trade.138 But this line of analysis ignores the evidence that in fact the adoption of the unilateral foreign tax credit by the U.S. in 1918 has led to a cooperative outcome in which double taxation is prevented and world welfare maximized.139 In a cooperative game the credit turns out to be superior to the See Robin Boadway & David Wildasin, Taxation and Savings: A Survey, 15 FISCAL STUDIES 19 (1994); Investment, Saving and Taxation, supra note 29, at 101. 133 OECD, TAXATION AND HOUSEHOLD SAVING (1994). The same conclusion is reached in the voluminous literature on the effect of cutting the capital gains tax on saving rates. See Bruce Anderson, Strategic Choice Taxation: A Solution to the Federal Revenue Crisis, 1995 COLUM. BUS. L. REV. 281, 314 (1995); John W. Le, Critique of Current Congressional Capital Gains Contentions, 15 VA. TAX REV. 1 (1995); George R. Zodrow, Economic Analyses of Capital Gains Taxation: Realizations, Revenues, Efficiency and Equity, 48 TAX L. REV. 419, 480 (1993); and Calvin H. Johnson, The Consumption of Capital Gains, 55 TAX NOTES 957, 970 (May 18, 1992). 134 See Michael J. Graetz, Revisiting the Income Tax vs. Consumption Tax Debate, 92 TAX NOTES TODAY 247-96 (1992); Marvin A. Chirelstein, Taxes and Public Understanding, 29 CONN. L. REV. 9 (1996). 135 In fact, no country grants an unlimited foreign tax credit (the U.S. credit has been limited since 1921) because it is an invitation to host countries to raise tax rates at the expense of the home country treasury. 136 See Martin Feldstein & David Hartman, The Optimal Taxation of Foreign Source Investment Income, 103 Q. J. OF ECON. 613 (1979). 137 See Frisch, supra note 66. See also NFTC Project, supra note 113. 138 See Eric W. Bond & Larry Samuelson, Strategic Behavior and the Rules for International Taxation of Capital, 99 ECON. J. 1099 (1989). 139 See Gordon, Can Capital Income Taxes Survive in Open Economies, supra note 32 (arguing that sourcebased taxation of capital is sub-optimal when capital is mobile). The burden of the tax must be born by immobile factors so it is more efficient to tax them directly and not deter inbound investment. However, the 132

31

exemption or deduction method (required by the CIN and national neutrality arguments respectively) because it attains the efficient allocation of capital without requiring rate harmonization or side payments.140 An influential article by Daniel Frisch has argued that the CEN/CIN debate is obsolete because it ignores changes in the world economy that took place since the 1960s. His main argument is that portfolio investments are more important determinants of efficient capital allocation than FDI, and therefore CEN should be implemented for the former but not for the latter.141 However, the fact that portfolio investment now exceeds FDI does not mean that the latter is unimportant in allocating investment, and applying CEN to portfolio investment should not rule out applying it to FDI as well. The location studies summarized above indicate that taxation plays an important role in the decisions of multinationals where to locate their investment capital.142 Other economists have suggested that the Musgrave model for CEN omits or oversimplifies other important considerations. First, it has been suggested that foreign business operations may be complementary with domestic activities that are particularly valuable due to intrinsic externalities, such as R&D or domestic job creation.143 But even if that were the case (and the evidence is not compelling), it would be better to subsidize such activities directly since not all MNEs engage in them to the same extent.144 Second, Devereux and Hubbard have argued that in the case of oligopolies firms may make greater profits if they can commit to expanding their output, and that this justifies treating such firms (e.g., Boeing) more favorably from a tax perspective.145 But this argument assumes survival of such source-based taxation can be explained by the U.S. enacting the credit as a “Stackelberg leader”, i.e., taking into account the reaction of other countries. 140 See Eckehard Janeba, Corporate Income Tax Competition, Double Taxation Treaties, and Foreign Direct Investment, 56 J. PUBLIC ECON. 311 (1995). 141 See Frisch, supra note 66. 142 An interesting issue that arises out of applying CEN to portfolio investment is that if the corporate and shareholder taxes are unintegrated, CEN for portfolio shareholders requires equalizing the corporate tax rate for the corporations they invest in which would require both CEN and CIN apply at the corporate level. See Devereux & Pearson, supra note 130. But if portfolio shareholders were given a credit for corporate taxes paid (similar to the “indirect” credit given to corporate shareholders under U.S. law) only CEN would be required at the corporate level to maintain shareholder level CEN. 143 See James R. Hines, Jr., No Place Like Home: Tax Incentives and the Location of R&D by American Multinationals (Faculty Research working article series, 1998); Joel P. Trachtman, International Regulatory Competition, Externalization, and Jurisdiction, 34 HARV. INT 'L L.J. 47 (1993). 144 And indeed R&D is subsidized through the tax system. See James R. Hines, Jr., On the Sensitivity of R&D to Delicate Tax Changes, in STUDIES IN INTERNATIONAL TAXATION 149, supra note 37. But see, Jonathan L. Mezrich, International Tax Issues of the U.S. Pharmaceutical Industry, 10 A KRON TAX J. 127 (1993) (arguing if a foreign country allows a deduction for expenses which the U.S. Treasury Department allocates to foreign source income, the U.S. corporation will pay "double" taxes on some portion of its income; these overall tax costs may arguably be a disincentive to locating R & D-intense operations in the U.S.). 145 See Michael P. Devereux & R. Glenn Hubbard, Taxing Multinationals (Apr. 1999) (Columbia University working paper).

32

(i) that there are important international oligopolies, (ii) that governments can precommit to their tax policies so as to encourage expanding output, and (iii) that treating foreign income more favorably for tax purposes encourages total output. None of these assumptions are free from doubt.146 Third, the Musgrave model for CEN assumes a first-best world, i.e., that there are no other economic distortions (e.g., via taxation), and therefore the value of an additional unit of capital equals its marginal product. But in fact, other taxes do exist, such as the personal income tax on dividends, which may justify lighter taxation of corporate income from abroad.147 But there is no particular reason to limit the relief to foreign source income; and if broader relief is justified, the appropriate solution is some form of integration rather than necessarily relief at the corporate level. 148 The most influential argument both historically and currently against adopting CEN for U.S.- based multinationals is not based on purely economic considerations but rather on political ones: It is argued that CIN is necessary to preserve the competitiveness of U.S. multinationals vis-à-vis multinationals headquartered in other countries.149 For example, assume that a U.S. multinational competes with a German multinational in a host country that does not tax either foreign investor. If the U.S. taxes the U.S. multinational currently on its profits from the host country while Germany exempts the profits of its multinational, this arguably results in a higher cost of capital for the U.S. multinational and places it at a competitive disadvantage.150 Whatever the overall merit of this argument,151 it implicitly assumes that only unilateral, non-cooperative action is possible in the field of international taxation. However, as the See James R. Hines, Jr., The Case Against Deferral: A Deferential Reconsideration, 52 NAT ’L TAX J. 385-404 (Sept. 1999). 147 See id.; Robert A. Green, The Future of Source-Based Taxation of the Income of Multinational Enterprises, 79 CORNELL L. REV. 18, 34 (1993); Bruce N. Davis & Steven R. Lainoff, U.S. Taxation of Foreign Joint Ventures, 46 TAX L. REV. 165 (1991). 148 Integration can be achieved either at the corporate or shareholder level, but in practice it is almost always done at the shareholder level (through dividend exemption or an imputation credit) to avoid automatic extension of integration to foreign shareholders. See Michael J. Graetz & Alvin C. Warren Jr., Introduction to Integrating Corporate and Individual Income Taxes, 1999 TAX NOTES TODAY 186-89 (1999); Stephen G. Utz, Tax Harmonization and Coordination in Europe and America, 9 CONN. J. INT 'L L. 767(1994); Integration of the Individual and Corporate Income Taxes: Reporter's Study of Corporate Tax Integration (American Law Institute Federal Income Tax Project, Alvin C. Warren, Jr., Reporter, 1993). 149 See Frisch, supra note 66; NFTC Project, supra note 113. 150 However, if the German multinationals sets its prices to maximize profits, it will charge the same as the U.S. multinational so that no competitive disadvantage need result. See Richard C. Henderson, Taxes, Market Structure, and International Price Discrimination, 10 J. INT ’L. L. BUS. 244 (1989); Mark P. Gergen & Paula Schmitz, The Influence of Tax Law on Securities Innovation in the United States: 1981-1997, 52 TAX L. REV. 119 (1997). 151 For a critique, see Frisch, supra note 66. Fundamentally, the problem is that the competitiveness argument assumes that the welfare of U.S. based multinationals is identical with U.S. national welfare. This, however, is a problematic assumption when all multinationals have both U.S. and foreign shareholders and employees. However, as argued in Part II above, ending deferral unilaterally may be counter-productive if it 146

33

cooperative ventures discussed in Part V make clear, it is quite conceivable that all OECD members (which currently are the home countries for about 85% of all multinationals) can be persuaded to adopt CEN-based policies, in which case the competitiveness issue is addressed in the short run.152 Thus, from the perspective of this article, the competitiveness issue is irrelevant because the solutions it discusses are all multilateral and are based on cooperative, concerted actions by the U.S. and its major trading partners. To sum up, CEN appears still to be the best guide for taxing cross-border investments, both in the case of portfolio investment and of FDI. Neither the economic nor the political arguments against it are persuasive. Thus, to the extent that tax competition leads to undertaxation of cross-border income compared to domestic income, as described in Part II, CEN leads to the conclusion that steps should be taken to curb such tax competition. However, as discussed in section (c) below, not all tax competition is equally harmful, and therefore CEN should be balanced against other considerations that favor allowing some forms of tax competition. (ii) Welfare Economics and Public Choice In this section, I shall focus on the economic arguments that have been made in favor of and against tax competition from a welfare economics perspective. The economic analysis of tax competition typically begins with Charles Tiebout’s classic 1956 article “A Pure Theory of Local Expenditures.”153 Tiebout addresses the problem of the provision of public goods in a local government context and shows that in that context the issue of knowing true voter preferences for public goods and avoiding free riders can be resolved. The key to Tiebout’s solution is the assumption that individuals can move and face a choice among localities offering fixed but different levels of public services financed through benefits taxes. In that context, Tiebout shows that the optimal outcome in terms of satisfying individuals’ preferences can be achieved: Individuals will sort themselves out geographically based on their differing preference levels, paying different levels of benefits taxes, and the final outcome will not be inferior to market-based outcomes for private goods. The implication of Tiebout’s analysis is that tax differentials among jurisdictions are positive because they maximize the ability of individuals to satisfy their preferences, and therefore actions to harmonize tax rates are counter-indicated. Tiebout’s analysis depends on two crucial assumptions: First, that individuals are completely mobile between taxing jurisdictions; and second, that all taxes are benefits taxes and represent payments for goods and services provided by government rather than taxes based on leads new multinationals to incorporate their parents outside the U.S. and existing U.S. multinationals to become subsidiaries of foreign multinationals. 152 In the long run, as discussed in Part V, such curbs on tax competition only for OECD members may lead to the rise on non-OECD based multinationals. From a world welfare perspective, CEN clearly should apply to all multinationals, not just U.S. multinationals. Because of the difficulties in applying CEN on a residence basis, Part V develops a source-based approach to CEN. 153 64 J. POL. ECON. 416 (1956).

34

ability to pay which have distributive goals.154 When either assumption is relaxed, Tiebout’s analysis becomes less valid.155 At the international level, both of Tiebout’s assumptions are problematic: As argued above, restrictions on immigration make movement of individuals between countries difficult and the main form of taxation is income taxes with explicit distributive goals. However, I will argue below that Tiebout’s analysis retains some validity even at the international level for generally applicable corporate tax rate differentials among countries. Following Tiebout, an extensive public finance literature has focused on tax competition at the local level. The focus in this literature is on the type of tax available to the local jurisdiction. As we have seen, Tiebout’s model assumes benefits taxes. What happens when distortionary taxes, such as taxes on capital, are used instead of benefits taxes or lump-sum taxes? Zodrow and Mieszkowski have shown that in that case the supply of public services will be less than optimal. They analyzed the supply of such services in a situation in which capital is mobile and the options available to local government are either a non-distorting lump sum tax or a tax on capital (e.g., a property tax). If so, because each community is concerned that raising the tax on capital will drive out capital, they would choose to levy only the head tax. But if the head tax has to be reduced (e.g., for political reasons), the result is a reduction in the level of public services below the optimal level.156 Zodrow and Mieszkowski’s analysis has generated several studies that focus on the negative effects of tax competition when taxes have to be levied for some reason on a mobile factor such as capital. For example, Wildasin has reconceptualized the analysis in terms of a fiscal externality: When a jurisdiction raises its taxes on capital, it causes a flow of capital to other jurisdictions that increases their tax revenues. The result is an under-supply of government services in the first jurisdiction driven by the wish to avoid the externality.157 154

Another assumption, discussed below, is that the provision of public goods creates no positive or negative externalities on other jurisdictions. 155 For a direct criticism of the Tiebout model in the international context, see Peggy B. Musgrave & Richard A. Musgrave, Fiscal Coordination and Competition in an International Setting, in INFLUENCE OF TAX DIFFERENTIALS ON INTERNATIONAL COMPETITIVENESS 59, 81-84 (1989) [hereinafter Fiscal Coordination]. Even Charles McLure, who, as discussed below, generally favors tax competition at the local level, admits that it may be harmful at the international level when distributive considerations are raised and mobility is limited. See International Considerations, supra note 22, at 101-102. 156 See George R. Zodrow & Peter Mieszkowski, Pigou, Tiebout, Property Taxation, and the Underprovision of Local Public Goods, 19 J. URBAN ECON. 356 (1986). The same result obtains when services are provided to business as long as perceived capital responsiveness to changes in the tax on capital does not fall too drastically as the level of public services increases. The Zodrow and Mieszkoeski model implicitly assumes that only capital is mobile between jurisdictions which is a realistic assumption at the international level but not at the national level. 157 See David E. Wildasin, Interjurisdictional Capital Mobility: Fiscal Externality and a Corrective Subsidy, 25 J. URBAN ECON. 193 (1989). The inefficiency can be corrected by a subsidy from a higher level of government, but Wildasin shows that very high subsidy rates (in the order of 40%) are called for. Obviously, at the international level there is, at present, no higher level of government to provide the subsidy. Wildasin also estimates the allocative inefficiency caused by inadequate local government spending; in his model it could amount to 20% of spending at the margin. This estimate depends on the elasticity of demands for public goods. It also depends on the type of public goods involved, see Michael

35

Bucovetsky and Wilson have extended this analysis to situations where a tax on wages, as well as a tax on capital, is possible.158 Since labor is relatively immobile, one could argue that the underprovision of public services resulting from the fear of driving capital elsewhere would not apply to taxes on wages (i.e., they would be more like Pigouvian head taxes or Tieboutian benefits taxes). However, they show that even if wage taxation as well as source-based taxation on capital is possible, the supply of public services continues to be inefficiently low. The reason is that a tax on labor tends to reduce labor supply in the country that taxes labor. Capital would then flow out of the taxing country, and as a result wages and labor supply would rise in other countries, creating the same kind of positive externality as a tax on capital. On the other hand, Bucovetsky and Wilson also show that if world-wide taxation of capital were possible, this would result in the efficient provision of public goods. If tax competition is harmful in these ways why do countries not attempt to coordinate tax policy? A series of studies by Wilson, Bucovetsky, and Kanbur and Keen has addressed this issue.159 They show that in a tax competition situation, the residents of smaller regions tend to benefit (because the small country would perceive capital elasticity to be higher and therefore undercut the larger country’s tax rate). Thus, even if all residents of both large and small regions taken together would benefit from tax coordination, in that overall tax revenue would rise to its optimal level, the residents of the smaller region would object because they gain more from tax competition. This line of analysis has clear implications in the international setting because smaller countries are more likely to be tax havens. It also favors restrictions on tax competition but points out some of the difficulties in achieving such restrictions.160 Thus, most of the public finance literature has tended to favor restrictions on tax competition from a utilitarian, welfare-maximization perspective. However, a different line of analysis by public finance economists yields opposite results. A fundamental assumption of the studies cited above is that governments are benevolent, i.e., that they seek to maximize the utility Keen & Maurice Marchand, Fiscal Competition and the Pattern of Public Spending, 66 J. PUBLIC ECON. 33 (1997) (showing that tax competition leads to the underprovision of public goods benefiting immobile consumers (e.g., recreational facilities or social services, or redistributional payments) and an overprovision of public goods benefiting mobile capital (e.g., infrastructure)). If governments can compete by providing public goods to industry, it can be shown that curbing tax competition requires curbing such subsidies as well. See Clemens Fuest, Interjurisdictional Competition and Public Expenditure: Is Tax Coordination Counterproductive?, 52 FINANZ A RCHIV 478 (1995). The OECD is in fact working on limiting such subsidies. See Marian Murphy & Udo Pretschker, Public Support to Industry, OECD OBSERVER 11 (Feb./Mar. 1997). 158 See Sam Bucovetsky & John D. Wilson, Tax Competition with Two Tax Instruments, 21 REG’L SCIENCE & URBAN ECON. 333 (1991). 159 See John D. Wilson, Tax Competition with Interregional Differences in Factor Endowments, 21 REG’L SCIENCE & URBAN ECON. 423 (1991); Sam Bucovetsky, Asymmetric Tax Competition, 30 J. OF URBAN ECON. 167 (1991); Ravi Kanbur & Michael Keen, Jeux sans Frontieres: Tax Competition and Tax Coordination when Countries Differ in Size, 83 A MER. ECON. REV. 877 (1993); John D. Wilson, Theories of Tax Competition, 52 Nat’l Tax J. 269 (1999). 160 The problem of tax havens is discussed in Part V.

36

of their residents. However, as the public choice literature shows, governments may also be conceived as Leviathans, i.e., maximizing the size of government revenues in the interest of government bureaucrats regardless of the wishes of the voters in their countries. From this perspective, tax competition may be beneficial because it acts as a constraint on the tendency of governments to grow.161 Charles McLure made a forceful statement of the public choice perspective on tax competition in an influential article written while he was Deputy Assistant Secretary for Tax Policy in the Reagan administration.162 McLure first summarized the anti-tax competition public finance literature and then criticized it for assuming that taxes have to be levied on mobile capital. If benefits taxes can be used and individuals are mobile, the Tiebout analysis indicates that the problem of underprovision of public services will not arise. McLure then argued that at the state and local level benefits taxes are both feasible and desirable. However, the core of McLure’s argument was that because of the Leviathan problem, “tax competition is, on balance, good:” It forces governments to be more attentive to the desires of voters. The first prong of McLure’s argument is problematic at the international level because of limited labor mobility and the relevance of distributive concerns.163 However, the Leviathan prong of the argument has proven to be quite influential.164 Two questions are typically addressed in this literature: First, can the Leviathan model or the benevolent model be empirically shown to be accurate? Second, if government is neither completely self-seeking nor completely benevolent, what are the implications for tax competition? On the first question, Oates has studied the empirical question whether the extent of fiscal decentralization, which enables tax competition to take place, is negatively correlated with the size of the public sector as the Leviathan model suggests. He studied the extent of fiscal decentralization in a sample of forty-three countries (based on IMF data) and in the forty-eight contiguous U.S. states. In both cases, there was no relationship between the degree of decentralization and the size of the public sector; if anything, greater decentralization tended to be correlated with larger government, contrary to the Leviathan hypothesis.165 Rodrik advances a hypothesis that may explain this last result: He shows that greater openness (and more tax competition) tends to be correlated with more developed social insurance programs as voters

The classic statement of this view comes from Brennan and Buchanan. See Geoffrey Brennan & James Buchanan, Towards a Tax Constitution for Leviathan, 8 J. PUB. ECON. 255 (1977); Geoffrey Brennan & James Buchanan, Tax Instruments as Constraints on the Disposition of Public Revenues, 9 J. PUB. ECON. 301 (1978); GEOFFREY BRENNAN & JAMES BUCHANAN, THE POWER TO TAX: A NALYTICAL FOUNDATIONS OF A FISCAL CONSTITUTION (1980). 162 Charles E. McLure, Tax Competition: Is What’s Good for the Private Goose also Good for the Public Gander?, 39 NAT ’L TAX J. 341 (1986). 163 McLure himself agrees. See International Considerations, supra note 22, at 101-102. 164 See, e.g., Fiscal Coordination, supra note 155 (responding to McLure); Fuest, supra note 157. 165 See Wallace E. Oates, Searching for Leviathan: An Empirical Study, 75 A MER. ECON. REV. 748 (1985). 161

37

demand government insurance from the external risks associated with an open economy.166 Thus, the empirical data are at best inconclusive and do not support the Leviathan model. Edwards and Keen have directly compared the Leviathan and benevolent views of government from the perspective of tax competition. They point out that both of those views are extreme cases of a more general (and more plausible) model in which policy makers attach some value both to the welfare of their citizens and to the surplus they are able to extract from them for their own uses. Using the model developed by Zodrow and Mieszkowski (who assumed a benevolent government), they asked whether tax competition is beneficial if policy makers act to maximize the welfare of citizens, to maximize tax revenues, or anything in between. They show that if coordination leads to a small increase in the tax on mobile capital, this unambiguously increases the welfare of the policy maker. It may, however, also increase the welfare of a representative citizen if and only if the elasticity of the tax base exceeds the policy maker’s marginal propensity to waste tax revenue. For example, if it is estimated that the elasticity of capital to tax increases is 0.4, then coordination is beneficial as long as government wastes less than 40% of the capital taxes it collects on benefiting bureaucrats rather than citizens. In addition, they also show that the benefits of coordination are larger the higher one estimates the deadweight loss from taxation because coordination makes the tax closer to an unavoidable lump sum tax. Thus, if the deadweight loss from the tax is 30 cents on the dollar, they calculate that coordination will be desirable as long as an increase of revenues by $1 does not increase socially unproductive public expenditure by more than 23 cents.167 The problem is that it is very hard to know how much tax revenue is wasted by government, and the definition of what constitutes waste is in the eye of the beholder. However, the analysis of Edwards and Keen indicates that even relatively low estimates of the elasticity of capital to taxation and of the deadweight loss from taxation require a very high estimate of government waste for the Leviathan model to be persuasive. To sum up, from an efficiency perspective, a persuasive argument can be made that undertaxation of cross-border capital flows is problematic because it leads to deadweight losses. Such losses decrease global welfare and are unlikely (except under extreme assumptions regarding the extent of governmental waste) to be adequately compensated by benefits flowing from increased tax competition. b. Equity: The Tax Mix Equity in international taxation has two aspects: inter-individual equity and inter-nation equity. Inter-individual equity considerations determine the proper rate of taxation on cross-

