Jump to ContentJump to Main Navigation
Fixing U.S. International Taxation$

Daniel N. Shaviro

Print publication date: 2014

Print ISBN-13: 9780199359752

Published to Oxford Scholarship Online: April 2014

DOI: 10.1093/acprof:oso/9780199359752.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (www.oxfordscholarship.com). (c) Copyright Oxford University Press, 2019. All Rights Reserved. Under the terms of the licence agreement, an individual user may print out a PDF of a single chapter of a monograph in OSO for personal use (for details see www.oxfordscholarship.com/page/privacy-policy). Subscriber: null; date: 15 February 2019

The Main Building Blocks of U.S. International Taxation

The Main Building Blocks of U.S. International Taxation

Chapter:
(p.31) 2 The Main Building Blocks of U.S. International Taxation
Source:
Fixing U.S. International Taxation
Author(s):

Daniel N. Shaviro

Publisher:
Oxford University Press
DOI:10.1093/acprof:oso/9780199359752.003.0002

Abstract and Keywords

This chapter reviews the basic U.S. international tax rules. It identifies five basic concepts needed to grasp the basic elements of how the U.S. taxes the income of multinational firms. First, there are the rules for determining corporate residence. Second are the rules for determining so-called source of income. Third are the rules concerning foreign tax credits. Fourth is deferral, or the lack of current inclusion for foreign source income that U.S. companies earn through foreign subsidiaries (referred to as “CFCs,” controlled foreign corporations, in the Internal Revenue Code). Deferral generally ends when the U.S. parent repatriates or otherwise realizes the underlying CFC income, such as by receiving a dividend or selling CFC stock. Fifth, if one chooses to list it separately from deferral rather than as part of the same broader concept, is subpart F of the Code. This is the rubric for a set of rules that potentially limit the scope of deferral by providing deemed dividend treatment for certain types of CFC income, thus making such items currently taxable to the U.S. parent.

Keywords:   international tax policy, corporate tax, income tax, multinational firms, controlled foreign corporations, tax credits, deferral, corporate residence, income source, tax law

THE U.S. INCOME tax rules for cross-border activity are complicated enough to require vast forests of print for those not yet entirely wedded to reading on computer screens. Suppose you wanted to review all of the income tax statutes that particularly address international tax issues. Internal Revenue Code sections 861 through 999, where for the most part they are found, takes up no fewer than 282 pages in a prominent recent edition.1 The Treasury regulations pertaining to those Code sections unfold luxuriantly across another 1,288 pages.2 The leading international tax treatise (Kuntz and Peroni 2013), while inhospitable to a page count because its many (and nonconsecutively page-numbered) chapters have continually updated inserts and replacement pages, sprawls sideways in its loose-leaf binders over nearly a foot of shelf space.

Over the years, these primary and secondary source materials keep growing like kudzu as ever more authorities emerge, addressing ever more permutations in applying the U.S. international tax rules. Intellectually gifted attorneys at leading U.S. law and accounting firms can and do devote large portions of their careers just to studying how all these rules apply to and are best utilized by their clients, without feeling at any point that they have now seen and know it all.

Nonetheless, this welter of complexity has a surprisingly simple core. To grasp the basic elements of how we tax the income of multinational firms, taking as given the rules that would be needed to tax income in a purely domestic setting, just four or five basic concepts—depending on how one counts—go far indeed. First, there are the rules for determining corporate residence. Second are the rules for determining what I will call (p.32) the source of income (defining this term somewhat more broadly than is customary in legal doctrine). Third are the rules concerning foreign tax credits. Fourth is deferral, or the lack of current inclusion for foreign source income that U.S. companies earn through foreign subsidiaries (“CFCs,” for controlled foreign corporations, in the lingo of the Internal Revenue Code). Deferral generally ends when the U.S. parent repatriates or otherwise realizes the underlying CFC income, such as by receiving a dividend or selling CFC stock. Fifth, if we choose to list it separately from deferral rather than as part of the same broader concept, is subpart F of the Code. This is the rubric for a set of rules that potentially limit the scope of deferral by providing deemed dividend treatment for certain types of CFC income, thus making such items currently taxable to the U.S. parent.

A. Corporate Residence

Just as all social groups have at least an incipient sense of “us” versus “them,” so all countries distinguish for multiple legal purposes between their citizens or residents on the one hand, and everyone else on the other. Typically only “we” qualify for specified benefits (such as the right to vote) and are subject to specified burdens or obligations (such as compulsory military service). Thus, it comes as no surprise to observe that income tax systems around the world, including in the United States, distinguish between those who are classified as domestic residents (or citizens) and those who are not.

The main reason this income tax classification matters is that generally only citizens and/or residents are potentially subject to domestic tax obligations with respect to income arising abroad. The United States cannot, after all, tax a foreigner on what it admits is FSI—by requiring, say, a German to pay U.S. tax on income earned in Germany or France. In addition, as we will see, in some respects the application of the source rules depends on whether or not a given person (either the taxpayer or the issuer of a financial instrument) is classified as a U.S. resident.

As applied to individuals, the residence concept requires relatively little elaboration. Most people unmistakably live within a single country’s borders, and thus are its residents and no one else’s. True, there are problems relating to temporary cross-border jobs, permanent expatriation, and the like, but these for the most part are fairly easily handled, such as by bright-line rules concerning the number of days one spends in a year in a given country.

A central fact about cross-border business activity, however, is that it is predominantly conducted by legal entities, such as corporations, that are themselves—largely for reasons of administrative convenience (Shaviro 2009a)—commonly treated as taxpayers distinct from their owners. This simple fact turns out to have far greater implications for the enterprise of international taxation than one might initially have expected, and one of the key reasons relates to residence. The idea of distinguishing between “us” and “them,” so easy and natural most of the time when one is applying income tax rules to (p.33) individuals, must also apply to corporations once they are being classified as taxpayers. But here it causes far greater problems, reflecting its lack of a meaningful core. How can an abstraction, such as a legal entity defined by pieces of paper, have a true “residence” in the same sense as a flesh and blood individual?

Suppose that there were no cross-border shareholding—that is, that Americans only owned stock in U.S.-incorporated companies, Canadians in Canadian companies, and so forth. Then not only would it be easy to assign each company a country of residence, but the fact that residence is determined at the entity level would make little difference. To be sure, taxation at the entity level rather than the owner level might affect such matters as the tax rate for a given quantum of income, as well as the deductibility of losses from one investment against income from another. However, if all individuals owned only home-country stock, entity-level taxation would not affect the question of what foreign source income is subject to domestic tax.

With cross-border shareholding, however, the issue of corporate residence gets both harder to resolve and more consequential. To begin with, the existence of cross-border shareholding weighs against basing corporate residence on the ultimate owners’ nationality, even when those from one country predominate. Residence for tax purposes tends to be an all-or-nothing proposition, creating line-drawing problems if companies have owners all around the world. Moreover, ownership can change rapidly, especially for publicly traded companies. And even in the absence of ongoing trading, the underlying reality at the ultimate owner level can be hard to discern within a complex corporate structure in which particular owners may either hold distinctive combinations of rights, or place multiple tiers of intervening legal entities between themselves and the company whose residence must be determined.

Accordingly, with the ultimate owners’ residence being hard to utilize as a decisive (or even a contributing) factor, corporate residence rules generally apply either of two approaches (Kane and Rock 2008, 1235). The first, followed in the United States, is to ask where the entity is formally incorporated. The second is to ask, with varying degrees of rigor, where the company’s “real seat” appears to be located—that is, the main place where it is managed, or has most of its assets, employees, and business operations. This standard can range from simply asking where board of directors meetings generally are held—creating a “mild stimulus to the economies of Canada, Bermuda, or our Caribbean neighbors” (Tillinghast 1984, 263) if they build nice resorts for the meetings—to looking more rigorously at the “location of the administrative headquarters or… the firm’s center of gravity as determined by the location of the employees and assets” (Kane and Rock 2008, 1235).

With corporate residence depending on an entity-level attribute, a fundamental legal divide emerges between entity-level and owner-level taxpayer residence. Thus, suppose Ann and Brian, who are American individuals, operate a multinational business through Cayco, a Caymans entity, while Carlos and Dolores, who are Spanish, do so through Cardol, Inc., a Delaware corporation. For purposes of the U.S. corporate tax, Cayco is a (p.34) foreigner despite being American owned, while Cardol is a U.S. taxpayer despite being foreign owned. This ends up having important effects on the functioning of worldwide residence-based corporate taxation.

Even just conceptually, the entity-level focus on corporate residence can sow confusion. For example, if Ann and Brian put money in Cayco, which uses it to fund their U.S. business, this would appear to be “inbound” investment, coming to the United States from abroad. Likewise, if Carlos and Dolores use Cardol as a vehicle for investing their Spanish savings in Spain, this would appear to be “outbound” investment by a U.S. firm. If we try to rely on legally determined corporate residence in order to ascertain what actually is happening in the world economy (for example, what cross-border capital flows are taking place), we risk being misled.

More importantly, however, placing entity-level attributes at center stage has powerful effects on the practical reach of the U.S. international tax rules. In particular, it means that Americans’ foreign business income generally will fall outside the reach of the residence-based U.S. corporate income tax when they invest through foreign corporations, whereas foreign individuals may face the U.S. worldwide corporate tax, through its impact at the entity level, when they invest abroad through U.S. corporations. What is more, if individuals are reluctant to expatriate physically, but are willing to invest through foreign rather than domestic entities, then focusing on entity-level attributes ends up increasing residence electivity, or investors’ ability to select the corporate residence regime that they prefer, independently of what they are actually doing economically.

One of the most obvious ways electivity issues can arise is via an attempted midstream change in a given company’s place of residence, or at least that applying to the underlying corporate businesses. In countries using headquarters rules, changing the parent’s residence is relatively straightforward. At worst, if the country engages in a meaningful “real seat” inquiry rather than just looking at where annual meetings are held, it may require that enough of the company’s high-level managers be willing to relocate abroad. The threat of such headquarters expatriations appears to have significantly influenced the U.K. government’s recent decision to shift to a more territorial international tax regime for its resident companies.

In the United States, however, since we are a place of incorporation rather than a headquarters jurisdiction, the equivalent transaction would involve transferring the foreign businesses and their assets to a new foreign corporation that, assuming one does not want anything else to change, presumably would have the same owners as the old U.S. parent company. (There is no similar need to transfer the enterprise’s U.S. assets and businesses, given that they would be taxed here on a source basis anyway.)

When done in a purely domestic setting, such asset (or stock) transfers from one corporation to another with identical ownership may result in a technical realization of any built-in gain from asset appreciation, creating the potential for taxation at both the entity and shareholder levels. For many decades, however, tax-free reorganization rules have (p.35) generally permitted qualification for tax-free treatment at both levels, so long as (and to the extent that) assets remain in corporate solution and there are requisite showings of business purpose and lack of forbidden shareholder-level tax avoidance purposes, such as a plan to “bail out” corporate earnings at favorable capital gains rates.

