Mitt Romney may "love data", but he has a serious problem with math. He has simultaneously promised to (a) extend the Bush tax cuts and then slash all rates by an additional 20 percent and (b) keep Uncle Sam's total take "revenue-neutral" while (c) ensuring "that high-income people would continue to pay the same share of the tax burden that they do today." Like guaranteeing that the sun will rise in the west and set in the east, Romney's pledge is literally impossible.
Nevertheless, when the nonpartisan Tax Policy Center predicted that his $5 trillion, 10-year plan would deliver a tax cut windfall for the top five percent while hiking the effective rates for everyone else, Romney developed a novel new argument. The numbers problem, Romney insists, can be magically solved by closing tax loopholes and deductions. But which ones, of course, will remain secret until after the election. So any attempt to guess which of Mitt's mystery tax expenditures would be ended is therefore "garbage."

If that tortured logic sounds like the Bush administration's claim that "waterboarding isn't torture because the United States doesn't torture," it should. Romney made his circular argument in an interview with Fortune:

FORTUNE: The Tax Policy Center says your revenue-neutral tax-cut scheme lowers the overall tax burden on the rich and raises it on the poor and middle class. Your advisor Eric Ferhnstrom calls their analysis a "joke." What's your response?
ROMNEY: I indicated as I announced my tax plan that the key principles included the following. First, that high-income people would continue to pay the same share of the tax burden that they do today. And second, that there would be a reduction in taxes paid by middle-income taxpayers. Those are the key principles of my plan that the Tax Policy Center chose to ignore. Instead they made various assumptions about what they thought I would do which are not in fact accurate. They made an assumption that I would reduce the home mortgage-interest deduction. I will not do that for middle-income taxpayers, as I have already indicated. There's an old expression in the computer world: garbage in, garbage out. They made garbage assumptions and they reached a garbage conclusion. My tax policy will continue to have a very clear direction. We are not going reduce the share of taxes paid by high-income individuals, and we're certainly not going to increase the taxes paid by middle-income taxpayers.

In his "Nine takeaways on Romney's tax plan," Ezra Klein described the problem with Mitt's triple promise and his charge that TPC's analysis was "garbage in, garbage out":

The Tax Policy Center bent over backwards to make Romney's promises add up. They assumed a Romney administration wouldn't cut a dollar of tax preferences for anyone making less than $200,000 until they had cut every dollar of tax preferences for everyone making over $200,000. They left all preferences for savings and investment untouched, as Romney has promised. They even tested the plan under a model developed, in part, by Greg Mankiw, one of Romney's economic advisers, that promises "implausibly large growth effects" from tax cuts. The fact that they couldn't make Romney's numbers work even when they stacked all these scenarios on top of one another shows just how impossible Romney's promises are.
The reason Romney's plan doesn't work is very simple. The size of the tax cut he's proposing for the rich is larger than all of the tax expenditures that go to the rich put together. As such, it is mathematically impossible for him to keep his promise to make sure the top one percent keeps paying the same or more.

Which is one of the reasons why neither Romney nor his new running mate Paul Ryan will publicly say "in the light of day" which of the roughly $1 trillion per year in tax breaks the Republican ticket proposes to close.
For months, Mitt Romney has been using the same dodge as Paul Ryan, promising only that the "1 percent keeps paying the current share they're paying or more." In March, economic adviser Glenn Hubbard confirmed Romney's cowardice, explaining "it is not his intention to take on any specific deduction or exclusion and eliminate it." As the New York Times reported:

"In order to limit any impact on the deficit, because I do not want to add to the deficit, and also to make sure we continue to have progressivity in our tax code, I'm going to limit the deductions and exemptions, particularly for high-income folks," Mr. Romney, a former governor of Massachusetts, said. "And by the way, I want to make sure that you understand, for middle-income families, the deductibility of home mortgage interest and charitable contributions, those things will continue," he said. "For high-income folks, we are going to cut back on that, so we make sure that the top 1 percent keeps paying the current share they're paying or more. We want middle-income Americans to be the place we focus our help, because it's middle-income Americans that have been hurt by this Obama economy."

But you simply can't get there from here. Much of the estimated $1.3 trillion in annual tax expenditures in 2015 (a figure larger than the entire 2012 budget deficit and equivalent to about a third of the $3.8 trillion in federal spending next year) benefit working and middle income Americans. For example, the home mortgage tax deduction was worth $89 billion in 2011. Tax-deferred 401K accounts cost the Treasury $63 billion. The Earned Income Tax Credit had a similar $63 billion price tag last year. As the Washington Post rightly highlighted last year, "ever-increasing tax breaks for U.S. families eclipse benefits for special interests."

But even following Romney's approach by "starting at the top" (that is, offsetting revenue losses from tax rate reductions by first eliminating tax expenditures for the highest-income groups) the Tax Policy Center showed his math simply can't work. Mitt's Magic Unicorn simply doesn't exist:

It is not possible to design a revenue-neutral plan that does not reduce average tax burdens and the share of taxes paid by high-income taxpayers under the conditions described above, even when we try to make the plan as progressive as possible.

Just as damning, TPC pointed out that Romney's plan isn't just frighteningly regressive, but anti-family as well:

Families with children currently receive 57 percent of the available tax expenditures examined in this exercise but 23 percent of the revenue reductions.

Romney's pledge of revenue-neutrality, the analysis found, "would require deep reductions in many popular tax benefits ranging from the mortgage interest deduction, the exclusion for employer-provided health insurance, the deduction for charitable contributions, and benefits for low- and middle-income families and children like the EITC and child tax credit."
That's a lot of pain for middle class voters, which means it would even more painful for Mitt Romney to talk about during an election year. Which is why, Mitt made clear beginning in March, he won't:

"So I haven't laid out all of the details about how we're going to deal with each deduction, so I think it's kind of interesting for the groups to try and score it, because frankly it can't be scored, because those kinds of details will have to be worked out with Congress, and we have a wide array of options."

As Ezra Klein responded, "Let's be clear on this: A tax plan that can't be scored because it doesn't include sufficient details is not a plan."
That won't stop Mitt Romney from running around in circles to defend it.