Large capital gains tax cuts, including a 20% maximum tax rate and indexing of gains for inflation, are a central element in the tax bill passed by the House of Representatives on July 27, 1997. Official estimates by the Joint Committee on Taxation find that these capital gains tax cuts, after a transitional period during which it is asserted that the tax cuts would actually increase tax revenues, would cost in excess of $37 billion over the fiscal 2003-2007 period. This a large amount, but the latest capital gains research indicates that it may actually substantially understate the likely revenue loss from the House plan.
Citizens for Tax Justice has performed its own independent estimates of the cost of the House-passed capital gains tax cuts. CTJ finds the likely cost of these capital gains tax cuts over the next decade to be $169 billion--far more than the official government estimates.
The House-passed tax plan includes a 20% top capital gains rate (10% for 15% bracket taxpayers) and indexation of cost basis for individuals, plus a 30% maximum capital gains rate for corporations. In predicting the expected cost of these provisions, estimators must take account of likely behavioral responses by taxpayers. Specifically, history suggests that investors respond to lower capital gains tax rates by increasing the volume of capital gains realized, including not only some gains that might not otherwise have been realized but also gains generated by investment shifts and through tax-sheltering. There is wide disagreement among economists, however, as to the magnitude of this response and accordingly, how much of the revenue loss associated with a capital gains tax cut might be recouped in the form of higher tax payments.
The revenue estimates prepared by Congress's Joint Committee on Taxation of the House capital gains provisions assume a realization response that is very large--both by historical standards and with respect to the volume of gains the economy can reasonably be expected to generate in the future. Recent economic analysis suggests this huge assumed response is overstated and that the potential revenue loss from these capital gains tax reductions is substantially greater than official estimates indicate.
Citizens for Tax Justice's estimate of the cost of the House capital gains tax cuts relies
on a more modest prediction of investor response that is more in line with current
research. CTJ finds that the likely revenue cost of the capital gains tax reductions is $169
billion over the next ten years--far higher than what is presently being assumed by
congressional score keepers. Prudent budgeting dictates that these more realistic
estimates should be incorporated into the budget debate.
Individuals with accrued capital gains are accorded special benefits under several provisions of current U.S. Tax Law. First, a maximum tax rate of 28% applies to any realized net long-term gain resulting in a de facto exclusion of about 30% of realized gains for the highest bracket taxpayers. (The top regular rate is 39.6% when the 10% high-income surtax is factored in.) This exclusion is not available to the majority of taxpayers, whose regular tax brackets do not exceed 28%.
Second, since capital gains are only taxed upon realization, taxpayers who hold on to their assets get the benefit of a tax-free deferral not enjoyed by other types of investments.
Third, since investors can choose when to realize a capital gain, they can pick a time when their tax rate is temporarily low or they can realize offsetting capital losses on other investments. This timing option has been shown to be particularly valuable.
Fourth, if the taxpayer decides to hold on to an asset until death, the basis of the asset is stepped-up and capital gains escape income taxation entirely.
The House tax plan contains provisions that, if enacted, would considerably increase capital gains tax preferences. The top capital gains tax rate on individual taxpayers would be reduced to 20%--equal to a 50% exclusion of most net long-term capital gains for top-bracket taxpayers. In addition, the plan provides for indexing the cost basis of capital assets for inflation for assets acquired in 2001 or later. Corporate taxpayers would be eligible for a 30% capital gains rate (rather than the 35% maximum regular corporate tax rate).
Just how much capital gains tax cuts cost the Treasury has been the topic of much debate among government, academic and private sector economists. Indeed, the Senate Finance Committee held hearings in 1989 and 1990 to inquire as to the reasons for the wide disparity in the estimates presented by the Congressional Joint Committee on Taxation and the Treasury's Office of Tax Analysis. Presently though, both Treasury and the Joint Committee are in agreement on one thing: the capital gains provisions in the House tax plan (as well as its Senate counterpart) will cost the U.S. Treasury many billions of dollars a year once they are fully in effect.
It is the view of Citizens for Tax Justice that the official Joint Committee estimates are
overly optimistic. They fail to take into account new economic research which suggests
that the long-run, permanent response to capital gains tax cuts is much lower than has
previously been assumed. In addition, the United States has no experience under a system
that offers both an exclusion and indexing, so estimates of taxpayer behavior that rely on
historical data have to be interpreted with caution. The purpose of this report is to
present CTJ's estimates of the cost of the capital gains provisions in the House tax plan
using what we feel to be a more reasonable estimate of the realization response to tax
cuts.
Despite the wide disparity in empirical estimates of the behavioral response of individuals to changes in the capital gains tax rate, most researchers are in general agreement in at least three areas: that whatever the response may be, it is likely to be greater in the period immediately after a tax change than several years down the road; that taxpayers respond to marginal tax rates rather than, say, average rates; and that this response should vary with the absolute level of the marginal rate.
It appears that the Joint Committee on Taxation's revenue estimates of the Contract's
capital gains provisions assume an "elasticity" of about -0.70.(1) Together with some other
questionable assumption, that apparently led the congressional analysts assume that
about ninety percent of the cost of the proposal would be made up in the form of new
taxes on increased capital gains realizations.(2) CTJ believes that this assumption is far
beyond what most economists would deem reasonable, and that it is outdated in light
of recent empirical analysis.
