Thursday, March 14, 2013

Tax policy colloquium, week 7: Desai and Dharmapala's "Competitive Neutrality Among Debt-Financed Multinational Firms

On Tuesday we did our last Tax Policy Colloquium before NYU's one-week spring break.  Our speaker was Dhammika Dharmapala, presenting a draft of the above-titled work that he and Mihir Desai are co-authoring.  No link here, as he asked us to take it down afterwards because it is still in very preliminary form.

A bit of the abstract may help to explain what issues the paper tackles and what conclusions it reaches.  For starters:

"Debt plays an important role in the financing of multinational corporations (MNCs). Interest expenses are typically tax-deductible in most corporate income tax systems, and there has been a growth of interest in recent years in the tax treatment of debt and its consequences. This paper discusses the optimal form that interest deductibility and associated restrictions should take in a multi-jurisdictional setting. We straightforwardly extend existing neutrality norms in international taxation to serve as a benchmark."

They use "competitive neutrality" as their benchmark, but they could just as well have used "capital ownership neutrality" (CON), which Desai in particular has written about before (along with Jim Hines).  For convenience, I will characterize them as using CON even though they keep on saying "competitive neutrality."

I must confess to not being happy with this approach, for reasons that my forthcoming international tax book will explain in more detail.  Complaint 1: It's a global welfare norm, and thus not necessarily of interest to policymakers (or citizens) in any particular country.  I think the way to go is by starting with unilateral national welfare (i.e., what's best for the particular country that is setting its rules for inbound and outbound investment, if no other country responds to its choices), and then asking how strategic interactions with other countries may change the analysis.

If you instead identify a best approach from the global welfare standpoint, it's conceivable that countries could execute a Pareto deal to achieve it, leaving everyone better-off.  Conceivable, but not too bloody likely.

OK, in principle it doesn't matter where you start, so long as you end up in the right place.  So why not begin with global welfare, as Desai and Dharmapala do, and then proceed towards national welfare by examining opportunities for cooperation?  But the problem is this.  As the paper shows, there are just so many margins that one could think about, in trying to maximize global welfare, that starting there almost inevitably lands one in a morass.  National welfare turns out to be simpler, or at least more readily simplified for basic analytic purposes, most of the time.  But again, that's jumping into my book and would need fuller illustration than I can provide here.

In their colloquium paper, Desai and Dharmapala are interested in particular in the question of how CON plays out when it is impossible to achieve accurate "tracing" of interest expense to the outlays that are related to them at the margin.  So you have the basic scenario that a taxpayer can, say, borrow in a high-tax country and invest the funds in a low-tax country.  This can cause negative-present-value projects on a pre-tax basis to make a whole lot of sense (from the taxpayer's perspective) after-tax.  Under CON, they want to achieve global tax neutrality for all multinationals with respect to the question of who borrows (anywhere) in order to hold a particular asset (somewhere).

The paper concedes that there are other relevant margins, such as those pertaining to the choice between debt and equity and to the choice of negative-present-value projects.  But they want to isolate the CON thread for separate analysis, on the view that this is worth knowing even if in the end one will have to optimize overall at all of the margins, which probably requires failing to fully optimize at any one of them considered in isolation.

More damaging still, in my view, is the fact that this is only CON between multinationals.  They concede that their approach might result in tax bias as between multinationals and purely national firms.  For example, suppose GE and an Indonesian firm would both pay tax on factory production in that country at the Indonesian rate, but that only GE can borrow in such a way as to get interest deductions at the higher US rate, rather than the Indonesian rate.  You could then get the scenario where GE is tax-favored relative to the Indonesian company.  In such a case, satisfying CON for their purposes, by reason of treating, say, GE and Siemens the same, might be an empty achievement.

OK, one might argue that GE and the Indonesian firm are likely to be active in different sectors, and thus not competing directly.  But even apart from the fact that these two sectors may be competing for global capital, suppose GE wants to use the Indonesian firm in its global production processes, and is deciding whether to buy it or use arm's length contracting.  CON as used in the paper therefore appears to fall quite short even apart from its being just one of many global efficiency strands and its not being of direct interest to any national policymaker.

Putting out of mind (for just a moment) all those limitations, where do they believe it leads?  They note that, generalizing the approach taken in a rather notorious paper on interest deductibility by Jim Hines, suppose the U.S. agreed to provide 35 cents of subsidy per dollar of interest expense by all multinationals investing anywhere in the world.  That would achieve CON, but it doesn't appear to be politically feasible (gee, I wonder why).  So they look at alternative approaches, including one in particular that they expressly describe as a thought experiment rather than a concrete proposal.

Suppose that all countries in the world agreed to impose a worldwide debt cap that functioned as follows.  While in general each multinational's affiliates in a given country would deduct their in-country interest expense, all countries would impose a worldwide debt cap, restricting the amount of the in-country interest deductions to the multinational group's total worldwide third-party interest payments.  Desai and Dharmapala argue that this would satisfy CON (in the between-multinationals sense) so long as the multinationals had enough taxable income in each country where they borrowed to deduct all of the locally incurred interest expense.

They concede that this might both create a huge global tax bias in favor of debt at the expense of equity, and make negative-present-value projects worth undertaking after-tax.  But they argue that one could address those concerns, without undermining the CON result, by having all countries agree to cap the allowable interest deduction at the same arbitrary fraction of global third-party interest expense.  Hence, at least for thought experiment purposes, they appear to regard their proposal as potentially meritorious across all of the relevant margins, so long as policymakers around the world pick the same arbitrary fraction in light of the proper weight of all the competing objectives.

While the proposed thought experiment solution is not unclever, I in the end can't help but regard the paper, with all due respect to its authors, as very effectively (though inadvertently) illustrating the argument in my forthcoming book that much of the international tax policy literature has gone badly off the rails.  I am skeptical that "alphabet soup," or the use of a single-bullet global welfare approach to guide the analysis, can lead anywhere that is very useful, no matter how intelligently it is executed.  And when I say "useful," I mean either in the sense of leading to important theoretical insights or of generating practical proposals that policymakers might find useful.  I hope and even believe that very different types of approaches to international tax policy analysis will become increasingly prevalent over the next few years.

No comments: