The IS Curve and the Transmission of Monetary Policy : Is there a Puzzle? *

Transcription

1 The IS Curve and he Transmission of Moneary Policy : Is here a Puzzle? * Charles Goodhar a and Boris Hofmann b a Financial Markes Group, London School of Economics b Zenrum für Europäische Inegraionsforschung, Universiy of Bonn ABSTRACT In his paper we assess he performance of he New Keynesian IS Curve for he G7 counries. We find ha here is an IS puzzle for boh he purely backwardlooking as well as for he forward-looking IS curve. The real ineres rae does no have a significanly negaive effec on he oupu gap. Based on an exended specificaion of he IS curve, also including asse prices and moneary aggregaes, we are able o resore a significanly negaive ineres rae effec on aggregae demand in all counries. This finding suggess ha a richer specificaion of he IS curve in empirical work may be necessary in order o obain an unbiased esimae of he effec of moneary policy on aggregae demand. * We hank Andrew Filardo, Jürgen von Hagen and Frank Smes and paricipans a he Sveriges Riksbank Conference on Asse Markes and Moneary Policy, Sockholm, June 2000 and seminars a he ZEI for helpful commens and suggesions. Correspondence: Boris Hofmann, Zenrum für Europäische Inegraionsforschung, Waler-Flex-Srasse 3, 533 Bonn, Germany, bhofmann@uni-bonn.de

2 . Inroducion The New Keynesian model of moneary ransmission has become a sandard ool for he analysis of moneary policy. The model reduces he economy o a wo equaion sysem, comprising an aggregae supply or Phillips Curve and an aggregae demand or IS Curve. The Phillips Curve reflecs ineremporally opimal price seing decisions by monopolisically compeiive firms (Taylor, 980, Roemberg, 982, Calvo, 983). The IS Curve represens he ineremporal Euler consumpion equaion. The heoreical version of he New Keynesian Phillips Curve and he New Keynesian IS Curve are purely forward looking. The Phillips Curve relaes inflaion o expeced fuure inflaion and he oupu gap. The IS Curve relaes he oupu gap o he expeced fuure oupu gap and he ex-ane real ineres rae. In empirical applicaions, however, backward-looking specificaions of boh he Phillips and he IS Curve are ofen preferred in order o mach he lagged and persisen responses of inflaion and oupu o moneary policy measures ha are found in he daa (Rudebusch, 2002). The performance of forward-looking versus backward-looking specificaions of he New Keynesian Phillips Curve has been subjec of exensive research effors over he las couple of years (Fuhrer, 997, Robers, 998, Gali and Gerler, 999). In conras o his, only few sudies have analysed he empirical performance of he New Keynesian IS Curve. Mos sudies rely on a purely backward-looking specificaion of he IS Curve 2. Rudebusch and Svensson (999) and Peersman and Smes (999) esimae backward-looking IS Curves for he US and he euro area respecively and boh find a significanly negaive real ineres rae elasiciy of around 0.. Nelson (200, 2002), on he oher hand, fails o find a significan effec of he real ineres rae on he oupu gap in he US and he UK. He refers o his finding as he IS puzzle. Basic references for he foundaions of he New Keynesian model are Clarida e al. (2000), Goodfriend and King (997), McCallum and Nelson (999a, b) and Roemberg and Woodford (999). 2 Excepions are Fuhrer (2000) and Fuhrer and Rudebusch (2002), who also allow for forward-looking elemens in he US IS Curve.