See DANI RODRIK, HAS GLOBALIZATION GONE TOO FAR? 59-62 (1997) [hereinafter GLOBALIZATION]. This issue is discussed more extensively in Part IV. 167 See Jeremy Edwards & Michael Keen, Tax Competition and Leviathan, 40 EUR. ECON. REV. 113 (1996). 166

38

border transactions.168 Inter-nation equity considerations determine the proper division of the tax base among countries.169 In this section I will concentrate on inter-individual equity because the focus is on the tax rate while inter-nation equity issues will be discussed in Part IV. The traditional inter-individual equity analysis of international taxation is as follows: Suppose individual A earns 100 of income from domestic sources while individual B earns 100 of income from domestic sources and 100 of income from foreign sources. Vertical equity requires taxing B more than A because of his greater income; but if foreign source income is excluded, B and A will be taxed the same. Thus, under-taxation of foreign source income is inconsistent with inter-individual equity. The problem with this argument, when couched in such general, impersonal terms, is that it tells us nothing about the parties’ ability to respond to taxation by shifting the source of their income. Suppose, for example, that A can easily shift her income to a foreign source (e.g., it is interest income) while B’s domestic source income is labor income which cannot so easily be moved. In that case, taxing B the same as A is not problematic from an equity perspective because A can easily ensure that she pays less tax than B by shifting the source of her income (or else other people will adjust their investments so that in equilibrium the after tax returns to A and B will be equal). In that case, as Boris Bittker has shown, the equity violation turns into a violation of efficiency.170 To make the equity argument, a broader framework is needed, which takes into account the relative mobility of capital and labor.171 In general, as argued in the introduction, the 168

Inter-individual equity is usually analyzed in terms of horizontal and vertical equity, although some have argued that the former is merely a subset of the latter. See Louis Kaplow, Horizontal Equity: Measures in Search of a Principle, 42 NAT 'L TAX J. 139 (1989); Richard A. Musgrave, Horizontal Equity, Once More, 43 NAT 'L TAX J. 113 (1990); Paul R. McDaniel & James R. Repetti, Horizontal and Vertical Equity: The Musgrave/Kaplow Exchange, 1 FLA. TAX. REV. 607 (1993). This distinction is not crucial for the following discussion which focuses on the broader distributive consequences of undertaxing capital, but it can be conceptualized as a vertical equity issue. 169 On inter-nation equity, see e.g., Richard A. Musgrave & Peggy B. Musgrave, Inter-nation Equity, in M ODERN FISCAL ISSUES 63 (Richard M. Bird & John G. Head eds., 1972) [hereinafter Inter-nation Equity]; Peggy B. Musgrave, International Tax Base Division and the Multinational Corporation, 27 PUB. FIN. 394 (1972) [hereinafter International Tax Base Division]; Nancy H. Kaufman, Fairness and the Taxation of International Income, 29 LAW & POL’Y INT ’L BUS. 145 (1998). 170 See Boris I. Bittker, Equity, Efficiency and Income Tax Theory: Do Misallocations Drive Out Inequities? 16 SAN DIEGO L. REV. 735 (1979). 171 This distinction is theoretically somewhat artificial because all capital is earned at some point primarily from labor. Thus, economists would emphasize that the argument (e.g., in regard to the consumption v. income tax) is really about whether differential tax rates should be applied to future versus current consumption. However, given the lopsided distribution of capital and thus of income from capital in most societies, I consider it meaningful to distinguish between the majority of citizens who earn only labor income (many of whom cannot afford to save any of their income, and therefore will not have significant savings or income from savings over their entire life cycle) and the small (in most countries) minority who earn significant income from capital. This distinction is particularly significant in the cross-border context because most capital investment overseas, and certainly most capital invested not through widely held mutual funds (where evasion is not an issue), is held by the richest segment of society. For example, in the

39

present age of globalization can be distinguished from the preceding one (from 1870 to 1914) by the much higher mobility of capital than labor. This is the result of three historical changes that took place between 1914 and 1980: First, during the 1920s extensive restrictions were placed on immigration by all developed countries, so that large movements of labor in response to taxation are very difficult.172 Second, during the 1980’s capital controls were relaxed in all developed and many developing countries so that they do not pose a significant barrier to capital mobility. Finally, technological changes from the development of electronic banking to the rise of the Internet have made it easy to transfer funds worldwide at the speed of light. As noted above, it is estimated that international capital movements now approximate $1 trillion a day, which exceeds the annual GDP of most countries. Given these changes and the difficulty of taxing foreign-source income from capital (as explained in Part II above), economists have long predicted that the tax base will shift from taxes on capital (the relatively mobile factors) to taxes on labor and land (the relatively immobile factor).173 Thus, the standard recommendation in the public finance literature for small, open economies is to refrain from levying any taxes on capital because its mobility means that the tax burden will be shifted to domestic labor and land, and it is more efficient to tax those factors directly.174 Can we in fact observe such a shift from taxes on capital to taxes on labor? A good place to start would be to consider the “tax mix”, i.e., the relative percentage of various types of U.S., 68% of all stocks are held by the wealthiest 5% of households. See Martha Starr-McLuer, Stock Market Wealth and Consumer Spending, Federal Reserve Board, Finance and Economics Discussion Series No. 1998-20. The income of that segment derives to a large extent from (previously accumulated) capital, not current labor. See Henry J. Aaron & Alicia H. Munnell, Reassessing the Role for Wealth Transfer Taxes, 45 NAT 'L TAX J. 119, 130-32 (1992) (noting that "the role of bequests in wealth accumulation is clearly controversial and unresolved ... The central finding is that intergenerational wealth transfers are of sufficient size to establish a potential role for wealth transfer taxes to affect the distribution of wealth."); Anne L. Alstott, Colloquium on Wealth Transfer Taxation: The Uneasy Liberal Case Against Income and Wealth Transfer Taxation: A Response to Professor McCaffery, 51 TAX L. REV. 363 (1996). On the other hand, the half of American families whose incomes are below $50,000 per year have less than $12,500 in total financial assets and thus must depend almost entirely on labor earnings for survival. See M ICHAEL J. GRAETZ & JERRY M ASHAW , TRUE SECURITY 75-76 (1999). It is in this sense that the current discussion focuses on the undertaxation of capital compared to labor income as a distributive issue. 172 The exception is highly skilled labor which finds it much easier to immigrate. Thus, tax competition raises the same issues for such highly skilled labor as for capital. See, e.g., INCOME TAXATION AND INTERNATIONAL M OBILITY, (Jagdish N. Bhagwati & John D. Wilson eds., 1989); Assaf Razin & Efraim Sadka, Tax Burden and Migration: A Political Economy Perspective, NBER Working Article 5850 (1996). The following discussion will assume that such high-skilled labor is akin to capital (which is realistic in the sense that it is hard to distinguish the return to capital and labor for entrepreneurs), and the equity analysis will focus on low-skill labor. 173 See, e.g., Gordon, Can Capital Income Taxes Survive in Open Economies, supra note 32; Vito Tanzi, Globalization, Tax Competition and the Future of Tax Systems, in STEUERSYSTEME DER ZUKUNFT 11 (Gerold Krause-Junk ed., 1996); TANZI, supra note 11, at 138-139. 174 See, e.g., Gordon, Can Capital Income Taxes Survive in Open Economies, supra note 32; Assaf Razin et al., supra note 29. This recommendation is modified if the taxes imposed by the host country can be credited in the home country. See the discussion of this issue in Part IV(b) below.

40

tax in total tax revenues, in OECD member countries from 1965 (i.e., before the relaxation of exchange controls) to the present. The data for 1965-1995 are summarized in the following Table:175 Table 1.

Type of Tax

Structure of Taxation in OECD Countries (Percent of total revenue) 1965 1970 1975 1980 1985 1990 1994 1995

Personal income 26 Corporate income 9 Social security 18 Property 8 General consumption 12 Other goods & services 26

28 9 20 7 14 22

30 8 22 6 13 19

31 8 22 5 14 18

30 29 8 8 22 23 5 6 16 17 18 14

27 8 25 6 18 15

27 8 25 5 18 15

On the basis of these data, the following observations can be made. First, the data should be interpreted against an overall tendency for government revenues to grow as a percentage of GDP. This is true not just for OECD member countries (from an unweighted average of 30% in 1975-80 to 34.3% in 1986-92) but also for most developing countries.176 This trend has continued in OECD member countries, where in 1994 average government revenues reached 37.6% of GDP.177 However, the rate of growth has been considerably less steep in the 1980’s and early 1990’s than in the 1970’s.178 As discussed in Part IV below, the overall growth in the tax burden can be explained in terms of the growing role of the state in providing social insurance. The slower recent growth can be explained in terms of the state facing political and practical obstacles to raising taxes to maintain further growth in social insurance programs. The growth in overall government revenues means that the 25% of total revenues attributed to social security (payroll) taxes in 1995 represents a significantly higher percentage of GDP than the 26% of total revenue attributed to “other goods and services” (i.e., excise taxes and tariffs) in 1965.

175

This table is taken JEFFREY OWENS & JACQUES SASSEVILLE, EMERGING ISSUES IN TAX REFORM 9 (1997), Table 1. More detailed data for 1975-1992, which also cover non-OECD members, can be found in A. WoldeMariam, Summary Tax Structure Tables, 1975-92, in TAX POLICY HANDBOOK 289-318 (Parthasarathi Shome ed., 1995). 176 Total government revenues as a percentage of GDP rose from 20.5% (1975-80) to 21.7% (1986-92) in African countries; from 16.6% to 18.9% in non-OECD Asian countries; and from 19.3% to 19.7% in nonOECD Western Hemisphere countries. Only in Middle Eastern countries was there a decline (from 37.5% to 28.3%, which includes non-tax oil revenues, but tax revenues also declined from 15.7% to 13.6%). See id. 177 See OWENS & SASSEVILLE, 7. GLOBALIZATION, supra note 166, at Table 4.1 shows that for a sample of industrialized countries (the European countries, Canada, Japan, and the U.S.) government expenditures reached 49% of GDP in 1994. 178 See OWENS & SASSEVILLE, supra note 175, at 7, Figure 1.

41

Second, revenues from both the corporate and personal income taxes have tended to be flat as a percentage of total revenues (although rising as total revenues rose). In OECD member countries, revenues from the personal income tax were 26% of total tax revenues in 1965 and 27 in 1995.179 Revenues from the corporate income tax in the same years were 9% and 8% respectively. 180 While in developing countries the corporate income tax is more important than the individual income tax, the same flatness can be observed for those countries as well.181 Because the income tax taxes both labor and capital, this flatness may mask a trend to shift the tax from capital to labor.182 The flatness is striking because in OECD member countries tax rates have tended to go down in recent years.183 Third, a trend in both OECD member countries and developing countries is the rise in taxes on consumption, which are usually considered as taxes on labor. In OECD member countries, general consumption tax (VAT) revenues have risen from 12% of total tax revenues in 1965 to 18% in 1995.184 In developing countries, general consumption taxes rose from 25.5% of total tax revenues in 1975-80 to 31.8% in 1986-92. Most of this increase can be explained by rises in the tax rate: Owens shows that almost every OECD member country that This flatness masks a slight rise in the late 1970’s (to a peak of 31% in 1980) and subsequent decline. See id. at 9. 180 See id. at 9. The average masks considerable inter-country variations: For example, in New Zealand revenues from the individual income tax declined from 62.6% of total tax revenues in 1975-80 to 53.6% in 1986-92 while revenues from the corporate income tax declined from 10.8% to 8.3%. In the U.S. revenues from the individual income tax were flat at 46.6% of total tax revenues in 1975-80 and 46.1% in 1986-92 while revenues from the corporate income tax declined from 14.7% to 9.8%. See Summary Tax Structure Tables (supra note 175). 181 In African countries revenues for the individual income taxes for 1975-80 and 1986-92 were 10.2% and 11.4% of total tax revenues while the figures for the corporate income tax were 17.5% and 17.1%. A small decline in revenues can be observed in Asian countries (from 15.4% to 14.7% in the individual tax and from 19.7% to 16.9% in the corporate tax). In Latin American countries there was a small decline in the individual tax (from 8.5% to 6.4%) and a small rise in the corporate tax (from 13.2% to 14%). See Summary Tax Structure Tables (supra note 175). 182 The individual income tax typically taxes income from both labor and capital, but the tax rate on capital even in developed countries varies tremendously with the type of asset. See Jeffrey Owens, Tax Reform for the 21 st Century, 97 TAX NOTES INT ’L 583, 591 (1997) (calculating effective tax rates on pensions, housing, equities, and certificates of deposit for OECD members). For 1983, Gordon and Slemrod have calculated that the U.S. collected no tax on capital, see Roger H. Gordon & Joel Slemrod, Do We Collect Any Revenue from Taxing Capital Income, in TAX POLICY AND THE ECONOMY 2 (Lawrence H. Summers ed., 1988). However, the changes made in the 1986 Act may have changed this analysis. The incidence of the corporate income tax is notoriously hard to assess; a recent estimate has put it in the long run at partly falling on capital and partly on labor, see U.S. Treasury, Report on the Integration of the Individual and Corporate Tax Systems: Taxing Business Income Once (Jan. 1992). This incidence may shift with changes in market conditions, including globalization (Harberger’s conclusion that the incidence is on all capital was based on a closed economy model), see Arnold C. Harberger, The Incidence of the Corporation Income Tax, 70 J. POL. ECON. 215 (1962). See the discussion of incidence and its implications in Part IV(b) below. 183 See OWENS & SASSEVILLE, supra note 175, at 11; Owens, supra note 182, at 585-86; Chennells & Griffith, supra note 56, at ch. 3. 184 To some extent this rise was offset by a decline in revenues from tariffs and excises (classified as other goods and services), but given the rise in overall government revenues, it represents a significant increase in the tax burden. 179

42

has a VAT (i.e., all OECD members, except the U.S.) has raised the standard rate since the tax was introduced.185 Fourth, in OECD member countries another trend is the rise in payroll (social security) taxes to fund social insurance programs. Such payroll taxes rose from 18% of total tax revenues in 1965 to 25% in 1995.186 In the majority of OECD member countries more revenue was raised from social security taxes in 1995 than from the income tax.187 This trend can be explained by the expansion of social insurance programs as a result of the aging of the population and increases in unemployment benefits, which are discussed in Part IV.188 The combination of these trends can be summarized as follows: From 1965 to 1995, governments all over the world increased substantially the proportion of GDP that they collected as tax revenues. As revenues from both the individual and corporate income taxes have been generally flat (as a percentage of total revenues) over this period, the increase in total tax revenues was financed by increases in consumption taxes (in all countries) and in payroll taxes (in developed countries). Since both consumption taxes and payroll taxes fall on labor, while income taxes may be imposed on both labor and capital, the data are consistent with the shift from taxing capital to taxing labor that has been predicted theoretically to occur as a consequence of globalization. These broad data on the tax mix can be supplemented with more specific data on trends in income taxation, which suggest that the effective income tax rate on income from capital may be declining. Effective tax rates can be calculated in three ways. First, using the model developed by King and Fullerton, the effective marginal tax rate (EMTR) can be calculated for a country by using a simulated investment project and assuming certain inflation and interest rates. The EMTR is the difference between the pre- and post-tax rate of return for the marginal investment. Second, a similar model can be used to calculate the effective average tax rate (EATR) which is relevant for a project that earns some economic rent (as most projects financed by multinationals tend to do). The EATR is the difference between the value of the project in the presence and absence of tax in a given country. 189 Finally, a different measure of effective tax rates is to attempt to calculate the actual average tax rate (ATR) based on financial data published by firms in their financial accounts. The ATR differs from the EMTR and EATR

The only exception are Canada and Switzerland which only introduced the VAT in the 1990’s. See id. at 12. Australia has just introduced a VAT, leaving the U.S. as the only VAT-less OECD member. 186 See id. at 9. The rise from 1975-80 to 1986-92 was less pronounced (from 24.9% to 25.6%) suggesting that most of the rise took place in the early 1970’s. 187 See id. at 13. 188 Payroll taxes are less important in developing countries where government spending on general programs and government employment are more significant than direct social insurance. In non-OECD countries, payroll taxes have tended to be flat from 1975-80 to 1986-92. 189 See Chennells & Griffith, supra note 56, at 37-38. 185

43

in that it is based on actual firm-level data and not on hypothetical assumptions, but the quality of the available data varies greatly from country to country.190 The advantage of all three methods is that they attempt to calculate effective tax rates and therefore take into account both changes in the statutory tax rate and changes in the tax base. Thus, if tax reform leads to a reduction in the statutory rate that is offset by widening the tax base (as most tax reforms in the 1980’s in OECD member countries aimed to do), effective tax rates should be unchanged.191 Chennells and Griffith report results from calculating EMTRs, EATRs, and ATRs for a sample of ten OECD member countries over the period 1979 to 1994. They found that domestic EMTRs declined from an average of 21.7 in 1979 to 20.5 in 1994. Domestic EATRs, which are more relevant for foreign direct investment, declined over the same period from 21.7 to 17.9. Finally, ATRs based on actual accounting data for six of the countries (Australia, France, Germany, Japan, the U.K., and the U.S.) declined from 40% in 1985 to 32.6% in 1994. Since these calculations are based on corporate income taxes, they can be assumed to fall on income which is mobile relative to labor income.192 Further confirmation of this trend for effective corporate tax rates to decline can be found in Harry Grubert’s data based on actual tax returns filed by U.S.-based multinationals showing taxes paid by their overseas affiliates.193 Grubert studied changes in the average affective tax rates in a sample of sixty countries for the period 1984 to 1992 and supplemented it by published financial data for the years after 1992. He found that ATRs fell from 32.9% in 1984 to 23% in 1992, a decline of almost ten percentage points. Statutory rates also fell but by less than effective tax rates (from 41.2% to 33.4%).194 Moreover, the decline in ATRs was largest in countries with a population of less than 15 million, which are likely to be more sensitive to capital flows, and smallest in countries with more restrictions of capital flows. Thus, the data are consistent with the hypothesis that capital mobility may be driving effective rates on income from capital down.195 190

Most ATR calculations are based on the income tax footnote as summarized in the COMPUSTAT database and represent the ratio of the net provision for current taxes to net pre-tax income. Data based on tax returns are superior but are not usually available. 191 For an overview of tax reforms in the 1980’s, see OWENS & SASSEVILLE, supra note 175; Chennells & Griffith, supra note 56, at ch. 2. 192 Chennells and Griffith argue that their data do not exhibit tax competition because there is little evidence of convergence. But, as noted in Part II above, until taxes reach 0% tax competition can lead to a general downward trend without any convergence taking place. 193 See Grubert, supra note 58. Grubert used Form 5471 which is filed for every CFC and includes foreign taxes paid and net income reported. 194 U.S. Investment Abroad, supra note 54, found a decline of more than 15% in ATRs between 1984 and 1992. Statutory tax rates fell by 14% indicating rate reductions that were not compensated by base broadening. 195 Grubert notes that there was little convergence and that reductions in tax were lower in the EU and argues that these points militate against the tax competition explanation. But as shown in Part II, tax competition does not imply convergence, and EU tax rates may have been kept up by the need to fund generous social insurance programs (as explained in Part IV, however, this ability to tax capital to fund social insurance may not last).

