Journalof Economic Perspectives—Volume 17,Number 2—Spring 2003—Pages 173– 186

PolicyWatch The ExpandingReach ofthe Individual Alternative MinimumTax

Leonard E.Burman,William G. Gale and JeffreyRohaly

This featurecontains short articleson topics that arecurrently on theagendas ofpolicymakers, thus illustratingthe role of economic analysis inilluminating currentdebates. Suggestionsfor future columns and commentson past ones should besent toC. EugeneSteuerle, c/ o Journalof Economic Perspectives, The Urban Institute, 2100M StreetNW, Washington, D.C. 20037.

Introduction

In January 1969,Treasury Secretary Joseph W.Barrinformed Congress that 155individual taxpayers with incomes exceeding $200,000 had paid no federal incometax in 1966. The news created a politicalŽ restorm.In 1969,members of Congressreceived more constituent lettersabout the155 taxpayers than about the VietnamWar. Laterthat year,Congress created a minimumtax to preventwealthy individualsfrom taking advantage oftaxlaws to eliminatetheir federal income tax liability. Both theoriginal minimum tax and itssuccessor, theindividual alternative minimumtax (AMT), have appliedin thepast toa smallminority of high-income y LeonardE. Burman isSeniorFellow, UrbanInstitute, Co-director, Tax Policy Center,and ResearchProfessor, Georgetown University Public Policy Institute,all inWashington, D.C. WilliamG. Gale isArjay andFrances FearingMiller Chair,Brookings Institution, and Co-director,Tax Policy Center,both in Washington, D.C. JeffreyRohaly is a Research Associate,Urban Institute, and Director of Tax Modeling at the Tax Policy Center,both in Washington,D.C. Theauthors may becontacted at [email protected] . 174Journal of Economic Perspectives

households. 1 Butbarring a change inlaw,this “class tax” willsoon bea “mass tax.” Currentprojections show thenumber of AMT taxpayersskyrocketing from one millionin 1999to 36 millionin 2010. 2 Without reform,virtually all upper middle- class familieswith two or more children will be paying the AMT bydecade ’s end. TheAMT isnotoriouslycomplex, and itsrecord on fairnessand ef Ž ciencyis mixed at best. Butbecause ofits widening reach, Ž xingthe AMT willbe expensive. By the end ofthe decade, repealingthe AMT willcost morethan repealingthe regular incometax. This paperexplains how ataxoriginally designed to target 155 taxpayerscould growto cover36 million, discusses economicissues related to the alternativeminimum tax and examinesoptions forreform. 3

How theAlternative Minimum Tax Works

Taxpayerswho maybe subject tothealternative minimum tax must calculate theirtax liability twice: once underregular income tax rules and again underAMT rules.If liabilityunder the AMT proveshigher, taxpayers pay thedifference as a surchargeto the regular tax. Technically,the difference paid istheir AMT. To calculatethe alternative minimum tax, taxpayersadd totheir regular taxableincome various items that areparadoxically called “AMTpreferences ” and that fallinto two categories. Exemptionpreferences allowtaxpayers a varietyof deduc- tions, exclusionsor credits in the regular tax, but arenot allowedin the AMT. Theseitems include personal exemptions, the standard deduction and itemized deductions forstate taxes and miscellaneousexpenses. Middle-income taxpayers arethe most likely to be hit by exemptionpreferences, which have littleto do with taxsheltering. As aresult,these adjustments aredif Ž cultto justify. Deferralpreferences allowtaxpayers to postpone regularincome tax payments or shelterincome by hastening deductions ordelayingincome recognition. The AMT ruleslimit the extent to which taxpayerscan usedeferrals by, forexample, allowing lessgenerous depreciation deductions. Comparedwith exemption preferences, deferralpreferences are more complex, have agreatertendency to affect high- income Ž lersand generateless AMT revenue. Once ataxpayeradds inallapplicable preferences and talliesincome, the next stepis tosubtract thealternative minimum tax exemption — currently$49,000 for

1 Aseparatealternative minimum tax appliesto corporations. See Lyon (1997). 2 Throughoutthis paper,we compare current or recentdata toprojections for 2010. We choose 2010 becauseall of the provisionsof the EconomicGrowth and Tax ReliefReconciliation Act of2001 (EGTRRA, the 2001tax cut) arescheduled to expire, under current law, bythe endof 2010. This creates considerableuncertainty aboutprojections after 2010.For example, if EGTRRA wereextended, 43 milliontaxpayers wouldface the AMT in2013. If EGTRRA wereallowed to expire in 2010,26 million taxpayers wouldbe on the AMT in2013. 3 This paperdraws from Burman et al. (2002). Thereader interested in learning more might also start with Graetzand Sunley(1988), Harveyand Tempalski(1997), JointCommittee on Taxation (1970, 2001a), JointEconomic Committee (2001), Karlinsky (1995), Kieferet al. (2002), Rebeleinand Tempalski(2000), Shaviro(2001) and Tempalski(1996). LeonardE. Burman, WilliamG. Gale andJeffrey Rohaly 175

