Tax policy colloquium on Michael Schler's "Rebooting Section 356"

On the surface, it's difficult to imagine two papers that we would cover at the NYU Tax Policy Colloquium as different as Joel Slemrod's "Buenas Notches" paper from last week and Michael Schler's "Rebooting Section 356" yesterday.

As discussed here, Slemrod's article is a brief exploration of general conceptual and design issues raised by the concept of discontinuity in system design. Schler's is a considerably longer discussion of a set of relatively technical issues in the reorganization provisions of the existing U.S. corporate income tax.

Nonetheless, I wasn't actually kidding (just exaggerating for effect) when I kept saying: This is the same paper as last week. Because essentially it's about how to handle discontinuity and line-drawing problems.

When I first saw Schler's title, "Rebooting Section 356," but hadn't seen the actual paper yet, I was surprised by the choice of topic, because I had been inclined (from inattention) to think that the subject must be relatively cut-and-dried. Section 356 generally provides that "boot," or non-qualifying consideration such as cash, received by the shareholders in otherwise tax-free corporate reorganizations, is taxable to the extent of gain realized (but otherwise not recognized for tax purposes) on the transaction. E.g., suppose you have stock in company A with tax basis of $70 and value of $100. If company A merges into company B, and you swap your A stock for B stock, presumably also with value of $100, it's a tax free exchange. But suppose you instead get B stock worth $90 plus $10 cash. The cash is taxable to the extent not in excess of the underlying gain (which is $30, from the excess of the $100 value received over the $70 basis of the A stock).

So what could be the big issue? It turns out there are several, but the main one concerns rules addressing a trick that taxpayers might otherwise be inclined to pull. Suppose you are planning to pay yourself a big cash dividend, but realize that stapling the distribution to an otherwise pointless corporate reorganization (such as with a shell entity) would permit you to get more favorable tax treatment on the cash received - as it would now be section 356 boot - than you would have gotten by simply doing the dividend. Then jumping through silly hoops, by arranging an otherwise pointless reorg instead of merely distributing the cash, may become appealing from a tax planning standpoint.

This scenario is sufficiently realistic that section 356 was amended from its original form many decades ago in order to provide that boot (such as cash) distributions that are overly dividend-like in their attributes will generally be treated as a dividend. If section 356 worked exactly in parallel to the rules otherwise used to detect disguised dividends (e.g., on a purported redemption by a company of its own stock), then taxpayers would have no reason to attach otherwise pointless reorgs to their cash distributions as a way of improving the tax treatment. But the rules are not exactly parallel, for the most part seemingly due to historical happenstance. Thus, in some situations the idea of stapling a pointless reorg to a cash distribution may continue to be appealing, and in cases where that happens this may wastefully (from a social standpoint) increase transaction costs associated with tax planning.

The notches analogy is that we have discontinuous rules in the tax law for various categories - e.g., "dividend," "stock redemption," etc. - and how we draw the lines may affect efficiency. Perhaps a comparison to David Weisbach's paper on line-drawing would have been even closer, but we hadn't covered that in the colloquium.

Perhaps the most prominent and important disparity in the paper, and certainly the one we focused on the most, is the "gain limitation" rule that applies to dividend treatment of boot under section 356, but not to the treatment of dividends generally. Consider the earlier example of a reorg using stock that has basis of $70 and value of $100. Only, suppose now that you get $60 new stock and $40 cash. Suppose further that the cash distribution is effectively a disguised dividend given the broader circumstances. But for the gain limitation rule, you'd have a $40 dividend even though the overall gain is only $30. But of course a straight cash sale of the stock - assuming that this would NOT have been a disguised dividend (e.g., the recipient is departing from the scene rather than retaining voting control) would have led to a tax on only $30 of gain.

Schler, in keeping with several prior commentators from the tax bar who have examined the issue, wants to tax the above shareholder on a $40 dividend even though the total gain is only $30. The reason: This is exactly how the dividend rules work. Thus, while he would be willing to consider revisiting the dividend rules generally, to limit the taxable amount to underlying stock appreciation, he thinks consistency with the dividend rule is the most important thing here, so that taxpayers determined to extract cash in excess of appreciation, but who don't actually want to sell, won't simply devise more complicated and wasteful transactions (by stapling a reorg to the dividend payment).

Whatever one thinks of the argument in this particular context, this is actually a move that the existing tax system frequently makes, and indeed must make barring more fundamental reform. Thus, consider the rules affording stock redemptions more favorable treatment than dividends. (More favorable in that they get basis recovery, although pre-2003 they also usually were taxable at a much lower rate.) If the tax code didn't have rules taxing entirely pro rata stock redemptions as disguised dividends, it would generally would never make tax sense for anyone with taxable shareholders to pay a simple, straightforward dividend. Transaction costs permitting, it would be a no-brainer to simply take the more circuitous route of a pro rata redemption. (Given, e.g., that it should make no difference to me whether I have 10 of the company's 100 outstanding shares, or 9 out of 90.) Economic substance / business purpose rules, along with almost anything that limits the effective electivity of favorable tax treatment (e.g., claiming a taxable realization only when this is to one's advantage, such as to realize a tax loss), also can be thought of in similar terms.

It's worth noting, however, the craziness of the underlying dividend rule (leaving aside administrative rationales for it). Taxing shareholders on dividends when they don't have appreciated stock would be perverse even if one generally favored double taxation of equity-financed corporate income. Thus, suppose Jones creates a new company, issues zero-basis stock to herself, earns $100 (after payment of corporate tax) through the company by exploiting a good idea, and then sells the stock to Smith for $100. Double tax results from taxing Jones on the capital gain. But now suppose Smith is silly enough to give himself a $100 corporate distribution. This is a dividend, since made out of the company's earnings and profits, so Smith pays a dividend tax despite having no stock appreciation (its pre-distribution value, just like its basis to him, is $100). Only when he subsequently sells the stock for 0, realizing a $100 loss given its basis, does the triple tax revert to being merely double (ignoring problems such as whether he can deduct the capital loss).

In short, while we tend to think of the dividend tax as a rule for treating shareholder-level gain as realized when cash comes out, in fact it can exceed the shareholder's unrealized gain (if any) because there is no "gain limitation" rule outside section 356.

OK, perhaps I've allowed myself to get more interested in this topic than non-hardcore-tax readers of this blog would find suitable. But I do think that there is a conceptually interesting broader issue pertaining to how one should think about disparities of this kind in rule-drawing, i.e., under what circumstances is broadening a "bad" rule (and an even worse one if one dislikes double taxation) better than permitting it to be avoided through wasteful maneuvers?

The main criticism I offered (Mihir Desai was the lead commentator, but had generally similar views) was that, while it's hard to tell what is optimal in this sort of setting, it's natural for someone in Schler's position as a practitioner to focus (and potentially over-focus, in comparative terms) on the waste that he can observe from over-complicated deals that exploit such disparities. In other words, while the disparity imposes real tax planning costs, other real costs or benefits may be less visible. Plus there is a natural aesthetics of wanting to achieve clean alignment between different rules, which one could imagine a benevolent legislator being unmoved by.

It was a good feeling to see that a paper quite different from our usual fare, and that some prospective attendees might have predicted would not work out well in our somewhat navel-gazing colloquium setting, actually led to a really good session. (Though biased, I feel I can say that because my expectations are high enough that I'm quite willing and able to be disappointed when it doesn't work out as well.)

Overall, I feel we've had a good semester, and I should add with really good students (perhaps collectively our best ever). Just one week to go, and in my book it will be summer.