This article first appeared in the September 5, 2011 edition of Tax Notes.
Edward D. Kleinbard is a professor at the University of Southern California Gould School of Law. He can be reached at email@example.com.
This report considers the tax consequences and policy implications of the phenomenon of stateless income. Stateless income is income that is derived for tax purposes by a multinational group from business activities in a country other than the domicile of the group's ultimate parent company but that is subject to tax only in a jurisdiction that is neither the source of the production factors through which it was derived nor the domicile of the group's parent company. Google Inc.'s "Double Irish Dutch Sandwich" structure is one example of stateless income tax planning in operation.
The report first shows that the U.S. tax rules governing income from foreign direct investments often are misapprehended: In practice they operate not as a worldwide system of taxation, but as an ersatz variant on territorial systems, with hidden benefits and costs when compared with standard territorial regimes. That claim holds whether one analyzes these rules as a cash tax matter or through the lens of financial accounting standards. This report rejects as inconsistent with the data any suggestion that current U.S. law renders U.S. multinational firms less competitive when compared with their territorial-based competitors.
Stateless income privileges multinational firms over domestic ones by offering the former the prospect of capturing "tax rents" -- low-risk inframarginal returns derived by moving income from high-tax foreign countries to low-tax ones. Other important implications of stateless income include reduced coherence in the concept of geographic source; the systematic bias toward offshore rather than domestic investment; the more surprising bias in favor of investment in high-tax foreign countries to provide the feedstock for the generation of low-tax foreign income in other countries; the erosion of the U.S. domestic tax base through debt-financed tax arbitrage; many instances of deadweight loss; and, essentially uniquely to the United States, the exacerbation of the lockout phenomenon, under which the price that U.S. firms pay to enjoy the benefits of dramatically low foreign tax rates is the accumulation of extraordinary amounts of earnings (as much as $1.4 trillion, by the most recent estimates) and cash outside the United States.
Economic policy conclusions that are useful in a world without stateless income do not follow once the presence of stateless income tax planning is considered. More specifically, stateless income planning undermines a critical assumption in the capital ownership neutrality model that has been advanced as an argument why the United States should adopt a territorial tax system.
The report concludes that policymakers must choose between a highly implausible outcome (a territorial tax system that deters stateless income base erosion) and a manifestly imperfect one (worldwide tax consolidation). Because the latter's imperfections can be reduced through the choice of tax rate, the report ultimately recommends a worldwide tax consolidation solution.
Copyright 2011 Edward D. Kleinbard. All rights reserved.
* * * * *
Edward D. Kleinbard is a professor at the University of Southern California Gould School of Law. E-mail: firstname.lastname@example.org
Copyright © 2011 Edward D. Kleinbard. All rights reserved.
* * * * *
A. Scope of Report
This report considers the implications of "stateless income" tax planning for the design of tax systems that address the taxation of returns from corporate foreign direct investment. The bulk of this report compactly summarizes an academic paper, "Stateless Income," that will appear in the Florida Tax Review.1 The report's economic analysis and policy recommendations are extended in a second academic paper, "The Lessons of Stateless Income," to appear in the Tax Law Review.2 Because many Tax Notes International readers are familiar with the relevant international tax rules, this report does not summarize their operation. The longer papers supply more background.
Tax Notes International has published many reports on subpart F, territorial tax systems, and large-scale proposals to change current U.S. tax law governing outbound foreign direct investment. The principal contributions of this report are its development of the theme of stateless income, its analysis of the consequences of that phenomenon for standard policy prescriptions, and its attempt to combine practical insights with current economic efficiency theories for the design of international tax systems.
B. Overview of the Policy Challenges
I use the term "stateless income" to mean income that a multinational group derives from business activities in a country other than the domicile (however defined) of the group's ultimate parent company but that is subject to tax only in a jurisdiction that is neither the source of the factors of production through which it was derived nor the domicile of the group's parent company.3 Stateless income can be understood as the movement of taxable income within a multinational group from high-tax to low-tax source countries without shifting the location of externally supplied capital or activities involving third parties. Stateless persons wander a hostile globe, looking for asylum; by contrast, stateless income seeks any of several zero- or low-tax jurisdictions, where it finds a ready welcome.
The techniques used to generate stateless income rely on norms embedded in virtually every tax system. As a result, stateless income is extremely difficult to curb. At the same time, its availability distorts U.S. firms' investment decisions -- such as the relative attractiveness of situating marginal investments in foreign jurisdictions or in the United States -- and erodes the domestic U.S. tax base.
The fundamental and widely shared international income tax norms that enable stateless income include the recognition of the separate tax personas of different juridical persons, even when commonly owned; the deductibility of intragroup interest, rents, and royalties; the freedom of a multinational enterprise to deal with a subsidiary as an independent actor for arm's-length contracting purposes; and the freedom to situate the economic rents attributable to unique business opportunities in a low-taxed affiliate. Stateless income also flourishes because of nations' collective failure to develop effective source rules for income derived from intangible assets in particular.
Multinational firms get at least two bites at the stateless income apple. First, they can situate in a low-tax jurisdiction returns from factors most plausibly situated in high-tax countries (for example, sales to local customers) by relying on the norms of freedom of contract within the group, the purportedly arm's-length nature of arrangements reached by a parent company and its wholly owned subsidiary (freshly capitalized by the parent), and ambiguities in the international consensus rules surrounding the source of returns to intangible assets. Second, multinational firms can use earnings-stripping strategies to move income tentatively situated in a jurisdiction with the most plausible claim to be the source of that income to another (low-tax) jurisdiction, typically through the creation of an item of intragroup deduction/income inclusion (for example, intercompany interest, rents, or royalties). That second-stage earnings-stripping strategy need not have any nexus to the generation of the income.
As used in this report, however, the term is reserved for strategies to reduce high-tax source country income. The term thus represents more than the problem of residence country base erosion through aggressive transfer pricing, although that activity plainly exacerbates its magnitude in practice.4 In practice, however, the policy recommendations this report makes respond to both issues, for two reasons. First, the technologies used in source and residence country base erosion overlap. Second, the report's ultimate goal of outlining a helpful approach to cross-border taxation in light of the stateless income phenomenon implicates the familiar question whether that proposed approach distorts investment decisions as between source and residence countries.
Stateless income also is not the same as "capital mobility." As traditionally understood, capital mobility involves a person's ability to locate real investments or third-party activity with a view to minimizing the resulting tax burden; it is "the elasticity of supply of a location-denominated factor with respect to its net [after-tax] reward in that location."5 Stateless income, by contrast, is the movement of taxable income within a multinational group without substantially determining how a firm does business with third-party customers.
Stateless income tax planning is divorced from actual market transactions. Instead, it undercuts the functions of markets in setting market-clearing after-tax returns on capital investments. If one accepts the premise that after-tax returns on business income converge on a single worldwide level,6 tax returns must diverge, with commensurately higher pretax returns in high-tax countries. Stateless income tax planning permits advantageously situated multinational firms to earn high-tax source country pretax returns and then to migrate those returns to a low-tax jurisdiction, thereby capturing low-risk supranormal returns, which this report labels "tax rents."
If left unchecked, a firm's ability to capture tax rents can lead it systematically to prefer even marginal investment abroad when compared with a competing domestic marginal investment.7 But stateless income creates other important challenges for U.S. tax policy as well. Those include reduced coherence to the concept of geographic source; the systematic bias toward offshore rather than domestic investment; the more surprising bias in favor of investment in high-tax foreign countries to allow for the generation of low-tax foreign income in other countries; the erosion of the U.S. domestic tax base through debt-financed tax arbitrage; many instances of deadweight loss; and, essentially uniquely to the United States, the exacerbation of the "lockout" phenomenon under which the price that U.S. firms pay to enjoy the benefits of dramatically low foreign tax rates is the accumulation of extraordinary amounts of earnings (an estimated $1.4 trillion or more8) and cash outside the United States.
As explained below, one policy implication that is inconsistent with the data is that current law disadvantages U.S. multinational firms regarding the effective foreign tax rates they suffer compared with their territorial-based competitors. Whether those tax burdens are measured by reference to actual cash taxes paid or to the financial accounting statements that are the lens through which shareholders and other stakeholders view publicly held firms, many U.S. multinational firms today have global effective tax rates closely comparable to those of their foreign-based competitors. Indeed, the most adroit U.S. firms have been so successful in stateless income tax planning that they end up being hoist with their own petard. They have removed so much income from their tax bases, both in the United States and in high-tax foreign jurisdictions, that they are running out of remotely feasible ways of reinvesting the huge sums accumulating in their low-tax subsidiaries.
Stateless income imposes substantial welfare costs such as the well-known lockout effect, whereby U.S. firms must leave earnings (and cash) in foreign subsidiaries to retain the benefits of stateless income. But it is not clear that these costs create any significant "competitiveness" problem for U.S. multinationals or lead to direct reductions in U.S. employment. Either of these hypotheses would require a showing that U.S. multinational firms are capital constrained in the United States, so that rational investment opportunities here cannot be pursued. But there is little evidence of that. Instead, the principal social costs of lockout seem to be that lockout in practice functions as a kind of "lock-in," in which shareholders cannot extract from firms the earnings and cash locked away to preserve the firms' stateless income results.
C. The Double Irish Dutch Sandwich
Recent news stories on the internal tax planning of U.S. firms like Microsoft, Forest Laboratories Inc., and Google Inc. have injected drama into the narrative by providing useful insights into how firms generate stateless income in practice. This section uses Google's "Double Irish Dutch Sandwich" structure to illustrate how stateless income tax planning relies on deeply embedded global tax norms and how it operates to disassociate taxable income from any connection with any location in which the value-adding activities that generated that income could plausibly be said to lie.9 The same story could be told of many other U.S. multinational firms.10
In 2003, a few months before its initial public offering, Google entered into a cost-sharing agreement with a newly organized wholly owned Irish subsidiary, Google Ireland Holdings (Ireland Holdings), under which Ireland Holdings acquired the rights to Google's search and advertising technologies and other intangible property for the territory comprising Europe, the Middle East, and Africa (EMEA). Google commenced its Irish operations in 2003 with five employees.11
Ireland Holdings made an undisclosed buy-in payment for rights to the technologies as they then existed, and it appears to have agreed under a cost-sharing agreement to bear future development costs in proportion to the size that the EMEA market bore to the worldwide market for the Google technologies.12 As apractical matter, that buy-in payment likely reflected in part the then-market capitalization of Google (which in turn would have been a good proxy for the value of its intangible assets). That value in turn presumably was much smaller than the value that might have been inferred post-IPO.13 Regardless, in 2006 Google negotiated an advance pricing agreement with the IRS that accepted the bona fides of the 2003 buy-in payments for the then-existing intangibles. The terms of the APA (like all such agreements) are not public.
The Google structure immediately after the company entered into the cost-sharing agreement is represented schematically in Figure 1.
Figure 1. Google Double Irish Dutch Sandwich 1 -- 2003
In a sense, the most remarkable aspect of the entire structure is shown in this schematic. It is countries' ready acceptance of the fantastic notions that (1) a wholly owned subsidiary has a mind of its own with which to negotiate arm's-length contractual terms with its parent; (2) capital provided to the subsidiary by the parent somehow becomes the property of an independent actor (the subsidiary) with which it can take business risks that for tax purposes are not simply assimilated into those borne by the parent (as both provider of the capital and ultimate economic owner of the assets acquired therewith); and (3) a multinational enterprise that exists as a global platform to exploit a core set of intangible assets is best analogized to wholly independent actors taking on limited and straightforward roles in a vertical chain of production or a horizontal array of product distribution. The second and third of these notions, in particular, transcend the question of transfer pricing -- in the second case, because of the international tax norm that equity owners are not required to include in income any minimum current return on their investment, and in the third case because the global assets and synergies that a multinational group exploits are attributes of the group as whole, not any one member.