marriedcouples and $35,750for singles. The remaining income is then taxedat  atterrates than underthe regular income tax. Thestatutory AMT taxrate of 26percent applies to the Ž rst$175,000 of net income above the exemption. For incomeover that level,a 28percent tax rate applies. Under the regular income tax (in2002), the same income would be taxed at ratesranging from 10 percent to 38.6percent. Many taxpayers ’ effectiveAMT rate,however, is signi Ž cantly higher, becausethe AMT exemptionitself phases out at a25percent rate over higher incomeranges. Thus, effectivemarginal tax rates can beas high as 35percent (1.25 times28 percent). The AMT exemptionsand taxbrackets are not indexedfor in ation.4

Class Tax to Mass Tax

Undercurrent law, about 36million people will be paying the alternative minimumtax by 2010, almost 14 times as many as in2002, as shown inTable 1. 5 Theincrease in coverage will occur inall but thevery lowest income classes. In 2002,for example, 3 percentof Ž lerswith income between $75,000 and $100,000 (in2001 dollars) faced theAMT; by2010, 79 percent of Ž lersin that incomerange willpay theAMT. TheAMT willbecome the de facto taxsystem for Ž lers in the incomerange from $100,000 to $500,000, 95 percent of whom will face the tax in 2010.At very high incomelevels, the share oftaxpayers on theAMT falls,because thetop AMTrateis lowerthan thetop marginaltax rate in theregular income tax. As aresult,as incomerises, eventually regular income tax liability overtakes AMT liability.Even so, in2010 more than one-quarterof tax Ž lerswith incomes above $1million will pay theAMT, up from15 percent in 2002. Becausethe alternative minimum tax does not allowexemptions for depen- dents ordeductions forstate taxes, it will impose particularly high burdens on taxpayerswith children and those inhigh-tax states. Becausethe AMT exemption forcouples is less than doublethe exemption for singles and becausethe tax bracketsare not adjusted formarital status, theAMT imposessigni Ž cant marriage penalties.In combination, theseissues can raiseAMT participation rates dramatically,

4 Thealternative minimum tax generallypreserves the lowertax rates oncapital gains inthe regulartax. Currentlaw limits tax rates onlong-term capital gains to10 percent for low- and moderate-income taxpayers and 20percentfor others. Thoselimits apply to both the regularincome tax and the AMT. 5 Unlessotherwise noted, all of the projectionsin this paperderive from the Tax PolicyCenter MicrosimulationModel. The model is based on data fromthe 1996public-use Ž leproduced by the Statistics ofIncomeDivision of the Internal RevenueService. The Ž lecontains about112,000 records with detailedinformation from federal individual income tax returns Ž ledin the 1996calendar year. Imputations basedon data fromother sources such as the CurrentPopulation Survey supplement the tax data. Thetax modelhas twocomponents: a statistical routinethat usesforecasts from the Congres- sionalBudget Of Ž ce to “age” orextrapolatethe 1996data tocreate representative samples of tax Ž lers forfuture years and adetailedtax calculatorthat computesthe regularincome tax and AMT liabilityfor all Ž lersin the sampleunder current law and underalternative policy proposals. For additional information,see Burman et al. (2002, appendix). 176Journal of Economic Perspectives

Table 1 AggregateAlternative Minimum Tax Projections

AMTParticipation

Pre-EGTRRA Current Law Law Percentageof EGTRRA IncomeTax Cut Taken 2002 2010 2010 Back by AMT,2010

AMT Taxpayers a Number(in millions) 2.6 35.6 17.9 — As percentageof alltaxpayers b 2.7 33.0 16.1 — As percentageof alltax Ž lers 1.9 24.2 12.1 36.3 As percentageof Ž lers, by AGI (thousands of2001$) 0–30 * 0.2 0.2 * 30–50 0.2 8.7 6.9 1.0 50–75 1.4 43.2 25.6 17.6 75–100 3.0 78.6 34.6 42.3 100–200 10.9 94.0 40.2 71.2 200–500 35.6 96.7 53.2 73.8 500–1,000 19.4 54.1 13.2 18.8 1,0001 15.4 26.9 12.3 8.4 AMT Revenue Dollars(billions) 13.0 141.4 47.0 — As percentageof incometax revenue1.4 9.9 3.0 — Percentageof AGIonAMT returns 8.9 55.5 26.4 — Costof incometax repeal($ billions)204.0 47.0 211.6 —

