NYU Tax Policy Colloquium, week 10: Steve Rosenthal on who owns shares of US companies' stocks, part 2

 My prior blogpost offered some background concerning the methodology and main findings in Rosenthal's & Burke's Who's Left to Tax? US Taxation of Corporations and Their Shareholders. Here I shift to some preliminary thoughts on what it might mean, and what else we might also want to know.

Let's start  by considering the multiple recent rounds of global corporate rate-cutting, including the unfunded 2017 US corporate rate cut from 35% to 21%. Paul Krugman has noted that the foreign stock ownership figures from Rosenthal and Austin 2016 suggest that this was a bit like handing $700 billion in short-term transition gain to foreign shareholders - rather generous for a foreign aid program. But while most tax policy experts felt that the rate cut was too large, especially given the lack of funding and of adjustments to the individual tax base, they largely agreed that the US corporate rate should be cut below its prior 35% level.

Allow for offsetting adjustments, and most tax policy experts would be happy to go lower still. For example:

--The destination-based cash flow tax, endorsed by such leading economists as Alan Auerbach and Michael Devereux, and briefly considered by the House Republicans on 2017, would lower the US corporate rate to zero. True, there would still have been a tax collected from corporations and other businesses (and sometimes given labels such as "modern corporate tax"), but as I discussed here, that was essentially a value-added tax (VAT), not a corporate income tax.

--The Toder-Viard corporate tax reform plan would lower the US corporate rate to 15%, replacing it (for publicly traded companies) with mark-to-market taxation of the shareholders. An earlier version of the proposal would have eliminated the entity-level tax.

--The Grubert-Altshuler corporate tax reform plan would likewise lower the US corporate rate to 15%, while expanding taxation of shareholders so that they were taxed on the deferral of gain when they realized dividends or capital gains

--The Kleinbard BEIT business tax reform plan would have shifted the taxation of ordinary returns (although not rents) from the entity to the owner level.

All of these proposals (and others) reflect two fundamental premises. The first is that the rising mobility of businesses' tax residence, places of business activity, and places of reported profit supports shifting taxes from the entity level to the owner level. The Rosenthal-Burke paper does not contest this line of thinking.

Second, these proposals (and others) reflect the premise that flow-through taxation of entities, so that their positive net returns are taxed directly to the owners at the owners' tax rates, is the ideal. After all, the entity is just a legal fiction or web of contracts, and the only (or at least most obvious) reason for taxing it directly is just administrative convenience.

Rosenthal's findings challenge this, in a sense. Suppose that the taxes directly borne by certain types of shareholders, leaving aside any that are borne via their ownership interests in the entity (if it is a tapayer), are simply too low. Suppose, perhaps, that we even were counting (or should have been) on the entity-level tax to impose some tax burden on them. Then simply eliminating the entity-level tax, without replacing the lost indirect tax burden, might be undesirable. And if there's no other convenient way to tax these parties similarly, the entity-level corporate tax might even prove to be (with suitable modifications) the best instrument available for the job.

For me, a logical next stage, going beyond the content that would logically fit in the Rosenthal-Burke article, is to think further about each of the groups in the not-directly-taxable 75% that owns most of the existing US corporate equity. Here are some quick preliminary thoughts for each category.

FOREIGN SHAREHOLDERS (40%) - If the US corporate tax rate were reduced to zero, only withholding taxes would continue to apply to this group's earnings through the US businesses in which they own stock. To be sure, for foreign shareholders one needs to start with the incidence question. Even with a robust US corporate tax, to what extent can they be expected to bear the incidence of the entity-level taxes thereby imposed? (Or, for that matter, of the withholding taxes.)

In the long run, one can't make foreign shareholder bear US tax liability under classic "small open economy assumptions." These include assuming full capital mobility, normal returns only (no rents), and that US can't affect worldwide after-tax returns. So, if they can do whatever they like and demand, say, a 6% after-tax return, then a 25% US tax would merely cause them to demand an 8% pre-tax US return, shifting funds elsewhere as needed. They'd be "paying" the tax (even if via the entity), but not bearing it insofar as they were earning the same 6% after-US-tax returns as previously.

This is a nice textbook model, and one worth having in the back of one's mind, but it is not always entirely applicable under real world conditions. The reasons for its often not being true (or at least completely true) may vary somewhat as between their foreign direct investment (FDI) in the US, and their foreign portfolio investment (FPI).

Rents - for FDI, more than FPI, it may be contradicted by the existence of US-specific rents. If a company is earning site-specific rents through its US operations, it will keep earning them here even if they are partly taxed away. The reason this point pertains mainly to FDI, and thus to the foreign owners' US business operations, is that portfolio investors will presumably earn just normal returns (absent some other source of advantage) if they merely buy and sell publicly traded shares of US companies that may themselves be earning rents. For example, if I buy Amazon stock today, the rents that it is expected to earn are already built into the stock price that I will have to pay.

