Statement on the budgetary implications of President George W. Bush's proposed tax cut
March 08, 2001
Mr. Chairman, there are three major points I would like to make this morning regarding the budgetary implications of President George W. Bush's proposed tax cut.
First, the estimated revenue loss from it is too high, because the published estimates do not take into account the macroeconomic impact of the tax cut. Second, I believe that the baseline revenue forecast is too low, meaning that the impact of the tax cut on the surplus is too high. Third, the tax cut is being treated as if it will be a permanent feature of our tax system for all time. In the event that budgetary or economic circumstances change, it can be changed in the future.
Scoring Tax Cuts
As this Committee well knows, the impact of tax changes on aggregate revenues is highly uncertain. I believe that a key reason for this is that the revenue estimators use static scoring methods that ignore the macroeconomic effect of tax changes. Use of dynamic scoring, which is permitted under House rules, would give a more accurate estimate of proposed tax changes.
The basic issue is this. Large tax cuts, of the sort currently being proposed, stimulate economic growth. There is really no debate about this in the economics profession. The only question is how much will growth be stimulated. Clearly, there is great disagreement about this, owing mainly to inadequate statistical techniques for measuring different sorts of tax cuts. For example, a tax rebate and a permanent tax rate reduction of equal dollar size on a static basis will have very different effects on growth, both in the long- and short-term. But it is very hard for existing macroeconomic models to capture these effects.
In recent years, there has been considerable discussion of tax scoring methodology in prominent academic journals. Almost all agree that dynamic scoring is feasible, although there are important questions still to be resolved about how to do it and when. (I append to my testimony a bibliography of academic research on dynamic scoring.) But the basic point is undeniable that the tax cut currently under consideration by Congress will raise growth by some degree, thus reducing its net revenue loss.
In a recent study, the Heritage Foundation, utilizing the well-respected WEFA econometric model, estimated that faster economic growth will recoup almost half the static revenue loss . Over a 10-year period, the static revenue loss is estimated at $1.8 trillion. But this loss falls to $939 billion once faster growth is taken into account.
The Heritage estimate is probably a little on the high side, I admit. But it is very common for economists to get about 35 percent revenue reflow from a tax cut of this type. For example:
In 1981, Richard Musgrave, dean of American public finance economists, told the Joint Economic Committee that the Reagan tax cut would recoup 30 to 35 percent of the static revenue loss.
In 1996, Lawrence Chimerine, chief economist of the Economic Strategy Institute, wrote that "credible evidence overwhelmingly indicates that revenue feedback from tax cuts is 35 cents per dollar" (Washington Post, 7-23-96).
Just a few days ago, Martin Feldstein, Harvard professor and president of the National Bureau of Economic Research, found that the Bush plan would only lose 65 percent of its officially estimated revenue loss (Wall Street Journal, 2-16-01).
I don't know what the correct reflow assumption is, but I know with certainty that it is not zero. Ideally, I would like to see the Joint Committee on Taxation and the Congressional Budget Office work together to come up with a dynamic estimate. Six years ago, the House amended its rules to allow such an estimate to be made, but in all the years since, no one has ever asked. I think now is the time. Even if the JCT comes up with an estimate more conservative than those cited above, it will still give a better picture of the impact of the tax cut on budget totals than we have now.
The failure to use dynamic scoring inflates the budgetary size of the proposed tax cut. But from the point of view of this Committee, the absolute budgetary levels are also of concern. In this respect, I suggest that the baseline revenue forecast, both from CBO and OMB, are underestimating future revenue growth. Hence, even if the tax cut loses as much revenue as the static forecast projects, revenues may still rise by a greater rate than projected. In short, there may well be more revenue available for a tax cut, debt reduction and spending growth than currently assumed.
Historically, revenue growth has closely tracked nominal GDP growth. Since 1963, for every one percent rise in nominal GDP, revenues have risen by 1.04 percent. Since the Tax Reform Act of 1986, the elasticity has been slightly higher at 1.18 percent. Since the 1993 tax rate increase, the elasticity has grown still more to 1.37 percent. In other words, since 1993 federal revenues have increased 37 percent faster than GDP. That is the reason why revenues as a share of GDP have risen from 17.6 percent in 1993 to 20.6 percent last year.