*Lessthan 0.05percent. Source: Urban-BrookingsTax PolicyCenter Microsimulation Model. a AMT taxpayers includethose with AMTliabilityon Form6251 and thosewith lostcredits. b Taxpayersare de Ž nedas returnswith positiveincome tax liabilitynet ofrefundablecredits. as spelledout inTable 2. By2010, among marriedcouples with two or more childrenand incomebetween $75,000 and $100,000,97 percentwill face the AMT. Moregenerally, the expansion ofthe AMT impliesthat by2010 more than 45percent of tax Ž lerswill be unaffected bymarginal tax rate changes inthe regularincome tax — half ofthose becausethey are on theAMT and half because theirincomes are too lowto owe regular income tax. Two-thirdsof married couples will Ž nd themselvesin that situation —86percent of them because ofthe AMT. Perhaps themost striking illustration of thegrowing scope oftheAMT isthat by2010, more than half ofall income will be taxedunder the AMT, and repealing theregular income tax would reduce tax revenues by less than repealingthe AMT. In 2010,total income tax revenues are projected to be $1,423 billion; that is, $1,282billion from the regular income tax and $141billion from the AMT (de Ž ned as theamount owedabove and beyond regulartax liability). AMT repealwould thus reducerevenues by $141billion in 2010.If theregular income tax were repealed, AMTrevenueswould increase dramatically to $1,376 billion, so thenet revenue loss TheExpanding Reach of the Individual Alternative Minimum Tax177

Table 2 AlternativeMinimum Tax Projectionsby Individual Characteristics

AMTParticipation

Current Law Pre-EGTRRALaw

2002 2010 2010

Percentageof Ž lerson AMT by: a Numberof children b 0 1.1 15.8 3.4 1 1.7 31.5 15.4 2 3.4 46.1 36.3 3 or more 8.7 53.5 53.4 State tax level c Low 0.8 20.6 8.9 Middle 1.1 25.1 12.0 High 2.9 25.2 15.2 Filingstatus Single 0.5 2.6 1.1 Married Ž ling joint 3.7 57.5 27.7 Headof household 1.3 10.1 8.4

Married Ž lerswith incomebetween $75,000 and $100,000 d

0 Kids 1.0 81.9 5.4 1 Kid 1.6 96.1 29.3 21 Kids 4.6 97.4 86.8

Source: Urban-BrookingsTax PolicyCenter Microsimulation Model. a Includesthose with directAMT liabilityon Form6251 and thosewith lostcredits. b Numberof children is de Ž nedas numberof exemptionstaken forchildren living at home. c State codesare not providedon the SOIpublic-use Ž lefor individuals with 1996AGI above $200,000. Figureshere include only those Ž lersfor which we have state-of-residenceinformation. d Incomerefers to AGI in 2001dollars. wouldbe only$47 billion ( 5 1,423 2 1,376).The number of AMTtaxpayerswould doubleto 71 million, as evenmoderate-income taxpayers would face AMT liability. Determiningthe causes ofthe expanding reach ofthe AMT istricky. For example,the AMT mighthave beenrepealed in 1979, but itwasn ’t. Thus, insome sense, failureto repealthe AMT in1979could beviewed as thecause ofprojected AMTgrowth.Despite this underlyingambiguity, we believe it makes the most sense tofocus ontwofactors: thelack of in  ation indexingin theAMT and the2001 tax cut. Asageneralrule, most major taxlegislation since 1980 has includedchanges intheAMT that broadlyconform to the reforms made in theregular income tax: forexample, the 1986 Tax Reform Act broadened the base ofboth taxes,and the 1993tax legislation raised marginal rates under both taxes. Thereare two major exceptionsto the general rule, though. TheEconomic Recovery Tax Act of 1981 cut taxesand indexedthe regular tax system for in  ation, but didnot indexthe AMT.Morerecently, the Economic Growth and TaxRelief Reconciliation Act of 178Journal of Economic Perspectives

2001(EGTRRA) cut theregular income tax, but didnot makesigni Ž cant, lasting changes tothe AMT. 6 Our simulationsshow that iftheAMT had beenindexed for in ation along withthe regular income tax in 1985,and ifEGTRRAhad not been enacted in2001, the number of AMT taxpayerswould have remainedroughly constant at 300,000through 2010,rather than risingto 36 million. The2001 tax act roughlydoubled the projected number of those who facethe alternativeminimum tax in 2010,as shown inthethird column ofTable1. Before this legislation,16 percent of taxpayers were slated to pay theAMT in2010 (with therise over time due primarily to the lack of in  ation indexing);after the 2001 legislation,33 percent will pay AMT.In addition, by2010,EGTRRA will more than doublethe share ofadjusted grossincome (AGI) subject tothe AMT, from 26percent to 55percent, and thelaw will triple the cost ofeliminating the AMT, from$47 billion to $141 billion. Ironically,just as the2001 tax cut madethe AMT problemworse, the AMT will undo someof theeffects of the 2001 act. By2010,the AMT will “take back” about 36percentof the overall income tax cut enacted through EGTRRA,including more than 70percentof thecut targetedto taxpayers with income between $100,000 and $500,000,as shown inTable 1. Theseeffects were understood at thetime; indeed, somehave claimedthat theAMT was leftunadjusted preciselyin orderto reduce artiŽ ciallythe revenue costs and mitigatethe reported distributional effects of the taxcut. Galeand Potter(2002) discuss theseissues further.

