Inspecting ECB QE with a dynamic macroeconomic model

Transcription

1 Inspecing ECB QE wih a dynamic macroeconomic model Romanos Prifis DG ECFIN, European Commission, and European Universiy Insiue Lukas Vogel DG ECFIN, European Commission Preliminary draf This version: 31 March 2015 Absrac: The paper analyses quaniaive easing (QE) in a dynamic general-equilibrium model which includes asses of differen ypes and mauriy. We explicily model asse purchases by he cenral bank and heir impac on he cenral bank's balance shee. In paricular, QE is capured by cenral bank purchases of long-erm governmen bonds financed by enhanced liquidiy provision o he privae secor. Wih imperfec subsiuabiliy beween asse classes, QE affecs he erm premium, sock prices, he exchange rae and he privae secor's saving decision. We use he model o assess he impac of he ECB's QE programme. Wih 50 basis poins erm-premium reducion he model generaes 1 per cen effecive euro depreciaion and raises real GDP in he euro area by 0.3 per cen and prices by 0.4 per cen unil he end of The impac on long-run inflaion expecaions remains modes in our seing. Fronloading QE srenghens is expansionary effec in he shor and medium erm, unless i is associaed wih an earlier exi. A significan increase in longerm inflaion (expecaions) requires he balance shee expansion o be very long lasing. JEL classificaion: E44, E52, F41 Keywords: quaniaive easing, porfolio adjusmen, asse prices, real secor Acknowledgemens: The views expressed in he paper are hose of he auhors and should no be aribued o he European Commission. Auhors deails: Romanos Prifis, DG Economic and Financial Affairs, European Commission, CHAR 14/244, B-1049 Brussels, romanos.prifis@ec.europa.eu, el.: ; Lukas Vogel, DG Economic and Financial Affairs, European Commission, CHAR 14/233, B-1049 Brussels, lukas.vogel@ec.europa.eu, el.:

2 1. Inroducion The European Cenral Bank (ECB) has announced is programme of quaniaive easing (QE) on 22 January 2015 and has sared asse purchases in he QE framework in March QE has can be defined as exension of he cenral bank's balance shee, where he cenral bank purchases financial asses and exens liquidiy (base money) o he privae secor. Operaing close o he zero bound, he ECB considered is "convenional" moneary accommodaion o be insufficien o address weak inflaion dynamics, declines in inflaion expecaions and sizeable economic slack in he eurozone. As a resul, balance shee inervenions were proposed in order o "achieve he price sabiliy objecive, given ha ineres raes have reached heir lower bound." The measures have he objecive of "decisively underpinning he firm anchoring of medium o long-erm inflaion expecaions, [and] he sizeable increase in [he] balance shee will furher ease he moneary policy sance [ ] srenghening demand, increasing capaciy uilisaion and supporing money and credi growh, and hereby conribuing o a reurn of inflaion raes owards 2%." (M. Draghi, Inroducory Saemen o he Press Conference, 22 January 2015). The ECB's QE programme has a volume of 1000 billion, corresponding o circa 10% of annualised euro area (EA) GDP. This amoun is o be purchased in monhly ranches of 50 billion up o Sepember Claeys e al. (2015) provide a comprehensive overview of he echnical deails of QE implemenaion. This paper looks a he impac of QE on financial variables and economic aciviy in he conex of a dynamic general-equilibrium model. In paricular, we use and exend he European Commission's QUEST model by explicily specifying he cenral bank's balance shee and by inroducing governmen bonds of differen mauriy. QE is capured by cenral bank purchases of long-erm governmen bonds financed by enhanced liquidiy provision o he privae secor. The model feaures imperfec subsiuabiliy beween he differen asses. In his framework of imperfec subsiuabiliy, QE affecs he erm premium, sock prices, he exchange rae and he privae secor's saving decision. Our paper is relaed o he lieraure evaluaing he effecs of QE using a DSGE model which ransmis he effecs of cenral bank balance shee measures hrough he so-called porfolio rebalancing channel. The DSGE lieraure accouning for such a channel can be undersood as consising of wo broad modelling environmens for he analysis of QE. For a review of complemenary heoreical frameworks sudying QE see Caglar e al. (2012). Vayanos and Vila (2009) and Harrison (2012) evaluae he effecs of large-scale asse purchases (LSAPs) using he preferred-habia seup. In such models he Euler equaion for households is adjused o allow for a linear combinaion of shor- and long-erm ineres raes. The long-erm ineres rae may hen deviae from is long-run expecaion by a preference erm ha increases in he relaive supply of asses wih long and shor mauriy. Insead, we choose o follow Chen e al. (2011) and Falagiarda (2013) (among ohers) and inroduce he scope of cenral bank balance shee measures in new-keynesian DSGE models by assuming ha asses are imperfecly subsiuable, moivaed hrough ransacion coss on porfolio adjusmen. The inroducion of porfolio adjusmen coss implies 2

