The long erm effecs of capial gains axes in New Zealand Andrew Coleman Mou Working Paper 09-13 Mou Economic and Public Policy Research Augus 2009
Auhor conac deails Andrew Coleman Mou Economics and Public Policy Research andrew.coleman@mou.org.nz Acknowledgemens I d like o hanks he New Zealand Treasury for heir funding of his work. I would like o hank Arhur Grimes for commens on his paper. Mou Economic and Public Policy Research PO Box 24390 Wellingon New Zealand Email info@mou.org.nz Telephone +64-4-939-4250 Websie www.mou.org.nz 2009 Mou Economic and Public Policy Research Trus and he auhors. Shor exracs, no exceeding wo paragraphs, may be quoed provided clear aribuion is given. Mou Working Papers are research maerials circulaed by heir auhors for purposes of informaion and discussion. They have no necessarily undergone formal peer review or ediorial reamen. ISSN 1176-2667 (Prin), ISSN 1177-9047 (Online). i
Absrac This paper develops a model of he housing marke incorporaing a consrucion secor, a renal secor, and a housing demand secor o examine he long erm consequences for he housing marke of differen ypes of capial gains axes. The secor is based on an overlapping generaions model of he economy ha included a deailed represenaion of he credi consrains and ax regulaions affecing households. The model suggess ha capial gains axes will raise rens, increase homeownership raes, rebalance he housing sock owards smaller houses, and increase he ne foreign asse posiion. The implicaions for welfare are much less clear, however, paricularly for young low income households ha will face higher rens. JEL classificaion E40, E58 Keywords Inflaion, credi consrains, capial income axes, housing markes, home-ownership raes, moneary policy Summary haiku Money debases; he Crown claims he widow's mie, and poor people ren. ii
Conens 1 Inroducion... 1 2 An inergeneraional model of housing demand... 3 2.1 The basic framework... 3 2.2 Modelling capial gains axes.... 7 3 Resuls... 9 3.1 The effec of inflaion... 9 3.1.1 Price and disribuional effecs of inflaion... 9 3.1.2 The welfare effecs of inflaion... 11 3.2 The effec of capial gains axes wih an exempion for owner-occupied housing....12 3.2.1 Taxing capial gains on leased propery a marginal income ax raes 13 3.2.2 Taxing capial gains on leased propery a a consan rae... 14 3.2.3 Taxing capial gains on leased propery: resuls for oher supply funcions... 14 3.3 Exemping he inflaion componen of ineres from income ax...15 3.4 The effec of an accrual capial gains regime applied o all agens...16 3.5 The welfare implicaions of capial gains axes...17 4 Discussion and conclusions...19 5 References...23 6 Tables...25 7 Figures...32 8 Appendix 1...33 iv
1 Inroducion This paper develops a model ha analyses he long erm effecs of capial gains axes on New Zealand s residenial propery markes. The model is an exension of he model developed by Coleman (2008) ha analysed how New Zealand s ax sysem migh be disoring he housing marke when he inflaion rae is posiive bu relaively modes. Tha paper argued ha even wo o hree percen inflaion raes may significanly lower homeownership raes, parly because young households face binding credi consrains whose effecs inensify as he inflaion rae increases, bu also because ineres income bu no capial gains are axed. The paper furher argued ha mos of he effecs on he housing marke semming from he ineracion of inflaion wih he ax sysem could be avoided by exemping he inflaion componen of ineres income from ax, or by reducing he inflaion rae o zero. However, a hird opion exiss: inroducing a ax on capial gains may reduce he disorionary effecs of he curren ax sysem. To assess his possibiliy, his paper invesigaes how various capial gains ax regimes affec housing markes when here is inflaion. The model is a version of he overlapping generaions lifecycle model pioneered by Modigliani and Brumberg (1980), and adaped o analyse housing issues by Oralo-Magné and Rady (1998, 2006). A he hear of he model is a dynamic, forward looking maximizaion problem in which agens make choices abou he ype of housing in which hey live, how much hey consume and save, and how much hey borrow and lend. These agens, who differ by income, age, and wealh, have choices over wheher o ren or buy, o live in large or small houses, or o share housing wih oher people. They face realisic bank imposed consrains on he amoun hey can borrow and he repaymen schedule hey face if hey a purchase a house, and hey face a ax sysem ha closely reflecs ha prevailing in New Zealand. Paricular aenion is paid o he various ways ha axes on housing income differ according o wheher one is an owner-occupier of housing or a landlord. House prices and rens are deermined endogenously in he model, and reflec he ineracion of decisions by households, landlords, and a consrucion secor o supply or demand housing. The model calculaes dynamic seady-sae pahs for house prices and rens, and a se of equilibrium housing supply and demand paerns ha depend on fundamenal parameers such as ineres raes, consrucion secor supply 1
elasiciies, he inflaion rae, and he pariculars of he ax sysem. The paper examines how hese prices and demand paerns change as axes and he inflaion rae change, and uses hese resuls o evaluae he consequences of differen possible ax sysems. The paper examines he effecs of four varians of a capial gains ax regime. While all four are accruals based, hey differ according o wheher owneroccupied housing is axed or exemp, and wheher capial gains are reaed as income and axed a households marginal income ax raes, or wheher capial gains are simply axed a a fla rae. Many of he resuls of he four varians are similar, alhough here are imporan differences, paricularly in he amoun of revenue ha is raised by he ax. In general, when he inflaion rae is moderae, capial gains axes lead o an increase in rens, an increase in he home-ownership rae, a small reducion in number of large houses in he economy, and an increase in he ne foreign asse posiion. However, he effecs on economic welfare are ambiguous, for many lowincome households suffer a welfare loss from he increase in rens. The simulaions suggess he welfare consequences will be worse for low income households if owner-occupied housing is exemp from he ax, alhough his resul is dependen on he revenue from a capial gains ax being refunded o households (including low income households) hrough a reducion in he GST rae. The primary purpose of he paper is o explore he possible economic consequences of differen ypes of axes, no o make a recommendaion as o heir desirabiliy or pracicaliy. Noneheless, he paper noes ha he welfare consequences of a capial gains ax applied o all households are similar o he welfare consequences of a fla rae propery ax. Similarly, he welfare consequences of a capial gains ax ha exemps owner-occupied housing are similar o he welfare consequences of a ax sysem ha exemps he inflaion componen of ineres income from income ax. Boh of hese alernaive ax regimes may be easier o implemen han a capial gains ax. Consequenly, i may be possible o devise alernaive ax regimes ha have similar effecs o a capial gains ax wihou some of heir perceived adverse consequences. The paper is organised as follows. Secion 2 oulines he srucure of he model. (The echnical deails of he model are conained in a lenghy appendix.) Secion 3 discusses he resuls of he simulaions, beginning wih an exploraion of 2
he welfare consequences of he effecs of inflaion on he housing marke, and concluding wih a discussion of he welfare consequences of differen capial gains ax sysems. Conclusions are offered in secion 4. 2 An inergeneraional model of housing demand 2.1 The basic framework The model is an exension of he model used by Coleman (2008) o analyse he effec of inflaion, credi consrains and New Zealand s ax sysem on he housing marke. In urn, i is based on he overlapping generaions housing model of Oralo-Magné and Rady (1998, 2006). The deails of he model are described in he appendix, bu he basic srucure of he model has four key pars: he demand for renal and owner-occupied housing; he supply of renal housing; and he oal supply of housing. The demand for housing is based on an ineremporal uiliy maximisaion model of consumer demand applied o a large number of agens who differ by age, income, and wealh. In he model, here are four cohors each conaining 400 agens, wih each agen passing hrough four disinc sages (wo young sages, one middleaged sage, and one sage in reiremen) before dying. The agens have differen exogenously deermined labour income, which follows a life-cycle paern. The agens consume a single non-sorable good, pay ax, save for reiremen, and have choices over differen ypes of housing a each sage of heir lives wheher hey share housing wih oher agens, ren a small house (an aparmen), buy a small house or buy a large house. The agens choose heir mos preferred housing opions, given heir age, wealh and afer-ax incomes, he cos of rening or buying differen houses, and heir abiliy o raise a morgage. Agens can borrow or lend a exogenously deermined ineres raes, alhough young agens face bank imposed credi consrains limiing he amoun hey can borrow. In he las period of life agens consume all wealh excep heir house, which is inheried by a younger generaion. The model is dynamic and house prices and rens can change hrough ime. Indeed, when choosing heir housing opions agens ake ino accoun boh he rae a which house prices appreciae and he ax reamen on any capial gains ha hey make. Sricly speaking, in he model house prices and rens comprise wo pars: 3
a price level a some base period ( = 0); and a price (or ren) appreciaion rae. The model calculaes he rae of propery price appreciaion as par of he process by which i calculaes equilibrium prices; while i is normally he general inflaion rae, i does no need o be. Agens are assumed o be forward looking, so when hey choose housing in a paricular period hey ake ino accoun no only heir curren income and curren housing prices, bu heir remaining lengh of life, fuure house prices, heir fuure income sream, and heir desired fuure housing paerns. The model includes a careful represenaion of he condiions imposed by banks on hose obaining morgage finance o purchase a house, including realisic consrains on he minimum deposi and he maximum morgage repaymen o income raio. These consrains mean ha young households may choose o ren raher han buy a house when inflaion and nominal ineres raes are high, because hey canno obain suiable financing. The uiliy maximisaion model generaes housing demand for each of he agens during each of heir life sages, for a given se of ren and house price pahs. These differen housing demand funcions are hen aggregaed ogeher. The resuling aggregae demand funcions describe how he demand o ren, he demand for small houses, and he demand for large houses varies as a funcion of he ren and he price of each ype of house, as well as all he basic parameers of he model such as income, ineres raes, and ax raes. Renal accommodaion is supplied by agens who become landlords. I is assumed ha enry ino he renal secor is compeiive, so landlords bid for houses and se rens a levels ha leave hem indifferen beween he afer-ax reurns from lending money and he afer-ax reurns from invesing in residenial propery. The marginal compeiive landlord is assumed o be a middle aged agen who is on he op marginal income ax rae. Paricular care has been aken o ensure ha axes in he model replicae he axes currenly imposed on housing in New Zealand. If house prices increase over ime, a capial gains ax will lower reurns o landlords, and, for a given level of house prices, rens will be higher han hey would oherwise have been. Prices are deermined endogenously in he model by equaing he oal demand for differen ypes of houses wih he supply of differen ypes of houses. 4
Cos funcions describing he coss of building large and small houses are specified exogenously in he model, and can ake any form. In his model, I focus on he case ha here are separae upward sloping supply curves for he quaniy of large and small houses, each wih approximaely uni elasiciies. An elasiciy of 1 is broadly consisen wih he long run increase in prices and he quaniy of houses in New Zealand beween 1960 and 2005. Two differen parameerisaions ha reflec house prices ha are relaively high or relaively low in comparison o income because of high or low consrucion coss are examined. Several oher combinaions of supply elasiciies have also been analysed, including he cases when he supply of boh classes of houses are eiher perfecly elasic or perfecly inelasic, and he case ha he supply of small houses is more elasic han he supply of large houses. A soluion o he model is obained by finding a se of prices ha equae he aggregae demand for differen ypes of housing wih he aggregae supply of hese ypes of housing. The prices are solved using a complex numerical rouine ha calculaes he housing demand for each of he 1600 differen households for a se of prices, and hen chooses a sequence of prices unil a se is found a which aggregae demand equals aggregae supply. For his equilibrium se of prices, overall demand paerns are calculaed. As Coleman and Scobie (2009) argue, he effec of axes, inflaion, and ineres raes on he housing marke depend on a few crucial elasiciies including (i) he elasiciy of he oal supply of houses o he price of houses (he elasiciy of he supply of housing); (ii) he elasiciy of he supply of renal housing wih respec o rens; (iii) he elasiciy of he demand for renal housing wih respec o rens and he prices of houses; and (iv) he elasiciy of he oal demand for housing wih respec o rens and he price of houses. The elasiciy of he supply of housing wih respec o prices is se equal o 1 in he main versions of he model discussed below, bu he resuls have also been analysed when his elasiciy is near zero or very large. The supply of renal housing is perfecly elasic wih respec o ren, because landlords are assumed o be perfecly compeiive and o supply renal housing unil he long run afer-ax reurn on renal accommodaion is equal o he afer-ax reurn on ineres income. The demand elasiciies are no direcly imposed, bu are implicily derived from he consumer maximisaion problem and depend on he basic parameers in he 5
model. These elasiciies can have a major effec on he model s resuls and warran furher discussion. The elasiciy of he demand for renal accommodaion o housing renals and he prices of houses is a measure of he exen o which households are prepared o subsiue beween rening and home-ownership. This will depend on he relaive uiliy households ge from sharing, rening, or owning a house. These parameers are explicily specified in he model; ypically, households are assumed o gain less uiliy from rening raher han owning, because hey can shape an owned house in heir own image, and less uiliy from living in shared accommodaion han living by hemselves. The subsiuabiliy beween renal and owned accommodaion will be greaer he smaller he differences beween rening and owning. The more willing households are o subsiue beween rened and owned housing, he less will be he uiliy loss from various housing marke imperfecions or policy inervenions. The elasiciy of oal housing demand, wih respec o house prices or rens, measures he exen o which new households form when prices change as adul children leave home, or as young aduls form households by hemselves raher han sharing wih a group of ohers. This elasiciy is imporan as i is he only mechanism by which oal housing demand can be alered. The more willing are households o share wih ohers, he smaller are he price changes necessary o equilibrae he housing marke. This paper depars from he earlier work by Oralo-Magne and Rady (2006) and Coleman (2008) by inroducing a mechanism o subsanially increase his elasiciy. In hese earlier papers, he only way households could share was for he younges households o remain a home wih heir parens. In his paper, he younges wo cohors can share by rening half a house. If hey do so, hey pay half he ren and obain uiliy which, while less han he uiliy of a whole rened house, can eiher be greaer han or less han half he uiliy of a rened house. This opion proves o be aracive o many households, paricularly o hose wih low incomes or seep life cycle earnings, resuling in a higher elasiciy of oal demand wih respec o boh ren and house prices han in hese earlier papers. As a resul, smaller price changes are needed o induce changes in oal housing demand han in hese earlier papers, and he welfare changes of policy inervenions are smaller. 6
The model analyses he way households climb a housing ladder over he course of heir lives. In large par heir ascen can be characerised by wo facors: he ulimae heigh hey reach and he speed ha hey aain ha heigh. The ulimae heigh is largely deermined by he raio of life-ime income o he user-cos of housing: people wih higher life-ime incomes will be able o afford larger houses han people wih low lifeime income. In he parameerisaions sudied a majoriy of people choose a large house in middle age, parly because he ax sysem favours home ownership as impued ren is no axed. The speed of ascen is deermined by he seepness of he earnings profile, ineres raes and he availabiliy of credi from banks, and he ax incenives facing boh households and propery invesors. In equilibrium, he mix of small and large houses in he economy is deermined boh by he lengh of ime spen climbing he housing ladder and he peak rung a household aains. Policies ha exend he lengh of ime climbing he housing ladder do no, however, necessarily reduce he demand for large houses because an agen can use he money saved by living in a small house while young o live for longer in a large house while old. 2.2 Modelling capial gains axes. When he inflaion rae is posiive, income from ineres paying asses is axed more heavily han income from oher forms of capial asses because he inflaion componen of nominal ineres earnings is axed, while capial gains are no. The asymmery of his reamen means ha he ax sysem generaes an incenive for agens o borrow and inves in asses ha appreciae over ime, poenially inducing agens o inves oo heavily in residenial housing asses, and lowering home ownership raes among younger and lower income agens. One possible mehod o eliminae he asymmerical ax reamen would be o exemp he inflaion componen of ineres income from income ax, as i is no income. Anoher poenial mehod would be o impose a capial gains ax on residenial propery and oher asses. Four differen capial gains ax schemes are considered. The firs scheme reas capial gains as income, and axes hese gains a a axpayer s marginal income ax rae. Owner-occupied housing is exemp. As landlords are assumed o be higher income, middle aged agens, he applicable ax rae is he op marginal ax rae, 33 7
