ECONOMICS SERIES SWP 2009/22. An SVAR Analysis of Monetary Policy Dynamics and Housing Market Responses in Australia

Transcription

1 Faculy of Business and Law School of Accouning, Economics and Finance ECONOMICS SERIES SWP 2009/22 An SVAR Analysis of Moneary Policy Dynamics and Housing Marke Responses in Ausralia IKM Mokharul Wadud Omar HMN Bashar Huson Joher Ali Ahmed The working papers are a series of manuscrips in heir draf form. Please do no quoe wihou obaining he auhor s consen as hese works are in heir draf form. The views expressed in his paper are hose of he auhor and no necessarily endorsed by he School or IBISWorld Py Ld.

2 An SVAR Analysis of Moneary Policy Dynamics and Housing Marke Responses in Ausralia IKM Mokharul Wadud Omar HMN Bashar Huson Joher Ali Ahmed Absrac: This paper examines he impac of moneary policy and a range of secor-specific and macroeconomic shocks on he Ausralian housing marke using quarerly daa for a period of The paper develops a srucural vecor auoregressive (SVAR) model based on conemporaneous resricions o analyse he dynamics of hese shocks. The resuls indicae ha supply of new houses in Ausralia rises wih higher real house prices; and ha house prices rise and fall wih higher inflaion rae and ineres rae, respecively. Dynamics of he impulse responses reveal significan effec of moneary policy on new house consrucions, real house prices, maerial coss and inflaion. Resuls also sugges ha housing oupu, real house prices and ineres raes respond significanly o shocks o housing supply, housing demand and o a number of oher variables. These resuls are expeced o shed some lighs on he curren policy environmen peraining o he Ausralian housing secor. Keywords: Moneary ransmission, Housing marke, Srucural VAR JEL Classificaion: R31, E52, E62, C51 School of Accouning, Economics and Finance, Deakin Universiy, Pigdons Rd., Waurn Ponds, Geelong, Vicoria 3217, Ausralia. Tel , Mokharul.wadud@deakin.edu.au Corresponding auhor. School of Accouning, Economics and Finance, Deakin Universiy, Pigdons Rd., Waurn Ponds, Geelong, Vicoria 3217, Ausralia. Tel , Omar.bashar@deakin.edu.au School of Accouning, Economics and Finance, Deakin Universiy, Pigdons Rd., Waurn Ponds, Geelong, Vicoria 3217, Ausralia. Tel , huson.aliahmed@deakin.edu.au 1

3 1. Inroducion Moneary ransmission effec on he housing secor has been an ousanding issue and a major area of ineres o researchers in recen years. The role he housing secor in any economy and he welfare implicaions of improved housing affordabiliy have been well documened in he lieraure (Mishkin, 1996, 2007; Fry e al., 2009; Vargas-Silva, 2008; McCarhy and Peach, 2002). However, he effecs of policy sraegies on he housing marke do no seem o be readily predicable in view of differences in response paerns of various secor specific and macroeconomic variables. In paricular, he effecs of moneary policy on he housing marke have been subjec o scruiny due o he exisence of a number of direc and indirec ransmission channels (Mishkin, 2007). Housing is generally considered as more ineressensiive han he economy as a whole and he degree of such sensiiviies can vary hrough ime and across counries (Berger-Thomson and Ellis, 2004). Over he las few decades, he Ausralian housing secor has expanded significanly wih high price and demand growh, much unlike many oher developed economies of comparable socio-economic seing including he UK, Canada and he US. The consequences of high price growh in Ausralian housing secor have been observed in erms of declining age-specific home ownership raes for younger households from he 1970s and increased affordabiliy problems. 1 Yaes (2008) suggesed ha an effecive soluion o housing affordabiliy problems lie in addressing he underlying demand and supply facors as well as in direcing policies o increase he supply of affordable renal housing. To assis wih home ownership and o promoe residenial invesmen, he Ausralian Federal governmen and he sae governmens made concered effors in he form of concessions provided by he governmens and public expendiure o suppor housing. 2 While high house price growh and he governmen s fiscal measures o suppor he housing secor in Ausralia have been of some concerns, he moneary ransmission in Ausralian 1 From 1960 o 2006, real house prices increased a an average rae of 2.7 per cen per annum, ouweighing he 1.9 per cen per annum growh in per household real income (Yaes, 2007 & 2008). Oo (2007) suggesed ha he growh in real house prices could be aribued o he combined effec of he growh in per household real incomes and he growh in number of households. 2 Esimaes reveal ha he oal governmen expendiure on housing and communiy increased more han en imes in 2009 over is 1974 levels. Some of he major governmen policies included inroducion of a new capial gain ax in 1985, a concessionary arrangemen in he exising capial gain ax in 1999 and The firs home owners gran (FHOG) in July 2000, funded by boh he Federal and he sae governmens, o compensae for increases in building and housing coss associaed wih he implemenaion of he goods and services ax (GST) Despie he apparen appeal of he FHOG, he scheme has been criicised by some auhors as he scheme ends o increase he house prices and disor equiabiliy (Wood e al, 2006). 2

4 housing has been subjec o limied research. In view of he sraegic impac of moneary policy on house prices, affordabiliy of housing and housing oupu driven by he marke dynamics, much of he underlying behaviour of he Ausralian housing marke could be explained hrough he moneary policy ransmission mechanism. The Reserve Bank of Ausralia (RBA) pursued moneary conracion in he 1970s and in he 1980s, followed by phases of expansions beginning from he early 1990s. Over he 2000s, he RBA coninued is inflaion argeing sraegies wih a view o sabilising he economy. 3 For he Ausralian housing marke hough, while he reduced ineres raes lowered he coss of borrowing boosing demand; high growh of house prices adversely affeced he home buyers. In addiion, here seems o be perceived disrupion in he Ausralian moneary ransmission mechanism in recen years, wih he commercial banks inadequaely responding o he RBA s reduced ineres raes. However, he overall efficacies of moneary policies are discernible hrough he ransmission of he moneary shocks affecing a range of housing marke facors including demand, supply, maerial cos and prices (Mishkin, 2007). Despie a plehora of sudies available examining he housing marke response o a variey of fiscal measures (Wood e al., 2006; Wood, 1999; Yaes, 2008; Dvornak and Kohler, 2003), here seems o an absolue dearh of sudies addressing issues of moneary ransmission in Ausralian housing. A major conribuion o he moneary ransmission and housing lieraure in he Ausralian conex known o dae has been made by Fry e al. (2009). However, his sudy was limied o examining he wealh and moneary policy effec on he overvaluaion in Ausralian housing and equiy markes. In his paper, we aemp o exclusively analyse he Ausralian housing wih a broader se of objecives and by inegraing he moneary ransmission mechanism wih a number of housing marke and macroeconomic variables. The major objecives of he sudy are: Firsly, o examine he dynamic effecs of moneary policy on he house prices and housing oupu in Ausralia; Secondly, o analyse he major facors explaining housing demand and supply; Thirdly, o analyse he reacion of moneary policy o he housing marke shocks; and Fourhly, o examine how shocks o various macroeconomic variables affec house prices and housing oupu. Hence, his sudy 3 The RBA adoped is inflaion argeing policies in he 1990s during he moneary expansion of he decade. To weaher he impac of he global financial crisis, he RBA reduced is cash rae arge o 4.25 per cen by he end of 2008 and o 3.25 per cen in The RBA raised ineres raes from around 8 o 10 per cen in he 1970s o abou 18 per cen by he end of he 1980s. This was followed by moneary expansions as he RBA lowered ineres raes from abou 11 per cen in he early 1990s o abou 5 per cen by he end of he cenury. 3

5 essenially adops a more holisic approach compared o he Fry e al. (2009) sudy, by examining he dynamics of a number of secor specific, macroeconomic and policy variables using a simple housing marke model. We develop a srucural vecor auoregressive (SVAR) model for Ausralian housing and analyse he effec of moneary policy, inflaion and a number of oher marke facors. The approach enails an examinaion of he policy efficacies including hose revealed hrough he moneary ransmission channels as well as an analysis of he dynamic inerdependence and conemporaneous relaionships involving a range of macroeconomics and housing marke variables. In paricular, he model developed herein disenangles he effecs (exernal, demand and supply shock, moneary policy shock, fiscal policy shocks) in a srucural form wih an emphasis on analysing how moneary ransmission affecs house price, housing demand, housing supply, raw maerial cos, inflaion and exchange rae hrough various channels. The resuls indicae ha in he shor run he supply of new residenial houses increases wih he increase in real house prices. Also, he real house price increases wih higher inflaion rae, and falls wih higher ineres rae. Evidences from he impulse responses show ha a conracionary moneary shock significanly affecs number of new houses, real house prices, maerial coss, housing oupu, inflaion rae and ineres raes. Also, shocks in housing demand, housing supply, inflaion raes, maerial coss and exchange raes significanly affec housing oupu. Furher, real house prices respond significanly o shocks in real oupu, housing supply and housing demand. Significan responses of he ineres raes are also observed o shocks o inflaion rae, governmen expenses, housing demand, housing supply, maerial coss and foreign ineres. The res of he paper is srucured as follows: Secion Two discusses he heoreical and empirical lieraure on moneary ransmission effec on housing marke. The model based on a srucural VAR framework used in his sudy has been presened in Secion Three. Secion Four discusses he empirical findings and Secion Six concludes highlighing some implicaions of our resuls. 4

6 2. Moneary ransmission effec and he housing marke Mishkin (1996) provides a comprehensive analyical framework on how he effecs of moneary policy are ransmied hrough various channels. These channels show he direc effecs on he user cos on housing capial, expecaions of fuure house-price movemens, and housing supply; as well as he indirec impac on housing oupu hrough sandard wealh effecs, credi-channel effecs on consumer spending, and credi-channel effecs on housing demand. Figure 1 provides a diagrammaic represenaion of he direc and indirec ransmission channels of moneary policy affecing he housing marke. 4 One of he direc effecs of moneary policy is ransmied on o he user cos on housing, defined as an opporuniy cos of occupying a house raher han rening. If he opporuniy cos of owning a house is lower han rening, demand for housing aciviies is likely o decline. 5 A simple analogy ha can be drawn ou of his conex is ha when moneary policy raises he shor-erm rae, long-erm rae also ends o rise driven by expecaions of fuure increase in he shor-erm rae; and hence, he average morgage rae also raises leading o higher user cos of capial on housing (Figure 1). Therefore, here will be a fall in demand for housing, which in urn would lead o a decline in housing consrucion and prices hereby lowering aggregae demand in he economy (Mishkin, 2007). The expeced real rae of appreciaion in housing price provides anoher way for moneary policy o affec housing aciviy. As poined ou earlier, he ighened moneary policy sofens housing price because he demand for housing decline hrough user cos of ransmission. To illusrae furher, expecaion of ighening moneary policy may likely lead o much lower real rae of house price appreciaion hrough rising user cos of housing, subsequenly leading o lower demand for housing aciviies. Case and Shiller (2003) have emphasized ha changes in hese expecaions can have an effec on he user cos of capial and hus on housing demand. 4 Noe ha Figure 1 is a concise represenaion of he broad picure of moneary ransmissions mechanism as i shows he ransmission channels affecing he housing secor only and no he macroeconomic variables. 5 The user cos of housing is illusraed hrough he following equaion: UC = Ph [{(1-)i π e }-{ π e h -π e }+ δ], where Ph is he relaive purchase price of new house, {(1-)i π e } is he afer ax real morgage rae as a form of cos of borrowing, { π e h -π e } is real house price appreciaion, δ is he depreciaion rae for housing, i is he morgage rae, π e h is he expeced rae of appreciaion of housing price and π e represens he expeced rae of inflaion. Therefore, user cos of capial on housing depends on wo imporan parameers namely afer ax real ineres rae and expeced appreciaion of housing price. 5