Within a few years, the structure had morphed. First, Ireland Holdings had become a dual resident company -- that is, for U.S. tax purposes it remained an Irish corporation (because that is its place of incorporation), but for Irish tax purposes it became a resident of Bermuda (because that is where its "mind and management" are centered). Second, Ireland Holdings had put the EMEA rights to the core technologies to work by licensing them to a subsidiary organized as a Dutch company (Google BV), which in turn had licensed them to a lower-tier subsidiary, Google Ireland Limited (Ireland Limited). Ireland Limited licenses the technologies throughout the EMEA territories and collects billions of dollars in advertising revenues from the use of those technologies.
Presumably, Google BV and Ireland Limited have each checked the box14 -- that is, made a special election relevant only for purposes of U.S. tax law not to be characterized as a corporation. Because each has a single owner and has elected not to be regarded as a corporation for U.S. tax purposes, each is treated as a disregarded entity -- a tax nothing -- for U.S. purposes, but continues as a juridical person for all non-U.S. tax purposes. Here one can see another fantastic element of international tax planning. Through a simple tax return election, a company can disappear from view for purposes of U.S. law, while remaining relevant for purposes of all other fiscal systems, thereby facilitating myriad tax system arbitrage opportunities.
Ireland Limited today has about 2,000 employees; it is unclear how many of them are engaged in the sale and marketing of Google products in the EMEA territory and how many are working as engineers in developing extensions of those technologies.15 Technically, a foreign subsidiary can perform its obligations under a cost-sharing agreement by hiring affiliates to do the work, using capital provided by the parent to pay those affiliates until it generates its own revenues. Again, one sees at work the fantastic idea that a subsidiary has both capital and an appetite for risk that can be separated from those of its parent.16
Figure 2. Google Double Irish Dutch Sandwich II -- 2009
The structure can be summarized in Figure 2.
Now the full stateless income generation machine can be seen. Income earned from the use of the Google intangibles by customers (or, to the extent relevant, affiliates) in high-tax countries streams directly to Ireland Limited as a component of Ireland Limited's advertising fees, without bearing source country tax, because the fees paid are deductible in the source country.17 While much of Ireland Limited's income presumably comes directly from third-party customers in the EMEA region, the same sort of structure can be used to strip out income from local affiliates that in turn serve local customers and then to move that income to Ireland. The net effect in either case is that income from the exploitation of the Google intangibles throughout the EMEA region is taxed in this first step only in Ireland. But that is not where the income ultimately comes to rest.
Ireland imposes a 12.5 percent corporate income tax on Irish resident companies. Ireland Limited is therefore subject to that tax rate on its net income, but it makes large deductible royalty payments to Google BV for the use of the core Google intangibles originally transferred in 2003 (and since extended by investments made under the internal cost-sharing agreement). Google BV in turn makes royalty payments almost as large to Ireland Holdings. The latter is a Bermuda company from an Irish perspective, and Bermuda has no corporate income tax.
Google BV exists because royalties paid directly from an Irish company to a Bermuda company (from Ireland Limited to Ireland Holdings) would be subject to an Irish withholding tax.18 That tax does not apply to royalties paid to a company resident in an EU member state, even one that is an affiliate and that apparently serves no purpose except the elimination of Irish withholding tax. The Netherlands does not impose withholding tax on the outbound royalties paid to Ireland Holdings, and it collects only a small tax (essentially a fee for the use of its tax system) on the modest spread between the royalties Google BV receives and those it pays on to Ireland Holdings. (It is normal in Dutch tax practice to negotiate this sort of spread in advance with the Dutch tax authorities.)
From a U.S. tax point of view, neither Ireland Limited nor Google BV exists at all. The United States sees only an Irish (not Bermudan) company (Irish Holdings) with a Bermuda branch where most of its net income comes to rest. The result is a near-zero rate of tax on income derived from customers in Europe, the Middle East, and Africa that is attributable to the high-value intangibles that encompass the bulk of Google's economic factors of production, and a very low rate of tax on returns attributable to the services of Google's Irish-based sales force.
This stateless income generation machine is referred to as a "Double Irish" structure because of the use of the two Irish firms. The "Dutch Sandwich" sobriquet follows from the insertion of Google BV as a sort of tax filler between the two Irish firms. The structure is easily replicable by others (and in fact has been reported to be widely used among U.S. technology firms19); there is nothing in it that relies on any unique business model or asset of Google's. From the perspective of sophisticated U.S. multinational firms, this arrangement is simply one tool among many in the stateless income planning tool kit.
II. An Ersatz Territorial Regime
A. The U.S. Tax System in Practice
The U.S. tax system is conventionally described as using a worldwide tax base, with the important exception that the net income, but not the net loss, of a foreign subsidiary generally is includable in the taxable income of its U.S. parent company only at some indefinite future date. (A true worldwide system would consolidate for tax purposes the operations of foreign subsidiaries with those of the parent company so that, for example, foreign losses could offset domestic income.) This is a false picture of the U.S. tax system in operation.
It is more accurate to say that in practice and as used by sophisticated multinational firms, the U.S. tax system operates as an ersatz territorial tax regime, with two twists.20 First, some extraordinary (that is, much larger than normal) repatriations of overseas profits to the U.S. parent are subject to U.S. taxation. As a result, the system strongly discourages extraordinary repatriations. Second, untaxed foreign income paid by foreign subsidiaries to the U.S. parent in the form of interest or royalty payments can be sheltered from U.S. tax through the use of unrelated foreign tax credits (which would not be the case in a well-designed territorial regime).
The United States of course fundamentally deviates from a worldwide tax norm by offering U.S. firms the opportunity for deferral, under which the active business earnings of a U.S. company's foreign subsidiary (but not a foreign branch) are not taxed in the United States until they are somehow repatriated to the U.S. parent.21 As the next section discusses, the residual tax the United States collects on repatriated income is surprisingly small.
As a result of deferral, the United States retains only a residual claim to tax the active business earnings of foreign subsidiaries, when that income is somehow made available to the U.S. parent, and then after allowable FTCs are claimed. The practical consequences of the deferral principle are dramatic. The accumulated earnings of foreign subsidiaries of U.S. resident parent companies that have not been taxed by the United States today exceed $1 trillion, after net extraordinary dividends in 2005 of about $312 billion in response to the one-year repatriation tax holiday offered by section 965.
Interest expense incurred by a U.S. corporation is fully deductible, but if it arises from debt that is deemed to support foreign assets, it is treated as derived from foreign sources.22 The net effect of these interest allocation rules is to reduce a U.S. firm's foreign income solely for U.S. tax purposes, while leaving unaffected its actual foreign tax liability. That result is thought to limit a firm's willingness to incur debt in the United States to fund foreign equity investment.
In practice, however, as long as a U.S. firm does not drive its effective foreign tax rate above the U.S. statutory rate after taking these interest expense allocation rules into account, the rules have no effect on the U.S. firm's tax liability. As a result, firms that successfully reduce their foreign tax bills through stateless income planning have a substantial ability to incur U.S. interest expense without harming their ability to use FTCs.23
The U.S. FTC, deferral, and subpart F rules interact in complex ways that often are underappreciated by analysts of the system. Critically, a U.S. firm can choose to defer or repatriate income from its foreign subsidiaries on a subsidiary-by-subsidiary basis. The FTCs that flow up to the U.S. parent in turn depend on the foreign tax burdens imposed on the subsidiary whose income is repatriated (which income in turn is calculated under U.S. principles). Moreover, FTCs are not linked to a specific item of income. Thus, excess credits from one item of income (that is, foreign tax imposed at a rate greater than the U.S. tax rate on that item of income) can be re-deployed to offset tentative U.S. tax on unrelated low-taxed foreign-source income.
A sophisticated U.S. firm manages the residual U.S. tax on repatriated foreign earnings by manipulating the complex interactions between the U.S. deferral and FTC rules in a manner that can be analogized to a tax distillery. The firm's tax director functions as the master distiller, confronted by hundreds of casks of foreign income, one for each category of income earned by each foreign subsidiary. Each cask sits waiting to be tapped by the master distiller as needed, and each dram of foreign income drawn from a cask brings with it a different quantum of FTCs. The master distiller takes instructions from the CFO as to how much cash must be repatriated to the United States each year, and he then sets about perfecting a blend of income and credits so that the residual U.S. tax on the resulting liqueur is as small as possible.
Through careful tax planning, the tax director can replenish the casks of high-tax and low-tax foreign income while keeping untapped income offshore and waiting to be drawn down as needed. The result, as summarized above, is a very low effective U.S. residual tax rate on regular repatriations to the United States. At the same time, the operation of the distillery tends to drive down the effective foreign tax rate associated with unrepatriated foreign earnings, because the purpose of the distillery is to strip out from indefinitely deferred foreign earnings all the FTCs that are needed to offset current repatriations of zero-taxed or low-taxed foreign income.
The typical corporate tax distillery is built to handle a specified maximum annual throughput of foreign income and associated FTCs. If business exigencies were to call for a very large repatriation in one year, the tax director's distillation apparatus would be overwhelmed and a substantial residual U.S. tax liability would be incurred. Thus, the right way to see the U.S. rules for taxing income from foreign direct investment as they apply to ordinary course operations is as a de facto territorial tax system, with a contingent (and firm-specific) residual tax liability associated with large-scale repatriations.
B. Current System's Revenue Collections
As a result of the interactions of the complex rules summarized above, the United States today imposes a very small cash tax burden on foreign income that is repatriated to the United States. In 2004 the United States collected $18.4 billion in tax from the foreign operations of U.S. multinationals.24 That figure includes not only taxes on dividends paid by foreign subsidiaries, but also subpart F income and interest and royalty income paid from controlled foreign corporations to U.S. affiliates. Yet in 2004 foreign subsidiaries of U.S. firms paid $47 billion in dividends to their U.S. parents, generated $48 billion in subpart F income taxable to U.S. owners, paid another $59 billion in royalties, and $12 billion in interest to U.S. affiliates -- altogether some $166 billion in total repatriations out of foreign earnings.25 The $18.4 billion in U.S. tax collections represents a U.S. tax rate of about 11 percent on that repatriated income.
At the same time, profitable foreign subsidiaries of U.S. firms (those subsidiaries that reported positive income for the year) retained outside the reach of the U.S. tax system $270 billion of net earnings for the year after foreign tax (and after dividends to the United States).26 As described below, the average effective foreign tax rate on those retained earnings was roughly 16 percent.
One important consequence of the design of the U.S. FTC rules is that royalty and interest payments received by U.S. affiliates from foreign subsidiaries today are both large and partially tax-free everywhere in the world, which is not the case under properly constructed territorial tax systems. These items bear little tax when they are received in the United States, because they generally are deductible in the source country and are in turn sheltered from tax in the United States through the blending of high-tax foreign income from other sources to shelter these zero-taxed items.27
For example, if the United States had used a standard territorial tax system in 2004, it would have collected a modest amount of tax on the $95 billion of dividend and subpart F income actually or constructively repatriated in that year.28 However, it would have collected roughly $25 billion (35 percent of $71 billion) on the royalty and interest income received by U.S. firms from their foreign subsidiaries -- some $6.6 billion more than it actually collected under the entire current "worldwide" system.
