Preannounced tax cuts and their potential influence on the 2001 recession

Transcription

1 Preannounced ax cus and heir poenial influence on he 2001 recession R. Andrew Buers and Marcelo Veraciero Inroducion and summary The 2001 recession differed from previous recessions in several ways. Firs, i was quie mild in erms of is associaed conracions in oupu and consumpion. Also, since oal hours worked fell sharply, labor produciviy remained relaively high. Furhermore, while business fixed invesmen plummeed (acually, much more han in a ypical recession), residenial invesmen and purchases of durable goods remained surprisingly srong. This is highly unusual: Typically, residenial invesmen and purchases of durable goods collapse during recessions, ofen leading he general conracion in economic aciviy by several quarers. Anoher disincive feaure of he 2001 recession was ha i was preceded by a presidenial elecion dominaed by ax cu discussions. The proposals of he wo major candidaes differed in crucial ways. While he Democraic candidae, Al Gore, promised cus ha would leave sauory income ax raes essenially unchanged (mos cus would come in he form of ax credis for paricular economic aciviies), he Republican candidae, George W. Bush, announced a plan ha would significanly reduce income ax raes across all income brackes. Thus, he oucome of he 2000 presidenial elecion promised o have a large impac on he ax raes ha households and businesses would face in he fuure. A basic hypohesis in his aricle is ha he wo ses of facs he unusual feaures of he 2001 recession and he ax cus promised during he 2000 presidenial elecion could be relaed. The raionale for his view is ha people and firms are forward-looking: Expecaions abou he fuure may have a significan effec on he decisions ha hey make oday. For insance, he anicipaion of higher demand may lead a producer o expand his producion capaciy, or he anicipaion of higher wages in a paricular occupaion may induce a worker o acquire specific skills. Anicipaed ax cus are no excepion. If ax raes are expeced o decrease in one year, he curren year becomes a relaively bad year for working and invesing. 1 Forward-looking households and businesses may hus decide o devoe less ime o marke aciviies, cuing back on ime worked in he marke and increasing ime worked a home and subsiuing business invesmen for home invesmen. These conracionary effecs of anicipaed ax cus could have played an imporan role in he paerns of aciviy observed during he 2001 recession. Of course, he anicipaed ax cus were no he only facor poenially influencing economic aciviy during he 2001 recession. A number of oher imporan shocks and policy responses also occurred in 2000 and For sarers, marke paricipans apparenly began o reevaluae he profiabiliy of many invesmen projecs in he high-ech secor. This and oher facors were refleced in a sharp decline in equiy prices saring he spring of In addiion, in 2001 here were he erroris aacks on Sepember 11, followed by he revelaion of he Enron scandal laer ha fall. Moreover, he Federal Reserve lowered is policy rae subsanially over he course of 2001, which influenced coss underlying household and business decisions regarding he purchase of durables and capial goods. In order o deermine he possible effecs of anicipaed ax cus, we consruc and analyze a heoreical model ha absracs from hese oher influences on he economy. The model will hus ell us if he anicipaed ax effecs can plausibly reproduce some of he paerns observed in he daa. However, i is imporan o poin ou ha since he oher facors are excluded, we canno use R. Andrew Buers is an associae economis and Marcelo Veraciero is a senior economis in he Economic Research Deparmen a he Federal Reserve Bank of Chicago. They hank Jeff Campbell, Spencer Krane, Anna Paulson, and Dan Sullivan for heir commens. Federal Reserve Bank of Chicago 17

2 he analysis o rank he relaive imporance of axes and hese oher influences on he economy over his period. The model we use is a version of he Greenwood and Hercowiz (1991) home producion model. In his model, he economy is populaed by a represenaive household ha values consumpion of a marke good and consumpion of a home good. The home good is produced using home capial and ime spen a home. The marke good is produced using business capial and ime worked in he marke. Oupu of he marke good can be consumed, invesed in business capial, invesed in home capial, or consumed by he governmen. The governmen finances is expendiures by axing capial and labor income. Moreover, he governmen is assumed o balance is budge every period. Admiedly, he model is quie simple. However, i capures imporan decision margins, and herefore, we consider i a useful saring poin for he analysis. Selecing model parameers o reproduce salien feaures of he U.S. economy, we find ha, while immediae ax cus generae a boom in economic aciviy, delayed ax cus iniially generae a recession. Our analysis underscores he imporance of aking forwardlooking behavior on he par of households and businesses ino accoun in considering he impac of policy alernaives. In paricular, aking his behavior ino accoun can help us undersand some of he paerns in aciviy observed during he 2001 recession. There are a number of papers ha have previously analyzed he effecs of anicipaed changes o he economic environmen (for example, Jaimovich and Rebelo, 2008, 2009; and Beaudry and Porier, 2007). However, he mos closely relaed is he one by House and Shapiro (2006). Their paper also evaluaes he effecs of he 2001 ax reform. However, hey focus on he effecs of phased-in ax cus from he ime ha he reform was signed ino law, and hey consider a model in which business capial is he only form of capial. Conrary o House and Shapiro (2006), we emphasize he anicipaory effecs of he reform before i was signed ino law, and inroduce home capial ino he analysis. Boh exensions allow us o analyze he sar of he 2001 recession and o evaluae wheher he model is able o generae he unusual srengh in home invesmen ha was observed during ha recession. In he nex secion, we presen he salien observaions from he 2001 recession. Then, we describe he ax reforms ha were promised during he 2000 presidenial campaigns, as well as he ax reform ha was acually implemened. Nex, we explain he model economy. We describe he compeiive equilibrium o be analyzed and how he model s parameers are seleced. Finally, we presen our resuls. The 2001 recession On November 26, 2001, he Naional Bureau of Economic Research (NBER) issued a saemen announcing ha he U.S. economy had reached a peak in business aciviy in March 2001 and had moved ino a recessionary period. 2 The NBER repor cied falling indusrial producion as he mos significan piece of evidence o sugges he economy had slowed. Poor real sales and employmen also provided evidence supporing he decision o announce a recession. While employmen peaked in March, in parallel wih he NBER peak dae, boh indusrial producion and sales had peaked six and seven monhs before ha dae, respecively. The NBER commiee menioned in is saemen ha earlier daes had been considered o reflec he divergen pahs of manufacuring and employmen, bu hese daes were dismissed because of he lower emphasis placed on he manufacuring and goods-producing secors of he economy. 3 On July 17, 2003, he NBER repored ha he economy had reached a rough in November of 2001, ending he recession. 4 The srengh of boh real gross domesic produc (GDP) and real personal income relaive o levels before he recession allowed he NBER commiee o conclude ha any fuure downurn in he economy would in fac be a separae recession and no a coninuaion of he 2001 recession. Neverheless, indusrial producion and employmen showed no sign of recovery. To gain a more deailed undersanding of he 2001 recession, figure 1 repors he pahs of oupu, consumpion, hours worked, business invesmen, and home invesmen leading up o and during he 2001 recession. For comparison, i also repors he average of hose pahs before and during he previous six recessions. 5 For consisency wih he model used laer on, residenial invesmen and personal consumpion expendiures on durables goods are combined ino a single measure denoed home invesmen. In urn, business invesmen is defined as privae nonresidenial fixed invesmen plus changes in invenories, and consumpion is resriced o consumpion of nondurables and services. Because our model economy will be closed, oupu is defined as GDP minus ne expors (ha is, gross domesic purchases). 6 All of hese variables are repored in real erms. For hours worked, we focus on a broad measure consruced by Presco, Ueberfeld, and Cociuba (2009), which includes miliary personnel. Because our model will have no growh componen, we derend each series using a deerminisic rend. 7 A quick glance a figure 1 indicaes ha, relaive o he sandard recession, he 2001 recession was highly aypical in several respecs. Oupu and consumpion (panels A and B) fell during 2001, bu no as much as 18 3Q/2009, Economic Perspecives

