CRS Reports on Costs of Modifying AMT

Congress enacted the precursor to the current individual alternative minimum tax (AMT) in the Tax Reform Act of 1969 to ensure that all taxpayers, especially high-income taxpayers, paid a minimum of federal taxes. Initially, the minimum tax applied to fewer than 20,000 taxpayers; in 2006 it applied to 3.5 million. Moreover, absent legislative action, the AMT will affect significantly more middle- to uppermiddle-income taxpayers: by 2007, up to 23 million taxpayers could be subject to the AMT.

The AMT's expanded coverage will occur because temporary increases in the basic AMT exemptions and provisions allowing taxpayers to apply certain tax credits against the AMT have expired. To keep the AMT from affecting more taxpayers in the future would, at a minimum, require (1) that the higher AMT exemption levels be maintained and that they be indexed for inflation, and (2) that all personal tax credits be allowed against the AMT. This "patch" is estimated to cost almost $60 billion in its first full year.

If the recent reductions in the regular income tax expire at the end of 2010, then the patch would cost $279 billion over the FY2008 through FY2012 period and $569 billion over FY2008 through FY2017. If the reductions in the regular income tax are extended past 2010, then the cost of the AMT patch rises to $1 trillion dollars over the same 10-year period. Repealing the AMT would be even more expensive.

While offsetting the costs of AMT relief may prove difficult, several approaches are being discussed. One approach would shift AMT taxes from middle-income to high-income taxpayers by some combination of higher AMT rates and taxing capital gains and dividends at the AMT rates. Proponents argue that including capital gains preferences in the AMT would be consistent with the original purpose of the minimum tax.

In other proposals, AMT relief would be financed by raising regular income tax rates; for example, raising the top three rates by 24%. Another approach might be to repeal or restrict some of the preference items under the regular income tax. However, income tax base broadening may be difficult because many of these provisions are both longstanding and popular.

Another approach to financing AMT relief might be to focus on reducing the tax gap. Although the tax gap is estimated to be in the neighborhood of $300 billion, the revenue yield from proposals to reduce the gap appears to fall far short of the amount needed to address the AMT.

This report will be updated to reflect legislative developments.

 Contents

Two main factors have caused the increase in the number of taxpayers affected by the AMT. First, the regular income tax is indexed for inflation, but the AMT is not. Over time, this failure to index has reduced the differences between regular income tax liabilities and AMT liabilities at any given nominal income level, differences that will continue to shrink in the absence of AMT indexation. Second, the 2001 and 2003 reductions in the regular income tax have further narrowed the differences between regular and AMT tax liabilities.² Because of these two factors, many taxpayers will find that their regular income tax liability has been so reduced relative to their AMT liability that, even though they have few (if any) tax preferences, they may still be subject to the AMT.

An increase in the basic exemption for the AMT and provisions allowing certain personal tax credits to offset AMT liability have temporarily mitigated the increase in the coverage of the AMT. For 2006, the AMT exemption was $62,550 for joint returns and $42,500 for unmarried taxpayers. Under current law the basic AMT exemption is scheduled to revert to $45,000 for joint returns and $35,750 for unmarried taxpayers in 2007. In addition, starting in 2007, several personal tax credits will not be allowed against the AMT.³

This report first reviews the cost of modifying the AMT to reduce its coverage. It then discusses a range of potential revenue offsets for these modifications, including alterations to the AMT and to the regular income tax. These measures include higher rates under both the AMT and the regular income tax, a reduction of preferentially treated items under the regular income tax, and measures to reduce the tax gap. Many of the revenue offsets examined are designed so that the burden of paying for AMT modifications is borne by those taxpayers for whom the AMT was originally intended (i.e., taxpayers at the highest income levels).

This grossed-up income becomes the tax base for the AMT. For tax year 2006, a basic exemption of $62,550 for joint returns, or $42,500 for single and head of household returns, is subtracted to obtain AMT taxable income. (These exemption levels are temporary and are scheduled to revert, in 2007, to their prior law levels of $45,000 for joint returns and $35,750 for unmarried taxpayers.) The basic AMT exemption is phased out for taxpayers with high levels of AMT income. A twotiered rate structure of 26% and 28% is assessed against AMT taxable income. The tax is 26% of AMT taxable income up to $175,000 and 28% of AMT taxable income in excess of $175,000. The taxpayer compares his AMT tax liability to his regular tax liability and pays the greater of the two.

It is important to note that even though a taxpayer may not be subject to the AMT, it can still affect his regular income tax liability. The reason is that, after 2006, some personal tax credits under the regular income tax are limited to the amount by which regular income tax liability exceeds AMT liability.⁴ Thus, a taxpayer who has a regular income tax liability of $5,000 and $1,000 of these affected personal tax credits will effectively see these regular income tax credits reduced in value by $300 if his AMT liability is $4,300.

