The Oxford International Tax Group, chaired by Michael Devereux, has just posted this chapter of a forthcoming book, entitled (i.e., the chapter) "Residual Profit Allocation by Income."
The cast of authors for the chapter includes Devereux, Alan Auerbach, Michael Keen, Paul Oosterhuis, Wolfgang Schoen, and John Vella. (Michael Graetz is also credited with having contributed to the group's deliberations about the chapter's proposal.) Happily, the committee here appears to have indeed designed a horse as intended, not a camel.
The basic idea is to replace current transfer pricing practice in international taxation with something that looks a bit like it but is actually (and by design) quite different. They propose to split a multinational's profits into the "routine" and "residual" components, allocate the former based on arm's length comparables, and assign the latter to market countries where sales occurred, proportionally to sales.
As the paper notes, routine versus residual may tend to overlap considerably with normal returns vs. rents, but they're not identical. "'Routine' profit is the profit a third party would expect to earn for performing a particular set of functions and activities on an outsourcing basis."
The approach resembles sales-based formulary apportionment, except that it only applies that approach to the residual element. The routine piece ends up being allocated to the countries where the underlying activities take place.
It even more closely resembles this proposal by Reuven Avi-Yonah, Kimberly Clausing, and Michael Durst. The big difference is that Devereux et al look to standard transfer pricing comparables for the routine piece, rather than assigning a fixed rate of return. They argue that this makes their proposal less distortive albeit more complex to apply.
While I obviously follow international tax policy stuff pretty closely, you can only get your hands fully immersed (so to speak) in so much of it. I've regarded the transfer pricing / formulary apportionment et al debate as an area in which I have leanings but don't regard myself as having invested fully enough to have the same degree of confidence about my views, as in areas that I have focused on more directly. Also there are empirical tradeoffs to worry about here that I don't believe the literature has as yet fully resolved.
That said, my inclination has been to strongly favor an approach like that in Avi-Yonah, Clausing, and Durst, even if a given country such as the U.S. at first adopts it unilaterally. But I understand that there are serious people who have looked at this issue and NOT liked it so much (e.g., Altshuler and Grubert). Clearly such a proposal leaves in place enough distortions and tax planning opportunities that it is at best a third-best; the case for it is that transfer pricing is at best a fourth-best. (Which is not to say that I have a particular candidate in mind for second-best.)
I'm also initially inclined to view Devereux et al as an improvement on Avi-Yonah et al. The transfer pricing comparables at issue here are for routine stuff, like say for advertising and manufacture. There will be no pretense or possibility of their applying to, say, Amazon's mega-profits from succeeding with the iPhone. They could well end up, much of the time, being about as formulary (e.g., using a fixed percentage markup) as Avi-Yonah et al, only this would be particularized to market conditions. Its retaining conventional, recognizable transfer pricing more as a matter of form than of substance (i.e., insofar as the big money is concerned) is arguably a further practical point in its favor.