Featured Articles

News Analysis: What Should the OECD Do About Base Erosion?

Posted on September 10, 2013 by Lee A. Sheppard

This article originally appeared in the September 9, 2013 edition of Tax Notes.

By Lee A. Sheppard -- lees@tax.org



Tivoli Gardens, a tiny amusement park situated in central Copenhagen, is either a charming reminder of a more innocent time or an accident waiting to happen, depending on your point of view. Copenhagen was the site of the 2013 International Fiscal Association annual congress, the official color of which was fuchsia pink (yes, your correspondent did don a pair of pink tights one day).

Opened in 1843 and continuously added to, Tivoli is a sentimental mishmash of Coney Island and some old sets from Puccini's Turandot. A mosquito-ridden pond in the middle of the park hosts a pirate ship. The park is ringed with bars and restaurants. Tivoli's Orientalism extends to the newly revamped and pricey Nimb Hotel, which has a series of vaguely Middle Eastern onion-topped minarets. Peacocks and ducks have the run of the place.

With the exception of the truncated carousel, bits of which decorate buildings in the park, most of the rides consist of strapping the customers to wooden chairs and swinging them around high above the ground. Make that very high in the case of the tower with chairs attached to wires. Some of the rides appear to be a century old -- like the whirling biplanes attached to rotating arms.

Rides like these haven't been built in America for five decades -- and for good reason. The park's fake Chinatown lies underneath the tracks for the roller coaster, called the Demon. You read that right.

European tax and corporate laws are like Tivoli -- quaint, charming, antiquated, attractive nuisances just waiting to be abused by sophisticated multinational tax planners. Indeed, the striking thing about the OECD base erosion and profit-shifting action plan is how many of the proposed solutions would essentially call for rewrites of European domestic laws.

For the uninitiated, attractive nuisance is a tort law concept that ascribes to the proprietor of a dangerous instrumentality a duty to restrict access to it on the view that it will inevitably lure people, particularly children, to hurt themselves. An unfenced swimming pool is an attractive nuisance. An aging amusement park ride could be considered an attractive nuisance, but the counterargument is that customers assume the risk, because the machine is being used as intended.

This article looks at a selection of the proposed actions in the action plan, making recommendations and predictions. Needless to say, the former may be more aggressive than the latter. The predictions differ from the hopeful statements of the OECD action plan. Some Europeans are getting cold feet, and it is a safe bet that the Americans will do nothing.

And like the headline, much of this article argues that concepts found in Anglo-Saxon common law would be a more flexible and efficacious way to handle recognized problems. The fate of the BEPS project, and indeed the OECD Centre for Tax Policy and Administration, rests on the follow-through.

Digital economy. As previously explained, the row about the digital economy started with the French observation that there should be some way to tax value added when Google and other remote actors scoop up local customer information and resell it to advertisers. To focus on the digital aspects of the transaction is to focus on the wrong thing. (Prior coverage: Tax Notes, Apr. 22, 2013, p. 364.)

Digital upload is a business-to-consumer transaction requiring no physical presence of the business (or at least not what the OECD article 5 commentary considers physical presence). It doesn't matter what the business is selling -- there is value creation wherever there is a customer base, regardless of data sharing. The presence of the immobile local consumer and the economic activity of the nonresident business should be the focus.

There is no good way to impose income taxes on Google without also taxing Louis Vuitton when it sells handbags in China. But if the French and other luxury goods purveyors dependent on Asian markets want to avoid this conflict by imposing a VAT or an excise on digital uploads, there is nothing the OECD and its income tax treaties can do to stop them.

Tivoli's fake Chinatown is underneath the roller coaster tracks. European domestic laws are endangering their tax bases. (Rony Zmiri/Bigstockphoto.com/)Prediction: This awkward issue has already been swept under the rug by being relegated to a study.

Arbitrage. Tax arbitrage is what our readers do. It's been one of the central problems of treaty-based relationships for a long time, but the OECD and various governments winked at and even encouraged it in order to minimize taxes on their national champions. The theory was that those national champions would not turn on their own governments or whittle their tax bills down to negative numbers. They did.

A case in point: repos. At the IFA USA/Germany joint meeting in Hamburg, there was a discussion of using a cross-border repo between the United States and Germany to get an arbitrage result and facilitate repatriation of profits to the United States. The subject of the repo was special shares issued by an affiliate. Practitioners consider this transaction technically sound because Germany regards the U.S. parent -- the obligor on the loan represented by the repo -- to be a third party.

