Should We Freeze Tax Expenditures? Yes, But Carefully

Len Burman, one of the nation’s leading experts on the tax system, proposes to freeze federal tax expenditures. His freeze would be a companion to the one proposed by President Obama in his most recent budget. Like the President’s freeze on domestic discretionary spending, Burman’s freeze would not begin immediately but would be implemented in the not-too-distant future. Burman, like the President, would not impose his freeze in a mechanical, inflexible way. Under Burman’s plan Congress would be obliged to trim some tax expenditures and eliminate others in order to maintain an overall cap on tax expenditures. An ideal freeze would lead to the abolition or modification of unjustified tax preferences.

It’s easy to agree with Burman on the necessity for shrinking the long-term gap between federal revenues and outlays. It’s also clear that a good place to start is with a thorough reexamination of the loopholes and tax preferences that allow some taxpayers and forms of income to be more lightly taxed than others. Tax preferences deserve the same careful scrutiny as public spending in regular government programs. If a tax preference is ineffective, counterproductive, or unfair, it should be eliminated or modified. Burman and I would undoubtedly agree on a long list of tax expenditures that in a sensible world would be trimmed, transformed, or abolished.

While I warmly endorse the idea of pruning wasteful tax expenditures, I think Burman may be exaggerating the likely scope of potential deficit reduction. The simplest way to curtail tax expenditures is to cut marginal tax rates. If the typical marginal rate were cut from 20% to 15%, the tax expenditure for a particular exemption or tax deduction would shrink one-quarter. Only in a supply-siders’ wildest fantasy, however, would a one-quarter cut in marginal taxes reduce the size of the deficit.
I see two other reasons to be cautious. First, the size of the tax expenditure “budget” cited by Burman probably overstates the true long-term cost of all tax expenditures. As the Office of Management and Budget points out, its tax expenditure estimates “… do not necessarily equal the increase in Federal revenues (or the change in the budget balance) that would result from repealing these special provisions.” One reason is that estimated tax expenditures on particular preference items may have reached their current level precisely because taxpayers have been encouraged to obtain certain kinds of favored income or incur certain kinds of expenses as a result of the tax preference. If the preference were eliminated, this kind of income or deductible expense would decline. Were Congress to eliminate the tax preference for retirement savings, for example, savings held in retirement accounts would probably begin to shrink, if only because some of the money currently held in the accounts would be taxed away.
The more basic issue is that several important tax preferences, such as the ones provided to capital income and worker savings, may be inherent components of an acceptable income tax system. These preferences may be politically necessary in order to make the income tax system palatable to middle class voters and economically efficient in maintaining appropriate incentives.
Consider how savings would be taxed in an income tax system that does not have any preference for retirement savings or capital income. A 21-year-old worker who earns a moderate wage and faces a 15% marginal income tax could consume all of her net earnings today or set aside part of her wages to pay for her future retirement. In the absence of any tax preference for capital income or retirement savings, she would pay 15% of her marginal earnings in taxes if she consumes all her wages today, or she could pay about 50% of her marginal earnings in taxes on the earnings she sets aside in a retirement account. The reason the latter tax rate is so high is that her investment income, if held outside of a tax-preferred account, would be taxed over and over again before her savings can be used to pay for consumption in old age. Current tax preferences for pensions and some kinds of capital income represent a realistic accommodation to an inherent shortcoming in a pure income tax system. Without some such accommodation, an income tax can severely distort consumers’ choices about whether current income should be consumed today or set aside for future consumption.
I am not arguing that tax preferences for retirement saving and capital income should be treated as sacred or inviolable. Their design can certainly be improved. My point instead is that even though some tax preferences may be departures from a pure income tax, they are nonetheless crucial for keeping the tax system politically acceptable and economically rational.