Equity

Thealternative minimum tax was originallymotivated by a minimalistnotion ofvertical equity —that high-incomepeople should pay at leastsome income tax each year.The logic of such agoalis questionable on purelyeconomic grounds, but itcommands substantial publicsupport, as the1969 letter-writing campaign suggests. Thealternative minimum tax has succeededin holdingdown thenumber of high-incometax Ž lerswho pay no federalincome tax. Weestimate that in2001, roughly100 tax Ž lerswith incomes above $1 million paid no federalincome tax, but at least700 high-income tax Ž lersowed no incometax before the AMT. 7 If the existenceof the AMT also discouragestaxpayers from attempting to shelter

6 Lindsey(2001) blamesthe 1993tax measuresfor much of the growthin AMT participation,but our estimatessuggest that thoseeffects weremuch smaller than thoseattributable to EGTRRA, inlargepart becausethe 1993act raisedthe highest incometax rates, whereasEGTRRA reducedmarginal tax rates. 7 Ourestimates are based on data from1996 extrapolated to 2001 income levels (Burman etal., 2002, appendix).Balkovic (2002) reportsthat the numberof nontaxpaying Ž lerswith incomeover $200,000 increasedby 54 percent between 1996 and 1999.The corresponding increase for those with real incomesover $200,000, measured in 1976 dollars, was 48percent.Thus, ourestimates may understate the numberof high-income nontaxpayers, butthe qualitativeconclusion that smallnumbers are involvedis likelyto be robust. LeonardE. Burman, WilliamG. Gale andJeffrey Rohaly 179

income,the number paying no incometaxes without an AMTcould have been much higher.Nevertheless, it is unclearwhy a fewmillion people need to pay the taxcurrently in order to stop severalhundred ora fewthousand frompaying no tax. Moreover,although theAMT ismore progressive than theincome tax, both theregular income tax and theAMT willbecome less progressive over time. The progressivityof the regular tax will decline because the2001 tax cuts increasingly beneŽ thigher-incometaxpayers over the course of thedecade (Burman, Maag and Rohaly, 2002;Gale and Potter,2002). The progressivity of theAMT willalso decline as itcomes to affect millions of middle-class families. Filers with income under $100,000(in 2001 dollars) will account for53 percentof AMT taxpayersin 2010, up from24 percent in 2002. Those Ž lerswill account for24 percentof AMT revenues, comparedwith 8 percentin 2002. Only 9percentof AMT revenueswill come from taxpayerswith incomes above $500,000 in 2010, compared with 33 percent in 2002. That incomegroup will account for30 percentof income tax revenues in 2002 and 26percentin 2010. Thus, theAMT ’sabilityto boost theprogressivity of the income taxwill erode in the future. Thealternative minimum tax also raiseshorizontal equity issues. Byreiningin taxshelters, the AMT signi Ž cantly reducesthe variance of average effective tax rates among taxpayerswith similar incomes. Our calculations show that theAMT reduces thevariance by 17 percent for taxpayers with incomes between $200,000 and $500,000in 2002. By 2010, the AMT willreduce the variance of effective tax rates byover 30 percent for such taxpayersand byalmost 20 percent for those earning between$50,000 and $75,000.However, a fullmeasure of horizontal equity must adjust fordifferences in ability to pay taxcreated by factors otherthan income; speciŽ cally,it might include adjustments forfactors likecharitable contributions or extraordinarymedical expenses that arenow writteninto the regular tax code. The AMTallowssome of these adjustments, such as deductions forcharitable contri- butions and casualty losses, but disallowsothers, such as childexemptions and deductions forcertain medical expenses. It also signi Ž cantly increasesmarriage penalties.Thus, ajudgment on how theAMT affectshorizontal equity will neces- sarilyinvolve considering which elementsof thecurrent tax code are necessary to re ectability to pay.

EfŽ ciency

Themost plausible economic rationale for a minimumtax of some sort is that itcould bea second-best backstop fora porous incometax. Byreining in unwar- rantedtax shelters that lawmakersfor some reason could not address directly,the taxmight reduce distortions and limittax sheltering. For example, by taxing interestincome from bonds that stateand localgovernments issue to support privateactivities like shopping centersor stadiums, incomethat isexemptfrom the regularincome tax, theAMT reducesthe subsidy affordedsuch investments 180Journal of Economic Perspectives