This transaction was popular between the United States and United Kingdom a decade ago despite the treaty. The British adopted highly specific antiabuse rules. (Prior analysis: Tax Notes, Apr. 4, 2005, p. 15.) The IRS also cracked down, holding the U.S. side to secured loan treatment for the repo (ILM 200645018). (Prior analysis: Tax Notes, Dec. 6, 2004, p. 1307.)

It is as bizarre as a 19th-century Tivoli whirligig ride to imagine that any other country would let this hokey repo transaction continue. There is no conceivable business reason for group members to be dealing in shares of an affiliate. But here we are in 2013, and planners are taking this sad double act to new stages.

It is ridiculous to make tax administrators and courts beat repos over the head with antiabuse rules and economic substance arguments when the entire financial and accounting world treats these transactions as loans, despite their being papered as sales. The OECD should recommend domestic legislation that says that repos are loans.

Germany recently changed its domestic law to disallow a dividend exemption when the payer was allowed to deduct the payment in its country of residence. Practitioners question whether it could violate the EU parent/subsidiary directive (90/435/EC). The Germans would like this rule to be included in an EU code of conduct for corporate taxation, according to Katharina Becker of the Bundesfinanzministerium, who spoke in Hamburg.

Some arbitrage opportunities are a function of local corporate law. Some European countries permit perpetual debt. Indeed, the use of a commissionnaire instead of a buy-sell distributor can be considered a form of arbitrage. As with the Danes and their amusement park, there are likely to be ancient, unknowable reasons for these rules. Countries can't be expected to change them.

But the tax law does not have to respect these treatments. European tax law is going to have to become a whole lot less form-based if European governments expect to collect tax from multinationals with parents in common law countries.

No longer should mermaids on the border of a document ensure treatment as debt. The action plan states that the OECD will develop model domestic law provisions that deny a deduction, deny an exemption, deny a double deduction, or establish a tiebreaker rule for entity classification.

Arbitrage often takes the form of hybrid entities. The U.S. check-the-box rule is not mentioned in the action plan (reg. section 1.7701-3). Is the United States going to restrict or repeal its check-the-box rules, widely blamed for making arbitrage zipless? Hell, no -- those rules were effectively reenacted by Congress and are regarded as holy writ. So other countries, like Denmark, will have to enact workarounds (section 2A of the Danish Corporate Tax Act). (Prior analysis: Tax Notes Int'l, Apr. 9, 2012, p. 169.)

Prediction: Some solutions will be offered, and taxpayers will find new tricks when the old ones are shut down.

Controlled foreign corporations. European countries have CFC rules modeled on, and often tougher than, U.S. CFC rules. At the IFA congress, the question was raised what the purpose of the rules is. This question is surprisingly difficult to answer.

The original impetus for the CFC rules was to prevent companies fobbing mobile passive income off to tax havens. But what is best practice now that EU-member tax havens and low-tax countries cannot be blacklisted? Switzerland, Ireland, and Luxembourg have no CFC rules, so entities in these countries are often interposed between European parents and their CFCs.

Cadbury Schweppes (C-196/04) puts a serious dent in the kind of CFC rules EU member countries can have. The European Court of Justice held that consistent with freedom of establishment, a company may decamp to pursue a more advantageous tax regime, and that act cannot give rise to a presumption of tax avoidance.

Cadbury requires any restriction on movement to be specifically directed toward "the creation of wholly artificial arrangements which do not reflect economic reality, with a view to escaping the tax normally due on the profits generated by activities carried out on national territory." The decision forbids any tax avoidance motive inquiry unless the taxpayer is given the opportunity to demonstrate that the arrangement is genuine business.

Cadbury's Irish CFC had no employees, no office, no nothing. A couple of guys with phones and desks and lunches at Dublin pubs every day would have kept the candy-maker out of court. Even though three employees would strike many as flimsy, that's enough people to run a group finance company, an intangibles holding company, or a multibillion-dollar reinsurance company.

Moreover, some havens are demanding local presence in return for tax favors. The Dutch government said it would expand minimum substance requirements in response to BEPS. Minimum substance means that half the directors must be resident. It will apply to all conduits, even those with advance clearance. Holding companies will have to show something akin to substance to be eligible for the participation exemption.