3 figure recession versus average recession A. Oupu 101 B. Consumpion quarers since peak C. Hours worked quarers since peak D. Business invesmen quarers since peak quarers since peak E. Home invesmen quarers since peak Average recession 2001 recession Noes: All variables are normalized (ed) o a he recession peaks. Time 0 indicaes he recession peak quarer. The average recession is based on hose ha occurred in 10 61, 19 70, , 10, 11 82, and , according o he Naional Bureau of Economic Research. Source: Auhors calculaions based on daa from he U.S. Bureau of Economic Analysis, Naional Income and Produc Accouns of he Unied Saes, from Haver Analyics. Federal Reserve Bank of Chicago 19

4 in he average recession. 8 Once he 2001 recession sared, hours worked (panel C) behaved similarly o he average recession. However, during he hree quarers leading up o he 2001 recession, here was a seady decline in hours worked compared wih he consan levels leading up o he average recession. For he 2001 recession, he decline in business invesmen (panel D) was much sharper and sared much earlier. Perhaps mos noably, he 2001 downurn had minimal effecs on home invesmen (panel E). On average, home invesmen s decline leads an upcoming recession, while here were no noiceable effecs boh before and during he 2001 recession. Laer on, we will show ha he anicipaion of fuure ax cus could have conribued o some of hese aypical feaures of he 2001 recession (for example, he relaively srong consumpion and home invesmen). In order o do his, we mus firs idenify reasonable esimaes for wo criical elemens of he analysis: 1) when economic agens began anicipaing he fuure ax cus and 2) wha was he paricular ax cu schedule ha economic agens were anicipaing. We will examine he 2000 presidenial campaigns and elecion, as well as he implemenaion of he Economic Growh and Tax Relief Reconciliaion Ac of 2001 (EGTRRA), o deermine hese wo elemens. The presidenial campaigns and he 2000 elecion On March 14, 2000, boh George W. Bush and Al Gore won heir respecive pary s nominaion o become he 43rd Presiden of he Unied Saes. Over he nex eigh monhs, boh candidaes campaigned and presened he American public wih heir own proposals o simulae he economy. 9 The Republican candidae, George W. Bush, ran on a plaform wih across-he-board marginal rae ax cus as one of is foundaions. Leading up o he elecion, he oal cos of he cus was esimaed a $1.3 rillion over he nine-year period In he plan, he 28 percen and 31 percen ax brackes would boh be dropped o 25 percen. The 15 percen bracke would be dropped o 10 percen, and boh he 36 percen and he 39.6 percen ax brackes would be reduced o 33 percen. 10 All of hese cus were proposed o be phased in saring in 2002, wih a furher reducion in 2004, and wih all of he effecs being implemened by Al Gore, he Democraic candidae and incumben Vice Presiden, proposed a ax plan more conservaive in cos (esimaed a $0 billion) and more geared o he low- and middle-income classes. His ax reform included raising he sandard deducion, wih ax breaks and deducions for savings accouns, child care, college uiion, and long-erm caregivers. A ax credi for new reiremen savings accouns was he mos subsanial of hese proposals. Alhough boh candidaes promised significan ax cus, heir proposals differed in a fundamenal way. While he Republican candidae promised reducions in marginal ax raes, he Democraic candidae promised cus in inframarginal axes. This is an imporan disincion, since reducions in marginal ax raes end o increase labor supply, while reducions in lump-sum axes have he opposie effec. These differences made i difficul for economic agens o adjus heir behavior o he prospecive ax cus. The reason is ha adjusing o one ype of reform would have produced large errors had he alernaive reform been implemened. I would be exremely difficul o deermine a each poin in ime he economic agens beliefs abou he likelihood of eiher ype of reform being implemened and, herefore, he fuure axes hey were anicipaing. However, i is convenien for he purposes of his aricle o make a number of no enirely implausible assumpions. Firs, since he elecion was so igh and he Florida recoun acually posponed is oucome for almos a monh afer elecion day (November 7), i seems reasonable o assume ha unil he end of 2000, economic agens were puing a / chance on eiher ype of reform being evenually implemened. Second, since adjusing o each ype of reform required such drasically differen ypes of labor supply responses, i is no implausible o assume ha agens waied unil he elecion oucome before making any changes o heir behavior. Third, we assume ha, once George W. Bush was declared he new Presiden of he Unied Saes, economic agens immediaely shifed heir expecaions abou fuure ax cus o wha had been promised during his campaign. Fourh, we assume ha once a slighly differen reform was laer implemened, economic agens adjused heir expecaions accordingly. In he nex secion, we describe he reform ha was acually implemened. The implemenaion of he 2001 ax cus The Economic Growh and Tax Relief Reconciliaion Ac was signed ino law by George W. Bush on June 7, As proposed in Bush s campaign for he presidency, he law s mos subsanial changes involved an across-he-board reducion in he marginal ax raes. A 0.5 percenage poin cu in he marginal raes for all ax brackes above he 15 percen rae became effecive immediaely. The law sipulaed subsequen cus in 2002, 2004, and 2006 of 0.5 percenage poins, 1 percenage poin, and 1 percenage poin, respecively, for each ax bracke (he only excepion being a cu of 2.6 percenage poins for he highes bracke in 2006). This schedule 20 3Q/2009, Economic Perspecives