In the international setting, however, concern about appreciated assets permanently leaving the reach of the U.S. tax system without ever being taxed is addressed by Internal Revenue Code section 367 and the regulations thereunder. This set of rules, though of “numbing complexity,” contain a “basic idea… [that] is nearly fathomable. It aims to preserve for U.S. taxation all gains that originally arose within the reach of U.S. taxation” (Isenbergh 2006, 216), such as by imposing a “toll charge” via immediate gain recognition (or the denial of valuable tax attributes such as the allowance of net operating losses against subsequent taxable income) in certain situations where appreciated foreign assets are reshuffled out of ultimate U.S. ownership.

Starting in the 1990s, the expatriation game changed sufficiently to render Code section 367 far less responsive than it had previously been to tax-planning motivations and potential U.S. revenue concerns. U.S. multinationals with foreign subsidiaries increasingly began arranging what were called corporate inversions. In a typical such transaction, the U.S. parent would be replaced at the top of the corporate group by a foreign company, located in a tax haven such as Bermuda or the Cayman Islands. Once the dust had settled, the new parent would own both the U.S. company that continued to conduct U.S. operations and, through a separate chain of ownership, all of the foreign operations. The latter would thereby cease to be owned by a U.S. company, and therefore would no longer be subject in any way, via the former U.S. parent, to the U.S. worldwide corporate tax regime (Desai and Hines 2002).

The point of these transactions was not to remove appreciated assets from the U.S. company, which is the main concern addressed by Code section 367, but rather to avoid having the foreign subsidiaries’ income ultimately subject to U.S. worldwide taxation via the former U.S. parent. A related goal was to facilitate tax planning to reduce the group’s U.S. source income from its U.S. operations. Such corporate inversions became a huge U.S. domestic political issue, however, when Stanley Works, a leading toolmaker that had been a U.S. company since 1843, announced in 2002 its intention to expatriate to Bermuda. Although this transaction ultimately fell through, Congress responded in 2004 by enacting the so-called anti-inversion rules of new Code section 7874.

Under these rules, along with applicable Treasury regulations that ultimately tightened their application,3 existing U.S. multinationals are now to a considerable degree stuck with their status as U.S. residents. Transactions in which at least 60 percent of the company’s stock remains in the same hands as previously, and in which its new legal home is not a country in which it afterwards is conducting “substantial business activities,” effectively are caught by the statute. However, even if these rules pervasively defeat fake expatriations that are merely tax planning arrangements, the expatriation of a given U.S. multinational is almost certain to be feasible if it is genuinely purchased by new (p.36) owners, be they private equity funds or distinct foreign companies with their own shareholders and managers.

B. Source

1. Underlying Conceptual Problems

All countries with income tax systems distinguish between income that is deemed to arise at home and that which is deemed to arise abroad. I will call this the issue of “source,” whether the question at issue is (1) classifying an item of gross income as either domestic source or foreign source; (2) determining whether a deduction reduces domestic source income or foreign source income; or (3) for a multinational group of corporate affiliates, determining how its global income is divided for tax purposes as between domestic and foreign members of the group.

As a matter of formal U.S. income tax terminology, only the first of these questions involves “source.” However, the other two questions sufficiently overlap with it, as a matter of practical consequences, to make my broader usage of the term “source” conceptually clearer, and therefore more convenient. One can easily see how the treatment of deductions, which the U.S. rules may describe as being apportioned or allocated between domestic source and foreign source income, rather than as themselves having a “source,” occupies the same substantive territory. However, the same is largely also true for issues in category (3). As we will see below, rules that determine the income split between, say, a U.S. parent and its foreign subsidiaries often effectively determine the amount of the group’s U.S. source income—or, failing that, the amount of group income that the U.S. will currently tax.4

In a world of multiple sovereigns, each supreme within a geographically defined realm, source-based income taxation comes as no surprise. Partly, this reflects legal convenience. In any case where one cannot, or at least does not, tax income on a residence basis, source is the only jurisdictional hook that one has left. But source-based taxation is not just legally convenient. The concept also has great intuitive power. Its ruling metaphor is that of people who are physically located in a jurisdiction and who are doing things (i.e., earning income) there.

Once we accept this ruling metaphor, the inevitability of source-based taxation, in a world where governments have physically defined realms of sovereignty, seems almost beyond question. Why wouldn’t a government impose taxes (including income taxes, if otherwise levied) on economic activity within its borders? After all, regulating or even banning particular activities within one’s borders is par for the course, and merely taxing them would seem to be a more modest endeavor. What is more, if residents are being taxed on their local economic activity, it may seem unfathomable that foreigners, once they have chosen to enter the jurisdiction, would be specially exempted. (p.37)

As we will see, however, the ruling metaphor of people physically located in a jurisdiction who are doing things there has a surprisingly weak connection to what source-based taxation actually means (or could mean) in practice. What is more, it will turn out that, even if source-based taxation could be made more workable than it actually is, the conclusion that it is a good idea would not necessarily follow. Indeed, if corporate income could be taxed directly to the individual owners, rather than initially just at the entity level, we’ll see that there would a compelling argument, not just that source-based taxation ought to disappear (at least in most cases), but even that perhaps it would largely disappear.

However, under actual real-world circumstances, including entity-level corporate taxation that makes residence-based taxation a shaky enterprise, even those who are normatively skeptical about source-based taxation tend to agree that it is politically inevitable (Shay, Fleming, and Peroni 2002, 89). Thus, in practice one has no choice but to set about devising rules for determining the source of income. It therefore is unfortunate that this task proves so challenging in practice. Given the point that the source of income simply is not a well-defined economic idea (Ault and Bradford 1990), any attempted implementation faces grave conceptual problems.

These problems are especially great once we start considering multinational firms, which are the main economic actors in international taxation. But even just in simple cases involving individuals, at least two core conceptual problems rear their heads.

Origin basis versus destination basis—Suppose we have a cross-border transaction in which neither party leaves his or her home. For example, I might be an architect in New York who uses the Internet plus Skype calls to design someone’s luxury apartment in Rome. For this design work, which might even include visual oversight of people’s labor at the job site, suppose that I am paid $100,000. Which country ought to be deemed the source of this income: the United States or Italy?

Part of the apparent perplexity in this example may relate to the manner in which modern technology permits me, so far as Italy is concerned, to be there and yet not quite there. But even in a more old-fashioned example—say, if I construct furniture that I ship to my Roman customers—both the U.S. and Italian governments might feel that they had a compelling source-based claim. After all, the income from the transaction ineluctably depends on conditions in both jurisdictions, and could not have been realized absent both.

For any export of goods or services from one jurisdiction to another, it is basically just a matter of convention how we define source. Under what is called an origin basis tax, the producer or exporting country—the United States, in the above examples—gets to levy the tax. Under what is called a destination basis tax, the consumer or importing country—Italy, given the apartment in Rome—instead gets to do so.

In principle, over the long run, these should yield the same global division of the tax base. A country that could not engage in trade would get to consume just what it produced, no more and no less. With trade, it can swap some of its production for that in (p.38) other countries. However, since trade is reciprocal, this should not alter the long-term equivalence between what its people produce and what they get to consume. Thus, at least roughly speaking, taxing domestic production and domestic consumption should eventually lead to about the same place.5

As it happens, consumption taxes, such as value-added taxes (VATs), typically use the destination basis, although some proposals to adopt a VAT-like U.S. tax would instead use the origin basis (see Shaviro 2004, 107). Income taxes generally use the origin basis, reflecting that (as we will see in chapter 6) use of the destination basis would create timing issues with respect to capital outlays, or those that are not supposed to be currently deductible. But, as we will see later in this chapter, many U.S. states, for purposes of their income taxes, use something that loosely resembles the destination basis to determine their share of multistate businesses’ U.S. income, and a prominent recent proposal (Auerbach 2010) would switch to the destination basis after substantially modifying the existing U.S. corporate income tax.

Services versus property—Adopting the origin-basis convention leads to a clear answer as to source when an individual, working in a given location, provides services to consumers anywhere in the world. But things can rapidly grow muddy if one is able to steer one’s efforts into the weeds of the long-standing income tax classification problem that lies at the boundary between selling (1) labor services and (2) property that one has created through one’s labor.

For sales of property, the U.S. tax rules used to rely simply on where title passes, but now apply a complex web of provisions in which a whole range of different factors may determine the outcome, depending on the type of property involved and other relevant circumstances.6 In some cases, the seller’s residence determines the source, meaning that a U.S. seller will inevitably have U.S. source income, at least on the initial sale (such as to an exporter). But there are various complicated exceptions—pertaining, for example, to inventory, depreciable personal property, and intangibles.

Even just for individuals, accordingly, the source rules can founder on conceptual problems that create planning opportunities. But the problems grow exponentially harder once we shift to the setting of multinational firms that conduct integrated operations around the globe. The reasons these problems are so hard include the following:

  • Multinationals’ profits often depend more on the value of the intangibles that they own than on the currently observable production activities that they conduct in particular locations. While the intangibles presumably reflect someone’s past labor or creativity, the time and place when their value arose (or became foreseeable) may be hard to identify.

  • As Ronald Coase (1937) famously observed, firms (including multinationals) arise in settings where transaction costs make command-and-control a more efficient operating mechanism than ongoing market transactions with respect to integrated economic production by multiple parties. Firms therefore generally (p.39) do not have observable internal market transactions based on arm’s-length prices. This makes it difficult to tell how profits or losses should be apportioned when a multinational firm engages in both U.S. and foreign operations. For example, there are no reliable market prices when one affiliate transfers goods and services to another. Likewise, internal financing choices, whether with respect to intra-group capital flows or borrowing from third parties, may not reliably indicate anything about where value and profits are actually being generated.

  • The lack of true and observable arm’s-length transactions extends not only to dealings between corporate affiliates, but also to those between employee-owners, such as founders, and the firm itself. Thus, consider Apple and Steve Jobs, during his many years at the helm, or Facebook and Mark Zuckerberg. Each was grossly “underpaid” by his firm for helping to create billions of dollars in expected profits, other than through the effect on the value of his shares. This, however, left the firm as the main taxpayer, giving it the opportunity to locate abroad, for tax purposes, profits that arguably reflected the work these individuals did while living in the United States.

2. The U.S. Rules’ Approach to Sourcing Multinationals’ Income

In describing how the U.S. rules address (and, many would agree, badly botch) the inevitably difficult task of determining the U.S. source and foreign source components of multinationals’ income, two main topics merit particular attention. The first is transfer pricing, or how one determines the intra-firm prices that U.S. and foreign affiliates are deemed to have charged each other. The second is how deductions incurred by the various members of a multinational group are treated.

