Why U.S. Business-Level Taxation Should Rely on a Mixture of Both Source-Based and Destination-Based Principles
Last week, I reviewed a recent article by David Hariton in TaxProf Blog. In that article, Hariton analyses gaming of destination-based business-level taxes — like the House Republicans’ destination-based cash-flow tax (DBCFT) proposal. Synthesizing Hariton’s analysis with what we know about how the existing U.S. source-based corporate income tax (CIT) is currently gamed, and also academic work of mine, implies a strong case for why the U.S. business-level tax regime should be reformed so as to incorporate some better mixture of both source-based and destination-based principles.
Below, I post the key portions of my TaxProf Blog post that support this conclusion. I then continue with some further analysis.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
The problems with our existing, source-based, corporate income tax (CIT) are well known. Two classic problems with our CIT are the incentives it creates to substitute debt financing for equity financing and to substitute the use of partnership forms for the use of corporate forms. The third major problem is transfer-pricing games. This problem has been rising rapidly in importance, and it is this problem that border-adjustment proposals are primarily designed to address.
The problem of transfer-pricing games largely arises because U.S.-based corporations can artificially inflate payments made to foreign subsidiaries, so that taxable profits are attributed to the foreign subsidiaries and thereby escape the U.S. CIT (at least so long as those profits are not repatriated to the U.S.). For a simple example, consider a U.S. technology company that moves intellectual property to an Irish subsidiary. Because it is difficult to estimate with any precision the value of intellectual property, the U.S. corporation can then readily claim that the intellectual property is extremely valuable. This enables the U.S. corporation to justify making large payments to the Irish subsidiary as charges for use of that intellectual property. Because those payments are deductible under the U.S. CIT, the U.S. corporation can thereby strip taxable earnings out of the U.S. and shift them to Ireland.
Variations on the above transaction are rampant. The DBCFT would combat these transactions by disallowing any deductions for imports or other payments made to foreign persons or entities. Thus, under the DBCFT, the U.S. corporation in the above example could not deduct any payments made to foreign subsidiaries, for intellectual property or otherwise. Consequently, the DBCFT should successfully eliminate the tax advantage from traditional transfer-pricing games.
Along with denying deductions for payments to foreign persons or entities, the DBCFT would also exclude from its tax base any revenue from exports or from other sales to foreign persons or entities. As Hariton explains in detail, it is this feature of the DBCFT that would give rise to a new form of tax gaming. Hariton elaborates numerous techniques that U.S. business taxpayers could use to effectively render sales that are ultimately paid for by U.S.-based persons or entities to nevertheless be treated as excludable exports under the DBCFT.
For a simple example, consider a U.S.-based corporation selling software to a multinational customer that has offices in both the U.S. and in Ireland. The DBCFT would create strong incentives for the parties to label this as a sale into Ireland — paid for by the Irish operations of the multinational customer. But if the multinational customer subsequently used that software in its U.S. operations, rather than its Irish operations, how could the IRS identify this so as to challenge the tax treatment of the transaction? As Hariton explains, even with anti-abuse rules, any plausible implementation of the DBCFT would open an enormous door for U.S. taxpayers to recharacterize sales as exports so as to remove those sales from the DBCFT tax base, even for sales that are mostly paid for and benefit U.S.-based customers.
Of course, Hariton is not the first commentator to analyze how the DBCFT might be gamed. Yet, most of the previous discussion of such gaming has focused on how the DBCFT rules could be employed to strip earnings out of other countries’ corporate income taxes, effectively making the U.S. a tax haven. Notably, Avi-Yonah and Clausing’s discussion of tax gaming primarily explains transactions of this sort. Accordingly, supporters of the DBCFT have responded to these prior analyses of DBCFT gaming by noting both (1) that this sort of tax gaming is not necessarily against the U.S.’s interests, and (2) that other countries adversely affected by the DBCFT could combat this sort of gaming by following the U.S. lead and making their corporate income taxes destination-based.
Hariton’s analysis of DBCFT gaming is thus importantly different from previous analyses, because the DBCFT gaming opportunities that Hariton demonstrates would directly undermine the U.S. tax base. Indeed, prior to reading Hariton’s article, my view had been that the DBCFT would probably be considerably less subject to tax gaming of the sort that would undermine the U.S. tax base, as compared to the existing U.S. CIT (assuming that the details of the DBCFT were well-designed and implemented). Hariton’s article has convinced me to rethink that prior view.
Overall, then, we know that the existing CIT is subject to major transfer-pricing games, and Hariton has now provided strong reason to expect that the DBCFT would be subject to major gaming based on recharacterizing domestic sales as exports. Yet it is important to note that these two sorts of tax-gaming transactions are fundamentally different in nature.
Because the DBCFT would disallow deductions for payments made to foreign persons or entities, the transfer-pricing games currently plaguing the CIT would not undermine the DBCFT. Conversely, because the CIT includes export revenues in its tax base, the major gaming transactions that Hariton explains would plague the DBCFT do not undermine the CIT. To reiterate for emphasis, these major tax gaming transactions are fundamentally different in nature — the tax gaming transactions that would undermine the DBCFT would not undermine the CIT, and vice versa.
In a pair of articles, I previously argued that when two forms of taxation are subject to fundamentally different forms of tax gaming, it is often possible to substantially mitigate the harm from overall tax gaming by employing a mixture of the two different forms of taxation. More recently, building on that work of mine, David Schizer has argued that we should tax both corporations and shareholders, rather than only one of these, because the tax gaming transactions employed by shareholders and by corporations are fundamentally different in nature.
The key logic behind these arguments is that, in most tax policy environments, we can reasonably assume that taxpayers will engage in tax gaming transactions that cost the taxpayers less per dollar of tax savings, before moving on to tax gaming transactions that cost the taxpayers more per dollar of tax savings, and that taxpayers should then generally stop engaging in further tax gaming transactions once the cost of incremental tax gaming would exceed the marginal tax benefit. Based on similar logic, it is often said that doubling a tax rate approximately quadruples the harm from tax gaming. A corollary, then, is that — to the extent that two forms of taxation induce different forms of tax gaming — employing a mixture of the two forms of taxation at lower rates can approximately cut the harm from tax gaming in half as compared to raising the same amount of revenue from only one of the forms of taxation.
There are many caveats to this conclusion and reasons why this conclusion might not always apply, as I discuss in my articles. Most importantly, levying two forms of taxation can often increase administrative and compliance costs as compared to using only one form of taxation, and this can potentially counteract the benefit from mitigating the harm of tax gaming.
Returning to comparing the CIT and DBCFT, it might be possible to implement a hybrid of these two forms of taxation that would not greatly increase administrative and compliance costs. For instance, it has been reported that Mindy Herzfeld of Tax Analysts has suggested a “split the pie” approach that would make imports only 50% deductible and exclude only 50% of exports from the tax base. If implemented successfully, this would halve the effective tax rate affecting both traditional transfer-pricing games and the new recharacterization-to-export games that Hariton explains, thereby halving the marginal tax benefit to taxpayers from engaging in both of these sorts of games and potentially also approximately halving the overall social costs of these forms of tax gaming, as compared to either a pure CIT or DBCFT.
Another approach would be to dramatically reduce the CIT tax rate, while also introducing a new value-added tax, and simultaneously cutting or eliminating payroll taxes to make the overall reform package revenue and distributional neutral. Like Herzfeld’s split-the-pie approach, this alternative would move the U.S. tax system toward relying on a mixture of both source-based and destination-based tax principles. This approach could thus also potentially (approximately) halve the incentives and social costs of combined source-based and destination-based tax gaming.
Notably, this latter approach would reform the U.S. tax system so that it would better match the tax systems of every other OECD country. Unlike the DBCFT proposal, this approach should also clearly be compliant with WTO rules.
— — — — — — — — — — — — — — — — — — — — — — — — — — — — — —
Overall, then, there is strong reason to infer that the existing U.S. business-level tax regime relies too heavily on source-based tax principles. This is primarily because: (1) the U.S. stands alone among OECD countries in not levying a VAT, and (2) the U.S. levies one of the highest corporate income tax rates of any OECD country. Moreover, piling onto the problems that the federal-level CIT would create on its own, U.S. state governments are incentivized to pile on and exacerbate these problems by levying their own piggyback CITs with high rates and narrow bases, as my co-author Darien Shanske and I have explained here and here.
Based largely on Hariton’s article, it now seems to me that the DBCFT proposal would go too far in the other direction — moving the U.S. business-level tax regime to over-rely on destination-based tax principles. Instead, business-tax reform should aim for a better mixture of both source-based and destination-based approaches.
What should this mixture look like? What are the best options for designing a business-level tax regime around a mixture of both source-based and destination-based tax principles? These are important questions that should be explored by future scholarship.