percen 1. The second scheme also axes capial gains a he sandard marginal income ax raes of 20 or 33 percen, bu in his case owner-occupied housing is axed. The hird scheme imposes a fla rae capial gains ax of 20 percen, wih an exempion for owner-occupied housing. This ax scheme is similar o ha which operaes in he Unied Saes of America. The fourh scheme is similar, excep all housing including owner-occupied housing is liable o a fla capial gains ax of 20 percen. The model s resuls when he capial gains of owner-occupied housing are axed are concepually problemaic. Almos no counries apply capial gains axes o owner-occupied housing, for a variey of economic and poliical reasons. One reason is ha hese axes are usually only imposed on realised gains when a house is sold, a rule ha migh deer households from moving from one locaion o anoher, perhaps in response o work opporuniies. Anoher reason concerns he financial hardship such a ax could cause when a household dissolves, perhaps because of divorce. These negaive effecs, which may have firs order welfare consequences, canno be modelled in his paper and are ignored. To avoid hese sors of issues, i is assumed ha if a capial gains ax is inroduced wihou an exempion for owner-occupied housing, i is imposed on an accruals basis: ha is, he household is liable for capial gains ax each year, wheher or no i is realised hrough he sale of he house. These resuls are used o provide a reference case for he effec of a capial gains ax. The rae of propery price appreciaion is an oucome of he model, and propery prices appreciae in real erms over ime when here is populaion or income growh, unless he supply of housing is perfecly elasic. In he resuls presened below, however, here is neiher income growh nor populaion growh and so propery prices increase a he inflaion rae. In hese circumsances a capial gains ax only axes he increase in nominal housing wealh ha is due o inflaion and hus i reduces he disorion ha arises from he asymmerical axaion of ineres income and oher asses. Noneheless, he ax sysem remains non-neural wih respec o he inflaion rae because ax raes on real capial income are an increasing funcion of he inflaion rae when he inflaion componen of ineres income is axed. 1 The op marginal ax rae in New Zealand is currenly 38 percen, alhough i was 33 before 2000. However, mos landlords could choose o pu a leased propery in a rus which is only axed a 33 8
The revenue raised by axing capial income and/or capial gains depend on boh he ax raes and he inflaion rae. In he model, any addiional revenue raised from changes in he ax sysem or changes in he inflaion rae are refunded hrough a change in he Goods and Services Tax (GST) rae, so ha he amoun of ax raised is invarian o he ax sysem. 3 Resuls 3.1 The effec of inflaion 3.1.1 Price and disribuional effecs of inflaion Inflaion has hree major effecs on he housing marke. Firs, i increases he rae a which propery prices increase in nominal erms, generaing nominal capial gains for he owners. Since nominal ineres earnings raher han real ineres earnings are axed, here is an incenive for owners of capial o inves in residenial housing when he inflaion rae is high. This incenive applies o boh landlords and owner-occupiers, so by iself inflaion does no necessarily lead o a decline in he home-ownership rae, alhough i may lead o over invesmen in residenial housing. Secondly, inflaion may lead o a reducion in nominal rens. This is because (i) afer ax real reurns from ineres earning asses decline as he inflaion rae increases, as he inflaion componen of ineres income is axed, and (ii) landlords ge a porion of heir reurn as capial appreciaion, and are prepared o pay more for houses or o accep less ren in order o become landlords. The balance beween lower rens and higher prices will depend on he supply elasiciy of housing. When supply is relaively elasic, and new consrucion limis he amoun propery prices increase, rens will end o fall. Thirdly, inflaion exacerbaes he credi consrains facing agens who borrow o buy houses. This is because bank imposed resricions on he amoun households can borrow are rarely adjused for inflaion, even hough nominal ineres raes increase when he inflaion rae rises. If banks do no increase he amoun credi-consrained households can borrow when nominal deb servicing paymens increase, i becomes more difficul for hese households o purchase houses percen. I have chosen o solve he model for a op marginal ax rae of 33 percen in par because his rae is ofen seen as a goal by poliical paries, and in par because of he way landlords can use russ. 9
(Modigliani, 1976; Kearl, 1979). In addiion, if rens fall, inflaion makes i aracive for young, credi consrained agens o ren raher han purchase and home ownership raes are likely o decline. Tables 1a and 1b show how inflaion affecs long erm housing marke oucomes in he model when here is no capial gains ax and here is an elasic supply of housing. In able 1a consrucion coss and house prices are approximaely 25 percen higher han in able 1b. The ables show how prices and rens, he number of houses, he fracion of people owning, and he seady-sae level of ne financial asses vary wih he inflaion rae. In boh cases, a 2 percen increase in he inflaion rae leads o a 6 percen reducion in rens, a 0.8% reducion in he number of houses in he economy, a 3-4 percenage poin increase in he fracion of he populaion rening, and a 2-3 percenage increase in he fracion of he housing sock ha is owned by landlords and leased. More people ren a all ages. Irrespecive of consrucion coss, here is a reducion in he number of small houses as he combinaion of falling rens and igher credi consrains induces more young households o share, reducing aggregae demand for housing. The effec on he number of large houses in he economy differs in he wo cases. When consrucion prices are high, rising inflaion increases he oal demand for large houses despie falling demand among younger cohors: here is an increase in he demand by older households, because of he ax advanages of using a house as a saving vehicle. However, when consrucion coss are low, mos people who wan o live in a large house can afford o do so for mos of heir lives, and inflaion has a very small effec on he quaniy of large houses. These resuls are subly differen han hose repored in Coleman (2008). Firs, he effec of inflaion on homeownership is smaller han repored here, because owning half a house is a more aracive opion han rening a whole house. Secondly, in ha model he oal demand for housing increased raher han decreased as he inflaion rae increased. The difference sems from he way he process of household formaion is modelled. In Coleman (2008), he oal number of households varies only because of changes in he number of adul children who lived wih heir parens. In ha case, a decline in ren araced children ou of he parenal home and increased he oal demand for housing. In his model, he oal number of households varies because of changes in he number of young aduls who live by 10