Transfer pricing—To give a general sense of the intra-group pricing problem, suppose that Acme Products’ U.S. and French business operations, if conducted by separate owners, would earn $10 million each, for a total of $20 million. Given, however, the synergies that the two affiliates happen to realize by reason of their being jointly owned—pertaining, for example, to shared management, financing, and access to valuable intangibles—suppose that the unified firm earns $25 million. One really has no principled basis for determining the source of this synergy income (Shaviro 2009a, 108). What is more, this example, rather than being unusual or contrived, is fundamental to the existence of multinationals, which arise precisely to exploit synergies from common ownership where this works better than market-based arm’s-length interactions (Avi-Yonah, Clausing, and Durst 2008, 5).

The U.S. tax rules, in common with those of most other countries, address these issues through transfer pricing, which involves attempting to require that companies’ affiliates in different jurisdictions report the prices they would have charged each other if acting at arm’s length. The starting point in transfer pricing is treating even wholly owned corporate subsidiaries as separate from their corporate parents—reflecting legal form, but (p.40) not practical economic reality. The question then becomes one of asking what prices they “really,” even if unobservably, must have charged each other when contributing goods or services to each other. Thus, suppose Acme United States spends $10 producing a widget, which Acme-France then sells locally for $40. The income split for the $30 of group-wide profit depends on the arm’s-length transfer price that Acme United States is deemed to have charged Acme-France. Thus, if the correct transfer price is $18, the income split would be $8 United States and $12 France, whereas if it is $23, the split is $13/$7.

If one can determine what price the U.S. seller would have charged an unrelated French buyer in an otherwise identical transaction, that supposedly comparable price must be used for tax purposes in the related-company setting. But if the non-arm’s-length relationship creates extra profit by reason of intra-group synergies, then by definition there cannot be market price evidence of how arm’s-length parties would have agreed to split it. Even if unrelated parties were magically handed the unavailable extra profit, subject only to their agreeing about how to split it, they would face a bilateral monopoly bargaining problem, the outcome of which might be highly unpredictable. Thus, if Acme United States and Acme-France could make $10 million each operating separately but $25 million total if operating in tandem, all we could really say about the hypothetical income split that they would reach at arm’s length is that each side would surely end up with somewhere between $10 million and $15 million of the total, with the exact breakdown depending on how the negotiations shook out.

In practice, therefore, transfer pricing is at once a puzzle without a correct solution and a sinkhole for substantial tax planning, compliance, and administrative expense (albeit providing gainful employment for numerous lawyers, economists, and accountants). Nonetheless, things would not be quite so dire if tax planning games were limited to keeping theoretically indeterminate synergy income disproportionately out of the United States. Instead, as we’ll see further in chapter 3, in practice the observed problems are considerably worse than this. For example, U.S. companies’ tax haven affiliates, which in practice often are little more than incorporated post office drops with their own bank accounts, often end up claiming large profit shares. Yet their true contribution to the firms’ pre-tax profits may actually be negative, since, while operating them has a positive cost, they may actually contribute nothing to the firms’ operations, other than the opportunity to reduce U.S. (and foreign) taxes.

Part of the blame for transfer pricing’s working so poorly in practice lies with its central “comparable price” idea. For decades, the core practice in transfer pricing has been to look for real-world transactions that were similar to those occurring within the firm for which one needs to determine the transfer price, except that they involved unrelated third parties. At one time, this approach may have worked reasonably well. Michael Durst (2007, 1048) notes that, back in the 1920s, “available transportation and communications technology did not permit close centralized management of geographically diverse groups. Therefore, members of multinational groups functioned largely as independent entities, and benchmarking their income or transactions based on uncontrolled comparables (p.41) probably made good sense.” With improved communications technology, however, not only did central headquarters acquire the ability closely to control worldwide operations in real time, but a kind of market segmentation occurred that was directly adverse to the tax law’s reliance on comparable arm’s-length transactions:

[I]‌n those industries and markets where common control poses advantages, it typically is economically infeasible to remain in the market using a noncommonly controlled structure (for example, by maintaining distributors that are economically independent of manufacturers). In those markets in which multinational groups operate (that is, in those markets in which transfer pricing issues arise), it is unlikely that reasonably close uncontrolled comparables can be found. For example, today… there are no independently owned distributors of mass-market automobiles in the U.S.; all of the distributors are owned by their manufacturers. (1049)

Companies were quick to react to the rise in tax planning opportunities. By the early 1960s it had become “standard practice” (Hellerstein 1963, 165) to establish tax haven subsidiaries that would attract substantial shares of multinational businesses’ worldwide income. Moreover, it was already clear by this time that intangible assets, such as “‘knowhow,’ secret processes, the use of patents, providing personnel and a whole variety of technical and advisory services” offered especially rich opportunities for income shifting to low-tax jurisdictions, as “the reasonableness of the prices paid for these services and property… are customarily shrouded in a good deal of mystery” (165).

While these problems have not changed in kind since the 1960s, they have grown vastly more acute, reflecting ongoing advances in worldwide economic integration, along with the ever-rising importance of intangibles in global economic activity. Intellectual property, not bricks and mortar equipment, is increasingly the dominant factor in global economic production. And the transfer pricing rules cannot entirely impede shifting such property to low-tax jurisdictions long before it has demonstrably exhibited the economic value that its developers expect it to have.

In this brave new world of unique high-tech products, there frequently are no plausible arm’s length comparables. Making things worse, many industries in which production has been globalized are wholly populated by multinational firms, such that even generally similar goods and services that arise within the stages of the production process are never sold at arm’s length. Then the search for comparable prices often amounts to little more than a creative and flexible game of “find the strained analogy.”

By reason of this problem, the comparable price method has largely been abandoned with respect to intangibles—unfortunately, however, without yielding great improvement. For a number of years, companies such as Apple and Google achieved astoundingly favorable results, with respect to new technologies developed in the United States, by exploiting transfer pricing regulations that pertained to “cost-sharing.” For example, with respect to the iPhone, Apple appears to have done the following (see Sullivan 2011b). (p.42) First, it gave money to its subsidiaries that were located in tax havens such as Bermuda. Next, it had the subsidiaries give this money back to the U.S. parent, in a supposedly arm’s-length deal in which they helped pay for developing the iPhone (through the work of engineers living almost entirely in the United States), in exchange for the parent company’s giving them the right to market the iPhone overseas. Such methods have permitted Apple to treat about 70 percent of its global income as foreign source (Duhigg and Kocieniewski 2012), even though nearly all of its key value-creating work appears to have been done by people living in the United States (Sullivan 2011b). While the cost-sharing regulations were revised in 2009, apparently with an eye to addressing such techniques, U.S. companies may still retain “almost frictionless flexibility” to achieve comparably favorable results, with respect to intangibles that are actually developed at home, through alternative permitted valuation methods (Brauner 2008, 157).

Dismay about transfer pricing has led many commentators to suggest its replacement by a method called formulary apportionment, under which observed or hypothetical intra-firm transactions would drop out of the picture altogether. Instead, the proportion of a multinational firm’s global income that was treated as U.S. source would depend purely on a mechanical computation that was based on the firm’s U.S.-to-foreign proportion of one or more of the following objectively determined factors: (1) its property (typically limited to tangible items that have a determinate location); (2) its payroll, that is, wage payments to employees in different countries; and/or (3) its sales.

In illustration, suppose that Acme Products had $12 billion of global income, and that exactly two-thirds of each of the above three factors was located at home. Under a formulary approach that was based on any one or more of the three factors, Acme would be treated as having U.S. source income of $8 billion, and FSI of $4 billion. Or suppose Acme was a multinational corporate group, structured so that each affiliate was active only in its home country. Then the technical result of using formulary apportionment might be that the U.S. parent had $8 billion of income, all of it U.S. source, while the foreign subsidiaries had $4 billion of income that was not directly taxable by the United States (albeit effectively taxable as FSI of the U.S. parent upon actual or deemed repatriation).

For many decades, U.S. states with corporate income taxes have generally used formulary apportionment, rather than transfer pricing, to identify the in-state component of U.S. firms’ national taxable income. Historically, it was most common for states to give equal weighting to the property, payroll, and sales factors. More recently, however, there has been a trend toward relying more (or even exclusively) on the sales factor, so as to avoid discouraging in-state investment and employment (see Avi-Yonah, Clausing, and Durst 2008). The European Union has also been moving toward the use of formulary apportionment by European firms, at least as a permissible method (see Weiner 2007).

As we will see in chapter 3, formulary apportionment, whether better or worse on balance than transfer pricing, is itself no panacea. For example, it can yield clearly incorrect or inappropriate results as to the source of income, and it is exploitable through tax planning (see, e.g., Altshuler and Grubert 2010). Moreover, even those who view it as (p.43) clearly superior to transfer pricing often assert that its adoption should await broad international agreements to employ a common formula—a scenario that currently appears remote at best.

Treatment of Deductions—Further difficulties in sourcing multinationals’ income, or otherwise affecting their domestic tax liability, pertain to deductions for outlays that may help produce gross income in multiple locations. Examples include (1) the central headquarters expenses of managing a worldwide group; (2) research and development (R&D) expenditures aimed at developing intangible assets that will be deployed in multiple countries; and (3) interest expense paid for the use of funds, either to third-party lenders or to affiliated companies that have made intra-group loans. The question presented is whether a given deduction should be taken against domestic or foreign source gross income, when it may have contributed to generating either or both.

Most countries base their treatment of deductions purely on where (and by which group member) the items were incurred, and do not attempt to match them against the particular source of the income that they helped to generate. Thus, all otherwise allowable expenses that are incurred locally, but only such expenses, are deducted against domestic income. The United States, however, departs in a key way from this practice. While likewise treating foreign affiliates as separate taxpayers, with the consequence that only the U.S. affiliates’ outlays are potentially deductible, the U.S. rules in some respects take account of foreign subsidiaries for purposes of determining whether the U.S. companies’ deductions will reduce their domestic source or their foreign source income.

In general, for a U.S. company’s deductions that may have benefited foreign operations but that have no specific causal link to a specific item of income, several distinct sets of complicated rules require allocation and apportionment on some pro rata basis as between reducing U.S. source and foreign source gross income. For most expenses of a U.S. company that relate equally to all of its gross income, there is very broad discretion regarding the “pro rata to what” question.

Suppose, for example, that a U.S. multinational’s headquarters provided support to all of its global operations. In allocating the headquarters expenses between U.S. source and foreign source income, the company could use the domestic to foreign ratio of, among other possibilities, its gross income, sales, assets, or expenses incurred that had a clear relationship to particular income items.7 If assets are used, the foreign to domestic ratio can depend on their basis for tax purposes or their fair market value, again at the taxpayer’s discretion.8

If income is being used to allocate deductions, the U.S. company’s CFCs enter into the picture insofar as they may yield foreign source income, such as by paying dividends. If asset ratios are being used, shares of the CFCs’ stock count as foreign assets of the U.S. company. Thus, suppose that Acme International is a U.S. multinational that incurs $10 million of headquarters expenses and elects to allocate them pro rata to asset basis. If the basis of its stock in its CFCs is 40 percent of its overall asset basis, then $4 million (p.44) of its headquarters expenses will be treated as reducing foreign source, rather than U.S. source, income.