3 ha agens are facing a cos whenever he relaive composiion of heir asse porfolio deviaes from is seady-sae level. We exend previous heoreical conribuions on QE operaing hrough he porfolio rebalancing channel by allowing for an inernaional dimension o our analysis. Unlike he laer papers, which have siuaed heir analysis in closed-economy models, we assume ha he EA is linked o a res of he world aggregae (ROW) hrough rade and exchange rae ies. As a resul, we can also sudy he inernaional ransmission of QE hrough he exchange rae channel. Furhermore, we aemp o quanify he impac of he specific ECB programme, which was announced and subsequenly launched in he firs monhs of Through he lens of our quaniaive assessmen of his specific scenario, we are also able o address several aspecs of QE design, such as he quesion of he preferabiliy of back-loading or fronloading of balance shee operaions. In addiion, he framework is applicable o relaed quesions, i.e. he cos of exiing QE, or he quesion of he effeciveness of QE as policy ool even when shor-erm ineres raes are no a he ZLB. The paper, firs, describes he modelling of QE and he ransmission channels o financial and real variables. I, second, provides a survey of empirical evidence on he effeciveness of QE wih a view of informing he model paramerisaion. Third, we presen simulaion resuls ha gauge he impac of he ECB's QE programme on ineres spreads, exchange raes and economic aciviy, noably real GDP and inflaion. Finally, we discuss poenial cross-counry heerogeneiy in he impac of EA QE and conclude. 2. Modelling QE QE denoes a moneary policy sraegy o increase he size of he cenral bank's balance shee. In paricular, he cenral bank purchases long-erm governmen bonds, wih he aim o reduce ineres spread beween shor and long mauriies (i.e. flaening he yield curve and finances his ransacion by addiional liquidiy provision o he privae secor. QE inends o affec privae-secor porfolio and saving decisions especially when shor-erm policy raes are already a or close o he zero lower bound (ZLB). We inroduced hese elemens of QE in a sandard version of he European Commission's QUEST macroeconomic model wih one producion secor (radables) and wo regions, namely he euro area (EA) and he res of he world (RoW). A deailed descripion of he QUEST model can, e.g., be found in Vogel (2014). The model exension inroduces long-erm governmen bonds (superscrip L) alongside shor-erm governmen deb (superscrip S) and he cenral bank balance shee. Following Woodford (2001), Chen e al. (2012), and Liu e al. (2015) long-erm governmen deb is modelled hrough bonds for which he nominal coupon c, which is a fracion of he principal, depreciaes over ime a rae b. The price in period of a long-erm bond issued in ( N P ) equals he discouned value of fuure paymens: T n N b (1) P 1 n0 (1 i) n c 3

4 where T is he mauriy period of he bond. Analogously, he price in period of a longerm bond issued in -1 ( P ) equals he discouned sum of ousanding paymens: O (2) P O T 1 n 1 n b 1n 0 (1 i) c If b (1 i) 1 and T is large, he price in of long-erm bonds issues in -1 corresponds (approximaely) o he price of newly issue long-erm bonds imes he depreciaion rae: (3) P P O N b Equaion (3) shows ha he price of he long-erm bond ha pays a declining coupon declines over ime a he rae b. We assume he long-erm bonds in he model o have 10- year mauriy, so ha, on average, one 40 h of he long-erm bonds maure in each quarer. The depreciaion rae on he coupon is se in line wih his o b Toal governmen deb consiss of long-erm bonds B and shor-erm bonds L S B : (4) B B B L S We assume ha long-erm bonds accoun for a consan share of he sock of governmen deb: (5) B L L s B The share of long-erm bonds is se o sl 0.5 in he simulaions, implying ha he average mauriy maches he value repored by Harwig Lojsch e al. (2011) for EA governmen deb. From he ousanding long-erm bonds he cenral bank: B are held by he privae secor and LH, B by L, CB (6) B B B L L, H L, CB The model inroduces imperfec subsiuabiliy beween shor-erm and long-erm bonds, which akes he form of quadraic porfolio adjusmen coss. In paricular, households have a preference for holding a mix of shor-erm and long-erm bonds, and deviaions from he arge value for he raio of long-erm over shor-erm deb generae quadraic adjusmen coss, which are scaled by he parameer b. The same formulaion of porfolio preferences/adjusmen coss has been used previously by, e.g., Andrés e al. (2004), Falagiarda (2013), Harrison (2012), and Liu e al. (2015). As shown by Schmi-Grohé and * Uribe (2003), an analogous formulaion for (ne) foreign asses relaive o he arge B closes he exernal par of he model. Privae households wih access o financial markes (superscrip r for Ricardian) face he following opimisaion problem: 1 1 The descripion here omis adjusmen coss in he real secor of he economy (price, wage, capial, and labour adjusmen coss) ha do no affec he firs-order condiions for porfolio holdings and savings. These adjusmen coss are presen in he full version of he model used for he simulaions. Deails on he specificaion of he real-secor adjusmen fricions can be found in, e.g., Vogel (2014). 4