The action plan mentions requiring substantial activity for access to a preferential regime. Thus the issue for CFC rule designers is not whether multinationals won't have enough substance in their haven affiliates, but what happens when they do and the result is still objectionable to source country governments.

A really big multinational with extensive operations can always throw a few bodies at a tax haven project. Cadbury situations happen because planners are too cheap and lazy to put people in their tax haven arrangements. A substance requirement would simply allow the better-heeled multinationals with more flexible business models to continue to avoid taxes.

As the IFA Copenhagen meeting, Mike Williams, director of business and international tax at HM Revenue & Customs, explained that CFC rules work best when every country has them. He doubted that EU countries without them could be persuaded to adopt them. Moreover, even if a minimum level of corporate income tax were desirable, CFC rules would not necessarily be the way to accomplish that, Williams argued.

In its corporate tax reform, the United Kingdom retained CFC rules as a backdoor approach to intangibles migration. Williams emphasized that the functioning of this regime depends on the efficacy of the transfer pricing rules for outbound transfers of intangibles.

The British are reluctant to have onerous CFC rules for fear of making British multinationals uncompetitive. Becker expressed a similar sentiment at the IFA Hamburg meeting. Germany wants to stay competitive and will move on BEPS only if other countries do, she said.

Prediction: Continued stalemate as participants fail to agree on the purpose of CFC rules and despair of the holdouts ever adopting them.

Interest deductions. Interest deductions are a classic case of the kind of problem that could be handled by domestic law. How is the OECD supposed to help countries draft domestic laws? It can't. The most it can do is recommend best practices.

Indeed, the German limitation of interest deductions to 30 percent of EBITDA is becoming the standard for Europe (section 8A of the Abgabenordnung). But it has shortcomings, particularly in the fat capitalization situation, in which a multinational with no debt uses internal interest deductions to strip income from operating affiliates.

So a consensus appears to be forming around a version of the British worldwide debt cap. Funny -- the cap was not regarded as particularly onerous when it was installed. It shouldn't have worked; it wasn't intended to. That it brought in any revenue at all is an indication of the degree to which British multinationals were packing interest deductions into Blighty.

The worldwide debt cap is a substitute for the failure of British CFC rules to treat inbound loans as dividends. It is not a straightforward comparison of each member's debt/equity ratio to the group's total debt/equity ratio. That would be a sensible approach. The British version gives far too much credence to intragroup loans.

The worldwide debt cap prevents British interest deductions from exceeding the group's worldwide interest expense. Because of the European Court of Justice holding in Bosal (C-168/01), it applies to purely domestic groups with intragroup debt, which are allowed to exclude the income represented by disallowed interest. The tested amount is the total intragroup interest expense in the United Kingdom, excluding British external interest expense, and the cap is the worldwide group's net external interest expense.

HMRC must determine the subject company's debt capacity -- that is, how much it would be able to borrow, as an independent company, from an outside lender. Debt capacity is then compared to the subject's actual borrowing from related companies, or under a related-party guarantee, to determine whether it is thinly capitalized. Outside loans are considered when the proceeds are on-lent around the group.

The OECD solution need not be so complicated. Like under the German rules, a debt cap should be compared to the individual member debt/equity ratio to the group's ratio. No member should be able to deduct interest attributable to a debt/equity ratio exceeding that of the group of which it is a member.

That would mean that there would be no interest deductions for intragroup lending in a fatly capitalized group. That result is as it should be. Intragroup loans are not loans, but capital contributions. American financial accounting recognizes that, but sadly, IFRS takes this lending at face value. A debt cap is a polite way to get to disallowance of deductions for internal loans.

Should an exception be made for group finance companies, as in the British CFC rules? No. They are income-stripping devices. If groups want to pretend for their own books that they lend money so each member has a measurable cost of capital, that's their internal bookkeeping issue. There should be no tax ramifications.

A debt cap still focuses on location of the borrower. As an OECD official argued in Copenhagen, the location of an interest deduction can be arbitrary. Even for a group without outside borrowing, the interest deduction is likely to be in the parent's home country or in a high-tax country with developed capital markets. But money is fungible, and borrowing benefits the whole group. That means that allowable interest deductions should be sprinkled around the group in a rational manner.