Given this highly consistent trend, broken only by recession periods, it is very odd that both CBO and OMB are projecting future revenues to rise by less than the growth of GDP. Over the next five years, CBO estimates that nominal GDP will grow 5.2 percent per year on average, whereas individual income tax revenues are projected to grow only 4.7 percent per year. In other words, CBO expects revenues to grow 10 percent more slowly than GDP will grow over the next five years. OMB is even more pessimistic about revenue growth, stating that it projects revenue growth "to lag GDP growth throughout a multi-year period to a degree generally experienced only during recession."
Both CBO and OMB base much of this pessimism on an anticipated decline in capital gains realizations, owing to the fall in the stock market. However, the principal reason why revenues have historically grown faster than GDP is because of the progressive nature of the tax code. Although the personal income tax is largely indexed to inflation, it is not indexed for real economic growth. As real incomes rise, workers are pushed up into higher tax brackets in the same way inflation used to. Thus as long as there is real GDP growth, we will still get bracket creep.
A reasonable approach to forecasting revenues would leave out the recent revenue bulge resulting from exceptionally large capital gains realizations. But I think that the post-1986 trend for revenues, in which they grew 18 percent faster than GDP, is far more likely than the excessively conservative forecasts of OMB and CBO. On this basis, the federal government can expect $2.1 trillion in additional revenue, over and above that presently estimated, between now and 2011. That is more than enough to pay for President Bush's tax cut and still run surpluses as large as current projected without the tax cut.
The foregoing analysis indicates that President Bush's proposed tax cut is easily affordable. Nevertheless, some members have expressed concern that unforeseen events might alter that view. They suggest that the tax cut should be subject to some sore of trigger mechanism. If budget surpluses don't emerge as expected, the tax cut would be canceled.
I think the trigger idea is utterly unworkable. The principal economic benefit of lowering marginal income tax rates is so that people will change their behavior in ways that will increase work, saving and investment. For example, someone might seek additional education or training in response to the prospect of keeping more of their future income. Adding a high degree of uncertainty to whether legislated lower tax rates will emerge defeats this purpose.
Of course, future tax rates are always uncertain to some degree. Congress can and does change rates and other features of this tax code from time to time. This is another reason why triggers are undesirable. In the event that economic or political circumstances change in the future, Congress can always pass new tax legislation, raising revenues if necessary.
Some would suggest that this is unrealistic, because it is easier to cut taxes than to raise them. However, this view conflicts with the experience of recent history. According to a recent Treasury Department study (http://www.ustreas.gov/ota/ota81.pdf), there have been 15 major tax bills since 1980. Of these, 11 were tax increases. Ronald Reagan, the arch tax cutter, signed into law 6 of them, including the Tax Equity and Fiscal Responsibility Act of 1982, one of the largest tax increases in history. And of the 4 tax cuts, only the Economic Recovery Act of 1981 was significant.
Therefore, I would suggest that raising taxes is not nearly as difficult as it is painted. True, many of these tax increases occurred as the result of tense budget negotiations. But in most cases, the tax increase portion of those budget packages were their least controversial element. Of course, there are exceptions, as in 1990 and 1993. But the point still holds that Congress has shown a willingness to raise taxes by large amounts in recent years, when it felt that budget circumstances warranted it. Those who favor a tax trigger implicitly assume that Congress cannot or will not do the same in the future.
I believe that the proposed tax cut is affordable and necessary. Without it, the tax burden will rise from a level that is already extraordinarily high by historical standards. I believe it will help improve the performance of the American economy in future years, and is appropriately designed both in size and structure.
Neither Mr. Bartlett nor the National Center for Policy Analysis have received any federal government grants or contracts in the current or prior two fiscal years.
Literature on Tax Scoring Methodology
Auerbach, Alan J. 1993. Public Finance in Theory and Practice. National Tax Journal, vol. 46, no. 4 (December): 519-526.