3 How can he IS puzzle be explained? Nelson (200) discusses various poenial explanaions for he IS puzzle, bu does no assess heir empirical relevance. The IS puzzle may arise because of (i) simulaneiy bias arising from forward looking moneary policy, (ii) misspecificaion because of omission of forward-looking elemens, which is srongly suggesed by heory, or (iii) because of mis-specificaion due o omission of oher deerminans of aggregae demand. The firs explanaion would invalidae any aemp o esimae IS curves empirically, bu here is lile empirical evidence of he moneary auhoriies being sufficienly good forecasers for his o be a plausible explanaion. The laer wo poenial explanaions would imply ha he IS puzzle migh be resolved by choosing an alernaive specificaion of he IS Curve. Tess of he empirical performance of he backward-looking New Keynesian IS curve are available only for he US, he UK and he euro area, bu no for oher major indusrialised counries. In his paper we aim o fill his gap by esimaing backward-looking IS Curves for he G7 counries. We find ha he real ineres rae coefficien is always insignifican, suggesing ha here is srong inernaional evidence of an IS puzzle in backward-looking specificaions of he IS Curve. In order o assess he empirical relevance of he poenial explanaions for he IS puzzle, we hen esimae forward-looking IS Curves and augmened backward-looking IS Curves also including oher asse prices and moneary aggregaes besides he shor-erm real ineres rae. We find ha adding forward-looking elemens does no, while adding oher asse prices does, help o resolve he puzzle. The forward-looking specificaion of he IS Curve yields a significanly negaive ineres rae elasiciy only for he US, he exended specificaion of he IS Curve for all G7 counries. The plan of he paper is as follows. In Secion 2 we discuss he poenial explanaions for an IS puzzle o emerge. In Secion 3 we esimae backward-looking, forward-looking and exended backward-looking IS curves for he G7 counries. Secion 4 concludes. 3

4 2. The IS Puzzle: Poenial explanaions There are various poenial explanaions for he empirical finding of an insignifican ineres rae elasiciy in simple empirical represenaions of aggregae demand such as he backwardlooking specificaion of he New Keynesian IS Curve. Firs, he esimaed ineres rae elasiciy may be downward biased because of simulaneiy bias arising from forward looking moneary policy. This criicism implies ha he validiy of any aemp o esimae srucural IS Curves can be quesioned and ha he analysis of moneary ransmission should raher focus on he effec of he exogenous or unsysemaic componen of moneary policy. In fac, parly as a resul of his criicism, he bulk of he lieraure on moneary policy ransmission is now based on VAR impulse response analysis, which aims a idenifying and simulaing he effec of moneary policy shocks, in order o disenangle he demand effec of moneary policy from is endogenous response o economic aciviy 3. A severe limiaion of he VAR approach is, however, ha i provides evidence only for he effec of moneary policy shocks, which accoun for a negligible share of overall ineres rae movemens, while nohing is learn abou he effec of sysemaic moneary policy measures. Anoher poenial explanaion for he IS puzzle is ha empirical IS curves are mis-specified. Firs, heory suggess ha he IS curve is forward looking and ha he relevan real ineres rae measure is he ex-ane real ineres rae, i.e. he nominal ineres rae less expeced fuure inflaion, raher han he ex-pos real ineres rae. Omission of forward-looking elemens in empirical IS curves may give rise o a downwards biased ineres rae elasiciy. Second, oher variables besides he shor-erm real ineres rae may affec aggregae demand. In empirical open economy exensions of he New Keynesian model (Ball, 998 and Svensson, 2000), he exchange rae appears as an addiional variable, besides he shor-erm real ineres rae, in he IS curve. A depreciaion of he real exchange rae makes domesic goods more compeiive and herefore has a posiive effec on ne expors. The exchange rae is ofen considered o be he mos imporan deerminan of aggregae demand besides he real 3 For a survey of he VAR lieraure see Chrisiano e al. (999). 4