7 Conracionary Moneary Policy (An increase in he Ineres rae) Raises Direc channel Indirec channel Reduces Reduces Raises Raises User Cos Housing Supply Housing Price Reduces Wealh Credi Consrain Reduces Reduces Demand for Housing Reduces Reduces Reduces Figure 1: Moneary Transmission Channels and he Housing Marke 6

8 From he supply side perspecive, any increase in shor erm ineres rae may have an immediae impac on housing consrucion cos, hereby housing oupu. Thus any increase in policy ineres rae is likely o influence housing aciviies. These ypes of argumens are suppored by McCarhy and Peach (2002). I is undersandable ha he fall in real house prices due o lower demand for housing as a resul of conracionary moneary policy will lead o a decline in asse (wealh) for individuals. The fall in individuals wealh (negaive wealh effec) would furher reduce housing demand leading o lower housing prices (Figure 1). In urn, individuals wealh oo declines lowering household consumpion and demand for housing. Households migh be credi consrained which may affec housing demand 6. Crediconsrained households are hose whose expendiure exceeds heir revenue. Therefore, any increase in shor-erm raes leads o ineres burden of he households and reduces heir cash flow. Mishkin (2007) suggess ha he relaionship leads o wo possible credi channels. The firs channel is hrough which he nominal, and no jus real, ineres rae can affec housing demand. Higher nominal raes, even when real ineress raes remain unchanged, reduce curren cash flow due o increase in morgage rae will lower afer ineres and ax income as resuls lowering cash flow. Hence subsequen decline in he demand for housing will be expeced. From credi channel perspecive, he cenral problem ha limis he demand and supply of housing is he resul of igh credi marke due o asymmeric informaion beween poenial house buyer and financial insiuion (Mishkin, 2007). Lenders are relucan o make loans because hey have difficulies idenifying he credi worhiness of he borrower. This can be analysed from wo perspecives. Firsly, he lenders are relucan o exend he loan o prospecive home loan borrower as hey lack informaion on credi worhiness of he borrower. Secondly, once loan is exended, he borrower migh underake aciviies ha may undermine he abiliy o repay he loan on ime. 6 For deails on credi channel and moneary policy shocks see Bernanke and Gerler (1995). 7

9 Review of he empirical sudies There has been a vas lieraure on modelling housing marke and various ransmission mechanisms, especially in he developed economies. For he US housing marke, a number of sudies examined he demand shocks (Jarocirinski and Smes, 2008) and supply shocks (Ludvigson e al., 2002 and Baffoe-Bonnie, 1998). In a recen sudy, Vargas-Silva (2008) examined he impac of moneary policy shocks on he US housing marke. Malpezzi and Maclennan (2001) esimaed he long-run price elasiciy of supply for new houses in he UK and he US. They esimaed ha in he pos-war period, he price elasiciy ranged beween six o 13 percen for he US and beween zero o one percen for he UK housing markes. Clearly, he shor run elasiciy of supply was lower for he UK housing marke, implying ha house prices are demand-driven in he shor run. Aoki e al. (2004) examined he responsiveness of UK house price o moneary policy shocks using VAR by including oupu, inflaion, oil prices, real broad money, shor erm ineres rae, consumer durable, and housing invesmen. Their findings sugges ha an increase in he shor erm rae by 50 basis poins leads o 0.8% fall in house price, and similar increase in shor rae leads o 0.1% fall in price for non-consumer durable. In a similar sudy, Iacoviello (2002) esimaed he srucural VAR model for 6 European counries using idenificaion of King e al. (1991) model. The sudy found ha he house prices in UK fall by up o 1.5% following a 50 basis poins ighening of moneary policy. However, he mos imporan facor ha was found o be affecing he house price was he aggregae demand shock. Iacoviello and Minei (2008) used a VAR model o examine he credi channel of moneary policy hrough he housing marke. In heir sudy, four differen vecor error correcion models (VECM) and a VAR for each European marke were used. They firs developed a VECM o idenify response o moneary policy shocks. The second VECM was used o look a he spread beween morgage rae and risk free rae; while he hird and he fourh models were used o examine he exernal finance mix of house loan (bank and non-bank). The findings suggesed ha he house price decline beween 0.7 o 1 per cen following 70 basis poin increase in shor ineres rae. In a more recen sudy, Elbourne (2008) adoped a srucural VAR model o undersand he effec of change in moneary policy on consumpion behaviour. 7 Using a wo-sep approach and a counerfacual simulaion, Elbourne (2008) 7 Elbourne (2008) used he model Kim and Roubini (2000) by including housing price. 8

10 suggesed ha abou 12-15% of he drop in consumpion on a conracionary moneary policy shock is channelled hrough changes in house prices. For he Ausralian housing secor, Abelson e al. (2005) developed and esimaed a long-run equilibrium model ha shows he real income deerminans of house prices and an asymmeric error correcion model o represen house price change in he shor run. In a relaed sudy, Dvornak and Kohler (2007) addressed he quesion of how changes in sock marke wealh and housing wealh affec consumpion expendiure in Ausralia. Their findings sugges ha housing wealh and sock marke wealh have a significan effec on consumpion. Oo (2007) examined he sandard economic facors o explain he growh of real house price in Ausralia and suggesed ha he variable morgage rae is one of he imporan variables ha affec he growh rae of house price of he capial ciies in Ausralia. As menioned earlier, he effec of moneary policy on housing marke has been examined by Fry e al. (2009). Using a SVAR model, heir esimaes involved idenifying he relaive effec of wealh and moneary policy causing overvaluaion in he Ausralian housing and equiy markes. Their findings sugges ha while he wealh effecs from porfolio shocks in equiy markes have been imporan drivers for overvaluaion in Ausralian housing, he effecs from moneary policy have been rivial. Their also found ha he some of he imporan facors driving he overvaluaion of Ausralian housing include he housing shocks arising from housing demand, he goods marke shocks arising from aggregae demand and he aggregae supply shocks. They showed ha he overvaluaion in pre-2006 period is dominaed by he housing demand shock while he overvaluaion in pos-2006 period is dominaed by goods marke shocks ransmiing hrough changes in aggregae demand and aggregae supply. In his paper, we have used a srucural VAR model of Ausralian housing marke wih a view o examining he effec of moneary policy hrough user cos, expeced price appreciaion, housing oupu, and credi channels. This paper also evaluaes he conemporaneous relaionship and shock ransmission mechanism involving a number of oher variables including housing demand, housing supply and coss, hereby shedding furher lighs on Ausralia s housing policy environmen. Hence, on empirical grounds, he scope and policy implicaions emanaing from he sudy are expeced o surpass hose of Fry e al. (2009). Furher, his sudy provides addiional insighs for he Ausralian housing marke compared o Elbourne s (2008) sudy. While Elbourne (2008) examined he price shocks in he UK housing marke, he marke forces behind such shocks were no idenified. In his sudy, price 9