hemselves raher han wih ohers. In his case a decline in rens could eiher lead o an increase in he oal demand for housing, as agens decide o sop sharing and ren by hemselves, or a reducion in he oal demand for housing, as agens decide o sop owning, and ren shared accommodaion insead. In he parameerisaions sudied in his paper he laer effec dominaes, so ha oal housing demand decreases as inflaion increases, credi consrains inensify, and rens decline 2. 3.1.2 The welfare effecs of inflaion In keeping wih earlier work by Modigliani (1976), Kearl (1979), and Feldsein (1996, 1997), Coleman (2008) argued inflaion reduced he welfare of many bu no all households because of binding credi consrains on young agens. In he model, inflaion has a large effec on young people because almos all agens would increase heir uiliy if hey could borrow more when young, eiher o smooh consumpion in he face of rising life-cycle income, or o buy a house, or boh. They do no borrow more because banks only make collaeral backed loans. In his environmen, inflaion has ambiguous effecs on welfare. Those agens who wish o purchase a house find inflaion ighens credi consrains, because nominal ineres raes increase and banks do no change heir lending erms and condiions o make an allowance for he way inflaion reduces he real value of he nominal ousanding deb. This makes i more difficul for he agens whose real incomes increase over ime o smooh consumpion, for hey have o reduce heir consumpion o make higher nominal ineres paymens if hey purchase a house. In conras, hose agens who ren benefi from inflaion, because i lowers he ren hey pay and enables hem o spend more while young han hey oherwise could. Wheher inflaion causes welfare losses or improvemens on average depends on he relaive size of he populaions ha ren and own when young. In urn, his depends on he raio of house prices o incomes. When consrucion coss are high, a large fracion of young people will wish o share accommodaion wih ohers raher han live in a house alone. In his case inflaion increases heir welfare, because i reduces he negaive effecs of borrowing consrains ha preven hem from smoohing heir consumpion hrough ime. When consrucion coss are 2 Wheher or no sharing is aracive will depend on he relaive uiliy of sharing compared o rening a whole house. The model was solved for several differen parameerisaions in which he uiliy from 11
lower, or social norms make i normal for young aduls o eiher live wih family or live by hemselves, inflaion lowers welfare by making i harder for young agens o buy heir firs homes. In he parameerisaions analysed in his paper, inflaion is on balance welfare enhancing because here are more agens who benefi from lower rens han here are agens who suffer from higher ineres paymens a he sar of he morgage. In Coleman (2008), he laer effec dominaed, so inflaion lowered welfare for mos people. This difference reflecs wo ses of parameer changes: he curren model allows agens o share housing wih each oher, raher han jus heir parens; and i has a higher house price o income raio, which reduces he araciveness of home ownership a young ages even when he inflaion rae is zero. In he real world, which of hese wo compeing effecs dominaes is an empirical maer. The answer will depend in par on he social mores and convenions of sociey, paricularly he accepabiliy of sharing housing wih non-family members. Inflaion causes one addiional welfare effec in he model: i changes he equilibrium number of houses and house prices, which changes he user coss of housing. If inflaion leads o an increase in oal housing demand, because lower rens enice adul children o leave home earlier, house prices will rise. This ends o lower he welfare of oher agens, because he user cos of housing rises and hese agens have less o spend on oher goods 3. This effec is an example of he negaive pecuniary exernaliy ha occurs when agens disregard he effec of heir acions on he prices paid by oher members of he economy. In conras, if oal housing demand falls in response o inflaion, house prices fall, and he welfare of oher agens increases. In he parameerisaions analysed in his paper, inflaion lowers oal housing demand, so here is a small posiive pecuniary exernaliy ha improves he welfare of all agens because of lower house prices. 3.2 The effec of capial gains axes wih an exempion for owneroccupied housing. Tables 2a and 2b show he long erm effecs of capial gains axes when he inflaion rae is wo percen, owner-occupied housing is exemp, and here is an elasic housing supply. The parameerisaions are consisen wih ables 1a and 1b, sharing half a house was eiher less han or more han half he uiliy from rening a whole house. In all of he parameerisaions analysed, an increase in inflaion led o a decline in he oal housing sock. 12
wih consrucion coss and house prices approximaely 25 percen higher in able 2a han in able 2b. The firs columns show he equilibrium values of various aspecs of he housing marke under curren ax regulaions. The second and hird columns shows how hese values change when capial gains are reaed as income and axed a marginal income ax raes (20%, 33%) and when capial gains are axes a a fla rae 20 percen rae. Noe ha mos landlords are high income agens, so he former scheme is effecively a fla rae capial gains ax wih a 33 percen ax rae. The fourh column shows how hese values change when he inflaion componen of ineres income is exemp from income ax. In each case, here is neiher income nor populaion growh and propery prices appreciae a he inflaion rae. Consequenly, when he inflaion rae is zero a capial gains ax has no effec. 3.2.1 Taxing capial gains on leased propery a marginal income ax raes The main effec of he (20%, 33%) capial gains ax is o increase rens by approximaely $1300 or 11 percen (able 2a) 4. There is also a small increase in house prices, by 0.6 0.8 percen. The laer occurs because he demand for propery increases: he increase in rens makes rening less aracive and here are a number of agens who cease living in shared renal accommodaion and purchase and live in a house by hemselves. The decline in rening is mos noiceable among older households. Among younger (cohor 0 and 1) households, rening only declines modesly, for credi consrains are sufficienly igh on mos low income agens ha rening is sill more aracive han home ownership, paricularly as mos of hese agens share renal housing and hus only experience half of he ren increase. In boh ables 2a and 2b, 2.5 3.5 percen of cohor 0 and cohor 1 cease rening. In conras, middle-aged and reired households almos compleely cease rening, because he capial gains ax raises he long run cos of rening above he cos of owning as only landlords pay he 3 Annual consumpion falls by approximaely he real ineres rae muliplied by he addiional housing cos. Higher house prices also lower he ne foreign asse posiion. 4 Small house prices are $225000, so he capial gain is $4500 when he inflaion rae is 2 percen; if he landlord has o pay 33 percen of his sum in ax, he ren has o be raised o make he same afer ax reurn as invesing in ineres earning asses. The $1300 increase in ren is no exacly equal o 0.33* $4500 for wo reasons. Firs, in he model he iming convenion is ha he landlord is paid ren and pays income ax a he sar of he period, bu pays capial gains ax a he end of he period. The aferincome-ax value of he $1300 ren increase is invesed for he lengh of he period (in his case 12.5 years); in his case he exra ineres is approximaely he same value as he income ax paid. Secondly, 13