5 ineres rae in small open economies where ne expors accoun for a large share of aggregae demand 4. The real exchange rae is no he only candidae for addiion o he IS Curve. Fuhrer and Moore (995) and Nelson (2002) sugges ha i is he ex-ane long-erm real ineres rae raher han he ex-pos shor-erm real ineres rae ha maers for aggregae demand. Fuhrer and Moore (995) show ha he ime series properies of he ex-ane long erm real rae resemble ha of he shor-erm nominal ineres rae, so ha he laer can be used as a proxy for he former in empirical analysis. Several empirical sudies (e.g. Bernanke and Blinder, 992, Fuhrer and Moore, 995, Wrigh, 2000) find a significan effec of he nominal ineres rae on he oupu gap. Nelson (2002) shows ha, given a sable demand funcion for base money, real base money growh proxies he ex-ane long-erm real ineres rae. He repors evidence suggesing ha real base money growh has a significanly posiive effec on he oupu gap in he US and he UK. Propery and share prices and broad moneary aggregaes may also affec aggregae demand via wealh effecs. A change in perceived lifeime wealh, caused by a change in asse prices or broad money, may induce consumers o change heir consumpion plans (Modigliani, 97). Evidence repored in Case e al. (200) suggess ha propery prices have a paricularly srong effec on household consumpion. Anoher, more indirec wealh effec of asse price movemens operaes via households and firms balance-shees. Households and firms may be borrowing consrained due o asymmeric informaion in he credi marke, which gives rise o adverse selecion and moral hazard problems 5. As a resul, households and firms can only borrow when hey offer collaeral, so ha heir borrowing capaciy is a funcion of heir ne worh, which in urn depends on asse prices. Empirical evidence 4 A Moneary Condiions Index (MCI), a weighed average of he shor-erm ineres rae and he exchange rae, is commonly used as a composie indicaor for he sance of aggregae demand by inernaional organisaions and cenral and privae banks. See Eika e al. (996), Ericsson e al. (998) and Gerlach and Smes (2000) for an exposiion and criical discussion of he MCI concep. 5 Basic references of his lieraure are Bernanke and Gerler (989) and Kiyoaki and Moore (997). For a survey see Bernanke e al. (998). 5

6 repored in Hofmann (200, 2002) suggess ha propery prices have a significanly posiive effec on credi aggregaes boh in he shor and long-run 6. If oher variables besides he real ineres rae also affec aggregae demand, he esimaed ineres rae elasiciy in he sandard IS Curve specificaion may be biased. The direcion of he bias will depend on he correlaion beween he ineres rae and he omied variable. We can illusrae his poin based on highly sylised model 7. Suppose ha aggregae demand is deermined by a relaionship of he form: () y = β r + β 2x + v, where r is a shor-erm real ineres rae represening he moneary policy insrumen and x is some oher variable ha also affecs aggregae demand, e.g. he real exchange rae or some oher asse price. If we would now esimae equaion () omiing variable x, he OLS esimaor of β would be given by (2) ˆ β β + cov( r, x )/ var( r ) =. β 2 The omission of variable x would herefore give rise o a biased esimae of he ineres rae effecs on oupu. Bu his is no necessarily a reason o worry. The coefficien β gives he direc effec of ineres raes on oupu. However, ineres rae changes may also be ransmied via variable x. E.g. an increase in he real ineres rae is expeced o have a negaive effec on asse prices and moneary aggregaes. If here is a relaionship beween x and r of he form (3) x = γr, 6 See also Borio, Kennedy and Prowse, 994, IMF, 2000, and BIS, The model is similar o Woodford (994). 6

7 hen he esimaor of β will be given by ˆ = β + β γ. Since γ is negaive, he esimae of β 2 β will be upwards biased. ˆβ will hen give an esimae no only of he direc ineres rae effec, bu raher of he oal effec of an ineres rae change comprising all channels of moneary ransmission. However, i may also be ha moneary policy successfully ries o neuralise he effec of x on oupu, so ha he relaionship beween x and r is given by: (4) r = λx. In his case we ge ˆ = β + β λ. λ is posiive, implying ha he esimaor is biased β 2 owards zero. In he exreme case where moneary policy compleely neuralises he effec of x on y, λ = β 2 / β and ˆ β = β β 0. = 3. Empirical Analysis The backward-looking IS curve In empirical applicaions (Fuhrer and Moore, 995, Rudebusch and Svensson, 999, Rudebusch, 2002), he forward looking heoreical IS Curve is ofen approximaed by a backward looking specificaion of he form: (5) y = α y + α 2 y 2 + β ( i π ) + ε, where y is he oupu gap, i is he shor-erm nominal ineres rae, π is he inflaion rae. Equaion (4) has become one of he sandard empirical models of aggregae demand in he US (Esrella and Fuhrer, 998, Nelson, 200). Rudebusch and Svensson (999, p.4) argue ha he real ineres rae erm on he righ hand side of equaion (5) is a simple represenaion of he moneary ransmission mechanism, which, in he view of many cenral 7

