Saturday, April 11, 2009

Tax Policy Colloquium on Desai & Dharmapala, Investor Taxation in Open Economies

This past Thursday we discussed the above paper proposing "global portfolio neutrality," or GPN, as a new entry in the alphabet soup of proposed international tax norms (joining CEN, CIN, CON, NN, and NON). GPN holds that national as well as worldwide efficiency is maximized by causing individuals to face the same tax rate on their portfolio holdings no matter where they invest. It would be satisfied by purely residence-based taxation of individuals on their portfolio holdings. Given source-based taxation of passive income, such as withholding taxes on dividends, the authors argue for granting unlimited foreign tax credits (FTCs), including refundable FTCs for tax-exempts. They also apply GPN to inbound investment, and suggest, therefore, that we not tax sovereign wealth funds (SWFs) that are not otherwise taxable outside the U.S. on their worldwide holdings.

The paper is partly a response to an earlier paper by Michael Graetz and Itai Grinberg, which - at least as interpreted by Desai and Dharmapala, although Graetz, attending the session, did not entirely accept this interpretation - argues that the US should merely allow deductions for withholding taxes paid abroad. The Graetz-Grinberg argument, at least as reported, was as follows. As per the Desai-Hines norm of capital ownership neutrality (CON), who owns a given business actually may matter a lot, in the sense of affecting profitability, in the case of conducting an active business. But for portfolio holdings there presumably is no effect, since the holder is assumed to be passive and to play no operating role in the business. Hence, while merely allowing a deduction rather than an FTC for foreign withholding taxes would distort ownership decisions - discouraging the holding of foreign stock - this doesn't matter since ownership is irrelevant here.

Desai and Dharmapala respond that ownership does so matter, for reasons of optimal diversification. People in a given country who are already "long" the national macro-economy because their human capital is tied up in it should be eager to diversify via exposure to foreign macro-economy risks (which are somewhat but not perfectly correlated with our own) by holding foreign stock. (BTW, the "home bias" that we continue to observe in stock ownership patterns, though it is declining, arguably shows irrational under-diversification although there also are some rational proposed explanations for it.)

So far, so good. But the trickier part, for me, is their argument that GPN establishes, as a matter of unilateral national self-interest, providing FTCs for foreign withholding taxes so long as our withholding tax rate is about the same as those applying abroad.

They have a logical and internally consistent argument for this, which would take too long to explain here but can be found at pages 23-24 of their paper (available here under April 9). But I was skeptical on the grounds that:

(a) It isn't actually unilateral if withholding tax rates have to be about the same.

(b) FTCs create bad incentives, from the U.S. standpoint, both for our taxpayers (who need not seek to economize on foreign taxes in deciding where to invest if we'll offer a credit anyway) and for foreign governments (who can think of the tax cost as passed on to the U.S. Treasury).

(c) The reason D&D require that withholding taxes be about the same is that foreign inbound investment will replace the outbound (e.g., if Americans sell U.S. stocks to hold foreign stocks, then someone else has to buy the U.S. stocks). Hence, we automatically pick up withholding tax revenues to replace the lost revenues from offering FTCs. But Alan Auerbach and I argued that this amounted to assuming that capital flows must be symmetric by asset class, which isn't necessary even if symmetry holds overall.

A further issue I raised is that investors care about true diversification as to underlying economic characteristics. But the determination of source for dividends received is essentially formal - under U.S. law, depending almost purely on where the issuer is incorporated. Does selling GE stock to hold Siemens stock have anything to do with diversification if both companies are active in the same places to the same degree? Diversification by source, as determined by the tax system with respect to passive income, could at the limit be no more meaningful than making sure you hold both stocks that are printed on red paper and those that are printed on blue paper.

A final criticism I offered on these issues (although actually, at the session, I stated it first) is that GPN, like all of the alphabet soup norms, addresses only one margin (portfolio diversification) whereas there are many. For example, can one really discuss the tax treatment of passive income held by U.S. taxpayers without considering U.S. corporations, the outbound passive investment of which is taxable without deferral under subpart F.

Despite these criticisms, this was one of the best papers and sessions of the semester. Among its virtues was addressing the significance of tax rate differences among investors within a jurisdiction. But here, although I largely sympathized with the analytical bottom line, I of course found a way to carp and cavil all the same.

With respect to refundable FTCs for tax-exempts, my main critique was that, since FTCs create incentive problems from the national welfare standpoint, it's logical to limit them. The way we actually do so, by offering a 100% marginal reimbursement rate (MRR) until one hits the credit limit and then 0% thereafter, seems a bit arbitrary and unlikely to be optimal. But it's hard to say if, overall, we are too generous or not generous enough. Hence, in any given case (such as the tax-exempts) in which one proposes to scrap the limit and permit more FTCs to be claimed, it is hard to be sure whether we are going in the right direction or not.

Finally, with respect to sovereign wealth funds (SWFs), I agreed that it is likely to be in the U.S. national self-interest not to tax them on inbound investment IF we are effectively a small open economy without the market power to impose some of the burden of the tax on them. They're distinctive in that, for various other inbound investors, the withholding taxes we impose might (a) not exceed the tax rate they would face anyway, and/or (b) be creditable by their home governments. But SWFs can't take advantage of FTCs because, effectively, they ARE the home governments that would be providing the credit.

But it seemed odd to me to call this GPN, given that (a) from a national welfare standpoint we don't care about affording them better diversification, and (b) we'd be quite happy to tax them and distort their portfolio choices insofar as we have enough market power to stick them with some of the economic incidence of the tax.

No comments: