Latest Scholarship

January 19, 2018

States Should Tax the Over Two Trillion Dollars About to Be Deemed Repatriated (But Many Are Not Going To If They Do Not Act)

Here is the Why and the How (roughly)

By Darien Shanske

[Cross-posted from Medium.]

Under its pre-2018 international tax regime, the United States attempted to tax the income of multinational corporations on the basis of their worldwide income. To take a non-random example, the United States sought to tax Apple on its income earned all over the world. The old US international tax regime did permit multinational firms to defer payment of tax on the income they earned oversees - defer until the firm brought the money home. So, until Apple Germany sent home its profits to Apple US, those profits would not be subject to tax. Naturally, Apple and other multinationals let a lot - like over 2 trillion dollars a lot - of income hangout abroad.

One strategy, used for instance in 2004, to bring this money home to the US was to offer a special low rate. Repatriate now and pay 5.25% rather than the usual 35% rate. The tax law just passed, commonly referred to as the Tax Cuts and Jobs Act (TCJA), but actually having no name, applies much stronger medicine. Kind of. The TCJA deems all of this income to be repatriated and then applies a tax rate of 8% or 15.5%. The Joint Committee on Taxation estimates that this provision will raise $338 billion over ten years (see bottom of page 566 of the PDF). To return to Apple, this one company alone expects to pay $38 billion on $252 billion in repatriated earnings.

Given that there was no good reason for the tax on this income to have been deferred, this deeming provision is arguably pretty sensible. Alas, several other aspects of this part of the law make it a travesty, as so much else is in The Act with No Name. First, if this income was going to be deemed returned anyway, why not subject it to the actual rate that was avoided (35%) or at least the new very low rate (21%)? Second, this is one-time money. The Obama Administration had planned to commit the money from repatriation to infrastructure, including capitalizing an infrastructure bank. Instead, the Act with No Name uses this one time money for short-term and likely ineffectual economic stimulus.

But all is not lost. As Daniel Hemel has already explained, states should tax the deemed repatriation. The basic reason to do so is the same reason that the states should generally act to undo as much of this law as possible. It is terrible policy; it is squandering our national wealth for no discernible reason. Let's return to infrastructure. On top of the lost opportunity costs from spending this one-time money wisely, and thanks to the Act's exploding the federal debt, the federal government has only made itself less able to serve as a partner to the states in financing infrastructure. States should act to use this one-time money to do the right thing and establish their own infrastructure banks or finance other capital projects. (One other idea: states should consider using the money to start a climate science institute the way California started a stem cell science institute when the federal government dropped the ball in that area during the Bush II years.)

The states should also tax these repatriated earnings because it is efficient for them to do so. Ordinarily, states need to worry about taxpayer response. If this were another tax amnesty, then a state might worry that local firms would not repatriate their foreign earnings at all if the state imposed too high a tax. Or perhaps firms might move. But these earnings are coming home no matter what, and this year. Further, deemed repatriation is a one-time event as the United States fundamentally changes its approach to international taxation. The United States is shifting to a territorial system. The US will not even attempt to tax multinationals on their worldwide income in the future. (To use jargon, this is a rather inelastic tax base.)

States can tax this deemed repatriation, but their current tax systems are not designed to do so - or at least not well. New York just recently reported that it expected to net very little from this repatriation (see pp. 28-29). The details must await another time (or at least the appendix to this blog post), but it should not be surprising that state tax systems are not designed to effectively deal with the deemed repatriation. Consider the tax rate. The rate that states apply to the deemed repatriation should be pretty high. Again, multinationals cannot avoid repatriating this income. But state corporate income tax rates are set taking interstate competition into account. A state that usually taxes corporate income at 5% might well consider a 20% rate on the deemed repatriation appropriate - after all a 20% rate captures most of the windfall given to the corporations by the federal Act.

An even bigger issue is that states generally permit multinational corporations to choose to have only their income generated from the US subject to tax. This is called a "water's edge" election. Again, in the usual context of interstate competition, this makes sense. The law governing this election is complex and differs between states. Suffice it to say that these laws in many cases will permit multinational corporations to avoid paying state corporate income tax on much of their repatriated earnings.

So states should pass new laws that explicitly cope with this situation. (I get into the weeds of what this law might look like in the appendix.)

But can states do this? The answer, I believe, is yes - but with an explanation. States cannot reach out and tax extraterritorial value, but states can tax an apportioned share of the business income of a multinational corporation. They can also tax the non-business income of a corporation at the place of commercial domicile. What does this mean? Take Apple and California. California can say that it is going to tax Apple on its worldwide income, but subject to a reasonable formula that apportions that income to California. Only income generated by Apple as a unitary business can be apportioned. Income earned by Apple in some other way, say as investment income, can be taxed by a business' commercial domicile. In this case, this would also be California.

Apportionment is generally done by a formula. States will typically choose an apportionment formula appropriate to their competitive position. A market state like California apportions the income of multi-state corporations on the basis of sales. A big resource state like Montana only uses the location of sales for 1/3 of its formula, but also uses the location of property and payroll. But in the context of this one time deemed repatriation provision, states need not be overly concerned with choosing a competitive formula. Thus an apt formula might be: The income from the deemed repatriation should be apportioned on the basis of historical sales or property and payroll or state population, which ever is higher.

(As a backstop, the law might provide that any income found not to be apportionable is still subject to allocation.)

But can the rate on this repatriated income be higher than for other corporate income? I think there are good arguments that it can be. After all, states often have different rates for different kinds of income and even different kinds of businesses. See, for example, here and here. And this is leaving to the side the different effective rates that corporations pay based on any credits they might receive - or any penalties that might be imposed. If states can and do have higher taxes on less mobile businesses, on financial businesses and can impose significant penalties, then it seems that a state can also impose a higher rate on this repatriated income, which is a kind of like an immobile windfall, kind of like a financial asset and, though accumulating this revenue was not illegal, the extremes of deferral that some firms went to can be viewed as a kind of behavior that a state could reasonably want to discourage by applying a rate higher than would have been applied if the income had been brought back earlier.

To sum up, states should impose a special tax on the deemed repatriation at a high rate with a favorable apportionment formula. There will be litigation, of course, but I think the states will win.

Appendix

A first cut at some in the weeds issues. First, one should remember that in a case challenging state taxes in this context, the burden is heavily on the taxpayer. This might turn out to be very important.

Second, one might wonder if states really can change their tax systems to reach this income. Interconnected corporations often dividend income to one another and, when they do, the receiving corporation is often entitled to a "dividends received deduction" on the theory that the corporation sending the income has already been subject to tax. One might think the situation is different when the dividend is coming from abroad, and the matter is tricky, but in at least most cases the Supreme Court has held that domestic and foreign dividends must get similar treatment. Thus, if this repatriation is just a big dividend, then states probably cannot subject it to a special rate. But, unlike in 2004, the deemed repatriation is not categorized by the federal law as a "dividend." Rather, the deemed repatriation is another type of income that multinational corporations can create - "Subpart F Income." Don't ask what that is, but do note that it is not a dividend. Many - perhaps most (feel free to email me with information on this)- states do not tax Subpart F income either, as is the case in NY, but the states could and without violating the equal treatment of dividends rule. For a model of what this could look like, look no farther than California's special rule for the taxation of Subpart F income. (See in particular Cal. Rev. and Tax Code Sec. 25110(a)(2)(A)(ii)). Yes, this is obscure stuff. At least one analysis by a consultant to the California Senate Committee on Governance and Finance seems to agree with my take. This is the analysis of SB-337 (Bates)). California's approach is still far from optimal, but it seems to be the very least that states should do.