44

Rodrik has calculated unweighted average effective tax rates on capital and labor for France, Germany, the U.S., and the U.K. for the period 1970-1991. The data indicate that taxes on both capital and labor in those countries went up in tandem from 1970 until about 1981, but since then taxes on capital have gone down while taxes on labor have continued to go up.196 Finally, there are a series of studies that used data from national income accounts to calculate tax rates. The advantage of this method is that it uses actual revenues to construct effective tax rates on consumption, capital, and labor separately. 197 Mendoza, Razin, and Tesar calculated tax rates for the G-7 countries from 1965 to 1988 and found that while taxes on consumption and capital income tended to be stationary, the tax rate on labor income has followed an increasing trend in all countries.198 In another study, Mendoza and his colleagues used the same methodology to calculate tax rates for eighteen OECD member countries for the period 1965-1991.199 They then calculated a regression relating those data to a measure of “lagged openness,” defined as the sum of imports plus exports divided by total output in the economy for the previous year. They found that taxes on labor respond positively to increases in lagged openness while taxes on capital respond negatively.200 That is, the more open to capital flows an economy becomes, the more it tends to shift the tax burden from capital to labor consistently with the prediction. What are the equity implications of these data? Since labor is generally less mobile than capital, a decline in taxes on capital and a rise in taxes on labor will generally involve a shift in tax burden which cannot be offset by income shifting and therefore tends to violate vertical equity. More broadly, since the rich save more than the poor, taxes on labor (consumption and payroll taxes) are generally more regressive (i.e., tax the poor more heavily) than taxes on capital (savings). Thus, a shift in tax burden from capital to labor would tend to render the tax system more regressive. Such a tax system is also less capable of redistributing resources from the rich to the poor.201 As a result, the overall distribution of income in a society compared to a

See GLOBALIZATION, supra note 166, at 65 (Figure 4.4). For this purpose, taxes on consumption are general taxes on goods and services and excise taxes. Taxes on labor are the portion of the individual income tax that falls on wages, social security taxes (both employee and employer portions), and payroll taxes. Taxes on capital are the portion of the individual income tax that falls on non-wage income, the corporate income tax, property taxes, and taxes on financial transactions. 198 See Enrique G. Mendoza et al., Effective Tax Rates in Macroeconomics: Cross-Country Estimates of Tax Rates on Factor Incomes and Consumption, 34 J. M ONETARY ECON. 297 (1994). 199 These countries are Australia, Austria, Belgium, Canada, Denmark, Finland, France, Germany, Italy, Japan, Netherlands, New Zealand, Norway, Spain, Sweden, Switzerland, the U.K., and the U.S. 200 See Enrique G. Mendoza et al., On the Effectiveness of Tax Policy in Altering Long-Run Growth: Harberger’s Superneutrality Conjecture, (1996) (CEPR Discussion Paper 1378). 201 A graduated consumption tax of the cash flow type can be progressive, but this does not apply to the VAT which typically has only one or at most two rates. In addition, the exclusion of income from savings from the tax base means that in most likely scenarios even a cash flow consumption tax will be less progressive than the income tax. Payroll taxes tend to be regressive because they are typically proportional and capped at a certain level of income. 196 197

45

proportional distribution, which is measured by the Gini coefficient, would tend to become more inequitable (i.e., the Gini coefficient would rise).202 The distribution of before-tax income has widened in the 1980s and 1990s in several OECD countries, including the U.K. and the U.S. Given the changes in tax patterns surveyed above, this means that the change in after-tax distribution of income is likely to be greater with an increase in the Gini coefficient.203 Nor is this pattern limited to OECD member countries. The evidence from Latin America is that as countries have opened their borders, income inequality has tended to rise.204 For example, in Chile, which led the trend of Latin American trade and investment liberalization, the Gini coefficient rose from 0.46 in 1971 to 0.58 in 1989 which is one of the largest jumps ever witnessed in a country over such a short period of time.205 To sum up: As capital mobility has increased since the relaxation of exchange controls in the early 1980’s, taxes on capital have tended to go down in both developed and developing countries while taxes on labor have tended to go up. Since labor is relatively less mobile than capital, it was generally not able to turn the inequity into an inefficiency by moving to countries with lower tax rates. Moreover, since capital income is disproportionately earned by the rich, the shift in tax burden from capital to labor has tended to make all societies less equitable (in terms of the distribution of income or wealth) than before. Remedying this situation calls for finding ways to tax capital despite its relatively high mobility, in the ways explored in Part V. c. Democracy: Two Types Of Tax Competition In the preceding two sections I have focused on the ways in which tax competition on income from capital can be harmful: It can lead to an inefficient global allocation of capital, and it can force governments to use relatively regressive forms of taxation and limit their ability to use taxes on capital for redistributive purposes. These efficiency and equity considerations all lead to the conclusion that tax competition should be limited. However, these considerations need to be balanced against another: Tax competition can reflect the different preferences of citizens in democratic countries for differing sizes of their governments. In fact, the level and type of taxation is one of the most important decisions made by voters in democratic countries, and it provides an image of the kind of society they wish to 202

The Lorenz curve is the relationship between the percentage of income and the percentage of individuals in a society. A straight Lorenz curve implies that the distribution of income in a society is completely equal (i.e., the bottom 50% of the people have 50% of the income). The Gini coefficient is the ratio of the area above the Lorenz curve to that below the diagonal (complete equality). A zero Gini coefficient means complete equality while a Gini coefficient of 1 means that all income accrues to one individual (complete inequality). See Howell Zee, Taxation and Equity, in TAX POLICY HANDBOOK 32, supra note 175. 203 See OWENS & SASSEVILLE, supra note 175, at 35. In the 1980's the Gini coefficient increased in all of the ten OECD countries studied in A TKINSON ET AL., INCOME DISTRIBUTION IN OECD COUNTRIES 49-51 (1995). 204 See Dani Rodrik, Understanding Economic Policy Reform, 34 J. OF ECON. LIT . 9 (Mar. 1996). 205 See id.

46

live in. From a welfare perspective, tax competition is beneficial to the extent it reflects these voter preferences. Thus, it becomes necessary to inquire whether it is possible to distinguish between types of tax competition that are likely to reflect such preferences and those that are less likely to do so. A clear case can be made for limiting the latter type of tax competition; the case for limiting the former type (tax competition that reflects preferences for the desirable size of government) is more problematic. We can start with the Tiebout model in which tax competition (or at least differences in tax rates among jurisdictions, since Tiebout assumes that these are fixed) is beneficial because it enables citizens to maximize their welfare by voting with their feet.206 On its face, this model is inapplicable to the problem that lies at the heart of this article, i.e., how to tax mobile capital in an open economy. The model is inapplicable because on the international level individuals cannot easily choose to move from one country to another and because taxes are used for redistribution as well as payments for government goods and services. However, the degree to which the Tiebout model is inapplicable can be exaggerated. First, even though citizens cannot move from country to country, in democratic countries they can vote, and given the salience of the tax issue in politics it seems likely that their votes actually effect the level of government spending.207 Thus, one can envisage a situation in which the world is divided into jurisdictions with different levels of public spending depending on the wishes of (the majority of) each country’s citizens. Second, the focus in tax competition has been on source-based corporate taxes.208 These taxes are much closer to the benefits taxes envisaged by Tiebout than is generally recognized. They are not a significant factor in redistribution (in fact, their incidence is both unknown and possibly shifting). Instead, they represent a payment to the source country for the costs it incurs to enable investment by a multinational as well as a way of assuring source countries (and especially developing source countries) an adequate tax base.209 See Tiebout, supra note 60. The level and form of taxation has been a crucial issue in recent elections in the U.S., Canada, Japan, and Australia. See Alison Mitchell, Republicans Looking Right and Left for the Best Tax Cut Plan, N.Y. TIMES, Mar. 22, 1999, at A13; Why are These Republicans Smiling?, ECONOMIST , Nov. 8, 1997, at 29; Robert S. Greenberger, Canadian Elections Promise Acid Test on Lower Taxes or Improved Services, W ALL ST . J., June 2, 1999, at A21; Hold On, Japan's Busy, ECONOMIST , Sept. 12, 1998, at 39; Richard J. Vann & Graeme S. Cooper, Report Of Australia's Business Tax Review And The Government's Response, 1999 W ORLDWIDE TAX DAILY 191-6 (Oct. 4, 1999). 208 Source-based taxation on portfolio income of individuals are also at issue but only as a backstop for residence-based taxation; if residence-based taxation of individuals could be assured, no problem would arise even if tax competition meant zero source based taxation. For corporations, however, residence-based taxation is implausible, as argued in Part V. In regard to individual investors, a different type of benefit analysis applies: Investors who reside in high-tax countries but earn untaxed overseas income are free riders on the benefits provided by the high tax country (i.e., they do not pay the price of civilization). 209 For an argument that source-based corporate income taxation is correlated with costs imposed see International Taxation of Electronic Commerce, supra note 34, at 520-21. For the inter-nation equity aspect of allowing developing countries to tax multinationals, see Part IV. 206 207

47

This leads to an important distinction: If a multinational corporation invests in a country with a small public sector, it is likely to have to incur more costs or earn lower profits than if the public sector were larger. For example, it may have to invest in training its workers because the government provides inadequate public education, or it may have to build its own transportation systems because the government does not provide adequate infrastructure. In that case, it would be inappropriate to penalize the multinational by taxing it at a higher rate (e.g., its home country rate) to offset the lower taxes levied by the host country. 210 If the multinational faces lower profits, the incentive to locate in the low-tax country would be limited, and therefore the negative effects of tax competition would be limited as well. But this argument holds true only if the public sector is smaller in general as a result of a generally applicable tax decrease. If, on the other hand, the government decides to maintain generally high taxes on immobile factors (such as labor and land) but to grant the multinational a targeted tax holiday (or equivalently to subsidize it with targeted expenditures on training and infrastructure), the multinational obtains a windfall. Having made the decision to invest somewhere, it can then choose the location that offers it the best tax and subsidy package with the awareness that because the general level of taxation on domestic labor and land is high it will not have to incur additional costs. Thus, this line of analysis suggests that a distinction should be made between tax competition in the form of generally applicable tax decreases that reduce the overall size of the public sector and tax decreases (or subsidies) that are limited only to foreign investors. In the former case, the tax reduction represents the wishes of the electorate and does not confer a windfall on foreign investors. In the latter case, the electorate is unlikely to be involved (since tax holidays to foreign investors are rarely the subject of political attention even if they are made public).211 Moreover, the tax holiday represents a windfall to the multinational since it benefits from a large public sector financed by overall high taxes. The same distinction can also be made from a Leviathan perspective. Tax competition may be beneficial when it reduces wasteful government spending but only if it actually reduces the size of government. Overall tax reductions can be beneficial from this perspective especially if mandated by the voters and if they result in smaller governments. Targeted incentives aimed at foreigners are not beneficial because they are typically too small in the aggregate to affect the overall size of government. In fact, targeted tax incentives can themselves be seen as a form of government waste: They enable elected officials to boast about job creation in the short run 210

While the tax paid by the multinational in its home country would be lower to the extent that its (deductible) costs in the low-tax country are higher, this is the equivalent of granting a deduction rather than a credit for the foregone host country tax. It would not compensate the multinational for its lower profit. 211 A rare exception to this rule was the benefit package granted by Alabama to attract Mercedes Benz which was widely criticized because the dollar amount per job created was three times the amount paid by any state to a foreign investor. This became an election issue and helped cost Governor Folsom his job. See Allen R. Myerson, O Governor, Won't You Buy Me a Mercedes Plant?, N.Y. TIMES, Sept. 1, 1996, at 3.

48

without engaging in an adequate cost/benefit analysis in terms of lost revenues in the longer run.212 The costs of targeted tax incentives (in lost revenues and congestion problems) are likely to be less salient politically than the immediate benefits in job creation, especially where unemployment is a central political issue. Thus, I would suggest that the line between harmful and beneficial tax competition be drawn between general tax reductions that apply to all taxpayers (domestic and foreign) and targeted tax reductions that are granted only to foreign investors. Harmful tax competition should thus include both production and headquarters tax havens since these are typically only granted to foreign multinationals. Harmful tax competition should also include reductions in withholding taxes that are aimed solely at attracting foreign capital, unless measures are taken to ensure residence based taxation. 213 Finally, harmful tax competition also includes traditional offshore tax havens with an overall low or zero tax rate, if the tax base is confined almost entirely to foreign investors.214 Such investors in traditional tax havens have no real business operations in the tax haven and therefore do not suffer a detriment from the small size of the public sector. However, harmful tax competition does not include overall reductions in the tax rate that apply to all taxpayers in a jurisdiction when the tax base includes significant numbers of domestic taxpayers. It may be objected at this point that the arguments made in the preceding sections against tax competition on efficiency and equity grounds apply to all forms of tax competition and not just to harmful tax competition (as defined above). After all, a generally applicable tax decrease can lead to a large flow of capital from one country to another and result in an inefficient allocation of capital and limit the ability of the home country to redistribute income equitably. However, as noted above, such considerations need to be balanced against the ability of voters in democratic countries to determine the desirable size of their public sector, even if this has potential adverse consequences for other countries. This is both a normative judgment and a practical one: As the failed attempts to impose minimum tax rates in Europe show, countries would be very reluctant to give up their right to set generally applicable tax rates which is a core attribute of sovereignty.215 On the other hand, countries would be more likely to give 212

Most studies of targeted tax benefits designed to attract investment in both the state and international context have found that, while they are effective in attracting the investment, the ultimate benefits to the host country are insufficient in the long run to compensate for the lost revenues. See Part IV. 213 Since individual income taxes should be levied primarily by the home country, it is acceptable to reduce withholding taxes if residence-based taxation is assured. 214 A line needs to be drawn to define what “almost entirely” means. Perhaps a tax incentive 90% or more of whose benefits accrue to foreigners would be harmful (traditional tax havens would generally cross this line). While marginal cases can be imagined, in most situations the distinction would be clear. 215 In 1975, the European Commission proposed a directive for uniform minimum and maximum corporate tax rates which was never adopted and was withdrawn in 1990, see Commission Communication to Parliament and the Council, Guidelines on Company Taxation (SEC 90) 601 final, Brussels, Apr. 20, 1990 (withdrawing Corporation Tax Directive, COM(75) 392 final). The Ruding committee made a similar proposal which was never seriously considered, see RUDING REPORT , supra note 16, at 202.

49

up their right to target tax incentives at foreigners if the main reason for granting the incentive is fear that otherwise the capital would flow to another country that does grant such incentives. In this kind of an assurance game situation, multilateral action is possible to achieve the best outcome for all concerned.216 The distinction proposed here between harmful and beneficial tax competition can be illustrated by two examples drawn from recent U.S. history. Ronald Reagan was elected president in 1980 on the basis of explicit promises to reduce the size of the federal government and cut taxes. In 1981, these promises were fulfilled through the largest individual and corporate tax decrease in U.S. history. In particular, the adoption of the accelerated cost recovery system and an investment tax credit meant that new capital investment by corporations would be more than expensed, resulting in negative tax rates on normal corporate profits. This, in turn, led to a proliferation of tax shelters for individuals, so that Gordon and Slemrod estimated that in 1983 the U.S. collected no tax whatsoever on capital income.217 The other result of the Economic Recovery Tax Act of 1981, together with the Reagan defense build-up, was a burgeoning budget deficit which was in fact predicted by the Congressional Budget Office. This deficit was financed by a flow of foreign capital into the U.S., which at least in part was the direct result of lower tax rates. Sinn has estimated that the ACRS and ITC were responsible for a flow of between $1 and $1.5 trillion into the U.S. from 1981 to 1984, or 7% of the world’s capital stock.218 This influx of foreign capital could have been predicted at the time of enactment since in the absence of an increase in domestic (private) savings and given a decline in aggregate (public and private) savings, additional investment could only be financed from abroad.219 Sinn has also argued that the consequences of this inflow were clearly beneficial to the U.S. and detrimental to the rest of the world: The first half of the eighties was characterized by enormous capital imports into the U.S. accompanied by a strong dollar and a high world interest rate level. Most countries suffered from this situation. Europe was driven into the worst recession of the post-war period, and the developing countries were shaken by one debt crisis after another. A number of countries were unable to meet their interest obligations, and a collapse of the world banking system was avoided only by strenuous effort. The U.S. alone seemed to 216

See Part V for a fuller discussion of how such cooperative outcomes can be implemented. See Gordon & Slemrod, supra note 182. The 1981 tax cuts were partially reversed in 1982; and in the 1986 Tax Reform Act , tax shelters were largely eliminated, the tax base was broadened, and rates were reduced. The 1986 Act was intended to be revenue neutral so it preserved the overall tax cut of 1981 (and budget deficits persisted through the 1980s). The 1986 Act led many foreign governments to reduce their rates and broaden their base. 218 See Hans Werner Sinn, United States Tax Reform 1981 and 1986: Impact on International Capital Markets and Capital Flows, in INFLUENCE OF TAX DIFFERENTIALS 25, supra note 22. 219 See International Considerations, supra note 22, at 7. However, McLure stated that “[r]elatively few seem to have appreciated the full international consequences of these actions.” 217

50

have benefited; despite the high interest rate, it enjoyed a significant consumption and investment boom.220 Thus, Sinn would support actions taken by Europe to curb such tax competition by the U.S., e.g., by taxing its residents on their U.S. source income. But it is highly unlikely that the U.S. would have agreed to cooperate in such a move (and thus forego the benefits of its tax reduction). Without cooperation from the U.S. it is highly unlikely that Europe could have succeeded in taxing its individual residents on their U.S. source income.221 Moreover, from a normative perspective the U.S. could argue that it had no obligation to maintain a larger public sector than its citizens desired and that its citizens were entitled to the tax cuts they voted for, even if they had negative incidental effects on Europe. The primary obligation of the U.S. government is the welfare of U.S. citizens who elect it, not of Europeans. On the other hand, consider the portfolio interest exemption enacted in 1984 to facilitate financing the same deficit. As described in Part II, this was a targeted provision aimed solely at foreigners. It was not widely discussed and was passed in Congress largely at the behest of the Treasury and U.S. banks.222 As a provision aimed solely at foreigners, it had a relatively small impact on the overall size of government (it is not even included in the tax expenditure budget). But the results for Latin American countries, which lost $300 billion in capital flight as a direct result of enacting the exemption, have been devastating.223 In part because of the lack of public exposure, these results were apparently not foreseen by the administration. 224 If they had been, perhaps another way could have been found to reduce the cost of borrowing (e.g., by maintaining the Antilles treaty or renegotiating the treaty with Japan). Thus, I would recommend that proposals to curb tax competition (such as the current proposals by the EU and OECD) should be evaluated in part based on the extent to which they observe the above distinction, as well as whether they are likely to prove effective in practice. In particular, proposals that would have limited the ability of the U.S. to enact ERTA 1981 are unlikely to succeed in practice; proposals that would have prevented the enactment of the portfolio interest exemption are much more likely to be both beneficial and successful. In Part V, I will attempt such an evaluation.

220

HANS-W ERNER SINN, CAPITAL INCOME TAXATION AND RESOURCE A LLOCATION 224-25 (1987). Most of the capital inflow was in the form of individual portfolio investment which, as argued in Part II, is very hard to tax unilaterally on a residence basis. 222 The banks also succeeded in inserting a provision denying the benefits of the exemption to foreign banks. See I.R.C. § 881(c)(3)(A). 223 See Capital Flight from Latin America, supra note 25. 224 See International Considerations, supra note 22, at 8. 221

51

IV.

The Costs and Benefits of Tax Competition from Each Country's Perspective

In Part III, I discussed the arguments for and against tax competition from a global perspective. However, international tax rules are not designed by an authoritative "world tax organization," but are set by each country separately with its own interests in mind. Thus, in this Part, I will attempt to analyze the costs and benefits of tax competition from the perspective of the countries affected by it, turning first to developed welfare states and then to developing and transition economies. Finally, I will attempt to analyze how tax competition affects the division of tax revenues among countries which relates to the concept of inter-nation equity. a. Developed Countries and the Fiscal Crisis of the Welfare State The welfare state, which began over a hundred years ago with Bismarck's social insurance scheme (financed by a comprehensive income tax), is facing a severe fiscal crisis. The fundamental problem is an aging population, resulting from the baby boom of 1946-1970 and an increasing life span, together with a decreasing number of births. The result is a significant increase in the "dependency ratio," defined as the ratio of the young (to age twenty) and the elderly (above retirement age for public pensions) to the working age population, in almost all OECD member countries. For example, in Japan the dependency ratio rises from about 40% in 1995 to a peak of 60% in 2045; in Germany it rises from 30% in 1995 to over 60% in 2035. Even in the U.S., which has relatively high fertility and immigration rates, the dependency ratio rises from about 20% in 1995 to 40% in 2035.225 In the OECD in general, the percentage of the population over 60 rises from 18.2% in 1990 to 27.0% in 2020 and peaks at 31.2% in 2050.226 The fiscal implications of this "old age crisis” are profound.227 The OECD has published fiscal scenarios for its members that depend on their starting position in 2000, the nature of the demographic changes in each country, and the exposure of government budgets to the effects of aging through existing arrangements for public pensions and public health spending.228 The results are that by 2030 all of the OECD members studied except Ireland will experience a budget deficit ranging from 0.5% of GDP in Belgium to 8.8% of GDP in Japan (the U.S. is in about the middle with a deficit of 3.8% of GDP).229 Total expenditures on social insurance programs (including pensions, health, unemployment, and others) in all OECD countries are

See Deborah Roseveare et al., Aging Populations, Pensions Systems and Government Budgets: Simulations for 20 OECD Countries 27, Figure 1 (OECD Economics Department Working Papers 168, 1996). 226 See International Considerations, supra note 22, at 3. 227 See W ORLD BANK, A VERTING THE OLD A GE CRISIS (1994); see also ESTELLE JAMES, PROTECTING THE OLD AND PROMOTING GROWTH (1996). 228 See Willi Leibfritz et al., Aging Populations, Pension Systems and Government Budgets: How Do They Affect Savings? (OECD Economics Department Working Paper 156, 1995). 229 See id. at Table 5. 225

52

scheduled to rise from 18.3% of GDP in 1990 to 25.5% in 2050 with a significantly higher percentage in Japan (27%) and Western Europe (33.4% of GDP in 2050).230 The tax implications of this phenomenon can be analyzed by using "generational accounting." Generational accounting estimates what the lifetime net tax rates will have to be on future generations compared with current newborns in order to finance government expenditures at their current levels (i.e., if no changes are made to government programs). Net tax rates are taxes paid minus transfers received. For the U.S., Kotlikoff estimates that the net tax rate required on future generations will be an (impossibly high) 84%. To equalize lifetime net tax rates of current newborns and future generations from 2006 onward, Kotlikoff estimates that it will be necessary to either (i) increase federal income taxes by 61%; (ii) cut transfer payments by 43%; or (iii) reduce federal purchases by (an impossible) 109%.231 For other countries the picture is similar: The generational imbalance for males, i.e., the increase in net tax payments for future generations required to finance future benefits at current rates, ranges from 27% in Germany to 446% in Italy.232 The OECD study concludes that "[w]here major fiscal pressures have been identified … some policy action will be needed to redress imbalances. Without such action, government dissaving could also lead to significant reductions in national savings in some countries."233 The policy options identified by the OECD are (i) increasing the workforce through immigration, (ii) increasing taxes or cutting other expenditures, or (iii) cutting health and pension programs. The first option is rejected as unrealistic given the numbers of immigrants required. The second option is likewise rejected because cutting other expenditures by the necessary amount would require, e.g., the U.S. to eliminate all defense spending and Japan to cut twice all spending on general public services and defense. As for raising taxes, "[g]iven concerns about existing overall tax burdens, the option of raising taxes is not considered further."234 Thus, the third option remains as the only viable alternative.

230

ILO ANNUAL REPORT (1999), at Table 3.1. For more data on the old age crisis and its fiscal implications see OECD, NEW DIRECTIONS IN HEALTH CARE POLICY 13-19 (1995); OECD, M AINTAINING PROSPERITY IN AN A GING SOCIETY 9-27 (1998). 231 See Laurence J. Kotlikoff, Kotlikoff Testimony at House Budget Committee Hearing on Generational Accounting and Government Spending, 96 TAX NOTES TODAY 52-33 (Mar. 14, 1996). All of these may seem politically impossible, but if the U.S. were to introduce a VAT, revenues are likely to exceed 60% of current federal revenues from the income tax (in other OECD members, VAT revenues almost equal income tax revenues). In general, it is unclear why a debate on the future of social insurance in the U.S. has to assume revenue neutrality, as in GRAETZ & M ASHAW , supra note 171. 232 See Leibfritz et al., supra note 228, at Table 11. 233 Id. at 17. The last point refers to the possibility that the government finance its budget deficit through borrowing. This is generally regarded as having adverse effects on the economy by crowding out private investment options. See Theodore P. Seto, Drafting a Federal Balanced Budget Amendment That Does What It Is Supposed to Do (And No More), 106 YALE L.J. 1449 (1997); A NATION IN DEBT : ECONOMISTS DEBATE THE FEDERAL BUDGET DEFICIT (Richard H. Fink & Jack C. High eds., 1987); and THE FEDERAL DEFICIT (Andrew C. Kimmens ed., 1985). 234 Leibfritz et al., supra note 228, at 20.