5 Table 1 Sauory and effecive marginal ax raes: Proposed and enaced Income ax brackes Effecive marginal ax raes $45,200 o $109,2 o $166,340 o $297,300 Labor ax Capial ax Dae $109,2 $166,340 $297,300 and above rae (τ n ) rae (τ k ) ( percen ) A. Pre-EGTRRA Before 2001:Q B. Campaign proposal 2001:Q1 2001:Q :Q1 2003:Q4 a :Q1 2005:Q4 a :Q1 and beyond C. EGTRRA 2001:Q3 2001:Q :Q1 2003:Q :Q1 2005:Q :Q1 2010:Q :Q1 and beyond a The campaign proposal by George W. Bush did no explicily sae he ax reducion schedule ha would be implemened in 2002 and Noes: EGTRRA means he Economic Growh and Tax Relief Reconciliaion Ac of This able shows he ax raes for he brackes above he 15 percen rae. The schedule for he effecive marginal ax raes on boh labor and capial is aken from House and Shapiro (2006). By using he same weighing procedure used by House and Shapiro (2006), we inerpolae he effecive marginal ax raes if George W. Bush s proposed ax schedule (panel B) had been enaced. Sources: Auhors calculaions based on daa from House and Shapiro (2006) and Congressional Budge Office (2001). would remain in effec unil 2011, when he ax raes would rever, or sunse, o heir pre-egtrra raes. In addiion, he law phased ou he esae ax and sen a ax rebae check of $300 o each individual. An increase in he child credi from $0 o $1,000 and relief from he alernaive minimum ax (AMT) and he marriage penaly rounded ou he bill. Alhough he rebae checks were highly visible in 2001, hey did very lile o affec marginal ax raes (House and Shapiro, 2006). 11 The effecs of he oher provisions on marginal ax raes will be discussed shorly. Relaive o he ax cus ha George W. Bush had proposed during his campaign, EGTRRA differed in several ways. Firs, he iniial cus o marginal raes of EGTRRA became effecive immediaely (in June 2001) and were reroacive o he beginning of he year. In proposals during he campaign, he Bush ax cus were no o ake place unil Nex, as seen in able 1, he ulimae percenage cus for some ax brackes signed ino law were slighly smaller han wha had been proposed o Congress. For he 31 percen and 39.6 percen ax brackes (panel A), proposed cus of 6 percenage poins and 6.6 percenage poins, respecively (panel B), were scaled back o 3 percenage poins and 4.6 percenage poins (panel C). In addiion, an explici sunse dae of January 2011 was pu on all marginal ax rae changes (Brumbaugh e al., 2002). A sudy adminisered by he Congressional Budge Office (2001, p. 34, boxes 2 3), or CBO, esimaed he effecive marginal ax raes (for boh labor and capial income) before and afer EGTRRA. Effecive marginal ax raes depend on oher feaures of ax law beyond he sauory rae, including he earned income ax credi, he child ax credi, and he AMT, among ohers. The analysis by he CBO aemped o ake hese oher provisions ino accoun when deermining he esimaed change due o EGTRRA. 12 The CBO s esimaes of effecive marginal ax raes of labor and capial income are repored in able 1 (he las wo columns of daa). We see ha, according o he CBO, he effecive marginal ax rae on labor fell from a pre-egtrra level of percen (panel A) o percen (fourh row of panel C) and ha he effecive marginal ax rae on capial fell from percen o percen, once EGTRRA was fully phased in (in 2006). 13 A more formal discussion on he axaion of capial can be found in he appendix. The effecive marginal ax raes esimaed by he CBO for he years 2000 and 2006 can be inerpolaed o all oher years by using he corresponding ax raes for he differen income brackes (he firs hrough fourh columns of panel C in able 1). 14 The resuls are shown in he las wo columns of panel C. This procedure can also be used o consruc he implici effecive marginal Federal Reserve Bank of Chicago 21

