What is Mitt Romney's effective or average tax rate?

He now says that it is 15%. But if that is his opening claim, I would say that the smart money is on an over-under that is considerably lower.

Romney appears to believe that he can get away with releasing only the 2012 return, which obviously is being prepared with current campaign circumstances in mind. I myself would regard it as preposterous if he is not compelled by political pressure to go several years back, in keeping with what is otherwise universal practice in presidential campaigns.

Would his tax returns for, say, 2007 through 2011 look about the same as that for 2012, given that he has been running for president (or preparing to) throughout this entire period? If he has made good decisions, clearly yes. But he appears so clueless about the difference between his circumstances and those of average voters that I suspect the answer is no. He may simply have figured: this (i.e., extremely aggressive tax planning) is how you do it, and how everyone does it.

But suppose he releases his tax returns. How should we interpret them? There will be a lot of discussion of what his average or effective tax rate is, and how that compares to that for more average taxpayers. Indeed, he himself engaged in this analysis by quasi-endorsing the 15% figure.

An initial question is whether to benchmark the number we end up with against average taxpayers' income tax liabilities, or their income plus payroll tax liabilities. And if we include payroll tax liabilities for average taxpayers, should that include the employer share? (This would require grossing up the income measure to which one's tax liability is being compared.)

There certainly is an argument for including payroll tax liabilities. But it is true that, when you earn wages that are subject to the Social Security portion of the payroll tax, you may also be earning expected retirement benefits from Social Security. (Technical note: you only pay Social Security tax on your first $110,000 of wages, and the formula for determining retirement benefits counts only your 35 highest-earning years, excluding amounts earned that were above the tax threshold.) So arguably your net payroll tax liability is lower than your gross payroll tax liability. Then there's also the lifetime perspective (Social Security and Medicare benefits versus payroll tax liabilities overall in present value, rather than current year). This is an important perspective but obviously well beyond what we can imagine focusing on here.

Now let's get to Romney's income tax return itself. One possibility would be to compare the bottom line income tax liability to his adjusted gross income (AGI). This may be the source of his 15% estimate, if he is paying mainly capital gains tax based on the highly controversial income tax treatment of "carried interest" paid to private equity managers.

But suppose he also has a bunch of tax shelter losses that reduce AGI. Then there would certainly be a strong case for grossing up AGI, for purposes of the effective rate measure, to reflect the noneconomic character of such losses.

Suppose further that, since so much of his income took the form of capital gains, he used an aggressive "strategic trading" strategy to generate offsetting losses. The basic trick is as follows. You hold a huge stock portfolio, figuring that some will go up in value while others go down. You then hold the winners and sell the losers, generating sizeable capital losses. In one sense, these are not fake - the stocks you sold actually did generate the losses claimed. But it may radically misrepresent your overall portfolio results.

Strategic trading has the potential to wipe out income tax liability for rich people with large stock portfolios. It is combated by the capital loss limitation, which limits individuals' net capital losses claimed to $3,000 per year. But this is not a constraint insofar as the income you want to shelter is taking the form of capital gains.

Thus, there would be an argument for grossing up Romney's income measure, for purposes of the effective tax rate computation, to disregard capital losses, insofar as we suspect that this is going on. But the right answer depends on the rest of his stock portfolio, which will not be directly observable from his tax return.

A related issue, which came up in public discussion of Warren Buffett's effective tax rate, pertains to unrealized asset appreciation generally. It's a bedrock rule in our tax system that this stuff generally isn't taxed. But it is a part of economic income, so if you are interested in tax liability relative to that, it oughtn't to be ignored.

Now let's take the unrealized appreciation issue one step further. Suppose Romney has huge unrealized gains that have accrued economically overseas, in particular in tax havens, without being currently taxable in the U.S. Suppose that his not including all this stuff in his income reflects the very aggressive, but (under current law) perhaps legally defensible, tax planning that reports I have seen on-line suggest is characteristic of Bain investments. This, like tax shelter losses, might be viewed as unrealized appreciation plus. Excluding it would give a false picture, making his tax liability seem higher than it really is as a percentage of economic income.

Finally, let's consider a deduction that would reduce taxable income but not AGI. I have seen reports suggesting that he gives millions of dollars in annual charitable donations to the Mormon Church. At least if they're done in cash rather than reflecting tax gimmicks such as the use of appreciated property, this really does reduce his remaining cash out of pocket. But on the other hand, it is a voluntary and discretionary outlay, best viewed as how he chose to spend his economic income rather than as a reduction thereto. And it does reduce his contribution to the public fisc. So arguably the income measure that we use to compute his effective tax rate should not be reduced (relative to AGI) by his charitable contributions. But on the other hand, the size of his charitable contributions may be relevant to the characterological conclusions that we would draw from getting to see his tax returns.

There are some complex issues here, not all of which have an absolutely clear right answer, and not all of which would be entirely illuminated if he consented to the level of disclosure that all other major presidential candidates have accepted for decades. But I say, release the returns and we can let the debate begin.