8 banks, likely involves nominal ineres raes (e.g., morgage raes), ex ane real shor and long raes, exchange raes, and possibly direc credi quaniies as well. They conclude ha equaion (2) appears o be a workable approximaion of hese various inermediae ransmission mechanisms. This argumenaion implies ha Rudebusch and Svensson assume ha he rue model for aggregae demand is given by equaion () and ha all oher variables appearing in he IS Curve besides he real ineres rae are deermined according o a relaionship as equaion (3). Rudebusch and Svensson (999) esimae equaion (5) for he US over he sample period 96:-996:4 and obain: y =.6 y 0.25 y 0.0 i π + η, (6) 2 ( ) (0.08) (0.08) (0.03) using he Congressional Budge Office s esimae of he oupu gap for y, he four quarer average of he Federal Funds Rae for i and is he four quarer average of he GDP deflaor inflaion rae for π. Peersman and Smes (999) esimae equaion (5) for he euro area over he sample period and obain very similar resuls. Nelson (200, 2002) esimaes backward looking IS curves very similar o equaion (5) for he US and he UK over he sample period For boh counries he fails o find a significanly negaive real ineres rae coefficien. No aemp has ye been made o assess he performance of equaion (5) for a larger group of counries. In he following we esimae equaion (5) for he G7 counries over he sample period using quarerly daa 8. As a measure of he oupu gap we use he percen gap beween real GDP and poenial real GDP, calculaed using a sandard Hodrick-Presco- Filer wih a smoohing parameer of Following Rudebusch and Svenssson (999) we use a four-quarer moving average of he shor-erm money marke rae for i and a four- 8 The sample period was chosen in order o avoid ha he large increase in shor-erm real ineres raes in he early 980s dominaes he esimaion resuls and also o make i mach he sample period of Nelson (200, 2002). 9 The Hodrick-Presco Filer rend was calculaed over he period The rend esimaes from oher rend filers, such as a bandpass filer, were very similar. 8

9 quarer moving average of he CPI inflaion rae for π. The daa were aken from he IMF and he BIS daabase. The equaions were esimaed separaely by OLS. The second lag of he oupu gap was reained only if i was significan a leas a he 0% level. The resuls are repored in Table. For each counry we repor he coefficien esimaes wih -saisics in parenheses, he 2 adjused coefficien of deerminaion ( R ) and he Durbin-Wason saisic (DW). The resuls sugges ha here is srong evidence of an IS Puzzle in he G7 counries. In no counry do we find a significanly negaive real ineres rae elasiciy. Only in Canada is he -saisic is close o he 0% significance level (-.64). Taken lierally, his resul means ha here is no significan link from he moneary policy insrumen o he real economy. In oher words, moneary policy is ineffecive. Table : OLS esimaes of he backward looking IS Curve, 982:-998:4 2 y y 2 i π R DW Canada.33 (2.9) (-3.852) (-.63) France 0.9 (5.77) (-.38) Germany 0.72 (7.90) (-0.58) Ialy (9.80) (-.25) Japan 0.78 (0.2) (-.2) UK.222 (0.73) (-2.9) (-0.45) USA.308 (2.2) (-4.37) (-0.66) Noe: The able repors he resuls obained from esimaing equaion (5) by OLS. T-saisics are in parenheses. 9

