The New York Times

May 5, 2005

What Should a Reconfigured Tax System Look Like?


TAX systems are like septic tanks: they need to be cleaned out every 10 years or so.

The last significant structural tax reform in the United States was in 1986, so we are long overdue for a thorough cleansing. Some economists have claimed that the changes in 1986 increased the economy's growth rate by as much as 1 percent in the years immediately afterward. But since then, virtually every piece of tax legislation introduced in Congress has added to the complexity and inefficiency of the tax system.

Last January, President Bush formed an advisory committee on income tax reform, which he asked to issue a report by the end of July 2005. Given the importance of tax reform to the economy, two noted tax economists, Alan J. Auerbach and Kevin A. Hassett, recently invited nine leading economic and legal scholars to prepare papers describing their own views. This collection of papers, "Toward Fundamental Tax Reform," can be ordered at

The authors span the political spectrum and have diverse views about the size of the public sector, how progressive the tax system should be and other value judgments. But as economists, they are all concerned with the same question: given that a government wants to raise a given amount of revenue, how can it raise it in a way that least distorts economic decisions?

Posing the question in this way naturally leads to an examination of marginal tax rates - the tax that individuals would face if their income increased slightly - since that is the relevant tax rate for deciding whether to engage in activities that could increase income.

Some might want to see marginal tax rates increase with income; some might like to see them stay relatively flat. But, as the authors point out, there is no coherent theory that advocates the "crazy quilt of blips" implied by the current rate structure.

Most economists would agree that it is better to have low tax rates on a broad base of activities rather than high taxes on a few activities. So, if you are going to tax income, say, it is better to tax income from all sources at the same rate and allow as few deductions as possible.

A system of this sort will have the least impact on choices among different sorts of income-producing activities. In addition, the fewer deductions there are, the less attractive it will be to engage in activities whose primary goal is reducing taxable income.

Consider, for example, everyone's favorite handout to the middle class: the deductibility of home mortgage interest. Many economists would argue that the favorable tax treatment of residences leads people to overinvest in housing.

If mortgage interest were not deductible, but overall rates were reduced so that people paid about the same total amount of taxes, then we would probably see a shift toward smaller and less expensive houses.

Needless to say, such a change would not be popular with many voters, so it is unlikely to be put in place anytime soon.

Of course, taxing income is not the only way to raise revenue. It might make more sense to tax consumption instead. Since income is equal to consumption plus savings, by definition, a consumption tax is equivalent to making savings tax deductible.

More savings translates into more investment and a higher living standard in the future, certainly desirable goals.

One could directly impose a consumption tax by using a Value Added Tax (VAT) as is done in Europe. But a less disruptive change that still captures some of the benefits of a consumption tax would be to unify the current messy system of tax-deferred savings, including I.R.A.'s, 401(k)'s, 403's and Keough plans. We do not really need all those different plans and having one, simple tax-deferred savings plan would make a lot of sense.

Many people essentially face a consumption tax now, given the plethora of tax-deferred plans, so simplifying the system should cause few disruptive changes in behavior. Simplification might even increase savings, since it would make the choices much less confusing.

Opponents of a consumption tax argue that it is regressive. But it does not have to be that way. There could be a generous personal exemption, which would substantially reduce taxes paid by low-income households. Or we could tax higher levels of consumption at higher rates by setting limits on the amount of savings that could be deducted, as we do now.

Unfortunately, such a change would discourage savings among the rich, who are most able to save more. One of the papers by R. Glenn Hubbard, a Columbia University economist, explores the distributional effect of a consumption tax in some detail and concludes that such a tax need not be strongly regressive.

Another source of tax complexity is the differing tax rates that businesses face. Single proprietors, partnerships, small and large corporations all face different sets of rules and rates. If there were a uniform set of tax rates for businesses, and these rates were closely related to individual tax rates, we would see significant savings in spending on accountants and lawyers.

The papers by David F. Bradford, a Princeton economist who recently died, and Robert E. Hall, a Stanford economist, describe how a unified tax system might work.

But the big problem with tax reform is how do you get there from here? It is true that the current tax code is a mess, but it is a mess for a reason. Every one of those loopholes has a constituency. Sometimes it is a broad constituency, like the one supporting the mortgage interest deduction. Sometimes it is a narrow but politically powerful constituency, like the agriculture or energy industries.

Any sort of change or simplification will require quite a bit of bipartisan cooperation and a substantial amount of logrolling. In theory, everyone can come out ahead with a simpler system. But in practice, finding a set of changes that are politically palatable will be difficult.

Hal R. Varian is a professor of business, economics and information management at the University of California, Berkeley.