(Leonard, 1998).Under certain assumptions, this could makethe tax system more efŽ cient. Although thenotion ofthe AMT as abase-broadening, rate-loweringtax was plausiblein thepast, itisnot today. In theearly years of the alternative minimum tax, shelterswere booming. Shelters served to reduce or eliminate taxes for many high-income Ž lersand typicallyworked by combining assets that generatedcapital gains and expensesthat weredeductible. Taxon capitalgains could bedeferred for years and faced alowstatutory rate when recognized.Deductions included highly accelerated depreciation, generous oildepletion allowances and interestpayments that largelyrepresented in  ation ratherthan thereal cost offunds (Graetz,1997). Thus, an investmentthat would losemoney before tax — because theincome including capital gains was lessthan the expense— could be proŽ tableafter-tax because expenseswere overstated for taxpurposes and capitalgains wereonly partially taxed. TheAMT likelylimited those sheltersand arguablyimproved economic neutrality by reducing the gener- osityof the deductions and taxingcapital gains at thesame rate as otherincome. Priorto 1985, about 85percent of AMT preferencesrelated to capital gains. However,the alternative minimum tax no longerfocuses mainlyon taxshel- ters.A much largershare ofits revenue now comesfrom run-of-the-mill provisions likethe disallowance of personal exemptions and standard deductions. TheTax ReformAct of1986 combined with the near-elimination of in  ation sharply curtailedtax shelter activity (Samwick, 1995). Because the 1986 tax reform taxed capitalgains at thesame rate as ordinaryincome, capital gains wereeliminated as an AMTpreferenceitem. When thedifferential between tax rates on capitalgains and ordinaryincome was reestablishedin 1990and expanded in1997,the role of capitalgains insheltering income rose, but capitalgains werenot reinstatedas an AMTpreferenceitem. Thus, thepreferential treatment of capital gains, thelinch- pin ofmany shelteringschemes, isnot addressed at allin the AMT. Finally,one of the enduring bits of conventional wisdomabout thealternative minimumtax is that, whateverits other faults, ittaxesa broaderbase ofincome at lowermarginal rates than theregular income tax. Thefacts arealmost exactly reversed;that is, theAMT oftenresults in lessincome subject totax but at higher marginalrates than underthe regular income tax. Forexample, a coupleearning $75,000with six children would have $43,150of taxableincome under the regular taxin 2002, assuming that theytook the standard deduction. (Forthis family, taxableincome would equal $75,000 minus $24,000in personalexemptions — eight times$3,000 per person —minus astandard deduction of$7,850, which equals $43,150.)Neither the personal exemptions nor standard deduction wouldbe allowedagainst theAMT, but thecouple would be entitled to an AMTexemption of$49,000, yielding income subject toAMT of$26,000 —lessthan taxableincome underthe regular tax. Theywould nevertheless owe AMT because theirmarginal taxrate under the AMT —26 percent—ismuch higherthan theirregular income taxbracket of 15 percent. Overtime, more and moretaxpayers will Ž nd themselvesin asimilarposition. TheExpanding Reach of the Individual Alternative Minimum Tax181

Theshare ofAMT taxpayerswith less income taxed in theAMT than intheregular incometax is projectedto rise from 66 percent in 2002to 87 percentin 2010.The share withhigher marginal tax rates under the AMT than underthe regular tax will risefrom 35 percent in 2002 to more than 90percent in 2010.

Complexity

TheNational Taxpayer Advocate (2001) and theInternal Revenue Service (2000)have calledthe alternative minimum tax one of the most dif Ž cult and complexareas of tax law. Many taxpayersmust keeptwo separate sets of books because ofthe deferral preferences —theAMT ruleson thetiming of income recognitionand deductions that differfrom regular income tax rules. These rules reducethe number of high-income tax Ž lersthat pay no incometax and thus serve an identiŽ ablegoal. The same goal could beadvanced much moresimply, however, byscaling back deferralpreferences in the regular tax, ratherthan requiring taxpayersto juggle two separate, complicated calculations. Much oftherest of AMT complexityappears tobe completely pointless. Most peoplewho must currently Ž llout theAMT formsend upowingno additional tax. Increasingly,the tax will impose greater compliance burdens on middle-class taxpayers,a groupthat was neverthe tax ’smaintarget. Moreover, the complexity also makespredicting marginal tax rates and understanding taxrules much more difŽ cult.

Options for Reform

Theunderlying goals of the AMT —requiringhigh-income people to pay some tax, deterringthe aggressive use of tax shelters and ensuringprogressivity — have widespreadpopular appeal, but thetax itself is replete with problems. A variety ofreform options could, tovarying degrees, keep the baby but throwout the bathwater.

Reducingthe Alternative Minimum Tax Simplyindexing the alternative minimum tax for in  ation wouldreduce the numberof AMT taxpayersin 2010 by 71 percent overall and by92 percent for middle-classtaxpayers, de Ž ned as those withAGI between$50,000 and $75,000. Indexingwould reduce revenues by $418 billion through 2013under current law (Table 3). Themiddle class could bealmost entirely removed from the AMT byalso allowingdependent personal exemptions and nonrefundable credits,such as the taxcredits for child care and education. In conjunction withindexing, these reformswould reduce the number of AMT taxpayersin 2010by 83percentoverall 182Journal of Economic Perspectives

Table 3 ReformOptions a

Percentagechange in AMT AMT taxpayers,by AGI(2001$) b Change in High-income taxpayers, revenue, Žlers owing 2010 500K–1 2004–2013 No tax, (millions) All 50K–75K million ($billions) b 2010 c