10 The forward-looking IS Curve In he previous secion we have argued ha one poenial explanaion for he IS puzzle is misspecificaion of he IS curve due o omission of forward-looking oupu erms. While here are various sudies esimaing forward-looking New Keynesian Phillips Curves 0, only a few sudies have ried o esimae forward-looking New Keynesian IS Curves. In he following we esimae an IS curve specificaion ha also incorporaes forward-looking oupu expecaions. Following Fuhrer and Rudebusch (2002) we choose a hybrid specificaion of he form 2 (7) y = α y + α y + µ E y+ β i Eπ + ) + ε 2 2 ( Equaion (7) differs from equaion (5) by comprising, in addiion o wo lags of he oupu gap, also he curren period s expecaion of nex period s oupu gap ( E y + ). Also, insead of he ex-pos real ineres rae an ex-ane period real rae measure is used, defined as curren period s shor-erm nominal ineres rae less curren period s expecaion of nex period s inflaion rae ( Eπ + ). We esimae equaion (7) by Generalised Mehod of Momens (GMM) using four lags of he oupu gap, he shor-erm nominal ineres rae and he inflaion rae as insrumens. The second lag of he oupu gap was again reained only if i was significan a leas a he 0% level. In Table 2 we repor he coefficien esimaes wih -saisics in parenheses. T-saisics were calculaed based on heeroskedasiciy and auocorrelaion robus Newey-Wes sandard 0 See for insance Gali and Gerler (999) for he US and Gali, Gerler and Lopez-Salido (200) for he euro area. Fuhrer (2000) and Fuhrer and Rudebusch (2002) esimae forward-looking specificaions of he IS Curve for he US and conclude ha here is lile evidence ha forward looking expecaions are an imporan deerminan of curren oupu. 2 Empirically, he purely forward-looking IS curve was found o be unable o mach he dynamics of aggregae oupu (Cogley and Nason, 995 and Esrella and Fuhrer, 998). For his reason, hybrid specificaions of he IS curve, including boh forward-looking and backward-looking elemens, are preferred in empirical analysis. Fuhrer (2000) shows ha such hybrid specificaions can be heoreically moivaed by habi formaion in consumpion. 0

11 2 errors. We also repor he adjused coefficien of deerminaion ( R ), he Durbin-Wason saisic (DW) and he J-saisic of he overidenifying resricions wih probabiliy values in parenheses. The resuls sugges ha omission of forward-looking elemens in he IS curve is no he cause of he IS puzzle. We find a significanly negaive real ineres rae elasiciy only in he US. In he oher counries he real ineres rae elasiciies are insignifican, in Canada i is even significanly posiive. Table 2: GMM esimaes of he forward looking IS Curve, 982:-998:4 2 E y + y y 2 i E π + R J-es DW Canada (3.7) (9.74) (2.2) (0.75) France (7.24) 0.54 (9.48) (-0.86) (0.64) Germany (5.45) (6.84) (0.75) (0.74) Ialy (5.44) 0.52 (6.99) (-0.8) (0.80) Japan (6.3) (9.87) (-.38) (0.50) UK (3.90) (8.3) (-3.59) (-0.66) (0.37) USA (7.79) (0.4) (.82) (-.88) (0.44) Noe: The able repors he resuls obained from esimaing equaion (7) by GMM. T- saisics are in parenheses.

12 An exended IS Curve In he previous secion we have shown ha if aggregae demand also depends on oher variables besides he real ineres rae and if hese variables are no a funcion of ineres raes according o equaion (3), bu if ineres raes are raher a funcion of hese variables according o equaion (4), hen he esimaed ineres rae effecs will be biased owards zero. In order o assess he empirical relevance of his argumen, we esimae an exended version of he IS Curve, adding one lag of he real effecive exchange rae (rex), he nominal shorerm ineres rae (i) and he annual rae of change in real residenial propery prices ( hp ), real share prices ( 2