5 (7) r r r 0 0 max L U ( C, N ) c C C N L, H S (1 ) P r P ( K (1 k ) K 1) P B b B 2 C 1 ( 1) LH, P P P 2 B 2 S * * * L, H N L, H B e B f e ( B B ) TR cb 1 bp B 1 0 * (1 i ) P (1 i ) P 2 P P P P 0 S * w r C B 1 e B 1 (1 ) W N k k k P D ( i 1 ( i 1 k ) 1 1) K 1 P P P P P k Ricardian households receive labour income, reurns on financial asses, income i from lending capial o firms ne of an (exogenous) risk/insurance premium given revenue uncerainy, and dividends D from firm ownership. K I (1 k ) K 1 is he capial sock as he sum of new invesmen I and he previous-period capial sock depreciaed a w k rae k. The governmen levies axes on income from labour, on corporae income c and on consumpion. The price in period of a shor-erm (1-period) bond of nominal value e B S B is B / (1 i ) S * *, wih i being he shor-erm nominal ineres rae. Analogously, / (1 i ) is he price in domesic currency of a foreign bond * B, where e is he nominal exchange rae defined as he value in domesic currency of one uni of foreign currency. The maximisaion problem (7) provides us wih he following firs-order condiions (FOC): r S L 1 P 1 1 N B (8) E ( ) E ( ) b P ( 1) S L, H B P 1i B r N S S S L 1 P P 1 b B 2 B B (9) E ( ) ( ) 1 ( 1) ( 1) L, H E N L, H b L, H L, H B bp 1 c P 2 B B B (10) r * * L 1 e P 1 1 e ( B B ) E ( ) ( ) * E * f B e 1 P 1 i P (11) r C L 1 P 1 P 1 E( ) E( ) C (1 k ) k ( k K P P i i ) (12) L C 1 k k r c c C (1 ) P U r P (13) r L N r U w (1 ) W P N Combining (8) wih (10), (11) and (12) illusraes he ransmission channels of QE o he real economy: 5

6 (14) 1 1 1i B e 1 i P S * * N B e e ( B B ) b P ( 1) E ( ) LH, * f 1 S C 1 N B P 1 (15) b P ( 1) E ( ) L, H C k k k 1 i B P (1 i ) ( i ) 1 k k S c c C 1 N B (1 ) P U 1 (16) b P ( 1) L, H c c C 1 i B (1 ) P U 1 1 The impac on asse prices of cenral bank purchase of long-erm bonds derives from privae invesor porfolio adjusmen coss, i.e. imperfec subsiuabiliy beween differen financial asses. For b 0, i.e. bonds of differen duraion are imperfec subsiues, he, effecs of reducing LH S B relaive o B in he household porfolio are similar o he impac of a reducion of he shor-erm ineres rae i in he FOC. Unconvenional moneary policy can, hence, mimic he effecs of shor-erm ineres rae reducions. In paricular, when he cenral bank inervenes by purchasing long-erm bonds, privae invesors ha aim a re-esablishing he porfolio mix of shor-erm and long-erm asses can respond by holding more corporae equiy and foreign bonds, and by reducions in savings. The firs response means a porfolio reallocaion owards equiy and foreigncurrency asses ha increases he price of corporae equiy (rising sock marke) and he price of foreign currency (exchange rae devaluaion). Equaion (14) shows ha QE leads o higher demand for foreign asses and depreciaion of he domesic currency (increase in e ) for given levels of i (resriced, e.g., a he ZLB) and i *. Equaion (15) illusraes he porfolio reallocaion from governmen bonds owards corporae equiy. Equaion (16) shows ha QE reduces privae saving in he shor erm similar o a reducion in he shorerm ineres rae. Concerning he ransmission o he real economy, rising sock markes reduce he financing coss of corporaions and, dampening he required reurn o capial, ranslae (under decreasing reurns o capial) ino sronger invesmen and capial accumulaion. Exchange rae depreciaion srenghens ne expors provided ha expor and impor demand are sufficienly price elasic. The second response of reduced savings o resore he preferred porfolio mix srenghens conemporaneous consumpion demand. The sock of governmen deb which is composed of shor-erm and long-erm bonds follows: B P L B 1 ( bp c) B 1 PGE TAX PR (17) B (1 i ) P P P P P P P where PGE, S N S N L CB TAX and CB PR are, respecively, governmen primary governmen expendiure (public consumpion, public invesmen, ransfers), oal ax revenue (labour, consumpion, and corporae axes), and he operaing profi of he cenral bank as an addiional source of governmen revenue. 6