Diversification - If foreigners are eager to hold some US stock as part of a global diversification strategy - given the size of the US economy - then work I have seen (perhaps by the likes of Mihir Desai, Dhammika Dharmapala, and Thomas Brennan?) suggests that the US might have sufficient power to impose some tax on them, in incidence terms. It depends on how much they value any diversification advantages that they can only (or best) achieve in this way.

Value of US incorporation - Suppose foreign shareholders are willing to pay something for the benefits of US incorporation - due, for example, to the quality of US corporate law (e.g., in Delaware), or due to its having brand or branding value. Then they'll pay something for it. 

Transition - Suppose current US corporate stock prices reflect, not only the currently applicable 21% rate, but also the expectation that this rate will continue indefinitely, and is likely to decline as rise. Then Congress suddenly and unexpectedly raises the corporate rate to 28%, which now becomes the new expected mean. Shareholders, including foreign ones, will suffer an immediate transition loss from the change in information. (Stock prices would decline, all else equal, with the revelation that the US government will be increasing its share of corporate profits - assuming, of course, that the expected rate is now higher, not just the nominal statutory rate.)

Imposing tax burdens on foreign shareholders this way raises, of course, the classic time-consistency problem. But there is little present danger that, if we (say) increased the rate to 28% under a Biden Administration, that we would acquire a costly reputation as serial expropriators. (Note: the chance of a 28% rate under Biden has surely been priced into stocks to a degree, but it surely isn't currently considered certain.)

Should we want to impose US tax burdens on foreign shareholders? - Clearly yes if one's framework is the welfare of US individuals. Indeed, from that standpoint, and in a unilateral setting, one would want the revenue-maximizing tax, subject only to counting the welfare costs of getting the revenue to US individuals, such as from deadweight loss that they bear or lost positive externalities (e.g., from inbound investment that raises labor productivity).

Even with an exclusive US welfare focus, the answer might change if one is being strategic, rather than thinking unilaterally. For example, we might regret imposing the tax if it sufficiently increased the extent to which other countries impose tax burdens on US individuals who own stock in foreign corporations.

RETIREMENT ACCOUNTS: Two important distinctions here lie between: (a) defined benefit (DB) and defined contribution (DC) accounts, and (b) tax-favored accounts that get expensing (like traditional IRAs) and those that instead get yield exemption (like Roth IRAs).

DB versus DC - For DB plans that are in midstream - that is, their funding and benefit levels have been pre-set - the corporate tax on the stocks they hold to help fund future benefits may fall on the benefit provider. It gets less of an after-tax return to use towards funding the benefits. But this may shift to the beneficiaries insofar as it increases the default risk. Also, expected corporate tax levels when a DB plan is being created or newly extended will presumably affect what the beneficiaries pay (such as via lower cash salary) and/or what they get. So in that case it won't fall primarily on the owners of the benefit provider.

Insofar as DB is on the way out, rather than still being newly set up like DC benefits, the midstream perspective may be an apt one to take - subject, however, to concern about default risk. This might conceivably reduce one's distributional concern about the tax burden being imposed through the corporate tax on fund assets, if we think the benefit providor's owners are likely to be richer than the beneficiaries.

In DC, by contrast, the retirees holding the corporate shares through the plans are presumably the ones affected. I'll discuss below what one might think about taxing (nominally exempt) retirement saving as such, but the paper notes that these people are very much tilted towards the top 1 percent. Distributionally therefore, and also with regard to ensuring that millions of Americans have adequate rather than inadequate retirement savings, imposing the tax burden via entity-level corporate taxation might be considered just fine.

(Traditional (expensing) vs. Roth (yield-exempt) - These two methods of exempting retirement saving from the normal reach of the income tax are equivalent under a well-known set of assumptions - including, for example, uniform tax rates in all periods and a normal rate of return that is available for one's marginal investments without relevant limitation.

In practice, however, the two methods can play out quite differently, and in ways that are relevant to how we think about imposing tax burdens (via the entity-level corporate tax) on the holders of traditional as compared to Roth accounts. In particular:

(1) Expensing yields better-than-exempt outcomes if the marginal tax rate in deduction years exceeds that in withdrawal years. (Worse than exempt, of course, if the tax rate variation goes the other way.) Intervening statutory rate changes can affect this, along with variations in the amount of one's current-year taxable income.

In the yield-exempt model, seemingly no tax rates matter, since the savings are neither deducted or included. But here the exemption promise lies in the future, and thus is more easily revoked (and more likely to be revoked) than the tax benefit from expensing, once it lies in the past.