U.S. Treas. reg. section 1.882-5 provides that rational allocation. It allocates interest between foreign banks and their U.S. branches based on the fair market value of its U.S. assets relative to its worldwide assets for purposes of determining the branch's separate U.S. interest deduction under the branch profits tax (section 882(c)). The regulation treats money as fungible and the branch as having a call on its parent's assets.

This regulation is formulary apportionment in all but in name. Europe has its own formulary apportionment project, the common consolidated corporate tax base (CCCTB). In Hamburg, Becker argued that a revived CCCTB project could help with the base erosion problem.

Prediction: A worldwide debt cap and apportionment of interest deductions among group members. No explicit nonrecognition of intragroup debt.

Harmful tax practices. How does the OECD gracefully talk about this awkward subject when the United Kingdom has reduced its corporate rate and gutted its corporate income tax rules? The refreshing thing about the British approach is that it is open and aboveboard -- or as much so as any legal scheme requiring the services of expensive lawyers to interpret can be.

The action plan identified corporate rate reductions and patent boxes as harmful. These are acts of desperation on the part of both the European havens and the home country governments of multinationals. It may be that preferential regimes for corporations are better addressed at the EU level. At the IFA Hamburg meeting, Becker argued that the harmful tax project needs a new set of standards and a new outlook.

A harmful practice, as the transparency proposal recognized, is secret rulings that enable taxpayers to get tax haven results despite a superficial tax system with an acceptable rack rate. Even a low-tax kick-in rule like the German CFC rules would not adequately address the sneaky practices of some havens.

Prediction: More wheel spinning, but at least the nonsense about countries competing for business has been suppressed for the time being.

Treaty abuse. The OECD will draft an antiabuse provision for treaties and clarify that treaties are not meant to facilitate double nontaxation. OECD commentary has never stated that treaties should be interpreted to ensure that double nontaxation does not occur.

By treaty abuse, the OECD apparently means the use of third countries to gain unwarranted treaty benefits. The idea of abuse is confined to a trilateral situation in which one party is a conduit. An antiabuse rule would not seem to be the optimal solution.

The action plan cites the commentary on article 1 of the OECD model treaty as a source of ideas to address three-party treaty abuse. Paragraphs 15 and 17 of that commentary respectively describe subject-to-tax and anti-conduit clauses that could be added to treaties. There's even a suggested low-tax kickout clause (paragraph 21.2).

The OECD is mulling various iterations of American-style limitation on benefits clauses, because les rosbifs have made it clear that they will not stand for a restrictive definition of beneficial owner along the lines of the OECD's recent white paper on the subject. U.S. LOB clauses typically exempt publicly traded companies -- the very companies whose tax avoidance is the subject of the BEPS project.

American LOB clauses come in two flavors: the big, complicated provision for friendly countries and the short, dirty, nasty one for countries suspected of becoming havens. There are no prizes for guessing which one is efficacious. The Netherlands is privileged to have the long form, which has been criticized for catching arrangements it is not supposed to catch .

Barbados and Cyprus have the short form in their U.S. treaties, and it works. The money shot in the Barbados treaty is contained in article 22(d)(ii), which allows treaty benefits only when:


    less than 50 percent of the person's gross income for that taxable year is paid or accrued, directly or indirectly, to persons who are not residents of that Contracting State . . . in the form of payments that are deductible for the purposes of the taxes covered by this Convention in the State of which the person is a resident (but not including arm's length payments in the ordinary course of business for services or tangible property).

Even this 49 percent allowable level of income stripping is generous. The working assumption must have been that planners leave no income in the country. The OECD could recommend a much lower level, such as 30 percent.

The trouble with the treaty exemption method for double tax relief is that exemption is generally considered absolute. It should always be conditional. The OECD should be considering subject-to-tax rules in treaties and domestic law that cover all conceivable situations of residence country exemption for taxable economic actors. Germany has a subject-to-tax rule in its domestic law.

The OECD model has an optional denial of benefits clause, in the form of a fourth paragraph to article 23A for exemption-method countries, that could be made mandatory (para. 56 of the OECD model commentary to article 23A).

Domestic law exemption in the residence country would not be covered by paragraph 4, and the source country would still have to grant exemption (para. 56.3). Likewise, U.N. model commentary recommends adoption of a fourth paragraph in article 23A by exemption-method countries to deny benefits for income that goes untaxed by the residence country, but only by reason of the treaty (para. 19 of the U.N. commentary on article 23A).