6 ax raes ha were proposed by George W. Bush during he presidenial campaign. These ax raes are repored in he las wo columns of panel B. The effecive marginal ax raes in panels B and C are he ones ha will be used o deermine economic agens expecaions a each poin in ime. In paricular, we will assume ha saring in 2001:Q1 agens were anicipaing he effecive marginal ax raes provided in panel B, bu ha in 2001:Q3 (ha is, afer he passage of EGTRRA) hey swiched heir expecaions o he effecive marginal ax raes lised in panel C. These expecaions will play an imporan role in he model economy o be described nex. The model The model economy consiss of hree secors: a household secor, a firms secor, and a governmen secor. The household secor is composed of a large number of idenical individuals ha supply labor and business capial o he firms. In addiion, hey produce a home good, using ime and home capial. The firms secor is consiued by a large number of idenical firms ha produce he marke good, using business capial and labor. The marke good is sold o he households, which use i for consumpion and invesmen. The governmen needs o purchase a cerain amoun of he marke good every period. These expendiures are financed wih a combinaion of capial income, labor income, and lump-sum axes. 15 In wha follows we describe he model in deail. The household secor The represenaive household has preferences described by he following uiliy funcion: 1) β ψ c ψ h ln + ( 1 ) ln, = 0 where c is consumpion of a marke-produced good, h is consumpion of a home-produced good, 0 < β < 1 is he subjecive ime discoun facor, and 0 < ψ < 1. The represenaive household is endowed wih one uni of ime. A he beginning of period, he household owns k unis of business capial and d unis of home capial. Boh ypes of capial can be accumulaed using a sandard linear echnology. In paricular, k 2) k 1 = 1 δ k + i and + k d d 3) d + 1 = 1 δ d + i, where i k is gross business invesmen, i d is gross home invesmen, 0 < δ k < 1 is he depreciaion rae of business capial, and 0 < δ d < 1 is he depreciaion rae of home capial. A he beginning of dae 0, he sock of business capial k 0 and he sock of home capial d 0 are given. The home good is produced according o he following producion funcion: ( ) α 1 α 4) h = d 1 n, where n is he amoun of ime spen in marke aciviies, and 0 < α < 1. Observe ha since he ime endowmen is equal o one, 1 n is he amoun of ime ha he household spends in home aciviies. The household s budge consrain is given as follows: k d n 5) c + i + i + T 1 τ w n k k k + 1 τ r k + τ δ k, where w is he wage rae, r is he renal rae on capial, τ n is he ax rae on labor income, τ k is he ax rae on capial income, and T is he lump-sum axes. Observe ha he household receives a ax depreciaion k k allowance given by τ δ k. Also observe ha he household uses is afer-ax labor and capial income o consume, o inves in business capial, o inves in home capial, and o pay lump-sum axes. The household akes he lump-sum axes T, he ax raes τ n and τ k, and he prices w and r as given. The household s problem is o maximize he uiliy funcion (equaion 1) subjec o equaions 4 and 5. The firms secor The represenaive firm produces he marke good using he following producion funcion: θ 1 θ 6) y = K N, where y is oupu, K is business capial, N is labor, and 0 < θ < 1. The firm solves he following saic profi maximizaion problem: 7) max {y r K w N }, subjec o equaion 6. Tha is, he firm maximizes he difference beween he revenues ha i receives from selling is oupu and he oal renal paymens on capial and labor. The firm akes he renal rae of capial r and he wage rae w as given. 22 3Q/2009, Economic Perspecives

7 The governmen secor The governmen needs o make a sequence of expendiures g { }. These expendiures are exogenous = 0 ha is, hey are deermined ouside he model. 16 However, he following governmen budge consrain mus be saisfied: n k k k g = τ w n + τ r k τ δ k + T. Tha is, governmen expendiures mus be financed wih ax revenues. 17 Marke clearing A equilibrium all markes mus clear. In paricular, k d 8) c + i + i + g = y. Tha is, consumpion of he marke good c plus oal k d invesmen i + i plus governmen expendiures g mus be equal o oupu y. Also, K = k, and N = n. Tha is, he renal markes for capial and labor mus clear. 18 Selecion of parameer values Wih consan governmen expendiures and ax raes, he model economy evenually seles ino a seady sae where consumpion, business capial, home capial, oupu, hours worked, and all prices are consan over ime. In wha follows, model parameers are chosen so ha his seady sae reproduces key observaions abou he U.S. economy. Since here are nine parameers o choose, we arge nine observaions. The parameers o be seleced are β, ψ, α, θ, δ d, δ k, g, τ n, and τ k. Before proceeding we need o idenify empirical counerpars for he differen ypes of capial. In wha follows we idenify he sock of home capial d wih he sum of residenial srucures and consumer durable goods. As a consequence, we associae home invesmen i d wih gross privae residenial fixed invesmen plus personal consumpion expendiures on durable goods (from he U.S. Bureau of Economic Analysis s naional income and produc accouns, or NIPAs). In urn, we idenify he sock of business capial k wih oal fixed asses minus residenial srucures. Tha is, k includes privae business srucures, equipmen, and sofware, as well as all forms of governmen capial. As a consequence, we associae business invesmen i k wih privae nonresidenial fixed invesmen plus governmen gross invesmen (from he NIPAs). Using annual daa from 17 hrough 2007 published in he NIPAs, we find ha he corresponding average annual invesmen raes i d /d and i k /k are equal o 9.1 percen and 8.5 percen, respecively, and ha he average invesmen oupu raios i k /y and i d /y are equal o 14.9 percen and 12.7 percen, respecively. 19 This provides four arge observaions. Two addiional arge observaions are he share of labor in naional income (equal o 70 percen) and he average fracion of ime spen working by he oal civilian noninsiuional populaion aged and miliary personnel (equal o 27 percen). 20 The firs observaion, which is sandard in he macroeconomics lieraure, is obained from he NIPAs. The second observaion, which corresponds o he period , is from Presco, Ueberfeld, and Cociuba (2009). The las hree observaions ha we use are associaed wih he governmen secor. The firs of hese observaions is a governmen expendiures raio g/y equal o 16 percen, which is he average over he period in he NIPAs. The oher wo observaions are he pre-egtrra effecive marginal ax raes on labor and capial ha were described in able 1 (p. 21). Table 2 liss he parameer values ha generae hese nine observaions when he model s ime period is se o one quarer. Resuls In his secion, we analyze he effecs of inroducing differen ypes of ax reforms o he economy calibraed in he previous secion. The purpose of he exercises is wofold: o compare he effecs of anicipaed ax cus wih hose of unanicipaed ax cus and o explore wheher anicipaed ax cus may have conribued o generaing some of he aypical feaures of he 2001 recession. In all cases we will assume ha in 2000:Q4, he economy was a he seady sae calibraed in he previous secion. The effecs of immediae ax cus The firs experimen is o evaluae he effecs of immediae ax cus ha is, a ax cu reform ha inroduces no delays beween he ime of is announcemen and he ime of is implemenaion. The experimen s purpose is o illusrae how he model works and o faciliae comparisons wih a delayed reform laer on. The paricular ax cus considered are he oal ax cus promised by George W. Bush during he presidenial campaign. In paricular, we assume ha in 2001:Q1 economic agens learn ha heir marginal ax rae on capial (τ k ) is immediaely and permanenly reduced from is pre-egtrra rae of percen o percen. Similarly, we assume ha Federal Reserve Bank of Chicago 23