capial gains ax. The oal effec of he capial gains ax on he renal marke depends on he number of people iniially rening, which depends on consrucion coss and house prices. When consrucion coss are high, a large number of agens ren under he curren ax rules and a capial gains ax reduces he fracion of agens rening by over six percenage poins; when hey are low, fewer middle aged and older agens ren, and he capial gains ax only reduces he fracion rening by hree percen. There are wo oher economic effecs. Firs, he simulaions indicae he capial gains ax raises lile revenue, for he GST rae only declines by 0.1 percenage poins. This is parly because he capial gains ax reduces he number of renal houses, and hus leads o a reducion in he income ax paid by landlords on heir renal income. Secondly, he simulaions sugges ha here is a small increase in he ne financial asse posiion of he economy. The amoun is larger when consrucion coss are high raher han low, and reflecs he increase in saving ha occurs as some households swich from being life-ime reners o middle-aged home-owners due o he increase in rens. 3.2.2 Taxing capial gains on leased propery a a consan 20% rae The resuls are qualiaively similar if he capial gains ax is applied a a fla 20 percen rae, excep he effecs are quaniaively smaller as he average ax rae is 20 percen raher han 33 percen. The increase in rens is only 60 percen as large, and here is a correspondingly smaller decrease in he fracion of agens ha ren and he fracion of houses ha are leased. Ineresingly, more revenue is raised under he lower fla rae capial gains ax when consrucion coss are high, as here is a much smaller decline in rening. Noneheless, he amoun of ax raised under eiher capial gains ax regime is small and he reducion in GST is less han 0.2 percen in eiher case. 3.2.3 Taxing capial gains on leased propery: resuls for oher supply funcions When he supply of small house is elasic bu he supply of large houses is inelasic, he resuls are similar o he case when boh he supply curves are elasic. Once again, he primary effec of he capial gains ax is o increase rens, reduce he propery prices increase by approximaely 1 percen once he CGT is inroduced, leading o a 1% or $100 increase in rens. 14
fracion of he populaion ha is rening and he fracion of he housing sock ha is leased, and increase he oal housing sock. When boh supply curves are inelasic, he resuls are a lile differen. In able 2c, he number of flas and houses is consan and, because here is an overall shorage, prices are high. As before, a capial gains ax raises rens and reduces he number of people rening, paricularly amongs hose who are middle aged or reired. However, house prices increase quie sharply, by four or five percen. The house price increase is needed o reduce he oal demand for housing, because an increase in rens wihou an increase in house prices leads o a reducion in he number of agens sharing renal accommodaion and an increase in he oal demand for housing. The only way o reduce he oal demand for housing is o raise house prices, and make i aracive o share. When he supply of housing is elasic, he price rise is no necessary, as new houses are buil o mee he addiional demand. If a rise in rens lead o a reducion in oal demand, because some young agens responded o he increase in rens by moving back o heir parens home, a capial gains ax could lead o a fall in house prices as well as an increase in rens. Consequenly, he way prices would behave in New Zealand if a capial gains ax were inroduced will depend on he size of he elasiciy of oal demand for housing o ren. 3.3 Exemping he inflaion componen of ineres from income ax Coleman (2008) argued ha he ax sysem would have less effec on he housing marke if he inflaion componen of ineres income were exemp from income ax. He argued ha by exemping he inflaion componen of ineres income from ax, and by only allowing he deducion of real raher han nominal ineres paymens, landlords would have less incenive o ener he propery marke when he inflaion rae was posiive, raising rens and home ownership raes. The fourh columns of ables 2a 2c show wha happens if neiher capial gains nor he inflaion componen of ineres income were axed. In all of he housing supply versions considered, he effecs on rens, prices, and home ownership raes are similar o wha happens if a fla 20 percen capial gains ax regime wih an exempion for owner-occupiers were inroduced. There are wo differences, however. Because ax revenue declines slighly when he inflaion componen of 15
ineres income is exemped from income ax, he GST rae has o be increased slighly, raher han cu. The increase is always less han 0.2 percenage poins, however, parly because he loss of ax on ineres income is offse by a reducion in he deducions allowable agains renal income. Secondly, here is a larger increase in he ne financial asse posiion han in eiher of he capial gains ax regimes considered. This is because exemping he inflaion componen of ineres income from ax raises afer ax real ineres raes, encouraging saving and capial accumulaion among working age agens 5. Since a capial gains ax does no affec afer ax ineres raes, afer ax reurns are higher when he inflaion componen of ineres is ax exemp han when a capial gains ax is inroduced. 