The same approach partly applies to U.S. companies’ R&D expenditures, but with an extra wrinkle. Rather than use the above approach for all of the R&D expenditures, a fixed percentage (25 or 50 percent, depending on the taxpayer’s choice of apportionment method) is sourced based on where they were predominantly incurred, with only the remainder being subject to the pro rata method. This part-fish, part-fowl R&D rule appears to be intended as a compromise between the usual pro rata rule and the U.S. agenda of encouraging domestic R&D, by permitting it to reduce domestic rather than foreign source income.

For interest expense—raising the most important deduction source issues given the ubiquity of both intra-group and third-party borrowing—the rules are generally similar, but in a key respect more complicated. Specifically, they look inside a U.S. company’s foreign subsidiaries, as the headquarters and R&D rules do not, for purposes of gauging the proportionality of domestic to foreign interest expense. This in turn must be done relative to assets, rather than involving taxpayer discretion to use, say, income or sales instead.

To illustrate how the interest allocation rules work, suppose that parent company Acme United States has exclusively domestic assets, with an adjusted basis of $1,000, except that it also owns the stock of a foreign subsidiary, Acme China, which the rules treat as a foreign asset. If Acme U.S.’s adjusted basis in its Chinese subsidiary’s stock is $500, then its overall assets are two-thirds domestic for purposes of the interest (or any other) allocation rules.

If the interest allocation rules worked like those for headquarters expenses, then two-thirds of Acme U.S.’s interest expense would be treated as reducing its foreign source, rather than its domestic, income. The interest rules depart from this approach, however, by considering Acme China’s interest expense for purposes of determining domestic to foreign proportionality.

Suppose Acme United States has interest expense of $80, while Acme China has interest expense of $25. The total is therefore $105, of which only two-thirds, or $70, can be treated as reducing domestic source income. Accordingly, while Acme United States’ deductible interest expense remains $80, the rules require treating $10 of that amount as reducing its foreign source taxable income.

Now suppose instead that Acme China’s interest expense totaled $55. Accordingly, the global Acme total is $135, and seemingly $90 of interest expense should be treated as reducing U.S. source income. This does not happen, however. Acme China is not a U.S. taxpayer, and thus its expenses have no direct effect on Acme United States’ taxable income under the U.S. rules. Instead, all that happens is that Acme United States need not treat any of its $80 of interest expense as reducing foreign source, rather than domestic, income.

By thus considering foreign subsidiaries’ interest expense for the limited purpose of assessing the group-wide domestic to foreign proportionality of U.S. companies’ interest (p.45) expense, the interest allocation rules take a step away from separate company accounting and in the direction of what might be called full global consolidation. The rules still fall short of the latter, however, not just because only the U.S. group members’ items are directly includable or deductible for U.S. tax purposes, but also by virtue of their only considering foreign subsidiaries of U.S. companies, as distinct from all foreign affiliates.

Thus, in the above example, suppose Acme China were the corporate parent and Acme United States the subsidiary. Or suppose Acme United States and Acme China were merely sibling affiliates in a group with a Bermudan parent. In either of these scenarios, since Acme United States would not own any of its foreign affiliates’ stock, its assets would be 100 percent domestic, and the affiliates’ interest expense would be entirely disregarded for U.S. income tax purposes. Thus, all of Acme United States’ $80 in interest expense would reduce its domestic source taxable income, even if its foreign affiliates had no debt or interest expense whatsoever.

To call the interest (and other) allocation rules exceptionally complex, both on their face and in how they apply to inevitably intricate facts, would be if anything an extreme understatement. While no other country attempts anything similar, many countries evidently share the U.S. concern about multinationals’ use of domestic interest deductions to fund operations that produce foreign source income, thus effectively stripping out income from the domestic tax base. To protect themselves, they commonly apply “thin capitalization” or “earnings-stripping” rules, under which interest deductions may be disallowed if excessive relative to equity or domestic gross income—and in particular when paid to non-residents (Ault and Arnold 2004, 411).9 While such rules do not directly pertain to source, and may apply to companies with purely domestic operations that are funded by a high ratio of debt to equity financing, they can reasonably be viewed as a functional substitute for the U.S. response to concern about excessive domestic as compared to foreign leverage.

3. The Tax Advantages of Being a Multinational

As both the transfer pricing and the interest allocation sagas make clear, the source rules permit a great deal of profit shifting by U.S. companies. Reported profits can readily migrate, not just to lower-tax countries in which significant productive activity may be occurring, but to tax havens that often lack sufficient resources, such as land and population, to support more than a trivial amount of such activity.

Suppose that all one needed to do, in order for profit shifting from home to foreign tax havens to be effective, was to establish a shell subsidiary with a postal box and a bank account, in some convenient jurisdiction such as the Cayman Islands. This might make the source-based U.S. corporate tax wholly unfeasible, even as applied to purely domestic businesses that were owned and operated by U.S. individuals. Shifting profits abroad, in this scenario, might function almost as if it were an explicit election for all U.S.-incorporated businesses, permitting their owners to avoid any U.S. (or other) tax on (p.46) their profits until they needed to withdraw the funds—and perhaps not even then, if they could simply borrow in consideration of the value of the stock.

In actuality, however, things are not quite so simple. Thus, Clausing (2011, 1580) finds that only about 30 percent of the domestic U.S. corporate tax base, rather than all of it, ends up escaping proper attribution under the source rules. This evident limitation on effective profit shifting both shows and reflects that the U.S. source rules are not entirely—even if they are often substantially—empty and manipulable. And it reflects in particular the fact that, to a considerable extent, successfully gaming the U.S. source rules requires that the taxpayer actually be a multinational company in practice, with discernible operations both in the United States and abroad, rather than just pretending to be a multinational on paper (as in the case where all that one owns overseas is a shell corporation in the Caymans).

Once a U.S. company has significant foreign operations, giving at least surface plausibility to the claim that a sizeable share of its profits may actually have been earned abroad, overstating the true foreign operating component may only be the first step. As Kleinbard (2011b, 754–755) notes, once profits have been labeled as FSI rather than as domestic source income, shifting them to a tax haven tends to become far easier. For example, a foreign subsidiary that is more than just a mail drop with a bank account can more easily claim a high transfer price on inter-company transactions with the U.S. parent. And once one has shifted the profits to an actually operating foreign affiliate, “straightforward earnings stripping technologies that are unavailable for domestic income can be used to move that income to a low-tax affiliate” (755).10

For this reason, actual foreign economic activity, even if it must be located in non-haven countries, serves for U.S. companies the role of tax-planning complement to activity in the United States. Or, as Kleinbard (2011b, 755) more colorfully puts it, U.S. multinationals have an “odd incentive… to invest in high-tax foreign countries, to provide the raw feedstock for the stateless income generation machine to process into low-taxed… [foreign] earnings” that may end up never being taxed anywhere.

Kleinbard (755–757) argues that U.S. (and perhaps other) multinationals are able, by reason of this practice, to derive “tax rents,” or above-normal after-tax returns that are not competed away by global capital flows. Ordinarily, if Country A has a lower tax rate than Country B, one would expect A to attract global capital at B’s expense until the marginal pre-tax returns that it was offering were sufficiently below those in B for the two countries’ after-tax returns to be equalized. After all, investors would have something to gain by shifting funds from B to A until this condition prevailed. However, if multinational firms, unlike purely domestic businesses, can uniquely escape this process without becoming the marginal investors that determine prevailing relative pre-tax returns in different countries, then they can get B’s higher pre-tax return without actually paying B’s tax rate.

Alternatively, multinationals might collectively dissipate these tax rents by competing with each other sufficiently to become the true marginal investors that determine (p.47) prevailing relative pre-tax returns. However, this would not relieve purely domestic businesses from facing a significant tax disadvantage in comparison to the multinationals. Either way, the Coase story, in which multinationals arise where they can lower the transaction costs associated with cross-border economic activity, ends up being supplemented, and perhaps even in some cases replaced, by a “tax synergy” story in which the tax advantages that can be derived from operating in multiple countries end up fueling the use of multinationals.

4. Determining the Source of Portfolio Income Earned by Investors

A corporation’s tax residence may matter for purposes of the source rules even when it is not itself the taxpayer. Consider its portfolio investors, or those owning non-controlling financial interests in it such as stock and debt, respectively yielding dividends and interest income. For those investors’ portfolio or passive income, the source question is even harder to resolve satisfyingly than it is with respect to the company’s active business income, which at least may have some discernible connection to where economic activity actually takes place.

Suppose the rules simply flowed through entity-level source determinations, making them applicable at the portfolio investor level as well. Thus, if I were a shareholder in Acme Products and received a $100 dividend, the source of this income would depend either on the company’s overall ratio between U.S. source and foreign source income over a given period, or else on how my funds (or those from the original issuance of my stock) were deemed to have been used.

Attempting to base the source of portfolio income on such entity-level determinations would be administratively challenging, to say the least, but in any case it generally is not attempted. Perhaps the logic of treating corporations as separate taxpayers from their owners encourages not looking too intently within the entity when its own tax treatment is not the thing at issue. Rather, the question of apparent interest might be thought of as where the investor notionally “went” in order to earn the portfolio income. However, given the difficulty of devising any substantively meaningful answer to this question, the U.S. rules, and those of numerous other countries, respond with arbitrary formalism.

In general, for U.S. federal income tax purposes, the source of interest or dividend income depends on the residence of the issuer. Thus, debt and equity issued by a U.S. corporation yields U.S. source interest and dividends, while that issued by a foreign corporation generally yields foreign source income (with an exception for dividends paid by foreign corporations that earn too high a percentage of their gross income in the United States).

There is a different rule, however, for income from notional principal contracts (NPCs), in which parties agree to offsetting payment obligations based on the returns that each would have earned on an underlying notional amount if invested in a particular way. Parties A and B might agree, for example, that A will owe B the amount that (p.48) $10 million would have yielded during a given year if invested in a fixed rate debt instrument, while B will owe A whatever the same amount would have earned if invested, say, in a given company’s stock (taking account both of dividend payments and changes in the market price of the stock). At the end of the period, they settle by netting the obligations against each other, leading to a single payment by the party that owed more. The source of any such payment is the residence of the recipient, rather than that of the payer.

There also are financial instruments that defy clear characterization as fitting any such present law category as equity, debt, or NPC. For income from such instruments, source questions may be impossible to resolve with any confidence, using existing precedent (Walker 2009).

There are two different senses in which these rules could potentially cause the legally determined source of income from portfolio investments to verge on being elective with the taxpayer, rather than depending, say, on where she is taking particular local economic risks. First, by choosing between issuers, one may be able to pick the preferred source outcome without much changing the actual set of economic risks that one is bearing. Second, one can choose between financial instruments that are governed by different source rules, again without much changing the actual economics. For example, as I further discuss in chapter 3, NPCs can be used in lieu of stockholding to bet that a particular company’s stock will do well, with the consequence of changing the applicable source rule so that it depends on one’s own place of residence, rather than that of the company.