8 he marginal ax rae on labor (τ n ) is immediaely and permanenly reduced from percen o percen (see firs and las rows of panel B in able 1, p. 21). We wan o emphasize ha his exercise is purely illusraive: As was described previously, George W. Bush did no promise ha hese ax cus would ake place immediaely bu ha hey would be phased in over a period of several years. Figure 2 shows he evoluion of he economy afer his reform. We see ha, in he model, he reform generaes a boom in economic aciviy. The lower income ax raes increase he reurns o working in he marke and o invesing in business capial. As a consequence, hours worked (panel C) and business invesmen (panel E) increase during he firs period of he reform. Also, he lower reliance on disorionary axes makes households feel richer, and hey respond by increasing heir consumpion (panel B). Observe ha during he firs period of he reform here is a sharp drop in home invesmen (panel F). The reason is ha he lower ax rae on business capial changes he desired mix of capial. In paricular, households wan o hold more business capial and less home capial. During he second period of he reform, business invesmen drops and home invesmen increases. The reason is ha once he correc capial mix is achieved during he firs period of he reform, boh ypes of capial sar growing a a more balanced pace. As business capial increases during he subsequen periods, oupu (panel D) and consumpion coninue o grow and hours worked sar o decrease. The effecs of delayed ax cus In he previous secion, we considered a scenario in which he oal ax cus promised by George W. Bush during his presidenial campaign were immediaely implemened. The scenario was highly unrealisic: In acualiy, his promise was o gradually reduce ax raes in 2002, 2004, and 2006, wih he oal ax cus aking full effec only by Here we consider he scenario in which no only he oal ax cus bu heir pace of reducion are he ones promised during he campaign. In paricular, he sequence of ax raes τ k and τ n inroduced are hose given by he las wo columns of panel B in able 1 (p. 21). The purpose of his experimen is wofold: Firs, i illusraes he effecs of preannouncing ax cus insead of implemening hem as surprises; second, i evaluaes he effecs ha migh have been obained had he ax reform promised by George W. Bush during his campaign been implemened. Figure 3 shows he evoluion of he economy saring in 2001:Q1, when economic agens firs learn ha ax raes will be reduced in he fuure. We see ha in he model economy he delayed reform generaes a Parameer Table 2 Parameer values Value β 0.68 ψ α θ 0.30 d d d k g 0.16 y τ n τ k Noe: The governmen-expendiures-o-oupu raio (g/y) is equal o recession during The reason is ha he anicipaed ax reducion makes 2001 a relaively bad year for working and invesing. Economic agens essenially ake a break from marke aciviies, subsiuing ime worked in he marke for ime worked a home and subsiuing business invesmen for home invesmen. Mos of he invesmen adjusmen akes place in 2001:Q1, when here is a sharp decline in business invesmen (panel E) and a sharp increase in home invesmen (panel F). Also observe ha consumpion (panel B) immediaely jumps o a permanenly higher level because he lower fuure ax raes make he represenaive household richer. Laer, in 2001:Q4, agens know ha axes are going o be cu he following quarer, so hey prepare for his by increasing business invesmen and decreasing home invesmen. This leaves agens a he sar of 2002:Q1 wih a higher sock of business capial and a lower sock of home capial, which are appropriae for he sharp subsiuion in ime worked a home for ime worked in he marke (panel C) and for he increase in oupu (panel D) ha subsequenly akes place. The effecs of EGTRRA The delayed ax reform scenario of he previous secion seems o be a plausible descripion of how prospecive ax cus may have affeced he economy hrough 2001:Q2. There are wo reasons for his. Firs, alhough George W. Bush was already announcing his inenions of cuing marginal ax raes during his 2000 presidenial campaign, i seems unlikely ha his may have had significan effecs on economic decisions before 2001:Q1. If economic agens had changed heir behavior in anicipaion of George W. Bush winning he elecion (and marginal ax raes being reduced), hey would have regreed i laer on had Al Gore become he new presiden (and marginal ax raes had remained unchanged). Given he high uncerainy abou he elecion oucome and given 24 3Q/2009, Economic Perspecives

9 figure 2 Effecs of immediae ax cus A. Tax raes rae Tax on capial Tax on labor B. Consumpion :Q3 01:Q1 01:Q3 02:Q :Q3 01:Q1 01:Q3 02:Q1 C. Hours worked D. Oupu 2000:Q3 01:Q1 01:Q3 02:Q1 2000:Q3 01:Q1 01:Q3 02:Q1 E. Business invesmen 200 F. Home invesmen 2000:Q3 01:Q1 01:Q3 02:Q :Q3 01:Q1 01:Q3 02:Q1 Noe: Panels B hrough F are normalized (ed) o a 2000:Q4. he risk of erring in eiher direcion, i seems reasonable o assume (as a firs approximaion) ha economic agens waied unil he elecion oucome before changing heir behavior. Second, i seems plausible o hink ha once forward-looking economic agens learned by he end of 2000:Q4 ha George W. Bush would become he new presiden, hey sared o adjus heir behavior in anicipaion of he ax cus announced during his presidenial campaign. While being a plausible descripion of he effecs of prospecive ax cus hrough 2001:Q2, he delayed ax reform scenario of he previous secion does no apply Federal Reserve Bank of Chicago 25