Auerbach, Alan J. 1996. Dynamic Revenue Estimation. Journal of Economic Perspectives, vol. 10, no. 1 (Winter): 141-157.
Bartlett, Bruce. 1996. The Case for Dynamic Scoring. Wall Street Journal (July 30).
Bartlett, Bruce. 1997. A Victory for Supply Siders. Wall Street Journal (January 23).
Bopp, Michael D. 1992. The Roles of Revenue Estimation and Scoring in the Federal Budget Process. Tax Notes, vol. 56, no. 12 (September 21): 1629-1652.
Congressional Budget Office. 1978. Understanding Fiscal Policy. Washington: U.S. Government Printing Office.
Congressional Budget Office. 1982. How Changes in Fiscal Policy Affect the Budget: The Feedback Issue. Washington: U.S. Government Printing Office.
Congressional Budget Office. 1995. Budget Estimates: Current Practices and Alternative Approaches. CBO Papers (January).
Congressional Budget Office. 1998. Projecting Federal Tax Revenues and the Effect of Changes in Tax Law. CBO Papers (December).
Engen, Eric, Jane Gravelle and Kent Smetters. 1997. Dynamic Models: Why They Do the Things They Do. National Tax Journal, vol. 50, no. 3 (September): 657-682.
Feldstein, Martin. 1994. The Case for Dynamic Analysis. Wall Street Journal (December 14).
Feldstein, Martin. 1995. Revenue Estimates Should Reflect the Effect of Taxes on Work and Saving. In House and Senate Budget Committees, Review of Congressional Budget Cost Estimating, pp. 191-194. Washington: U.S. Government Printing Office.
Feldstein, Martin. 1997. How Big Should Government Be? National Tax Journal, vol. 50, no. 1 (June): 197-213.
Gravelle, Jane G. 1994. Dynamic Revenue Estimating. CRS Report for Congress, 94-1000 S (December 14).
Joint Committee on Taxation. 1997. Joint Committee on Taxation Tax Modeling Project and 1997 Tax Symposium Papers. Washington: U.S. Government Printing Office.
Kies, Kenneth J. 1995. The Revenue Estimating Process-Letting in the Light and Letting Out the Hot Air. Tax Notes, vol. 69, no. 3 (October 16): 373-375.
Lackman, Abraham M., and John M. Bryan. 1997. Does Dynamic Scoring Work? Examining the Repeal of the New York State Real Property Gains Tax. 1996 Proceedings of the National Tax Association, pp. 308-313.
Lyon, Andrew B. 1996. Should We Be Afraid of Dynamic Revenue Estimates? 1995 Proceedings of the National Tax Association, pp. 207-213.
Mauskopf, Eileen, and David Reifschneider. 1997. Dynamic Scoring, Fiscal Policy, and the Short-Run Behavior of the Macroeconomy. National Tax Journal, vol. 50, no. 3 (September): 631-655.
Schwartzstein, Linda A. 1996. Smoke and Mirrors: Tax Legislation, Uncertainty and Entrepreneurship. Cornell Journal of Law and Public Policy, vol. 6, no. 1 (Fall): 61-93.
Slemrod, Joel. 1998. Methodological Issues in Measuring and Interpreting Taxable Income Elasticities. National Tax Journal, vol. 51, no. 4 (December): 773-788.
Sunley, Emil M., and Randall D. Weiss. 1992. The Revenue Estimating Process. American Journal of Tax Policy, vol. 10, no. 2 (Fall): 261-298.
Triest, Robert K. 1998. Econometric Issues in Estimating the Behavioral Response to Taxation: A Nontechnical Introduction. National Tax Journal, vol. 50, no. 4 (December): 761-772.
Vasche, Jon David, and Hoang Nguyen. 1994. The Treatment of Feedback Effects in Revenue Impact Analyses. Tax Notes, vol. 65, no. 5 (October 31): 599-618.
Williams, Brad, Kristin Szakaly and Jon David Vasche. 1996. Dynamic State Revenue Impact Analysis: A View From California. 1995 Proceedings of the National Tax Association, pp. 214-221.