On Tuesday, Leslie Robinson of the Tuck Business School at Dartmouth presented the above-titled paper, co-authored with her colleague Katharina Lewellen, and available here.
This continued our annual tradition of having a paper by an accounting professor (we also try to do one each year by a political scientist, this year to be Larry Bartels on April 23), along with some economists, a mix of junior and senior law professors, and often a tax practitioner.
The paper looks at multinational firms' internal structures. In particular, when do they have "complex" structures in which, rather than having the U.S. parent directly own, say, a sub in every country where it operates (this would be a "flat" structure), they have tiered structures in which some subs own other subs. The aim is to get a sense, from the overall empirics regarding "owner subsidiaries" and the particular vertical pairings that one finds, regarding what might be the likely causation, be it tax or something else, for internal complexity.
One unfortunate limit in the data is that one can't observe which firms are "hybrids," i.e., firms that the taxpayer elects to treat for U.S. tax purposes as legally non-existent branches of their direct owners. Hybrids are useful so that one can, say, pay interest from a Germany subsidiary of an ultimate U.S. parent into the Caymans for German tax purposes, without the U.S. thereby saying: Aha, receipt of interest by the Caymans affiliate means that the U.S. parent has subpart F income.
The paper finds evidence of both tax motivations and others in the use of complex internal structures. But a lot of the others were probably also tax. For example, affiliates that trade with each other often have a vertical ownership structure. But those dealings may be tax-motivated transfer pricing. And affiliates with a lot of cash on hand tend to invest equity in other affiliates. But they may have had cash on hand due to profit-shifting games within the group, and then they use the cash to own other affiliates directly so that it doesn't have to be repatriated taxably to the U.S. parent first.
A natural policy response would be to entirely ignore internal structuring of multinational firms, and tax the entire global entity (whether it has a U.S. parent or not) as a unitary business. For U.S. firms, this could involve replacing deferral with a lower tax rate for foreign source income generally (since the question of what tax rate we want to impose on foreign source income is distinct from that of whether internal firm structure should matter). It might also involve wholly ignoring intra-firm cash flows of all kinds (who borrows from each other or a third party lender, who pays royalties to whom, etc.). This would put us in a formulary apportionment world, which isn't ideal either. But I see that (if done right) as a way of making the effective election to shift profits costlier than it is when we take account of meaningless intra-firm entity lines and events.
A different type of response would be to tax-penalize complex structures directly, on what I call the Three Stooges rationale.
(Moe hits Curly on the head. Curly: "Whatja do that for? I didn't do nothin'!" Moe: "That's in case you do something later when I'm not around.")
This could either involve directly relating U.S. tax burdens to some measure of complex rather than "flat" internal structuring, or more consistently giving adverse tax consequences to intra-group cash flows, including those that use hybrids. The corporate literature on "pyramiding" and responses thereto may be of interest here.