3.4 The effec of an accrual capial gains regime applied o all agens Table 3 shows wha happens o he housing marke if capial gains axes are applied o all households on an accruals basis, eiher as a fla rae (20%) or a marginal income ax raes (20%, 33%).The resuls are for he case ha he housing supply is elasic and consrucion coss are high, bu he resuls for oher housing supply parameers are qualiaively similar. The capial gains ax leads o an increase in rens. When he capial gains ax rae is a fla 20 percen, here is lile effec on he quaniy of housing rened, however, because owner-occupiers are also liable for capial gains ax, so he cos of owning a house rises by a similar amoun. There is a significan swich from large houses o small houses, however, as he capial gains ax raises he user cos of large houses by more han he user cos of small houses. This reduces he demand for large houses a all ages. The subsiuion beween large and small houses also occurs because of a sizeable drop in he GST rae ha makes he consumpion of goods relaively more aracive han he consumpion of housing. The capial gains ax revenue is much larger han when owner-occupied housing is exemp, and he GST rae declines by over wo percenage poins raher han 0.2 percenage poins. When capial gains are axed as income a marginal income ax raes, here is a subsanial decline in he number of middle-aged and reired households rening. 5 While he model has a seady sae saving rae of zero, as people run down he asses hey accumulae while working when hey are reired, he economy s ne asse posiion increases when he saving rae among working age people increases. 16
This is because rens increase by more han he amoun of capial gains ax ha would be paid by low income reners, so i is cheaper in he long erm for low income households o purchase raher han ren. Indeed, he reducion in he number of households rening is similar o when owner-occupied housing is exemp from capial gains ax. In his model, he effecs of a capial gains ax applied o all households are very similar o he effecs of a fla rae propery ax. This is because propery prices increase a he inflaion rae, so a fla rae capial gains end up axing houses a a rae ha is proporional o value. Coleman and Grimes (2009) discuss he effecs of inroducing a propery ax a greaer lengh. 3.5 The welfare implicaions of capial gains axes The above analysis suggess ha capial gains axes raise rens, increase home-ownership raes, cause a subsiuion owards smaller houses, and improve he ne foreign asse posiion because hey reduce he disorions caused by he ineracion of inflaion wih he ax sysem. Wheher capial gains axes raise welfare, however, will depend on wo hings: he exen o which inflaion reduces welfare, because of is negaive effecs on credi-consrained owner-occupiers; and/or he exen o which inflaion enhances welfare because of is posiive effecs on crediconsrained reners. In he parameerisaions of he model sudied in his paper, he welfare losses o he reners exceed he benefis o he owners, for here are more young reners han young owners. In his case, a capial gains ax will end o have negaive welfare effecs as i raises rens, alhough his need no be he case. Figure 1 shows how differen ax schemes affec lifeime welfare for people wih differen income levels when he supply of housing is elasic and consrucion coss are low. The figure shows he average change in uiliy for each income decile 6. Three poins sand ou. Firs, a capial gains ax scheme ha exemps owner-occupied housing has lower welfare for mos people han one which does no. 6 The effecs on he welfare of he lowes decile are no shown as hey largely reflec he inheriance arrangemens in he economy. In hese simulaions odd numbered agens receive no inheriance, bu even number agens inheri he houses of he wo reired agens wih he same rank in he income disribuion. When here are no capial gains axes, mos decile one agens ren hroughou heir lives and neiher leave an inheriance nor receive one. When a capial gains ax is inroduced, many of hese agens buy a house, and bequeah i in old age. The logic of he model means ha even-numbered 17
Secondly, a capial gains ax scheme ha exemps owner-occupied housing reduces welfare for mos low income people, because of he increase in rens. Thirdly, a capial gains ax scheme ha exemps owner-occupied housing has similar welfare properies as a ax scheme ha exemps he inflaion componen of ineres income from ax. These hree resuls occurred in mos of he parameerisaions sudied, even hough he exac naure of he welfare changes depends on a number of facors such as he housing supply elasiciies and he way people inheri propery. Several feaures of hese resuls are of ineres. Firs, a capial gains ax scheme wihou an exempion for owner-occupied housing has very similar properies o he fla rae propery ax scheme analysed by Coleman and Grimes (2009). As discussed above, his is no surprising, for in he model axing capial gains on an accrual basis when he inflaion rae is consan is like having a fla rae propery ax. Since he welfare effecs are similar, and since he effecs on rens, prices, and home-ownership raes are similar, a fla rae propery ax could be a subsiue for a capial gains ax if i were believed he inflaion rae would coninue o be low and sable. Given he poliical difficulies of inroducing a capial gains ax on owner-occupied residenial housing in oher counries, a fla rae propery ax may be an aracive opion. Secondly, he welfare properies of capial gains regimes ha do or do no exemp owner-occupied housing are significanly differen, even hough hey have a similar effec on rens. The differences are caused by wo facors. Firs, when owneroccupied housing is axed, much more ax is colleced. The consequen cu in he GST rae parially compensaes reners for he rise in rens, and leads direcly o an improvemen in heir welfare. In addiion, here is a reducion in he oal demand for propery, so house prices fall relaive o he case ha only landlords pay he ax. This leads o a reducion in he direc user cos of housing o all agens. This provides a gain o all agens in he economy excep he firs generaion, who suffer a capial loss. Thirdly, a ax regime ha exemps he inflaion componen of ineres income from ax has similar welfare properies as a ax regime ha axes capial gains agens also inheri one or more houses, and are much beer off. Alhough his effec dominaes he welfare calculaions for he lowes income decile, This resul is no emphasised in his paper. 18