53

The conclusion regarding taxes seems too broad. The key questions are what types of taxes are involved and in which countries. For the U.S., for example, since overall taxation is only about 30% of GDP, there appears theoretically to be considerable scope for raising additional revenues (e.g., through a VAT) before reaching European levels of public expenditure (about 50% of GDP). However, in Europe the scope for raising taxes is much more limited. Value added tax rates in many European countries now exceed 20% which is about as high as a consumption tax can go before the incentive to consume elsewhere becomes too significant and enforcement costs swallow the additional revenues.235 Most social insurance is currently funded through payroll taxes, but given the high level of unemployment in Europe, it would seem counterproductive to raise payroll tax rates even further.236 In addition, high payroll tax rates may also create inefficient incentives for employers to hire part time or foreign workers whose wages are not subject to the same rates.237 This still leaves the option of raising more revenues through the income tax. The impediment to doing so is the risk of exacerbating current distortions in the choice of whether to work and whether to save or consume which are inherent in Europe's current high rates of income tax. However, as discussed in Part II above, income from cross-border investments is currently taxed at much lower effective rates than generally applicable ones. Therefore, to tax such income at rates that approximate the rates on domestic investment would both advance world efficiency and equity and also raise more revenues to help address the fiscal crisis described above.238 This rationale may explain why (as discussed below) both the EU and the OECD are currently showing such interest in curbing "harmful tax competition" even though there may be insufficient evidence of actual erosion of the tax base. The potential for erosion certainly exists, and OECD governments are acutely aware that at least until 2050 or so they will need every penny they can collect in taxes to sustain the promises made to the baby boom generation. Since the need for additional revenues is predictable on demographic grounds, and the potential for tax base erosion resulting from globalization exists, it is not surprising that the developed countries are taking actions to shore up the income tax base even before any significant erosion has occurred.

See OECD, CONSUMPTION TAX TRENDS Table 1.5 (1997) (showing rates at 20% or above in 10 OECD member countries including all the Nordic countries). 236 This point is the main conclusion of the OECD Jobs Study. See OECD, THE OECD JOBS STUDY: TAXATION, EMPLOYMENT , AND UNEMPLOYMENT (1995); OECD, IMPLEMENTING THE OECD JOBS STRATEGY (1997); OECD, M AKING W ORK PAY: TAXATION, BENEFITS, EMPLOYMENT AND UNEMPLOYMENT (1997). 237 See Jack M. Mintz, Is National Tax Policy Viable in the Face of Global Competition?, 1999 W ORLDWIDE TAX DAILY 128-20 (1999). 238 Although, as discussed above, it is unclear how much revenue can be raised in this way and it is unlikely that sufficient revenues are available even on the most optimistic scenario. For developing countries the numbers may be more significant. 235

54

However, this analysis will appear to many readers to beg the question: Why should the welfare state not be drastically reduced? Governments may be coerced to maintain existing programs by political realities, but from a normative perspective, it is less clear that government should be doling out billions in pensions and health care expenditures for the elderly, especially if much of this could be done through the private sector. Two answers can be given to this question, both of which ultimately depend on normative value judgments that not everybody may share. The first answer is the one given by Dani Rodrik: Failing to maintain existing social insurance programs risks causing a political backlash against globalization and a return to 1920s-style protectionism and isolationism.239 Rodrik points out that there exists an empirical relationship between economies becoming more open and the increase of spending on social insurance programs. Calculating the relationship between openness and the size of government yields an "unmistakable positive association" which is not confined only to OECD countries.240 The explanation lies in the role of social insurance in minimizing exposure to increased external risk. Opening the economy to competition from abroad increases risk to both income and consumption because of increased specialization and the difficulty of diversifying human capital.241 Theoretically, such increased risk could be offset by decreased exposure to domestic risks (e.g., by the ability to invest abroad), but empirically Rodrik shows that for 105 countries increased openness results in a net absolute increase in risk both to income and to consumption. 242 The risk is expressed both in terms of income insecurity (e.g., loss of employment) and income volatility (extreme fluctuations as a result of more frequent job shifts). As a distributive matter, the increased risk probably falls on the less mobile factors in society (primarily unskilled and semi-skilled labor) as indicated by increasing income disparities.243 Social insurance programs can reduce such risks. The most immediate link is to unemployment insurance which can cushion against increased risk to income levels and job volatility resulting from globalization. But pensions and health care coverage can also play a major role in reducing lifetime uncertainty over income which creates an increased willingness to tolerate exposure to external risks.244 If social insurance programs are cut as a result of the fiscal crisis described above, such buffers against the risks inherent in globalization disappear. Such cuts can give rise to massive See GLOBALIZATION, supra note 166, at 49-69. For a critique of Rodrik's approach see Paul B. Stephan, Book Review, 18 NW . J. INT 'L L. & BUS. 246 (1997). 240 GLOBALIZATION, supra note 166, at 52. 241 See id. at 55. 242 See id. at 57. 243 See id. at 58. In the 1980's, the Gini coefficient increased (i.e., income inequality increased) by 0.5-3% in all of the 10 OECD member countries studied in A TKINSON ET AL., supra note 203, at 49-51. See also Petter Gottschalk & Mary Joyce, The Impact of Technological Change, Deindustrialization, and Internationalization of Trade on Earnings Inequality, in POVERTY, INEQUALITY, AND THE FUTURE OF SOCIAL POLICY 197-230 (Katherine McFate et al. eds., 1995). 244 See GLOBALIZATION, supra note 166, at 58. 239

55

social resistance as exemplified by the French strikes of 1996 (occasioned by the need to cut social insurance programs to meet the Maastricht budgetary guidelines for monetary union). It is no accident that as Europe moved closer to economic integration in the 1990s, voters in almost all European countries elected center-left governments on a platform of protecting social insurance programs.245 At the same time, European voters gave increasing power to extreme right wing parties committed explicitly to fight against globalization.246 The danger, as Rodrik points out, is that if social insurance programs are eroded, this would produce a retreat from the commitment to openness which characterized the 1980's and early 1990's.247 Assuming (as most critics of the welfare state do) that globalization itself is a positive phenomenon, such an outcome would be unfortunate. Nobody wants to go back to the previous retreat from globalization which took place in the 1920's and led to a worldwide depression and a world war. The second argument in favor of social insurance programs is a purely normative one. To make it, one needs to distinguish between social insurance and social assistance, both of which are parts of the modern welfare state.248 Social assistance is government aid designed to ensure that every citizen is assured a minimal level of resources for living. Social insurance, on the other hand, is designed more minimally to ensure that citizens who are unable to work because of age (too young or too old), disability, health, or temporary unemployment can replace the income they had while working.249 Social insurance, unlike social assistance, is connected to employment status and (typically) to contributions paid by the employee. This article focuses on social insurance rather than social assistance for two reasons: First, social insurance is more clearly linked to the pressures resulting from globalization since globalization may result precisely in increased income insecurity and volatility for those who have (or had) jobs. Second, social insurance enjoys far broader acceptance among OECD members than social assistance: All OECD member countries have some form of social insurance programs and those programs are very popular, while only some OECD members have social assistance programs and those programs enjoy much lower popular support.250 Social insurance is popular precisely because in a market economy people may lose their source of income due to factors beyond their control. Nobody can prevent being too See Martin Walker, Europe's Third Way Labs; Tony Blair's Formula For Reform Is Contagious On The Continent, NEW DEMOCRAT , May/June 1999, at 8; Unsocialist Realism, ECONOMIST , Oct. 3, 1998, at 66. 246 See Tony Judt, The Social Question Redivivus, FOREIGN A FF., Sept./Oct., 1997, at 95; Serenella Sferza, What is Left of the Left?, 128 DAEDALUS 101 (1999). 247 See GLOBALIZATION, supra note 166, 9; see also Dani Rodrik, Trade, Social Insurance and the Limits of Globalization (NBER Working Paper 5905, 1997). 248 See the definitions in TONY EARDLEY ET AL., SOCIAL A SSISTANCE IN OECD COUNTRIES (1996); JOHN DITCH ET AL ., COMPARATIVE SOCIAL A SSISTANCE (1997). 249 See GRAETZ & M ASHAW , supra note 176. 250 See EARDLEY ET AL., supra note 253, at 12. For a measure of support in the U.S., see Virginia P. Reno & Robert B. Friedland, Strong Support But Low Confidence, in SOCIAL SECURITY IN THE 21ST CENTURY 17894 (Eric R. Kingson & James H. Schulz eds., 1997). 245

56

young or too old to be able to earn a living, nor is it possible to forestall illness, disability, or unemployment caused by factors beyond one's control. From an ex ante perspective, we all face a similar likelihood of suffering from such disruptions to income: "there but for the grace of God go I". Thus, it is not surprising that social insurance programs designed to prevent such risks from threatening a reasonable standard of living are very widespread. As Graetz and Mashaw state, "the social insurance systems of capitalist democracies have a solid claim to represent the greatest triumph of twentieth century domestic public policy… social insurance … is a crucial underpinning of a vibrant market economy."251 Moreover, for a lot of social insurance programs the private market is unable to provide adequate protection. This can occur for a variety of reasons. First, the risk can be too uncertain for private insurance companies to actuarially predict. Second, risks can covary in ways that threaten all policyholders at once, like the demographic changes described above. If the risk cannot be adequately diversified, private insurance may become impossible to obtain. Third, adverse selection may occur if the potential insured know the risks much better than the insurers resulting in poor risks being overrepresented in the insurance pool. This could result in such high premiums that all low and moderate risks are priced out of the market. Finally, there could be cases of extreme moral hazard, when the insured changes her behavior to increase the risk.252 Government insurance can alleviate these problems, at least theoretically, by a combination of taxation and regulation. The uncertainty problem can be overcome by paying compensation when the risk materializes and taxing to pay the costs at that point. Covariance of risk can be addressed because governments do not go out of business; they can borrow and then tax to meet their debts. Adverse selection disappears with a government mandate to insure everybody, and moral hazard can be reduced by effective regulation (e.g., requiring the unemployed to seek work).253 Globalization results (as argued above) in increased job insecurity, income volatility, and income disparities that exacerbate rather than reduce the need for government-provided social insurance. The risks posed by globalization are inherently uncertain and co-varying, falling on large numbers of people at once in an unpredictable pattern. In addition, both adverse selection and moral hazard problems are inherent to those risks. Insurance companies are typically less able to assess an employee's chances of being displaced by global competition than the employee (or her employer who in a typical private market provides the insurance). And moral See GRAETZ & M ASHAW , supra note 171, at 6, 12. It seems unrealistic, given the data in note 171 above, to expect most people to meet even predictable declines in income purely from their own savings. 252 See id. at 20-21; Thompson & Upp, The Social Insurance Approach and Social Security, in SOCIAL SECURITY IN THE 21ST CENTURY, supra note 250. For a discussion of the appropriate mix of public and private in providing social insurance from a European perspective, see also M ARTIN REIN & ESKIL W ADENSO , ENTERPRISE AND THE W ELFARE STATE 1-33 (1997); OECD, PRIVATE PENSIONS AND PUBLIC POLICY (1992). 253 See GRAETZ & M ASHAW , supra note 171, at 22-23; see also OECD, BENEFIT SYSTEMS AND W ORK INCENTIVES 49 (1998). 251

57

hazard is the major problem in providing unemployment insurance since the employee is able to take many steps to mitigate or exacerbate the risk. Nor is unemployment insurance the only type of social insurance affected by globalization. The largest social insurance expenditures are for pensions and health care costs. While these depend crucially on demographic trends that are unrelated to globalization, the risk of inadequate income in old age or upon falling ill also depends on what private savings are available. If employees have to save primarily to cover periods of temporary unemployment, they may not have enough left over for temporary or permanent loss of income resulting from disability or old age. Moreover, in a system like the U.S. in which both pensions and health insurance are tied to permanent full-time employment, there is a direct link between the decline of full time, unionized jobs with benefits attached that is connected to globalization and the increasing need for government mandated social insurance.254 Thus, there is a strong normative basis for the desire of the developed countries to preserve their existing social insurance programs in face of the dual risks posed by demographic realties and globalization. To do so, it is necessary to prevent the erosion of the individual and corporate income tax base which is still the biggest source of revenue in almost all OECD member countries. At least in Europe, increasing other taxes (i.e., the VAT and payroll taxes) to cover any shortfall in income tax revenues is not a viable alternative for the reasons discussed above.255 Moreover, since one result of globalization is an increase in income disparity, using the income tax to reduce such disparities by the provision of social insurance programs (which themselves have a redistributive aspect) seems particularly appropriate. Given the extent of the fiscal crisis described above, no amount of changes in the income tax is likely to produce sufficient revenues to keep existing programs unchanged in all OECD member states. In particular, the imposition of effective income taxation on crossborder transactions, as advocated in this article, would not suffice to solve the fiscal deficit. Moreover, it is not clear that existing programs should not be changed in ways that are consistent with their normative underpinnings. Most obviously, more means testing should be introduced and European countries are largely moving in this direction.256 But there is a strong normative case in favor of ensuring that the income tax base not be eroded by globalization in For a searing critique of the U.S. system, see GRAETZ & M ASHAW , supra note 171, at Part II; see also Michael Graetz, The Troubled Marriage of Retirement Security and Tax Policies, 135 U. PENN. L. REV. 851 (1987); Jeffrey Lehman, To Conceptualize, To Criticize, To Defend, To Improve: Understanding America's Welfare State, 101 YALE L. J. 685 (1991). For the vulnerable position of retirees in the U.S. and Canada see OECD, THE TRANSITION FROM W ORK TO RETIREMENT 119 (1995). 255 Even in the U.S., it can be argued that only by using general tax revenues (i.e., primarily from the income tax) can existing social insurance programs be preserved. See GRAETZ & M ASHAW , supra note 171, at 386. However, the U.S. social insurance system is grossly inadequate (as Graetz and Mashaw recognize) and to expand it to adequate levels would require adopting a VAT as every other OECD member country has done. 256 See EARDLEY ET AL., supra note 248, at 23; but see Eric R. Kingson & James H. Schulz, Should Social Security be Means Tested, in SOCIAL SECURITY IN THE 21ST CENTURY, supra note 250, at 41-61 (arguing against means testing on political grounds). 254

58

order to leave sufficient revenues to fund whatever social insurance programs can and should be retained. b. Developing Countries and Tax Incentives The analysis above suggests that it is in the interest of developed countries to prevent tax competition by developing countries from eroding their income tax base and that they have adequate normative grounds for wishing to do so. But what about the developing countries? Is it not in their interest to engage in tax competition to attract foreign investors, and would or should they resist any attempt by the developed countries to prevent them from doing so? Several arguments need to be addressed here. First, it can be argued that developing countries need the tax revenues less than developed countries because they are not welfare states and are less affected by the fiscal crisis described above. Second, an economic case can be made that developing countries should refrain from taxing any foreign investment. Finally, tax competition may be the only way for developing countries to attract crucially needed foreign investors even if there is a cost involved (i.e., they need the revenues and cannot obtain them otherwise). (i) Do Developing Countries Need the Revenues? A common misperception is that the fiscal crisis described above is a problem unique to the members of the OECD. In fact, the increase in dependency ratios is expected to take place in other geographic areas as well as fertility rates go down and health care improves. Outside the OECD and the transition economies, the dependency ratio starts in the single digits in the 1990s but rises to just below 30% by 2100.257 Nor is it accurate to say that social insurance programs are unique to the developed countries. A study by the International Labour Office shows that in Central and Eastern European countries total spending on social insurance is projected to reach 25.6% of GDP by 2050, i.e., the same level as in the OECD (25.5%). In Poland and Slovakia, for example, spending is scheduled to exceed 30% of GDP.258 Outside the transition economies direct spending on social insurance is much lower, but other forms of government spending (e.g., government employment) effectively fulfill a social insurance role. In Latin America, for example, direct government spending on social insurance is much lower than indirect spending through government employment and procurement programs.259 Moreover, it seems strange to argue that developing countries need tax revenues less than developed countries because they have less developed social insurance programs. If one accepts the normative case for social insurance, it applies to developing countries with even See International Considerations, supra note 22, at 3. This excludes Central and Eastern Europe where the ratios are as high as in Western Europe. 258 ILO, supra note 230, at Table 3.2. 259 See K. SUBBARAO ET AL., SAFETY NET PROGRAMS AND POVERTY REDUCTION: LESSONS FROM CROSSCOUNTRY EXPERIENCE , ch. 8 (1997). 257

59

greater force because of widespread poverty which means that losing a job can have much direr consequences.260 But the need for revenues in developing countries goes far beyond social insurance. In some developing countries revenues are needed to insure the very survival of organized government as the Russian experience demonstrates. In other, more stable developing countries, revenues are needed primarily to provide for adequate education (investment in human capital) which many regard as the key to promoting development.261 One could argue, however, that in many developing countries government inefficiency and corruption are widespread and that many developing country governments are atrocious dictatorships, so that it seems misguided to argue that they deserve more revenues. While this is undoubtedly true, it is a far less accurate description of most third world governments now than twenty years ago. Democracy is more prevalent now than in 1980, especially in Latin America and East Asia but even in Sub-Saharan Africa; and corruption is declining (in part due to recent OECD efforts to combat it and to constraints imposed by globalization itself). Moreover, this argument seems more relevant to direct aid to developing countries than to designing tax rules. While aid can be targeted at those countries whose governments seem able to use it best, it is doubtful whether tax rules can be tailored in the same way.262 It seems inappropriate to refrain from designing general tax rules that enable all developing countries to raise needed revenues merely because in some cases those revenues may be used for inappropriate ends, like lining some official's pockets or enabling a dictator to wage war more effectively. In fact, insofar as the tax rules are directed to foreign investment, those investments are more likely over time to flow to countries with less corruption and more democratic governments.263 (ii) Should Developing Countries Tax Foreign Investors? A more powerful argument in favor of allowing developing countries to engage in tax competition is the one made by economists who argue that in general small open economies should refrain from taxing foreign investors.264 The argument is as follows: If an economy is small, this means that it has to accept worldwide prices (including interest rates) as given. If it is open, this means that foreign investors can choose to invest in it or elsewhere. Since foreign 260

The UN has estimated that access to basic social services for all people in the world could be ensured for a mere $30-$40 billion a year. Two-thirds of this amount is to be raised by developing countries who are to increase spending on social insurance to about 20% of GDP. See W ORLD SUMMIT FOR SOCIAL DEVELOPMENT , supra note 97. 261 See, e.g., OECD, HUMAN CAPITAL INVESTMENT : A N INTERNATIONAL COMPARISON (1998); ECONOMIC ST AND SOCIAL DEVELOPMENT INTO THE 21 CENTURY 3-40 (Louis Emmerij ed., 1997). The recent UN summits have set a goal of complete access to primary education by boys and girls by 2015. 262 The U.S. experience in attempting to tailor tax rules to practices of developing countries (e.g., supporting terrorism) is not encouraging, at least if the purpose was to deter the behavior. See James R. Hines, Jr., Tax Avoidance: American Participation in Unsanctioned International Boycotts (NBER working article no 6116, 1997). See Miller, Third World Views of Ends and Means of U.S. Tax Policy, in UNITED STATES TAXATION AND DEVELOPING COUNTRIES 83, 91-102 (Robert Hellawell ed., 1980). 263 DANI RODRIK, THE NEW GLOBAL ECONOMY AND THE DEVELOPING COUNTRIES: M AKING OPENNESS W ORK 96-99 (1999). 264 See International Tax Competition, supra note 124.

