Taxing Multinationals: Exploring a New Approach

Published on March 27, 2023 Updated on July 26, 2023

on the March 27, 2023

Focus on research

Taxing Multinationals: Exploring a New Approach

The international tax treatment of multinationals has fallen into disrepute. Over the last few years, the 140 or so members of the G20/OECD-led Inclusive Framework have put considerable effort into finding ways to fix it. In a recent paper, my co-authors and I explore alternative arrangements — ’Residual Profit Allocation’ — that have been proposed to bring the system back to some coherence. And (while we do not claim causality) those arrangements have strong similarities with those that have now secured wide political agreement. 

Sources of Disrepute

Multinationals are widely perceived as not paying enough tax, and not paying it in the right places. This reflects two features of the century-old system of international corporate taxation. One is that multinational groups are taxed not on their total profit, but instead, each affiliate within the group is taxed separately as if it were an independent business. The result is that the multinationals can engage in ‘profit shifting’: arranging transactions between these affiliates—through lending, manipulating transfer pricing, and a host of more complex devices — to arrange that their profits turn up where they will be taxed at low rates rather than at high ones. 

The second is that the mere act of selling something in a country does not create a liability to corporate income tax there. Whereas extensive profit shifting is generally recognized as abusing the spirit of the age-old tax principles, this feature — no corporate taxation without some physical presence — is precisely as was intended. In recent years, however, the massively increased ability to do business remotely, not least delivering by services over the internet (and acquiring valuable information in doing so), has made the test of physical presence increasingly flimsy — both technically and in public perception — as a persuasive basis for taxation.

How RPA can help, in principle

Both problems can be addressed by Residual Profit Allocation (RPA). This involves three steps. First, allocate to each affiliate what the lawyers call a ‘routine return:’ this broadly corresponds to what economists think of as a normal rate of return and could be calculated as, for instance, a fixed return on capital or sales. Second, calculate the multinational’s ‘residual profit’ as the amount by which group-wide profits exceed affiliates’ total routine earnings: translated for economists, this is super-normal profit or rent. Third, use some mechanical formula to share residual profits across countries. Our paper focuses on the case in which each country’s share equals the proportion of the multinational’s sales made there: a ‘destination-based’ RPA.

That, in principle, addresses both of the problems above. There is much less, if any, scope for profit shifting. In particular, moving profit between affiliates has no effect on residual profit, and so also no longer has any effect on the multinational’s tax bill. And selling into a country creates a liability to tax there, even without any physical presence. 

And in practice?

None of this means that destination-based RPA is the best alternative available, and our paper does not argue that it is. Instead, the aim is to use data on over 8,000 of the world’s largest multinationals and other information to explore what a destination-based RPA might look like, and imply, in practice. 

Residual profits, it turns out, are large: something like 60 percent of the total earnings of the multinationals in our sample. They are also highly concentrated: about 100 multinationals account for 30 percent of total residual profits. And about one-third accrues to multinationals headquartered in the US. 

In terms of tax revenue, the negation of incentives to shift profits to low-tax countries means that an RPA is likely to increase global revenue. An additional (and, it so happens, positive and larger) effect arises from allocating global profits to destination (or ‘market’) countries. Overall, switching to a destination-based RPA — assuming unchanged tax rates — raises additional revenue of around USD 94 billion (about 4 percent of worldwide corporate tax revenue). 

Underneath this total, the changes for some countries are far more significant: in France, corporate tax revenue goes up by around 25 percent; in Ireland, it falls by about 75 percent. Importantly, there is a systematic tendency for developing countries to benefit. 

The impact on investment will also vary across countries. Eliminating profit shifting eliminates an opportunity that companies have had to reduce the tax cost of, and so will discourage, investing. In high-tax countries, however, this may be mitigated or even offset because part of the residual generated by investing there will be allocated for taxation to lower-tax countries. Overall, it is hard to make a case that a destination-based RPA will generally increase investment.

What next?

Inclusive Framework members are now considering introducing a scheme — ‘Pillar One’ — with similarities to a destination-based RPA. It would identify as residual profit any earnings in excess of a 10 percent profit margin (on sales) and allocate that for taxation across countries in proportion to sales. But there are differences. Only 25 percent of residual profit would be allocated this way; this new element would be overlaid on, not replace, current arrangements; and it would be combined with a global minimum tax (‘Pillar 2’). This is a highly complicated set of arrangements, and indeed the adoption of Pillars One and/or Two is far from assured. 

What is assured, however, is that the admission into the policy debate of the fundamental departures from long-established principles that Pillar One embodies and which RPA ideas pursue to a logical conclusion, has changed the terms of the debate around international taxation irrevocably.