A further important feature of the U.S. tax rules for U.S. source passive income responds to potential collection difficulties. In order to sidestep the constraint of lacking direct income tax jurisdiction over foreigners whose only U.S. connection lies in their owning the stock or debt of U.S. companies, the United States imposes a 30 percent withholding tax, collected from the payer, on their U.S. source dividend income (and certain interest). However, various bilateral tax treaties between the United States and foreign countries reciprocally reduce or eliminate the withholding tax for each other’s residents.

Tax planners often respond to the presence of these treaties through treaty shopping, or attempting to qualify under an especially favorable treaty even if the true beneficiaries have little or no connection with the treaty’s non-U.S. signatory. This in turn is combated by anti-treaty shopping provisions, both in the treaties and in U.S. tax law. These provisions typically attempt to ensure that the foreign country residence claim has some underlying economic substance, as opposed to simply reflecting one’s having, say, incorporated a conduit entity there that passes through all of its perfectly matched income and deductions.

C. Foreign Tax Credits

1. Mechanics and Incentive Effects of the Credit

Suppose a U.S. company conducts branch operations abroad. The income from these operations is currently taxable at home because it was earned by the company directly, (p.49) rather than by a legally distinct foreign subsidiary. In computing U.S. taxable income, the allowable business expenses of these foreign operations, such as for labor, fuel, and rent, are merely deductible. However, the foreign income taxes paid where the operations take place generally give rise to U.S. foreign tax credits, rather than merely to deductions. The company therefore effectively gets a 100 percent, dollar-for-dollar, refund from the U.S. government of its foreign taxes paid, assuming that no foreign tax credit limits apply.

To illustrate the importance of the distinction between credits and mere deductions, we can start with an example where foreign taxes are zero. Suppose that Acme Products’ foreign operations have gross receipts of $140 and deductible expenses of $40, yielding net income of $100. If Acme pays no foreign taxes, it would owe the U.S. $35 of tax at the currently prevailing U.S. corporate rate, and end up with $65 after paying that tax. Had Acme incurred one more dollar of deductible expenses, reducing its net income to $99, the U.S. tax would have been only $34.65, and Acme would have ended up with $64.65. Thus, given the effect on its U.S. tax liability of earning a dollar less, Acme’s extra one-dollar outlay ended up costing it only 65 cents. One could say, therefore, that it has a marginal reimbursement rate (MRR) for business expenses of 35 percent—equal, of course, to its marginal tax rate (MTR).

Now let’s change the example so that Acme also pays foreign taxes that are creditable in the United States. Suppose initially that the foreign tax rate is 20 percent. Acme therefore pays foreign taxes of $20 on its $100 of pre-foreign tax net income. Because the foreign taxes are creditable, they reduce Acme’s U.S. tax liability from $35 to $15, leaving it with the same $65 after paying all taxes as if the foreign rate were zero. Or suppose the foreign tax rate had been 30 percent, increasing Acme’s foreign tax liability to $30. Acme would pay only $5 after claiming foreign tax credits, and therefore still would end up with $65. In short, so long as U.S. foreign tax credits are allowable immediately and in full, Acme has a foreign tax MRR of 100 percent.

Where effectively applicable, a 100 percent MRR for foreign taxes eliminates all cost consciousness, or incentive by a U.S. taxpayer to care how much it pays. Indeed, with a 100 percent MRR, a U.S. taxpayer would lose out, after U.S. tax, from paying one dollar out of its pocket in order to avoid $1 billion in foreign taxes. Similarly, suppose that a third party asked a U.S. taxpayer to pay that party’s billion-dollar foreign tax bill, in exchange for just a dollar, and that the U.S. taxpayer would indeed be allowed to treat this tax payment as creditable, immediately and in full. Even if the dollar it received was taxable at home, the U.S. company, while losing $999,999,999 before U.S. taxes were considered, would end up 65 cents better off after U.S. tax. As we will see in chapter 3, there actually have been real-world transactions based on something resembling this idea, although the U.S. foreign tax credit rules use multiple mechanisms to impede them.

The story is more complicated where a U.S. company operates abroad through a foreign subsidiary, creating the possibility that deferral of the subsidiary’s FSI will apply. This affects both the mechanics for claiming foreign tax credits and, at least potentially, the true MRR. Let’s start with the mechanics. Given that a foreign company (even if (p.50) wholly U.S. owned) is not itself a U.S. resident, the foreign income that it earns is never literally or directly taxable in the United States. Rather, the U.S. parent is taxable when it realizes income, in its capacity as a shareholder, from the subsidiary. This could involve, for example, its receiving a dividend—or, as we will see, a deemed dividend under the so-called subpart F rules that limit the U.S. tax benefit of deferral.

Given this formally indirect approach to imposing U.S. tax on a foreign subsidiary’s foreign source income, the subsidiary’s business expenses are not literally or directly deductible by the U.S. parent. Rather, they are implicitly deductible, in the sense that they reduce the amount of taxable income that can be repatriated to the U.S. parent. This implicit deductibility, in turn, has two elements. First, every dollar of business expense that the subsidiary has to pay (such as for wages or rent) thereby becomes unavailable to be paid to the U.S. parent. Second, under U.S. tax law generally, amounts paid by companies to their shareholders are taxable as dividends only insofar as they do not exceed the company’s earnings and profits (E&P). For this reason, U.S. parents are supposed to keep track of their foreign subsidiaries’ E&P, and a distribution in excess of E&P merely reduces the parent’s basis in the payer’s stock.11

This brings us to the foreign subsidiary’s foreign tax payments. If they, too, received no direct consideration in U.S. tax law other than reducing available cash along with E&P, the result would be effective denial of foreign tax credits to U.S. companies that operate abroad through foreign subsidiaries rather than mere branch operations. To prevent this result, the rules allow U.S. parents to claim a deemed-paid or indirect foreign tax credit when they include actual or deemed dividends from their foreign subsidiaries.

To illustrate, suppose in the prior example that the $100 of pre-foreign tax income had been earned by Acme’s foreign subsidiary, rather than through branch operations, and that the subsidiary had paid $20 of income tax in the source country, leaving it with $80. If the subsidiary then paid $80 to Acme as a dividend, the deemed-paid foreign tax credit would be used to create the equivalent of direct creditability. Specifically, Acme would treat the foreign tax attributable to the dividend amount both as grossing up its U.S. taxable income and as generating an allowable credit. Thus, upon receiving the $80 of cash, Acme would have U.S. taxable income of $100, generating pre-credit U.S. liability of $35. However, the $20 foreign tax credit would reduce its U.S. tax bill to $15, leaving it with the same $65 of free cash as in the branch example.

Considered by itself, the deemed-paid or indirect foreign tax credit preserves tax equivalence between branch and subsidiary operations abroad, including with respect to the multinational group’s 100 percent MRR for foreign taxes paid. Deferral, however, makes the story more complicated. While deferral is a tax benefit that subsidiaries enjoy but branches do not, it potentially reduces the value of foreign tax credit claims, for reasons that I discuss in chapter 3. Indeed, if one never repatriates particular foreign source income, then the associated foreign tax credit will itself presumably never be claimed, thus reducing the MRR (as well as the U.S. MTR for the income) to zero. (p.51)

Despite deferral’s potential mitigating effect on foreign tax MRRs, foreign tax creditability remains so potentially generous that the existence of a plethora of provisions limiting its availability comes as no surprise. These provisions are of two main types: foreign tax credit limits, which prevent the credits from offsetting more than the U.S. tax otherwise due on particular foreign source income; and a host of what I call gatekeeper rules providing that particular taxes (or claimed taxes) are not creditable.

2. Foreign Tax Credit Limits

Under a long-standing though frequently changing set of rules, allowable foreign tax credits are limited to the amount of U.S. tax otherwise due with respect to one’s foreign source income. Excess credits are neither refundable (in the sense of generating a net cash payment from the U.S. Treasury) nor allowable to offset the U.S. tax due on U.S. source income. They can, however, be carried over to the previous taxable year or any of the next ten taxable years.

To illustrate the operation of the foreign tax credit limit, suppose that, in the earlier example, the foreign tax due on Acme’s $100 of foreign source income was $40, and that Acme had no other FSI. While it still would get to use $35 of U.S. foreign tax credits to reduce its U.S. tax liability on this income to zero, the extra $5 of credits would be disallowed—even if Acme also had U.S. source income on which it owed U.S. tax. Only if Acme had other FSI, unsheltered by otherwise available foreign tax credits, at any time during the carryover period would these credits be usable against its U.S. tax liability.

Foreign tax credit limits provide the reason why the source of a U.S. company’s taxable income can affect U.S. tax liability even if a given item (such as interest expense under the interest allocation rules) will appear on the company’s U.S. tax return in any event. Thus, when interest deductions of a U.S. company are treated as foreign source under the interest allocation rules, the reason this matters is that it reduces the limit, potentially causing the disallowance of credits that otherwise could have been claimed. Where this happens, the rules’ effect turns out to be arithmetically equivalent to directly disallowing the interest deductions that triggered such disallowance (Shaviro 2001). For U.S. companies that need not be concerned about credit disallowance, however, the interest allocation rules are effectively irrelevant.

Absent any ability to claim disallowed excess credits in another year, the credit limit, at the moment when it is reached, suddenly shifts the MRR for foreign taxes from 100 percent (ignoring deferral) to 0 percent. It thus (as in the interest allocation example) more generally has the effect of dividing U.S. taxpayers into two groups: (a) those that are “excess credit,” and thus may have an MRR as low as zero; and (b) those that are “excess limit” (i.e., have foreign source income that is unsheltered by foreign tax credits), and thus may have an MRR as high as 100 percent. This divide can powerfully influence foreign investment patterns as between U.S. firms, such as by inducing excess-credit firms to invest in low-tax countries so that their excess credits will become usable, while (p.52) excess-limit firms may prefer to invest in high-tax countries given their ability to make full use of the credits generated there.

As both sides of this divide suggest, one obvious taxpayer response to foreign tax credit limits is to seek “cross-crediting” opportunities. Suppose initially that the foreign tax credit limit applied solely with respect to the sum total of all one’s foreign source income. Thus, if Germany had a higher tax rate than the United States, causing disallowance of foreign tax credits when a U.S. company invested there, all the company would need to do is invest as well in low-tax countries, thus lowering the average foreign tax rate paid to the U.S. rate or lower.

To illustrate, recall the earlier example where Acme paid $40 of tax on $100, and suppose this happened in Germany. If Luxembourg had a 10 percent rate, Acme might respond by shifting $20 of income that it otherwise would have earned in the United States to arise instead (at least for tax purposes) in Luxembourg, on which it would pay $2 of Luxembourg tax. Absent the German tax problem (and ignoring deferral), this would have failed to reduce Acme’s worldwide tax bill, given the residual U.S levy on lightly taxed foreign source income. Now, however, Acme’s overall foreign source income would be $120, on which it would have paid $42 of foreign taxes—exactly the amount creditable under the U.S. rules, given the 35 percent rate and the foreign tax credit limit. Thus, Acme would in effect have “liberated” its excess foreign tax credits by using the low-taxed Luxembourg income to soak them up.