Indexthe AMT after 2002 10.4 270.9 292.3 22.8 2418 200 1 Removemiddle-class exemption 6.0 283.1 298.3 27.8 2474 200 preferences d 1 Removeother major exemption 0.3 299.1 299.7 290.3 2650 400 preferences e Repeal 0.0 2100.0 2100.0 2100.0 2700 1500 Revenue-neutralretargeting f 6.7 281.2 297.3 67.2 7 100

Source: Urban-BrookingsTax PolicyCenter Microsimulation Model and authors ’ calculations. a Changein revenuein Ž scalyears; othertabulations in calendar years. b Baselineis current law (EGTRRA expireson December 31, 2010). c High-incomeis de Ž nedas having AGIgreaterthan $1million in 2001dollars. d Indexesand allowsdependent exemptions and personalnonrefundable credits. e Takessteps inthe previousplan and allowsdeductions for expenses and taxes and repealsthe AMT exemptionphaseout. f Allowsdependent exemptions and personalnonrefundable credits. The preferential rates forcapital gains underthe AMT wouldbe repealed, the 28percent AMT ratewould be increasedto 35 percent and the AMTexemptionphaseout would be repealed, all effective for 2003. The sunset ofthe AMT exemptionincrease would be repealedand the exemptionand ratebracket threshold would be indexed beginningin 2005. and morethan 98percent among those withincomes between $50,000 and $75,000. Combiningthese reforms with two additional steps —repealingthe phase out ofthe AMT exemptionand allowingdeductions forstate and localtaxes and miscellaneousexpenses —wouldeliminate all of themajor exemptionpreferences and virtuallyend theAMT forall but veryhigh incometax Ž lers.The number of AMTtaxpayerswould fall by more than 99percent relative to current law. The additional steps primarilybene Ž thigh-incomehouseholds; forexample, they wouldreduce the number of AMT taxpayerswith income between $500,000 and $1million by over 90 percent. The additional measuresare also expensive.The ten-yearrevenue cost wouldbe $650 billion, more than 50percent greater than indexingalone. Fromhere, outright repeal of the AMT isa smallstep, consisting mainlyof the eliminationof the deferral preferences. Relative to the plan above, repealwould cost just $50billion more over the decade and reducethe number of AMTpayers in2010by an additional 0.3million. Repeal of the deferral preferences would be signiŽ cantly regressive,however, with very large tax cuts goingto the highest- incomehouseholds (Burman etal., 2002). Repealingthe deferral preferences would also signi Ž cantly increasethe num- berof high-income Ž lerswho pay no incometax. Thenumber of Ž lerswith income LeonardE. Burman, WilliamG. Gale andJeffrey Rohaly 183

above$1 millionwho wouldpay no incometax would rise from 400 under the plan aboveto 1,500 under repeal; similarly, the number of Ž lerswith income above $200,000who pay no incometax would rise from 6,000 under the plan aboveto 17,000with repeal. These Ž guresshow thepower of the deferral preferences in reducingthe number of high-income Ž lerswho pay no tax. Evenmore nontaxpay- erswould exist (and therevenue costs wouldbe larger) if AMT repealunleashed arash ofnew tax shelters.

RevenueNeutral Reform Allof the plans noted abovewould signi Ž cantly reducerevenues. 8 In the currentbudgetary environment, such changes maybe neither feasible nor desir- able(Auerbach etal., 2003).AMT reformcould be Ž nanced byretargeting the tax orby coupling AMT repealwith income tax changes. One wayto retargetthe AMT at veryhigh incometaxpayers and aggressivetax sheltererswould be to allow dependent exemptions and personalnonrefundable taxcredits, eliminate the AMT exemptionphase out and indexthe exemption from its2004 level starting in 2005. These reforms could bepaid forby increasing the 28percentAMT bracketto 35 percent (which wouldincrease taxes only for those withincomes above the AMT exemptionphase out, an incomelevel of $330,000 for couplesafter 2004) and eliminatingthe preferential rates for capital gains under the AMT. As shown inTable 3, undercurrent law, this proposal wouldraise about $7billionover the next ten years. The proposal wouldbe highlyprogressive, cutting overalltaxes on those withincomes under $500,000 and raisingtaxes on higher income Ž lers.It wouldreduce the number of AMT taxpayersin 2010 by 81 per- cent—and by97 percentfor those withincomes between $50,000 and $75,000.But itwouldincrease the number of AMT taxpayersamong those withincomes between $500,000and $1million by two-thirds. Although not shown inthe table, it would morethan triplethe number of taxpayers with incomes over $1 millionsubject to the AMT. If thereis to be an alternativeminimum tax, this option has much torecom- mendit. Allowing preferential capital gains taxrates under the AMT isa major sourceof complexity and sheltering.Many individualtax shelters are designed to exploitthe difference between the tax rates on capitalgains and thetax rates on otherincome and expense —mostnotably thehigher effective tax rate on interest expense(Burman, 1999).The deduction forinvestment interest is the most im- portant factorexplaining the nontaxation ofhigh-income returns (Balkovic, 2002). Thus, taxingcapital gains thesame as otherincome under the AMT could arguably enhance efŽ ciency,equity and simplicityand raiserevenue that could beused to reducethe number of AMT taxpayers.