C. Arbitrage and Domestic Base Erosion
The U.S. regime for the taxation of foreign direct investment not only collects little revenue (and as suggested above, possibly less than what would be collected under some scenarios if the United States were to switch to a well-designed territorial system), but also exposes the U.S. corporate tax on the domestic tax base of U.S. multinationals to systematic erosion through straightforward tax arbitrage strategies.
Current law has the pernicious effect of implicitly encouraging domestic leverage to fund a firm's domestic cash needs while leaving low-taxed foreign earnings abroad. This strategy allows U.S. multinational firms to operate in a quasi-territorial tax environment by supplying the U.S. parent company with cash to fund its domestic operations from two sources: the low-taxed stream of regular course foreign operations (as described above) and domestic borrowings. The attendant increase in domestic interest expense in turn is allocated in part against foreign operations for purposes of the FTC limitation rules described earlier. Nonetheless, as long as the firm's foreign earnings are sufficiently low-taxed before taking into account the increase in the firm's foreign effective tax rate from the application of those expense allocation rules, the limit is not binding.
A U.S. multinational firm's systematic use of domestic borrowing to replicate the cash flow advantages of other firms in territorial regimes (where foreign earnings can be repatriated costlessly) erodes the U.S. corporate tax base, because the firm's interest expense is deductible in the United States while the foreign earnings are not included. The combination of deferral with stateless income planning and incomplete domestic expense allocation rules -- which often are not binding -- thus leads to classic tax arbitrage, no different in character than if taxpayers could incur tax-deductible interest to invest in uncapped IRA accounts.
III. How Large Is Stateless Income?
There is powerful evidence that multinational firms substantially reduce their aggregate worldwide tax burdens -- and thereby increase their incentives to retain earnings outside the United States -- through stateless income planning. Because that is clear to anyone working in the field, this subsection only briefly reviews some of that evidence, looking at both "cash" tax liabilities (the tax liabilities shown as due on the taxpayer's actual tax returns) and financial accounting data. The evidence strongly indicates that U.S. firms are operating in a tax environment not very different from that of foreign competitors in territorial tax systems.
A. Cash Tax Liabilities
The exact tax liabilities of U.S. multinational firms are confidential, because corporate tax returns, like individual ones, are not released to the public. Fortunately, the IRS Statistics of Income division publishes tax data on CFCs biennially. The data include the aggregate earnings and profits of all CFCs having positive E&P for the year and the foreign income taxes paid or accrued by these profitable foreign companies for that year.29
The SOI public data for 2006 (the most recent year for which data have been released) show that CFCs with positive earnings in that year had pretax E&P of $587.8 billion. Those firms paid or accrued foreign income taxes of $96.6 billion for that year.30 The data therefore suggest that U.S. CFCs actually paid or accrued foreign taxes for their 2006 economic income at an effective rate of 16.4 percent. The same figure for 2004 was comparable, at 15.7 percent. To put that 15.7 percent effective foreign tax rate in context, the Government Accountability Office calculated that for 2004 the weighted average U.S. domestic effective tax rate for large profitable U.S. corporations was 25.1 percent; the median was 31.8 percent.31
Some U.S. multinationals are fortunate enough to have foreign tax rates materially lower than the 16 percent average effective tax rate. For example, Microsoft Corp.'s financial statements in its 2010 annual report indicated that it has $29.5 billion in permanently reinvested earnings outside the United States (that is, after-foreign-tax earnings of foreign subsidiaries that the company does not currently intend to repatriate to the United States).32 Microsoft also noted that the tax cost of repatriating those earnings to the United States would be $9.2 billion.33 These numbers suggest that Microsoft's permanently reinvested foreign earnings had an effective foreign income tax rate in the neighborhood of 4 percent.34
Extremely low effective foreign income tax rates like Microsoft's theoretically could be explained if most countries had commensurately low statutory corporate income tax rates. But that is a false hypothesis. Working with firm-specific confidential Treasury data, Treasury Department economist Harry Grubert and Prof. Rosanne Altshuler calculated that for 2002, U.S. multinational firms faced an average foreign statutory tax rate of 29 percent, weighted by the firms' foreign incomes.35 Koninklijke Philips Electronics NV (Philips), a major Dutch multinational industrial group, reported in its 2007 annual report that the weighted average statutory tax rate of all the jurisdictions where it did business was 26.9 percent.36
Royalty and Interest Paid by CFCs
Rents, Royalties, and License Fees Interest
Paid to U.S.- Paid to U.S.-
Year Related Parties Paid CFC to CFC Related Parties Paid CFC to CFC
2004 $59,275,141,484 $13,489,657,755 $12,419,547,764 $42,039,808,030
2006 $66,719,388,821 $12,659,524,687 $25,139,162,746 $67,012,282,063
Source: IRS, SOI division, by electronic communication with the author in
response to a request for information.
There is strong circumstantial evidence of stateless income tax planning in the high interest and royalty payments made by U.S. firms' foreign subsidiaries (technically, CFCs) to other foreign subsidiaries. Table 1 details the data for 2004 and 2006, as prepared by SOI.
As can be seen, in 2006 CFCs of U.S. parent firms made some $80 billion in (presumptively) deductible royalty and interest payments to other CFCs. This sum vastly understates the actual amount of those payments because it ignores payments by a disregarded entity -- a subsidiary of a CFC that for U.S. tax purposes is treated as having no separate juridical existence, but that is very much alive and counted as a company for local tax purposes. The Google facts described earlier are a real-life example of enormous (presumably, multibillion-dollar) royalty streams among foreign affiliates of a U.S. multinational group that work to accomplish stateless income goals but are invisible for U.S. tax purposes.
Another indicator of widespread stateless income tax planning is the role that a few very low-tax jurisdictions, such as Bermuda, the Cayman Islands, Ireland, and Switzerland, play in the foreign effective corporate income tax rates of U.S. firms.37 This concentration of U.S. multinational firms' reported incomes in a few relatively small foreign economies whose only common feature is their low tax rates belies the notion that U.S. firms' low effective foreign tax rates in the 2002-2006 period were attributable simply to tax preferences that were generally available in countries with high nominal rates.
The results of stateless income tax planning are that at the end of 2008, the foreign subsidiaries of U.S. firms held more than $1 trillion in retained low-taxed earnings (net of the $312 billion in special dividends that qualified for the one-year repatriation holiday afforded by section 965). To the same effect, Grubert found in a recent study that from 1996 to 2004 (the period immediately preceding the one-time repatriation tax holiday), the share of U.S. firms' worldwide income that was retained by foreign subsidiaries each year climbed from 17.4 to 31.4 percent.38
B. Financial Accounting Evidence
A firm's current cash tax liabilities often are not as important as its audited financial accounting statement provisions for taxes, because U.S. generally accepted accounting principles are the lens through which investors judge public firms.39 Indeed, investors have little choice in the matter: A firm's U.S. corporate income tax return is confidential, while GAAP financial statements of publicly held firms are not. And here again one sees evidence that U.S. multinational firms enjoy very low effective foreign tax rates that can logically be explained only through stateless income tax planning.
GAAP accounting for taxes is an odd mixture of different concepts. Generally, the tax reconciliation table in a firm's tax footnote to its financial statement is used to calculate a hypothetical tax burden equal to the statutory rate (35 percent) applied to the GAAP (not tax) measure of income. Differences between the actual U.S. tax burden and this hypothetical figure must be accounted for, either as temporary differences (for example, differences in depreciation accounting conventions) or as permanent differences (for example, irreversible differences between the GAAP and tax accounting measures of income, such as tax-exempt bond interest income). Temporary differences are reflected on the financial statements as deferred tax liabilities (when tax deductions run ahead of the corresponding GAAP measure of an expense) or deferred tax assets (the converse). Permanent differences are reflected as a reduction in the firm's tax expense, and therefore its effective tax rate.
Examples of material permanent differences are relatively scarce, with one principal exception: GAAP does not require any deferred tax liability to be established for the contingent residual U.S. tax liability that might be incurred on the repatriation of "permanently reinvested" low-tax foreign earnings.40 A better term for this might be "indefinitely reinvested" -- as long as a firm can demonstrate that it has no plan to repatriate foreign income and no need to do so, it doesn't have to provide for the potential liability for it on its GAAP financial statements.41 The permanent reinvestment decision can be made on a subsidiary-by-subsidiary basis, and the decision to treat accumulated foreign earnings as not permanently reinvested can be reversed (and an accounting benefit booked) in appropriate circumstances.
These rules mean that low-taxed permanently reinvested earnings bring down a firm's GAAP tax expense. They also mean that firms that defer the repatriation of active foreign earnings are not penalized relative to competitors in territorial systems, when viewed through the lens through which investment decisions ordinarily are made.
Two recent complementary empirical studies confirm the intuitive heuristic that GAAP accounting for taxes on foreign earnings significantly affects the repatriation decision. In one, Jennifer Blouin, Linda Krull, and Leslie Robinson, working with confidential Bureau of Economic Analysis data, conclude that "our empirical tests tell a consistent story; [GAAP] reporting incentives [for permanently reinvested earnings] deter the repatriation of foreign earnings."42 In the other, John Graham, Michelle Hanlon, and Terry Shevlin report the results of an extensive survey of firm tax executives; the authors conclude that "the ability to not recognize the U.S. income tax expense on foreign earnings in financial statements . . . is an important consideration in real corporate investment decisions regarding location of operations and whether to repatriate foreign earnings to the U.S. or reinvest the foreign earnings overseas."43
Some studies have suggested that the market in fact discounts stock prices for the U.S. residual tax that firms actually disclose in their financial statements as estimates of the cost of repatriating their permanently reinvested earnings.44 Even if the market does discount these stocks, recent corporate practice seems to tilt in favor of not quantifying estimated repatriation tax costs. For example, out of the 30 constituent members of the 2010 Dow Jones industrial average, only three disclosed their 2007 estimated tax costs to repatriate their permanently reinvested earnings.
In sum, from the perspective of investors, the U.S. global tax regime often operates much like a territorial system. For example, in 2007 (the last year before the financial crisis) the effective GAAP tax rate for the global operations of General Electric Co. and its GAAP-consolidated subsidiaries was 15.1 percent.45 (This means, of course, that GE's effective foreign income tax rate for the year was far lower, since the 15.1 percent figure represents an average of foreign and U.S. income tax rates on their respective proportions of firm income.) The non-inclusion of any GAAP liability for U.S. taxes on foreign operations accounted for 15.2 percentage points of the difference between the statutory rate of 35 percent and the reported global tax rate of 15.1 percent.46
IV. Policy Implications of Stateless Income
A. The Fruitless Search for Source
The artificiality of the global norms that define the source of income is a well-known problem, and solutions are not obvious.47 But territorial tax systems require their resolution, because the source rules that are adopted determine which jurisdiction has the right to tax the income. Source rules thus are central to the operation of territorial tax systems. They also are important for the U.S. tax system, although they do not play the central role they do in territorial regimes, because source rules drive a U.S. taxpayer's ability to claim FTCs.
Stateless income tax planning compounds the meaninglessness of income tax source rules. Even if a multinational enterprise's income is sourced in the first instance by every country according to some economically rational set of agreed principles, stateless income tax planning simply extracts the income from the source country (for example, through deductible interest, royalty, or fee payments) and deposits it in a more tax-friendly locale. Google's income from sales to German advertisers is deducted from German income tax returns, while Google Ireland has no permanent establishment in Germany to which that income is attributable. As a result, Google's income from providing advertising services in Germany effectively is untaxed in Germany. That income is sourced in the first instance to Ireland as the domicile of the putative owner of the intangible assets that give rise to the advertising income. And then, in a second step unrelated to the wisdom of the first-level source rule, that income migrates to Bermuda via the Double Irish Dutch Sandwich mechanism described earlier.