All this might make one more eager to tax (at the entity level) traditional than Roth-type retirement saving, if one expects lower future than current tax rates (e.g., due to retirement effects on current-year income). But other differences between the two systems may lean in the other direction.

(2) Under the expensing but not the yield-exempt method, (a) rents, exceeding the normal return and available only in finite quantity, are taxed at retirement, and (b) within the system itself, good luck increases one's tax burden, while bad luck reduces it. Both of these differences, I'd argue, add to the relative desirability of taxing yield-exempt retirement saving, such as via the entity-level corporate tax.

The difference as to rents is why David Bradford, in his classic 1980s Blueprints work, limited yield-exempt to what he called arm's length - e.g., your or my buying Facebook stock today, as opposed to Mark Zuckerberg's founding the company. Rents are one of the main things we want to tax, and if yield -exempt, unlike expensing, is shielding them then we might be glad at least to throw in some entity-level corporate taxation in the middle.

Then there's the luck piece. Ignoring risk premia for simplicity, suppose I have a 10% expected return, in that I will either win 40% or lose 20%, with equal probability. Say that, at a 25% tax rate, I expense $100 and then include at retirement either $140 or $80. So my $25 tax saving upfront is offset either by a $35 tax liability if I win, or $20 if I lose. This gives me a kind of insurance through the tax system.

The tax policy literature (including some of my contributions to it) tends to say: (i) Why not let people bet if they want to?, and (ii) If they want then we can't stop them, e.g., because they will scale the bets up or down to get to the same after-tax position anyway.

But if there's one thing that the empirical literature about retirement saving has taught us, it's that people don't actually do this, at least in tax-favored retirement accounts. (Sophisticated players with lots of money and good financial advice may indeed do it.) For example, we see that people scarcely respond to incentives in this realm, although they do to nudges, and that (per work by Brigitte Madrian) they don't invest nominally more under traditional than Roth IRAs, even though they need to do so if they want to keep the true-after tax retirement saving & payoff constant.

Thus, at the risk of tax policy community apostasy, I'd say that the insurance aspect here may both be desirable and resilient, increasing the motivation to impose a bit of it via entity-level corporate income taxation.

But should we tax retirement saving? - The paper says yes, because tax-favored retirement saving is so upward-tilted in terms of who holds it. The empirical evidence suggesting low salience of and responses to taxing it would  tend to push in the same direction. On the other hand, one might both regret its imposition on those who are not really saving enough for retirement, and believe (as David Bradford did) in tax neutrality as between working period and retirement period consumption.

Taxing nonprofits - For nonprofits such as Harvard with significant stockholdings, lowering the corporate tax rate increases their overall federal tax subsidy for reasons that have nothing to do with how much subsidy we do or should want to provide. In that sense, it seems anomalous. On the other hand, it's true that the optimal subsidy level is hard to determine. Might it be too low, rather than too high? I rather think not, as to the nonprofits that would be most affected (e.g., Harvard, as opposed to a soup kitchen). But this is admittedly a matter for debate.

Does all this mean we should move aggressively to increase effective tax rates under the US corporate tax - The paper notes that all the issues around mobility that have generated corporate tax rate reduction around the world (and academic suggestions to go even further, with suitable replacements for the entity level tax) cannot be wished away merely because we might like some of what the tax long did to be preserved. But I would say that, if one agrees with the paper that the entity-level tax's burden on foreign shareholders and the rich, through one mechanism or another, ought to be preserved or even restored to the extent possible, then there are two distinct paths forward to consider. (And one might do some of each.)

Preserving and restoring entity-level corporate income tazation - Here the logical steps to take - and they are complementary with each other- might include:

1) Raising the US corporate rate,

2) Expanding US taxation of foreign source income, such as by raising the tax rate in GILTI (by addressing both the inclusion percentage and the current 10% QBAI exclusion),

3) Making the source rules harder to avoid, such as via the use of sales-based formulary apportionment or something like it,

4) Imposing tougher anti-inversion rules and/or an exit tax for companies that cease to be US residents, and

5) Expanding the corporate residence rules, e.g., to apply to companies that are either incorporated or headquartered in the US.

The Biden campaign has called for #1 and 2 from the above (along with a book income minimum tax), and if they get the chance should also take a look at # 3 to 5.

Doing it a different way - For foreign individuals with US economic ties that make taxation feasible, one might explore optimal tariffs, along with the usual menu of choices in increasing taxation of the rich. (E.g., wealth taxes, inheritance or other gratuitous transfer taxes, expanded mark-to-market, realization at death, greater income tax rate graduation, etc.)

Let us hope that policymakers who care about the welfare of the United States and its people get a chance to think about these choices in the near future.