7 The operaing profi of he cenral bank equals he sum of base money issuance and ineres income minus he curren expendiure on buying long-erm bonds, where he laer equals he change of he value of long-erm bonds on he cenral bank's balance shee: (18) PR M cb P B P B CB L, CB N L, CB N L, CB 1 b 1 Under he cenral bank's budge consrain (18), purchases of long-erm governmen bonds can be financed eiher by increasing liquidiy (money issuance), or by reducing he cenral bank's operaing profi. 2 ; In line wih he sandard definiion of QE and he ECB announcemen, he noe focuses on he firs opion of enhanced liquidiy provision. Purchases of long-erm bonds by he cenral bank can be modelled as endogenous response o he economic environmen (e.g., he economy's posiion in he business cycle or he slope of he yield curve) similar o a Taylor rule, or as exogenous pah : CB (19) B (1 ) B B (1 ) F(.) L, CB L, CB L, CB CB L L 1 L The simulaions do no assume an endogenous response, i.e. F(.)=0, and calibrae he exogenous pah o he ECB programme insead. Affecing governmen revenues direcly (ineres rae on governmen deb, cenral bank profi) or indirecly (ax revenues), QE also affecs he evoluion of governmen deb o some exen. The budge-closure rule in he model ensures he sabilisaion of governmen deb around is arge level. In he simulaions we assume ha deb sabilisaion operaes hrough he adjusmen of he labour ax rae. Alhough he modelling of porfolio decisions focuses on privae households as financial invesors, i can also be undersood as including pension and invesmen funds ha ac on behalf of households. Banks as financial inermediaries are no included in he model, bu hey can be hough of as facing a similar decision problem based on a preferred porfolio mix. When he cenral bank buys long-erm governmen bonds from banks, he laer can respond by buying more bonds from households (secondary marke), buying more equiy and foreign asses, and providing more loans o equilibrae heir porfolio srucure. The empirical lieraure on unconvenional moneary policy has lised a number of channels hrough which QE can affec macroeconomic variables (see Krishnamurhy and Vissing-Jorgensen, 2011). The relevance of he individual channels is an empirical maer and likely o vary across counries and ime. Our modelling approach incorporaes he following: The purchase by he cenral bank of long-erm bonds reduces he reurn on hese bonds for given expecaions abou fuure shor-erm ineres raes, i.e. i lowers he erm premium, which is he difference beween he reurn on he long-erm bond and he expeced sequence of fuure shor-erm raes ("erm premium channel"). Porfolio rebalancing in he privae secor in response o cenral bank purchases of (safe) long-erm governmen bonds lead o higher demand for (riskier) asses. In paricular, demand for corporae equiy and real esae may increase and srenghen corporae and consrucion invesmen ("safey channel"). 2 A hird opion, in general, is for he cenral bank o sell oher asses from is porfolio o serilise he impac of is inervenion on he cenral bank balance shee. 7

8 QE srenghens he demand for foreign-currency asses and leads o a depreciaion of he domesic currency. The devaluaion improves he price compeiiveness of domesic oupu and he rade balance, if rade is sufficienly price elasic and openness owards non- EA counries significan ("exchange rae channel"). QE in is original definiion means an expansion of he cenral bank balance shee ha implies growh in base money ("liquidiy channel"); our model does no incorporae implicaions for he credi supply by banks, however. If QE depreciaes he exchange rae and simulaes demand and oupu, i also increases inflaion expecaions. A rise in inflaion expecaions reduces he real ineres rae and should srenghen ineres-sensiive demand in he economy, paricular when he zero bound on nominal shor-erm raes is binding ("inflaion channel"). Balance-shee operaions ha change he marke value of long-erm governmen bonds affec he financing condiions of he governmen. Using he financing-cos reducion o lower axes or increase primary expendiure would add demand simulus in he shor erm. Using he budgeary saving due o lower financing coss o pay down governmen deb reduces he fuure ax burden on he privae secor ("fiscal channel"). Absen in he QUEST se-up, alhough i plays a prominen role in he general discussion of QE effecs, is he "signalling channel", according o which QE communicaes a commimen by he cenral bank o leave ineres raes low for an exended period of ime. In he model, he pah of fuure risk-free ineres raes depends on he moneary policy rule. To he exen ha QE succeeds in simulaing aggregae demand and sabilising inflaion expecaions, fuure policy raes are likely o rise. Also absen in he QUEST framework are he "prepaymen risk channel" and he "defaul risk channel" (Krishnamurhy and Vissing-Jorgensen, 2011), which, according o Demary and Hueher (2015), are of minor relevance in he EA conex. 3. Empirical evidence on QE The vas majoriy of he empirical lieraure on unconvenional moneary policy measures has focused on ransmission hrough he porfolio rebalancing or erm-premium channel. The lieraure ends o concenrae on ineres rae effecs raher han considering also he ransiion from financial markes o aciviy in he real secor. Frazscher e al. (2013) invesigae he domesic effecs and global spillovers of he Fed's QE1 and QE2 unconvenional moneary policy operaions on porfolio decisions, asse prices and exchange raes, hus providing evidence on boh he erm-premium channel and he exchange rae channel of our model. They find ha QE1 measures boosed bond and equiy prices in he US and led o an appreciaion of he dollar. On he oher hand, QE2 was ineffecive in lowering yields worldwide and boosed global equiy prices leading o depreciaion of he dollar. The firs sage of QE hus generaed a porfolio rebalancing ou of emerging marke economies ino he US, whils QE2 generaed a rebalancing from he US ino EME's. Furher evidence suggesing ha QE has boosed GDP and inflaion via he porfolio rebalancing channel includes Joyce e al. (2011), and Joyce and Tong (2012). Boh papers use even sudy mehodologies o esimae he impac on bond yields of he Bank of England's firs ranche of 200bn of asse purchases in They find ha he effec was approximaely equal o 100bps. 8