The U.S. rules respond by seeking to limit cross-crediting. The method they use is creating separate “baskets” of foreign source income, the taxes on which can only be used against income from the same basket. Thus, suppose Acme pays $40 of foreign tax on $100 of foreign source active business income, and $2 of tax on $20 of foreign source passive income (just as in the Germany-Luxembourg example). Since the U.S. rules generally place active business income and passive income in separate baskets,12 only $35 of the active business foreign tax credits would be allowed, and Acme would owe $5 of U.S. tax ($7 gross minus the $2 credit) on the income in the passive basket.

The motivation for dividing foreign source income into active and passive baskets is that income in the latter basket, since it can be earned almost anywhere given competitive global financial markets, is far easier to move for tax purposes to low-tax countries. For example, only so much active business income can plausibly be claimed to arise in the Cayman Islands, given its limited population and industrial capacity. But in theory, all one’s passive income could be located there for tax purposes, if one feels comfortable with investing one’s entire portfolio in Caymans banks and companies (that could in turn, as an economic matter, have investments around the world, thereby providing substantial diversification).

Thus, allowing taxpayers to pool low-tax passive income with potentially higher-tax active business income would make cross-crediting especially easy. Luckily for U.S. companies, however, there are various low-tax jurisdictions—for example, Ireland and Singapore—in which one may reasonably be able to place significant business assets, (p.53) relying on transfer pricing and financing patterns to increase the amount of income treated as arising there.

The attempted rigor of the U.S. response to cross-crediting has fluctuated substantially over the years. From 1932 through 1976, the U.S. rules contained various versions of a per-country limitation—preventing, for example, the use of excess French tax credits to offset the U.S. tax on German income. The rules have separated passive income from active business income since 1962. There has also been ebb and flow in the creation of other separate baskets designed to isolate particular categories of income and thereby limit cross-crediting. For example, the Tax Reform Act of 1986 created no fewer than nine separate FSI “baskets,” as between which cross-crediting was impermissible. At present, however, only the active versus passive distinction remains.

3. Gatekeeper Rules That Make Particular Items Non-Creditable

In addition to providing credit limits and at least modestly addressing cross-crediting, the U.S. rules respond to the potential for a 100 percent foreign tax MRR by providing that particular taxes, or claimed taxes, are not creditable. Examples of this approach include the following:

“Specific economic benefit” rule—The credit, by treating foreign taxes paid so much more favorably than other overseas business expenses, creates an incentive for U.S. taxpayers to seek to convert what would otherwise be merely deductible (or capital) outlays into creditable income tax payments. Suppose, for example, that building an overseas factory would cost Acme U.S. $100 million. Arranging instead to pay the local government a $120 million “tax” in exchange for its furnishing the building would potentially benefit both Acme and that government, at the expense of U.S. individuals. Treasury regulations respond by providing that payments to a foreign government, even when collected pursuant to its taxing power, are not creditable if received in exchange for a “specific economic benefit,” or one that is “not made available on substantially the same terms to substantially all persons who are subject to the income tax that is generally imposed by the foreign country.”13

Income taxes only—Another apparent policy response to the foreign tax credit’s extreme generosity is its being expressly limited to foreign income taxes, or those whose “predominant character… is that of an income tax in the U.S. sense.”14 Relevant factors in making this determination include whether a given tax generally uses realization accounting, requires inclusion of gross receipts, and allows deductions so as to produce a measure of taxable net income. This rule may help in screening out disguised fees for particular benefits, and it also ensures that other countries’ VATs will not be creditable against U.S. income tax liability. Otherwise, however, its rationale is not entirely clear.15

Other miscellaneous rules—Various other specific rules address particular gambits that respond to the foreign tax credit’s 100 percent MRR. For example, only a “compulsory payment” of foreign taxes, rather than one that is effectively voluntary, can give rise to a (p.54) foreign tax credit.16 Likewise, foreign tax liabilities that could potentially be challenged are not creditable until the U.S. taxpayer “exhausts all effective and practical remedies” with the foreign government.17 To address real-world cases in which U.S. taxpayers would buy foreign stocks just before a declared dividend was paid, and sell it immediately afterwards, in order to be able to claim foreign tax credits for the withholding tax that the host country was levying, there is now a 15-day minimum holding period for the foreign stock in such cases.18 To address so-called splitter transactions, in which taxpayers devised creative new ways to bring home foreign tax credits without having to repatriate the associated FSI, there is now a statute that suspends the credits if there has been a “foreign tax credit splitting event.”19 All this is an ongoing cops-and-robbers process, as new transactions necessitate the enactment of new rules limiting the allowance of foreign tax credits.

Anti-abuse and economic substance rules—The IRS also periodically responds to new foreign tax credit planning gambits by attempting to apply anti-abuse rules, such as the common law (and now statutory) requirement that tax-effective transactions have requisite economic substance and business purpose. Indeed, on such grounds the IRS unsuccessfully challenged both withholding tax seeking and splitter transactions,20 which legislative responses could address only prospectively. The need for anti-abuse rules is of course not unique to the foreign tax credit area, but the 100 percent MRR that credits can offer adds to the economic stakes.

D. Deferral

As noted above, U.S. multinationals operating abroad can use either a branch or a subsidiary structure. A branch is simply a set of operations in the source jurisdiction, conducted directly by the multinational firm through on-site employees and other agents (typically described as giving rise to a “permanent establishment” within the jurisdiction—a status that nonresident firms are often eager to avoid). A subsidiary, by contrast, is a separate legal entity, incorporated in the source jurisdiction but, in the pure case, wholly owned by the multinational firm.

For most practical business purposes, it makes no difference whether one operates through a branch or a subsidiary. Either way, the multinational firm has complete control over the local operations and ultimately owns all of the profits. However, using a subsidiary may permit the multinational to take advantage of limited liability with respect to the local operations. Its use may also have other (typically favorable) domestic legal consequences, insofar as the local government treats resident corporations differently from foreign ones and accepts that foreign multinational parents are not directly subject to its jurisdiction. For these among other (such as tax) reasons, multinationals typically use subsidiaries rather than branches in their overseas operations. The one big exception is banks, which regulators may require to use branches so that the entire global group will support local solvency. (p.55)

However close the branch and subsidiary modes of operation may seem economically, they look very different doctrinally from the standpoint of a realization-based income tax. Foreign branch profits ineluctably are part of the domestic parent’s current operating results. So far as realization is concerned, a U.S. company’s office in Paris is no different than an office in Phoenix. In the case of a profitable foreign subsidiary, however, all that technically has happened to the domestic parent is that its stock in a legally distinct entity presumably has appreciated. Thus, the realization doctrine, unless expressly modified to ignore the parent-subsidiary line, suggests ignoring what happened at the subsidiary level until the parent either sells the stock or receives dividends.

For this reason, foreign subsidiaries’ non-U.S. income has never directly been includable in the U.S. parents’ income until something happens to trigger domestic realization. The very fact that this is called deferral, however, seems to imply a harsher critique than one might normally expect from the application of something so familiar and widespread as realization doctrine. This choice of term presumably reflects both the well-understood arbitrariness of the branch versus subsidiary distinction and the fact that, in the eyes of the U.S. tax system, the subsidiary would itself have taxable income, rather than merely unrealized appreciation, if it were being classified as a U.S. taxpayer by reason of its (at least entity-level) ultimate U.S. ownership.

While deferral arose for formal or doctrinal reasons, its retention to this day is a deliberate policy choice, reflecting Congress’s reluctance to choose either pure worldwide taxation at one pole or exemption at the other pole. Deferral thus is effectively rationalized as a tax preference, given the otherwise worldwide character of U.S. residence-based taxation. Even proponents tend to view it as half a loaf toward achieving territoriality, rather than as making sense on its own terms.

Deferral’s formalistic origin still matters, however, in that it causes the tax preference of choice to have a particular structure that surely no one would have deliberately chosen from a design standpoint. Its two main features are that one only gets to defer one’s foreign earnings if one (a) earns them through a subsidiary and (b) avoids repatriating them. While (a) is immaterial to the extent that a given U.S. multinational would operate abroad through subsidiaries anyway, (b) has important practical implications that essentially no one in U.S. international tax policy debate considers desirable. After all, why would one want to discourage U.S. companies from bringing money home that they then might use domestically?

As it happens, the question of why and how deferral discourages the repatriation of foreign earnings is more complicated than it may initially seem. Thus, suppose that Acme Bermuda has $100 million of wholly untaxed profits, which would cause Acme U.S. to pay $35 million of current year tax liability if foreign subsidiaries were treated like foreign branches. Suppose that all of Acme’s global funds earn a 10 percent return, and that Acme would be forced for business reasons to repatriate its funds in Bermuda next year. Is there any benefit from the one-year delay, resulting here from deferral, in collection of the economically accrued $35 million liability? At a first approximation, the answer is no. (p.56) After all, in a year Acme Bermuda will presumably have $110 million (given the assumed 10 percent return), thus triggering a repatriation tax of $38.5 million. The present value of this tax at a 10 percent discount rate—$35 million—is exactly the same as that of the tax that deferral permits Acme to avoid today.

We will see in chapter 3 why deferral nonetheless is a substantial tax benefit. The answer turns on varying tax rates between years, and the fact that Acme might never have to repatriate its Bermuda funds, at least during the lifetime of the repatriation tax that is currently part of U.S. law. For now, however, suppose we simply assume that deferral highly benefits U.S. multinationals, and thus that they have “trapped earnings” abroad.

Under this assumption, almost the only plausible rationale for using such a zombie relic of formalistic realization doctrine, rather than a deliberately designed tax preference to achieve a compromise outcome somewhere in between pure worldwide taxation and territoriality, is political. The ongoing stalemate or ceasefire in place regarding U.S. international tax policy makes retaining deferral convenient, and suggests that any effort to replace it could lead to a political bloodbath (at least among competing lobbyists and their pet legislators) as the prospective winners and losers from rival replacement proposals jockeyed for position.

One peculiar political side effect of deferral’s perverse structure is that both sides in the worldwide versus territorial debate can point to “trapped earnings” abroad as an important reason for adopting their preferred changes. There would be no possible problem of trapped earnings if foreign subsidiaries’ income was taxable in the United States either immediately or never. In 2004, on the ground that trapped earnings were a problem, proponents of exemption won an important battle, though not (as it turned out) the war, through the enactment of a “tax holiday” providing that, for one year only, qualified foreign dividends would be taxed in the United States at only a 5.25 percent rate, rather than the usual 35 percent. The provision triggered a huge response, with the amount repatriated greatly exceeding official expectations. In the end, $312 billion worth of dividends came home for 2005, the holiday year (Redmiles 2008, 103), as compared to only $30 billion that would have been expected otherwise (Kleinbard and Driessen 2008).

