How Faulty Official Figures Greatly Overstate the Cost of the Bush Tax Plan

Report Taxes

How Faulty Official Figures Greatly Overstate the Cost of the Bush Tax Plan

March 6, 2001 9 min read

Authors: Daniel Mitchell, D. Wilson and William Beach

Official estimates--even from the Bush Administration--greatly overstate the revenue "cost" of President George W. Bush's tax reduction plan. The reason: Official "bean-counters" typically assume that changing tax rates causes little or no change in work and investment.

In reality, of course, quite the opposite is true. Major changes in tax policy do cause changes in taxpayer behavior, and when taxpayers alter their behavior, they reshape the pool of income from which government draws its revenue. Lower tax rates lead to a bigger tax base, which leads to some degree of revenue feedback, lowering the net cost of the tax rate reduction. When this reflow effect is applied to President Bush's $1.6 trillion tax plan, the actual net revenue cost is less than $1 trillion, or 53 percent of the official estimate.

The Bush Administration's deliberate decision to understate the benefit of the tax relief and ignore the reflow effect may make it difficult for opponents to argue that the President's plan is based on a "rosy scenario"; but using static estimates in this way also will ensure that lawmakers have to rely on inaccurate data. "Static" revenue estimates produce the worst-case scenario and overstate the effect that tax relief will have on future tax revenues. When President Ronald Reagan cut marginal tax rates in 1981, for example, the net revenue reduction turned out to be a fraction of the official estimates. This is why "dynamic" analyses--based on revenue estimates that take behavioral changes into consideration--are better predictors of actual revenue effects.

Across-the-board tax rate reductions always result in a substantial "supply-side" effect. Lower rates encourage more work, saving, investment, and entrepreneurship, and this additional economic activity means more jobs and higher incomes. In other words, some of the tax revenue that is lost when rates are reduced will be recaptured because there is more income to tax.

This does not mean, of course, that all tax relief "pays for itself " or that this supply-side effect is instantaneous. Only in rare instances, such as in capital gains tax reduction, are supply-side effects that large. But it does mean that it is silly and misleading to estimate the impact of tax rate changes on tax collections without calculating the degree to which improved economic performance generates additional revenue that offsets the static revenue loss from lowering tax rates.

This discussion of the estimates underlying the President's plan has important implications for the budget debate in Washington this year. For example, in his address to Congress on February 9, President Bush proposed significant tax relief that includes lowering tax rates for all taxpayers, phasing out the federal death tax, easing the marriage penalty, and other changes in the tax code. Altogether, these changes will affect the incentives people have to engage in productive activity.

Yet most revenue-forecasting models do not estimate how much new tax revenue would be produced as a result of additional economic activity. Indeed, two commonly cited tax models rely solely on static estimates. The staff of the Joint Committee on Taxation (JCT) and the labor-funded Citizens for Tax Justice (CTJ) both argue that the 10-year change in revenue is generally equal to the accounting change ($1.372 trillion and $1.920 trillion, respectively).2

The JCT and CTJ static estimates are reminiscent of similar calculations of capital gains rate changes produced in 1989 by the Congressional Budget Office (CBO) for Senator Robert Packwood (R-OR), then Chairman of the Senate Finance Committee. When asked how much more revenue a 100 percent tax on capital gains declarations would produce compared with a tax of 28 percent, the CBO dutifully reported that increasing the tax 3.6 times would produce 3.6 times more in revenue.3 In other words, taxing away all of the profits from the sale of appreciated assets (like equities) would not in any way change either the stockholder's behavior or the number of times in a year that the investor would trade stocks. This is unrealistic.

What really happens? The effects of the tax policy changes during the last century repeatedly demonstrated that significant reduction in tax rates and the tax burden increases economic activity and produces new tax revenue. A recent Heritage Foundation analysis of the Bush tax plan shows just such an economic response.4 Rather than reducing revenues by $1.8 trillion over 11 years as a static estimate would predict, the net tax reduction would be considerably smaller because behavioral changes will generate $846 billion in offsetting revenues. Thus, the all-important net change in revenues (net of the changes in taxpayers' behavior) is $939 billion.5

Martin Feldstein of Harvard University also has estimated that the Bush tax cut would result in a substantial supply-side effect. According to his calculations, additional economic growth would boost tax revenues by $400 billion to $600 billion. Counting this supply-side effect, Feldstein estimates that the revenue impact of the Bush tax cut is actually $1.2 trillion or less.6 Perhaps even more important, he estimates that the economy would reap big benefits, since lower tax rates would reduce incentives to engage in inefficient tax planning activities that cause resources to be misallocated.