10 A. Tax raes rae Tax on capial Tax on labor figure 3 Effecs of delayed ax cus B. Consumpion C. Hours worked D. Oupu E. Business invesmen 200 F. Home invesmen 0 Noe: Panels B hrough F are normalized (ed) o a 2000:Q4. afer 2001:Q2. The reason is ha he acual ax reform passed by Congress and signed ino law in June 2001, EGTRRA, differed in significan ways from he ax reform ha George W. Bush had announced during he campaign: 1) The oal ax cus were smaller (alhough hey would sill ake full effec by 2006), 2) a sunse provision was incorporaed, and 3) small ax cus were already given for he year 2001 (reroacively o he beginning of he year). So, in a hird scenario ha follows, we will assume ha economic agens are surprised by he acual passage of EGTRRA and ha hey revise heir expecaions accordingly. In paricular, we will assume 26 3Q/2009, Economic Perspecives

11 ha in 2001:Q3 economic agens sar o believe ha he fuure sequence of ax raes τ k and τ n will be given by he las wo columns of panel C in able 1 (p. 21). Figure 4 shows he complee evoluion of he economy. By consrucion, he pah ha he economy follows hrough 2001:Q2 is idenical o ha of he delayed ax reform scenario. However, saring in 2001:Q3, he pah is significanly differen. Since now households find ou ha some ax cus already ake place in 2001, hey immediaely shif hours worked a home o hours worked in he marke (panel C) and increase he amoun of oupu produced (panel D). Because of he subsiuion oward marke aciviies, we also see ha in 2001:Q3 here is an increase in business invesmen (panel E) and a drop in home invesmen (panel F). Consumpion (panel B) drops, however, because agens learn ha hey are no as rich as hey iniially believed: EGTRRA now incorporaes a sunse provision. In anicipaion of furher ax cus ha will ake place in 2002:Q1, business invesmen remains relaively high in 2001:Q4 and home invesmen remains relaively low. Once he ax raes are reduced in 2002:Q1, here is an addiional increase in hours worked and oupu, while business invesmen and home invesmen sabilize around heir pre-egtrra levels. Adjusmen coss We saw in he previous secion ha in he model economy, he expecaions of fuure ax cus during he early par of 2001, followed by he acual implemenaion of EGTRRA, generae a shor-lived recession during 2001 in which hours worked and oupu fall, while consumpion remains relaively srong. These are feaures observed in he acual 2001 recession (see figure 1, p. 19). However, home invesmen is exremely srong during he early par of he year and exremely weak during he second half of he year (panel F in figure 4). The opposie is rue wih business invesmen (panel E in figure 4). These large swings in invesmen are highly counerfacual (see figure 1, p. 19). In order o improve he performance, we inroduce adjusmen coss o he model economy. In paricular, we assume ha i is cosly o change boh ypes of invesmens from heir levels in he previous period. Under his assumpion, he household s budge consrain (equaion 5) becomes k d φk k k 2 φd d d 2 c + i + i + i i + i i + T n 1 τ k k k w n + 1 τ r k + τ δ k, and he marke-clearing condiion (equaion 8) becomes k d φk k k 2 φd d d 2 c + i + i + i i + i i + g = y, where φ k 0 and φ d 0. In wha follows, we will assume ha φ k = 0.01 and φ d = These adjusmen coss are quie small. To see his, consider saring from he seady sae calibraed previously and doubling he amoun of business invesmen i k and home invesmen i d. As a fracion of oal oupu, he associaed adjusmen coss urn ou o be and φ k 2 φ d 2 k 2 i y d 2 i y = % = %, respecively, which are small numbers indeed. 21 Figure 5 reproduces he experimen ha we performed in he previous secion o measure he effecs of EGTRRA, bu subjec o he small adjusmen coss described here. We see ha he behavior of he model economy resembles he broad feaures described in figure 4, bu wihou he large swings in he invesmen componens. 22 We conclude ha he model wih adjusmen coss broadly reproduces some of he aciviy paerns observed during he 2001 recession. During he firs year afer he iniial seady sae, he model economy goes ino a recession wih low levels of hours worked, oupu, and business invesmen (panels C, D, and E) bu wih relaively srong consumpion and home invesmen (panels B and F). The recovery saring in he second half of 2001 seems o be oo srong, hough. However, his is no surprising, since he model absracs from imporan subsequen shocks o he economy, such as he Enron accouning scandal and Sepember 11 erroris aacks, and he cyclical propagaion of earlier shocks, such as he high-ech bus and he associaed sock marke decline of The 2008 presidenial elecion So far we have focused on he effecs of anicipaed ax cus following he 2000 presidenial elecion. In his secion, we consider he mirror image: he afermah of he 2008 presidenial elecion. During he 2008 presidenial campaigns, he Republican candidae, John McCain, promised o make permanen he ax cus ha were implemened by he George W. Bush adminisraion. Conrary o ha campaign proposal, he Democraic candidae, Barack Obama, made i clear ha, a leas for high-income Federal Reserve Bank of Chicago 27

12 figure 4 Effecs of EGTRRA A. Tax raes rae Tax on capial Tax on labor B. Consumpion C. Hours worked D. Oupu E. Business invesmen 200 F. Home invesmen 0 Noes: EGTRRA means he Economic Growh and Tax Relief Reconciliaion Ac of Panels B hrough F are normalized (ed) o a 2000:Q4. individuals, he would le hose ax cus expire wih he exising sunse provision. Wih Barack Obama winning he presidency in 2008 and his pary reaining conrol of Congress, economic agens are likely o have concluded ha ax raes would parly rever in 2011 o heir pre-egtrra levels. In wha follows, we illusrae he effecs of his ype of anicipaed ax increase. Since i is exremely difficul o deermine wha ax increases migh evenually be implemened and wha impac hey migh have on effecive marginal ax raes, we make an exremely simplisic assumpion: Tha all ax raes will 28 3Q/2009, Economic Perspecives