Since U.S. tax law generally defines dividends as distributions made out of E&P, deferral requires U.S. shareholders to keep track of their foreign subsidiaries’ E&P, as defined under U.S. tax law principles. The amount of such foreign E&P currently exceeds $2 trillion.21 Its ongoing rapid growth reflects, not only the actual profitability of U.S. companies’ foreign operations, but the companies’ ability to label profits, wherever earned, as foreign source.

However significant the actual economic benefit of deferral may be, the managers of publicly traded companies often care more about accounting benefits, or the ability to state high reported earnings on their financial statements. In the words of a knowledgeable insider, “saving taxes is all very nice, but earnings per share make the world go round” (see Shaviro 2009d, 449). Thus, deferral might not affect U.S. multinationals’ behavior as much as it actually does, if it did not permit them to state higher current earnings. (p.57)

As it happens, generally accepted accounting principles (GAAP), which govern U.S. financial reporting, do not in fact permit companies to treat deferred taxes more favorably than current ones. For reasons that are difficult for non-accountants to understand, GAAP generally ignores time value considerations—even though any rational investor surely understands that a dollar today is worth more than a dollar in the future. Thus, in the Acme Bermuda example above, Acme would have to report a $35 million tax expense, for purpose of determining current year earnings, even if it anticipated benefiting from the deferral indefinitely. Only if it stipulated, to its auditors’ satisfaction, that Acme Bermuda’s funds were permanently reinvested abroad, would it get any accounting benefit as to current earnings. But with such a stipulation, it would be able to drop its reported U.S. tax liability, for financial accounting purposes, all the way from $35 million to zero.

Given the magnitude of this discontinuous benefit, it has become common practice for U.S. multinationals to report that the profits they have treated for tax purposes as arising in tax havens are permanently reinvested abroad. In practice, this may cause the companies to have “trapped earnings” abroad after all—trapped, that is, by the managers’ reluctance to take the earnings hit that would result from contradicting their prior stipulation to the auditors that a U.S. repatriation tax would never be paid.

Despite the tax and accounting stakes, taxable repatriations of U.S. companies’ foreign source earnings do in fact happen. In such cases, however, the key factor determining the magnitude of the U.S. federal income tax hit is how much in the way of foreign tax credits can be claimed. For example, if you bring home $100 million of foreign earnings but this frees up $35 million of foreign tax credits, the resulting U.S. tax liability is zero.

Needless to say, companies do not leave the U.S. tax hit purely (or even substantially) to chance. Instead, a “forced move” repatriation of earnings from either a high-tax or a low-tax country frequently prompts offsetting repatriations, designed to optimize the overall U.S. tax consequences. The common result from this interaction between deferral and the foreign tax credit rules is tax planning and compliance costs that are very high relative to the U.S. tax revenues that end up being generated.

E. Subpart F

The first clear proof that deferral has become a deliberate U.S. policy choice, not simply the logical consequence of a realization-based income tax, came in 1962, when the Kennedy administration proposed almost completely repealing it. In support of making the earnings of controlled foreign corporations (CFCs) immediately taxable to their U.S. parents, the administration argued that CFCs were effectively identical to foreign branches, and noted that deferral is simply a tax preference if one accepts worldwide taxation and branch equivalence. In denouncing deferral, the administration made classic anti-tax preference arguments, based on issues of fairness and tax neutrality, and also (p.58) expressed concern that tax-favoring U.S. companies’ outbound investment would reduce domestic U.S. investment and jobs.

The attack was in a sense two pronged. While the administration targeted deferral generally,22 much of its rhetoric and argumentation focused more narrowly on U.S. companies’ use or abuse of tax havens. These generally were small countries (such as Luxembourg or Bermuda) with two linked characteristics: extremely low (if any) income tax rates; and a lack of sufficient domestic resources, including population, to support significant levels of local economic production. In light of these characteristics, taxable income that U.S. companies attributed to havens might reasonably be suspected of reflecting economic activity that actually took place elsewhere. Under the source rules, however, such income shifting was hard to combat, given transfer pricing opportunities plus the ability to use intra-group debt to place interest deductions in high-tax countries (including the United States) and the offsetting interest income in tax haven subsidiaries.

Not surprisingly, this proposal prompted intense opposition from U.S. multinationals. In the main, however, rather than challenging the logic of branch equivalence or defending realization doctrine as applied to wholly owned foreign subsidiaries, the companies responded with their own competing policy claims—about worldwide taxation, as distinct from deferral as such. In particular, they argued that increasing the U.S. tax burden on outbound investment would gravely handicap their ability to compete abroad with non-U.S. companies that only faced local source-based taxes.

The companies were perhaps less eager than the Kennedy administration to discuss tax planning designed to shift income to tax havens. They could, however, note two defensive rationales in favor of these practices: (1) that foreign competitors were also shifting income to the havens, thus making it a part of remaining competitive; and (2) that thereby saving foreign, as distinct from U.S., taxes, was something that the U.S. Treasury should welcome, rather than condemn.

Congress responded to this closely fought cage match by splitting the difference. Rather than generally ending deferral, it enacted a set of rules commonly known as subpart F, under which specified types of CFC income are taxed currently to U.S. shareholders, through the mechanism of being treated as deemed dividends. These rules arguably focus on tax havens, but only indirectly. Rather than distinguishing between, say, countries like Germany and those that are more like Bermuda, the subpart F rules target income that is relatively mobile (and thus easy to locate in a tax haven), along with indicia of tax planning designed to shift income from one country to another.

For purposes of subpart F, CFCs are defined as foreign corporations that are more than 50 percent owned by U.S. shareholders, counting only such shareholders who own (including through attribution from related parties) at least 10 percent. Those owning less than 10 percent not only are disregarded in measuring U.S. control, but are not treated as receiving deemed dividends even if the ownership threshold is otherwise met for the foreign company. (p.59)

In defining the circumstances that give rise to deemed dividends, subpart F is quite wide ranging. For example, it addresses paying illegal foreign bribes, as well as participating in international boycotts, such as by Arab countries of Israel.23 Its principal effect, however, is to create the following three main categories of items that give rise to a deemed dividend:

  1. (1) passive income, such as interest and dividends on portfolio financial assets held by the CFC. One important application of this category is to intercompany financial flows, such as interest income that one CFC earns on debt owed to it by the U.S. parent or another CFC. Such related-party passive income generally is taxable under subpart F—reflecting the concern about income shifting—even though, from the standpoint of the entire worldwide group, one might view it as a wash. The subpart F rules do, however, provide for disregarding intra-group interest payments if the borrower and lender are in the same country.24

  2. (2) Overseas tax planning is even more clearly the focus of a second set of rules defining subpart F income. These rules address the “base company” scenario, in which a conduit entity, located in a tax haven, effectively diverts taxable income from higher-tax countries by serving as an ostensible middleman, using transfer pricing to claim a share of the overall worldwide group profit that exceeds its true value added.

    To illustrate, suppose that Acme Products produces goods for the European market, most of which would be sold in high-population but high-tax countries such as France and Germany. Rather than shipping the goods directly to its French and German affiliates, Acme might insert a Luxembourg subsidiary in the middle that ostensibly provided, say, storage and marketing services, and arrange the reported transfer prices such that this affiliate ended up with a significant slice of the group’s overall income from its export activities. This would increase Acme’s benefit from using transfer pricing to shift taxable income abroad. But even if the United States did a good enough job of transfer pricing enforcement to squelch such shifting entirely, Acme could still benefit abroad from the arrangement, by transfer-pricing income out of France and Germany.

    The subpart F rules address such arrangements as follows. When a CFC earns sales income with respect to property that is neither produced nor consumed within the country where the CFC operates (e.g., Luxembourg in the above example), such income is generally taxable as a deemed dividend to the U.S. parent.25 A similar rule applies when a CFC receives income from the performance of services when it is neither in the country where the services were performed, nor that where they were consumed.26 As a result, Acme will likely have no tax incentive to establish a Luxembourg affiliate, given that each (p.60) dollar of taxable income shifted out of France and Germany would show up as current U.S. taxable income.

  3. (3) A final major category of inclusion under subpart F pertains to investments by CFCs in U.S. property, such as real estate or stock in U.S. corporations. Such investments, when made by CFCs with unrepatriated profits, give rise to a deemed dividend even though they do not reflect direct receipt of the profits by the U.S. parent. One rationale for this rule might be that the case for perpetuating deferral, if based on permitting foreign operations to fund their growth through retained earnings, no longer applies when the earnings have come back to the United States. A second might be that treating U.S. investment as a taxable repatriation makes it harder for U.S. companies to avoid the residual U.S. tax indefinitely. For example, if they are concerned that keeping the funds abroad would expose them to foreign economic risk that they would prefer to avoid, U.S. real estate investment might be an attractive alternative, but for its giving rise to subpart F income. However, the rule’s efficacy in this regard is greatly reduced by the fact that a U.S. company’s foreign subsidiaries can hold U.S. dollars or U.S. government and private securities, thus effectively bearing U.S. in lieu of foreign economic risk no less than if they held U.S. real estate, without this being treated as a taxable repatriation (Linebaugh 2013, 1).

What should we make of subpart F as a whole? However murky and disparate its rules may be, it actually has a surprisingly coherent (whether or not persuasive) conceptual core. Again, the basic idea is that, when a U.S. multinational invests abroad, it should not face pre-repatriation U.S. tax so long as it not only keeps the money abroad but (a) is actually using retained earnings to expand its active business operations, rather than to hold an investment portfolio that can easily be placed in a tax haven; and (b) does not appear to be avoiding meaningful source-based taxes abroad by shifting its income from the “true” place where it was likely earned. The rules make no effort to answer the hard question of whether, for these purposes, it is U.S. or foreign taxes that either directly or ultimately are being saved. And, while seemingly aimed at tax haven income, they address this issue purely by reference to the type of income that is involved (or the structure used in earning it), rather than by focusing directly on where the taxpayer actually reports it.

However conceptually clear subpart F’s main structure and purposes might be, it is considerably less pellucid in practice. Indeed, it is “widely regarded as one of the most complex and difficult pieces of legislation in existence” (Isenbergh 2004, 70:8). This complexity has multiple causes, even beyond the difficulty of identifying either passive income or arrangements that raise suspicion about income shifting abroad. One set of issues relates to the need to keep track of foreign subsidiaries’ E&P, so that one can determine whether deemed distributions are dividends. Moreover, for deemed dividends no less than actual ones, one must determine the foreign taxes that are attributable to them, (p.61) while continually coordinating to make sure that neither earnings nor foreign taxes paid are counted twice. Finally, because U.S. shareholders who own less than a 10 percent interest, as well as foreign shareholders, are not subject to the receipt of deemed dividends under subpart F, one must take care to “assign the right earnings to the right owners at the right time. … The upshot [of all this] is a very busy statute” (Isenbergh 2004, 70:8).