13 The exended IS Curve is of he form: (8) y = α y + α 2 y 2 + β( i π ) + β hp 4 + β sp 5 + β m 6 + β rex base 2 + β m 7 + β i broad 3 + β y 7 us + η Since he variables of he righ hand side of equaion are obviously correlaed, he esimaing equaion suffers from mulicollineariy which was refleced by insignifican -saisics when he full specificaion was esimaed. For his reason we esimae equaion (8) saring wih he full specificaion and hen progressively eliminae he leas significan variable unil all reained variables were significan a leas a he 0% level. In Table 3 we repor he resuls of his esimaion procedure. The resuls are quie clear cu. The shor-erm real ineres rae is now significan in all counries excep for he UK and he change in real propery prices is significan a leas a he 0% level in all seven counries. In he UK, aggregae demand is influenced by he shor-erm nominal ineres rae raher han he shor-erm real rae. The US oupu gap is significan only in Canada and Ialy. We do no find evidence of a significan effec of he real exchange rae, real share prices and real base money growh in any counry. Broad money growh is significan only in he UK. Thus, he real ineres rae and real propery prices are he mos imporan deerminans of aggregae demand. The general paern is ha he shor-erm real ineres rae and he change in real residenial propery prices ener significanly he IS Curve. The exended specificaion yields significan real ineres rae coefficiens, suggesing ha he IS puzzle disappears when a richer specificaion of he IS curve is chosen. For he UK i appears ha he IS Curve should be modelled in erms of he shor-erm nominal rae and no he shor-erm real ineres rae. 3

14 Table 3: OLS esimaes of he exended IS Curve, Sample 982:-998:4 y y 2 i π rex i hp base sp m Canada 0.94 (6.6) (-2.50) (-.65) (2.4) (2.52) France (6.40) (-.79) (2.26) Germany 0.67 (7.57) (-.77) (2.80) Ialy (5.68) (-2.54) (.82) (2.57) Japan (4.87) (.96) (-.85) (2.49) UK 0.82 (8.25) (-2.57) 0.06 (.66) (2.7) USA.2 (0.02) (-.80) (3.6) Noe: The able repors he resuls obained from esimaing equaion (8) by OLS. Insignifican variables were eliminaed. T-saisics are in parenheses. broad m us y 2 R DW

15 4. Conclusions In his paper we assess he performance of he empirical New Keynesian IS Curve, which is widely used for he analysis of moneary policy. We show ha he sandard backward-looking specificaion of he IS Curve does no work in he G7 counries. The real ineres rae elasiciy is always insignifican, implying ha he cenral bank can no sabilise he economy. Nelson (200) erms his finding as he IS Puzzle. We hen assess empirically he wo mos promising explanaions for he IS puzzle. Firs, misspecificaion due o omission of forward-looking elemens in he IS curve. Second, misspecificaion due o omission of oher deerminans of aggregae demand besides he real ineres rae. The evidence appears o favour he laer raher han he former explanaion. Excep for he US, we do no obain significanly negaive ineres rae elasiciies from esimaing forward-looking IS curves, while exended IS Curves yield significanly negaive real ineres rae elasiciies for all G7 counries, excep for he UK, where i is he shor-erm nominal ineres rae raher han he real rae ha affecs he oupu gap. Besides he ineres rae, propery prices are found o ener wih a significanly posiive elasiciy he IS curves in each counry. Real share prices and he real exchange rae are no found o have a significan effec on he oupu gap. The same applies o moneary aggregaes, excep for he UK where real M4 growh has a significanly posiive effec on he oupu gap. Thus, he resuls from esimaing exended IS curves for he G7 counries sugges ha he sandard specificaion of he IS curve is no sufficien o properly idenify he effec of ineres raes on aggregae demand and ha a broader framework, also aking ino accoun he demand effecs of oher variables, especially propery prices, migh be more appropriae. However, some qualificaions are in order. Firs, i should be noed ha he exended IS curve is no derived from an opimising model bu is specified raher ad-hoc. The exended IS curve may herefore reflec reduced form raher han srucural relaionships. More work on he heoreical underpinnings of an exended IS curve of he form esimaed in his paper is needed. 5