The result is that in a world imbued with stateless income tax planning, there can be no meaning at all to source, because transactions one or more steps removed from a firm's original value-adding operation redirect that income to friendlier locales. The efforts to date devoted to clarifying source rules largely overlook how these second- or third-step internal transactions -- all perfectly consistent with arm's-length standards and other bedrock global tax norms -- significantly erode the value of those principles.
B. Capture of Tax Rents
Global capital markets are liquid and efficient, and many countries have eliminated or greatly reduced barriers to foreign investment in their local economies. Moreover, for most direct and portfolio investment, source country net income tax is effectively the final tax on cross-border investment income. As a result, economists generally expect that global after-tax returns on corporate marginal investments will converge, because foreign and local investors will provide capital to jurisdictions in which after-tax marginal returns exceed world norms, and they will withdraw capital from those in which returns are below normal.48 But corporate income tax rates differ around the world, which means that pretax marginal returns necessarily must diverge if after-tax returns do not.
Stateless income tax planning offers multinational firms, but not wholly domestic ones, the opportunity to convert high-tax source country pretax marginal returns into low-tax country inframarginal (supranormal) returns, by redirecting pretax income from the high-tax country to the low-tax country. Multinational firms can thus be said to capture "tax rents."49 Their inframarginal returns stem not from some unique high-value asset, but from their unique ability to move pretax income across national borders.
For example, assume that the United States has a corporate tax rate of 35 percent, Sylvania's tax rate on domestic income is 25 percent, and Freedonia imposes a 10 percent tax rate on domestic income. Moreover, because capital is globally mobile and capital markets are efficient, after-tax normal returns on capital invested in business firms are the same around the world. Assume this global after-tax rate is 5 percent. What this implies is that pretax normal corporate returns will vary from country to country to reflect differences in tax burdens. Pretax corporate returns in the United States will be 7.7 percent, while in Sylvania they will be 6.67 percent, and in Freedonia 5.56 percent.
A U.S. firm, faced with earning a 5 percent after-tax return on a marginal investment, will opt instead to invest not in low-tax Freedonia but in high-tax Sylvania, and then through stateless income tax planning will move the Sylvanian pretax 6.67 percent return to Freedonia. After Freedonian income taxes on that 6.67 percent marginal return, the U.S. firm will enjoy an after-tax marginal return of 6 percent, rather than the global prevailing 5 percent rate. The incremental 1 percent return that comes without any incremental risk is an example of tax rents.
At least as applied to U.S.-domiciled companies, tax rents are easier to harvest from foreign jurisdictions than they are from a multinational firm's country of residence.50 U.S. firms prefer marginal investments in foreign high-tax countries to investments in the United States because the former are more easily used in stateless income planning. The income is already foreign source, and straightforward earnings stripping technologies that are unavailable for domestic income can be used to move that foreign-source income to a low-tax affiliate.
The best counterargument is that capital, like nature, abhors a vacuum, and that foreign investors will replace domestic firms as investors in the U.S. domestic markets.51 But this argument confuses U.S. investment with U.S. taxable income.52 To a foreign-domiciled multinational firm, the United States is just another source country, and a particularly high-tax one at that. Thus, it may be that foreign multinational firms replace any missing U.S. investment, but the empirical issue goes beyond that. The analysis must also consider whether foreign firms are themselves unschooled in the art of stateless income planning when the United States is the source country. Despite some statutory protections against earnings stripping, there are thus good reasons to believe that the United States is a net loser when its tax system encourages domestic firms to invest disproportionately outside the country and (as the next subsection discusses) to finance domestic cash flow needs through U.S. borrowings that erode the U.S. tax base.
In sum, the net effect of the tax rents phenomenon is an odd incentive for U.S. firms to invest in high-tax foreign countries to provide the feedstock that the stateless income generation machine will process into low-taxed permanently reinvested earnings. The tax rents that are generated are retained outside the United States to preserve their value.
This last point, when combined with the arbitrage possibilities described in the next subsection, effectively answers the question often posed by the private sector as to why the United States should care if U.S.-domiciled multinational firms minimize their foreign income tax liabilities. The answer is that the pursuit of tax rents, combined with the erosion of the domestic tax base through leverage, leads to distorted investment decisions by domestic firms and sharply reduced domestic tax revenue collections.
C. Domestic Base Erosion: Tax Arbitrage
At the same time that U.S. firms capture tax rents through stateless income tax planning, they meet many of their funding needs (including dividends and stock repurchases) through domestic U.S. borrowing. The resulting interest deductions erode the U.S. corporate tax base through a classic tax arbitrage operation in which the inframarginal returns on offshore investments fall outside the U.S. tax net while interest expense is deducted on debt that might not be incurred if those returns were repatriated (and the income included in the U.S. tax base). The U.S. tax base is shifted outside the United States through domestic leverage incurred to support foreign earnings, genuine foreign earnings in turn migrate to low-tax locales, and those low-taxed foreign earnings can compound free of U.S. tax indefinitely.
This arbitrage operation is not a theoretical abstraction. At the end of its quarter ending December 31, 2010, Microsoft had $29.5 billion in permanently reinvested earnings and worldwide held $41 billion in cash and short-term investments. In February 2011 it borrowed $2.25 billion in the U.S. capital markets. It was recently reported in the financial press that Microsoft issued this debt to fund dividends and stock buybacks to avoid any repatriation tax, because 80 to 90 percent of its cash and short-term investments are held outside the United States.53 The article also suggested that this pattern is becoming more common among U.S. technology companies generally.
As previously mentioned, the code contains a few special rules -- in particular, the FTC interest expense rules -- whose nominal purpose is to limit this arbitrage. In practice, however, these limitations often do not constrain the full deductibility of U.S. interest expenses, again in part because of the work of the tax director as master blender of the tax distillery, in this case by bringing back enough very low-taxed foreign income (including interest and royalty income) to offset the allocation of U.S. expense.54 Given that the United States has comparatively high corporate statutory tax rates, it would be extraordinary to think that U.S. firms, having successfully captured tax rents through the operation of their stateless income mechanisms, would not complete the tax minimization circle by funding their global cash needs through U.S. domestic borrowings. The net effect is to compound the benefits of stateless income tax planning by migrating to stateless status (through domestic interest deductions) what would have been U.S. taxable income.
D. Capital Ownership Neutrality Theory
Since its inception, the debate on how best to tax income from foreign direct investment has been dominated by the search for the tax equivalent of an alchemist's stone -- more specifically, some sort of economic efficiency benchmark by which competing tax structures could easily be evaluated and the optimal one selected. The well-known result has been a stream of competing neutralities, with names like "capital export neutrality," "capital import neutrality," "national neutrality," and the like. The resulting sterile "Battle of the Neutralities," as I have described the literature, is both tiresome and unproductive.55
Undaunted by the failure of the economics profession to coalesce around a single benchmark, Mihir A. Desai and James R. Hines Jr. in 2003 introduced a new criterion for evaluating international tax systems, which they dubbed "capital ownership neutrality."56 The Desai and Hines article was an important contribution because it brought to the debate a more sophisticated theory of the firm than had been assumed in some earlier work, squarely recast the issues in terms of the United States as simply one country among many, and addressed the reality that the United States (or any other developed country) is simultaneously a capital exporter and a capital importer.
In a nutshell, the Desai-Hines model begins with a (largely implicit) set of assumptions. First, all countries are open economies; as a result, investments can flow freely from one to another. Moreover, the net income from a business is taxed only to the firm earning it and only in the country where the factors of production that generate the income are located. Finally, the identity of the source country is unambiguous. Those assumptions taken together lead to the observation that after-tax marginal rates of return on direct investments are the same everywhere in the world.57
Desai and Hines plausibly argue that, within this model, a territorial tax system is optimal. The reason is that different corporate tax rates around the world will have been capitalized into the prices for business assets and for entire firms; this tax capitalization is the phenomenon that we understand in the domestic context (for example, when analyzing tax-exempt municipal bonds) as implicit taxation.58 For example, if in the United States tax rates are 35 percent and the pretax rate of interest is 10 percent, we might expect that tax-exempt bonds would yield 6.5 percent. An investor who bought a taxable bond would pay a 35 percent actual tax rate on her interest income. If the same investor bought the tax-exempt bond, she would suffer an implicit tax at the same 35 percent rate, reflected in the lower coupon that she was willing to accept to reflect the bond's favorable tax status. In either case, her after-tax return on her investment would be 65 percent of the pretax rate of return.
As applied in the Desai-Hines model, the phenomenon of perfect tax capitalization of prices for business assets and entire firms would mean that, while corporate tax rates around the world might differ, the sum of the explicit and implicit tax burdens on investors would be the same everywhere, as long as a country did not impose any incremental tax on the foreign income of its corporate residents. In other words, source country taxation would be sufficient to ensure that firms had no tax reason to prefer foreign over domestic investment, or vice versa, because the after-tax returns to a marginal investment anywhere in the world would be the same. The addition of residence country taxation would thus simply distort investment decisions, by driving down the after-tax returns on foreign acquisitions below the world market price.
The capital ownership model has intuitive appeal and leads to a conclusion that U.S. multinational firms find extremely agreeable. As a result, it has received a great deal of attention as an intellectually rigorous justification for the United States to adopt a territorial tax system -- indeed, one without any allocation of U.S. expenses against foreign income.
Unfortunately, the stateless income phenomenon completely vitiates the policy relevance of the capital ownership neutrality model. As previously demonstrated, U.S. multinational firms are awash in low-taxed foreign income that has been spirited away from high-tax countries. Because there is no practical meaning to the concept of source, there is nothing on which the tax capitalization mechanism can operate. That is, if all we cared about was the price of a sheep farm in Ireland, it might be the case that the farm's market price would implicitly capitalize Ireland's low tax rate. But once we acknowledge that stateless income planning enables firms to capture tax rents, by moving high-tax source country pretax income to a low-tax jurisdiction, then the entire tax capitalization hypothesis that is the bedrock of capital ownership neutrality falls apart. The adoption of a territorial tax system without elaborate safeguards, including expense allocations, would lead to a systematic preference for foreign over domestic investment, and a systematic elimination of the U.S. domestic corporate tax base through the arbitrage mechanism described above.
E. Comparative Tax Rates
The United States has the second highest federal statutory corporate tax rate among the world's largest economies.59 Relying in part on that fact and in part on their assertion that the United States imposes a worldwide tax on the income of U.S. multinational firms, many U.S. firms have argued that the U.S. tax system makes them "uncompetitive" against foreign multinationals operating with territorial tax regimes.60 The data indicate otherwise.
First, the gap between U.S. and world corporate tax rate norms is sometimes overstated. Many analysts find it convenient to rely on an annual OECD data set for this purpose.61 Using this source, the simple unweighted average of 2010 corporate tax rates among the 30 OECD countries, excluding the United States, was 25.6 percent. (In 2006 it was about 28 percent.62) This data set must be applied cautiously, for three reasons.
First, the data set includes subcentral government taxes on corporate income. This explains why the U.S. rate is said to be 39.2 percent. It is appropriate to include subcentral government taxes when comparing the competitive tax environment of U.S. domestic firms with that of foreign domestic firms or when measuring the foreign tax burden on inbound investment in a country, but it is inappropriate to include U.S. subcentral government taxes when measuring an actual or hypothetical U.S. statutory tax burden on U.S.-domiciled multinational firms contemplating an outbound investment. That is because foreign income generally is not taxed by the U.S. states.63 The right statutory rate comparison in that case is the total (central and subcentral) foreign tax rate to the federal U.S. statutory rate (35 percent).