F. Bilateral Income Tax Treaties

One last feature of U.S. international tax law that is worth briefly mentioning is the existence of bilateral income tax treaties, which the United States presently has with more than sixty nations. Under U.S. law, treaties have no higher standing than statutory law, and thus give way if contradicted by a later-in-time enactment, but this is relatively rare, reflecting that the treaties and the Internal Revenue Code generally reflect a shared or at least consistent vision.

While the treaties are not entirely uniform, their terms, along with those of other bilateral income tax treaties around the world, tend to have various common features. In general, the treaties state as their principal goal the “avoidance of double taxation and the prevention of fiscal evasion with respect to taxes on income” (United States Department of the Treasury 2006). In other words, they generally seek to ensure that cross-border activity involving the treaty partners is taxed exactly once, rather than twice or not at all. A key tool for preventing evasion is information sharing.

To prevent double taxation, the treaties generally use two distinct approaches. For business activity in a source country that is conducted by a foreign entity through a permanent establishment, they require the residence country to offer either foreign tax credits (effectively rebating the source country tax) or else exemption. By contrast, for passive income earned from portfolio investments in the source country, they often lean toward exemption (or at least a reduced rate of withholding tax) by the source country.

Even with these rules in place, however, double taxation or double non-taxation may result if source and residence countries apply different source rules, or otherwise measure income differently. Thus, suppose that Acme Products earned $100 million worldwide, all of it in either the United States or France. If both countries claim that $60 million was earned within their own borders, $20 million of Acme’s global income may effectively be taxed twice (if only in the sense that the U.S. foreign tax credit limit, if Acme is a U.S. company, will reflect FSI of only $40 million). Accordingly, fully avoiding double taxation and double non-taxation would require tax base coordination, not just foreign tax credits or exemption by one country or the other.

Tax treaties generally do not prevent tax base or source rule differences from arising. While they may offer mechanisms for abating it, such as by providing for consultation between the treaty partners’ “competent authorities” to address issues such as transfer (p.62) pricing, these mechanisms generally have limited scope. In particular, they can only address purely bilateral issues, such as the transfer price that is applied to a transaction between a given multinational’s affiliates in the treaty countries. Source rule differences of broader geographical scope—for example, those pertaining to interest deductions, which generally are not country specific, as they merely distinguish domestic source from foreign source income—generally cannot be handled this way.

Given the limited scope of bilateral sourcing reconciliation efforts even if the competent authorities are eager to reach agreement, a logical next step—if countries placed a high enough priority on ensuring that income wax taxed exactly once—might be the creation of a multilateral tax authority, akin to the World Treaty Organization (WTO) in trade matters, with a mandate to assure that income is generally included somewhere exactly once. In fact, however, while various multilateral institutions—for example, the Organization for Economic Coordination and Development (OECD), the International Monetary Fund (IMF), and the United Nations’ Committee of Experts on International Cooperation in Tax Matters—may offer advice and coordination in support of the one-tax end result, nothing like the WTO exists in international tax, nor does the creation of any such institution appear imminent (or likely to attract U.S. support). Even more limited initiatives, such as arranging common global approaches to formulary apportionment or the sourcing of interest deductions, have proven unachievable.

G. Summary

U.S. international taxation uses five core concepts: residence, source, foreign tax credits, deferral, and subpart F exceptions to deferral. Each, in different ways, rests on shaky foundations and leads to difficulties in practice. The residence concept works well, for the most part, with respect to individuals, but is poorly suited to the task of classifying legal entities such as corporations. The source of income lacks fundamental economic meaning, but is even worse in practice than one might expect from its theoretical limitations. The main tools used to exploit it are transfer pricing and the strategic use of borrowing to shift net income from high-tax to low-tax jurisdictions. Foreign tax credits can eliminate all cost consciousness by U.S. taxpayers with respect to foreign tax liabilities, and prompt a need for complex rules limiting their use or allowability. Deferral helps to ensure that tax planning costs are high relative to the revenue raised. The subpart F rules that limit deferral’s reach, mainly in situations where one might suspect that income is being shifted to tax havens, create still more pyramiding complexity in practice.

(p.63) Endnotes

(1) . I am referring to the winter 2013 edition of the Commerce Clearing House (CCH) compilation, where Code sections 861 through 999 go from pages 2551 through 2833. This admittedly includes not just the text of current statutes, but information about recent amendments thereto.

(2) . In the winter 2013 CCH compilation of income tax regulations, those pertaining to Code sections 861 through 999 appear at pages 45,940 through 46,359; 48,801 through 49,470; and 50,801 through 50,998, making for a total of 1,288 pages.

(3) . The issuance of IRS Notice 2009-78, 2009-2 C.B. 452, played an important role in tightening the anti-inversion rules’ application, addressing gaps and defects in the initial proposed regulations.

(4) . Even if income that is assigned to the U.S. parent ends up being classified as FSI, it will be currently taxable under the U.S. rules, and the source determination will matter solely for purposes of applying the foreign tax credit limitation. Income of a foreign subsidiary that avoids any direct U.S. presence is not directly taxable under the U.S. rules—although, as we will see below, in some cases it may in effect be currently taxable to the U.S. parent, by reason of its giving rise to a deemed dividend under subpart F.

(5) . This assumes that one is taxing production or consumption by the same people—and not, for example, taxing consumption by foreign visitors or failing to tax that by one’s own residents when they travel abroad.

(6) . See IRC section 865.

(7) . See Treas. Regs. sections 1.861-8(b) and 1.861-8T(c).

(8) . See Treas. Regs. section 1.861-8T(c)(2).

(9) . The United States also has a thin capitalization rule, but not an especially rigorous one. See Code section 163(j).

(10) . These earnings-stripping technologies rely on the existence and limited reach of subpart F, discussed later in this chapter and in chapter 3.

(11) . Once the basis of the subsidiary’s stock has been reduced to zero, further distributions generally are treated as giving rise to capital gain.

(12) . See generally Code section 904(d).

(13) . Treas. Reg. §1.901-2(a)(2)(ii)(B). Absent a generally imposed income tax in the foreign country, the regulation instead defines a specific economic benefit as one that is “not made available on substantially the same terms to the population of the country in general.” Id.

(14) . Treas. Reg. §1.901-2(a)(1)(ii).

(15) . The rule providing that only foreign income taxes are creditable was recently the subject of a U.S. Supreme Court decision. PPL Corp. v. Commissioner, decided on May 20, 2013. This decision directly affected only a few U.S. taxpayers, and its broader significance to U.S. international tax law appears to be limited.

(16) . Treas. Reg. §1.901-2(a)(1)(i).

(17) . Treas. Reg. §1.901-2(e)(4)(ii).

(18) . IRC section 901(k).

(19) . IRC section 909.

(20) . See Compaq Computer Corp. v. Commissioner, 277 F.3d 778 (5th Cir. 2001); IES (p.64) Industries, Inc. v. United States, 253 F.3d 350 (8th Cir. 2001); Guardian Industries Corp. v. United States, 477 F.3d 1368 (Fed. Cir. 2007).

(21) . Linebaugh 2013 notes a recent estimate of $1.7 trillion in U.S. public companies’ foreign earnings, counting only that portion, which (as discussed below) the companies have classified as permanently reinvested abroad.

(22) . The Kennedy administration proposed retaining deferral for investment in developing countries, but as a deliberate tax preference to promote their economic advancement.

(23) . See IRC sections 952(a)(3) and (4).

(24) . See IRC section 954(c)(3)(A).

(25) . See IRC section 954(d).

(26) . See IRC section 954(e).

Notes:

(1) . I am referring to the winter 2013 edition of the Commerce Clearing House (CCH) compilation, where Code sections 861 through 999 go from pages 2551 through 2833. This admittedly includes not just the text of current statutes, but information about recent amendments thereto.

(2) . In the winter 2013 CCH compilation of income tax regulations, those pertaining to Code sections 861 through 999 appear at pages 45,940 through 46,359; 48,801 through 49,470; and 50,801 through 50,998, making for a total of 1,288 pages.

(3) . The issuance of IRS Notice 2009-78, 2009-2 C.B. 452, played an important role in tightening the anti-inversion rules’ application, addressing gaps and defects in the initial proposed regulations.

(4) . Even if income that is assigned to the U.S. parent ends up being classified as FSI, it will be currently taxable under the U.S. rules, and the source determination will matter solely for purposes of applying the foreign tax credit limitation. Income of a foreign subsidiary that avoids any direct U.S. presence is not directly taxable under the U.S. rules—although, as we will see below, in some cases it may in effect be currently taxable to the U.S. parent, by reason of its giving rise to a deemed dividend under subpart F.

(5) . This assumes that one is taxing production or consumption by the same people—and not, for example, taxing consumption by foreign visitors or failing to tax that by one’s own residents when they travel abroad.

(6) . See IRC section 865.

(7) . See Treas. Regs. sections 1.861-8(b) and 1.861-8T(c).

(8) . See Treas. Regs. section 1.861-8T(c)(2).

(9) . The United States also has a thin capitalization rule, but not an especially rigorous one. See Code section 163(j).

(10) . These earnings-stripping technologies rely on the existence and limited reach of subpart F, discussed later in this chapter and in chapter 3.

(11) . Once the basis of the subsidiary’s stock has been reduced to zero, further distributions generally are treated as giving rise to capital gain.

(12) . See generally Code section 904(d).

(13) . Treas. Reg. §1.901-2(a)(2)(ii)(B). Absent a generally imposed income tax in the foreign country, the regulation instead defines a specific economic benefit as one that is “not made available on substantially the same terms to the population of the country in general.” Id.

(14) . Treas. Reg. §1.901-2(a)(1)(ii).

(15) . The rule providing that only foreign income taxes are creditable was recently the subject of a U.S. Supreme Court decision. PPL Corp. v. Commissioner, decided on May 20, 2013. This decision directly affected only a few U.S. taxpayers, and its broader significance to U.S. international tax law appears to be limited.

(16) . Treas. Reg. §1.901-2(a)(1)(i).

(17) . Treas. Reg. §1.901-2(e)(4)(ii).

(18) . IRC section 901(k).

(19) . IRC section 909.

(20) . See Compaq Computer Corp. v. Commissioner, 277 F.3d 778 (5th Cir. 2001); IES (p.64) Industries, Inc. v. United States, 253 F.3d 350 (8th Cir. 2001); Guardian Industries Corp. v. United States, 477 F.3d 1368 (Fed. Cir. 2007).

(21) . Linebaugh 2013 notes a recent estimate of $1.7 trillion in U.S. public companies’ foreign earnings, counting only that portion, which (as discussed below) the companies have classified as permanently reinvested abroad.

(22) . The Kennedy administration proposed retaining deferral for investment in developing countries, but as a deliberate tax preference to promote their economic advancement.

(23) . See IRC sections 952(a)(3) and (4).

(24) . See IRC section 954(c)(3)(A).

(25) . See IRC section 954(d).

(26) . See IRC section 954(e).