8 If the 2001tax cut weremade permanent, the revenuelosses due to AMT reformwould rise. Indexing and repealwould reduce revenues by about $600 billion and $1trillion, respectively, through 2013. Theseestimates omit the addeddebt service costs that the governmentwould owe if revenuefell. 184Journal of Economic Perspectives

Instead, iflegislators could Ž nd therevenue and resolveto Ž nance AMT repeal,the best option wouldbe to incorporate directly into the regular income tax whateverAMT provisionsare deemed good taxpolicy, while adjusting therates and taxin the regular income tax to achieve revenue neutrality and distributional neutrality.For example, if the regular income tax deductions forstate and local taxesand depreciationrules are too generous, reformers could eliminatethese provisionsfor all taxpayers, not just forthose payingthe AMT. Inevitably,many taxpayerswould face higher marginal and averagetax rates under a revenue- neutralpackage. 9 Butreplacing the hodgepodge of implicit taxes created by the AMTwithwell-designed explicit taxes under the regular income tax could makethe taxsystem fairer, simpler and moreef Ž cientwithout spawning sheltersor sacri Ž c- ingtax revenues. Alternatively,a natural sourcefor revenue to pay forreform is the2001 tax cut, which not onlyreduced revenues, but exacerbatedthe AMT problem.The 2001 tax cut phases inslowlyover time. Burman etal. (2002)show that freezingthe income and estatetax cuts at their2002 levels, and foregoingthe future cuts, wouldroughly pay forindexing the AMT, but wouldfall more than $200billion short ofpaying for repeal.A partialfreeze could also becombined with AMT retargeting,along the linesnoted above.

WhyNot Repeal the Regular Income Tax? By2010, it would cost moreto repeal the alternative minimum tax than to repealthe regular income tax. Somecommentators have suggested,with varying degreesof seriousness, that theregular tax be repealedand theAMT kepton. We believethis wouldbe a major mistake. If thealternative minimum tax were to become the only income tax, itis unlikelythat itcould (orshould) remainin its current structure. The tax is not indexedfor in  ation. It islacedwith marriage penalties and childpenalties. A tax based on ability-to-payshould have an adjustment forfamily size, which suggests that personalexemptions should beallowed. Those who favora  at incometax should beclear that thealternative mini- mumtax is not ashortcut tothat goal.After all, the AMT includesall the aspects ofthe regular income tax that arenot explicitlyerased by the adjustments under theAMT. Thealternative minimum tax includes four different tax brackets (in- cludingthe effect of the phase out oftheexemption) at ratesof 26,31.25, 35 and 28percent. It wouldcertainly not besimple.

9 It wouldbe possibleto mimic the distributionof averagetax burdensunder current law even with AMT repeal.If taxpayers ineach bracket were to make the sameaggregate tax paymentin 2010 after the eliminationof the alternativeminimum tax as they wouldunder current law, the 10percent tax rate wouldhave tobe raisedto 10.7 percent, the 15percenttax rateto 16.3 percent, the 25percenttax rate to29.2 percent, the 28percent tax rateto 33.8 percent, the 33percent rate to 43.0 percent and the 35percent tax ratewould be lowered to33.2 percent. That is, the statutory regularincome tax ratewould have torise for those in the bottom Ž vebrackets and fall inthe highest incomegroup. TheExpanding Reach of the Individual Alternative Minimum Tax185

Conclusion

Lack of in ation indexingin thealternative minimum tax expands thereach ofthe tax each year.Meanwhile, the phase inof the 2001 tax cuts willsteadily reduceregular income tax burdens overtime. Caught amidthese trends, onein threeAmerican taxpayers will soon besqueezed by a problematictax that almost none ofthem were ever meant to pay. To date, neitherpolitical party has been willingto shoulderthe responsibility for addressing the problem. But as thereach ofthe alternative minimum tax expands toencompass evermore taxpayers, the politicalbene Ž ts ofseeking out asolutionwill expand as well. y Theauthors wish to thank Brad De Long,Peter Orszag, Timothy Taylor and Michael Waldmanfor very helpful comments.