Second, the simple unweighted average of OECD statutory rates mixes rates imposed by economies of greatly disparate size. In general, here is an inverse relationship between the size of an economy and its corporate tax rate. In 2010 the unweighted average of the five largest OECD economies other than the United States was roughly 32 percent, and the unweighted average of the next six economies was 28 percent.64 Giving equal weight to the smallest 19 economies (where U.S. firms by definition face smaller markets) misstates the tax burdens fairly attributable to a multinational firm's global economic opportunities (if undistorted by stateless income planning).
Finally, the OECD data set does not include non-OECD countries, particularly the BRIC countries -- Brazil, Russia, India, and China. These are very important markets, of course. Their 2010 unweighted average corporate tax rate was 28.25 percent.65
More fundamental to this report, U.S.-domiciled multinational firms do not in fact bear a 35 percent tax burden on their non-U.S. income. The data summarized earlier demonstrate that residual U.S. tax today on actual repatriations is a small fraction of total foreign earnings. At the same time, U.S. multinational firms are able to use stateless income tax planning techniques to drive down their cash foreign tax liabilities and their GAAP financial accounting effective foreign tax rates on unrepatriated earnings to levels far below the foreign tax statutory average.
Taken together, these facts point to many U.S. firms operating in an environment much closer in practice to territorial systems -- indeed, superior to them regarding intragroup interest, royalties, and license fee income. In the same vein, the recent study by Grubert discussed earlier concludes that from 1996 to 2004, there was no meaningful correlation between lower foreign tax rates and the growth rate of U.S. firms.66 From this he concludes that "the importance of low taxes on foreign income for U.S. 'competitiveness' does not, at least on this evidence, have much empirical support."67
The lockout effect refers to the fact that a firm's benefits from stateless income planning are contingent on the firm not repatriating more foreign earnings than its tax distillery can process. Because so many U.S. firms have developed multibillion-dollar pools of low-taxed foreign permanently reinvested earnings, they are compelled as a practical matter to keep a large percentage of their foreign earnings and cash outside the United States solely to avoid this residual tax.
For the reasons described in the preceding subsection, the real tax problem for the managers of U.S. multinational firms that are able to engage in widespread stateless income tax planning is not any U.S. tax burden on retained foreign earnings (whether measured by cash tax liability or financial accounting presentation) or even significant current U.S. taxation of ordinary-course cash repatriations of low-taxed foreign-source income. Instead it is the extraordinary accumulation of profits and cash in foreign subsidiaries and the inability of most firms' tax distilleries to absorb a very large repatriation dividend.68 This distorts behavior (for example, by encouraging firms to borrow in the United States or to make relatively unproductive investments outside the United States) and leads to deadweight loss.
Two recent news stories illustrate the problem. After suggesting (plausibly, in my opinion) that the lockout effect drives U.S. firms to make foreign acquisitions simply because they need some use for the cash they have accumulated outside the United States, one story quotes the CEO of Cisco Systems as saying that "Cisco has $30 billion of its $38 billion in cash parked abroad because of higher U.S. taxes."69 The other story says that eBay has 70 percent of its cash outside the United States and as a result is hunting for acquisitions in Europe.70
The magnitude of the lockout effect's deadweight losses has been the subject of spirited debate.71 As a practical matter, those losses no doubt increase disproportionately the amount of a firm's low-taxed permanently reinvested earnings.72 Nonetheless, it might be helpful, particularly when the competitiveness banners are unfurled in policy debates, to distinguish between the deadweight losses to firm managers for ongoing business activities and the deadweight losses to investors in those firms.
One way to focus on the first question is to ask whether U.S. firms are capital constrained by virtue of having to satisfy their funding needs by particularly costly borrowing in the United States rather than by repatriating cash from abroad. There is little statistical or anecdotal evidence to support such a capital constraint story for the major U.S. multinational firms that account for the bulk of U.S. firms' income from foreign direct investment.73 Many large firms with low effective foreign tax rates have very low debt-to-assets ratios or do not need to borrow at all.74 There is scant evidence that when those firms do borrow domestically, they suffer punitively high borrowing costs.75
The 2004 section 965 repatriation experience implicitly supports the idea that large U.S. multinational firms with substantial permanently reinvested earnings are not capital constrained in the United States. In response to the one-year repatriation tax holiday, U.S. firms repatriated $312 billion in cash dividends in excess of their normal aggregate dividend repatriation rate (about $50 billion/year).76 A subsequent study concluded that this gigantic influx was not correlated with repayments of domestic debt or with incremental investment in domestic property, plant, or equipment (as would be expected if large U.S. multinational firms were capital constrained in the United States). Instead, the repatriation dividends were strongly positively correlated with stock buybacks.77
Perhaps the most that one can say about the costs imposed by the lockout phenomenon on the managers of U.S. firms is that those firms that have been extraordinarily successful in stateless income tax planning have ended up being hoist by their own petard. They have removed so much income from the tax base in both the United States and in high-tax foreign jurisdictions that they now are running out of remotely feasible ways of reinvesting those huge sums accumulating in their low-tax subsidiaries.78
Another way of stating this conclusion is that the lockout effect really operates as a kind of lock-in effect: Firms retain more earnings (in this case overseas) than they can profitably redeploy, to the frustration of their shareholders. The result is not that firms forgo investments but that shareholders are unable to optimize their portfolios. Profits earned by successful multinational firms are retained in relatively low-yielding liquid investments or reinvested in suboptimal foreign acquisitions, all because of their great success in stateless income tax planning on the one hand and the lockout phenomenon on the other. Shareholders would prefer that the cash be distributed to them, but companies cannot afford to comply.
In short, U.S. multinational firms themselves are not disadvantaged materially by the lockout effect but their shareholders are. The ultimate reward of successful stateless income tax planning from this perspective should be massive stock repurchases, but instead shareholders are tantalized by glimpses of enormous cash hoards just out of their reach. There is genuine deadweight loss involved, but it has little to do with "competitiveness" or job creation in the United States.
G. Summary of Implications
Despite their protestations, U.S. multinational firms in fact enjoy substantially all the benefits of their territorial tax competitors, including the opportunity to use stateless income tax planning to capture large tax rents (or to drive down their effective foreign tax rates into the single digits, which is the same thing by another name) -- with one exception. That is the lockout effect, which leads U.S. firms to hold extraordinary amounts of cash equivalents or to make suboptimal investments outside the United States, solely to preserve the efficacy of their stateless income generation machines.
The United States' unique combination of a quasi-territorial tax regime, its enfranchisement of stateless income tax planning through idiosyncratic rules like check-the-box, and the lockout effect leads to deadweight losses, but those losses do not appear to overlap with the usual formulations of "competitiveness" concerns. The U.S. tax system causes U.S.-domiciled multinational firms first to prefer investments in foreign high-tax countries over investments in the United States (to set the stage for stateless income tax generation); second, to establish low-tax affiliates of sufficient size and activity to serve as receptacles of stateless income; third, to invest time and resources in operating the tax planning mechanisms required to create and defend stateless income generation; and fourth, to retain the resulting earnings and cash in those low-taxed receptacles to preserve both the cash and the financial accounting gains inherent in the production of stateless income. The results are distortions in original investment decisions, in the distribution of earnings, and in reinvestments, as well as wasteful expenditures to maintain the apparatus.
The lockout phenomenon generally is the consequence of low effective foreign tax rates and the current law deferral rules. Stateless income tax planning in turn pushes a firm's effective foreign tax rate downward further still. Preserving the benefits of stateless income by accepting lockout distorts firm behavior in welfare-decreasing ways for the simple reason that U.S. multinational firms must find some non-U.S. use for their permanently reinvested foreign earnings, which can distort their investment decisions and hurt the firms' ultimate owners.
The lockout phenomenon also has the pernicious effect of implicitly encouraging domestic leverage to fund cash needs while leaving low-taxed foreign earnings abroad. This strategy allows U.S. multinational firms to compete in a quasi-territorial environment (by preserving the benefits of stateless income tax planning through deferral and financial accounting treatment of those earnings as "permanently reinvested"), but it erodes the U.S. corporate domestic tax base, because the interest expense is deductible in the United States while the foreign earnings are not includable. Deferral, combined with stateless income planning and incomplete domestic expense allocation rules, which often are nonbinding, thus leads to U.S. tax base erosion and the quarantining of much of the firm's cash outside the United States. And for foreign-based multinationals, stateless income tax planning technologies can be applied to the United States as a source country, thereby reducing U.S. domestic tax revenues directly.
V. Where Do We Go From Here?
The taxation of foreign direct investment is a difficult exercise in balancing a wide range of competing constraints: commercial, economic, political, and administrative. There is no magic economic efficiency compass that points to one inevitable answer. We must muddle forward, trying to mitigate as best we can first-order tax-induced distortions in corporate behavior, while at a minimum protecting the domestic corporate tax base.79 If we can collect a little tax revenue along the way without making things much worse along other dimensions, that would not be such a bad thing, either.
Even without a simple benchmark to apply, it is reasonably clear to most observers that the current tax system is unsustainable. It is extraordinarily complex and, as this report has demonstrated, it is easily gamed -- and not just in the sense that aggressive transfer pricing leads to tax leakage. Firms easily can redirect their income from high-tax foreign locations to low-tax ones and then repatriate steady streams of cash at low all-in effective tax costs, by using the many containers, pipes, and valves of the tax distillery to blend low-taxed income (whether from stateless income planning or from interest and royalties received from foreign affiliates) against FTCs harvested from the income left offshore.
The fact that U.S. multinational firms today hold some $1.4 trillion in offshore retained earnings (despite the repatriation of $312 billion in incremental dividends in the 2004 tax holiday season) is just one testament to the magnitude of the tax-induced distortions of current law. Lockout is both a symptom of a failing system and a serious practical problem in its own right.
Ironically, there are two diametrically opposite solutions that resolve the lockout problem. The United States could adopt a territorial tax system, under which income from bona fide foreign direct investment would not be taxed at all; alternatively, it could adopt a true residence-based tax system, under which U.S. firms would be required to consolidate the earnings of foreign subsidiaries. The former resolves lockout by never taxing foreign earnings; the latter does so by taxing them as they are earned. In either case, the decision to repatriate foreign earnings does not add to the U.S. tax bill.
"The Lessons of Stateless Income" wrestles at length with the competing merits of those approaches.80 Each relies on a fundamentally artificial premise, and the choice between them requires a great deal of careful consideration -- not just of economics, but also of the legal constructs and institutions through which tax consequences are defined and enforced.
The core artificiality of territorial taxation lies in postulating that a tax administrator can reliably determine the source of a multinational firm's items of income -- the geographic location of the factors of production that have generated each income item. In the first instance, the inquiry is made virtually impossible by the importance of intangible assets as drivers of a modern multinational firm's profitability and by the problem of assigning to one source or another the income attributable to the internal synergies of the multinational firm synergies, which explain the prevalence of the multinational firm as a business organization. And what little reliable information might be derived from an inquiry into source in the first instance is completely eviscerated by the ability of stateless income tax planning to migrate income from a high-tax location to which there is an economic nexus to a low-tax one where there is none. As this report has demonstrated, the phenomenon of stateless income is pervasive, and that in turn implies the practical intellectual bankruptcy of any concept of source.
In a genuine worldwide tax system, by contrast, the artificiality lies in the conceit that a multinational enterprise is resident in one "home" country, in a manner analogous to an individual's citizenship in a particular jurisdiction. Because a corporation in the first instance exists only as a legal construct, not a natural one, the concept piles one false analogy on top of another.