11 shocks are clearly separaed o have originaed from demand and supply of Ausralian housing. 3. Economeric Modelling The srucural VAR modelling requires a pruden choice of variables for inclusion in he model reflecing he heories and he empirical objecives. Hence, in he VAR model we mainly included he variables ha play crucial role in Ausralian housing marke. In paricular, a demand and a supply equaion for Ausralian housing marke have been specified in order o deermine hese imporan variables. A Srucural VAR has he following general form: A0Y = A 1(L)Y +Bε [1] Here Y represens n-vecor relevan variables; A 0 and B are n x n marices; and A q i 1(L) A1iL i1 marix. represens marices polynomial in he lag operaor wih A1i being an n x n ε is an n-vecor of serially uncorrelaed, zero mean srucural shocks wih an ideniy conemporaneous covariance marix, E ε ε ] I. These srucural shocks, ε [ among ohers, include housing demand shock, housing supply shock and moneary policy shock. Provided ha A 0 is nonsingular, solving for represenaion: Y yields he reduced form of VAR Y = A A (L)Y + Α Bε or Y =C(L)Y +u [2] where 1 C(L) A 0 A1(L) and u Α 1 0 Bε 10

12 or Α0u = Bε [3] As he conemporaneous effecs are no direcly esimaed in reduced form VAR, hese effecs (if any) are supposed o be included in he VAR residual erms. Thus, he residual series described by he vecor u may be correlaed wih each oher due o he conemporaneous effec of he variables across equaions, alhough hey are sill presumed o be whie noise. Equaion [1] is he srucural model of he VAR. The reduced form represened in [2] is more familiar. The main echnique involved here is o esimae equaion [2] and recover he parameers and he srucural shocks, ε in [1] from hese esimaes. Having idenified he srucural shocks, we can hen find he impulse response of a variable o a one-ime shock o any variable included in he model. The main problem wih his approach is of idenificaion since he number of parameers ha need o be esimaed in he srucural model is larger han ha of he esimaed reduced form model. In order o solve he idenificaion problem, resricions are ypically imposed on he elemens of A 0 and/or B. In our analysis, we use he conemporaneous relaions in he Ausralian housing marke, he moneary policy reacion funcion and he domesic and exernal secors of he Ausralian economy in general o impose hese resricions. The housing marke model Based on he discussions in previous wo secions, we inroduce following simple model for housing in Ausralia: h mc rhp s 1 2 hs b [4.1] hs h ry rhp r fr rg ex b [4.2] d hd hd h s h [4.3] d where h = supply of housing, h = demand for housing, mc = housing maerial coss, s d rhp = real house price, ry = real GDP in Ausralia, r = shor erm nominal ineres rae, = inflaion rae, fr = foreign ineres rae, rg = real ne governmen spending for housing, and ex = nominal exchange rae agains he US dollar. Variables are expressed in VAR 11

13 residual form and all variables excep r,, and fr are in log form. 8 Here represen housing supply and demand shocks, respecively. hs and hd s d Re-wriing he equilibrium condiion in [4.3] as h h h and invering [4.2], we may rewrie he housing marke equaions as, h mc hpr b [5.1] hs 1 2 hs hd rhp 1fr 2ry 3 4rg 5r 6h 7ex bhd [5.2] In he above model, he firs equaion posulaes ha real house price and housing raw maerial coss are he main deerminans of housing supply in Ausralia. Housing demand, on he oher hand, is assumed o be deermined by he real income level, real house price, domesic and foreign ineres raes, he inflaion rae, governmen s ne spending on housing and he exchange rae. A higher maerial cos is expeced o reduce he housing supply, whereas an increase in real value of houses would increase he supply of new houses. A higher income level would increase he housing demand, whereas an increase in house price would decrease he demand for housing. The effec of changes in he ineres rae on housing demand may no be prediced direcly. While higher ineres rae is expeced o reduce housing demand; demand may acually increase during he impac period if people perceive his increase in he ineres rae as a signal for furher increase in he rae in he fuure. Similarly, effec of foreign ineres rae on housing demand is indeerminae. An increase in he foreign ineres rae increases he reurns on invesmen in alernaive asses, which may lead o a reducion in domesic housing demand. However, foreign residens may find Ausralia a more aracive place o buy a house if he Ausralian ineres rae is lower compared o he foreign ineres rae, in which case, domesic demand for housing may increase. A higher level of Gov. Spending on housing is expeced o increase housing demand. Like demand for any oher goods, housing demand is expeced o decline when inflaion goes up. Currency appreciaion reduces he prices of impored goods, which may increase consumers purchasing power and hence demand for housing. 8 These reduced form variables essenially form he vecor u. Deails of each of he variables are presened in secion