When discussing the long-run (beyond 2010) revenue implications of modifying the AMT, it is critical to specify whether it is assumed that the 2001/2003 tax cuts are allowed to expire after 2010 as scheduled, or whether it is assumed that the tax cuts will be extended beyond 2010. Allowing the 2001 tax cuts to expire as scheduled will reduce the costs of modifying the AMT. Extending the tax cuts beyond 2010 substantially increases the costs of modifying the AMT. If the 2001/2003 tax cuts are extended then, as a rough estimate, the cost of most options for modifying the AMT would almost double. The revenue effects of several modifications to the AMT are shown in Table 1.

To keep the AMT from affecting more taxpayers in the out years than it did in 2006 would, at the least, require maintaining higher exemption levels and indexing the AMT for inflation. According to the Congressional Budget Office (CBO), this option (which is often referred to as a "patch" and assumes allowing personal tax credits against the AMT, and indexing the basic exemption levels of $62,550/$42,500 along with the AMT tax brackets for inflation after 2006) would reduce revenues by almost $60 billion in its first full year.

If the recent reductions in the regular income tax expire as scheduled at the end of 2010, then the patch will cost $279 billion over the period FY2008 through FY2012 (five-year cost) and $569 billion over the FY2008 through FY2017 period (10-year cost). However, if the reductions in the regular income tax are extended past 2010, then the cost of the AMT patch as outlined by CBO rises to $1 trillion dollars over the same 10-year period. If these changes are debt financed (not paid for by either raising other taxes or reducing expenditures), then the cost of the patch rises significantly because of the higher net interest outlays.

Repealing the AMT is even more expensive. The five-year cost of repeal could, at a minimum, be around $436 billion. The 10-year cost of repeal could range from around $842 billion to $1.4 trillion, depending on whether the reductions in the regular income tax are extended past their 2010 expiration. Again, if the repeal was debt financed, the cost of debt servicing would engender considerably more cost.

The additional policy options outlined in Table 1 would be less expensive than the patch or outright repeal of the AMT, but these options would mean more taxpayers would fall under the AMT in the future than in 2006.

                  Table 1. Revenue Costs of Modifying the AMT

  Table 2. Percentage of Taxpayers with AMT Liability by Adjusted Gross Income

On the other hand, there will be a significant expansion of AMT coverage for taxpayers with adjusted gross incomes (AGIs) in the $50,000 to $500,000 range if the basic AMT exemption reverts to its lower prior-law levels and certain credits are not allowed against the AMT. For example, in 2005 16.7% of taxpayers with AGIs in the $100,000 to $200,000 range were subject to the AMT. By 2010, over 95% of taxpayers in that income range will be subject to the AMT.

There are several reasons why a smaller proportion of high- income taxpayers, especially those with AGIs over $500,000, will be less affected by the AMT in the future than their lower-income counterparts.

First, the marginal tax rates high-income taxpayers face under the regular income tax tend to be higher than the marginal tax rates they face under the AMT. Under the regular income tax, the top two marginal income tax rates are 33% and 35%. The top rate under the AMT is 28%.

Second, under the regular income tax, personal exemptions are phased out for high-income taxpayers. So the fact that personal exemptions are not allowed under the AMT does not have a significant effect on high-income taxpayers who do not get personal exemptions under the regular income tax.

Third, under current law, the basic AMT exemption is phased out for taxpayers filing joint returns when AMT taxable income exceeds $150,000. For joint returns in 2007, the basic AMT exemption is fully phased out at AMT taxable income levels in excess of $330,000. Hence, many high-income taxpayers do not get a basic exemption under the AMT. Therefore, the scheduled reduction in the basic AMT exemption will have little or no effect on their AMT liabilities.

Finally, high-income taxpayers, unlike their lower-income counterparts, generally derive a significant percentage of their total income from capital gains and dividend income. These items receive preferential tax treatment under the AMT as well as the regular income tax. Under both, the tax rate on long-term capital gains and dividend income is limited to a maximum rate of 15%.

Income from dividends and capital gains is an important source of income at high income levels. Table 3 below shows the share of income in interest, dividends, and capital gains by income class. For example, in the highest adjusted gross income class (over $1 million), on average 30.3% of income is from capital gains and 4% is from dividends currently taxed at 15%. In contrast, taxpayers in the $75,000 to $100,000 AGI range derive 1.1% and 0.8% of their income from capital gains and dividends, respectively.

    Table 3. Interest, Dividends, and Capital Gains by Adjusted Gross Income

The revenue offsets are grouped into four main categories: offsets derived from (1) changes to the structure of the AMT, (2) changes to the individual income tax rate structure, (3) income tax base broadening, and (4) reducing the tax gap.

Offsetting AMT Revisions with Changes to the Structure of the AMT

One approach to dealing with the expanding scope of the AMT would shift the AMT tax burden away from taxpayers with AGIs in the $50,000 to $500,000 range and towards higher-income taxpayers. This approach appeals to some because it helps restore the AMT to its original purpose -- ensuring that high-income taxpayers pay what many argue is their fair share of the federal tax burden.

Revisions within the AMT could take many forms but three revisions are of special interest. AMT tax rates could be increased or dividends and capital gains income could be included in the AMT tax base at regular AMT tax rates, or both. Either increases in the AMT tax rates or a combination of increases in AMT tax rates and subjecting capital gains and dividend income to full AMT tax rates could raise enough revenue to offset the costs of an AMT patch.

Increasing the top rate of the AMT is one option for shifting the burden of the AMT to higher-income taxpayers. For example, in a paper written in 2002, researchers at the Urban/Brookings Tax Policy Center (TPC) proposed a revenueneutral AMT revision that included increasing the basic AMT exemption, lowering the income level at which the AMT rate increases, eliminating the phaseout of the AMT exemption, and raising the top AMT tax rate to 35%.⁵

The Citizens for Tax Justice (CTJ) recently presented a proposal to tax capital gains and dividend income at ordinary AMT tax rates rather than the lower dividend and capital gains rates applicable under the regular income tax.⁶ According to the CTJ analysis, such a change would pay for 87% of the cost of the AMT exemption increase (making the higher exemption levels permanent and indexing them) for the next four years. They also calculate that including capital gains and dividends in the AMT combined with an increase in the top AMT tax rate to 29% would make the proposal revenue neutral.

CTJ argues that this change would restore the AMT to its original purpose. The principal target and most critical preference in the initial minimum tax was the capital gains preference.

A more recent study by researchers at the TPC discusses a variety of revenue neutral revisions over the ten-year budget horizon (FY2007-FY2016) that would shift the burden of the AMT to higher- income taxpayers.⁷ One proposal for an AMT patch, which is very similar to the CTJ approach, would be paid for by including capital gains and dividend income under the AMT and subjecting them to AMT tax rates. Combined with a 3% increase in the AMT tax rates (to 26.8% and 28.9%) this option would more than pay for itself over the next four years, according to TPC projections.

However, the CTJ and TPC estimates assume capital gains realizations are fixed. If realizations decline in response to these tax increases, then other measures would be needed to make up the revenue loss. Both studies acknowledge this issue, although CTJ argues that capital gains realization responses are small. With a realization response, including capital gains as a preference item in the AMT will not raise as much revenue as these two estimates indicate. The Joint Committee on Taxation (JCT) takes an expected decline in realizations into account in its estimates of the revenue effects of changing tax rates on capital income. Based on the JCT estimate of the effect of the tax rate changes on realizations in the late 1980s, only about 43% of the static gain from increasing the tax rate on capital gains and dividend income would actually be realized. Recent evidence suggests the realization response would be such that up to 80% of the static change in revenue would be realized. (This issue is discussed in more detail in the appendix.)

It is worth reiterating that, in large part, Congress enacted the first minimum tax largely to recapture the capital gains income that was excluded under the regular income tax. In 1969, 50% of a long-term capital gain was exempt from tax under the regular income tax. The long-term capital gain income that escaped tax under the regular income tax was included as a preference item under the original minimum tax and subjected to minimum tax rates. Prior to 1969, many high-income taxpayers used the preferential tax rates on capital gains income to reduce their federal income tax liabilities to levels that Congress deemed inappropriate for their levels of economic income.⁸

Offsetting AMT Revisions with Individual Income Tax Rate Increases

Another option to pay for AMT revisions is raising the individual income tax rates. The TPC study contains a variety of scenarios where rates are increased to pay for the AMT revisions. For instance, increasing the top three individual income tax rates by 24% over the next four years would offset the cost of repealing the AMT. This would require raising the top three rates from 28%, 33% and 35% to 34.8%, 41%, and 43.5% respectively.

To pay for the AMT patch, the TPC estimates that the top income tax rates would have to be increased by 2% from 28%, 33%, and 35% to 28.6%, 33.7%, and 35.8%. In addition, this option would require taxing dividends and capital gains income at regular AMT tax rates. Over the first few years, this option actually increases total revenues by about $28 billion, according to TPC projections. (Note that these options do not account for any capital gains realization responses.)

Offsetting AMT Revisions with Income Tax Base Broadening

It is possible to pay for repeal or a patch to the AMT through income tax base broadening. There are many provisions in the income tax law that narrow the tax base compared to a tax base that approximates economic income, but many of these tax benefits are both popular and long established.

There are several legislative proposals that would use base broadening to offset AMT reforms. Senator Ron Wyden and Representative Rahm Emmanuel introduced legislation (S. 1927 and H.R. 5176) in the 109th Congress that would have repealed the AMT and provided for overall tax reform. The cost of this legislation would have been offset through income tax base broadening and increases in the individual income tax rates.

In the 110th Congress, Senator John Kerry has introduced legislation (S. 102) that would partially offset the cost of a one- year AMT patch by rolling back the lower rates for dividends and capital gains for 2009 and 2010. Under this legislation, the maximum tax rate on capital gains income would increase from 15% to 20% and dividend income would be taxed at regular income tax rates. The lower rates on capital gains and dividends are set to expire after 2010, and so increasing these rates for 2009 and 2010 would only have a short-run revenue effect for scoring purposes.

Based on revenue estimates done by the JCT in 2003, changing the tax rates on capital gains and dividends appears to increase revenues in the neighborhood of $20 billion per year for dividends and $4 billion for capital gains. These capital gains estimates involve some timing effects, and were estimated at a time when gains were smaller. (Capital gains realizations are very volatile.) Based on the most recent estimate of capital gains liabilities, $75 billion for 2005,⁹ a static gain from increasing the rate by a third (from 15% to 20%) would be $25 billion. The actual revenue gain would be about 40% of the static cost, or $10 billion if the realization response is around 0.7 (a 10% reduction in tax leads to a 7% increase in realizations); 60% or about $15 billion for a realization response of 0.46 range, 78% or about $20 billion if the realization response of 0.25 range, and 90% or about $23 billion for responses of 0.11. These effects would not cover the full cost of the patch, but could provide some revenue.

There are several sources of information on possible tax base broadeners. Discussions of base-broadening provisions, along with their revenue effects, are found in various tax expenditure documents.¹⁰Table 4 reports the largest tax expenditures for individuals in the tax expenditure list, and while many of them would provide sufficient revenue to pay for the AMT patch, most of them are longstanding provisions that are quite popular. There are many other smaller tax expenditures as well as significant tax expenditures in the corporate income tax which could also be considered. Because there are more than 160 separate tax expenditures, a discussion of them is beyond the scope of this report.

           Table 4. Largest Tax Expenditures for Individuals, FY2006

CBO has also provided a list of possible base broadeners that in many cases would modify rather than eliminate tax provisions.¹² For example, they discuss proposals for a 2% floor for state and local tax deductions and for charitable contributions, which would respectively raise $26.6 billion and $19.9 billion in FY2009. Floors already exist for some itemized deductions (casualty losses and medical expenses) and they tend to simplify tax compliance, especially in the case of charitable contributions. The CBO study from 2005 also discussed a proposal to limit health benefit deductions which would raise $30.3 billion in FY2007, and lower the ceiling on mortgage interest deduction from $1 million to $400,000, which would raise $4.9 billion in FY2009. A proposal to allow the benefits of itemized deductions at a 15% rate is estimated to yield $53.5 billion in FY2009. CBO also discussed proposals that primarily affect corporations, such as eliminating graduated tax rates ($3 billion in FY2009) and lengthening depreciable lives ($12.9 billion in FY2009).

Offsetting AMT Revisions by Reducing the Tax Gap and Greater Tax Compliance

The tax gap is the difference between the taxes legally owed and the taxes actually collected. The latest estimate of the tax gap (for 2001) indicated a gross tax gap of $345 billion.¹³ If a significant fraction of that gap could be collected, then it would be possible to finance the AMT revisions with tax compliance measures. (The net gap, after collections IRS expects to make with audits and other measures, is estimated to be $290 billion).

The discussion below summarizes some legislative options aimed at closing the tax gap.¹⁴ It is not comprehensive, as there are numerous small changes in rules and requirements that could alter tax compliance; these changes are discussed in some of the sources cited in this section. Unfortunately, there are no estimates for the revenue gains for many of the measures and those that are available indicate that the revenue gains are small.

Increased Withholding or Third Party Reporting. Tax gap estimates indicate that compliance is greatest with direct withholding, rather then with third party reporting. Wages, which are subject to withholding, have a 1% non-reporting rate. Interest, dividends, social security payments, pensions, and unemployment payments, which are generally subject to third party reporting, have a 4.5% nonreporting rate. For income that has some (but not substantial) reporting -- such as partnership/S corporation income, alimony, reportable exemptions, deductions, and capital gains -- the rate is 8.5%. Income not subject to reporting (which includes proprietorship income, rents, and royalties) has a non-compliance rate of 54%.¹⁵ Businesses with cash transactions are estimated to have a non-compliance rate of 81%.¹⁶

A number of possible revisions to increase third party reporting include the following:

• Withholding of non-employee compensation for independent contractors (a rule recently imposed on governments), or withholding if no taxpayer identification number is provided.

• Requiring credit card companies to report payments made to merchants.

• Requiring information reporting on services provided by corporations (presently corporations, including small corporations, are exempt).

• Requiring brokers to report basis on capital gains.¹⁷

• Requiring the reporting of proceeds of auction sales.

• Requiring reporting of real estate taxes by state and local governments.

• Requiring more detailed reporting of mortgage interest, including information on whether the loan is refinanced or exceeds the mortgage limit.

• Adopting a due diligence rule for preparers relating to offshore banks and similar operations.

• Reporting distributions to partners and S corporation shareholders in personal service businesses.

• Requiring corporations to report interest deductions limited by earnings stripping.

It appears that these changes have a limited ability to close the tax gap. For example, the gap arising from the failure to report basis (generally, the acquisition cost of the asset) on capital gains is estimated at $11 billion, and misreporting of gain is estimated at 36% for assets where basis is not reported, as compared to 13% for mutual funds where net capital gains are reported.¹⁸ These differences seem to imply a potential gain of several billion dollars. The Administration estimates in their FY2008 budget, however, that a basis reporting requirement after 2008 would yield a revenue gain that is less than $500 million per year by FY2012, and would provide only $200 million in cumulative revenue through FY2010.

For the first three items in the list above, withholding on independent contractors without a taxpayer identification number, reporting credit card sales, and information reporting on payments to corporations, the revisions would yield respectively $80 million, $1,174 million, and $832 million for FY2012. They would yield respectively $119 million, $1,211 million, and $1,099 million cumulatively through FY2010. (Note that full withholding on payments to independent contractors would yield more revenue, perhaps considerably more).

Information reporting on sales of tangible personal property (reporting proceeds of auctions) would raise $220 million in 2012, and $233 million through 2010. The Administration also proposed increased reporting on non-wage payments by governments, yielding $27 million by 2012 and $155 million through FY2010. All of the Administration's FY2008 compliance provisions, which include most of the significant items on the list above, would yield slightly over $3 billion through 2010.

The provision enacted in the Tax Increase Prevention and Reconciliation Act of 2006, which required withholding on independent contractor payments by governments, had a one-time increase of $6 billion (reflecting timing effects), but afterwards raised about $200 million a year. It is possible that expansion to withholding on all independent contractors would yield more revenue, but it would likely be a controversial proposal since it would affect many taxpayers (both payers and recipients).

For most of the proposals that have been considered to improve compliance and reduce the tax gap through third party reporting, the revenue raised is a small fraction of the revenue needed to pay for the AMT patch or other AMT reforms.

Tax Shelters. Many proposals to restrict tax shelters would have relatively small revenue effects, but there are some provisions that could help offset the cost of an AMT patch. Some of these provisions have been contained in prior legislative proposals and some in general tax reform proposals.

Economic Substance Doctrine. Even when transactions meet the letter of the tax law, tax benefits may be disallowed by the courts if the activity is found to be a type of sham transaction; in the particular case of tax shelters the related issues of economic substance or business purpose are often used.¹⁹ That is, if an activity does not have economic substance or there is no business purpose, the tax benefits are disallowed.

Several bills that were introduced in the 109th Congress would have codified the economic substance doctrine. For instance, a provision in S. 2020 was projected to eventually gain about $2.5 billion in revenue annually.

S. 2020 would have recognized these doctrines in the tax law itself and provided several specific guidelines. For example, if a court found the economic substance doctrine to be relevant, the bill provided that the taxpayer must meet both the objective test of economic substance and the subjective test of having a non-tax business purpose to keep the tax benefit. Requiring both is referred to as a conjunctive rule, while requiring either is referred to as a disjunctive rule.

This legislation sought to strengthen the rule and to bring more uniformity to court decisions. Some court cases have required both aspects to be met and others only one. The bill also set out specific rules for determining when the taxpayer meets the economic substance test of demonstrating profit potential, by requiring that the return outside the tax benefits exceeds the riskless rate of return. The bill would also require that the transactions must be a reasonable means of achieving the business purpose.

There has been some controversy about how much revenue codifying the economic substance doctrine would actually raise.

Earnings Stripping. Reducing U.S. tax with debt or other deductible payments to related firms is referred to as "earnings stripping." Because of the potential for abuse, the current tax code has a restriction on deductibility of interest for thinly capitalized U.S. firms (with more than 60% of assets held in debt and with more than 50% of earnings paid in interest).

H.R. 2896 in the 108th Congress would have tightened the general earnings stripping rules by eliminating the asset share test (i.e., disallowing interest based only on the interest as a share of income) and lowering the interest share (to 25% for ordinary debt and 50% for guaranteed, or 30% overall). This provision would have raised about $2.7 billion over 2004-2013 and would have brought in almost $400 million annually by 2013.

Some interest in earnings stripping developed from concerns about corporate inversions. A corporate inversion occurs when a U.S. company sets up a foreign incorporated firm to become the parent corporation, making the current U.S. firm the subsidiary corporation. Proposals to penalize corporate inversions have been considered in a number of previous Congresses, but some have not been enacted. A provision considered but not adopted in 2005 would have raised revenues by about $50 million annually.

International Tax Shelters. Earnings stripping and corporate inversions can be used as tax shelters by shifting income to low tax countries. There are numerous ways to accomplish this income shifting: by shifting debt to high tax countries, by shifting intangible income (such as that related to innovations and marketing), and through inter-company pricing. There are many possible legislative responses to these tax sheltering activities, although the revenue consequence of each is likely small.²⁰

Robert McIntyre,²¹ testifying before the Senate Budget Committee, summarized the recommendations that came out of that committee, along with some additional revisions that he suggested. Among them is a proposal to tax currently the income of foreign subsidiaries of U.S. companies. This proposal would have policy implications beyond reducing tax shelter opportunities. Under current income tax law, this income is not taxed until repatriated. Estimates suggest that taxing this income now rather than waiting until it is repatriated could result in a revenue gain of around $6 billion per year. Another proposal to curtail this form of tax sheltering activity is to provide an apportionment formula for allocating profits (allocate profits based on physical assets, employment, sales, or some combination).

Summing Up: Tax Gap Issues. The IRS estimates that there is a significant tax gap but the legislative measures that could be undertaken to close that gap appear limited. For many measures, there are no estimates of the revenue that could be generated. The revenue estimates that have been done tend to suggest only small revenue gains.

When assessing provisions aimed at reducing the tax gap it is critical to differentiate between potential collections (as measured by the tax gap estimates) and the amount of revenue that might realistically be collected through provisions aimed at increasing tax compliance.

Empirical studies had produced widely ranging estimates. There were two types of empirical studies: times series studies that examined aggregate changes in capital gains realizations over time as tax rates changed, and cross-section or panel studies that compared many individuals' realizations within a time period. The time series studies yielded a much smaller and less variable estimate of the realizations response than did cross sections studies.²²

There were difficulties with both types of studies. For example, time series studies confronted difficulties in controlling for other factors, such as falling stock transaction costs and changes in potential realizations. The cross section and panel studies also had some significant limitations. Perhaps the most serious of these potential problems was the likelihood that much of the response might reflect a transitory effect. High income taxpayers, who tended to have variable income, would time their realizations to coincide with periods when tax rates were low, so that the relationship between high realizations and low tax rates may have merely been a timing response, and did not reflect a response to a permanent change. The seriousness of this problem and the variable, and in some cases, extremely large, elasticities in these studies led the Joint Committee on Taxation (JCT) and Congressional Budget Office (CBO) to rely on their own time series estimates for their revenue estimating and forecasting purposes.²³

Most estimates of capital gains realizations responses were made using nonlinear forms. A popular one, and one used by JCT and CBO, was a semi-log function, where the elasticity (percentage change in realizations divided by percentage change in tax rate) rose proportionally with the tax rate.²⁴ At a tax rate of 0.22 (which is roughly the midpoint between the 15% capital gains tax rate and the top AMT rates), at the time, the JCT estimate of elasticity was 0.76, but after accounting for a portfolio response was 0.70. The first-year response was 70% larger at 1.20, so that, at least for a cut in the capital gains tax, realizations would be expected to rise so much the first year that projected revenue would increase.

Subsequent to the adoption of this elasticity, a number of studies and events suggested that the elasticity estimates should be lowered. First, the importance of transitory responses was highlighted by the huge surge of realizations that were eventually documented following the proposed capital gains rate increases in the 1986 Tax Reform Act. This observation increased concerns about transitory effects in cross section and panel studies. Second, a study that used the limits to the behavioral response (since realizations can never exceed accruals), suggested elasticities that were likely to be under 0.5.²⁵ Third, Burman and Randolph published a major new study using cross section data that controlled for transitory effects by using the variation in state tax rates to measure the permanent response to differential tax rates.²⁶ They found a very low permanent elasticity of 0.18 at a 16 % average tax rate, a number that was not statistically significant. This measure implied an elasticity of 0.25 at a 22% rate, using the semi- log estimating function. When weighted by population, they found a smaller elasticity, 0.06, with a 12% average marginal tax rate, implying a 0.11 elasticity at 22% rate. They found a transitory elasticity of 6.4. Finally, as data were added to the aggregate time series in CBO's continuing studies, the elasticities fell, and are currently at 0.46 at a 22% rate.

A subsequent study by Auerbach and Siegel confirmed the relatively low elasticities in the Burman and Randolph study (finding an elasticity of 0.33, and also finding elasticities at virtually zero for very wealthy and very sophisticated taxpayers).²⁷

The elasticity makes a great deal of difference for the share of the static revenue gain from capital gains that is estimated to be collected. Measuring all elasticities at a 22 percent rate, with the JCT's 0.70 elasticity, only 27% of the static revenue gain from moving the tax rate from 15% to the top rate of 28% would be collected. With CBO's 0.46 elasticity, 49% would be collected. Burman and Randolph found that 71% would be collected with the 0.25 elasticity and 86% would be collected with the 0.11 elasticity. This realizations elasticity, of course, only applies to capital gains. In the aggregate, 22% of the total of capital gains and dividends reflects dividend payments. Thus, if this share also applies to the AMT, the total shares collected would begin at 43% for the 0.70 JCT elasticity (0.22 plus 0.78 times 0.27), by 60% if one assumes the CBO elasticity, and 77% and 89% for the two Burman and Randolph measures.

These calculations apply only before 2011, or assume that the dividend and capital gains tax benefits adopted in 2003 will be made permanent. If one assumes that the dividend and capital gains relief enacted in 2003 will expire, then there will be no revenue gain from dividends at all and the capital gains tax rate will begin at a higher level of 20%. In that case, the share of static revenue collected will be smaller for example, 22% at the original JCT elasticity and 46% at the CBO elasticity, but about the same for the Burman and Randolph measures. The absolute value of the static gain would also be smaller than under the assumption that these provisions are made permanent, but total federal revenues would be larger.

² The Economic Growth and Tax Relief Reconciliation Act of 2001 (P.L. 107-16) and the Jobs and Growth Tax Relief Reconciliation Act of 2003 (P.L. 108-27).

³ Starting in 2007, the dependent care credit, the credit for the elderly and disabled, the credit for interest on certain home mortgages, the HOPE Scholarship and Lifetime Learning credits, the credit for certain nonbusiness energy property, the credit for residential energy efficient property, and the D.C. first- time home buyer credit will not be allowed against the AMT. The child tax credit, the adoption tax credit, and the savers credit will continue to be allowed against the AMT in full.

⁴ See previous footnote.

⁵ Leonard E. Burman, William Gale, Jeffrey Rohaly, and Benjamin H. Harris, The Individual AMT: Problems and Potential Solutions. Available at [http://taxpolicycenter.org/ publications/urlprint.cfm?ID=7912].

⁶ See Citizen's for Tax Justice, "A Progressive Solution to the AMT Problem," Dec. 2006, available at [http://www.ctj.org/].

⁷ Leonard E. Burman, William G. Gale, Gregory Leiserson, and Jeffrey Rohaly, Options to Fix the AMT, Tax Policy Center, Jan. 19 2007, available at [http://www.taxpolicycenter.org/home/]. This study examines a variety of revisions to the AMT that would be revenue neutral over a ten-year budget horizon. However, the baseline for their estimates is current law, which assumes that the reductions in the regular income tax expire after 2010. If the tax cuts are extended, either in part or in total, then these options to reform the AMT would not be revenue neutral.

⁸ U.S. Congress. Joint Committee on Internal Revenue Taxation. General Explanation of the Tax Reform Act of 1969, Dec. 3, 1970, p. 105. It is worth noting that after 1986, when the preferential tax treatment of capital gains income was repealed under the regular income tax, the number of taxpayers subject to the AMT fell from around 60,000 in 1986 to around 11,000 by 1989.

⁹ Congressional Budget Office, letter to Chairman Charles E. Grassley, Feb. 23, 2006.

¹⁰ Both the Administration and the Joint Committee on Taxation provide annual lists which can be found respectively in the budget documents and on the Joint Tax Committee's website. For a more detailed discussion of these provisions, see Senate Budget Committee, Tax Expenditures: Compendium of Background Material on Individual Provisions, Prepared by the Congressional Research Service, S. Prnt. 109-072, Dec. 2006.

¹¹Simple, Fair, and Pro-Growth: Proposals to Fix America's Tax System, Nov. 2005, which can be found at [http://www.taxreformpanel.gov/]. For a review see CRS Report RL33545, The Advisory Panel's Tax Reform Proposals, by Jane G. Gravelle.

¹² Congressional Budget Office, Budget Options, Feb. 2007.

¹³ U.S. Treasury Office of Tax Policy, A Comprehensive Strategy for Reducing the Tax Gap, Sept. 26,2006.

¹⁴ These proposals come from a variety of sources, including the Treasury study cited above, past administration budget proposals, and prior congressional proposals. They also include statements made at a hearing of the Senate Finance Committee on the tax gap, on July 26, 2006: testimony of Michael Brostek, Government Accountability Office; J. Russell George, Treasury Inspector General for Tax Administration; and Nina Olsen, Taxpayer Advocate, posted at [http://finance.senate.gov/sitepages/hearing072606.htm]. They also include reports by the Joint Committee on Taxation, including JCS-2- 05, Options to Improve Compliance and Reform Tax Policy, Jan. 27, 2005, and an unnumbered report, Additional Options to Improve Tax Compliance, Aug. 3, 2006. Finally, many of these proposals are included in the Administration's FY2008 revenue proposals. Their revenue estimates can be found in General Explanations of the Administration's Fiscal Year 2008 Revenue Proposals, Department of the Treasury, Feb. 2007.

¹⁵ U.S. Treasury Office of Tax Policy, A Comprehensive Strategy for Reducing the Tax Gap, Sept. 26, 2006.

¹⁶ Statement of J. Russell George, Treasury Inspector General for Tax Administration, before the Senate Finance Committee, July 26, 2006.

¹⁷ Currently gross proceeds are reported, but not basis. To institute this rule would require some type of rule about how basis is established when some but not all of a type of assets (for example, shares of particular stock) are sold. Currently, a variety of rules are available.

¹⁸ Government Accountability Office, Capital Gains Tax Gap: Requiring Brokers to Report Securities Cost Basis would Improve Compliance if Related Challenges Are Addressed, GAO-06- 603, June 2006.

¹⁹ For a general background on the economic substance doctrine see Joseph Bankman, "The Economic Substance Doctrine," Southern California Law Review, vol. 74, Nov. 2000, pp. 5-30; Gerald R. Miller, "Corporate Tax Shelters and Economic Substance: An Analysis of the Problem and Its Common Law Solution," Texas Tech Law Review, vol. 34, 2003, pp. 1015-1069; and Martin J. McMahon, Jr. "Economic Substance, Purposive Activity, and Corporate Tax Shelters," Tax Notes, Feb. 25, 2002, pp. 1017-1026. See also CRS Report RL32193, Anti-Tax Shelter and Other Revenue-Raising Provisions Considered in the 108th Congress, by Jane G. Gravelle, and CRS Report RS22586, The Economic Substance Doctrine: Recent Significant Legal Decisions, by Erica Lunder.

²⁰ See, for example, the report released by the Permanent Subcommittee on Investigations of the Committee on Homeland Security and Government Affairs, U.S. Senate, Tax Haven Abuses, The Enablers, the Tools and Secrecy, Aug. 1, 2006.

²¹ Proposals made by the chairman and ranking member of that committee are summarized in testimony of Robert McIntyre before the Senate Budget Committee, Jan. 23, 2007.

²² The issues surrounding the debate at that time are discussed in CRS Report 90-16, Can A Capital Gains Tax Cut Pay for Itself?, by Jane G. Gravelle. This archived report can be requested from the Congressional Research Service. It is also reprinted in Tax Notes, vol. 48, July 9, 1990, pp. 209-219.

²³ The Administration also had an estimate that was somewhat higher than that of the JCT and CBO but did not come from a explicit estimating equation. It is not clear how the value was chosen.

²⁴ Holding all other variables constant, under this formula, G = Ae-bt, where G is gains, A is a constant reflecting other variables and fixed amounts, b is the coefficient of the realizations response, and the elasticity is bt.

²⁵ CRS Report 91-250, Limits to Capital Gains Feedback Effects, by Jane G. Gravelle. This archived report is available from the Congressional Research Service. It is also reprinted in Tax Notes, vol. 51, Apr. 22, 1991, pp. 363-371

²⁶ Leonard E. Burman and William C. Randolph, "Measuring Permanent Responses to Capital-Gains Tax Changes in Panel Data," The American Economic Review, vol. 84, Sept. 1994, pp 794-809.

²⁷ Alan J. Auerbach and Jonathan M. Siegel, "Capital Gains Realizations of the Rich and Sophisticated," The American Economic Review, Papers and Proceedings, vol. 90, May 2000, pp. 276-282. This study also reported a very large 1.7 elasticity; however, discussions with researchers at the Congressional Budget Office (where this research was performed) indicate subsequent concerns that this measure is reflecting transitory effects.