Monday, March 09, 2009

Rightside-Up Benefits, part 2

This past Friday, I posted some thoughts on the "upside-down benefit problem," which arises from the fact that tax deductions are more valuable to people as their incomes rise. Focusing on the deduction for charitable contributions, I created a hypothetical taxpayer named Zoe, with taxable income of $25,000, who would have only 15% of her contributions subsidized by the tax code. Anwar, on the other hand, declared $425,000 in taxable income and would thus receive a 35% subsidy. Some of the comments on my post offered interesting questions and suggestions about the issues raised by a plan from the Obama administration to partially fix this problem, and I thought I would offer some further thoughts inspired by those comments.

First, some clarifications. The current system provides two ways (other than rate cuts) for Congress to reduce people's taxes. Tax credits are straight-up reductions in a person's tax due. If I receive a $500 tax credit for, say, child care, then my taxes go down by $500. By contrast, tax deductions reduce taxable income, which is then multiplied by a person's marginal tax rate (their current tax bracket) to determine their tax saving. A person who receives a $500 deduction will reduce her taxes by $50 if she is in the 10% bracket but by $140 if she is in the 28% bracket and $175 if she is in the top 35% bracket.

A credit system with a flat rate would set a uniform percentage of total deductions to be credited against total tax liability, severing the link between the benefits from deductions and one's tax bracket. For example, a 25% credit would simply mean that every $4 of deductions would result in $1 of reduced tax liability, no matter the taxpayer's current bracket. (I also described a progressive crediting system, but I'll leave that aside here.)

The Obama plan would go only part of the way toward a pure credit system with a flat rate. Under his plan, someone like Anwar would reduce his taxes by $28 for a $100 charitable donation, because that is the maximum amount that can be credited. In other words, at income levels starting with the 28% bracket, there will be no connection between one's tax bracket and the tax savings that result from deductions. At lower income levels, the connection (and thus the upside-down benefit) would still exist. Therefore, Zoe's taxes would still go down by only $15 if she gave a $100 donation (and if she itemizes).

Some people on the comments board discussed various possibilities such as setting a uniform credit rate at the top bracket rate, which would not reduce the tax benefit from deductions for the highest incomes from current levels but would increase them for everyone else. (Again, even this only works for people who itemize, which leaves out a lot of middle- and almost all lower-income taxpayers.) I am relatively agnostic about this, though I note that setting the credit rate at the top income level is the most expensive option. I also see no reason to tie the credit rate to tax rates at all.

The broadest point that I wanted to make is that reducing people's taxes for any of the things for which we currently allow taxes or credits is the equivalent of a direct subsidy from the government. We could partially reimburse people for making charitable contributions, for example, simply by having them apply to an agency other than the IRS and then sending them a check. Running this through the tax code is, as I noted, an accident of history. While this accident has the advantage of being administratively quite inexpensive, it has the disadvantage of making people think of these things as "tax benefits" rather than just "government benefits" or "public subsidies." This connection to taxes thus muddles the debate by bringing in all of the white-hot rhetoric about the supposed evils of taxes. It need not be so.

Second, Mike -- proving that he remembers a lot more tax law than I remember con law -- pointed out that the Obama plan creates an interesting non-neutrality. People who are in the highest income tax brackets would have an incentive to provide services directly to charity rather than to earn income in their jobs and then give the money to a charity, because cash deductions will only save 28% while in-kind donations avoid 35% for the highest income taxpayers. Very true, as a theoretical matter. Mike makes no suggestion that this is empirically significant, but I will simply offer my hunch that this will not be a big deal. We are only talking about people whose taxable income in 2008 is roughly $165,000, which means that their gross income would typically be above $200,000. I find it hard to believe that this would be the type of taxpayer who would be likely to change from cash donations to labor donations on the basis of a few percentage points of difference. I could be wrong, but as my GW tax colleague Sarah Lawsky put it: "How many people are we talking about here? Ten?"

More to the point, this example applies to charitable contributions but not to the other major items for which we allow tax deductions. According to 2007 data, "charitable contributions other than education and health" (such as soup kitchens, but also including symphonies and opera companies) cost the Treasury just over $34 billion. The mortgage interest deduction cost almost $88 billion, and that cannot be gamed in the way that Mike describes. More broadly, we should not be focused only on charitable contributions, though this is mostly my fault for using that to motivate my examples.

Third, some commenters pointed out that there are some deductions that are available because they are necessary to allow us to measure income correctly. Miscellaneous employee expenses, for example, are properly taken out of income because they are part of the cost of producing income. We could, therefore, use this as an opportunity to think carefully about which deductions are part of computing income accurately and which are genuine subsidies that happen to be run through the income tax system, subjecting only the latter to the Obama rule or its extension. Apparently, the Obama plan would apply to all deductions as currently defined.

Even though I like theoretical purity as much as the next tax nerd, I am not especially concerned about this distinction, for two reasons. For one thing, the equivalence between phasing out or limiting deductions for higher-income earners and raising effective rates is easy to demonstrate arithmetically. If the Obama plan amounts to raising or lowering effective rates indirectly by limiting the deduction of certain expenses, that will hardly be the first time that we have deviated from a pure system. Raising the top rates after sorting out the deductions that should not be limited is an option, but the bottom line would be the same.

Moreover, we also violate the income-measurement aspect of the tax code for other reasons, such as ease of administation. We limit deductions for miscellaneous employee expenses to amounts exceeding 2% of adjusted gross income, meaning that we currently miscalculate income for literally everyone. We do so because of the saving in time and effort for people who are freed from keeping receipts for everything. More generally, I tend to be worried when we make tax changes that create a new type of exception to general tax principles or that significantly increase the size of an existing exception. Based on currently available information, neither of those concerns appears to be raised by the Obama plan or even by the two more thorough-going alternatives that I endorsed on Friday. The big changes are the de-coupling of benefits from tax rates and the mild increase in progressivity, which is why I am excited about this plan.

These thoughts are not meant to be definitive but only to suggest that I view the admitted impurities of switching from a deduction system to a credit system as not worrisome enough to make this a bad idea. Far from it. As always, however, further evidence and analysis could change my mind.

-- Posted by Neil H. Buchanan