14 The SVAR Sysem and Daa Descripion The above model of housing marke assumes ha variables ha influence housing demand and supply are exogenous o h and rhp. Bu as a maer of fac, hese variables may also be affeced by housing supply and demand. Also noe ha he above equaions are in residual form where he lag effecs were removed from he equaions. In order o explore he dynamics of he relaionships and o capure he endogeneiy of he variables, here we have inroduced a Srucural VAR model. In he VAR sysem we have included he variables ha are imporan for he Ausralian housing marke. As suggesed by [5.1] and [5.2], he srucural VAR sysem includes following nine variables 9 : RY, Π, RG, R, MC, H, RHP, EX, FR where RY = Real GDP (a prices) Π = Quarerly CPI inflaion RG = Real Gov. Expendiure on housing and communiy (a prices) R = 90-day Bill Rae (quarerly average) MC = An index for cos of maerials used in housing ( =100) H = Number of privae secor housing approvals. RHP = Real house Price Index adjused for CPI ( = 100) 10 EX = Exchange Rae in erms of he US dollar per uni of Ausralian Dollar. FR = he US Federal Funds Rae All variables excep for Π, R and FR are expressed in naural logarihms. Quarerly daa covering he period 1974Q2 2008Q4 have been used for he VAR esimaion. The Reserve Bank of Ausralia (RBA) and Ausralian Bureau of Saisics (ABS) were he main sources of daa. All daa are seasonally adjused. Wherever possible, seasonally adjused daa have been colleced from he main source. We have used Census X-12 mehod on raw daa ha were no seasonally adjused in he main source. Following sandard pracice we include he variables in heir levels (or log-levels) in he VAR model despie he fac ha mos of hese variables are nonsaionary. The main problem wih 9 We have used he capial leers o denoe he variables. Noe ha in Equaions , we used small leers o represen he reduced forms of hese variables. 10 The price index of esablished homes was only available from he ABS sources from he Q2, For period prior o 1986, we used he CPI housing figures. 13

15 using nonsaionary variables in a VAR model is ha he sandard errors of he coefficiens are no esimaed precisely, however, VAR coefficien esimaes will sill be unbiased. As inference of he reduced form VAR coefficiens is no our prime objecive, and given he usefulness of he VAR coefficiens due o heir unbiasedness, here is no poin ransforming he variables o induce saionariy. The whole VAR sysem can be decomposed ino four separae blocks; namely, he domesic goods marke block, he policy block, he housing marke block and he exernal block. Goods marke equaions are represened by equaions of RY and. Policy equaions are described by RG and R, as hey are direcly affeced by policy acions of fiscal and moneary auhoriies, respecively. The housing marke block comprises H, RHP and MC. Finally, FR and EX are included o represen he exernal effecs. For choosing he lag order of he VAR sysem, we primarily looked a crieria like Akaike Informaion Crierion (AIC), Final Predicion Error (FPE), Hannan-Quinn Crierion (HQC) and Schwarz Crierion (SC). For a maximum of 10 lags, AIC suggess 10 lags, FPE suggess 2 lags, and HQC and SC sugges a lag order of one. We have chosen a lag order of wo, as his was found o be he minimum lag a which residuals in each equaion saisfied he whie noise propery. As Ausralia is a small open economy, i is unlikely ha an Ausralian variable would affec he US ineres rae. Considering his, we included only he own lags in he equaion for FR. The Feasible Generalized Leas Squares mehod has been used o esimae his VAR sysem wih asymmeric lag orders. Idenificaion Once we esimae he VAR sysem, he nex sep is o impose resricions on he conemporaneous relaions beween he reduced form VAR residuals and he srucural shocks (as described by [3]). Our idenificaion scheme is based on equaion [5.1] [5.2] as well as moneary policy reacion funcion and he srucure of he Ausralian economy. This can be represened in marix form as: 14

16 1 fr x ε x 1 ry x ε x x 1 x ε x x x 1 rg x ε x x x 1 x r x ε x x x x 1 x mc x ε h x ε rhp x ε x x x x x x 1 ex x ε fr ry π rg r mc hs hd ex [6] As menioned before, variables in he lef hand side of he sysem represen relevan residuals obained in he reduced form VAR (reduced form represenaion of he variables). On he righ hand side of he sysem, he nine srucural innovaions ( ε' s ) basically represen shocks o he foreign ineres rae, domesic oupu, inflaion, gov. spending on housing, moneary policy, housing maerial coss, housing supply, housing demand and he exchange rae, respecively. In he above sysem of equaions, he sevenh and eighh equaions simply represen he conemporaneous relaions in he housing marke described by [5.1]-[5.2]. For oher equaions we use x o denoe he coefficiens o be esimaed. The firs four variables are assumed o affec housing demand and/or supply, wihou any feedback effec, i.e., hese variables are no assumed o be affeced by housing demand or supply. These four variables are arranged recursively wih he order appeared in [6]. We experimened wih alernaive orderings for hese four variables; however, he resuls were largely similar o wha we obained and repored in his paper. Noe ha we have no aemped o disenangle he aggregae demand and supply shocks, as his is no consisen wih our main objecive. The specificaion of he fifh equaion deserves some explanaion, as his represens he moneary policy reacion funcion. Here we have assumed ha he Ausralian Cenral Bank reacs (by changing he ineres rae wihin he same quarer) o he movemens in inflaion and o he changes in real oupu level. I is also assumed ha he domesic ineres rae is affeced conemporaneously by he foreign ineres rae and he exchange rae. To mainain he ineres pariy, domesic ineres rae may move in he same direcion of he movemen in he foreign ineres rae wihin he same quarer. Changes in he exchange rae may affec he 15