The decision to adopt an intelligent territorial tax system carries with it a concomitant obligation to tame the stateless income beast -- to do less is not only to abandon any contingent claim to taxing the income from foreign direct investment, but also to expose the domestic income tax base to continued erosion. Contrary to the suggestions of many U.S. multinational firms in their push for a territorial tax system without any constraints on stateless income planning or any safeguards to protect the domestic income business tax base, several other countries that have adopted territorial tax systems are concerned about these very issues and are exploring a number of important strategies (such as thin capitalization statutes that apply to the parent company of a multinational group) to address the stateless income problem.81 It mischaracterizes the work of those jurisdictions to argue that they all let their multinationals stay out late and party with whomever they please, so we should, too. That argument also fails to address the important economic inefficiencies that would result if the United States were to adopt what can be called a cartoon version of territorial taxation, in particular the tax-induced incentive to prefer foreign investment in high-tax jurisdictions over U.S. investment, in order to generate tax rents.
From the other direction, a true worldwide tax consolidation regime would require careful thought as to how one would define what constitutes a U.S. resident corporation. The concern, of course, is not so much for the immediate future (section 367 and a beefed-up section 7874 could protect the United States from a mass exodus of U.S. corporations) but for the long term.82 A worldwide tax consolidation regime also would require resolving a number of technical issues, such as the correct threshold for consolidation.83
As a practical matter, a worldwide tax consolidation system is impossible to imagine without a concomitant to a much lower statutory corporate tax rate than current law's 35 percent. A lower rate relieves some of the pressure on the definition of residence, and has the correlative virtue of lowering the tax burden of wholly domestic U.S. corporations, thereby addressing an authentic tax-induced global competitiveness concern that has remained surprisingly out of the spotlight, when compared with the attention given the alleged competitiveness problems faced by U.S. multinational firms in respect of their foreign tax rates.
Unsurprisingly, at the end of the day I favor a worldwide tax consolidation system.84 Worldwide tax consolidation promotes capital export neutrality. Thanks to the work of Desai and Hines, among others, we now understand that capital export neutrality is not the only relevant economic efficiency benchmark, but that does not make it wholly meaningless.
Along the more pragmatic decision margins that should dominate the political economy analysis, worldwide tax consolidation substantially militates against a U.S. multinational firm's use of aggressive transfer pricing tax return positions to strip income out of the United States, because in general there is no tax advantage to doing so. For the same reason, a worldwide tax consolidation system directly addresses the problem of rampant stateless income tax planning, through which income migrates from high-tax foreign jurisdictions to low-tax ones. In the absence of demonstrably watertight definitions of the source of income, all territorial tax regimes remain extremely vulnerable on both counts.85
Worldwide tax consolidation also eliminates current law's open invitation to engage in tax arbitrage. Moreover, a move to true worldwide tax consolidation along with a substantial reduction in corporate tax rates would reduce the corporate tax burdens borne by wholly domestic businesses (or multinational ones unable to indulge in wholesale stateless income planning). To paraphrase Winston Churchill's thoughts on democracy as a form of government, worldwide tax consolidation is a terrible idea -- except in comparison to all the alternatives.
1 The working version of this paper is available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1791769.
2 The working version of this paper is available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1791783.
The domicile of a multinational enterprise's ultimate parent company is referred to in the literature as the "residence" country. A country other than the residence country in which a multinational group derives business or investment income is referred to as the "source" country.
4 Examples of recent papers emphasizing how current arm's-length transfer pricing rules lead to the erosion of residence country tax revenues include Yariv Brauner, "Cost Sharing and the Acrobatics of Arm's Length Taxation," Univ. of Fla. Legal Studies Research Paper No. 2010-19 (2010), available at http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1651334; and Harry Grubert, "Foreign Taxes, Domestic Income, and the Jump in the Share of Multinational Company Income Abroad: Sales Aren't Being Globalized, Only Profits" (Dec. 1, 2009), available at http://web.gc.cuny.edu/economics/SeminarPapers/spring%202010/Grubert_March16.pdf.
5 Joel Slemrod, "Location, (Real) Location, (Tax) Location: An Essay on Mobility's Place in Optimal Taxation," 63 Nat'l Tax J. 843 (2010). Slemrod alludes to stateless income with his concept of tax mobility; this report argues that stateless income is even more pervasive than Slemrod's paper might suggest.
6 George Zodrow, "Capital Mobility and Capital Tax Competition," 63 Nat'l Tax J. 865, 881 (2010).
7 In general, the report takes as a given that it is easier to migrate pretax income from a high-tax foreign country than from a high-tax residence country. So, for example, it assumes it is easier for U.S. firms to move income economically earned in Germany to Ireland than it is to move income from sales to U.S. customers to Ireland. Aggressive transfer pricing strategies for intangibles might appear to be an exception, but that technique is at least as valuable with respect to high-tax foreign countries as it is for moving income out of the United States, and in any event is not implicated by a marginal investment. The allocation of global interest expense is a better counterexample.
8 J.P. Morgan & Co., North America Equity Research, U.S. Equity Strategy Flash, June 27, 2011.
9 The facts that follow are drawn principally from Jesse Drucker, "Google 2.4 Percent Rate Shows How $60 Billion Lost to Tax Loopholes," Bloomberg, Oct. 21, 2010, available at http://www.bloomberg.com/news/2010-10-21/google-2-4-rate-shows-how-60-billion-u-s-revenue-lost-to-tax-loopholes.html, as supplemented by inferences drawn from Joseph S. Darby III and Kelsey Lemaster, "Double Irish More Than Doubles the Tax Savings: Hybrid Structure Reduces Irish, U.S. and Worldwide Taxation," 11 Practical U.S./International Tax Strategies 2 (May 15, 2007). Since Google's tax planning is not transparent to outside observers, there could be some slight mischaracterizations of details in the text, but these would not detract from the points made therein.
10 As one example roughly contemporaneous with Google's Double Irish Dutch Sandwich, see Jeffrey L. Rubinger and William B. Sherman, "Holding Intangibles Offshore May Produce Tangible Tax Benefits," Tax Notes Int'l, Mar. 7, 2005, p. 907, Doc 2005-2977 , or 2005 WTD 45-13 , proposing a complex structure involving Norwegian companies to achieve comparable results.
11 Google blog post dated Oct. 6, 2004, available at http://googleblog.blogspot.com/2004/10/dublin-go-bragh.html ("A year ago, Dublin became the first location for Google's regional operations outside the U.S. We designed it to serve Google customers across multiple time zones and languages spanning Europe, the Middle East and Africa. There were just five of us in 2003. Today we've built a team of 150.").
12Veritas Software Corp. v. Commissioner (Symantec), 133 T.C. 297 (2009), Doc 2009-27116 , 2009 WTD 236-42 , offers an important window into how cost-sharing agreements were constructed at times proximate to the formation of Ireland Holdings. In Veritas, the Tax Court accepted as correct the $118 million cost-sharing buy-in payments made by an Irish subsidiary of a U.S. parent company beginning in 1999 against an IRS argument that the correct amount for the buy-in payment was $1.675 billion. For brief summaries, see, e.g., Kerwin Chung et al., "Tax Court Rejects IRS's Cost-Sharing Buy-In Analysis," Tax Notes Int'l, Jan. 4, 2010, p. 12, Doc 2009-27440 , or 2009 WTD 240-2 ; Stephen Blough et al., "Veritas Vincit," Tax Notes, Feb. 15, 2010, p. 839, Doc 2010-1954 , or 2010 TNT 31-5 . More recently the IRS announced that it would not appeal Veritas. Cindy Hustad and Alan Shapiro, "IRS Decides Not to Appeal Veritas; Action on Decision Issued," Tax Notes, Dec. 20, 2010, p. 1342, Doc 2010-26116 , 2010 TNT 243-6 . The Treasury regulations covering cost-sharing arrangements were revised in 2009. The new regulations appear to give the IRS more leeway to insist that buy-in payments like those at issue in Veritas take account of the value of transferred platform intangibles as a long-lived continuing foundation that provides incremental value to subsequent research and development work.
13 There is no publicly available information on the size or calculation of the buy-in payment or on the operations of Ireland Holdings before the cost-sharing agreement was entered into. The text's description relies on my general experience and conversations with market professionals and therefore may not strictly comport with Google's case. I believe, however, that the presentation is a fair summary of practice in this area in general.
14 Reg. section 301.7701-3(a)(2). That is the structure proposed in Darby and Lemaster, supra note 9, at 2. Like all other federal income tax return materials, check-the-box filings are not publicly available.
15 In a 2008 video interview, John Herlihy, the manager of Google Ireland, described Google's Irish operations as the company's second largest office. At the time, Google Ireland had 1,350 employees, 900 of whom worked on the online sales team, 250 in technology side, and 200 apparently in corporate-support-type functions for the EMEA operations, available at http://www.youtube.com/watch?v=pYZsLLMQZxM&NR=1&feature=fvwp. See also Google's description of its Irish operations, available at http://www.google.ie/intl/en/jobs/dublin/ ("What we do in Dublin is help millions of Google users and customers right across Europe, the Middle East, and Africa (EMEA) to get the most from our products. Google's Dublin office is the EMEA Operations Headquarters. That means we support everyone who uses our products: the search engine that we are most known for, plus consumer products like Gmail and Calendar, advertising products like AdWords and AdSense, right through to business solutions for major corporations. In Dublin we also build on our existing products and create new ones, employing some of the finest engineering talent in the world. Many of the Dublin-based teams are engaged in supporting other Google offices across the EMEA region, working in areas like finance, payroll, legal, and HR.").
16 Treasury regulations governing cost-sharing agreements were revised in 2008 to adopt the investor model of arm's-length pricing. Reg. section 1.482-7, as amended by T.D. 9441, Doc 2008-27341 , 2009 WTD 1-24 . This model emphasizes the idea that an affiliate that contributes only cash to a cost-sharing agreement built around existing high-value intangible assets should make buy-in payments that leave the affiliate with only a normal return on its operations. See Joint Committee on Taxation, "Present Law and Background Related to Possible Income Shifting and Transfer Pricing," JCX-37-10 (July 20, 2010), at 25-29, 111-114, Doc 2010-16144 , 2010 WTD 139-29 . But the regulations do not reject the idea of a cash-box subsidiary participating in a cost-sharing agreement in the first instance, and they might be expected only to lead to transfers of intangible assets at an earlier stage of development. Moreover, cash-box subsidiaries can contract with, and license intangible assets from, their U.S. parent. Those transactions are not ignored for U.S. tax purposes. Id. at 115-116.
17 Within the European Union, member states cannot impose source country withholding tax on royalties paid to a company resident in another state. Moreover, Ireland has a good tax treaty network whose treaties often reduce the tax rate on royalties paid between firms in the two treaty countries to zero.
18 Darby and Lemaster, supra note 9, do not discuss the role of the Dutch firm, either because they viewed it as a proprietary twist on the basic Double Irish idea or because it had not yet come into vogue. The article by Drucker, supra note 9, does discuss it.
19 Drucker, supra note 9.
20See generally J. Clifton Fleming Jr., Robert J. Peroni, and Stephen E. Shay, "Worse Than Exemption," 59 Emory L. J. 79 (2009) (concluding that the U.S. international tax system can be as generous or more generous than a well-designed territorial system. Id. at 149). See also Fleming, Peroni, and Shay, "Some Perspectives From the United States on the Worldwide Taxation vs. Territorial Taxation Debate," 3 J. Austrl. Tax Teachers Ass'n 35, 44 (2008); Fleming, Peroni, and Shay, "Deferral: Consider Ending It, Instead of Expanding It," Tax Notes, Feb. 7, 2000, p. 837, Doc 2000-3632, or 2000 TNT 25-66 ; Lawrence Lokken, "Does the U.S. Tax System Disadvantage U.S. Multinationals in the World Marketplace?" 4 J. Tax'n Global Transactions 43 (2004).