In theory, both of these approaches should yield similar results. But in practice the two approaches have come to wildly different conclusions: elasticities based on time series analyses have been uniformly lower than those obtained from cross-sectional studies. Indeed, revenue estimates relying on cross-sectional analysis typically show actual revenue gains from lowering the capital gains rate. But quite the opposite is true when a time-series approach is used. The reason for this disparity, analysts have found, is that the results of the cross-sectional studies seem to be dominated not by tax law changes, but rather by temporary changes in particular taxpayers' situations. In other words, events such as a large business loss or an extraordinary medical deduction may temporarily reduce a particular taxpayer's capital gains rate independent of changes in the tax laws affecting capital gains. Thus, while cross-sectional analysis has the advantage of focusing on the actions of particular taxpayers, it has the disadvantage of failing to isolate changes in tax law from other events. Time-series analysis does a better job of isolating tax law changes, but may insufficiently reflect individual taxpayer responses to those changes.
Recent research has attempted to combine the best of the two approaches and reconcile their disparate results. Like traditional cross-sectional analysis, this new research examines how a group of individuals (a "panel" in statistical jargon) responds to changes in capital gains tax rates. But like time-series analysis, the new research takes its measurements over a period of time, and tries to weed out factors unrelated to tax law changes. In particular, it makes a crucial distinction between changes that are perceived as permanent by the taxpayers and those that are transitory in nature (such as large business losses or other temporary deductions). When adjustments are made to account for these transitory effects, permanent elasticities are lowered dramatically. A recent academic study has reported that "the permanent elasticity is not significantly different from zero."(3)
The analysis conducted by CTJ takes a middle-ground approach: we predict that
approximately one-third of the static revenue loss would be recouped in the form of
higher realizations over the budget horizon. While this behavioral response is somewhat
lower than that used by the Joint Committee, CTJ's estimated response remains
significantly higher than what current research suggests. It should also be noted that our
estimates are for the combined effect of the 50% exclusion and indexing. Since the United
States has never operated under such a system before, even CTJ's estimate may prove to
be overly optimistic.
Figure 2 above compares how projected realizations would change under the House tax plan Contract with three different predictions of taxpayer response: no long-term response (the baseline); Joint Tax's -0.7 elasticity; and CTJ's estimate.
In this context, the level of realizations that would be consistent with an elasticity of -0.70 appears so far outside the realm of historical experience as to appear implausible. This scenario would require a permanent increase in gains realizations so huge that it would dwarf even the record, temporary peak of 1986 in real terms. Indeed, the -0.70 elasticity assumption would require a staggering $1.4 trillion in additional capital gains realizations over the next decade compared to the historical baseline.
We believe that the behavioral response predicted by CTJ--which is, by the way, still quite large--is much more in line with both the historical record and mainstream economic thinking on this subject than are other approaches.
Table 1 below shows CTJ's detailed estimates of the cost of the House-passed capital
gains tax cuts over the fiscal 1997-2007 period. In total, we find that the capital gains tax
reductions would reduce revenues by $169 billion over the ten-plus years.
Summary and Conclusions
The 1997 House tax plan contains proposals which significantly reduce the amount of taxes paid by wealthy individuals and corporations on the sale of appreciated capital assets. The official estimates of these provisions is that they will increase tax revenues by $2.7 billion over the 1997-2002 period, and then lose $37.5 billion over the following five fiscal years. Even based on these estimates, one must question the soundness of such a fiscal policy because of its ever-growing cost in future years. Citizens for Tax Justice has concluded that the official figure are probably grossly optimistic. We find that a more reasonable estimate of the revenue cost of these proposals is $169 billion over ten years.
Back To Reports
1. See Joint Committee on Taxation, Explanation of Methodology Used to Estimate Proposals Affecting the Taxation of Income from Capital Gains (JCS-12-90), Mar. 27, 1990, Appendix A, p. 54.
2. "Elasticity" is the term economists use to characterize the responsiveness of individual investors to changes in prices--or taxes in the case of changes in the capital gains tax rate. Simply put, a capital gains elasticity is the percentage change in realizations as a fraction of the percentage change in the tax rate. (For simplicity, the discussion in this note ignores portfolio shifts and new tax-sheltering.) For example, if realizations were not affected by a tax change at all, then the elasticity would be zero. Similarly, if realizations increased in direct proportion to the tax change--so that revenues remained unchanged--then the elasticity would be -1.0. Roughly speaking, a useful shorthand way of summarizing the elasticity of a capital gains proposal is in terms of how much of the "static" (i.e., zero elasticity) revenue effect disappears as a result of increased realizations. Thus, if capital gains taxes were cut in half, then the "static" revenue loss would be about 50% of pre-tax-cut revenues. If the predicted elasticity were -0.3, then that static revenue loss would be reduced by about 30%. The official estimate of the House plan shows an even bigger reduction in the "static" loss over 10 years because of even higher elasticity assumptions in the early years and an extremely questionable assumption that many people will voluntarily pay taxes on accrued but unrealized gains in calendar 2001 to take advantage of indexing for existing assets in later years.
3. Leonard E. Burman and William C. Randolph, "Measuring Permanent Responses to Capital-Gains Tax Changes in Panel Data," The American Economic Review, Sept. 1994. See also Jane G. Gravelle, "Capital Gains Tax Proposals," Testimony before the Senate Finance Committee, Feb. 15, 1995 ("the revenue cost of a fifty percent exclusion . . . may be twice as large as [previously] estimated").