9 In an empirical sudy ha analyses he effecs of he Fed's QE1 on he U.S. economy, Baumeiser and Benai (2010) find a 4 per cen posiive impac on real GDP in 2009Q1 and a 0.4 percenage poin impac on inflaion in 2009Q2. Chung e al. (2011) find ha a 50 bps decline in he erm premium in he U.S. (QE2) has led o an increase of real GDP by approximaely 2% in 2012 and an increase of inflaion by 0.7 percenage poins in However, he cenral focus of he empirical lieraure has relied on cenral bank announcemens, raher han acual implemenaion of QE. As Frazscher e al. (2013) sae "an imporan cavea is ha [...] announcemens do no imply any change in supply of e.g. US Treasury securiies a he ime he announcemens are made, bu hey merely indicae ha such a change will occur a some poin in he fuure, o some degree, and wih cerain probabiliy" (p.10). As a resul, his mehodology essenially capures a change in expecaions abou fuure asse prices raher han curren asse prices. Bauer and Rudebusch (2013) conribue o his issue by quanifying he saisical uncerainy surrounding esimaes derived from he workhorse models in empirical income finance and, using a suiable mehodology, find ha he primary effec of he Fed's QE operaed under announcemens, leading o lower marke expecaions of he fuure policy pah. As a resul, heir analysis suggess a more prominen role for QE o operae hrough he signalling channel. Engen e al. (2015) aemp o exploi he implemenaion lag in QE operaions and use survey daa from he Blue Chip Economic Indicaors o idenify changes in privae secor percepions. They hen use he survey-based esimaes of fuure ineres raes as inpu for heir model's Taylor ineres-rae rule and simulae Fed QE operaions for he US economy. Given ha privae-secor percepions implied a more accommodaive moneary policy, his placed downward pressure on real long erm ineres raes. Togeher wih he downward pressure on erm premia from asse purchases, hese expecaion effecs have eased financial condiions and suppored economic recovery. In anoher recen paper wih well-suied idenificaion, Trebesch and Zeelmeyer (2014) analyse he porfolio rebalancing and liquidiy channels of QE hrough he lens of he ECB individual-level bond purchases, revealed for Greece, afer he ECB did no ake par in he Greek sovereign deb resrucuring of Idenificaion is achieved from he crosssecional variaion in ECB purchases. The resuls sugges ha a purchase of 10% of a Greek bond is associaed wih a yield drop of bps. The oal drop in yields aribuable o he ECB purchases was esimaed o be bps, wih he effecs being more pronounced a shor mauriies. In an aemp o disenangle announcemen and implemenaion effecs of QE, Chrisensen and Rudebusch (2012) use a srucural model o decompose he declines in governmen bond yields following announcemens by he Fed and he Bank of England. They are able o disincly capure effecs relaed o changes in expecaions abou fuure moneary policy (signalling channel) and changes in erm premiums (erm premium channel). The resuls sugges ha for he US he main effec of QE operaed by affecing policy expecaions, whereas for he UK, yield declines were driven solely by erm-premium reducions. The cenral body of he empirical lieraure has aemped o assess he impac of QE by analysing he response of non-corporae asses, such as governmen bond prices and 9

10 yields, and has no focused on invesor-side behaviour and he spillover o privae-secor asses. In a recen paper by Joyce e al. (2014) he second issue is addressed. The sudy invesigaes he behaviour of UK insurance companies and pension funds before and during he financial crisis. The resuls sugges ha he Bank of England s QE policy achieved sizable porfolio rebalancing whereby invesors ha reduced heir bond holdings reinvesed some of he earnings ino corporae bonds. Greenwood and Vayanos (2010) look a how he supply and mauriy srucure of governmen deb has affeced bond yields and expeced reurns. They esimae he effecs of bond excess reurns and yield spreads on measures of he mauriy srucure of he public s holding of US governmen deb. They also find evidence of a porfolio rebalancing channel, which is sronges for longer-horizon excess reurns. In a similar spiri, Gagnon e al. (2011) esimae wheher long-erm yields are relaed o he 10-year Treasury erm premium and find ha QE1 should have reduced he erm premium by approximaely 52 bps and he 10-year Treasury yield by approximaely 82 bps. D'Amico and King (2010) analyse he effecs of he Fed's purchase program of $300 billion of U.S. Treasury coupon during 2009 and find ha each purchase operaion, on average, reduced he yields in he secor purchased by 3.5 bps on he days when hese purchases occurred. Hamilon and Wu (2012) invesigae he effec of QE in he US and find smaller esimaes for he effecs of QE1 on he 10-year Treasury yield and erm spreads han Gagnon e al. (2011), Greenwood and Vayanos (2010), and D Amico and King (2010). Furhermore, he esimaes for QE2 sugges he paradoxical resul ha he second round of asse purchases should have raised Treasury yields and erm premia. Using a novel mehodology, Li and Wei (2012) re-asses he effecs of he Fed's QE and conribue o he lieraure by being he firs o employ an arbirage-free erm srucure model. This modelling framework allows differeniaing beween long-erm ineres raes and he expeced average fuure shor-erm raes. Their esimaes quanify he effecs of he firs and second QE programmes by he Fed and find a combined effec of approximaely 100 bps on he 10-year Treasury yield. The approach hus lends validiy o he erm premium channel. On he inernaional dimension, Dahlhaus e al. (2014) invesigae he inernaional spillovers from US QE ono he Canadian economy using a FAVAR framework. They find ha he implemenaion of QE increased U.S. oupu by 2.3% and Canadian oupu by 2.2 % on average beween 2008Q4 and 2013Q3. A he same ime, U.S. long-erm spreads declined, and asse prices increased, leading o a rise in invesmen and consumpion. The effecs of US QE were spilled over o Canada hrough a financial channel, whereby Canadian long-erm spreads decreased, and asse prices increased. The Canadian nominal exchange rae also appreciaed wih respec o he US dollar, leading o an increase in Canadian demand for US expors. Finally, Lim e al. (2014) examine he gross financial flows o developing counries beween 2000 and 2013, by focusing on he effecs of he US Fed's QE. They find evidence ha QE operaed hrough porfolio rebalancing, signalling and liquidiy channel, o increase gross inflows ino developing counries by 3%. 10

