Today's Washington Post reveals a truly audacious stunt that the Treasury Department pulled in late September, essentially repealing on its own motion Code section 382 as applied to banks. The ruling through which it did this is available here, and it appears to be aptly described as flat-out repeal of the provision so far as banks are concerned.
Background for non-tax geeks (or tax non-geeks): companies can't deduct their net losses (you pay zero tax, but don't get a refund, whether your income for the year is zero or minus $10 billion). But losses thus rendered unusable can be carried over to other taxable years and used to offset taxable income in those other years. Loss companies therefore stagger around carrying "net operating losses" (NOLs) that they can use whenever they have offsetting positive taxable income, but in some cases they are unlikely to have such income any time soon or perhaps ever.
The NOLs are a tax asset, however, potentially making the companies (even if otherwise they are dogs) attractive to companies that have profits they would like to shelter. Section 382 greatly limits this little game by sharply reducing the ability to use the losses when you (i.e., another company) acquire a loss company.
The policy merits of this provision are decidedly mixed. For starters, why should losses be nonrefundable? An alternative approach to existing law would say that, if $10 billion of taxable income generates $3.5 billion of tax liability, then a $10 billion loss should generate a $3.5 billion negative tax, i.e., payment by the Treasury to the unfortunate taxpayer. Absent such a rule, the tax law discourages risk-taking and increases effective tax rates via a "heads we win, tails you lose" approach to tax liability. This is unambiguously bad policy (ignoring a complicating consideration that I'll add in a moment), and recent research by economist Alan Auerbach suggests that it has increasing adverse effects on corporate tax burdens because of greater dispersion in economic outcomes that makes losses more common.
There is, however, one decent rationale for loss nonrefundability. It serves as a backstop on the extent to which companies can derive tax benefit from generating fake tax shelter losses. Consider Enron. They were losing tons of money, but creating tons of positive taxable income through sham transactions that they used to generate bogus financial accounting income. They then offset the fake income with fake losses from preposterous tax shelters. But they could only use the shelters to drive their taxable income to zero. Absent nonrefundability, they could have kept on going and forced the government to pay them huge sums annually through the tax system.
Nonrefundability is thus defensible, which is not to say that it is clearly correct or unproblematic, so long as we remain unconfident that claimed tax losses are true economic losses. Section 382 then serves as a backstop by preventing the use of mergers as an endrun around the loss limit (create fake losses, then sell them to someone who can use them). It is a bad provision insofar as it tightens the constraint on using real losses that ought to be refundable, a good one insofar as it limits the abuse scenario, and a bit of a good one insofar as it prevents otherwise inefficient and socially undesirable mergers from being done simply for tax reasons (as the price of getting to buy the losses).
What's the balance of merits overall? Hard to say. But it is troubling to see the Treasury unilaterally repealing it as to banks, really beyond its proper authority (at least in normal circumstances) even if they can get away with it.
Even given the financial crisis, I don't think they should have flat out repealed section 382 as to banks. More limited relief, directed to the current financial situation and the next few years, would have made more sense and been less fundamentally improper. Perhaps policy enthusiasm for the broader repeal played a role.
Then again, given the change in Administration, it's plausible to me that the overreach will be allowed to stand just for now (again, perhaps justifiably given the financial crisis) and then reversed. And/or Congress can reverse it with deferred implementation of the rule restoring the provision's applicability to banks.
It will be interesting to see, in the years ahead, whether the Treasury reverses other Bush-era regulatory giveaways to corporate taxpayers, as there were indeed a lot of them and some at least were dubious (though often also defensible) on policy grounds. A recession might not be the right time to do any of this, but the new Administration should be around for a while, and the current Treasury has certainly left us here with a precedent for sharply reversing course, just because one wants to, when the time seems right.