There's been a moderate to-do in the press recently regarding plans by Senator Hatch, the chair of the Senate Finance Committee, to release a corporate integration plan. Here, for example, is a May 18 BNA article, "Hatch Pushes Corporate Integration Despite Revenue Concerns," and more recently Tax Notes reported that Hatch is just waiting now for the Joint Committee on Taxation revenue score.
An interesting article, just posted on SSRN by Ed Kleinbard, provides some useful analysis and background. As the title suggests - "The Trojan Horse of Corporate Integration" - it is fair to say that Kleinbard is skeptical of the plan, which he notes represents a surprising shift in direction for DC policymakers, who had not recently been focusing on corporate integration.
Rather than repeat his analysis, which is well worth an independent look from all interested readers, let me just throw a gloss on a couple of aspects, as to which I'll emphasize different aspects than he does.
Suppose the plan, although its full details have not yet been publicly announced, takes the following form. Corporations get a full dividends-paid deduction, although, like deductions generally, it would not be refundable to the extent in excess of taxable income. But the amount of the dividends-paid deduction will also give rise to a withholding tax liability that the corporation will remit to the U.S. Treasury on behalf of shareholders that receive the dividends.
For simplicity, let's just focus on the current year, leaving aside the details that the plan will no doubt have regarding the treatment of excess distributions (i.e., dividends in excess of taxable income), carryovers between taxable years, etc. The main points of interest to me here emerge out of the basic one-year model where dividends DON'T exceed other taxable income.
To illustrate in the simplest way possible: Suppose that Acme Products has $10 million of taxable income. Assuming for simplicity a flat 35% corporate rate, if it did nothing further, it would pay $3.5 million of corporate tax. But instead, it pays $10 million of dividends to its shareholders. This zeroes out taxable income, so it pays the Treasury zero on its own behalf that is denominated an entity-level corporate tax. BUT - if the withholding tax rate is 35%, it pays $3.5 million just as if there had been no dividend payout, only this is now denominated a withholding tax on the shareholders. So they get $6.5 million, not $10 million - just as they would have if there were no dividends-paid deduction and the company had nonetheless decided to pay out all of its after-tax earnings as an immediate dividend.
What's happened so far at the entity level, by reason of the proposal, is purely a re-labeling. The company still remitted $3.5 million of tax, just as it would have in the absence of a dividends-paid deduction (or if it paid no dividends). But the taxes it paid are now deemed to be a shareholder-level tax, rather than an entity-level tax. Thus, as I further discuss below, it's plausible that the accountants, in their wisdom, would decide that paying the dividends (and hence the withholding tax) caused Acme's financial accounting income for the year to be $10 million, rather than $6.5 million. So the company has more reported earnings by reason of a labeling convention!
What happens at the shareholder level? This depends on further details of the proposal once announced. Under present law, obviously, the dividends would be taxable if the shareholder was. (But, as recent work by Steven Rosenthal and Lydia Austin shows, in practice about 75% faces no shareholder-level tax.) Under the proposal, the shareholder-level effect would depend upon (a) the relationship between the shareholder-level rate and the withholding tax rate and (b) the question of whether, and if so when, the withholding tax was refundable when in excess of the shareholder-level tax. I gather that a key part of the plan is to deny refundability to various or all tax-exempts, thus sticking them with the full 35% rate, which of course merely perpetuates the fact that they effectively get stuck with the entity-level tax under present law.
As Michael Graetz and Al Warren have noted in their work on the subject, this can be the same as having an imputation credit system of corporate integration. Nonrefundability (or refundability, as the case may be) can be made a feature under either system. But calling it a withholding tax, instead of using the usual framing, would presumably increase reported financial statement income.
At the risk of unduly repeating the example from above, let's start with Case A, involving standard imputation framing. Again, Acme earns $10M, remits $3.5M to the U.S. Treasury, shareholders get $6.5M (subject to whatever else happens at the SH level). But as there was formally an entity-level corporate tax, financial statement income is $6.5M.