ax on leased residenial propery. Again, i may be poliically easier o inroduce such a ax regime han a capial gains ax. Figure 2 shows he welfare effecs of differen ax regimes when he supply of housing is elasic bu consrucion coss are high. The resuls are similar, alhough here is a downwards spike in he 6h decile ha reflecs he effec of inheriance arrangemens. (In his case i reflecs he change ha occurs when people sar inheriing large houses raher han smaller houses.) Once again, he welfare consequences of a capial gains ax ha includes owner-occupiers are beer han a ax ha does no; he welfare consequences of capial gains ax regimes ha exclude owner-occupiers are negaive for low-income agens because he axes increase rens; and he welfare consequences of capial gains axes ha exemp owner-occupiers are similar o he welfare consequences of ax regimes ha exemp he ineres componen of ineres income from ax. The welfare losses for low income agens are higher when consrucion coss are high, parly because more people ren bu also because rens are higher and hus increases in rens cause more severe cus in consumpion. 4 Discussion and conclusions This paper has explored some of he consequences of inroducing a capial gains ax on residenial propery in New Zealand. I has done his in he conex of a sylised model ha aemps o undersand he facors ha deermine housing marke oucomes in he long erm. The model focuses on hree main facors: he cos of supplying new housing; he financial incenives facing landlords; and he ax and financial incenives facing households as hey choose differen housing opions over he course of heir lives. The model suggess ha a capial gains ax will have he following effecs: i will lead o an increase in rens; i will lead o a reducion in he number of people rening, and an increase in homeownership raes; i will lead o an increase in he ne foreign asse posiion; and i will lead o a decline in he fracion of large houses in he economy. I is possible ha homeownership raes could rise by several percen if a capial gains ax were inroduced, wih a similar sized increase in he ne foreign asse posiion. 19
Two oher resuls seem general. Firs, a capial gains ax ha exemped owner-occupied housing would raise lile revenue, whereas one ha applied o all households would raise enough o allow a sizeable reducion in he GST rae. For his reason, low income households ha ren are beer off when a capial gains ax does no have exempions. Secondly, he increase in rens and he increase in homeownership raes will be larger if he capial gains ax raes on owner-occupied residenial propery are lower han hose on leased residenial propery, eiher because he former is specifically exemped from capial gains ax or because landlords ypically have higher marginal ax raes han households who ypically ren. Beyond hese general oucomes, he paper demonsraes ha he welfare implicaions of a capial gains ax depend a lo on he deailed srucure of he economy. I maers wheher he supply of housing is elasic or inelasic. I maers wheher consrucion coss are high or low. I maers wheher people prefer o own raher han ren. I maers wheher young people respond o ren increases by sharing wih more people, or by deciding o buy a house hemselves. Indeed, some of hese facors maer so much ha hey deermine wheher a capial gains ax is largely beneficial or harmful. I is boh a weakness and a srengh of he modelling approach ha i canno be more definiive abou he welfare effecs of a capial gains ax. From a echnical perspecive, he weakness is clear: a model ha delivers differen answers when he housing choice se is srucured differenly makes i difficul o know wheher he model s oucomes are robus or conrived. The srengh is more suble: he modelling approach suggess ha he welfare effec of differen policies depend a lo on several deep parameers in he models, suggesing empirical research on he naure of hese parameers is imporan before policies are inroduced. This paper has ignored many of he pracical and poliical issues ha would have o be solved if capial gains axes were o be inroduced. While he simulaions of he model sugges a capial gains ax ha includes owner-occupied housing has beer welfare properies han a capial gains ax ha does no, he poliical and pracical difficulies of inroducing an accruals based capial gains ax should no be underesimaed. Applying a capial gains ax only o realised gains has is own problems, noably he incenives i generaes o remain in unsuiable houses or living arrangemens in order o avoid he ax. Ye he simulaions also sugges ha 20
a fla rae propery ax has many of he same properies as an accrual based capial gains ax wih no exempions, and if a capial gains ax is desired bu no considered pracical his may be a suiable alernaive. The similariy beween hese wo axes will be greaer if nominal propery price appreciaion is dominaed by inflaion raher han real facors, and if he inflaion rae is relaively sable. Mos OECD counries ha have capial gains axes exemp owneroccupied housing from he ax and only ax leased residenial propery when a sale is realised. This is a much more sraighforward ax o implemen han an accruals based ax, bu sill removes some of he housing marke disorions ha arise from axing differenly ha inflaion componen of ineres earnings and he inflaion componen of capial gains. Noneheless, he simulaions sugges he effecs of his ype of capial gains ax could be largely replicaed by exemping he inflaion componen of ineres income from ax, a sraegy ha may be easier o implemen in pracice. Such a sraegy would have he added advanage ha real afer-ax ineres raes and reurns o capial are unaffeced by he inflaion rae. The key issue underlying he whole paper is wheher he effecs of moderae inflaion on he housing marke largely lowers or improves welfare. In line wih earlier work, his paper idenifies wo ways ha inflaion affecs welfare. Firs, inflaion makes i more difficul for people o purchase a house, or upgrade o a bigger house, because nominal ineres raes increase and banks do no change heir lending crieria o recognise he way inflaion erodes he real value of he exising deb. This is he familiar issue of morgage il, which lowers welfare (Modigliani, 1976.) Secondly, inflaion leads o lower rens if ineres earning asses are axed more heavily han capial gains. This improves he welfare of hose who are credi consrained and ren. Wheher he effec of inflaion on he housing marke improves or lowers welfare herefore depends on he fracions of he populaion who find i eases raher ighens he credi consrains hey face. In his paper, I have focussed on parameerisaions in which inflaion raises welfare for many agens, while in earlier work I focused on he case ha inflaion lowered welfare. In pracice, his is an empirical quesion, he answer of which will depend on he cos of housing and he way agens value rening and home-ownership. If, under he curren ax sysem, inflaion lowers rens and raises he welfare of many people, policies ha counerac he effecs of inflaion on he housing marke will end o lower welfare. 21