13 A. Tax raes rae Tax on capial Tax on labor figure 5 Effecs of EGTRRA wih adjusmen coss B. Consumpion C. Hours worked D. Oupu E. Business invesmen 200 F. Home invesmen 0 Noes: EGTRRA means he Economic Growh and Tax Relief Reconciliaion Ac of Panels B hrough F are normalized (ed) o a 2000:Q4. rever o heir pre-egtrra levels by 2011 and ha all economic agens believe ha his will be he case. While his is a highly unrealisic assumpion, i will suffice for our illusraional purposes. Specifically, we assume ha in 2008:Q4 he economy was a he seady sae corresponding o he low marginal ax rae on capial (τ k ) of percen and on labor (τ n ) of percen inroduced by he George W. Bush adminisraion (see fourh row of panel C in able 1, p. 21). We also assume ha in 2009:Q1 economic agens find ou ha marginal ax raes on capial (τ k ) and labor (τ n ) will be permanenly increased o heir Federal Reserve Bank of Chicago 29

14 figure 6 Effecs of revering o pre-egtrra ax raes A. Tax raes rae 0.38 B. Consumpion Tax on capial Tax on labor :Q3 09:Q1 10:Q1 11:Q1 12:Q1 2008:Q3 09:Q1 10:Q1 11:Q1 12:Q1 C. Hours worked D. Oupu 2008:Q3 09:Q1 10:Q1 11:Q1 12:Q1 2008:Q3 09:Q1 10:Q1 11:Q1 12:Q1 E. Business invesmen 200 F. Home invesmen :Q3 09:Q1 10:Q1 11:Q1 12:Q1 2008:Q3 09:Q1 10:Q1 11:Q1 12:Q1 Noes: EGTRRA means he Economic Growh and Tax Relief Reconciliaion Ac of Panels B hrough F are normalized (ed) o a 2008:Q4. pre-egtrra levels of percen and percen, respecively, saring in 2011:Q1. For comparabiliy, we inroduce he same adjusmen coss considered in he previous secion. Figure 6 presens he resuls. We see ha during 2009 and 2010 he model economy experiences a marked boos in economic aciviy, wih hours worked and oupu increasing monoonically over ime (panels C and D). Business invesmen (panel E) is srong during 2009 bu weak during The opposie is rue for home invesmen (panel F). Consumpion (panel B) drops immediaely because economic agens feel poorer from he higher expeced ax raes. The reason for he early economic boom is ha agens realize ha he period before he ax 30 3Q/2009, Economic Perspecives

15 increase is relaively aracive for working and invesing. Once he ax increase akes place in 2011:Q1, here is a sharp drop in hours worked and oupu. So, according o he model, he curren recession would have been even worse in erms of hours worked and oupu had here no been expecaions of fuure ax increases. However, hese expecaions migh be conribuing o he weakness in consumpion and residenial invesmen ha we currenly see. The model expecs a furher downward influence on economic aciviy in 2011 once ax raes are increased. Conclusion In his aricle, we used a sylized model economy o invesigae he hypohesis ha some of he unusual feaures of he 2001 recession may have been influenced by he ax cus promised during he 2000 presidenial campaigns. We found ha he model is consisen wih his hypohesis: In he model economy, anicipaed ax cus generae a mild recession wih relaively srong consumpion and home invesmen, bu wih weak hours worked and business invesmen. Because he 2008 presidenial elecion also had a significan impac on fuure ax raes, bu in reverse, we also used our model o illusrae he possible effecs on he economy of anicipaed fuure ax increases. Boh of hese applicaions illusrae a more basic resul in economic heory: Tha anicipaed fuure changes in economic policy migh have large effecs on curren economic aciviy. NOTES 1 The argumen ha expeced ax cus one year down he road decrease invesmen during he curren year is based on he assumpion ha invesmen affecs he sock of capial raher quickly (say, wihin one quarer). If here were long gesaion lags in building capial, anicipaed ax cus in one year may acually increase invesmen during he curren year. The assumpion ha capial is quickly buil will be mainained hroughou his aricle. 2 See 3 The NBER looks a more han gross domesic produc (GDP) in deermining when a recession sars, so ha he iming can differ from ha of declines in GDP. According o revised esimaes, he annualized growh raes in GDP were +2.1 percen, 0.5 percen, +1.2 percen, and 1.4 percen in 2000:Q4, 2001:Q1, 2001:Q2, and 2001:Q3, respecively. However, a he ime ha he NBER declared ha he recession had sared, hese numbers were +1.9 percen, +1.3 percen, +0.3 percen, and 0.4 percen, respecively. 4 See 5 The six previous recessions had occurred in 10 61, 19 70, , 10, 11 82, and , according o he Naional Bureau of Economic Research. 6 The source of all daa unless oherwise specified is he U.S. Bureau of Economic Analysis s Naional Income and Produc Accouns of he Unied Saes (NIPAs) from Haver Analyics. 7 The deerminisic rends are deermined by regressing he log of he differen variables agains ime. 8 Oupu sared o decline a few quarers before he sar of he 2001 recession, as i did before some previous recessions. However, once he 2001 recession sared, oupu did no fall as much as during he average recession. 9 Prior o he campaigns, wo pieces of legislaion defined he exising ax srucure: he Omnibus Budge Reconciliaion Ac of 1993 and he Taxpayer Relief Ac of While he laer only had effecs on capial gains axes, he former creaed he 36 percen and 39.6 percen income ax brackes and se he sauory raes a he levels seen in able 1, panel A (p. 21). 10 Bush (2001). 11 According o House and Shapiro (2006), only households wih axable income below $12,000 acually experienced any reducion in heir marginal ax rae as a resul of he new 10 percen bracke. Though his change can be expeced o have had large effecs on average ax liabiliies, marginal raes remained relaively unaffeced. 12 The CBO simulaed income ax liabiliy for each reurn in a sample of all ax reurns filed in he Unied Saes. The analysis hen calculaed marginal ax raes by adding $1,000 o he earnings on each reurn and recompuing he amoun of income ax owed. The difference beween he wo ax liabiliies, divided by $1,000, equals he effecive marginal ax rae. 13 Our effecive marginal ax raes on capial are roughly wice as large as he CBO s esimaes, since our noion of capial does no include residenial srucures, while he CBO s does. See he appendix for a brief discussion regarding he reamen of housing capial. 14 Among oher hings, his implicily assumes ha he AMT is changing in a similar way. 15 By allowing for lump-sum axes, he opimal ax sysem is o se he capial and labor income axes o zero and rely exclusively on lump-sum axes. However, he focus of our analysis is no on opimal axaion bu on he effecs of acual ax raes. 16 Alhough he 2000 elecion oucome may have had implicaions for prospecive governmen expendiures, we absrac from hese. 17 Observe ha we are assuming ha governmen expendiures are unproducive. However, his could be modified wihou alering he analysis by assuming ha governmen expendiures ener he uiliy funcion in a separable way. 18 Observe ha here is no renal marke for home capial: All home capial is direcly owned by he household secor. This is a limiaion of he model economy. In pracice, a significan fracion of residenial srucures are rened, and he income generaed is subjec o axes. Anoher limiaion of he model is ha he sock of home capial is no axed a all, alhough in he U.S. economy, housing is subjec o propery axes. 19 Since ours is a closed economy, he measure of oupu ha we use is GDP minus ne expors. 20 The implied Frisch elasiciy of labor supply is equal o 2.7. This is somewha lower han he Frisch elasiciy of labor supply used by Presco (2004) in his cross-counry analysis of labor income axes bu much higher han economeric esimaes based on microeconomic daa. However, recen research has shown ha he large elasiciy of labor supply used by he macro lieraure can be reconciled wih he micro evidence hrough heerogeneiy in labor supply (for example, Chang and Kim, 2006; Rogerson and Wallenius, 2009; and Gourio and Noual, 2007). 21 We make no claim ha hese adjusmen coss are empirically plausible. However, hey improve he model s performance quie significanly. 22 The invesmen measures repored in figure 5 (p. 29) include he adjusmen coss. However, in pracice his does no maer because he adjusmen coss are exremely small. Federal Reserve Bank of Chicago 31