17 fuure expeced exchange rae, which in urn may conemporaneously affec he domesic ineres rae. The sixh equaion shows ha he raw maerial coss in he housing marke depend conemporaneously on domesic and foreign ineres raes, real aggregae oupu level, inflaion and he exchange rae. The las equaion exhibis ha he exchange rae depends on all he variables included in he model excep house price and raw maerial coss. 4. Resuls The esimaed coefficiens of he housing marke model presened by (5.1) and (5.2) are repored in Table 1. Noe ha he relevan sandard errors are given in parenheses. The esimaed parameers in [6] are presened in Table A.1 in he Appendix. Table1: Esimaed Coefficiens in he Housing Marke Model h RHP (0.8443) * (0.6830) (0.1548) (0.1770) ** (0.2450) (0.0043) * (0.0023) (0.0868) (0.0352) b hs ** (0.0034) b hd ** (0.0023) Noe: Figures in parenheses are he relevan sandard errors. ** Significan a he 1 percen level. * Significance a he 5 percen level. 16

18 In Table 1, he coefficiens for real house price in he firs equaion; and for he domesic ineres rae and inflaion rae in he second equaion are found saisically significan. I is hus eviden ha in he shor run real house price is he main deermining facor for he number of new houses buil in Ausralia. The posiive significan value of 2 confirms ha he supply of new residenial houses increases wih he increase in real value of houses in he impac period. The maerial coss appear o exer expeced negaive impac on he housing supply. However, such effecs are very insignifican, which implies ha changes in maerial coss would no affec he supply of new houses in Ausralia. While his resul is a bi anomalous on heoreical grounds, i is perceivable ha empirically, growh in he consrucion of he new houses in Ausralia has no relied significanly on changes in he maerial coss, hereby limiing shifing of he housing supply funcion. The esimaes of he equaion of he real house price show ha he real house price increases wih he inflaion rae, which suggess ha nominal house price increases by more han proporionaely compared o an increase in he general price level. The real house price decreases wih an increase in he ineres rae significanly in he conemporaneous period. This finding suggess ha a conracionary moneary policy may have immediae dampening effec on housing demand. Noe ha he esimaed parameers of boh he housing supply and demand shocks are highly significan. However, he housing supply shock is found more volaile han he housing demand shock, as he sandard deviaion of he housing supply shock (0.0496) is found much larger han he sandard deviaion of he housing demand shock (0.0170). In order o examine he dynamic changes in he number of new house approvals and real house price in response o various macro shocks including he moneary policy shocks, we have esimaed various impulse response funcions, which are discussed in he following secions. These impulse response funcions have been grouped based on shock (for moneary policy) and impulses (for oher variables). The Moneary policy shocks We firs examine he impulse responses due o moneary policy shocks, since such shocks are of significan ineres o assess policy dynamics in he housing marke. Figure 2 shows he impulse response funcions of he ineres rae shocks (conracionary moneary policy) on he 17

19 range of variables affecing he Ausralian housing marke. As shown in he figure, a conracionary moneary shock immediaely increases he number of new houses followed by a gradual decline abou a year or so, lasing for a couple of years before he effec fades away. Figure 2: Effecs of Conracionary Moneary Policy ε r H ε r RHP ε r MC ε r RY ε r π ε r EX ε r RG ε r R Noe: The doed lines represen he upper and he lower bands of 95% boosrapped confidence inervals compued using 1000 repeiions of Hall s percenile mehod. Noe ha despie heir shor lived naure, boh hese flucuaions are significan. The iniial increase in he number of new houses could reflec he effec of expecaions of furher moneary conracion, posiive wealh effecs due o price increases wih higher demand, 18

20 ongoing building conracs ec., which primarily imply increased housing oupu along he housing supply funcion. Once he impacs of hese facors are realised, he housing oupu declines significanly, as an anicipaed oucome of he moneary conracion. For similar reasons, moneary shock causes an insananeous and significan increase in he real house prices wihin a year before he prices decline (Figure 2). For he UK housing marke, Elbourne (2008) found ha real house prices fall immediaely due o moneary policy shock. Alhough he iniial response of he real house prices in Ausralia conradics wih ha of he UK, i is imporan o realise ha increased house prices is ypical of Ausralian housing marke and any insananeous increase in prices could be affeced by expecaions on furher increases in ineres raes, which in urn would pu upward pressure on he real house prices. Over a longer ime frame hough, precisely afer he 8h quarer following he shock, as shown in Figure 2, he real house prices in Ausralia end o decline, presumably reflecing lower housing demand due o he user cos ransmission mechanism (Case and Shiller, 2003). The coss of maerials also increase significanly wih moneary policy shocks beween four o eigh quarers prior o heir dissipaions. Much of he rise in he maerial coss could be aribuable o increased renal coss of consrucion equipmens and reduced supply by firms producing building maerials promped by fall in invesmens. The responses of real oupu due o conracionary moneary policy shock are also well pronounced in Figure 2, as real oupu declines significanly wih reduced aggregaed demand lasing for abou hree years. In line wih hese oupu changes, he inflaion rae also declines significanly beween wo o four years responding o he moneary conracion. The response of he exchange raes follows a prediced paern, wih appreciaion in he Ausralian dollars due o increased ineres raes and resulan capial inflow from overseas. While he exchange rae appreciaions gradually dissipae, he process susains for a prolonged period of abou five years (Figure 2). Elbourne (2008) found similar response paerns of exchange rae o moneary shock for UK housing marke. The effecs of governmen expendiure on housing and communiy seem o remain rivial for a couple of years afer which such responses are significanly negaive and slowly end o rever o neuraliy over he nex six o seven years. Essenially, his phenomenon indicaes he decline in governmen expenses aligned wih moneary conracions and he ransmied reducion in aggregae demand and economic growh, which ends o erode away over ime. Lasly, he effec of moneary conracion on ineres raes seems o be mosly significan. The immediae response of he ineres raes is posiive wih he raes rising for more han a year following he policy shock. The response 19

21 evenually becomes negaive and gradually weakens over years, much alike he response paern of he public expendiure on housing. Figure 3: Impulse Responses of Housing Oupu ε ry H ε π H ε rg H ε s H ε d H ε mc H ε ex H ε fr H Noe: The doed lines represen he upper and he lower bands of 95% boosrapped confidence inervals compued using 1000 repeiions of Hall s percenile mehod. 20