21 The United States taxes on a current basis some categories of passive investment income or highly mobile income earned by foreign subsidiaries of U.S. firms' subpart F income. Over the last several years, the scope of the subpart F system has been reduced, so that increasing amounts of U.S. firms' foreign earnings can qualify as active business income and therefore be eligible for deferral. See Lokken, "Whatever Happened to Subpart F -- U.S. CFC Legislation After the Check-the-Box Regulations," 7 Fla. Tax Rev. 185 (2005). This scale-back of the subpart F system has greatly increased the ability of U.S. firms to operate in a quasi-territorial environment and to generate stateless income.
22 Section 864(e); reg. section 1.861-9 and -9T.
23 More technically, by driving down its foreign effective tax rate before considering interest expense, a firm can incur more interest expense in the United States without bumping into the section 904 ceiling on FTC use. The lower effective foreign tax rate (pre-U.S. interest expense) creates more capacity to harmlessly absorb the fraction of U.S. interest expense that is allocated against foreign-source income.
24 The figure represents the 35 percent U.S. statutory tax rate applied to the aggregate excess limitation income reported by those U.S. firms with excess limitation for the year. (Personal correspondence with Harry Grubert, Treasury Department.) Grubert and Rosanne Altshuler, "Corporate Taxes in the World Economy: Reforming the Taxation of Cross-Border Income," in Fundamental Tax Reform: Issues, Choices, and Implications, 319, 331-333 (2008), at 326-327, identifies several shortcomings with this approach to measuring the effective tax burden on foreign income. Since these shortcomings point in opposite directions and since no better data exist, it is necessary to use this measure. Those authors also analyze the components of the U.S. residual tax on foreign income for 2000, when the revenue from that tax totaled $12.7 billion.
25 For the first two figures, see Statistics of Income data for CFCs for 2004, available at http://www.irs.gov/taxstats/bustaxstats/article/0,,id=96282,00.html. The data are measured using U.S. tax principles rather than U.S. generally accepted accounting principles. The last two figures came in response to my request for information from SOI.
26 The data in the text assume that dividends are paid first out of current earnings, so that dividends paid in 2004 can be presented as distributed out of 2004 earnings. The data do not show how much cash was retained by foreign subsidiaries of U.S. firms, in part because CFCs can distribute cash out of previously taxed income (basically, subsidiary income previously taxed to the U.S. parent under subpart F). Those distributions are excludable from the U.S. parent company's taxable income. Section 959(a). In 2004 CFCs distributed $43.8 billion of previously taxed income.
27 Fleming, Peroni, and Shay, "Worse Than Exemption," supra note 20; Lokken, "Territorial Taxation: Why Some U.S. Multinationals May Be Less Than Enthusiastic About the Idea (and Some Ideas They Really Dislike)," 59 SMU L. Rev. 751, 759-770 (2006).
28 If one imagines that subpart F income would be defined in a territorial tax system like its current definition, even under that hypothetical territorial tax regime U.S. tax would be owed on the $48 billion of subpart F income includable in the income of U.S. shareholders after taking into account FTCs attributable to that income. If one assumes that the $48 billion in subpart F income brought with it FTCs at the global average of 16 percent, the residual U.S. tax would be about $11 billion.
29 For greater detail on the calculation of this information, see Lee Mahoney and Randy Miller, "Controlled Foreign Corporations, 2004," 28 SOI Bull. 49, 58-59 (Summer 2008).
E&P is a technical tax term of art, and for this purpose can be understood as a measure of income calculated using fundamental tax norms like the realization principle but with more economic measures of key items (such as depreciation) than would apply when calculating taxable income for a domestic income tax return.
30 Technically, these taxes include taxes paid to U.S. possessions. Id. at 59.
32 Microsoft 2010 annual report, financial statements, note 13. The next subsection describes the concept of permanently reinvested earnings in more detail.
33 Some of this $9.2 billion repatriation tax cost might be attributable to foreign withholding taxes, but those taxes ordinarily are fully creditable in the United States. As a result, the division of the repatriation tax cost between foreign withholding tax and U.S. residual income tax does not affect the calculation summarized in the following sentence in the text.
34 9.2/29.5 = 31 percent, implying that FTCs associated with the repatriation of all permanently reinvested earnings would amount to only about 4 percentage points.
35 Grubert and Altshuler, supra note 24, at 322-323 (29 percent effective statutory rate for foreign subsidiaries in 2003). Grubert and Altshuler concluded that in 2002, U.S. manufacturing firms' average effective foreign tax rate stood at 16 percent. Id.
36 Philips 2007 annual report, note 6 to U.S. generally accepted accounting principles financial statements and note 42 to international financial reporting standards financial statements.
37 Kimberly Clausing, "Multinational Firm Tax Avoidance and Tax Policy," 57 Nat'l. Tax J. 703, 714 (2009) (showing importance of Bermuda, Ireland, Luxembourg, Switzerland, and other low-tax countries as the situs of U.S.-domiciled multinational firms' profits); Martin Sullivan, "Extraordinary Profitability in Low-Tax Countries," Tax Notes Int'l, Sept. 8, 2008, p. 793, Doc 2008-18118 , or 2008 WTD 175-8 ("Low-tax Ireland is particularly prone to high profitability"); Sullivan, "U.S. Multinationals Shifting Profits Out of the United States," Tax Notes Int'l, Mar. 17, 2008, p. 910, Doc 2008-4725 , or 2008 WTD 49-5 ; Sullivan, "A Challenge to Conventional International Tax Wisdom," Tax Notes Int'l, Dec. 11, 2006, p. 841, Doc 2006-24455 , or 2006 WTD 239-5 (30 percent of the pretax profits of foreign affiliates of U.S. firms were located in very low-tax countries, a figure greatly disproportionate to employment or physical capital there); Altshuler and Grubert, "The Three Parties in the Race to the Bottom: Host Governments, Home Governments and Multinational Companies," 7 Fla. Tax L. Rev. 153, 170, 182 (2005) (finding that from 1997 to 2002 there was almost 100 percent growth in the income of foreign affiliates of U.S. parent companies in seven major low-tax countries -- Bermuda, the Cayman Islands, Ireland, Luxembourg, the Netherlands, Singapore, and Switzerland -- and that this income represented roughly 40 percent of worldwide income from equity investments).
38 Grubert, "Foreign Taxes, Domestic Income," supra note 4, at 2.
39 John R. Graham, Michelle Hanlon, and Terry Shevlin, "Real Effects of Accounting Rules: Evidence From Multinational Firms' Investment Location and Profit Repatriation Decisions," 49 J. Acct. Res. 137, 141 (2011). This report argues that GAAP accounting for taxes dominates cash tax costs.
GAAP accounting now requires firms to set out their cash tax payments for a year. This category is not the same as the tax liabilities shown as due on the firm's tax returns for the year, because the financial accounting category is a simple record of cash flows. Tax payments for prior years are conflated, for example, with estimated payments for the current year. As previously noted, this report uses the phrase "cash taxes" to mean the tax liabilities shown as due on the taxpayer's tax returns for the year.
40 Other, much smaller examples are the section 199 domestic production deduction and the research credit.
41 A more technical description would be that the facts drive the financial accounting result, but that the company controls the facts, including those concerning its future plans.
42 Jennifer L. Blouin, Linda K. Krull, and Leslie A. Robinson, "Is U.S. Multinational Intra-Firm Dividend Policy Influenced by Reporting Incentives?" (Feb. 2011), at 6, available at http://ssrn.com/abstract=1468135. The authors also find that public companies are more sensitive to the accounting benefits of permanently reinvested earnings than are private firms, which is consistent with the point made earlier in the text that financial accounting is the lens through which stakeholders view public firms.
43 Graham, Hanlon, and Shevlin, supra note 39, at 140.
44See Mark Bauman and Ken Shaw, "The Usefulness of Disclosures on Untaxed Foreign Earnings in Firm Valuation," 30 J. Am. Tax. Assoc. 53 (2008) ("This result is due to estimated repatriation tax amounts exhibiting downward bias, and less accuracy for actual repatriation tax effects, relative to firm-disclosed repatriation tax amounts."); Julie H. Collins, John R.M. Hand, and Douglas A. Shackelford, "Valuing Deferral: The Effect of Permanently Reinvested Foreign Earnings on Stock Prices," in International Taxation and Multinational Activity 143 (2000). Lisa Bryant-Kutcher, Lisa Eiler, and David Guenther also found evidence that firms' stock prices were discounted for disclosed repatriation tax costs, but only if those firms also had accumulated high levels of excess foreign cash -- presumably in an effort to avoid repatriation taxes. See Lisa Bryant-Kutcher, Lisa Eiler, and David A. Guenther, "Taxes and Financial Assets: Valuing Permanently Reinvested Foreign Earnings," 61 Nat'l Tax J. 699, 701 (2008).
At least one study concludes that the market does not discount stock prices for the unreported tax liability from permanently reinvested earnings. Dan Dhaliwal and Linda Krull, "Permanently Reinvested Earnings and the Valuation of Foreign Subsidiary Earnings" (unpublished, 2006). Also, while the Collins, Hand, and Shackelford model concludes that stock prices are hurt by disclosed but unquantified tax liabilities, it does not estimate with statistical significance the size of this effect. See Collins, Hand, and Shackelford, at 155-156.
45 GE 2009 annual report, note 14 to consolidated financial statements, p. 93 (showing 2007 as well as 2009 effective tax rate data), available at http://www.ge.com/ar2009.
47See, e.g., Richard J. Vann, "Taxing International Business Income: Hard-Boiled Wonderland and the End of the World," 1 World Tax J. 291, 305-343 (2010). Michael J. Graetz, "A Multilateral Solution for the Income Tax Treatment of Interest Expenses," 62 Bull. for Int'l Tax. 486 (2008), eloquently describes the artificiality of source rules that apply when determining the includability or deductibility of interest and recommends in effect a global multilateral treaty to apportion interest expense on pure fungibility of assets principles to all members of an affiliated group of companies, without regard to the identity of the affiliate that borrowed the funds. In other words, Graetz proposes the worldwide adoption of a formulary income standard, but applied only to interest expense.
48 This is the standard view in economics presentations. See infra note 57.
49 It might be argued that multinational firms are so successful in generating stateless income that their investment behavior changes global asset prices by bidding up prices for high-tax country assets. If multinational firms were the price setters in corporate investments around the world and they paid no tax anywhere (or conversely, paid residence country tax on everything), one might see convergence in pretax rather than after-tax risk-adjusted corporate net incomes (just as should be true for interest income today).
This scenario seems implausible for several reasons. First, all domestic investors and all portfolio investors (whether domestic or cross-border) are post-corporate tax investors. Moreover, since much cross-border investment today is portfolio investment, there is no particular reason to assume that direct investment by multinational firms sets asset prices. Second, not even this report argues that all multinational firms convert 100 percent of cross-border investment income into zero-taxed returns. Third, as developed in Section V, the ability to generate stateless income is a form of status tax arbitrage, which means it is available only to some investors competing for a particular investment. (Indeed, as effective tax rate studies show, it is not even a status equally distributed among all multinational firms.) Fourth, investment opportunities that yield normal returns often are relatively fungible or can be replicated through greenfield construction. As in the domestic market for municipal bonds or tax shelters (see Section V), it seems implausible to think that market forces alone would be sufficient to vitiate the tax rents story developed in the text.