11 An assessmen of he ECB's January 2015 QE programme is sill ousanding. The effecs on asse prices are, in general, difficul o idenify from he daa, because i would require knowledge of he relevan ime horizon, i.e. when did agens sar anicipaing which volume and ime pah of QE, and conrolling for simulaneous domesic and inernaional evens ha, e.g., affec he ime pah of risk premia, expeced fuure shor-erm ineres raes, and exchange rae dynamics. Table 1 provides some crude evidence. I shows he change in he yield of inflaionindexed 10-year bonds for Germany, France and Ialy over hree differen ime inervals. The bond yields have fallen by basis poins beween May 2014, when here was already some expecaion of EA QE, and January 2015, when he ECB finally announced he programme. Yield changes are smaller if shorer ime horizons are considered. The fall in yields beween end-december 2014 and end-january 2015, e.g., amouns o only basis poins. Table 1: Changes in bond yields for seleced EA counries If QE has conribued o he decline in long-erm bond yields i has no reversed he fall in inflaion expecaions over he same horizon. Insead, inflaion expecaions based on 10- year swap raes have declined unil end-january 2015 (Table 2). This does no exclude ha QE has miigaed he fall in expeced inflaion, which has been associaed primarily wih falling commodiy prices, i.e. ha inflaion expecaions would have fallen by more wihou he QE perspecive. However, by end-january 2015, he rend of falling expeced inflaion had no reversed. Table 2: Changes in swap raes and expeced inflaion More recenly, inflaion expecaions have sabilised in February and risen somewha during he firs half of March A conribuion of QE o his paern is plausibly, bu oher facors, such as he parial recovery of oil prices during February 2015, may also have played a role. 4. Simulaion resuls Change in yield of 10-year inflaion-indexed governmen bonds (% poins) 1 Germany France Ialy 30/05/14-31/01/ /10/14-31/01/ /12/14-31/01/ Indexaion o French CPI for France and o EA HICP for ohers. Source: Bloomberg Change in overnigh rae for 10-year ineres swaps (% poins) Nominal Real Inflaion 1 30/05/14-31/01/ /10/14-31/01/ /12/14-31/01/ Inflaion is he break-even value given nominal and real swaps. Source: Bloomberg This secion presens he simulaion resuls for implemening he ECB's QE in our modified version of QUEST. The simulaed QE pah corresponds o he announcemen of pur- 11

12 chasing 1000 billion long-erm bonds in monhly seps of 50 billion unil Sepember A gradual exi from QE is assumed hereafer. The long-erm bond purchase is financed by an increase in liquidiy supply. The value of porfolio adjusmen coss b deermines he impac of QE on he erm premium. This impac on he erm premium is very difficul o idenify from daa, because i requires knowledge of he relevan ime horizon, i.e. when did agens sar anicipaing which volume and ime pah of QE, and addiional domesic and inernaional facors such as he ime pah of (oher) risk premia and he pah of expeced fuure shor-erm ineres raes. In line wih he informaion in Tables 1 and 2 we assume an impac of QE on he erm premium of -50 basis poins for he oal volume of he ECB's programme. Given he ECB programme we se he value of b o generae a 50 bps erm premium decline. The remaining model parameers follow he sandard QUEST calibraion o mach long-erm raios from naional accouns and average adjusmen frequencies repored in empirical research. The zero bound on shor-erm policy raes is assumed o bind during Table 3 repors simulaion resuls for he ECB's pah during he duraion of he QE programme. Higher privae consumpion and invesmen demand, which is associaed wih lower savings and porfolio rebalancing, raise real GDP growh by 0.2 pp in he firs year and 0.1 pp in he second year. Table 3: ECB pah of 2-year QE 2015A 2016A in per cen: GDP.REAL_PCER CONSUMPTION_PCER INVESTMENT_PCER EXPORTS_PCER IMPORTS_PCER PGDP_PCER CPI_PCER NEER_PCER in percenage poins: GDP.GROWTH_ER INFL.PGDP_ER INFL.CPI Noe: Resuls are deviaions from a no-qe baseline The conribuion of ne expors (exchange rae depreciaion) is modes given adjusmen fricions ha reduce he shor-erm elasiciy of expor and impor demand wih respec o he nominal exchange rae. Inflaion rises by 0.2 pp in boh years in response o sronger oupu growh and exchange rae depreciaion. The simulaion repored in Table 3 assumes ha ax-payers benefi from he posiive impac of QE on he governmen deb-o-gdp raio (ax base growh, denominaor effec) hrough a moderae reducion in he labour income ax ha sabilises governmen finances a he pre-qe level. The alernaive ha posiive fiscal effecs are used o reduce he sock of governmen deb leads o very similar resuls, because he implied ax reducion in he scenario of Table 3 remains emporary and below 0.3 pp. 12