60

investors will not accept an after-tax yield that is lower than the yield available elsewhere (which, by definition, the small economy cannot affect), they cannot be made to bear the burden of any tax imposed by the capital importing country. Therefore, the tax will necessarily be shifted to less mobile factors in the host country, such as labor and/or land, and it is more efficient to tax those factors directly. Thus, it is argued, a small open economy should only tax domestic labor and land, and (if it can) domestic capital invested either domestically or overseas, but not foreign capital. This argument seems quite valid as applied to portfolio investment which can earn the worldwide rate of interest free of tax in many locations around the world. Thus, the only effective way to tax such portfolio investment is on a residence basis, although as an administrative matter coordinated source-based backup withholding may be needed to enforce the tax. But in regard to FDI the argument seems less valid for two reasons. First, a precondition for the argument to apply is that a foreign tax credit not be available for the tax. If a foreign tax credit is available in the home country of the investor, the investor is unaffected by the host country tax (since it would have to pay it to either the home or host jurisdiction). Thus, if host country refrains from taxation it is merely giving away revenue to the home country. This situation is relatively rare for portfolio investment because (as argued above) home countries are rarely able to tax such investment, and therefore the credit is irrelevant.265 However, in the case of direct investment the credit is available in many cases, and therefore host countries should not refrain from taxing the investor.266 The second argument is even more powerful because it applies even in the absence of a credit (i.e., where the home country grants an exemption or deferral). The analysis above assumes that the host country is small and one of its assumptions is that the investing MNE can earn similar returns elsewhere. This, in turn, assumes a competitive market situation. However, an extensive literature on MNEs suggests that typically they exist in order to earn economic rents, i.e., super-normal returns that result from an imperfectly competitive market.267 In a perfectly competitive market, it typically makes more sense to operate through independent distributors given the transaction costs involved in operating directly in a foreign markets. However, where there are rents involved, it may make sense to engage in FDI to keep those rents to oneself. Except in the case of banks in which case host country taxation makes sense. See, e.g., the Mexican-U.S. tax treaty in which the withholding tax rate on interest (4.9%) is precisely geared to the amount available for credit to U.S. banks (because a 5% or greater tax would be in a separate basket). 266 See TIMO VIHERKENTTA, TAX INCENTIVES IN DEVELOPING COUNTRIES AND INTERNATIONAL TAXATION (1991). This raises the question why home countries grant the credit which has puzzled economists. See Bond & Samuelson, supra note 138. The answer lies in the history of the credit which was heavily influenced by the MNEs and the wish to avoid double taxation on a reciprocal basis. See Michael J. Graetz & Michael M. O’Hear, The “Original Intent” of U.S. International Taxation, 46 DUKE L.J. 1021 (1997). In any case, now that we already have a credit system that advances world welfare, it seems strange to argue in favor of abandoning it for a system that at best only enhances national welfare. 267 See THE NATURE OF THE TRANSNATIONAL FIRM , (Christos N. Pitelis & Roger Sugden eds., 1991). 265

61

If an investor earns rents in a country, it is no longer "small" in the economic sense. That is, there is a reason for the investor to be there and not elsewhere. Therefore, any tax imposed on such rents (as long as it is below 100%) will not necessarily drive the investor to leave even if it is unable to shift the burden of the tax to labor or landowners. Thus, the standard advice to host countries (refrain from taxing the foreign investor) does not apply. This argument clearly holds in the case of rents that are linked to a specific location such as natural resources or a large market. But what if the rent can be earned in a large number of potential locations? For example, assume the rent is a result of possessing a unique intangible (an "ownership" advantage, to use the terminology developed by Dunning).268 Such an intangible can be a patent or other forms of know-how (such as a manufacturing technique that reduces costs drastically). In order for the rent to be earned, the MNE must invest somewhere but there are many available locations. In this case, the host country will not be able to tax the rent if the MNE can credibly threaten to go elsewhere, although once the investment has been made the rent can be taxed. This situation, which is probably the most common,269 would require coordinated action to enable all host countries to tax the rent earned within their borders. (iii) Do Developing Countries Need to Offer Tax Incentives for Investment? The standard advice given by international institutions like the World Bank and the IMF to developing countries is to refrain from offering tax incentives to foreign investors. Thus, for example, the IMF's Tax Policy Handbook states that "[t]he IMF maintains a widely held view that tax incentives of all sorts have proved to be largely ineffective, while causing serious distortions and inequities in corporate taxation."270 Thus, for example, an IMF mission to Indonesia in the 1980's successfully advocated the abolition of all tax incentives given to foreigners. Similarly, a World Bank study of Central and Eastern Europe recommended eliminating all tax incentives and instead levying a low corporate tax rate on all activities.271 In attracting foreign investors, it is argued, the country's overall business climate (stable government, well-educated workforce, good infrastructure) are much more important than specific tax incentives.272

See JOHN H. DUNNING, EXPLAINING INTERNATIONAL PRODUCTION (1988). See THE NATURE OF THE TRANSNATIONAL FIRM , supra note 267. 270 Janet Stotsky, Summary of IMF Tax Policy Advice, in TAX POLICY HANDBOOK 279, supra note 175, at 282. 271 See JACK M. M INTZ, CORPORATE INCOME TAXATION AND FOREIGN DIRECT INVESTMENT IN CENTRAL AND EASTERN EUROPE, 10 (Jack M. Mintz & Thomas Tsiopoulos eds., 1992); OECD, TAXATION AND FOREIGN DIRECT INVESTMENT 30 (1995) ("In most econometric studies on the question, the foregone tax revenue [from tax incentives] has exceeded the increase in the desired investment."). 272 See OECD, PROMOTING FOREIGN DIRECT INVESTMENT IN DEVELOPING COUNTRIES 53 (1993) ("Do not build a programme around tax holidays."). See also Jack Mintz, Corporate Tax Holidays and Investment, 4 W ORLD BANK ECON. REV. 81 (1990) (arguing that tax holidays are ineffective in many cases). 268 269

62

This view is also widely shared by academics, who argue that there is little evidence of the effectiveness of tax incentives in attracting foreign investment. For example, Peter Enrich has argued in the U.S. state tax context that "for as long as states have been racing to outdo each other in the tax breaks they offer, economists and other researchers have been searching for evidence of the influence of states' tax policies on the vitality of their economies. The conclusion from these efforts has been, at best, inconclusive… All the evidence points to a single conclusion: State tax incentives are a thoroughly unproven tool for promoting economic development."273 If one could accept this view, then it would be relatively easy to argue (as Enrich does) that constraining the ability of countries to engage in tax competition does them no harm. But in fact, the evidence for the impact of taxation on business location incentives is stronger than Enrich suggests. Even in the state context, more recent studies with better methodology have tended to conclude that taxes do play an important role in business location decisions.274 In the international context, where tax rates are much higher, the evidence for the impact of taxation is decisive. As Hines and others have shown, taxes play a crucial role in determining investment location decisions, at least when comparing otherwise equivalent locations.275 Moreover, it is hard to argue that in some cases tax incentives may have played an important role in improving the welfare of a nation's citizens. The most frequently cited case is Ireland, which was an economic basket case in the early 1980s but now has the fastest growing GDP in Europe with income per capita exceeding the U.K. This growth was the result of massive foreign investment, aided by factors like an educated, English speaking work-force, and an elaborate system of targeted tax holidays.276 Thus, how can it be argued that developing countries should not grant tax holidays to foreign investors even if they generally need the revenues if doing so is the only way to attract mobile business? Developing countries can certainly argue that they need the investment, and that the investment would not happen without the tax incentive.277 For example, a recent case

Peter D. Enrich, Saving the States from Themselves: Commerce Clause Constraints on State Tax Incentives for Business, 110 HARV. L. REV. 377, 389-392 (1996). 274 See JOHN D. DONAHUE, DISUNITED STATES ch. 6 and App. A, 171-182 (1997) (reaching opposite conclusions from Enrich although he supports the same policy conclusion). 275 See Lessons From Behavioral Responses to International Taxation, supra note 82, at 305-322; Eric Bond, Tax Holidays and Industry Behavior, 63 REV. OF ECON. AND STAT . 88 (Feb. 1981); Michael Bosking & William G. Gale, New Results on the Effects of Tax Policy on the International Location of Investment, in THE EFFECTS OF TAXATION ON CAPITAL A CCUMULATION 201 (Martin Feldstein ed., 1987). 276 See OECD, OECD ECONOMIC SURVEYS 1998-1999: IRELAND (1999); Rob Norton, The Luck of The Irish, FORTUNE, Oct. 25, 1999, at 194; Eve of an emerald era Ireland, FIN. TIMES (London), Sept. 22, 1998, at survey p.1. 277 In some cases, however, it can be questionable whether the investment is needed. After all, a dollar of investment is the same whether it comes from overseas or from domestic sources. See Dani Rodrik, The Global Fix, 219 NEW REPUBLIC 17-19 (Nov. 2, 1998). In some situations it may be better to promote domestic investment even if FDI is more glamorous. See Bernard Avishai, Israel's Future: Brainpower, 273

63

study has focused on Intel's decision to invest $300 million in Costa Rica. It is clear that Intel was attracted primarily by non-tax factors, such as political stability, a general commitment to economic openness, and an excellent educational system, and that it was not granted any specific tax breaks or other subsidies. But it is also clear that the investment would not have taken place without Cost Rica's general regime of tax holidays for foreign investors given the availability of such holidays in Costa Rica's competitors (principally Mexico).278 More generally, several studies of determinants of business location have concluded that the tax incentives are crucial given the availability of such incentives elsewhere. For example, the path-breaking study of Stephen Guisinger and his colleagues concluded that tax incentives are crucial because otherwise the host country could not compete. The definition of competitiveness used by Guisinger was "would a foreign investment project have located in a particular country if that country had eliminated its incentives and disincentives while other countries maintained incentive policies at existing levels."279 Using that definition, they found that the incentives were essential. However, the crucial assumption made by Guisinger and his colleagues is that such countries have "no opportunity for cooperative agreements limiting incentive policies."280 The solutions discussed below presuppose precisely such a multilateral mechanism. Thus, it can be argued that given the need for tax revenues developing countries would in general prefer to refrain from granting tax incentives if only they could be assured of remaining competitive. When they do grant tax incentives, it is precisely because of competition from other developing countries. Thus, Israel gave Intel $600 million under threat that it will locate its investment in Ireland, and Indonesia recently reintroduced its tax incentives under pressure of the Asian crisis of 1997-98 and the incentives offered by Malaysia and Singapore.281 Thus, restricting the ability of developing countries to compete in granting tax incentives does not truly restrict their autonomy or counter their interests.282 That is the case whenever they grant the incentive only for fear of competition from other developing countries and would not have granted it but for such fear. Whenever competition from other countries drives the tax incentive, eliminating the competition does not hurt the developing country and may aid its revenue raising efforts (assuming it can attract investment on other grounds). If, on the other hand, the tax reduction is not driven by competition then a country remains free to lower its generally applicable tax rates.

High Tech - and Peace, HARV. BUS. REV., Nov./Dec. 1991, at 50; Debra K. Rubin & Neal Sandler, Rebuilding a Nation, 240 ENGINEERING NEWS-RECORD 30 (1998). 278 See DEBORAH SPAR, A TTRACTING HIGH TECHNOLOGY INVESTMENT : INTEL 'S COSTA RICAN PLANT (1998). 279 GUISINGER ET AL., INVESTMENT INCENTIVES AND PERFORMANCE REQUIREMENTS 39 (1985). 280 Id. at 39. 281 See Robert Lenzner, Investing, Not Giving, FORBES, Dec. 18, 1995, at 106; Michael Sun, Exploding the Myth of the Free Ride, INVESTORS DIG., Nov. 16, 1996, at 15. 282 For a similar conclusion in the U.S. state context, see Enrich, supra note 273; DONAHUE, supra note 274.

64

The Irish case illustrates this distinction. Ireland was able to benefit from its tax incentives precisely because it was the only EU member to offer them. That is, Ireland had no real competition given the reality of tariff barriers and the high tax rates maintained by other EU members. But Ireland could have achieved the same result by maintaining a lower general tax rate and has now in fact done so as its preferential regimes have come under fire from the EU. In a typical tax competition scenario, on the other hand, a developing country is competing directly against the preferential regime of another developing country. In that case, lowering the tax rate generally is typically not an option because of the revenue needs of both countries. Competition limited to foreign investment is feasible, but it still hurts both countries in lost revenues especially if the incentives cancel each other out.283 In that case, and it would be in the interest of both to see it eliminated if a cooperative solution could be found. In summary, from the perspective of a typical developing country, the revenue loss from granting targeted tax holidays appears unlikely to be adequately compensated by the benefits flowing from the investment. The reason for this is that a typical developing country is not competing against every other country in the world for the location of investment. Rather, it is competing against a limited subset of countries with characteristics similar to itself. When a multinational enterprise decides to look for an investment location, it engages in a process of analysis reminiscent of Fernand Braudel’s three levels of historical change.284 It first eliminates countries that are unsuitable on the basis of very long-term factors, such as geography and climate. Then, it eliminates countries that are unsuitable because of shorter-range factors that are only to a limited extent under government control, such as labor costs, education levels, or political instability. Finally, the multinational conducts an auction among the remaining countries, which focuses on the factors under direct government control, such as tax holidays and other forms of subsidy. In that auction, the developing country is likely to find itself forced to offer subsidies merely because other countries offer them, and not because it would choose to do so on the basis of a cost/benefit analysis independent of tax competition. Two additional points need to be made from a developing country perspective. The first concerns the question of tax incidence. Since the tax competition that is most relevant to developing countries concerns the corporate income tax, it is important to attempt to assess the incidence of that tax in evaluating the effects of collecting it on the welfare of the developing country. Unfortunately, after decades of analysis, no consensus exists on the incidence of the corporate tax. While the older studies have tended to conclude that the tax is borne by shareholders or by all capital providers, more recent studies have suggested that the tax is borne 283

It is conceivable that the additional revenues generated by taxing employees or suppliers of the foreign investor more than make up the revenue lost in granting the preferential regime; but if the investment would have been made anyway for non-tax factors if there was no tax competition, those revenues could have been generated as well. In general, there is no evidence for direct investment being motivated solely or even primarily for tax reasons; rather, taxes are a factor in deciding among locations that enjoy similar other advantages. That is the sense in which taxes are said to be crucial in investment decisions. See Tax Policy, supra note 48; GUISINGER, supra note 279. 284 See FERNAND BRAUDEL, THE MEDITERRANEAN AND THE MEDITERRANEAN WORLD IN THE AGE OF PHILIP II ( tr. Sian Reynolds, 1992), introduction.

65

to a significant extent by consumers or by labor.285 Another possibility is that the tax on established corporations was borne by those who were shareholders at the time the tax was imposed or increased, because thereafter it is capitalized into the price of the shares.286 It is unlikely that this debate will be decided any time soon (in fact, the incidence may be shifting over time, especially as globalization may enable corporations to shift more of the tax burden to labor). However, from the perspective of a developing country deciding whether to collect taxes from a multinational, three out of the four possible alternatives for incidence (current shareholders or capital providers, old shareholders, and consumers) are largely the residents of other jurisdictions, and therefore from a national welfare perspective the developing country gains by collecting the tax. And even if some of the tax is shifted to labor in the developing country, it can be argued that as a matter of tax administration it is more efficient (as well as more politically acceptable) to collect the tax from the multinational than to attempt to collect it from the workers. Finally, it should be noted that a developing country may want to collect taxes from multinationals even if in general it believes that the private sector is more efficient is using the resources than the public sector. That is because in the case of a foreign multinational, the taxes that the developing country fails to collect may indeed be used by the private sector, but in another jurisdiction, and therefore not benefit the developing country. One possible solution, which is in fact employed by developing countries, is to refrain from taxing multinationals while they re-invest domestically, but tax them upon remittance of the profits abroad. However, such taxation of dividends and other forms of remittance is subject to the same tax competition problem that we discussed above. Thus, it would appear that overcoming the tax competition problem is in most cases in the interest of developing countries, and the question remains how to do so in the face of the collective action problem described above. c. Inter-Nation Equity and the Division of Global Tax Revenues Having established that both developed and developing countries need tax revenues, is there some way of deciding how to divide the available revenue between them? This question relates to the concept of inter-nation equity which was first developed by Peggy Musgrave.287 As Musgrave used it, the concept related primarily to the question of entitlement, i.e., which country was entitled to the revenue, and was used to defend the prevalent practice of sourcebased taxation against the preference for residence-based taxation that goes with capital export neutrality.288 See, e.g., JOSEPH PECHMAN, FEDERAL TAX POLICY 141-148 (5TH ed. 1987) and the sources cited therein. The 1992 Treasury report on integration concluded that the tax falls on shareholders in the short run but on all capital or partly on capital and partly on labor in the long run. U.S. Department of the Treasury, Integration of the Individual and Corporate Tax Systems: Taxing Business Income Once (1992), at 326. 286 See Pechman, id. 287 See International Tax Base Division, supra note 169; Inter-nation Equity, supra note 169, at 64; Kaufman, supra note 169. 288 See Inter-nation Equity, supra note 169. 285

66

The problem with this formulation of inter-nation equity is that it is vague and does not easily lead to practical recommendations. Even a preference for source- over residence-based taxation is not very meaningful unless one can develop a coherent notion of the source of income. However, most income has multiple sources, and economists have generally concluded that assigning income to a single source is not a meaningful exercise (albeit a necessary one for legal and tax purposes).289 If that is the case, then any generally acceptable sourcing rule is fine as an economic matter, but a concept of inter-nation equity that is based on sourcing does not give us any guidance in designing such a rule. That is, inter-nation equity does not really tell us how much income a country is "entitled" to tax. Perhaps some progress can be made if one attempts to relate inter-nation equity to its more familiar cousin, inter-individual equity.290 It is widely accepted that redistributive income taxation can be justified by considerations of vertical equity and the declining marginal utility of income.291 But there appears to be no sound theoretical reason to restrict redistribution to members of any single tax jurisdiction. If there were a world taxing authority, it would be justified and required to redistribute on a worldwide basis. Given that there are many taxing jurisdictions, redistribution in practice takes place within countries, and this can be justified on political grounds ("no taxation without representation"). Explicit redistribution among countries is rare especially given the widespread dissatisfaction with foreign aid. But this analysis suggests that in designing international tax rules, it should be a goal to take into account the relative wealth and poverty of the countries involved. In federal states, it is common to have an explicit redistributive mechanism from the richer to the poorer provinces.292 If there were a world federal taxing authority but many countries, a similar mechanism could be justified to redistribute revenues from richer to poorer countries. This could be done by direct transfers but also indirectly through tax rates. For example, whenever a transaction took place between two jurisdictions, the poorer one could explicitly be allowed to retain a larger share of the revenue (e.g., in proportion to GDP).293 In practice there is no world taxing authority, nor is one likely to be established in the near future.294 But some of the current practice of international taxation can be interpreted as reflecting concern for the relative revenue needs of poorer countries. In particular, the widespread acceptance of the right of the source country to levy its tax first and of the See Ault & Bradford, in TAXATION IN THE GLOBAL ECONOMY, (Assaf Razin and Joel Slemrod eds., 1990), at 31. 290 On the distinction, see Kaufman, supra note 169. 291 See Kaplow, supra note 168. For a recent exposition of the utilitarian case for progressive taxation, see Martin J. McMahon Jr. & Alice G. Abreu, Winner-Take-All Markets: Easing the Case for Progressive Taxation, 4 FLA. TAX REV. 1 (1998). 292 See Brian J. Arnold, Canada, in COMPARATIVE INCOME TAXATION: A STRUCTURAL A NALYSIS 25 (1997); Dr. Albert J. Radler, Germany, in COMPARATIVE INCOME TAXATION 49. 293 See PEGGY RICHMAN (Now Musgrave), TAXATION OF FOREIGN INVESTMENT INCOME—A N ECONOMIC A NALYSIS (1963). 294 See Part V below. 289

67

imposition of the burden of alleviating double taxation on the residence country reflects in part the position of the poorer (capital importing) countries in the 1920s. This position was accepted by the League of Nations and (subject to certain limits such as the permanent establishment concept) enshrined in the model treaties which are the foundation of the current international tax regime.295 Thus, I would argue that the concept of inter-nation equity can be given practical meaning if it is interpreted as embodying explicit distributive goals in designing international tax rules. More specifically, when a choice is presented between two alternative rules, one of which has progressive and the other regressive implications for the division of the international tax base among poorer and richer countries, the progressive rule should be explicitly preferred to the regressive one. In the absence of a world tax authority able to redistribute tax revenues, and given the paucity of direct international transfers from developed to developing countries, such a concept of inter-nation equity has the best chance of achieving meaningful distributive goals. The problem of tax competition, as described in Part II above, is a good illustration of such a dilemma. The problem breaks down into two components: portfolio and direct investment. For portfolio investment, the home country of the invested capital may be either a developed or a developing country while the host country is typically a developed country. For direct investment, the home country of the invested capital is typically a developed country while the host country may be either a developed or a developing country. 296 Tax competition, as described in Part II, is undermining the ability of both home and host jurisdictions to tax income from cross-border portfolio and direct investments. In each case, it is possible to design solutions that emphasize either home (residence) or host (source) taxation. For portfolio investment, for example, one could have either a final withholding tax which would accrue to the host jurisdiction or a refundable withholding tax which would be waived upon proof that the income was reported to the home jurisdiction. The first solution would favor host countries, the second home countries.297 As the home countries are on average poorer than the host countries, inter-nation equity would suggest preferring the second solution over the first one. Similarly, for direct investment, it is possible to design a regime that allocates corporate tax revenues entirely to the home (residence) country. For example, corporations could be assigned a residence based on the residence of their shareholders and the corporate tax See Graetz & O'Hear, supra note 266, Reuven S. Avi-Yonah, The Structure of International Taxation: A Proposal for Simplification, 74 TEX. L. REV. 1301 (1996); International Taxation of Electronic Commerce, supra note 34. 296 See W ORLD INVESTMENT REPORT , supra note 98. 297 The EU, as described below, has opted for a two-track method in which countries could choose between exchange of information which favors home countries and a refundable withholding tax which likewise favors home countries but to a lesser extent (some investors would prefer to forego the refund if the rate is lower than the home country rate). The OECD prefers pure exchange of information. 295

68

revenues allocated to those countries.298 Or a form of corporate-shareholder integration could be implemented that eliminates the corporate tax base and allocates all revenues to the home country of the shareholders.299 Alternatively, the corporate tax could be abolished and shareholders taxed on their holdings on a mark-to-market basis.300 All of these solutions favor home over host country taxation. On the other hand, solutions to tax competition could also be developed that seek to maintain the existing preference for host country (source) taxation of direct investment. The solution advocated in Part V below is one such solution. From an inter-nation equity perspective, this solution is to be preferred precisely because in the direct investment context host jurisdictions are poorer than home jurisdictions.301 If one accepts the need for redistribution in general, taking such distributive considerations into account explicitly in the international tax context would be a significant step forward since almost all of the current literature on international taxation omits any discussion of such issues.302

See Richard L. Doernberg, Electronic Commerce and International Tax Sharing, 98 W ORLDWIDE TAX DAILY 60-43 (Mar. 30, 1998); Gerald W. Padwe, AICPA Submits Exposure Draft On Integration Of Corporate And Shareholder Tax Systems, 93 TAX NOTES INT ’L 6-13 (Jan. 11, 1993). 299 See Alberto Giovannini & James R. Hines Jr., Capital Flight and Tax Competition: Are There Viable Solutions to Both Problems?, in EUROPEAN FINANCIAL INTEGRATION 172-210 (A. Giovannini & C. Mayer eds., 1991). 300 See Joseph M. Dodge, A Combined Mark-to-Market and Pass-Through Corporate-Shareholder Integration Proposal, 50 TAX L.REV. 265-372 (1995); Daniel I. Halperin, Will Integration Increase Efficiency?--The Old and New View of Dividend Policy, 47 TAX L.R. 645-651 (1992). 301 This is particularly true given the relative importance of the corporate tax to developing countries compared with its relative insignificance for most developed countries. See above. 302 But see Chang Hee Lee, Impact of E-Commerce on Allocation of Tax Revenue Between Developed and Developing Countries, 18 TAX NOTES INT 'L 2569 (June 21, 1999). 298

69

V.