Case B, dividend deduction plus withholding tax. In economic substance, everything is exactly the same (keeping in mind that SH-level tax effects can be made the same under both proposals). That is, we still have Acme earning $10M, remitting $3.5M to the Treasury, and SHs getting $6.5M. But now financial statement income is reported as $10M, because Acme merely remitted a tax that was formally defined as someone else's (i.e., the shareholders') obligation.
Here is another way to do the same thing. Congress passes an imputation proposal, rather than a dividend deduction proposal. But it commands the FASB to command accountants who are applying GAAP not to deduct entity-level corporate taxes from reported earnings insofar as any dividends generating imputation credits are paid out. So long as all the details are conformed properly, it can come out exactly the same. (Hence Kleinbard's question in his write-up: "Financial Accounting - How Stupid Are We?")
Suppose that I, unlike Kleinbard, wanted to argue in favor of the plan. What would I say? Probably I'd say that the point is as follows. Suppose that you like imputation but have observed that it isn't going anywhere, and that you think the third approach that I described above (commanding FASB to mandate non-deductibility from reported earnings of what are formally entity-level taxes) is optically unfeasible, By doing the proposal instead of imputation, you grease the wheels for its passage by giving corporate managers a big financial accounting benefit that - if (in Kleinbard's terms) we are indeed stupid enough - will make them big supporters. (BTW, whom should we think of as the relevant "we" in practice? Marginal investors? Big players? Someone else?)
Kleinbard doesn't buy this line of argument, partly because he's skeptical that "we" are that stupid, and partly because he notes that the most sophisticated thinking about capital income taxation has moved on from corporate integration that focuses on dividend payouts to more fundamentally addressing the taxation of capital income in general.
So let's throw another potential argument for the proposal onto the hopper. It's thought by some that this will have a positive impact on the "trapped earnings" problem for U.S. companies that have massively shifted their profits into tax havens, and also declared much of these profits to be "permanently reinvested earnings" (PRE), The PRE designation, if accepted by the accountants, permits one to score the deferred U.S. taxes at zero for financial accounting purposes, rather than at full value without regard to deferral (and to the prospect that they might never become payable at the currently applicable U.S. repatriation tax rate).
So let's go back to Acme. Suppose that Acme has zero U.S. taxable income, because all its earnings are through foreign tax haven subsidiaries. Say that it has $10M of PRE that faced zero in foreign taxes, and that it's unwilling to bring this money home, even though it wants to pay dividends to its shareholders, because then it would pay $3.5M in U.S. taxes that would come as a negative adjustment to earnings given the PRE designation.
Now the thinking is as follows. If we enact the Hatch plan, its $10M in taxable income from the repatriation is perfectly offset by a $10M dividends paid deduction. It still remits $3.5M of tax, and the shareholders still only get $6.5M due to the withholding tax, but now Acme avoids the negative earnings hit because the accountants now believe that this tax was merely remitted on behalf of someone else (i.e., the shareholders). No negative adjustment to PRE, so Acme is happy.
In this scenario, the Treasury also ostensibly is happy. After all, even under current law with the two levels of tax (subject to the Rosenthal-Austin point), how much in tax revenues did they actually expect to get? Zero in the current period, if Acme wouldn't have repatriated. And even in the long run, who knows - given, for example, the possibility that Congress will enact another repatriation holiday (a la 2004) at some point.
So, is everyone better off? Well, possibly not the shareholders. After all, given the chance of a future holiday, etc., they may have stood an excellent chance of, at some point, getting their hands on MORE than $6.5M (in present value) out of the $10M that Acme has squirreled away in the tax haven. Indeed, the boon to the Treasury, if it materializes, would appear to reflect managerial focus on maximizing reported earnings, as distinct from actually doing what's best for the shareholders. So perhaps the best argument for the proposal in the end (which Kleinbard acknowledges), is that it leverages managerial indifference to shareholder welfare into a mechanism for benefiting the public fisc.
If one believes that "we" really are as "stupid" as this argument for the proposal posits, does this potentially tip the balance in its favor? After all, if the managers are indifferent to shareholder welfare in various respects, why not have the fisc take advantage? Kleinbard does, however, offer various counter-arguments. I'll leave final bottom-line judgments to the readers.