50 For example, if a U.S domestic affiliate of a U.S. multinational group pays interest to a foreign affiliate, that income will constitute subpart F income. Section 954(a)(1) and (c)(1)(A). When a foreign affiliate in a high-tax jurisdiction pays interest out of active business earnings to an affiliate in a low-tax jurisdiction, that interest income is not subpart F income, because of the look-through rules of section 954(c)(6).
The section 954(c)(6) look-through provision is temporary and was recently extended by Congress.
51 James R. Hines Jr., "Reconsidering the Taxation of Foreign Income," 62 Tax L. Rev. 269, 280 (2008), summarizes the research underlying this counterargument.
52Id. at 278 ("To a first approximation there is little effect of additional foreign investment on domestic tax revenue."). Hines offers no evidence in support of this assertion. It may be that he assumes that investment and taxable income generally are closely positively correlated. A principal theme of this report, by contrast, is that stateless income tax planning and analogous strategies used by U.S.-domiciled multinational groups for the U.S. tax base have substantially disassociated investment from taxable income.
In one fairly recent study on the earnings stripping rules of section 163(j), Treasury concluded that the evidence for the proposition that foreign-controlled domestic firms systematically stripped income out of the United States was ambiguous. Treasury, "Report to the Congress on Earnings Stripping, Transfer Pricing and U.S. Income Tax Treaties" (Nov. 2007), at 3, Doc 2007-26269 , 2007 WTD 230-21 ("As discussed below, it is not possible to quantify with precision the extent of earnings stripping by foreign-controlled domestic corporations generally. However, there is strong evidence of earnings stripping by the subset of foreign-controlled domestic corporations consisting of inverted corporations (i.e., former U.S.-based multinationals that have undergone inversion transactions).").
The Treasury study has been treated skeptically. See, e.g., Stephen E. Shay, "Ownership Neutrality and Practical Complications," 62 Tax L. Rev. 317, 322 (2009). Its conclusions also appear to be at least partially inconsistent with those reached in a contemporaneous report by the GAO, "Tax Administration: Comparison of the Reported Tax Liabilities of Foreign- and U.S.-Controlled Corporations 1998-2005," GAO-08-957 (July 2008), Doc 2008-17519 , 2008 TNT 157-15 ("FCDCs reported lower tax liabilities than USCCs by most measures shown in this report." Id. at 3.). The GAO report acknowledges that there are several non-tax-related factors, such as the average age of foreign- and domestic-controlled domestic corporations, that might explain some of the differences in results.
53 Richard Waters, "Tax Drives U.S. Tech Groups to Tap Debt," Financial Times, Feb. 7, 2011, at 15, col. 6. In the same vein, Microsoft's plans to acquire Skype Software SARL (a Luxembourg-based company) have been explained as a tax-efficient use of the firm's vast hoard of offshore cash. "Microsoft Structured Acquisition of Skype to Avoid U.S. Taxes," available at http://thinkprogress.org/2011/05/13/microsoft-skype-tax-havens/.
54 Hines, "Foreign Income and Domestic Deductions," 61 Nat'l Tax J. 461, 463-464 (2008). ("Taxpayers whose foreign income is lightly taxed by foreign governments, and who, therefore, owe residual U.S. tax on that income, receive the benefit of full domestic deductibility of expenses incurred in the United States.").
55 Kleinbard, supra note 3, at 555-556. See also Kleinbard, supra note 2, Section II.
56 Mihir A. Desai and Hines, "Evaluating International Tax Reform," 56 Nat'l Tax J. 487 (Sept. 2003). The authors further developed their thesis in a series of subsequent papers, including Desai and Hines, "Old Rules and New Realities: Corporate Tax Policy in a Global Setting," 57 Nat'l. Tax J. 937, 957 (2004); Desai, "New Foundations for Taxing Multinational Corporations," Taxes (Mar. 2004); Hines, "Foreign Income and Domestic Deductions," 61 Nat'l Tax J. 461, 463-464 (2008); and Hines, "Reconsidering the Taxation of Foreign Income," 62 Tax L. Rev. 269, 280 (2008).
57 This is a standard assumption in economics presentations. See, e.g., Altshuler, "Recent Developments on the Debate on Deferral," Tax Notes, Apr. 10, 2000, p. 255, Doc 2000-10559, or 2000 TNT 69-97 ; Michael Devereux, "Taxation of Outbound Investment" 24 Oxford Rev. of Econ. Pol'y 698, 701-702 (2008); Zodrow, "Capital Mobility and Capital Tax Competition," supra note 6, at 881.
58 Kleinbard, supra note 2, Section III develops this analysis at considerable length.
59 Japan planned to reduce its national total (central and subcentral government) corporate tax rate to 34.5 percent on April 1, 2011, but the earthquake and tsunami disaster caused them to postpone the tax cut.
60 Competitiveness is not a concept that is well developed in the economic literature. For two recent efforts to better incorporate the term into economic analysis, see Eckhard Siggel, "International Competitiveness and Comparative Advantage: A Survey and a Proposal for Measurement," 6 J. Ind. Competition Trade 137 (2006); and Michael Knoll, "The Corporate Income Tax and the Competitiveness of U.S. Industries," 63 Tax L. Rev. 771 (2009).
U.S. multinational firms can be said not to be deeply troubled by any terminological ambiguity. To them, an "anticompetitive" measure is any cost that along any dimension might be greater than the comparable cost faced by a firm not domiciled in the United States. It has been my experience that no amount of pro-competitive factors can ever outweigh the damage imagined to be done by a single anticompetitive one.
61See http://www.oecd.org/dataoecd/26/56/33717459.xls for 2010 data.
62 OECD, "OECD in Figures, 2009 Edition," at 58, Table of 2006 Comparative Income Tax Rates (28.1 percent OECD simple average of maximum corporate statutory rates, including subnational taxes on corporate income).
63 No state directly taxes foreign income under its general corporate income or franchise tax. Three states (Idaho, Montana, and North Dakota) require global consolidation and apportionment of income. If firms report consistently higher profits on a separate company basis outside the United States than they do inside, the effect of this rule may be to increase firms' tax liabilities in those states. Finally, three states (California, Utah, and West Virginia) permit worldwide consolidation and apportionment at the taxpayer's election.
64 Author's calculations. The five largest OECD economies outside the United States are Japan, Germany, the United Kingdom, France, and Italy. The next six are Canada, Spain, Korea, Mexico, Australia, and the Netherlands. See Christopher Heady, "Directions in Overseas Tax Policy," in Melbourne Institute, Australia's Future Tax and Transfer Policy Conference, ch. 2, at 8, available at http://taxreview.treasury.gov.au/content/Content.aspx?doc=html/conference_report.htm.
65 Author's calculation from data in KPMG Corporate and Indirect Tax Survey 2010, available at http://www.kpmg.com/LU/en/IssuesAndInsights/Articlespublications/Pages/KPMG%27sCorporateandIndirectTaxRateSurvey2010.aspx. My calculation uses the standard (nonpreferential regime) maximum corporate income tax rate, which is consistent with the OECD method.
66 Grubert, "Foreign Taxes, Domestic Income," supra note 4, at 19.
68 Fritz Foley et al., "Why Do Firms Hold So Much Cash? A Tax-Based Explanation," 86 J. Fin. Econ. 579 (2007) (U.S. tax rules for foreign direct investment induce U.S. firms to accumulate excessive cash).
69 Serena Saito, "U.S. Technology Companies Go Shopping Abroad to Avoid Taxes," Bloomberg Businessweek, Sept. 2, 2010, available at http://www.businessweek.com/magazine/content/10_37/b4194031986280.htm.
70 Richard Waters, "Tax Drives U.S. Tech Groups to Tap Debt," Financial Times, Feb. 7, 2011, at 15, col. 6.
71See Desai and Hines, "Old Rules and New Realities: Corporate Tax Policy in a Global Setting," 57 Nat'l. Tax J. 937, 957 (2004) (Desai and Hines estimate $10 billion in indirect efficiency losses attributable to retained earnings resulting from residual tax on dividends); Grubert, "Comment on Desai and Hines, 'Old Rules and New Realities: Corporate Tax Policies in a Global Setting,'" 58 Nat'l. Tax J. 263 (2005) (rejecting Desai and Hines's proposition because dividends account for a relatively small amount of revenue); Desai, "Reply to Grubert," 58 Nat'l. Tax J. 275 (2005); Grubert, "MNC Dividends, Tax Holidays and the Burden of the Repatriation Tax: Recent Evidence," Oxford University Center for Business Taxation, working paper (2009).
72 Bryant-Kutcher, Eiler, and Guenther, supra note 44, at 702-703 ("Since U.S. tax law provides an incentive for foreign subsidiaries to defer repatriation of cash, managers must trade off the negative impact of U.S. repatriation taxes on firm value with the lower benefits that come from reinvesting foreign earnings in financial assets.").
73Cf. id. at 705 n.16 ("Only seven percent of U.S. firms that accumulate excess cash outside the United States to avoid the U.S. repatriation tax appear to be constrained in their access to capital markets.").
74 As examples from the Dow Jones industrial average companies Hewlett-Packard and, until recently, Microsoft.
75 For example, as previously described, Microsoft reported $29.5 billion in permanently reinvested earnings at June 30, 2010 (the end of its fiscal year). At the end of its fiscal 2011 second quarter (Dec. 31, 2010), Microsoft reported holding $41.2 billion in cash, cash equivalents, and short-term investments. (As previously noted, GAAP financial statements do not describe the location of these items within a multinational group.) In February 2011 Microsoft borrowed $2.25 billion in the public capital markets, including $1 billion of 5.3 percent notes due in 30 years and $500 million of 4 percent notes due in 10 years.
77 Dhammika Dharmapala, C. Fritz Foley, and Kristen J. Forbes, "Watch What I Do, Not What I Say: The Unintended Consequences of the Homeland Investment Act," National Bureau of Economic Research Working Paper Series (June 2009), available at http://www.nber.org/papers/w15023.pdf (concluding: "Repatriations did not lead to an increase in domestic investment, employment or R&D, even for the firms that lobbied for the tax holiday stating these intentions."). Ironically, the 2004 legislation prohibited the use of dividends eligible for the special repatriation holiday to fund stock buybacks. The paradox is solved once one discovers that the prohibition did not incorporate any fungibility of money concept, so that firms could both accomplish their corporate finance objectives and comply with the law by segregating different pools of cash for different corporate expenditures.
78 Foley et al., supra note 68 (U.S. tax rules for foreign direct investment induce U.S. firms to accumulate excessive cash).
79 Grubert and Altshuler, "Corporate Taxes in the World Economy: Reforming the Taxation of Cross-Border Income," Fundamental Tax Reform: Issues, Choices, and Implications, 319, 331-333 (2008).
80 Kleinbard, supra note 2.
81 Kleinbard, supra note 2, Section V considers those developments in detail, and more generally explores the difficulties in designing territorial tax systems that are robust to stateless income tax planning.
82 Kleinbard, supra note 2, Section VI is devoted to this and related issues.
84 Kleinbard, supra note 2, Section VI greatly expands on the reasoning.
85 For that reason, it is likely that the most promising way to design a territorial tax system that deters the sort of tax planning that this report addresses would be to couple such a system to a comprehensive formula apportionment of income. See Kleinbard, supra note 2, Section V. But to date (with the possible exception of the European Union's recent work on a common consolidated corporate tax base), no such formula has been comprehensively specified and vetted.
END OF FOOTNOTES