Solutions303 a. EU Proposals and Critique

On December 1, 1997, the Council of the European Union (EU) adopted a “package to tackle harmful tax competition in the European Union.”304 The package was based on the work of the Commission of the EU, and in particular the Tax Working Group headed by Fiscal Affairs Commissioner Mario Monti. The package was ultimately embodied in three measures: (1) A non-binding “code of conduct” on business taxation;305 (2) a draft directive on taxation of savings;306 and (3) a draft directive on taxation of cross-border interest and royalty payments.307 The code of conduct represents a political commitment that is not legally binding but was endorsed by the ministers of finance of the member states. The draft directives require adoption by unanimous consent under Article 100 of the Treaty of Rome. (i) The Code of Conduct. The code of conduct represents in some ways the most interesting of the three proposals and has had a significant impact both on the actions of member states and on the related work of the OECD. 308 The code is a “political commitment,” not a legally binding document, but it has the support of all member states at least on paper. Whether that support will be translated into action remains to be seen. The code addresses business tax measures “which affect, or may affect, in a significant way the location of business activity in the Community.” “Business activities” are defined broadly as including all activities carried out within a company or a group of companies. The tax measures affected include any measure applied by law, regulation, and administrative practice, whether through the nominal tax rate, the tax base, or “any other relevant factor.”309 In

303

The discussion below does not consider the possibility that the mobility of capital could or should be restricted directly. Most economists would reject such proposals, and nobody advocates them for longerterm investments. For the development of the OECD's view, see OECD, CONTROLS ON INTERNATIONAL CAPITAL M OVEMENTS (1982); OECD, EXCHANGE CONTROL POLICY (1993); OECD, OPEN M ARKETS M ATTER: THE BENEFITS OF TRADE AND INVESTMENT LIBERALIZATION (1998). But see Bo Sandermann Rasmussen, International Tax Competition, Tax Cooperation and Capital Controls (1997) (working paper, Aarhus Universitet, Dept. of Econ.) (arguing that even though capital controls are inferior to cooperation, the threat of imposing controls may be useful in persuading countries to cooperate in tax matters). 304 Conclusions of the ECOFIN Council Meeting on 1 December 1997 Concerning Taxation Policy, 98/C 2/01, OJ 1/6/98 [hereinafter Conclusions of the ECOFIN Council]. 305 See id. at Annex 1. 306 See COM (98) 295 final, 98/0193 (CNS), 5/20/98. 307 See COM (98) 67 final, 98/0087 (CNS), 3/6/98. 308 For the OECD proposals see section (b) below. 309 Code of Conduct, supra note 305, at C 2/3.

70

addition, the ministers bound themselves to re-examine subsidies delivered outside the tax system.310 The code covers all tax measures “which provide for a significantly lower effective level of taxation, including zero taxation, than those levels which generally apply in the Member State in question.”311 Such measures are considered “potentially harmful.” In determining whether they are harmful, several other factors are considered including: a. Whether the tax measures apply only to non-residents; b. Whether the measures are “ring-fenced,” i.e., segregated from the domestic market to prevent erosion of the domestic tax base; c. Whether advantages are granted without any real economic activity or substantial economic presence in the Member State;312 d. Whether the State follows the OECD transfer pricing guidelines (to prevent artificial allocation of profits to the activity benefiting from the measure); e. Whether the measure lacks transparency (e.g., is granted by covert administrative action).313 The code envisions two forms of enforcement. First, member states will engage in a self-review process and commit themselves not to introduce new tax measures that are deemed harmful under these criteria (“standstill”) and to eliminate existing measures within a 2-5 year period (“rollback”).314 Second, member states will inform each other about their preferential tax regimes and any member state may request an opportunity to discuss and comment on the tax measures of other member states. For this purpose, a review group will be formed which will assess those measures and forward a report to the Council which may publish it.315 Thus, the idea is to pressure member states that have established either production tax havens (such as Ireland) or headquarters tax havens (such as Belgium) to abolish them by exerting political pressure through the review process. In fact, soon after the promulgation of the code Ireland announced that it was phasing out its preferential tax regimes and replacing

See Conclusions of the ECOFIN Council, supra note 305; Code of Conduct, supra note 305, at J; Nigel Tutt, European Commission to Propose Guidelines on Providing State Aid Through Tax System, 98 TAX NOTES INT ’L 135-3 (July 15, 1998). The Treaty of Rome, arts. 92-94, limits state aids. For the particular problem of regional development programs, see Code of Conduct, supra note 305, at G. 311 Code of Conduct, supra note 305, at A. 312 Germany took the position that this applies to the targeted granting of advantages for international mobile activities where they are not granted for non-mobile activities. See Conclusions of the ECOFIN Council Meeting, supra note 305, at C 2/2. 313 See Code of Conduct, supra note 305, at B 1-5. 314 See Code of Conduct, supra note 305, at C, D, Conclusions, C 2/1-2. The Council may approve a longer period in some circumstances. 315 See Code of Conduct, supra note 305, at E-I. For the formation of the Code of Conduct Group, see Report of the 2072nd Council meeting (Brussels, 3/9/98), 6619/98, C/98/61. 310

71

them with a single, lower rate of corporation tax applicable to all companies (whether domestic or foreign owned).316 The code also requires member states to “promote” the adoption of similar principles in other countries and, in particular, to apply them to “dependent or associated territories,” such as the Channel Islands.317 The U.K. has in fact begun to pressure its territories to adopt the principles of the code.318 (ii) Taxation of Savings. The EU Council in the same meeting (chaired, ironically, by Luxembourg) adopted a resolution on taxation of savings which would apply a new regime to interest income paid to individual residents of the EU by paying agents located in another member state.319 The Commission in May 1998 proposed a draft Directive on this topic for adoption by the Council.320 Under Article 100 of the Treaty of Rome, the Directive has to be adopted by unanimous consent. The draft Directive proposes a “coexistence” model based on two options, only one of which can be chosen by a member state: Either to cooperate in an exchange of information program or to levy a 20% withholding tax on interest payments made by paying agents within its territory to individual residents of another member state.321 Under the exchange of information system, the member state agrees to provide automatically, at least once a year, information on all interest payments made by paying agents in its territory in the preceding year to individual beneficial owners residing in every other member state.322 Under the withholding tax system, the member state agrees to levy a 20% withholding tax on all interest payments made by paying agents within its territory to individual beneficial owners residing within the EU. However, the withholding tax is not levied if the beneficial owner provides a certificate drawn up by his country’s tax authorities attesting that they have been informed of the interest to be received.323 The withholding tax must be credited against tax liability in the beneficial owner’s country of residence.324

See Edward Troup & Paul Hale, EU Initiatives on Tax Harmonization: Do as I Say, Not as I Do?, 98 TAX NOTES INT ’L 197-3 (Oct. 13, 1998). 317 Code of Conduct, supra note 305, at M. 318 See Jim Kelly, Relegation Threat Looms, FIN. TIMES (London), July 15, 1999, at 2. 319 See Annex 2, Taxation of Savings, supra note 306, at C 2/6. 320 COM (98) 295 final, 98/0193 (CNS), 5/20/98. 321 Draft Directive, supra note 320, at Art. 2. 322 See id. at Art. 7. 323 See id. at Art. 8. 324 See id. at Draft Directive, Art. 10. If the withholding tax exceeds tax liability a refund is provided. 316

72

This draft Directive has encountered significant opposition. Industry groups have argued that it will result in an outflow of savings outside the EU, e.g., to Switzerland.325 More significantly, both Luxembourg and the U.K. have opposed the Directive as likely to have an adverse impact on the Eurobond market.326 Luxembourg announced that it would not support the Directive unless the EU’s main trading partners accept parts of the plan.327 (iii) Taxation of Interest and Royalties. This proposal is the least controversial element of the package. The draft Directive calls for the abolition of withholding taxes on interest and royalty payments made within the EU between “associated companies,” defined by a 25% ownership threshold.328 However, the reduction of tax does not apply where the interest or royalty payment is subject to a corporate tax rate in the country of residence that is lower than the generally applicable rate or benefits from a preferential reduction in the tax base.329 (iv) Critique. A fundamental problem of the EU tax competition package (as well as the OECD measures discussed below) is that they do not contain a full-fledged normative grounding for the measures they propose to adopt. The Council resolutions mentions three reasons for tackling harmful tax competition: preventing distortions in the single market, preventing “excessive” losses of tax revenue, and “getting tax structures to develop in a more employment-friendly way.”330 This last ground is further explained as relating to the increasing tax burden on labor vis-à-vis capital, resulting in increased taxes on employment and concurrent loss of jobs.331 Moreover, the transition to the EMU is cited as increasing the mobility of capital within the EU.332 However, none of these reasons is developed in any detail leaving the package open to criticism as opportunistic and unbalanced.333 See Impact Assessment Form attached to the Draft Directive, supra note 320 (statements of the European Mortgage Federation and the Banking Federation of the European Union); European Finance Groups Respond to ECOFIN Cross-Border Savings Directive, 98 TAX NOTES INT ’L 178-H (Sept. 15, 1998) (International Securities Markets Association and International Primary Markets Association). IPMA estimated that about 5% of existing Eurobond issues might become callable if the plan were adopted. See Troup & Hale, supra note 316. 326 See Council Conclusions, supra note 305, at C 2/3 (U.K. position); European Union Finance Ministers to Debate Tax Savings Plan, 98 TAX NOTES INT ’L 183-8 (Sept. 22, 1998); U.K. Chancellor Renews Attack on EC’s Eurobond Tax Plan, 98 TAX NOTES INT ’L 186-6 (Sept. 25, 1998). 327 Luxembourg Says No Agreement Reached on EU Savings Tax, 98 TAX NOTES INT ’L 188-7 (Sept. 29, 1998). 328 COM (98) 67 final, 98/0087 (CNS), 98/C 123/07 (Mar. 6, 1998), Art 1. 329 See id. Art. 7. This article is to be reviewed in three years. 330 Council Conclusions, supra note 305, at C 2/1. 331 See Draft Directive, supra note 320, at 3; European Parliament, Resolution on the Commission Communication, A Package to Tackle Harmful Tax Competition in the European Union, A4-0228/98 (June 18, 1998). 332 See Draft Directive, supra note 320, at 3. 333 See Troup & Hale, supra note 316. 325

73

In light of the normative argument developed in Part III above, the EU package seems quite reasonable. The package addresses the problems of inefficiency, inequity, and tax base erosion in a way that is adequately balanced with concern for the ability of member states to define the desirable size of their public sector. In particular, the distinction between generally applicable tax reductions and measures aimed only at foreigners is maintained, and has proven effective, as the Irish example indicates. In addition, the savings Directive is appropriately aimed at insuring taxation in the country of residence of the saver. The EU package does, however, suffer from a major limitation: It only applies within the EU. Thus, as far as the preferential regimes addressed in the code of conduct are concerned, only regimes in EU member states are affected. However, as described in Part II, both production tax havens and headquarters tax havens proliferate outside the EU. The code does envision measures aimed at “promoting” its principles in third countries but outside the dependent territories has little impact.334 This does not mean, however, that it has no impact. Given the advantages of locating within the EU to trading with EU member countries, the code of conduct has significant bite because production or headquarters tax havens located outside the EU will not benefit from the same advantages as those in Ireland or Belgium.335 Nevertheless, from a global perspective the reach of the code is quite limited, and given the reduction in transportation costs, it seems likely that multinationals will be able to locate their activities in production tax havens outside the EU. A more significant problem faces the draft Directive on savings which may fail to be adopted because of the fear of driving capital to Switzerland or the U.S. The draft Directive attempts to limit this problem by focusing only on payments to EU residents (so that EU countries can still attract capital from other countries without having to impose a withholding tax). This reduces its attractiveness from a global perspective. In addition, the focus on paying agents rather than debtors means that U.S. issuers raising capital in Europe will to some extent not have a direct competitive advantage over European issuers.336 Finally, the directive envisions the EU taking measures “to promote the adoption in third countries of equivalent measures relating to payments of interest to Community residents.”337 However, as the reaction of Luxembourg makes clear, requiring the imposition of a 20% withholding tax does risk shifting savings of EU residents (e.g., German residents) to non-EU countries (such as Switzerland). In addition, even if the Directive becomes effective, it would only alleviate and not solve the problem of under-taxation of interest income. A 20% rate is far from sufficient to See Code of Conduct, supra note 305, at M. It may, however, encourage companies to locate in countries like Israel which has a free trade agreement with the EU but is not a member. 336 See Draft Directive, supra note 320, at 4. The Directive also notes that the Eurobond market is to a large extent composed of institutional investors and non-EU residents who would not be affected by the proposal. 337 Draft Directive, supra note 320, at 20. 334 335

74

compensate for non-taxation in the home country (at rates of 50% or higher).338 However, the EU felt that a higher rate would be more likely to drive capital out of the Community.339 In this way, the proposal resembles the Nordic “dual tax” system, in which capital is taxed at a much lower rate than labor as a way to ensure that it is at least subject to some tax. 340 It is a start, but an insufficient one. The obvious solution is to expand the scope of the EU proposals to non-member countries. Some ways to do so are discussed in Part V(c). b. OECD Proposals and Critique (i) The OECD Report. The OECD report on “Harmful Tax Competition: An Emerging Global Issue” was approved by the OECD Council on April 9, 1998, with two abstentions (Luxembourg and Switzerland).341 It was written by a “Special Sessions on Tax Competition” chaired by France and Japan under the auspices of the Committee on Fiscal Affairs. The introduction explains the goals of the Report as: [T]o develop a better understanding of how tax havens and harmful preferential tax regimes, collectively referred to as harmful tax practices, affect the location of financial and other service activities, erode the tax bases of other countries, distort trade and investment patterns and undermine the fairness, neutrality and broad acceptance of tax systems generally.342 The OECD Report is thus much narrower than the EU package. It does not address the taxation of cross-border interest flows; that is to be the topic of a different report that was scheduled to be issued in 1999 by a Working Party on Tax Evasion and Avoidance.343 In addition, it is focused entirely on “geographically mobile activities, such as financial and other service activities, including the provision of intangibles.”344 The Report explains that the issue of tax competition for “real” investments will be examined separately in the future but does not commit itself to a date.345 Hugh Ault, who served as academic consultant to the OECD and 338

This assumes that there are no significant deductions associated with earning the interest income which seems realistic for individual taxpayers. 339 See Draft Directive, supra note 320, at 3. The French argued for a 25% rate. See Conclusions, supra note 305, at C 2/2. 340 See Timo Viherkentta, A Flat Rate Tax on Capital Income: The Nordic Model, 93 TAX NOTES INT 'L 4911 (Mar. 15, 1993). 341 See OECD REPORT , supra note 23, at 65. The OECD is an organization of 29 countries which are mostly developed countries, but it includes some developing countries (e.g., Mexico and South Korea) and some transition economies (Czech Republic, Hungary, Poland). 342 Id. at 8. 343 See id. at 10. 344 Id. at 8. 345 See id. at 8.

75

participated in drafting the Report, has written that this restriction was made “in the interest of creating a manageable work plan.”346 On the other hand, the OECD Report because of its provenance applies to a much broader geographical area than the EU package. In addition to twenty-seven of the twenty-nine members of the OECD, the Report envisages that the OECD will attempt to persuade nonmember countries to abide by its recommendations. Three regional seminars with non-member countries have been held, and the Report proposed that a high-level meeting open to nonmember countries be organized in 1999.347 In addition, the OECD Report puts much greater emphasis on traditional tax havens than the EU package. The OECD Report is divided into three chapters. Chapter 1 provides an overview of the ways in which the process of globalization has put pressures on current international tax arrangements. Chapter 2 attempts to distinguish between harmful and acceptable tax competition and lists factors that are to be used to identify harmful tax practices. Chapter 3 includes nineteen specific recommendations for countering harmful tax competition as well as Guidelines on Harmful Preferential Tax Regimes, a procedure to identify tax havens, and proposes the creation of a Forum on Harmful Tax Practices.348 Chapter 1 attempts to set out a normative basis for the Report. Economic distortions are barely mentioned and CEN is not mentioned at all. Instead, the focus is on tax base erosion and the change in the tax mix: These schemes can erode national tax bases of other countries, may alter the structure of taxation (by shifting part of the tax burden from mobile to relatively immobile factors and from income to consumption) and may hamper the application of progressive tax rates and the achievement of redistributive goals.349 However, none of these points is set out in any detail (the whole chapter is only 5.5 pages long). In particular, no evidence is adduced to support the assertion that the tax base of OECD members is eroding (“[t]he available data do not permit a detailed comparative analysis of the economic and revenue effects” of tax competition).350 The Report does attempt to Joann M. Weiner & Hugh J. Ault, The OECD’s Report on Harmful Tax Competition, 51 NAT ’L TAX J. 601 (1998), 602. 347 See OECD REPORT , supra note 23, at 10. 348 See id. at 11-12. 349 Id. at 14. See also id. at 37 (“Governments cannot stand back while their tax bases are eroded through the actions of countries which offer taxpayers ways to exploit tax havens and preferential regimes to reduce the tax that would otherwise be payable to them.”) This emphasis is unfortunate given the paucity of the evidence and the impression given of government bureaucrats trying to protect their cherished tax revenues. It certainly does nothing to combat the Leviathan critique. 350 Id. at 17. The Report does note that FDI by G7 countries in the Caribbean and in the South Pacific islands increased more than five-fold to over $200 billion in 1985-1994, a rate well in excess of the growth of total outbound FDI. See id. at 17. 346

76

distinguish between regimes that are intended to “poach” the tax base of other countries and regimes that “reflect different judgments about the appropriate level of taxes and public outlays or the appropriate mix of taxes in a particular economy.”351 But there is no systematic effort to distinguish on a normative basis between harmful and beneficial tax competition, such as attempted in Part III above. The line drawn by the OECD can, however, be deduced from the factors listed in Chapter 2 for identifying tax havens and harmful preferential tax regimes. The main distinction drawn is between, on the one hand, tax havens (with no or nominal income tax) and preferential regimes in countries that have generally applicable taxes, and on the other hand, generally applicable tax rates that are lower than those of other countries.352 The Report acknowledges that all three situations may have negative effects from the perspective of a high tax country but explicitly excludes the third category from the scope of the Report.353 A major contribution of the OECD Report is listing factors to be used in identifying tax havens which were not a focus of the EU package (there are no traditional tax havens in the EU). Those factors are: a. b. c. d.

No or only nominal taxes (a necessary condition); Lack of effective exchange of information; Lack of transparency; No substantial activities by the taxpayer in the jurisdiction.354

These objective factors are superior to the previous attempt by the OECD to identify tax havens based on a reputation or “smell test.”355 While the Report does not attempt to list tax havens, the Forum on Harmful Tax Competition which it sets up has begun to do so and has asked Bermuda to explain why it should not be labeled a tax haven.356 The Report also lists factors to be used in identifying “harmful preferential tax regimes.” Those are similar to the factors in the EU Code of Conduct except that the OECD list is limited to financial and service activities. The factors are: a. No or low effective tax rate (a necessary condition); b. “Ring fencing” (i.e., separation from domestic tax base); c. Lack of transparency; Id. at 16. As noted below, the emphasis on country motivation may also be unfortunate. See id. at 19-20. 353 See id. at 20. The Report acknowledges, however, that some CFC regimes may apply in the third category as well and states that “[i]t is recognized that countries retain their right to use such rules in such situations.” Id. at 20, 41. 354 See id. at 23. 355 See OECD, INTERNATIONAL TAX A VOIDANCE AND EVASION, FOUR RELATED STUDIES 22 (1987). 356 See OECD Panel Reportedly Met in Paris to Address Tax Haven Inquiry, 98 TAX NOTES INT ’L 201-7 (Oct. 19, 1998). 351 352

77

d. e. f. g. h. i. j. k. l.