Large, significant economic "feedbacks" of revenues from tax rate cuts have occurred before. For example, after Congress in 1964 enacted President John F. Kennedy's tax legislation, which cut marginal tax rates from a high of 90 percent to 70 percent, revenues flowed into the U.S. Treasury at a rate of more than 4 percent above the level that forecasters had predicted.7 In other words, all of the reductions in tax revenue that the "accountants" had expected came back in the form of new revenues plus an additional 4 percent.8

When President Ronald Reagan significantly reduced tax rates in 1981 from a high of 70 percent to 50 percent, the economy returned about 76 percent of the static reduction in the form of new revenues. A static view of the Reagan tax cuts would place the revenue shortfalls at around $330 billion. In fact, when analysts compared real revenue under the new economy after the Reagan tax relief with expected revenue under the old economy (and laws), they found that the "loss" was just $78.9 billion.9 This lower "cost" of the Reagan tax plan occurred because taxpayers were making more taxable income because the economy grew as a result of the tax policy changes. From the taxpayers' standpoint, their incomes rose, their economic opportunities improved, and their financial security (i.e., savings) flourished.10

In the real world, changes in tax policy do affect economic behavior. By clinging to static revenue forecasts, lawmakers are relying on flawed numbers that significantly understate the beneficial impact of lower tax rates. The Bush economic team probably adopted this ultra-cautious approach to avoid becoming ensnared in a debate over "rosy scenario" budget numbers.

This may or may not have been a wise decision, but the White House should at least use the fight over tax relief to begin educating legislators about static vs. dynamic scoring. It should point out that the Bush tax plan is based on a worst-case scenario and that there is every reason to believe that the net reduction in taxes will be far smaller than the official--and misleading--estimate of $1.6 trillion.

William W. Beach is Director of the Center for Data Analysis at The Heritage Foundation. Daniel J. Mitchell, Ph.D., is McKenna Senior Fellow in Political Economy, and D. Mark Wilson was a Research Fellow, in the Thomas A. Roe Institute for Economic Policy Studies at The Heritage Foundation.

1. The authors gratefully acknowledge the assistance of Heritage intern Matthew Mitchell in the preparation of this paper.

2. The Joint Committee on Taxation issued estimates of then-presidential candidate George W. Bush's tax plan on May 2, 2000, and September 28, 2000. See Joint Committee on Taxation, "Analysis of George W. Bush Tax Reduction Proposal for Representative Bill Archer (R-TX)," May 2, 2000, and letter to Representative Charles Rangel (D-NY), September 28, 2000. The JCT estimates that the Bush plan would reduce federal revenues by $1.372 trillion from fiscal year (FY) 2002 through FY 2011. Citizens for Tax Justice published its estimates of candidate Bush's tax plan in August 2000; see CTJ, "Summary and Analysis of George W. Bush's Tax Plan: Updated August 2000," at  CTJ argues that the tax plan would cost $1.92 trillion over the period FY 2002 through FY 2011.

3. The current economics staff of the Congressional Budget Office likely would be the first to condemn their predecessors' crude tax analysis.

4. See D. Mark Wilson and William W. Beach, "The Economic and Budgetary Effects of President Bush's Tax Relief Plan," Heritage Foundation Center for Data Analysis Report No. CDA01-01, February 22, 2001. The Heritage Foundation's Center for Data Analysis uses the WEFA U.S. Macroeconomic Model to estimate the net effects from tax policy changes. The WEFA model is one of the oldest and most respected large, structural models of the U.S. economy. Fortune 500 companies and major government agencies routinely use this model to estimate how public policy changes will affect general levels of economic activity. WEFA, Inc., formerly the Wharton Econometric Forecasting Associates, is located in Pennsylvania.

5. The Heritage analysis indicates that a retroactive Bush tax plan would reduce revenues by about $1.8 trillion over 11 years rather than the 10-year estimate of $1.6 trillion that is commonly heard in the press. See Wilson and Beach, "The Economic and Budgetary Effects of President Bush's Tax Relief Plan," for an explanation of how this 11-year estimate was made.

6. See Martin Feldstein, testimony before the Committee on Ways and Means, U.S. House of Representatives, 107th Cong., 1st Sess., February 13, 2001, at

7. See Bruce Bartlett, "The Kennedy Tax Cut," unpublished manuscript, Table 1.6 and accompanying text; copy available upon request.

8. If the economy had continued on the same course and the government had collected a smaller portion of workers' paychecks, then revenues would indeed have declined precipitously--nearly $12 billion over the next four years, in fact. But that is not what happened. The Heritage analysis indicates that a responsive economy had, by 1966, produced revenues significantly better than static analysis predicted. Indeed, by 1968, federal revenues were $9 billion more than forecasted using static methods.

9. This calculation relies on the model created by Lawrence B. Lindsey in The Growth Experiment: How the New Tax Policy Is Transforming the U.S. Economy (New York: Basic Books, 1990). The word "loss" appears in quotation marks because it has a special meaning here. The Treasury was still bringing in $325.7 billion worth of income taxes in 1985, which was more than the $301.9 billion it had brought in the year before. So even though tax revenues increased after 1983, some still call it a loss because revenues did not increase as much as they might have if rates had been left higher.

10. For more details on the benefits of the Reagan tax cuts, see Peter Sperry, "The Real Reagan Economic Record: Responsible and Successful Fiscal Policy," Heritage Foundation Backgrounder No. 1414, March 1, 2001.


Daniel Mitchell

Former McKenna Senior Fellow in Political Economy

D. Wilson


William Beach

Senior Associate Fellow