Wednesday, November 14, 2007

U.S. international taxation

Today at NYU Law School we held the session on international taxation mentioned in my blog entry from yesterday, with Jim Hines presenting his new paper, “Reconsidering the Taxation of Foreign Income,” followed by comments from Alan Auerbach, Mitchell Kane, and Stephen Shay, followed by general discussion. The following is a partial report and some very preliminary thoughts. (I am likely to publish about this topic at some point down the road.) But let me just mention up front that the event seemed (to my biased eyes at least) to be a big success - 70 attendees, lively discussion that could have kept right on going if people didn't have to leave, and most importantly it may have advanced people's thinking about the issues, as it certainly did mine.

Jim argues that U.S. (and other countries’) tax policy with respect to outbound investment should be guided by capital ownership neutrality (CON) rather than capital export neutrality (CEN), with the claimed implication in practice that the U.S. should exempt active business income earned abroad by U.S. corporations. Under CEN, the prescription would instead be to move towards full U.S. taxation of all worldwide income of resident corporations (and individuals), albeit subject to allowing foreign tax credits.

CON focuses on not distorting ownership patterns, as would happen if a German firm rather than a U.S. firm owned a given investment in China, despite the U.S. firm’s expecting a higher pre-tax return from the investment, because the German company earns more after-tax by reason of Germany’s imposing less tax than the U.S. does on top of whatever China levies. CON advances worldwide efficiency by increasing pre-tax profitability, taxes being ignored for this purpose as they are a cost to the taxpayer but a transfer from the social perspective (since the taxing government gets the money and does something with it). CON is understood to matter a lot in a world where the theory of the firm, as pioneered by Ronald Coase, suggests that ownership arrangements are economically very important. (You have a firm instead of arm’s length market arrangements between partners in the productive process where this increases efficiency.)

CEN focuses on not distorting where investments are made. If Ireland has a lower tax rate than the U.S., investment will tend to shift from the U.S. towards Ireland, leading to the selection of some Irish investments that are more appealing after-tax than their U.S. alternatives despite having lower expected pre-tax returns. Same point about pre-tax profitability being the proper guide since taxes are a transfer not a social cost.

So why should we think the CON margin is more worth pursuing than the CEN margin? Not because it is inherently more important, which Jim doesn’t claim (noting only that CON is indeed important), but on the ground that the U.S. can benefit from unilaterally pursuing CON but not CEN. This is a part of the analysis that needs to be developed more.

Proponents of CEN usually define unilateral pursuit of national self-interest by invoking national neutrality (NN), under which the home country would fully tax outbound investment by its companies without foreign tax credits – permitting only deductibility for foreign taxes paid. The foreign tax credits then emerge either out of benevolence or (more plausibly) reciprocity between nations. As I argued in my recent Tax Law Review piece on international taxation, moving towards CEN by increasing the U.S. tax burden on multinationals is often defensible in terms of national self-interest because it also moves towards NN.

Jim rejects this nationalistic ground for moving towards CEN and NN via what I am inclined to call the “musical chairs” or “row of shops” hypothesis. NN is motivated by the concern that, if we tax outbound investment by U.S. companies less than their domestic investment, we lose revenue because their investments relocate from the US to abroad. But suppose the amount that will be invested in the U.S. is fixed so far as the U.S. tax regime for U.S. firms is concerned. If the U.S. tax rules induce a U.S. firm to invest abroad, someone else will make the investment here. Or the U.S. firm will raise more capital in worldwide capital markets and make both investments, not just one or the other. He invokes recent empirical research in support of this view, suggesting that home and foreign investment by U.S. multinationals seem to be complements rather than substitutes.

I call this the musical chairs theory because it’s as if the music is playing, and the companies are marching around all the chairs (i.e., investment choices), and everyone ends up getting a seat somewhere. If we induce the U.S. firm not to take the seat here, someone else will take it instead. This is of course the benign (or should I say progressive schools) version of musical chairs, where you have just enough for everyone rather than being one short.

The row of stores metaphor for this story is inspired by Bleecker Street, which I sometimes pass on my way from home to school, in which every storefront is bound to be rented by someone – it’s just a question of who ends up where. Bad theory so far as Bleecker Street is concerned, by the way – there are lots of boarded-up storefronts still seeking tenants, some of which have been there for years. (A puzzle: why are the rents apparently so high if so many stay vacant for so long?)

Again, Jim doesn’t deny that low taxes attract investment. The claim here is that, if the U.S. tries to move towards CEN, rather than everyone doing it, all it does is create clientele effects, whereby other nations’ firms replace U.S. firms as the makers of particular foreign investments. So no motivation for the U.S. to address CEN alone, and indeed we get a piece of the worldwide CON welfare loss if U.S. firms (still these days predominantly owned by U.S. individuals) make a bit less money due to the inefficient reallocations. Unclear how much we gain from following CON, however, even granting that in this model we have nothing to gain from unilaterally following CEN.

Jim rejects distributional or fairness-based reasons for taxing the worldwide income of U.S. firms by analogy to tax-exempt bonds. Say the interest rate on taxable bonds is 10%, the marginal tax rate (MTR) is 30%, and tax-free municipal bonds pay 7%. Then there is no distribution problem by reason of the preference (assuming the pre-tax interest rate is fixed), because muni bond holders pay a 30% implicit tax that is the same as everyone else’s explicit tax. All earn 7% after-tax. By analogy, investment in low-tax Ireland pays an implicit tax in the form of a lower pre-tax return (by reason of the CEN-violating shift of investment into Ireland), so there’s no reason for distributional concern about its being (to exaggerate relative to actual Ireland) tax-exempt.

The big problem with the muni bond argument is that it only works with a single MTR for all investors. So, if in actuality we taxed corporate income on outbound investment at the individual level, so we could apply Bill Gates’ MTR to his investments and a lower rate to yours and mine, exemption for foreign source income would involve sacrificing this potential to apply the desired MTR to each investor. But since we generally tax outbound corporate equity investment purely at the flat corporate rate, we aren’t getting that rate differentiation anyway under the current system or even one revised to accord more with CEN.

There is lots more one could say (and that I perhaps will say) about the paper and the topic – including details of excellent comments by Auerbach, Kane, and Shay – but given the length of this post I will omit them at least for now. But one last point concerns passive income, earned through portfolio investment. Jim agrees that the U.S. should tax all worldwide income of U.S. residents. This is potentially a big concession, making one wonder about the broader principle. It has a rationale, relating to the point that passive income doesn’t have a meaningful location in the same sense as active business investment (e.g., there is no limit to the funds that could be described to the tax authorities as deposited in Caymans banks). So the musical chairs hypothesis does not apply. But still, it might have big implications for the overall analysis – especially considering the murkiness of the active-passive distinction (which Jim conceptualizes as, “does ownership or control matter here?” – a matter of degree, of course) – and considering as well the murkiness of the source concept.

Source is not an economically well-defined idea. Consider, for example, the synergies obtained by operating as a multinational rather than through arm’s length dealings between firms in different nations. Where exactly does the synergy income arise? And this is not just an implementation question – it undermines the underlying idea on which source-based taxation ostensibly rests.

No comments: