The nonpartisan Urban Institute publishes studies, reports, and books on timely topics worthy of public consideration. The views expressed are those of the authors and should not be attributed to the Urban Institute, its trustees, or its funders.
No. 5 in the Series, "Tax Policy Issues and Options"
The practice of requiring well-to-do Americans to pay a minimum tax was developed more than three decades ago. In January 1969, then-Treasury Secretary Joseph W. Barr informed Congress that 155 individual taxpayers with incomes exceeding $200,000 had paid no federal income tax in 1966. The news set off a political firestorm. Members of Congress were deluged with more constituent letters about the untaxed 155 in 1969 than about the Vietnam War. Later that year, Congress created a minimum tax to prevent wealthy individuals from taking undue advantage of tax laws to reduce or eliminate their federal income tax liability.
Historically, both the original minimum tax and the individual alternative minimum tax (AMT) that replaced it applied to only a small minority of high-income households. But under current law, this "class tax" will soon be a "mass tax." Current projections show the number of AMT taxpayers skyrocketing from 1 million in 1999 to 36 million in 2010. Without
reform, virtually all upper-middle-class families with two or more children will be paying the AMT by decade's end. This expansion will occur because the AMT is not indexed for inflation and because the 2001 tax cut reduced the regular income tax without making long-term cuts to the
The steep growth in the AMT would not justify alarm if it made taxation more fair, efficient, and simple. But the AMT's record on fairness and efficiency is mixed, and its structure is notoriously complex. For these reasons, the AMT must be reformed, if not eliminated, even though
fixing the AMT will be expensive. Indeed, by the decade's end, repealing the AMT will cost the Treasury more than repealing the regular income tax.
How did a tax originally designed to target 155 taxpayers grow so dramatically? What are the economics of the AMT? And what are the options for reform?
How the AMT Works
Taxpayers subject to the AMT must calculate their tax liability twice: once under regular income tax rules and again under AMT rules. If liability under the AMT proves higher, taxpayers pay the difference as a surcharge to the regular tax. Technically, the difference paid is their AMT.
To calculate their AMT, taxpayers add to their regular taxable income two categories of items called AMT preferences. Exemption preferences, which can be deducted from income under the regular income tax, are disallowed in the AMT. These items include personal exemptions,
the standard deduction, and itemized deductions for state taxes and miscellaneous expenses. Middle-income AMT taxpayers are the most likely to be hit by exemption preferences.
Deferral preferences allow taxpayers to postpone regular income tax payments by
hastening deductions or delaying income recognition. The AMT rules limit the extent to which taxpayers can use deferrals by, for example, allowing less generous depreciation deductions. Compared with exemption preferences, deferral preferences are more complex, tend to affect high-income filers, and generate less AMT revenue.
Once taxpayers add in all applicable preferences and tally income, they subtract the AMT exemptioncurrently $49,000 for married couples and $35,750 for singles. The resulting income level is taxed at flatter rates than under the regular income tax. The statutory AMT tax rate of 26 percent applies to the first $175,000 of net income above the exemption. For income over that level, a 28 percent tax rate applies. (Under the regular income tax, the same income
would be taxed at rates ranging from 10.0 percent to 38.6 percent in 2002.) Many taxpayers' effective AMT rate, however, is significantly higher, because the exemption phases out at a 25 percent rate over higher income ranges. The AMT parameters are not indexed for inflation.1 For a simple example of one family's AMT calculations, see the box.
"Class Tax" to "Mass Tax"
Under current law, the number of AMT taxpayers will soar over the next decade. The dramatic rise can be traced to prior changes, or lack thereof, in the regular income tax and the AMT. Most major tax legislation since 1980 has included changes in the AMT that broadly conform
to the reforms made in the regular income tax. Two notable exceptions, however, are the last two major tax cuts, which slashed the regular income tax without making conforming changes to the AMT. The Economic Recovery Tax Act of 1981 indexed the regular tax system for inflation but did not do the same for the AMT. The Economic Growth and Tax Relief Reconciliation Act of 2001 (EGTRRA) cut the regular income tax, but only made minor or temporary AMT adjustments.
Under current law, about 36 million people will be on the AMT by 2010, almost 14 times as many as in 2001. Without EGTRRA, the number of AMT taxpayers in 2010 would have been about 18 million. If the AMT had been indexed for inflation along with the regular income tax in 1981,
and if EGTRRA had not been enacted in 2001, only about 300,000 people would have to pay the AMT in 2010 (figure 1).
The focus of the tax will also shift, with a greater share of the middle class paying the AMT. In 2002, 1.4 percent of filers with incomes between $50,000 and $75,000 and 3 percent with incomes between $75,000 and $100,000 will face the AMT (all income classes are measured in 2001 dollars). By 2010, those figures jump to 43 and 79 percent, respectively (table 1).
Calculating the Bradys' AMT
The Bradys, a married couple with six children, have an income of $75,000 from salaries, interest, and dividends. Under the regular income tax, the Bradys can deduct $24,000 in personal exemptions for themselves and their children. They can also claim a $7,850 standard deduction. For the regular tax, their taxable income of $43,150 places them in the 15 percent tax bracket, and they owe $5,873 in taxes before calculating the AMT (or tax credits, which are allowed against both the AMT and the regular tax in 2002).2
To calculate AMT liability, the couple adds their preference itemspersonal exemptions of $24,000 and the standard deduction of $7,850to taxable income and subtracts the married-couple exemption of $49,000, yielding $26,000 in income subject to AMT. That amount is
taxed at the lower AMT rate of 26 percent, for a tentative AMT liability of $6,760. The AMT equals the difference between the couple's tentative AMT and their regular income tax, or $887.
Several points about this example are worth noting. First, the Bradys are on the AMT because they have a large family, not because they are rich or aggressive tax shelterers. Second, the Bradys' AMT situation is about as simple as it gets; they have no deferral preferences, no itemized deductions, no capital gains, no AMT credits from previous years, and no other complicating factors. Third, the Bradys will receive no long-term benefit from the 2001 tax rate reductions, because their income tax liability is set by the AMT, not the regular income tax.3 Finally, as long as the AMT is not indexed to inflation, the Bradys'
future tax payments as a share of their income will rise, even if their real (inflation-adjusted) income does not change.
|BOX 1. AMT Calculation for the Bradys
Married couple, filing jointly, with 6 children, 2002
|Calculate Regular Tax (before AMT)
||Calculate Tentative AMT
||Add preference items
| 8 personal exemptions @
|Taxable income =$43,150
||Subtract AMT exemption
||Taxable under AMT
||Tax (tentative AMT)
||(AMT tax bracket)
|Tax first $12,000 is taxed at 10%
next $34,700 is taxed at 15%
next $66,150 is taxed at 27%
|AMT = the excess of tentative AMT over regular income tax
AMT = $6,760 - 5,873 = $887
The AMT will become the de facto tax system for filers with incomes between $100,000 and $500,000: 95 percent of these taxpayers will face the AMT in 2010. At higher income levels, the share of taxpayers on the AMT falls, because the top AMT rate is lower than the top regular tax rate. Even so, in 2010, most filers with incomes between $500,000 and $1 million, and more than one-quarter of tax filers with incomes above $1 million, will pay the AMT.
In particular, the tax will hammer families with children and those that live in high-tax states. The AMT does not allow parents to claim exemptions for their children. In addition, it imposes marriage penalties, because the exemption for couples is less than twice the level for singles, and the tax rate brackets are not adjusted for marital status. By the end of the decade, couples will be more than 20 times as likely as singles to pay the AMT (not shown in the
table). Among married taxpayers with two or more children, 85 percent will face the AMT in 2010, including 99 percent of such families with incomes between $75,000 and $500,000. Because the AMT does not allow deductions for state taxes, filers in high-tax states are also more likely to face the AMT.
The 2001 tax cut will double a person's odds of being on the AMT by 2010. Before EGTRRA, 16 percent of taxpayers were slated to pay the AMT in 2010; post-EGTRRA, 33 percent will pay AMT. The 2001 law raised the likelihood of AMT liability by nearly 18 percentage points for filers
with incomes between $50,000 and $75,000 and by between 40 and 54 percentage points for filers with incomes between $75,000 and $1 million. In addition, by 2010, EGTRRA will more than double the share of adjusted gross income (AGI) subject to the AMT, from 26 percent to 55 percent, and the law will triple the cost of eliminating the AMT, from $47 billion to $141 billion. Indeed, by 2008, repealing the AMT would cost more than repealing the regular income tax.
As these figures show, EGTRRA worsened the AMT problem. Ironically, the AMT will also undermine the tax cut. By 2010, the AMT will "take back" about 36 percent of the overall income tax cut enacted through EGTRRA, including more than 70 percent of the cut targeted to taxpayers with incomes between $100,000 and $500,000.
The original purpose of the AMT was to reduce the number of high-income households that paid no federal income tax in a given year and to address overall fairness concerns, especially those related to aggressive or egregious tax sheltering schemes. In minimizing the number of
non-taxpaying high-income households, the AMT has been a success. Partly because of the AMT, the number of high-income filers that pay no income tax has not changed much since 1970. In 2001, an estimated 100 tax filers with incomes above $1 million paid no federal income tax, but at least 700 high-income tax filers owed no income tax before the AMT. The number paying no income taxes under AMT repeal would have been even higher if AMT repeal led to more tax sheltering activity.
More broadly, the AMT raises the overall progressivity of the income tax, though both the regular income tax and the AMT will become less progressive over time. The regular tax's progressivity will decline because the 2001 tax cuts increasingly benefit higher-income taxpayers over the course of the decade.
The AMT will also become less progressive, with millions of middle-class families becoming subject to it. Filers with incomes under $100,000 will account for 53 percent of AMT taxpayers in 2010, up from 24 percent in 2002. Those filers will account for 24 percent of AMT revenues,
compared with 8 percent in 2002. Only 9 percent of AMT revenues will come from taxpayers with incomes above $500,000 in 2010, compared with 33 percent in 2002. That income group will account for 30 percent of income tax revenues in 2002 and 26 percent in 2010. Thus, the AMT's ability to boost the progressivity of the income tax and the estate tax will erode just as the AMT becomes more and more burdensome to the middle class.
|TABLE 1. AMT Projections
||Percent of Income Tax
Cut Taken Back by
|Number (in millions)
|As percent of all taxpayersc
|As percent of all tax filers
|As percent of filers, by AGI
(thousands of 2001 $)
|As percent of income tax revenue
|Percent of AGI on AMT returns
|Cost of income tax repeal (billions)
|Source: Urban-Brookings Tax Policy Center Microsimulation Model.
NA = not applicable
a. Current law includes the effect of the Job Creation and Worker Assistance Act of 2002.
b. AMT taxpayers include those with AMT liability on Form 6251 and those with lost credits.
c. Taxpayers are defined as returns with positive income tax liability net of refundable credits.
In theory, a good tax system combines a broad base with low marginal tax rates. The AMT, however, often results in the opposite: a smaller tax base and higher marginal tax rates than the regular tax. As noted, tax rules broaden the AMT tax base by including preference items excluded from the regular tax base, but they shrink the AMT tax base by providing an AMT
For example, while the couple in our example (see box) has $43,150 of taxable income under the regular tax, they have an AMT tax base of just $26,000, because the AMT exemption ($49,000) is greater than the sum of preference items ($31,850). And their marginal tax rate under the AMT26 percentis much higher than their regular income tax bracket of 15 percent.
More and more taxpayers will find themselves in a similar position over time. The share of AMT taxpayers with more income being taxed in the regular tax than in the AMT is projected to rise from 66 percent in 2002 to 87 percent in 2010. The share with higher marginal tax rates under the AMT than under the regular tax will rise from 35 percent in 2002 to more than
90 percent in 2010.
Even if flawed as a stand-alone tax, the AMT could serve as a useful backstop to the regular tax. For example, the AMT's taxation of private activity bond interest income, which is exempt from the regular income tax, reduces the subsidy afforded such investments and could improve efficiency. This outcome, however, depends on two assumptions: that the subsidies in the regular
tax are bad policy, and that the AMT efficiently offsets the flawed subsidies. Although these assumptions might apply in particular cases, few analysts would argue they apply in any general sense.
The National Taxpayer Advocate and the Internal Revenue Service have called the AMT one of the most difficult tax law areas to comply with and administer. Complexity might be justifiable if it allowed policymakers to produce a fairer, more efficient tax system, but the AMT
produces questionable policy gains at best.
Because AMT rules on the timing of income recognition and deductions differ from regular income tax rules, taxpayers must keep two separate books. Moreover, any revenue gains from the recalculations are largely offset by a second set of complicated rules. These rules allow taxpayers to claim future income tax credits equal to the benefits lost through the first set of rules. Thus, on net, the rules compound the tax's complexity while contributing little to tax
revenue. The complicated rules do, however, serve an important policy goal: They reduce the number of high-income filers paying no income tax in a given year. But a simpler solution would be to scale back such preferences in the regular tax rather than requiring taxpayers to juggle two separate, complicated calculations.
Other sources of AMT complexity often serve no real policy goal. Most people who must fill out the AMT forms end up owing no additional tax. Increasingly, the tax will impose greater compliance burdens on middle-class taxpayers, a group that was never the tax's main target.
And the complexity also makes predicting marginal tax rates more difficult. Taxpayers cannot make informed economic choices if they do not know the rate at which additional income will be taxed.
Options for Reform
The underlying goals of the AMTrequiring high-income people to pay some tax, deterring the aggressive use of tax shelters, and ensuring progressivityare all sound, but the tax itself is replete with problems. We consider four AMT reform options that could help preserve the AMT's goals but would stem its explosive growth, reduce its complexity and distortions, and better shield the middle class. For comparison, we show the effect of repealing the AMT
altogether. We also discuss issues surrounding the financing of AMT reform and
consider modifications to the regular income tax and the estate tax that could complement AMT repeal.
Four Ways to Reform the AMT
Option 1: Index the AMT to inflation. Indexing the AMT to inflation after 2002 would
reduce the number of AMT taxpayers in 2010 by 71 percent. The number of AMT taxpayers with AGI of $50,000 to $75,000 would fall by more than 90 percent (table 2). However, under current law, this option would cost $370 billion in revenues through 2012 and $440 billion in total budget costs (including added interest payments).4
Option 2: In addition to inflation indexing, allow dependent exemptions and personal nonrefundable credits. This option would reduce the number of AMT taxpayers in 2010 by
83 percent. In addition, it would virtually eliminate the tax among individuals with incomes of $50,000 to $75,000.
Option 3: In addition to inflation indexing and the other measures in option 2, repeal the AMT exemption phaseout and allow deductions for state and local taxes as well as for miscellaneous expenses. Like option 2, these reforms would ease AMT requirements without creating aggressive sheltering opportunities. This plan virtually ends the AMT, reducing the number of AMT taxpayers by more than 99 percent. The additional measures, however, primarily benefit higher-income taxpayers. Relative to option 2, option 3 provides much larger tax cuts to households with incomes over $200,000. It is also much more expensive.
Comparison option: Repeal the AMT. Repealing the AMT after 2002 would add $788 billion to the public debt over the next decade. If EGTRRA were extended through 2012, the cost of repeal would rise to $950 billion. Despite the growing share of middle-class taxpayers that will pay at least some AMT, repeal would be very regressive, with the largest tax cuts going to the highest-income households. Under repeal, approximately 1,300 filers with incomes over $1.1 million (2001 dollars) and more than 17,000 with incomes above $200,000 would owe no federal income tax in 2010. These figures are several times greater than the figures projected under current law and under the other options outlined here. The main difference between reform option 3 and repeal is the effect of the deferral preferences. These provisions play an important role in ensuring that high-income tax filers pay at least some tax.
|TABLE 2. Reform Options
||10-Year Cost ($ billions)
||Number of AMT Taxpayers 2010 (millions)
||Percentage Reduction in AMT
Taxpayers, by AGI
|With No Other
|With a Partial EGTRRA
Freeze at 2002 Levelsa
|Index of the AMT after 2002
|Source: Urban-Brookings Tax Policy Center Microsimulation Model and authors' calculations.
NA = not applicable
a. Freeze EGTRRA income tax rate cuts and estate tax changes at 2002 levels. The top four statutory income tax rates would be 27.0, 30.0, 35.0, and 38.6 percent.The top estate
tax rate would be 50.0 percent, the 5 percent surtax would be repealed, the unified credit would be $1 million, and state tax credit rates would be reduced by 25 percent. Other
features of EGTRRA would be phased in as scheduled.
b. Option 2 indexes and allows dependent exemptions and personal nonrefundable credits.
c. Option 3 takes steps in Option 2 plus allows deductions for expenses and taxes and repeals the AMT exemption phaseout.
d. Option 4 indexes the AMT exemption for inflation, raises the top AMT rate to 35 percent, repeals the phaseout of the AMT exemption, and lowers the AMT rate bracket
thresholds (to $65,000 for married couples filing jointly, $48,750 for singles and heads of household, and $32,500 for married individuals filing separately), all effective after
2004. It also indexes the AMT rate bracket thresholds after 2010 and allows personal nonrefundable credits regardless of AMT liability after 2003.
Paying for AMT Reform
As our calculations show, AMT reform is generally expensive as well as regressive. One way to offset the revenue and distributional impact of AMT reform would be to freeze last year's tax cut at 2002 levels. Eliminating the reductions in income and estate tax rates scheduled to take
effect between 2003 and 2010, as well as abandoning the repeal of the estate tax in 2010, would raise revenue and remove the most regressive features of the tax cut.5
But even the combination of freezing the cuts in upper-income tax rates at their 2002 levels and holding the estate tax to its 2002 level could not fully finance the AMT reform options
described here. Each option would still require billions of dollars: about $31 billion for option 1; $79 billion for option 2; $228 billion for option 3; and $285 billion for outright repeal. Alternatively, we examine a revenueneutral reform of the AMT alone.
Option 4: Retarget the AMT to upper-income households without sacrificing tax revenues. This option would index the AMT exemption to inflation starting in 2005, allow taxpayers to use all personal credits against the AMT, eliminate the phaseout of the AMT
exemption, raise the top AMT tax rate to 35 percent, reduce the threshold at which the higher tax rate takes effect to $65,000 from $175,000 for married couples, and index the threshold for
inflation starting after 2010.
By 2010, these adjustments would reduce the number of AMT taxpayers by more than 50 percent relative to current law and redirect the tax toward high-income households. The number of AMT taxpayers with incomes between $50,000 and $75,000 would fall 90 percent, but the share with incomes between $500,000 and $1 million would rise 81 percent. The share with incomes over $1 million would rise 260 percent (not shown in table).
Repealing the AMT and Reforming the Regular Income Tax
To offset the expense and effects of repeal, reformers could incorporate AMT provisions that are deemed good tax policy into the regular income tax. For example, if deferral preferences were seen as necessary to deter tax shelters, they could be incorporated into the regular income
tax. If reformers conclude that deductions for state and local taxes are unwarranted, they could eliminate these itemized deductions for all taxpayers, not just for those paying the AMT. Such a reform package could make the regular income tax simpler (by reducing the number of allowable deductions). Still, unless Congress raised tax rates across the board, this reform package could not completely pay for the elimination of the AMT.
The AMT will soon affect 1 in 3 American taxpayers. Lack of inflation indexing creates automatic annual AMT tax increases. Meanwhile, the phase-in of the 2001 tax cuts will gradually reduce regular income tax burdens. Squeezed on both sides, more and more taxpayers face a problematic tax most of them were never meant to pay.
AMT reform is a question of when and hownot if. One reform approach would be to eliminate the AMT's exemption preferences while keeping its deferral preferences. This strategy would remove the AMT for almost all taxpayers and would hold down the possibility of high-income
tax filers paying no tax, but it would also be both expensive and regressive. Outright repeal, an alternative approach, would be even more expensive and regressive. Moreover, without other changes, repeal would allow many high-income tax filers to avoid paying any income tax. Alternatively, policymakers could retarget the AMT to the highest income groups and use the revenue generated to eliminate the tax for most middle-income taxpayers.
Absent political constraints, the best approach would couple AMT repeal, or a significant rollback, with changing the regular income tax and the estate tax in ways that preserve the goals of the AMT without imposing the costs. This kind of reform could improve equity and efficiency, simplify taxes without inviting more shelters, and help maintain revenues and progressivity.
To date, however, neither political party has been willing to shoulder responsibility for addressing the problem. But fixing the AMT should hold appeal across the ideological spectrum. Those who believe higherincome taxpayers should pay a lower share of the tax burden than they currently do will find that AMT growth undermines Congress's ability to enact high-income tax cuts. Those who favor progressive taxation will see that the AMT's future growth will fuel taxpayers' dissatisfaction with the current systemperhaps enough to spur passage of the flat tax or other radical tax reforms. Under a flat tax, individuals would pay no direct tax on their capital income. Then, millions of wealthy citizensnot just 155would pay no individual taxes.
1. The AMT generally preserves the lower tax rates on capital gains in the regular tax. Current law limits tax rates on long-term capital gains to 10 percent for low- and
moderate-income taxpayers and 20 percent for others. Those limits apply to both the
regular income tax and the AMT.
2. Under either the regular income tax or the AMT, the Bradys are entitled to a child tax credit of $3,600 ($600 per child). The credit reduces their tax liability but does not affect the difference in tax burdens under the AMT and the regular tax.
3. In the example, however, the Bradys will pay lower taxes than before EGTRRA,
because the increased child tax credit is available against the AMT as well as the regular tax. Other tax credits, such as those for educational expenses, are only allowable
against the AMT through 2003.
4. In table 2, the "effect on budget" is based on the assumption that the option is financed by government borrowing. It thus includes the interest payments the government would have to make on the additional debt.
5. See Burman, Maag, and Rohaly (2002).
Burman, Leonard E., Elaine Maag, and Jeff Rohaly. 2002. "EGTRRA: Which Provisions Spell the Most Relief?" Tax Policy Issues and Options No. 3. Washington, D.C.: Urban-Brookings Tax Policy Center.
Burman, Leonard E., William G. Gale, Jeffrey Rohaly, and Benjamin H. Harris. 2002. "The
Individual AMT: Problems and Potential Solutions." Tax Policy Center Discussion Paper No. 5. Washington, D.C.: Urban-Brookings.
About the Authors
Leonard E. Burman is a senior fellow in the Income and Benefits Policy Center at the Urban Institute and codirector of the Tax Policy Center.
William G. Gale is a senior fellow and Arjay and Frances Fearing Miller Chair at the Brookings Institution and codirector of the Tax Policy Center.
Jeffrey Rohaly is a research associate in the Income and Benefits Policy Center at the Urban Institute and director of tax modeling for the Tax Policy Center.
Benjamin H. Harris is a senior research assistant in the Economic Studies program of the Brookings Institution.
About the Series
The Tax Policy Center (TPC) aims to clarify and analyze the nation's tax policy choices by providing timely and accessible facts, analyses, and commentary to policymakers, journalists, citizens, and researchers. TPC's nationally recognized experts in tax, budget, and social policy
carry out an integrated program of research and communication on four overarching issues:
fair, simple, and efficient taxation; long-term implications of tax policy choices; social policy in the tax code; and state tax issues.
A joint venture of the Urban Institute and the Brookings Institution, the TPC receives support from a generous consortium of funders, including the Ford Foundation, the Annie E. Casey
Foundation, the George Gund Foundation, and the Charles Stewart Mott Foundation.
The views expressed do not necessarily reflect those of the Urban Institute, the Brookings
Institution, their boards of trustees, or their funders.
This brief is an abridged version of "The Individual AMT: Problems and Potential Solutions" (Tax Policy Center Discussion Paper No. 5), which contains citations and documentation of the results in this brief.
The authors thank Peter Orszag, Gene Steuerle, and Jerry Tempalski for helpful discussions, Deborah Kobes for expert research assistance, and the Ford, Casey, Mott, and Gund Foundations for research funding through grants made to the Tax Policy Center.