22 The dynamics of he housing oupu, real house prices and ineres raes We will now examine how he oher shocks may affec housing oupu, real house prices and he ineres raes. Impulse response funcions of he number of new houses o various shocks oher han he moneary shocks are presened in Figure 3. I is observed ha he oupu shock and he Gov. spending shock have insignifican effecs on he number of new houses buil. Boh housing demand and supply shocks have an immediae posiive impac on he housing oupu and he significan posiive effecs las for abou four quarers. I is eviden from he figure ha hese wo shocks have only emporary impacs on housing consrucion. A oneime posiive shock o inflaion sars o influence he housing oupu adversely in he second quarer of he shock and he significan negaive effec persiss unil he end of he 7 h quarer. Alhough we didn explicily idenify aggregae demand and supply shocks, i is apparen ha he inflaion shock we specify here can be comparable o an adverse supply shock. An adverse supply shock will reduce he real economic aciviy including he building of new houses in he economy 11. I is ineresing o observe ha a posiive shock o housing maerial coss increases he number of new houses significanly afer wo quarers of he shock and he effecs remain significan for abou five more quarers. One explanaion for his is ha higher maerial coss may reflec higher qualiy maerials used in he housing consrucion. The exchange rae shock appears o have a brief significan effec on housing consrucion. An appreciaion of Ausralian currency makes he impored housing maerials of same qualiy cheaper, which may in urn increase he number of new houses buil. Lasly, housing oupu responses o shocks in foreign ineres raes seems o be insignifican (Fig. 2). Figure-3 gives impulse responses of real house price o he various shocks. Wih a oneime shock in real oupu, he real house price increases immediaely. This growh is significanly posiive afer he 8 h quarer and sabilises a around 1 per cen afer abou hree years. Such responses appear o be consisen wih heories, since increased real oupu induces o higher 11 In appendix A.1, Figure A.1a shows he impulse response funcions of real oupu o various shocks. The negaive response of oupu o an inflaion shock confirms he idea ha he idenified inflaion shock is dominaed by he adverse supply shocks. 21

23 Figure 4: Responses of real house prices o various shocks ε ry RHP ε π RHP ε rg RHP ε s RHP ε d RHP ε mc RHP ε ex RHP ε fr RHP Noe: The doed lines represen he upper and he lower bands of 95% boosrapped confidence inervals compued using 1000 repeiions of Halls percenile mehod. consumpion demand (e.g., spending for housing) leading o increased real house price. The inflaion shock, on he oher hand, has negaive and insignifican impac on real house price. As we know ha higher price level or expecaion abou higher fuure price levels raises ineres rae. This may lead o lower purchasing power, and as a consequence, demand for housing will decline puing downward pressure on he real house prices. 22

24 Figure 4 also shows ha he fiscal policy shock represened by he shock o he governmen expendiure exers negligible impac on he real house prices. The governmen spending on housing and communiy creae greaer demand for housing, and hence may pu upward pressure on he real price. We documen a posiive impac on real house price hough i is no saisically significan. The housing supply shock has insignifican and negaive impac on real house prices. However, real house prices end o decline significanly afer abou four years. I is perceivable ha increased housing supply hrough ouward shif in supply curve for housing pus downward pressure for real price of housing. Conversely, he demand shock significanly raises he real house price, which lass for abou 12 quarers. A cursory look a he graph reveals ha he magniude of he demand shock on he house prices is raher large compared o he oher shocks, wih he rae of price growh rising close o 3 per cen by he 4 h quarer. Maerial cos is expeced o be a significan price deerminan, as per sandard heoreical predicions. Our findings sugges ha here is a posiive impac of maerial cos shocks on real house price up o quarer 12. However, such responses of he house prices appear o be mainly insignifican, excep for he 6 h quarer or so. Over a longer erm, here appears o be some dampening (bu insignifican) effec on he house prices. Overall, real house prices in Ausralia seem o be largely insensiive o shocks in he maerial coss. This is consisen wih general observaion in Ausralia ha house price rises significanly more han he increase in average household income, house building consrucion cos and average ren. Finally, boh he exchange rae shock and foreign ineres rae shock pose lile impac on real house prices, which implies ha he exchange rae is no a fundamenal deerminan of house prices in Ausralia. Wih is inflaion argeing sraegy adoped in he early 1990s, he RBA changes he official cash rae if he inflaion rae goes beyond he range of 2-3 percen. Various shor and long erm ineres raes are supposed o change in line wih he changes in official cash rae; however, here are oher facors ha may also conribue o he movemens in ineres raes. In paricular, we would like o see if he ineres raes are affeced by he shocks semming from he housing marke. 23

25 Figure 5: Ineres rae responses o various shocks ε ry R ε π R ε rg R ε s R ε d R ε mc R ε ex R ε fr R Noe: The doed lines represen he upper and he lower bands of 95% boosrapped confidence inervals compued using 1000 repeiions of Hall s percenile mehod. Figure 5 shows he impulse response funcions of he shor erm ineres rae o various macroeconomic and housing marke shocks. A posiive shock o he aggregae oupu level does no appear o have any significan effec on he ineres rae. The ineres rae ends o decline somewha noably beween eigh o 10 quarers only, which is an expeced response o higher real oupu. A oneime shock o inflaion leads o immediae significan increase in he ineres rae, rising up o abou 0.4 per cen. The ineres rae keeps on rising significanly for abou a year followed by 24

