Not long ago, transfer pricing was a subject for tax administrators and one or two other specialists. But recently, politicians, economists and businesspeople, as well as NGOs, have been waking up to the importance of who pays tax on what in international business transactions between different arms of the same corporation. Globalisation is one reason for this interest, the rise of the multinational corporation is another. Once you take on board the fact that more than 60% of world trade takes place within multinational enterprises, the importance of transfer pricing becomes clear.
Transfer pricing refers to the allocation of profits for tax and other purposes between parts of a multinational corporate group.mConsider a profitable UK computer group that buys micro-chips from its own subsidiary in Korea: how much the UK parent pays its subsidiary – the transfer price – will determine how much profit the Korean unit reports and how much local tax it pays. If the parent pays below normal local market prices, the Korean unit may appear to be in financial difficulty, even if the group as a whole shows a decent profit margin when the completed computer is sold. UK tax administrators might not grumble as the profit will be reported at their end, but their Korean counterparts will be disappointed not to have much profit to tax on their side of the operation. This problem only arises inside corporations with subsidiaries in more than one country; if the UK company bought its microchips from an independent company in Korea it would pay the market price, and the supplier would pay taxes on its own profits in the normal way. It is the fact that the various parts of the organisation are under some form of common control that is important for the tax authority as this may mean that transfers are not subject to the full play of market forces.
Transfer prices are useful in several ways. They can help an MNE identify those parts of the enterprise that are performing well and not so well. And an MNE could suffer double taxation on the same profits without proper transfer pricing. Take the example of a French bicycle manufacturer that distributes its bikes through a subsidiary in the Netherlands. The bicycle costs €900 to make and it costs the Dutch company €100 to distribute it. The company sets a transfer price of €1000 and the Dutch unit retails the bike at €1100 in the Netherlands. Overall, the company has thus made €100 in profit, on which it expects to pay tax.
But when the Dutch company is audited by the Dutch tax administration they notice that the distributor itself is not showing any profit: the €1000 transfer price plus the Dutch unit’s €100 distribution costs are exactly equal to the €1100 retail price. The Dutch tax administration wants the transfer price to be shown as €900 so that the Dutch unit shows the group’s €100 profit that would be liable for tax. But this poses a problem for the French company, as it is already paying tax in France on the €100 profit per bicycle shown in its accounts. Since it is a group it is liable for tax in the countries where it operates and in dealing with two different tax authorities it cannot just cancel one out against the other. Nor should it pay the tax twice.
In a bid to avoid such problems, current OECD international guidelines are based on the arm’s length principle – that a transfer price should be the same as if the two companies involved were indeed two independents, not part of the same corporate structure. The arm’s length principle (ALP), despite its informal sounding name, is found in Article 9 of the OECD Model Tax Convention and is the framework for bilateral treaties between OECD countries, and many non-OECD governments, too.
The OECD Transfer Pricing Guidelines provide a framework for settling such matters by providing considerable detail as to how to apply the arm’s length principle. In the hypothetical French-Dutch bicycle case, the French MNE could ask the two tax authorities to try to reach agreement on what the arm’s length transfer price of the bicycles is and avoid double taxation. It is likely that the original transfer price set by the MNE was wrong because it left all the profit with the manufacturer, while the Dutch proposal erred on the other side by wanting to transfer all the profit to the distributor.
But all of this assumes the best possible world, where tax authorities and MNEs work together in good faith. Yet transfer pricing has gained wider attention among governments and NGOs because of another risk: that it could be used to shift profits into low tax jurisdictions even if the MNE carries out little business activity in that jurisdiction. This leads to trade distortions, as well as tax distortions.
No country – poor, emerging or wealthy – wants its tax base to suffer because of transfer pricing. That is why the OECD has spent so much effort on developing its Transfer Pricing Guidelines. While they help corporations to avoid double taxation, they also help tax administrations to receive a fair share of the tax base of multinational enterprises. But abuse of transfer pricing may be a particular problem for developing countries, as companies might take advantage of it to get round exchange controls and to repatriate profits in a tax free form. The OECD provides technical assistance to developing countries to help them implement and administer transfer pricing rules in a broadly standard way, while reflecting their particular situation.
Applying transfer pricing rules based on the arm’s length principle is not easy, even with the help of the OECD’s guidelines. It is not always possible – and certainly takes valuable time – to find comparable market transactions to set an acceptable transfer price. A computer chip subsidiary in a developing country might be the only one of its kind locally. But replacement systems suggested so far would be extremely complex to administer.
The most frequently advocated alternative is some kind of formulary apportionment that would split the entire profits of an MNE among all its subsidiaries, regardless of their location. But proponents of such alternatives not only have to show that their proposals are theoretically “better” but that they are capable of winning international agreement. Not easy, since the very act of building a formula makes it clear what the outcome is intended to be and who the winners and losers will be for a given set of factors. Tax authorities would naturally want the inputs to reflect their assessment of profit. Questions like how to apportion intellectual capital and R&D between jurisdictions would become contentious.
Such problems would make it very difficult to reach agreement on the inputs to the formula, particularly between parent companies in wealthy countries and subsidiaries in poorer ones. ALP avoids these pitfalls as it is based on real markets. It is tried and tested, offering MNEs and governments a single international standard for agreements that give different governments a fair share of the tax base of MNEs in their jurisdiction while avoiding double taxation problems. Moreover, it is flexible enough to meet new challenges, such as global trading and electronic commerce. Governments so far appear to agree: much better to update the existing system than start from scratch with something new.
• For more on the OECD’s tax work: www.oecd.org/taxation
• Transfer Pricing Guidelines for Multinational Enterprises and Tax Administrations, OECD, 2001.
©OECD Observer No 230, January 2002
(Corrected 3 July 2008)