16 Second, he evidence presened in his paper should no be inerpreed as suggesing ha moneary policy should reac mechanically o propery price movemens. I can easily be shown ha in an empirical model comprising an exended IS curve of he form esimaed in his paper, a Taylor-rule designed o sabilise oupu and inflaion will acively respond o propery prices in order o off-se he effec of propery price movemens on aggregae demand. However, his conclusion would be highly misleading, since asse prices are endogenous, which mus be aken ino accoun when evaluaing he opimal response of moneary policy o asse price movemens. The opimal policy response will depend on he driving force of asse price flucuaions (Gerlach and Smes, 2000, Smes, 997) and he effec of ineres raes on asse prices. The finding of a significan effec of propery prices is herefore no sufficien o conclude ha moneary auhoriies should direcly respond o propery price movemens. In order o assess he opimal response of moneary policy o asse prices, a richer empirical model, also including behavioural equaions for asse prices, would be needed. 6

17 References Ball, L., (998), Policy Rules for Open Economies, NBER Working Paper No Bernanke, B. and A. Blinder, (992), The Federal Funds Rae and he Channels of Moneary Policy Transmission, American Economic Review, 82, pp Bernanke, B. and M. Gerler (989), Agency Coss, Collaeral and Business Flucuaions, American Economic Review, 79, pp 4-3. Bernanke, B., M. Gerler and S. Gilchris (998), The Financial Acceleraor in a Quaniaive Business Cycle Framework, NBER WP BIS (200), 7 s Annual Repor. Borio, C., Kennedy, N. and S. Prowse (994), Exploring Aggregae Asse Price Flucuaions across Counries: Measuremen, Deerminans and Moneary Policy Implicaions, BIS Economic Papers No. 40. Case, K., J. Quigley and R. Shiller (200), Comparing Wealh Effecs: The Sock Marke versus he Housing Marke, Cowles Foundaion Discussion Paper No Yale Universiy. Chrisiano, L., Eichenbaum, M, and C. Evans (999), Moneary Policy Shocks: Wha have we Learned and o Wha End?, in: J. Taylor and M. Woodford (eds.), Handbook of Macroeconomics, Norh Holland. Clarida, R, Gali, J. and M. Gerler (999), The Science of Moneary Policy: A New Keynesian Perspecive, Journal of Economic Lieraure, 37, pp

18 Cogley, T. and J. Nason (995), Oupu Dynamics in Real-Business-Cycle Models, American Economic Review, 85, pp Eika, K.H., Ericsson, N.R. and R. Nymoen (996) Hazards in Implemening a Moneary Condiions Index, Oxford Bullein of Economics and Saisics, 58, pp Ericsson, N., Jansen, E., Kerbeshian, N. and R. Nymoen (998), Inerpreing a Moneary Condiions Index in Economic Policy, in: Bank for Inernaional Selemens, BIS Conference Papers Vol. 6 Esrella, A. and J. Fuhrer (998), Dynamic Inconsisencies: Counerfacual Implicaions of a Class of Raional Expecaions Models, Working Paper No. 98-5, Federal Reserve Bank of Boson. Fuhrer, J. (997), The (Un)Imporance of Forward-Looking Behavior in Price Specificaions, Journal of Money, Credi, and Banking, 29, pp Fuhrer, J. (2000), Habi Formaion in Consumpion and Is Implicaions for Moneary-Policy Models, American Economic Review, 90, pp Fuhrer, J. and G. Moore (995), Moneary Policy Trade-Offs and he Correlaion Beween Nominal Ineres Raes and Real Oupu, American Economic Review, 85, pp Fuhrer, J. and G. Rudebusch (2002), Esimaing he Euler Equaion for Oupu, Working Paper 02-3, Federal Reserve Bank of Boson. Gali, J. and M. Gerler (999), Inflaion Dynamics: A Srucural Economeric Analysis, Journal of Moneary Economics, 44, pp Gali, J. Gerler, M. and D. López-Salido (200), European Inflaion Dynamics, European Economic Review, 45, pp Gerlach, S. and F. Smes (2000), MCIs and Moneary Policy, European Economic Review, 44, pp