Lack of effective exchange of information; An artificial definition of the tax base; Failure to adhere to international transfer pricing principles; Exemption of foreign source income from tax; Negotiable tax rate or base; Existence of secrecy provisions (bank secrecy, bearer debt); Access to wide network of tax treaties; Promotion as tax minimization vehicle; Encouragement of tax-driven operations.357

The first four factors are identified as “key” factors and the other factors are less significant.358 In addition, the Report lists three questions that can be used to identify whether a preferential tax regime is harmful: 1. Does the tax regime shift activity from one country to the country providing the preferential tax regime, rather than generate significant new activity? 2. Is the presence and level of activities in the host country commensurate with the amount of investment or income? 3. Is the preferential tax regime the primary motivation for the location of the activity?359 However, little guidance is given on how to answer these questions. In particular, as the debates about whether outbound FDI displaces jobs in the home country show, the first question is very hard to answer empirically.360 The second question may also be debatable.361 The third question, which hinges on motive, may be the hardest of all.362 As a result, I doubt whether these questions will add much to the objective factors identified above. In particular, a preferential tax regime that is identified as such under the objective factors should be regarded See OECD REPORT , supra note 23, at 27-34. See id. at 27. 359 See id. at 34-35. 360 See EDWARD M. GRAHAM & PAUL R. KRUGMAN, FOREIGN DIRECT INVESTMENT IN THE UNITED STATES (1989). 361 For a similar debate in the state context see Allied-Signal v. Dir., Taxation Div., 504 U.S. 768 (1992) (holding there must be a minimal connection between the interstate activities and the taxing state); Quill Corp. v. North Dakota, 504 U.S. 298 (1992) (holding that business was solicited through catalogs and flyers was enough to constitute a presence. 362 Numerous cases have developed a principal purpose test in order to ease the determination of this very question. See e.g., Dittler Bros., Inc. v. Commissioner, 642 F.2d 1211 (T.C. 1981) (holding that the transfer of cash and property to a foreign corporation in exchange for stock was in pursuance of a plan having as one of its principal purposes the avoidance of Federal income taxes); Airport Grove Corp. v. U.S., 431 F.2d 739 (5th Cir. 1970) (holding disallowance of surtax exemptions because of the principal purpose of the formation of the corporation was to obtain the benefits of multiple surtax exemptions, thereby avoiding federal income tax under I.R.S. § 269); Pitcher v. Commissioner, 84 T.C. 85 (1985) (holding exchange of all outstanding shares of one corporation for another was not a tax-free transaction when a principal purpose was avoid taxation). 357 358

78

as harmful even if it is impossible to show that it is intended to poach the tax base of another country.363 Chapter 3 of the OECD Report is the most important. It contains nineteen recommendations divided into three categories: recommendations concerning domestic legislation which each member may adopt unilaterally, recommendations concerning tax treaties, and recommendations for intensification of international cooperation.364 In addition, the chapter contains “Guidelines for Dealing with Harmful Preferential Tax Regimes in Member Countries.” These guidelines resemble the EU Code of Conduct: They represent a non-binding political commitment by the OECD members (but, importantly, not Luxembourg and Switzerland) to eliminate their preferential tax regimes. The guidelines contain three R’s: to refrain, to review, and to remove. The twenty-seven member states bind themselves to refrain from adopting new harmful tax practices (as defined in Chapter 2). They also commit to review their existing measures and report to the Forum on their harmful tax practices within two years. Finally, the member states commit to remove those practices within five years (with a final deadline of December 31, 2005).365 Finally, the Report recommends establishment of the Forum on Harmful Tax Practices which has in fact been established.366 The main goals of the Forum are: (i) to establish a list of tax havens; (ii) to review potential harmful tax practices by member countries (which may argue that their practices are not harmful) and to coordinate responses to harmful practices by nonmembers; and (iii) to “encourage actively non-member countries to associate themselves with” the Guidelines.367 Any member may request review of the practices of another member by the Forum which may issue a non-binding opinion. As Weiner and Ault note, the establishment of the Forum may be “the most important achievement” of the Report because it is the “first broadly mandated international institutional structure directly responsible for the evaluation and coordination of existing and proposed tax measures.”368 While the opinions of the Forum are not legally binding, they are likely to have a significant impact on member states who will be found to have violated their commitment to the guidelines.369 Luxembourg and Switzerland abstained from the vote on the Report and are therefore not bound by it.370 Luxembourg’s main objection was to the fact that the Report “[b]y See Troup & Hale, supra note 316 (questioning the appropriateness of tests based on country motivation). Taxpayer motivation is not much better. 364 The most important of these recommendations are for member countries to adopt controlled foreign corporation and foreign investment fund rules, and restrict the application of the exemption method, in a fashion consistent with the desirability of curbing harmful tax practices. 365 See OECD REPORT , supra note 23, at 56-57. Ireland has agreed to eliminate its Dublin center for international financial activities by that date. 366 See id. at 66 (Council instruction); see Weiner & Ault, supra note 346. 367 OECD REPORT , supra note 23, at 57. 368 Weiner & Ault, supra note 346, at 606. 369 Id. 370 See OECD REPORT , supra note 23, at Annex II (Statements by Luxembourg and Switzerland). 363

79

voluntarily limiting itself to financial activities … adopts a partial and unbalanced approach.”371 In addition, Luxembourg noted as a pre-emptive measure that the underlying philosophy of the report (emphasizing exchange of information) cannot be extended to taxation of interest; it proposes the EU’s co-existence model instead.372 Since any member can block Council recommendations, this threat is credible. Switzerland also objected to the focus on financial activities and to the implied rejection of the co-existence model. In addition, Switzerland objected to using lower tax rates as a criterion for identifying harmful preferential tax regimes, even though this criterion is never sufficient by itself. The Swiss argued that “[t]his results in unacceptable protection of countries with high levels of taxation.”373 However “[a]fter having seriously considered the possibility of exercising its veto,” Switzerland decided to abstain “in order not to prevent [the] adoption [of the Report] by other OECD member countries wishing to do so.”374 (ii) Critique. The OECD Report is a major achievement: It represents the first attempt to limit harmful tax competition that is based on a broad consensus by twenty-seven nations, including developing countries (the EU package only applies to fifteen developed countries). It is particularly remarkable that Switzerland and Luxembourg, who as financially-oriented tax havens are the most affected by the recommendations, chose not to veto the Report. This probably reflects their judgment that doing so would subject them to an unacceptable level of criticism in the other OECD member countries, and so reflects the fact that a broad consensus on the principal goals of the Report has indeed developed. Thus, the OECD Report is an indispensable first step in the process of limiting harmful tax competition. The Forum that it establishes promises to provide the institutional framework for further work on this area and may (as suggested in Part VI) become the initial kernel of a “world tax organization.” The development of any further limits on harmful tax competition must realistically be based on what has been achieved by the OECD, and the achievements of the OECD are the most promising signal that further work can usefully be done in this area. However, the OECD Report does suffer from several problems and limitations, some of which may be an inevitable result of the political compromises needed to achieve consensus among a diverse group of twenty-seven countries. Four limitations may be mentioned. First, the normative grounding (Chapter 1) is far too cursory to be persuasive. In addition, the emphasis on tax base erosion (instead of the beneficial uses of tax revenues) leaves the Report open to the Leviathan critique that it is merely an attempt by governments of high tax countries

Id. at 74. See id. 373 Id. at 77. 374 Id. at 78. 371 372

80

to protect their tax revenues even if their citizens would benefit from lower taxes.375 This is particularly troubling since the OECD does not have the data to back up the tax base erosion argument. It would have been better to set out more fully the arguments based on efficiency and equity considerations which are only briefly mentioned. Second, the line drawn between acceptable and harmful tax competition is the right one (i.e., it distinguishes between generally applicable tax reductions and tax regimes that are limited to foreign investors as well as traditional tax havens). However, the emphasis on whether the country intended to adopt a “beggar thy neighbor” policy is unfortunate because it invites fruitless inquiries into relative motivation. It would have been better to focus on objective factors and to draw the line based on considerations of democracy and of the benefits and costs to the investors. Third, the self-imposed limitations of the OECD Report are problematic, although they were probably politically necessary. 376 The topic of taxation of interest was deferred (originally to 1999); but given the threat of a Luxembourg veto it is unclear whether a report on this topic can be adopted, at least if it adheres to exchange of information as the main mechanism.377 More troubling is the limitation to financial activities and services since in this case there is not even a timetable for extending the Report to real activities. As argued in Part II, given the proliferation of production tax havens and the reductions in transportation costs, real activities are just as subject to harmful tax competition as financial and service activities even if they are less mobile in the short run. Finally, the most important limitation of the OECD Report is similar to the main problem of the EU package: Its direct application is restricted to OECD members (excluding Luxembourg and Switzerland). Although this scope is broader than the EU, it is far from sufficient. How can the scope be extended to non-member countries? The increasingly wide adoption of the OECD transfer pricing guidelines by non-member countries and the increasing use of the OECD model treaty even by developing countries are encouraging indications that the suspicion of the OECD as the “rich countries’ club” have abated somewhat. After all, the OECD now includes many countries that are classified by the World Bank as developing. 378 Thus, it seems likely to assume that the Forum will be able to persuade many non-member countries to abide by the Guidelines especially if they envisage For a non-persuasive version of this critique, see Arthur W. Wright, Review: OECD Harmful Tax Competition Report Falls Short, 98 TAX NOTES INT ’L 158-11 (Aug. 17, 1998). See also the responses of Eric Osterweil, In Defense f the OECD REPORT on Harmful Tax Competition, 98 TAX NOTES INT ’L 182-8 (Sept. 21, 1998) and OECD Fiscal Affairs Committee Chair Jan Francke, The 1998 OECD REPORT on Harmful Tax Competition: Just Right, 98 TAX NOTES INT ’L 187-11 (Sept. 28, 1998). 376 I suspect that the reasons Luxembourg and Switzerland wanted to include real activities in the Report was not just that they are mainly havens for financial activities, but also that they knew that such an inclusion would delay the Report. 377 Withholding taxes are mentioned as a “topic for further study.” See OECD REPORT , supra note 23, at 60. 378 These are Turkey, Mexico, the Czech Republic, Hungary, Poland, and Korea. 375

81

joining the OECD at some point.379 However, as the examples of Luxembourg and Switzerland indicate, other non-member countries that have preferential tax regimes, as well as the traditional tax havens, are unlikely to cooperate.380 Thus, the key question in evaluating the Report is whether the recommendations contained in it will enable the OECD to curb tax competition by non-cooperating non-member countries. In this regard, the main recommendations are to expand the use of CFC and foreign investment fund (FIF) regimes.381 The other recommendations, which focus on administrative measures, are likely to be helpful but not decisive.382 CFC and FIF regimes (such as Subpart F and PFIC in the U.S.) are ways of subjecting low-taxed foreign income to current tax in the taxpayer’s country of residence. Specifically, CFC regimes apply to corporate taxpayers and require them to include currently in their income certain types of income of their controlled foreign subsidiaries. The types of income typically covered are either classified by category (e.g., passive income) or by source (e.g., income from low-tax jurisdictions). FIF regimes apply to individual shareholders and require them to report currently (or pay an interest charge on) income from foreign investment vehicles. Relying on CFC and FIF regimes is problematic in several ways. First, existing regimes are filled with loopholes: Even Subpart F, which is the strictest CFC regime, has broad exceptions for active business income and for income derived from the active conduct of a financial business.383 The OECD Report merely recommends that member countries “consider” applying CFC and FIF regimes “in a fashion consistent with the desirability of curbing harmful tax practices.”384 The risk is that in the absence of a stronger commitment by all OECD

379

Mexico adopted the transfer pricing guidelines and Korea changed its definition of permanent establishment as the price of admission to the OECD. See John A. McLees et al., Tax Advisors' Forum: Extension of Mexican Assets Tax Complicates Transfer Pricing For Maquiladoras, 95 TAX NOTES INT ’L 88-1 (May 8, 1995); Ken Cook & Yasuko Masaki, Korea's New International Tax Coordination Law, 96 TAX NOTES INT ’L 14-4 (Jan. 22, 1996). 380 This is acknowledged in the Report. See OECD REPORT , supra note 23, at 20. The Report argues that countries with preferential regimes are more likely to want to cooperate because they have revenues to protect, but the Luxembourg and Switzerland cases raise doubts on the validity of this prediction. 381 See id. at recommendations 1 and 2. Recommendation 3 is similar and applies to countries with exemption systems. 382 These recommendations are to require information reporting on foreign holdings (4), transparency of rulings (5), applying transfer pricing rules (6), limiting bank secrecy (7), exchange of information (8), limitation of benefits in treaties (9), applying anti-abuse rules to treaties (10), exclusion of types of entities or income from treaties (11), terminating treaties with tax havens (12), coordinated enforcement (13), assistance in recovery of tax claims (14), listing tax havens (16), limiting links to tax havens (17), developing principles of tax administration (18), and engaging in dialogue with non-members (19). See id. 383 This last provision, enacted on a “temporary” one-year basis in 1997 and extended in 1998, runs directly contrary to the spirit of the OECD Report. 384 OECD REPORT , supra note 23, at 41, 42.

82

members countries will not be able to resist the lobbying pressures of their multinationals to restrict CFC rules in the name of competitiveness.385 Second, even if OECD members could adopt broad CFC and FIF rules that apply to all the activities identified as harmful tax practices, it is doubtful whether that would suffice to curb harmful tax competition. In regard to FIF rules, the problem (as explained in Part II) is an administrative one: Individuals resident in one OECD member can invest in another OECD member through a tax haven entity and escape the exchange of information net. It is not clear whether the “greater and more effective” exchange of information envisaged by the OECD Report can overcome this problem. 386 As for terminating tax treaties with tax havens and otherwise limiting treaty benefits, it would have no effect as long as countries maintain their nontreaty based exemptions from withholding on interest earned by non-residents.387 In regard to CFC rules, the problem is not enforcement but scope. Even if CFC rules were to apply to all tax havens and preferential tax regimes, they still would apply only to multinationals whose parent is a resident of an OECD member. Currently, about 85% of the parent corporations of the world’s MNEs are resident in OECD member countries.388 However, as argued in Part II, if CFC rules were significantly expanded it is quite likely that in the long run existing MNEs would seek to shift their residence to non-OECD member countries (or to Luxembourg or Switzerland). In addition, the parent corporations of new MNEs will be incorporated or managed and controlled from outside the OECD. In particular, the various headquarters tax havens of the world, many of which are outside the OECD (e.g., Singapore), are likely to welcome any expansion of CFC rules by the OECD. This shift may take some time but it is likely to happen for the reasons identified in Part II, and the OECD is likely to be powerless to stop it. The OECD Report does recommend as a “topic for further study” a review of the residency rules. It acknowledges that both country of incorporation and country of management and control are “easily manipulated.” Instead, it recommends focusing on the residence of the shareholders.389 However, most multinationals currently are trading on several exchanges so that it would not be feasible to establish a single country of residence this way and residence will constantly shift as shares are bought and sold (which CFC rules will apply?). In addition, given 385

See the debate around Notice 98-11 and the current effort to limit Subpart F to passive income, NFTC Project, supra note 113. 386 OECD REPORT , supra note 23, at recommendation 8; see also OECD REPORT , supra note 23, at recommendation 4 (adoption of foreign income reporting rules and exchange of information). An interest charge would work only if the income is eventually distributed to an OECD member country rather than enjoyed through consumption overseas. 387 See id. at recommendations 9 (limitation of benefits), 11 (limiting treaties by entity or type of income), and 12 (treaties with tax havens). 388 See W ORLD INVESTMENT REPORT , supra note 98. 389 See OECD REPORT , supra note 23, at 60. For a similar suggestion, see Doernberg, supra note 298; Howard E. Abrams & Richard L. Doernberg, How Electronic Commerce Works, 14 TAX NOTES INT ’L 1573 (May 12, 1997).

83

the widespread use of nominees, it is far from clear that the residence of the shareholder of a publicly traded MNE can be known. 390 Thus, the solutions offered by the OECD Report to the problem of harmful tax competition represent an indispensable first step toward limiting harmful tax competition, but they are incomplete. The next section represents a preliminary attempt to develop a better approach, which combines the best elements of the EU package and the OECD Report with some suggestions of my own. c. Recommended Proposals The following two proposals build on the work of the EU and the OECD described in the previous section. The first proposal addresses the under-taxation of cross-border portfolio investment while the second proposal addresses the same issue for direct investment. In both cases, the emphasis is on actions that can be taken by the OECD without securing the approval of traditional, production or headquarters tax havens which is unlikely to be forthcoming. The proposals therefore focus on ways to impose a tax by those countries (OECD members) that are most likely to want to tax the income. In the case of portfolio investments, because lowrisk, high-return investments are likely to be found only in OECD member countries, the host country into which the investment is made can levy the tax. In the case of direct investment, because countries in which goods and services are consumed are likely to wish to tax (there are few, if any, “consumption tax havens”), the tax should be primarily levied by the (OECD member) country of consumption. In both cases, however, the tax can be shared or even remitted completely if the income is subject to tax elsewhere. In the case of portfolio investment, the tax is refunded if the income is reported to tax authorities in the home country of the investor even if it is subject to a low tax rate there. In the case of direct investment, the tax base can be shared with countries in which production takes place even if subject to a generally applicable lower tax rate in those countries. Thus, the distinction between harmful and acceptable tax competition is preserved. (i) A Uniform Withholding Tax on Portfolio Investment As noted above, the OECD Report does not address the issue of taxation of portfolio investment leaving that topic to a further report originally scheduled to be issued in 1999. Meanwhile, the EU package does contain a draft Directive on this topic, but its application is 390

A more feasible suggestion is to abandon the corporate level tax and to focus on taxing the shareholders either by imputation (which is administratively cumbersome) or by marking their shares to market. See Robert A. Green, The Future of Source-Based Taxation of Income of Multinational Enterprises, 79 CORNELL L. REV. 18 (1993); Dodge, supra note 300. But as I have argued elsewhere, source countries are unlikely to give up their right to the corporate tax when shareholders are residents of other countries. See International Taxation of Electronic Commerce, supra note 34. Nor, under accepted principles of internation equity, should they. See Inter-nation Equity, supra note 169. If corporate level taxes are retained, one would have to combine mark-to-market (or a PFIC-like interest charge) with foreign tax credit rules which would be exceedingly complex.

84

limited to payments made within the EU to individual EU residents. Moreover, even this draft may have to be abandoned in the face of opposition from Luxembourg and the U.K., who fear that their lucrative Eurobond markets may flee to Zurich or New York if the Directive were adopted.391 This situation does, however, present a golden opportunity. As described in Part II, the problem of non-taxation of cross-border interest flows stems to a large extent from the unilateral enactment of the portfolio interest exemption by the U.S. in 1984. That enactment was driven by the need to finance a growing budget deficit and by the fear that any tax withheld on portfolio interest flows from the U.S. would simply be shifted forward to the U.S. borrowers (including the U.S. Treasury). Under current conditions, the reasons that led to the enactment of the portfolio interest exemption no longer exist.392 The U.S. is enjoying a budgetary surplus and because it is perceived as a safe haven in financially troubled times it is likely that U.S. borrowers would not have to bear the cost of any withholding tax on interest. Thus, the U.S. could probably afford to repeal the portfolio interest exemption immediately without suffering adverse consequences. Further work by economists in this area is clearly desirable. Even if the repeal of the portfolio interest exemption on a unilateral basis were found to lead to adverse consequences for the U.S., the prospective OECD report and the EU position render multilateral action much more likely. As observed in Part II, the non-taxation of crossborder interest flows is an assurance game: Each player (the EU, the U.S., and Japan) refrains from taxing for fear of driving investment to the others, even though they would all benefit from imposing the tax. However, it is well established that such assurance games can be resolved if parties can signal to each other in a credible fashion their willingness to cooperate.393 The EU draft Directive represents just such a signal. The EU is telling the U.S. that it is willing to go forward with taxing cross-border interest flows, and even Luxembourg and the U.K. are willing to cooperate if the U.S. and Japan agree to follow the EU lead. Thus, if in the context of the OECD report on taxation of interest the U.S. and Japan were to commit themselves to taxing cross-border interest, the assurance game could be resolved and a new, stable equilibrium of taxing rather than refraining from tax be established.394 See EU Savings Tax Gains Strength Before ECOFIN Meeting, 1999 W ORLDWIDE TAX DAILY 195-5 (Oct. 8, 1999). 392 See the discussion in Part II. 393 See OLSON, supra note 19. 394 The U.S. role in this regard would be more of a follower rather than a leader. It is unfortunate that in enacting the portfolio interest exemption the U.S. abandoned its role as a positive leader in international tax affairs, as exemplified in being the first to adopt the foreign tax credit and CFC rules. But even if it follows Europe, the U.S. would play a decisive role in breaking the current impasse. Japan is likely to cooperate since it is a major loser from the current regime (the portfolio interest exemption was enacted to attract Japanese investors that wish to avoid tax in Japan) and a leader in the OECD effort to limit harmful tax competition. 391

85

The prospects for agreement in this area are particularly good because only a limited number of players need to be involved. The world’s savings may be parked in traditional tax havens, but the cooperation of such tax havens is not needed. To earn decent returns without incurring excessive risk, funds have to be invested in an OECD member country (and more particularly, in the EU, the U.S., Japan, or Switzerland). Thus, if the OECD member countries could agree to the principles adopted by the EU in its draft Directive, most of the problems of taxing cross-border portfolio interest flows could be solved. My proposal is therefore as follows: The OECD should implement on a coordinated basis the principles contained in the EU draft Directive on taxation of savings. However, while in the EU context exchange of information could play a large role because there are few traditional tax havens in the EU, in a global context withholding taxes have to be the primary means of enforcement. As noted above, traditional tax havens with strong bank secrecy laws render it very difficult to have effective exchange of information among OECD member countries. If the investment is made through a tax haven intermediary, exchange of information is likely to be useless unless the tax authorities in the payor’s country can know the identity of the beneficial owner of the funds that are paid to the tax haven intermediary. I would therefore propose that instead of the “co-existence” model of the EU, the OECD adopt a uniform withholding tax on cross-border interest flows which should also be extended to royalties and other deductible payments to portfolio investments (e.g., payments on derivatives).395 To approximate the tax rate that would be levied if the payment were taxed on a residence basis, the uniform withholding tax rate should be high (at least 40%). However, unlike the withholding taxes that were imposed before the current “race to the bottom” started in 1984, the uniform withholding tax should be completely refundable. To obtain the refund, as in the EU draft Directive, a beneficial owner need only show the tax authorities in the host countries a certificate attesting that the interest payment was reported to the tax authorities in the home country. No actual proof that tax was paid on the interest income is required: From an efficiency, equity, and revenue perspective, it is sufficient that the home country authority has the opportunity to tax the income from overseas investments in the same way as it taxes domestic source income.396 Thus, in accordance with the distinction drawn in Part III(e), even if the home country has a generally applicable low tax rate on its residents (or even a zero tax rate as

395

It may also be advisable, but is less crucial, to extend the tax to non-deductible payments like dividends. See Avi-Yonah & Swartz, supra note 13. 396 It may be possible for developing countries to obtain a direct transfer of the funds from the OECD member country which they would then credit against domestic tax liability. The OECD member would retain a small percentage of the tax as a fee for its collection assistance. For a precedent for this type of procedure, see Gaza-Jericho Agreement, May 4, 1994, Israel-Palestine, 36 I.L.M. 551 (establishing system under which Israel collects taxes from Palestinians working within its borders and remits 75 percent of the revenue to the Palestinian National Authority).

86

long as it applies to all bona fide residents), the resident could obtain a refund by reporting the income to the tax authorities in his home country.397 Both the proposed withholding tax and the refund mechanism would not require a tax treaty. However, it would be possible for countries to reduce or eliminate the withholding tax in the treaty context when payments are made to bona fide residents of the treaty partner. In those cases, the exchange of information in the treaty should suffice to ensure residence-based taxation. Since most OECD members have tax treaties with most other OECD members, the proposed uniform withholding tax would generally apply only to payments made to non-OECD member countries (including traditional tax havens). Were this type of uniform withholding tax enacted by OECD members, it would go a very long way toward solving the problem described in Part II of under-taxation of crossborder portfolio investments by individuals. Such under-taxation is unacceptable from an efficiency or equity perspective. Moreover, unlike the under-taxation of direct investment, this type of under-taxation is illegal (which is why it is so hard to assess its magnitude). By adopting a uniform withholding tax, the OECD could thus strike a major blow at tax evasion which (as described in Part II above) is a major problem for most developing countries and some developed countries (including OECD members) as well. (ii) Consumption-Based Taxation of Multinationals The OECD proposals on taxing multinationals on their income from international operations are adequate in the short run. If every OECD member country were to adopt an effective CFC regime that would apply to all foreign source income of “its” multinationals, this would solve 85% of the problem because that is the percentage of multinationals whose parents are currently resident in OECD member countries.398 Moreover, the OECD Report presents an effective answer to the common argument made by multinationals against CFC regimes which was developed in Part III(a): If all OECD member countries adopt effective CFC regimes, then the multinationals resident in any given OECD country would not be at a competitive disadvantage vis-à-vis the multinationals resident in any other member country.399 397

This tax applies to individuals and thus residence is a meaningful concept. If an individual chooses to live in a Caribbean tax haven, she can under the proposal avoid tax on his foreign-source investment income. Presumably, living in a tax haven entails giving up on certain public services in exchange for the lower taxes. 398 It should be noted, however, that current CFC regimes do not observe the distinction between generally applicable rate reductions and targeted tax havens that is made in the OECD Report. See OECD REPORT , supra note 23, at 41 (“CFC rules may also apply in situations which do not involve harmful tax practices as defined in this Report. It is recognized that countries retain their right to use such rules in such situations.”). 399 This is why current arguments based on competitiveness to restrict the scope of Subpart F are wrong. See NFTC Project, supra note 113. An interesting precedent can be found for this type of situation in another field of law: U.S. multinationals have argued since 1977 that the Foreign Corrupt Practices Act restricts their ability to effectively compete overseas against foreign multinationals. These complaints have now lost their force since the OECD adopted strict anti-corruption guidelines in 1996.