26 insignifican declines (Figure 5). The immediae and significan rise in he ineres raes due o an inflaionary shock reflecs he RBA s inflaionary argeing sraegies as well as possible increases in he money demand. A shock semming from Gov. spending on housing has a srong significan effec (posiive) on he ineres rae, which persiss for abou 20 quarers. This finding has an imporan policy implicaion; higher gov. spending increases he money demand leading o quie persisen increase in ineres raes, however, as we have poined ou in he previous subsecion, his Gov spending plays very lile role in boosing he housing marke. I is observed ha a housing marke shock, irrespecive of wheher i is coming from demand or supply sides, increases he ineres rae. In boh cases, he ineres rae sars o increase significanly afer one or wo quarers of he shocks and he posiive responses persis for abbo quarers. I is undersandable ha ineres rae rises due o a housing demand shock hrough he increase in demand for housing loans. However, he increase in he rae o a housing supply shock is relaively hard o raionalize. One possible explanaion may be ha higher level of housing consrucions generaes addiional income for he people in he housing indusries; which in urn increases aggregae demand and he ineres rae. A posiive shock o he raw maerial coss iniially decreases he ineres rae significanly. A higher maerial cos is comparable o an adverse supply shock, which may lead o an increase in inflaion and aggregae oupu level. The ineres rae may decline if he Cenral Bank reacs o lower inflaion and he oupu level. In line wih he ineres pariy condiion, i is observed ha a posiive shock o he foreign ineres rae increases he domesic ineres rae immediaely and his effec is significan and quie persisen. A posiive shock o he exchange rae (appreciaion) appears o have long-lasing negaive effec on he ineres rae. The main reason for his may be ha an appreciaion of he currency causes he fuure expeced appreciaion, which in order o fulfil he ineres pariy condiion, leads o reducion in he domesic ineres rae. The effec of one period posiive shock o foreign ineres rae on he domesic ineres rae also seems o be significan (Figure 5). Domesic ineres rae seems o rise significanly responding o a shock in he foreign ineres rae over he enire span of ime presened in he figure (abou seven years), wih increased growh rae up unil he 8 h quarer. This is an expeced oucome since higher ineres overseas would raise domesic ineres rae slowing capial ouflow hereby mainaining ineres rae pariy. 25

27 5. Conclusions This sudy develops an economeric model for Ausralian housing marke and examines he moneary ransmission effecs on he housing marke by esimaing a 9-variable SVAR model. The idenificaion scheme used in his paper disenangles he effecs of various macroeconomic, housing marke and policy shocks while examining he role of hese shocks in Ausralian housing marke and he overall Ausralian economy. The housing marke model based on he conemporaneous relaions shows ha he shor erm ineres rae is he main deerminan of real house prices in Ausralia. The quaniy of new houses consruced is primarily affeced by he inflaion rae and real house prices. The evidence from impulse responses suggess ha a conracionary moneary shock immediaely increases he number of new houses followed by a significan negaive effec on housing consrucion for abou a year or so, before he effec fades away. A posiive moneary policy shock is also observed o significanly raise maerial coss and real house prices for a shor period of ime. These resuls sugges ha a conracionary moneary policy shock iniially increases he housing demand reflecing public s expecaion for furher increase in he fuure ineres raes and house prices. However, in he long run, he conracionary policy dampens he housing demand and ulimaely causes he real house prices o decline. The robusness of our idenificaion scheme is eviden from he convenional paerns observed in he responses of aggregae oupu, inflaion, he exchange rae, he ineres rae and governmen spending on housing o he idenified moneary policy shock. An imporan finding of his sudy is ha he RBA ends o reac o he shocks semming from housing marke. This is eviden from he finding ha he shor erm ineres rae increases following boh housing demand and supply shocks and decreases o a posiive shock o housing maerial coss. Apar from he moneary policy shock, a number of oher shocks are found o significanly affec real house prices and housing oupu in Ausralia. Real house prices are influenced posiively by he shocks originaed from housing demand and aggregae real economic aciviies and negaively by housing supply shocks. In his regard, he effecs of real oupu and housing supply shocks are found o be more persisen compared o he effec of a housing demand shock. While boh housing demand and supply shocks posiively affec he consrucions of new houses, a shock o inflaion is found o be adversely affecing he housing oupu. The shock o he Governmen spending for housing appears o have very insignifican effec on housing marke aciviies. 26

28 References: Abelson, P., Joyeux, R., Milunovich, G., Chung, D., Explaining house prices in Ausralia: Economic Record 81 (special issue), S Aoki, K., Proudman, J., Vlieghe, G., House prices, consumpion, and moneary policy: a financial acceleraor approach. Journal of Financial Inermediaion 13 (4), Baffoe-Bonnie, J., The dynamic impac of macroeconomic aggregaes on housing prices and sock of houses: a naional and regional analysis. The Journal of Real Esae Finance and Economics 17 (2), Berger-Thomson, L., Ellis, L., Housing consrucion cycles and ineres raes, Reserve Bank of Ausralia Research Discussion Paper No. 08. Bernanke, B., Gerler, M., Inside he black box: he credi channel of moneary ransmission. Journal of Economic Perspecives 9 (4), Case, K. E., Shiller, R. J., Is here a bubble in he housing marke? Brookings Papers on Economic Aciviy 2, Dvornak, N., Kohler, M., Housing wealh, sock marke wealh and consumpion: a panel analysis for Ausralia. Economic Record 83 (261), Elbourne, A., The UK housing marke and he moneary policy ransmission mechanism: an SVAR approach. Journal of Housing Economics 17 (1), Fry, R. A., Marin, V. L., Voukelaos, N Overvaluaion in Ausralian housing and equiy markes: wealh effecs or moneary policy? CAMA Working Paper No. 10. Iacoviello, M., House prices and business cycles in Europe: a VAR analysis. Boson College Working paper No Iacoviello, M., Minei, R., The credi channel of moneary policy: evidence from he housing marke. Journal of Macroeconomics 30(1), Jarocinski, M., Smes, F. R., House prices and he sance of moneary policy. Federal Reserve Bank of S. Louis Review 90 (4), Kim, S., Roubini, N., Exchange rae anomalies in he indusrial counries: a soluion wih srucural VAR approach. Journal of Moneary Economics 45(3), King, R., Plosser, C., Sock, J., Wason, M., Sochasic rends and economic flucuaions. American Economic Review 81(4), Ludvigson, S., Seindel, C., Leau, M., Moneary policy ransmission hrough he consumpion-wealh channel. Federal Reserve Bank New York Economic Policy Review 8 (1),