13 Table 4 looks a he quesion of wheher here is quaniaively imporan non-lineariy in he QE effecs wih regard o he saring condiions. Compared o Table 3, Table 4 places QE in a more adverse macroeconomic conex. In paricular, real GDP growh is lower by 1 pp in 2015 and 1 pp in 2016 (driven by weaker domesic demand) compared o he non- QE baseline of he Table 3 simulaion. The resuls in Table 4 sugges a moderaely higher posiive impac of he same volume of QE on real oupu and inflaion. The more posiive impac is associaed wih sronger euro depreciaion (improving expor demand) and a prolonged period of lower ineres raes (srenghening consumpion and invesmen demand). Table 4: ECB pah in more adverse environmen 2015A 2016A in per cen: GDP.REAL_PCER CONSUMPTION_PCER INVESTMENT_PCER EXPORTS_PCER IMPORTS_PCER PGDP_PCER CPI_PCER NEER_PCER in percenage poins: GDP.GROWTH_ER INFL.PGDP_ER INFL.CPI Noe: Resuls are deviaions from a no-qe baseline Table 5, finally, addresses he quesion of wheher he ime pah of QE implemenaion affecs is macroeconomic effecs. I repors resuls for a scenario of acceleraed QE in which he volume of 1000 billion long-erm bonds is bough by 2015q4 raher han 2016q3. As in he case of 2-year QE in Table 3, exi sars in 2017q1. Table 5 indicaes a more pronounced impac of fronloaded QE, noably on inflaion, which derives from he fronloading of privae-secor porfolio rebalancing and is implicaions for he exchange raes and privae secor demand. 3 3 Fronloaded QE wih an earlier exi, saring in 2016q1 insead of 2017q1, would, on he oher hand, weaken he GDP and inflaion effecs compared o Table 3. 13

14 Table 5: QE fronloading 2015A 2016A in per cen: GDP.REAL_PCER CONSUMPTION_PCER INVESTMENT_PCER EXPORTS_PCER IMPORTS_PCER PGDP_PCER CPI_PCER NEER_PCER in percenage poins: GDP.GROWTH_ER INFL.PGDP_ER INFL.CPI Noe: Resuls are deviaions from a no-qe baseline 5. Geographic heerogeneiy The discussion so far has focused on aggregae EA effecs. Differences in pre-qe governmen bond yields and economic condiions across EA Member Saes raise he quesion of wheher QE effecs will differ across counries and, noably, beween he "core" and he "periphery" of he EA. From he porfolio rebalancing perspecive, cross-counry differences can resul from counry-specific supply elasiciies for long-erm bonds ha he cenral bank is going o buy. If, e.g., privae holders are more willing o sell "periphery" bonds o he cenral bank, bu hold on o heir "core" bonds, he ECB purchase would imply less erm-premium reducion in "periphery" compared o "core" counries. Hence, QE would be less effecive in he EA periphery in paricular in he case of significan home bias in financial markes. Regionally differeniaed effecs may also derive from uneven financial consrains across counries. The direcion is, however, ambiguous and depends on he precise naure of he consrains. Where larger pars of he household secor have no access o capial markes (no porfolio reallocaion or savings response), immediae QE effecs will be weak. If weak privae demand is due o binding credi consrains, however, QE effecs may become more effecive. In paricular, QE-relaed growh in he value of pledgeable collaeral (housing wealh, firm value, and financial wealh) could ease credi consrains for households and firms. Exra liquidiy from QE and he rebound of asse values may also increase credi supply by banks, in paricular where he laer are operaing a he leverage consrain. To he exen ha banks remain fragile and undercapialised in "periphery" counries, however, addiional liquidiy hrough QE may no be convered ino more credi supply o he privae secor. Finally, real effecs hrough he exchange rae channel differ obviously due o heerogeneiy in rade elasiciies and openness o non-ea rade, alhough no necessarily across he core-periphery dimension. The rade effecs may feed back ino financial condiions, i.e. via higher sock marke gains of more expor-oriened firms and counries. 14