References

Auerbach,Alan J. etal. 2003. “BudgetBlues: Graetz, MichaelJ. 1997. TheDecline (and Fall?) TheFiscal Outlook and Optionsfor Reform, ” in ofthe Income Tax. NewYork: W.W. Norton& Agendafor the Nation. HenryJ. Aaron, James Company. Lindseyand PietroNivola, eds. Washington, Graetz, MichaelJ. andEmil Sunley. 1988. D.C.: BrookingsInstitution, forthcoming. “MinimumTaxes and ComprehensiveTax Re- Balkovic,Brian. 2002. “High-IncomeTax Re- form,” in UneasyCompromise: Problems of a Hybrid turns for1999. ” Statisticsof Income Bulletin. Income-ConsumptionTax. HenryJ. Aaron, Harvey Spring,pp. 7 –58. Galperand JosephPechman, eds. Washington, Barr,Joseph W. 1969. “Statement byJoseph D.C.: BrookingsInstitution, pp. 385 – 419. W. Barr,” in Hearingson the 1969 Economic Report Gruber,Jon and Emmanuel Saez. 2000. “The ofthe President before the Joint Economic Committee. Elasticityof Taxable Income: Evidence and CommitteePrint. Washington, D.C.: U.S. Gov- Implications. ” NBERWorking Paper No. 7512, ernmentPrinting Of Ž ce, p. 46. January. Burman,Leonard E. 1999. TheLabyrinth of Harvey,Robert P. andJerry Tempalski. 1997. CapitalGains Tax Policy: AGuidefor the Perplexed . “TheIndividual AMT: Why itMatters. ” National Washington D.C.: BrookingsInstitution Press. TaxJournal. 50:3, 453–73. Burman,Leonard E., ElaineMaag and Jeff InternalRevenue Service. 2000. “Annual Re- Rohaly. 2002. “TheEffect ofthe 2001Tax Cut portfrom the Commissionerof the Internal Rev- onLow-and Middle-IncomeFamilies with Chil- enueService on Tax LawComplexity. ” June 5. dren.” Tax Notes. July8, 96:2, pp. 247 –70. JointCommittee on Taxation. 1970. “General Burman,Leonard E. etal. 2002. “TheIndivid- Explanationof the Tax ReformAct of1969. ” ualAMT: Problemsand PotentialSolutions. ” JCS-16-70,December 3. NationalTax Journal. September, 55:3, pp. 555 – JointCommittee on Taxation. 2001a. “Study 96. Alsoavailable as Urban-BrookingsTax Policy ofthe OverallState ofthe FederalTax System CenterDiscussion Paper No. 5, September. and Recommendationsfor Simpli Ž cation, Pur- Gale,William G. andSamara R.Potter. 2002. suant toSection 8022(3)(B) ofthe FederalTax “An EconomicEvaluation of the Economic System.” JCS-03-01,April. Growthand Tax ReliefReconciliation Act of JointCommittee on Taxation. 2001b. “Esti- 2001.” NationalTax Journal. September,55:1, pp. matedBudget Effects ofthe ConferenceAgree- 133– 86. mentfor H.R. 1836[1]. ” JCX-51-01,May 26. 186Journal of Economic Perspectives

JointEconomic Committee. 2001. “The Alter- the Theoryof Optimal Income Taxation. ” Review native MinimumTax forIndividuals: AGrowing ofEconomic Studies. April,38:114, pp. 175 –208. Burden.” May. NationalTaxpayer Advocate. 2001. “National Karlinsky,Stewart. 1995. “AReporton Re- TaxpayerAdvocate ’sFiscalYear 2001 Annual formingthe AlternativeMinimum Tax System. ” Reportto Congress. ” Mimeo,December. AmericanJournal of Tax Policy. 12:1, pp. 139 –50. Rebelein,Robert and Jerry Tempalski . 2000. Kiefer,Donald et al. 2002. “TheEconomic “Who Pays the IndividualAMT? ” OTA Paper87, Growthand Tax ReconciliationAct of2001: June. Samwick, AndrewA. 1995. “Tax Sheltersand Overviewand Assessment ofEffects onTaxpay- PassiveLosses After the Tax ReformAct of ers.” NationalTax Journal .55:1, pp. 89 –117. 1986.” NBERWorkingPaper No. 5171, July. Leonard,Paul A. 1998. “Tax-InducedSegmen- Shaviro,Daniel. 2001. “Tax SimpliŽ cation and tation inthe Tax ExemptSecurities Market. ” the AlternativeMinimum Tax. ” Tax Notes. May QuarterlyJournal of Business and Economics. 37:4, 28, 91:10, pp. 1455 – 468. pp. 27– 47. Slemrod,Joel. 1990. “OptimalTaxation and Lindsey,Lawrence B. 2001. “GovernorBush ’s OptimalTax Systems. ” Journalof Economic Perspec- Proposaland the AlternativeMinimum Tax. ” tives.4:1, pp. 157 –78. Tax Notes. 86:4, pp. 553 –57. Steuerle,Gene. 2001. “MovingBeyond the Lyon,Andrew B. 1997. Crackingthe Code: Mak- Fight Overthe AlternativeMinimum Tax. ” Tax ingSense of the Corporate Alternative Minimum Tax. Notes. June18, 91:14, pp. 2067 – 068. Washington, D.C.: BrookingsInstitution. Tempalski,Jerry. 1996. “TheIndividual Mini- Mirrlees,James A. 1971. “An Explorationin mum Tax.” Mimeo,U.S. Treasury.