87

However, in the longer run, it is likely that multinationals will be able to establish the residence of their parent in non-OECD member countries (or in OECD members that do not subscribe to the report like Luxembourg or Switzerland). In that case, the application of CFC regimes is likely to ineffective. The solution envisaged by the OECD report is to re-define residence based on the residence of shareholders; but in the case of a publicly traded multinational whose shares trade on many exchanges, this proposal is likely to be ineffective or very complex to administer in practice.400 It is therefore proposed instead that the OECD should adopt a regime based on taxing multinationals as an initial matter in the country in which consumption of the goods or services provided by the multinational takes place.401 The advantage of choosing those countries (the “Demand Jurisdiction”) as the locus of initial taxation is that large consumer markets are unlikely to be tax havens and are likely to want to impose the tax on foreign importers as well as on domestic sellers. This latter point can be seen if one considers the popularity of the destination principle for consumption taxes such as the VAT or the U.S. retail sales taxes. Such taxes are imposed on a destination basis without the need for a coordinating tax treaty. 402 The first step toward imposing such a tax would be to modify the permanent establishment threshold that is embodied in tax treaties and domestic legislation. The current threshold is based on physical presence but, as explained in Part II above, that concept is obsolete and is likely to lead to under-taxation of multinationals who can sell into a jurisdiction without having a physical presence therein. Thus, as I have proposed elsewhere, a different type of threshold is required, one that will not be linked to physical presence.403 Such a threshold could be a de minimis amount of sales into the jurisdiction as suggested by Walter Hellerstein and others in the state sales tax context.404 For example, the rule could be that if a seller has gross sales of $1 million or less from a given tax jurisdiction (adjusted for inflation) it would not be subject to taxation at source.405 400

See the discussion in Part V(c) above. For a similar proposal for electronic commerce, see International Taxation of Electronic Commerce, supra note 34. That proposal however was not limited to harmful tax competition. The OECD Report does mention as a “topic for further study” the possibility of denying deductions for payments to tax havens, citing a Spanish precedent. See OECD REPORT , supra note 23, at 59-60. But this proposal is not elaborated. 402 They are thus operationally independent from the international tax system. See Charles E. McLure, Jr. & George R. Zodrow, The Economic Case for Foreign Tax Credits for Cash Flow Taxes, 51 NAT ’L TAX J. 1-22 (1998). Note that imposing income taxes on a destination principle will not violate the WTO rules if those taxes are imposed on domestic producers and are rebated if countries in which production takes place impose a tax. See Victoria P. Summers, The Border Adjustability of Consumption Taxes, Existing And Proposed, 96 TAX NOTES INT ’L 107-8 (June 3, 1996). 403 See International Taxation of Electronic Commerce, supra note 34. 404 See Walter Hellerstein, State Taxation of Electronic Commerce, 52 TAX L. REV. 425 (1997). Aggregation rules will be required to prevent manipulation of the threshold. 405 My recommendation is for a gross sales threshold, rather than a net income threshold, for administrability reasons. A net figure would require that the tax administration know the taxpayer’s income from sales into a jurisdiction which it would typically not have the information to determine while the gross 401

88

A key issue is obviously the determination of what would constitute a sale into a jurisdiction. Fundamentally, the inquiry should be where the goods or services sold are consumed. However, since making that determination is difficult, some simpler proxy like a billing address can be used. In the income tax context, a billing address may be adequate for most individual customers since (i) they are unlikely to consume the product elsewhere, and (ii) they do not have an incentive to provide a false address because they typically do not bear the tax burden. For business customers, both of these factors may be absent because the billing address may be anywhere (including in particular tax havens), and in the case of large customers collusion is a possibility. However in that case, since the customer is physically present in the taxing jurisdiction, an audit may be able to determine the actual location in which a product was used. Moreover, businesses may have an incentive to locate the destination of goods or services provided to them in a high-tax jurisdiction because that would maximize the value of the deduction they would take for the value of such goods or services in determining their own income tax liability. The proposed withholding tax regime for income derived from active business operations to be adopted by the OECD on a multilateral basis would have the following rules: First, a gross withholding tax is imposed on sales (and services) provided into the Demand Jurisdiction at a rate equal to the corporate tax rate in the Demand Jurisdiction. As explained above, a sale is defined as being into the Demand Jurisdiction if the goods or services are consumed therein under rules similar to a destination-based VAT. This tax can be imposed by the Demand Jurisdiction through forbidding merchants from selling goods to its residents unless procedures for withholding the tax are in place.406 Second, to obtain a refund or reduction of the gross tax the taxpayer has to file a return showing its deductions (including cost of goods sold). Thus, the function of the gross withholding tax is to force taxpayers to file a return in a jurisdiction where they have no physical presence.407 It may be advisable, for the reasons stated above, to allow at this point a full refund of the tax if the total gross sales fall below the expanded threshold. The remaining rules thus apply only if the threshold is exceeded.408 amount can be determined from the records of other parties (the customers). For the same reason, a threshold based on a percentage of total sales worldwide seems impracticable since it requires knowledge available only to the taxpayer. 406 Because the proposal is for coordinated OECD action, countries need not fear driving sellers away by imposing the tax since most large markets will be covered. 407 This is similar to the rule for U.S. real estate transactions. See I.R.C. § 1445. 408 Having a threshold will enable some income to escape taxation, but a threshold is necessary to shield small businesses and to prevent situations where the cost of compliance exceeds the profit derived from the transaction. For a fuller discussion, see International Taxation of Electronic Commerce, supra note 34.

89

Third, the Demand Jurisdiction disallows deductions to related and unrelated parties that are located in jurisdictions that have tax incentives targeted to foreigners unless those parties file a return and pay tax to the Demand Jurisdiction. This rule preserves the distinction made above between jurisdictions that engage in harmful tax competition (defined under the above criteria to mean tax incentives targeted at foreigners) and jurisdictions with generally applicable low tax rates.409 Thus, all income that is not taxed by another source jurisdiction under targeted tax incentives is allocated to the Demand Jurisdiction. However, income that is taxed at a lower rate than the rate in the Demand Jurisdiction because of generally applicable tax reductions will continue to benefit from those lower rates.410 This rule is similar to the VAT rule that allows deductions for inputs only for purchases from registered VAT payers. It also reflects some aspects of the current international tax regime, for example, the denial of tax sparing credits, the anti-treaty-shopping rule, and, most recently, the hybrid entity rules.411 All of these rules reflect the view that a source or residence country should not reduce its tax unless it can be sure that the income is really subject to tax somewhere. However, the rule should be limited to countries that engage in harmful tax competition and is therefore more restricted and allows for more flexibility than the abovementioned rules. It should be emphasized that under the proposed regime only income that is untaxed in the production jurisdiction because of production tax havens, or that does not belong to any jurisdiction, is allocated to the Demand Jurisdiction.412 Thus, taxation by the Demand Jurisdiction should be viewed primarily as a way of preventing harmful tax competition because it is less likely to be a tax haven than either residence or production jurisdictions. Finally, the Demand Jurisdiction refunds the difference between the gross tax and the net tax per the return.

409

A look-through rule will be needed to prevent sellers from selling into the Demand Jurisdiction via independent, low profit distributors located in a jurisdiction that does not engage in harmful tax competition. In most cases, however, MNEs like to control the distribution of their products (otherwise they would sell to independent distributors rather than engage in costly FDI). See id. for further elaboration of this point. 410 To prevent countries from engaging in harmful tax competition through targeted subsidies (rather than tax reductions), a rule similar to the anti-subsidies rule of I.R.C. § 901(i) would be needed. Under such a rule a targeted subsidy would be treated as equivalent to a targeted tax break. 411 See I.R.C. § 894(c). 412 In the case of related parties, transfer pricing principles (based on a profit split) should be applied and only the residual profit allocated to the Demand Jurisdiction. See International Taxation of Electronic Commerce, supra note 34, at 545-50, for further elaboration of this point.

90

The net tax is then credited by the residence jurisdiction of the corporation (if it imposes a corporate tax).413 If the residence jurisdiction has an integrated tax regime, credit for the net tax should be allowed at the shareholder level. While this system appears quite complex, it is fundamentally similar to the rules for a destination-based VAT except that the tax base is net income and not consumption. These rules seem to operate quite well (with a few notorious exceptions such as banks) and to not require an elaborate tax treaty network to allocate the tax base among countries. The rules can be adopted by the OECD Demand Jurisdictions acting in concert but may even be adopted by a majority if some members dissent. Moreover, the information required by the Demand Jurisdiction to implement the rules relates to the tax systems of other countries (which it can find out about) rather than to the specific tax characteristics of the individual taxpayer (which it is harder to ascertain).414 The OECD can adopt this proposal as a way to limit harmful tax competition in those situations in which rules based on corporate residence (i.e., CFC regimes) do not work. The key to the proposal, just like the withholding tax proposal, is that the tax base is shared with those countries that do not engage in harmful tax competition. However, the tax base is shifted from production tax havens and headquarters tax havens toward the Demand Jurisdiction. In terms of inter-nation equity, the effect of the proposal may seem regressive: Tax revenues are shifted from developing countries to the richest countries in the world (the richer a country is, the higher its demand for goods and services, and thus its role as a Demand Jurisdiction). However, if the analysis in Part IV is correct, and developing countries engage in tax competition mostly for fear of losing investments to other developing countries, the likely effect of the proposal would be to stop this race to the bottom. If so, developing countries could benefit by imposing their taxes on multinationals that produce therein without fearing that other countries could undercut them by offering production tax havens. Generally applicable tax reductions would be available as a policy tool but for the reasons stated above they are unlikely to be a major determinant in attracting investments (because profits are likely to be lower if overall public services are lower). Thus, the result of the proposed OECD actions could prove to be beneficial, rather than harmful, to developing countries (who could use the extra revenues as discussed in Part IV). Investments would be attracted to countries based on their comparative advantages in non-tax areas and on the basis of generally applicable tax rates rather than because of harmful tax competition.

413

No credit should be given for taxes imposed (through the denial of deductions) to parties that are unaffiliated with the resident taxpayer. 414 Many countries have CFC regimes that hinge on whether the other country imposes a sufficient tax rate. For example, Australia takes even sub-country tax variations into consideration. See HUGH J. A ULT , COMPARATIVE INCOME TAXATION: A STRUCTURAL A NALYSIS 387 (1997). In all cases the burden of proof will be on the taxpayer to show that it is subject to tax.

91

However, from a developing country perspective it would be advisable to combat harmful tax competition through a multilateral forum in which the developing countries are represented directly rather than through the OECD. The potential for developing such a forum and its likely institutional setting are discussed in Part VI.

92

VI.

Conclusion: Is There a Need for a World Tax Organization?

The preceding analysis can be summarized as follows: Tax competition has resulted in a significant potential for avoidance of income taxation on the income from cross-border portfolio and direct investments which may lead to significant tax base erosion and revenue losses (Part II). Such avoidance has negative implications for global efficiency and equity, but any actions to restrict tax competition need to be balanced against maintaining the ability of democracies to determine the size of their governments (Part III). Moreover, from the perspective of the countries involved, unrestrained tax competition threatens the social insurance safety net in developed countries at a time when it is facing a severe fiscal crisis and is generally not in the best interest of developing countries as well. In balancing the competing revenue needs of developed and developing countries, inter-nation equity indicates preferring the latter over the former (Part IV). Part V of the article then proceeds to first discuss and criticize the current efforts to restrict tax competition by the EU and the OECD and to develop alternative solutions. These solutions are designed to be implemented by the OECD because currently OECD member countries are both the destination of most portfolio investment and the largest markets. Thus, the OECD can effectively implement solutions that depend on withholding taxes to be levied by the source country (for portfolio investment) and by the demand jurisdiction (for direct investment). But the purpose of the withholding taxes proposed is to enable residence jurisdictions to levy taxes on portfolio investment, and source jurisdictions to levy taxes on direct investment, without the limitations imposed by unrestricted tax competition. In both cases, this is the outcome most favorable to developing countries in accordance with inter-nation equity. However, relying on the OECD to restrict tax competition in this manner ultimately suffers from two significant drawbacks. First, it can be envisaged that in the longer run significant markets for both portfolio investment and retail sales will develop outside the OECD. When that happens solutions that rely on OECD enforcement will lose their effectiveness unless those emerging markets were to join the OECD. While several developing countries have joined the OECD recently (e.g., South Korea and Mexico), it is hard to imagine China or India doing so in the near future. More importantly, relying on the OECD to implement solutions to the tax competition problem, even if those solutions are tailored to benefit developing countries, may not be acceptable to those countries. Even though the OECD has made a huge effort to include nonOECD members in the tax competition project, it is still identified as the rich countries' club. Thus, it is hard to believe that developing countries will be able to shed their suspicions that the OECD will not act in their interests even if it can actually be made to do so. In fact, the effort by the OECD to develop a multilateral agreement on investments (MAI) foundered precisely because developing countries and left-leaning non-governmental organizations coordinated a campaign against it as representing the interests of the rich countries and "their" MNEs.

93

Thus, as noted above, it would be preferable if the effort to restrict harmful tax competition (in the ways described in Part V above) were pursued by a multilateral body that included the developing countries. No such body currently exists, although several scholars have proposed setting one up.415 But in fact, there is a natural candidate for the job which already is in place: the World Trade Organization (WTO). For several years now, scholars have come to recognize that the traditional distinction between tax and trade issues has become untenable in theory and in practice.416 Theoretically, taxes can serve as a barrier to the free flow of trade exactly in the same way as tariffs (the traditional target of trade law). In practice, the WTO and its predecessor the GATT always had jurisdiction over export subsidies delivered in the form of tax reductions and had struck down such regimes on several occasions (including most recently the U.S. Foreign Sales Corporation provisions).417 But the distinction has become much more tenuous in practice as the GATT and the WTO moved from formal tariffs to non-tariff barriers. In particular, the inclusion of the General Agreement on Services (GATS) and of the agreement on TradeRelated Investment Measures (TRIMS) in the Uruguay Round agreements meant that the WTO now has jurisdiction in areas that relate directly to taxation: GATS because the delivery of services (unlike goods) often requires direct investment and TRIMS because it relates directly to investment.418 In the forthcoming Seattle Round, an effort will be made to revive the MAI in the WTO context, which will draw it even closer to the tax area.419 Thus, the WTO presents as the natural candidate to be the "world tax organization." In fact, it is hard to see how the WTO can fulfill its role of ensuring the free flow of trade and reducing non-tariff barriers without having jurisdiction over tax matters. In addition, from the See TANZI, supra note 11, at 140; Bird, supra note 31, at 297-99; John K. Sweet, Comment, Formulating International Tax Laws in the Age of Electronic Commerce, 146 U. PA. L. REV. 1949 (1998); Kenneth J. Vandevelde, The Political Economy of a Bilateral Investment Treaty, 92 A M.J.INT ’L.L. 621, 641 (1998). 416 See Joel Slemrod, Free Trade Taxation and Protectionist Taxation (NBER Research Working Paper No. 4902, 1994); Robert A. Green, Antilegalistic Approaches to Resolving Disputes Between Governments: A Comparison of the International Tax and Trade Regimes, 23 YALE J. INT 'L L. 79 (1998) [hereinafter Antilegalistic Approaches]. 417 Article III:2 of the GATT explicitly establishes the principle of non-discrimination in internal taxation, and Article XVI of the GATT prohibits any direct or indirect subsidy that results in the sale of a product for export at a lower price than is charged for a similar product in the domestic market; see GATT Doc. L/4422 (Nov. 12, 1976) at 1 (Chair informed Council of Panel's Composition on Feb. 17, 1976) reprinted in GATT, B.I.S.D., (23d. Supp.) 98 (1977). See also WTO, United States—Tax Treatment For "Foreign Sales Corporations”, WT/DS108/R (99-4118) (Oct. 08, 1999) (“Since the European Communities is complaining about an export subsidy, the WTO is the appropriate forum”); Philip J. Jelsma, Note, The Making of a Subsidy, 1984: The Tax and International Trade Implications of the Foreign Sales Corporation Legislation, 38 STAN . L. REV. 1327 (1986). 418 Although in both cases the application to income taxes has been restricted at the insistence on the U.S.; see discussion below. 419 The effort will be led by the EU and Japan but expects pressure from environmental and labour groups, both of which lead the fight against the OECD’s attempt to form a MAI. See Kym Anderson, The WTO Agenda For the New Millennium, 75 ECON. REC. 77-88 (Mar. 1999). 415

94

perspective developed above the fact that the WTO includes representatives from almost all the developing countries gives it an obvious advantage over the OECD even if the solutions it implements are exactly the same as the OECD-based ones proposed above. But there are several serious objections to including tax matters in the jurisdiction of the WTO. First, it has been argued that the WTO lacks sufficient tax expertise.420 However, that problem can be remedied by hiring a sufficient number of tax experts to sit on the WTO's panels. In fact, as the WTO has expanded its jurisdiction to non-tariff matters, its staff already includes tax experts who also understand trade issues. Robert Green has advanced a more serious objection arguing that the costs of imposing the WTO's legalistic dispute-resolution mechanism outweigh any benefits.421 Green argues that the need for the WTO to resolve trade disputes legalistically is based on two features that are typically lacking in the tax context: retaliation and lack of transparency. Retaliation is a feature of repeated prisoners' dilemma type games and insures that players have an incentive to cooperate.422 In an assurance (stag hunt) game, both players cooperate if they can be assured of the other player's cooperation. 423 In the first case an organizational setting is needed to manage retaliatory strategies while in the second it is needed to provide the information needed for the assurance to exist. However, in the context of tax competition it would seem that both retaliation and lack of information are serious problems. As described above, in the case of portfolio investment the U.S. began a race to the bottom by abolishing its withholding tax, and other countries responded (i.e., retaliated) by abolishing their own taxes. In the current situation no country dare re-impose its tax without adequate assurance that other countries will follow. Similarly, for direct investment, countries have adopted tax incentives or adopted deferral and exemption rules for their resident MNEs in response to the actions of other countries and fear changing such policies without assurance that others will follow suit. Thus, whether these developments are characterized as prisoners' dilemma or assurance games, they seem to present precisely the kind of problem that only a multilateral organization with rule-making power can effectively resolve.424

See Robert E. Hudec, Reforming GATT Adjudication Procedures: The Lessons of the DISC Case, 72 M INN. L. REV. 1443 (1988); Antilegalistic Approaches, supra note 416, at 123 n. 240; William M. Considine, The DISC Legislation: An Evaluation, 7 N.Y.U. J. INT 'L. L. & POL. 217 (1974). 421 See Antilegalistic Approaches, supra note 416. 422 See id.; see also ROBERT A XELROD, THE EVOLUTION OF COOPERATION (1984). 423 The assurance game has a payoff structure in which the best outcome is if both countries cooperate while in the prisoner's dilemma the best outcome is if you defect and the other side cooperates. See Antilegalistic Approaches, supra note 416. 424 The race to the bottom in international taxation in the 1980s resembled a prisoners' dilemma, in which one country (the U.S.) preferred to defect while others cooperate in order to draw investment to it. But the current situation is more like an assurance game, in which the U.S. and other OECD members would prefer cooperation above all other outcomes. 420

95

However, Green also raises another objection to giving the WTO authority over taxes which in practice is likely to be far more potent: the problem of sovereignty. Countries are wary of giving up their sovereignty over tax matters which lies at the heart of their ability to exercise national power. This concern is particularly acute in the U.S. and almost led to the failure of the entire Uruguay Round as the U.S. insisted at the last minute to exclude direct taxes from the purview of the GATS.425 Green argues that if the WTO dispute resolution mechanism were given authority over tax issues, this may lead to widespread noncompliance especially given the perception that the WTO is non-transparent and lacks democratic legitimacy. 426 But there exists a solution to this problem as well. Under the GATT regime, all decisions had to be reached by consensus, i.e., with the agreement of the party whose regime is at stake. Under the WTO rules, on the other hand, all dispute settlement rulings are binding unless there is a consensus not to implement them, i.e., when even the complaining party agrees to refrain from action. I would argue that the former rule is more appropriate for tax matters than the latter because it gives the loser a veto if it feels that its sovereignty is truly at stake. Similar rules exist for tax matters in both the EU and the OECD. But, as the DISC case in the GATT and the adoption of the tax competition report by the OECD show, a country will typically reserve its veto power only to those cases in which the adverse result is truly perceived as a severe limit on its sovereignty. In other cases, the stigma of disapproval is sufficient to ensure cooperation. The purpose of this article has been to show that as a result of globalization and tax competition tax rules can no longer be set by countries acting unilaterally or by bilateral tax treaties. In a world in which capital can move freely across national borders and MNE's are free to choose among many investment locations, the ability of any one country (or any two countries in cooperation) to tax (or otherwise regulate) such capital is severely limited. Any such unilateral attempt will be undercut by other countries and will probably not be even attempted in the name of preserving national competitiveness. Thus, a multilateral solution is essential if the fundamental goals of taxation or other regulation are to be preserved. Private market activities that span the globe can only be regulated or taxed by organizations with a similar global reach. This article has attempted to delineate some of the ways in which such global governance can be achieved in the area of capital income taxation. Achieving this goal will not be easy given the expected resistance of both private actors eager to preserve their freedom from taxation and of governments concerned about preserving their sovereign ability to set their own tax rules. But it is not impossible. Moreover, since preserving the ability of nations to tax

See Antilegalistic Approaches, supra note 416. See id.; see also Joel P. Trachtman, The Domain of WTO Dispute Resolution, 40 HARV. INT 'L L. J. 333 (1999) (describing factors to be weighed in choosing between rules and standards in the WTO context). But it should be noted that the WTO already has exercised jurisdiction over matters such as food safety, intellectual property, and similar issues that also involve sensitive sovereignty issues. 425 426

96

income from capital is essential to the achievement of several fundamental goals, like the preservation and development of an adequate social safety net, it must be tried.

97