Prediction: A model LOB clause that is more complicated than it needs to be and less effective than it should be.

Permanent establishment. The Demon roller coaster of the project. As previously noted here, PE is the most important question the OECD will address.

PE is a crucial element of the model treaty, which is designed to limit source country tax jurisdiction over foreign businesses. It has been under attack for years, from both sides -- from multinationals that abuse it by compartmentalizing and from developing countries that want to reclaim their jurisdiction.

At the very minimum, the OECD should do two things. First, it should remove the misbegotten paragraph 6 from article 5, killing off any notion that an independent agent is not a PE. That step would redefine PE as common law agency. The action plan specifically mentions commissionnaire structures. The best way to shut that down is to define agency PE as common law agency.

If some person, related or not, is acting on behalf of a nonresident business, the latter should have a PE, regardless of the agent's ability to conclude contracts. If the contracts are ratified, that should be good enough. The messy factual inquiry should be reduced, not moved to another area. The 2003 commentary crept toward this result without going far enough.

Second, the OECD should put a clause in the PE article stating that there should be no stacking of the various exceptions to PE contained in paragraph 4 of article 5. Thus article 5(4)(f), which appears to allow a combination of exceptions, should be removed. No business should be eligible for an exception unless that single exception is the only exposure to the country.

All of the exceptions should be repealed. Some consideration is being given to cutting back on the preparatory and auxiliary exception of article 5(4)(e) and (f). This exception should just be eliminated. A fixed place of business should nail down a PE, full stop.

The United States and Europe are anxious to restrain India and other like-minded countries that want to expand the concept of PE to reach economic nexus. While it is better to have India and like-minded countries in the room, their aggressive audits will not stop. The main purpose of discussing PE is to prevent these non-OECD-member countries from bolting from the base erosion discussion.

Expansion frequently goes under the guise of services PE. The OECD's 2008 climb down on services PE was the beginning of the end of the argument that continuous provision of services without a physical office should not be taxed. The next frontier, as the digital economy argument illustrates, is inbound services with no physical presence of the service provider.

Now, economic nexus is the future, as the experience of the American states shows. But the BEPS project is intended to postpone the inevitable in the hope of repairing the dying international consensus and treaty network. What the proponents of delay fail to understand is that the PE concept will only survive if it is expanded. Narrow PE is already moribund.

Prediction: Nothing will be accomplished. There is too much disagreement. European countries cling to the older, narrow interpretations of PE to benefit their national champions.

Transparency. As this article was being written, the G-20 were expected to endorse country-by-country reporting, a concept that migrated from mineral extraction to the Dodd-Frank Wall Street Reform and Consumer Protection Act to the tax law. What this gesture says is that eventually multinationals will have to stop treating return filing as an exercise in hiding information from governments.

The BEPS project also contemplates some government transparency. Readers, European governments make money by selling copies of their laws. It is going to be a difficult job getting them to consent to disclose every grubby deal they make with taxpayers. The U.S. government does not disclose advance pricing agreements, which are a crucial part of the very sort of multinational planning that the BEPS project seeks to address.

Prediction: Some more transparency than we have at present.

Multilateral agreement. Lawyers who pooh-pooh the prospects for a multilateral tax agreement should recognize that the OECD model treaty itself is a highly inefficient form of multilateral agreement. Changes to the treaty and its commentary require member consensus -- and then don't become effective because of the slow pace of renegotiation and the reluctance of many courts to adopt ex post changes to commentary.

Thus the multilateral agreement is a recognition that good ideas for reform need teeth. Those who don't want a multilateral agreement should just get off the bus. Although the prospect of unhappy countries terminating treaties worried some speakers at IFA, tax treaties are not necessary for foreign direct investment.

Will the multilateral agreement happen? It has to, otherwise the whole BEPS project will go the way of the unpopular new article 7 and its disastrous cousin, the authorized OECD approach to allocation of profits to PEs. The OECD model, which is already a European treaty, will die of disuse if it does not become ambulatory.

Prediction: The multilateral agreement will be adopted despite rosbif skepticism.

Will the BEPS project succeed or fail? Look at it this way: The OECD will do more to reform corporate income taxation in the short run than the U.S. Congress. In the event of failure, the OECD should think about asking participating countries to designate their national champions, whose financial statement income could be allocated and taxed in the countries where they do business.