19 Goodfriend, M. and R. King (997), The New Neoclassical Synhesis and he Role of Moneary Policy, NBER Macroeconomics Annual 997, pp Hofmann, B. (200), The Deerminans of Privae Secor Credi in Indusrialised Counries: Do Propery Prices Maer?, BIS Working Paper No. 08. IMF (2000), World Economic Oulook, May 2000 Kiyoaki, N. and J. Moore, (997), Credi Cycles, Journal of Poliical Economy, 05, pp McCallum, B. and E. Nelson (999a), Performance of Operaional Policy Rules in an Esimaed Semiclassical Srucural Model, in: J. Taylor (ed.), Moneary Policy Rules, Universiy of Chicago Press, pp McCallum, B. and E. Nelson (999b), Nominal Income Targeing in an Open-Economy Opimizing Model, Journal of Moneary Economics, 43, pp Modigliani, F. (97), Moneary Policy and Consumpion, in: Federal Reserve Bank of Boson, Consumer Spending and Moneary Policy: he Linkages Nelson, E. (200), Wha does he UK s Moneary Policy and Inflaion Experience ell us abou he Transmission Mechanism?, CEPR Working Paper No Nelson, E. (2002), Direc Effecs of Base Money on Aggregae Demand: Theory and Evidence, Journal of Moneary Economics, 49, Peersman G. und F. Smes (999), The Taylor Rule: A Useful Moneary Policy Benchmark for he Euro Area?, Inernaional Finance,, pp Robers, J. (998), Inflaion Expecaions and he Transmission Moneary Policy, FEDS Working Paper , Federal Reserve Board, Washingon DC. 9

20 Roemberg, J. (982), Sicky Prices in he Unied Saes, Journal of Poliical Economy, 60, pp Roemberg, J. and M. Woodford (999), Ineres Rae Rules in an Esimaed Sicky Price Model, in: J. Taylor (ed.), Moneary Policy Rules, Universiy of Chicago Press, pp Rudebusch, G. (2002), Assessing Nominal Income Rules for Moneary Policy wih Model and Daa Uncerainy, Economic Journal, 2 (April), pp Rudebusch, G. and L. Svensson, (999), Policy Rules for Inflaion Targeing, in: J. Taylor (ed.), Moneary Policy Rules, Universiy of Chicago Press for NBER. Smes, F., (997), Financial Asse Prices and Moneary Policy: Theory and Evidence, CEPR Discussion Paper No. 75 Svensson, L. (2000), Open-Economy Inflaion Targeing, Journal of Inernaional Economics, 50, pp Taylor, J. (980), Aggregae Dynamics and Saggered Conracs, Journal of Poliical Economy, 88, pp -23. Woodford, M. (994), Non-Sandard Indicaors for Moneary Policy: Can Their Usefulness be Judged from Forecasing Regressions?, in N.G. Mankiw (ed.), Moneary Policy, Universiy of Chicago Press. Wrigh, S. (2002), Moneary Policy, Nominal Ineres Raes and Long-Horizon Inflaion Uncerainy, Scoish Journal of Poliical Economy, 49, pp

21 Daa Appendix Definiions and sources of residenial propery price series Unied Saes Definiion Single-family house price index Source OFHEO and Naional Associaion of Realors Japan Naion-wide land price index Japan Real Esae Insiue Germany Average sales price of owner occupied dwellings in Frankfur, Munich, Hamburg and Berlin Ring Deuscher Makler France Residenial house price index Bank of France Ialy Naional house price index Bank of Ialy/ Il consulene immobiliare Unied Kingdom All dwellings price index Deparmen of he Environmen Canada Muliple lising service price index of exising homes Bank of Canada