16 APPENDIX: Calibraion of he ax rae on capial ( k ) When he Congressional Budge Office (2001) esimaes he effecs of EGTRRA on effecive marginal ax raes for labor and capial income, i includes residenial srucures in is noion of capial. Because owner-occupans of residenial srucures exclude heir implici gross receips (i.e., he renal value of he home) from axable income... [and may] deduc morgage ineres and propery ax paymens if hey iemize heir deducions, he CBO concluded ha owner-occupied housing capial is subsidized (Congressional Budge Office, 2005). Since his subsidy is no given o enan-occupied housing capial, he CBO concluded ha his form of capial is axed. Using he CBO s esimaes of he effecive ax raes on boh enanoccupied (18.2 percen) and owner-occupied ( 5.1 percen) housing capial income, as well as he proporion of each ype of capial in oal housing capial, we obain ha he h housing capial income ax rae τ is given as follows: 1 h τ = (% Tenan-occupied housing) (0.182) + (% Owner-occupied housing) ( 0.051) = (0.20) (0.182) + (0.80) ( 0.051) = Since he services from consumer durables are no h axed and τ 0, we have ha he effecive marginal ax rae on capial income esimaed by he CBO is equal o: τ k CBO = k k d τ k d + + k + d 0, where k represens business capial and d represens home capial. Using he average raio (k + d) / k over he period , we have ha he ax rae on business capial income is hen given by: τ k k k + d τcbo d ( ) ( ) The fracions of each ype of housing capial are obained from Congressional Budge Office (2005), p. 19, able A-1. references Beaudry, Paul, and Franck Porier, 2007, When can changes in expecaions cause business cycle flucuaions in neoclassical seings?, Journal of Economic Theory, Vol. 135, No. 1, July, pp Brumbaugh, David L., Jane G. Gravelle, Seven Maguire, Louis Alan Talley, and Bob Lyke, 2002, 2001 ax cu: Descripion, analysis, and background, CRS Repor for Congress, Congressional Research Service, No. RL30973, December 9. Bush, George W., 2001, The Presiden s agenda for ax relief, Whie House, repor, Washingon, DC: U.S. Governmen Prining Office, February, available a hp://georgewbush-whiehouse.archives. gov/news/repors/axplan.hml. Chang, Yongsung, and Sun-Bin Kim, 2006, From individual o aggregae labor supply: A quaniaive analysis based on a heerogeneous agen macroeconomy, Inernaional Economic Review, Vol. 47, No. 1, pp Congressional Budge Office, 2005, Taxing capial income: Effecive raes and approaches o reform, repor, Washingon, DC: U.S. Governmen Prining Office, Ocober., 2001, The budge and economic oulook: An updae, repor, Washingon, DC: U.S. Governmen Prining Office, Augus. Gourio, François, and Pierre-Alexandre Noual, 2007, The marginal worker and he aggregae elasiciy of labor supply, Boson Universiy, mimeo, July. Greenwood, Jeremy, and Zvi Hercowiz, 1991, The allocaion of capial and ime over he business cycle, Journal of Poliical Economy, Vol. 99, No. 6, December, pp House, Chrisopher L., and Mahew D. Shapiro, 2006, Phased-in ax cus and economic aciviy, American Economic Review, Vol., No. 5, December, pp Jaimovich, Nir, and Sergio Rebelo, 2009, Can news abou he fuure drive he business cycle?, American Economic Review, forhcoming., 2008, News and business cycles in open economies, Journal of Money, Credi, and Banking, Vol. 40, No. 8, December, pp Presco, Edward C., 2004, Why do Americans work so much more han Europeans?, Federal Reserve Bank of Minneapolis Quarerly Review, Vol. 28, No. 1, July, pp Presco, Edward C., Alexander Ueberfeld, and Simona Cociuba, 2009, U.S. hours and produciviy behavior using CPS hours worked daa: 1947-III o 2008-IV, Federal Reserve Bank of Minneapolis, mimeo. Rogerson, Richard, and Johanna Wallenius, 2009, Micro and macro elasiciies in a life-cycle model wih axes, Journal of Economic Theory, forhcoming. 32 3Q/2009, Economic Perspecives