15 6. Conclusions The paper analyses quaniaive easing (QE) in a dynamic general-equilibrium model which includes asses of differen ypes and mauriy. We exend he European Commission's QUEST model and explicily incorporae asse purchases by he cenral bank and heir impac on he cenral bank's balance shee. In paricular, QE is capured by cenral bank purchases of long-erm governmen bonds financed by enhanced liquidiy provision o he privae secor. Wih imperfec subsiuabiliy beween shor-erm and long-erm asses, QE affecs privae-secor porfolio allocaion and saving decisions and, by consequence, erm premia, sock prices, he exchange rae and he privae demand. The model has hen been used o assess he impac of he ECB's QE programme, i.e. he announced purchase of 1000 billion long-erm bonds in monhly seps of 50 billion up o Sepember 2016, under he assumpion of a erm premium reducion by 50 basis poins, which is in he order of magniude of experience wih QE in oher counries and wih some early evidence from EA financial markes. Wih 50 basis poins erm-premium reducion he model generaes 1 per cen effecive euro depreciaion and raises real GDP in he euro area by 0.3 per cen and prices by 0.4 per cen unil he end of The impac on long-run inflaion expecaions remains modes in our seing. Fronloading QE srenghens is expansionary effec in he shor and medium erm, unless i is associaed wih an earlier exi. Increasing long-erm inflaion (expecaions) significanly requires he balance shee expansion o be very long lasing. The noe has also discussed poenial differences in QE effecs beween "core" and "periphery" counries. If privae invesors are more willing o sell "periphery" bonds o he cenral bank and hold on o heir "core" bonds, QE would imply less erm premium reducion in "periphery" compared o "core" counries and, hence, be less effecive in he former. The impac of financial consrains on QE effecs is ambiguous and depends on heir precise form. Finally, real effecs hrough he exchange rae channel differ obviously due o heerogeneiy in rade elasiciies and openness o non-ea rade, alhough no necessarily across he core-periphery dimension. References: Andrés, J., D. López-Salido, E. Nelson (2004): Tobin's imperfec asse subsiuion in opimizing general equilibrium, Journal of Money, Credi and Banking 36, Bauer, M., G. Rudebusch (2013): The Signalling Channel for Federal Reserve Bond Purchases, Federal Reserve Bank of San Francisco, Working Paper Baumeiser, C., L. Benai (2010): Unconvenional Moneary Policy and he Grea Recession: Esimaing he Impac of a Compression in he Yield Spread a he Zero Lower Bound. European Cenral Bank Working Paper 1258 Caglar, E., Chadha, J. S., Meaning, J., Warren, J., and Waers, A. (2012): Nonconvenional moneary policies: QE and he DSGE lieraure, in: Chadha, J., S. Holly (eds.): Ineres Raes, Prices and Liquidiy: Lessons from he Financial Crisis, chaper 11. Chen, H., V. Cúrdia, A. Ferrero (2012): The macroeconomic effecs of large scale asse purchase programmes, Economic Journal 122,

16 Chrisensen, J., G. Rudebusch (2012): The Response of Ineres Raes o US and UK Quaniaive Easing, Economic Journal 122, F385-F414. Chung, H., J.-Ph. Lafore, D. Reifschneider, J. Williams (2011): Have We Underesimaed he Likelihood and Severiy of Zero Lower Bound Evens? Federal Reserve Bank of San Francisco Working Paper No D Amico, S., T. King (2011): Flow and sock effec of large scale reasury purchases, Federal Reserve Board Finance and Economics Discussion Series Dahlhaus, T., K. Hess, A. Reza (2014): Inernaional Transmission Channels of U.S. Quaniaive Easing: Evidence from Canada, Bank of Canada Working Paper Demary, M., M. Hueher (2015): Does he ECB s unconvenional moneary policy endanger he exi from he curren low ineres rae environmen? IW Policy Papers 7, Cologne Insiue for Economic Research. Engen, E., Th. Laubach, D. Reifschneider (2015). The Macroeconomic Effecs of he Federal Reserve s Unconvenional Moneary Policies, Finance and Economic Discussion Series , Board of Governors of he Federal Reserve Sysem. Falagiarda, M. (2013): Evaluaing Quaniaive Easing: a DSGE approach, MPRA Paper 49457, Universiy Library of Munich. Frazscher, M., M. LoDuca, R. Sraub (2013): On he Inernaional Spillovers of US Quaniaive Easing, ECB Working Paper Gagnon, J., M. Raskin, J. Remache, B. Sack (2011): Large-scale asse purchases by he Federal Reserve: did hey work? Federal Reserve Bank of New York Saff Repors 441. Greenwood, R., D. Vayanos (2010): Price Pressure in he Governmen Bond Marke, American Economic Review 100, Hamilon, J., J. Wu (2011): The effeciveness of alernaive moneary policy ools in a zero lower bound environmen, Journal of Money, Credi, and Banking 44, Harrison, R. (2012): Asse purchase policy a he effecive lower bound for ineres raes, Bank of England Working Paper 444. Harwig Lojsch, D., M. Rodríguez-Vives, M. Slavík (2011): The size and composiion of governmen deb in he euro area, ECB Occasional Paper 132. Joyce, M., A. Lasaosa, I. Sevens, M. Tong (2011): The financial marke impac of quaniaive easing in he Unied Kingdom, Inernaional Journal of Cenral Banking 7, Joyce, M., Z. Liu, I. Tonks (2014): Insiuional invesor porfolio allocaion, quaniaive easing and he global financial crisis, Bank of England Working Paper 510. Krishnamurhy, A., A. Vissing-Jorgensen (2011): The Effecs of Quaniaive Easing on Ineres Raes: Channels and Implicaions for Policy, Brookings Papers on Economic Aciviy 43,

17 Li, C., M. Wei (2012): Term Srucure Modelling wih Supply Facors and he Federal Reserve s Large Scale Asse Purchase Programs, Federal Reserve Board Finance and Economics Discussion Series Lim, J., S. Mohapara, M. Socker (2014): Tinker, Taper, QE, Bye? The Effec of Quaniaive Easing on Financial Flows o Developing Counries, World Bank Policy Research Working Paper Liu, P., H. Mumaz, K. Theodoridis, F. Zanei (2015): Changing macroeconomic dynamics a he zero lower bound, mimeo. Nelson, E. (2013): Key aspecs of longer-erm asse purchase programs in UK and US moneary policy, Oxford Economic Papers 65, Schmi-Grohé, S., M. Uribe (2003): Closing small open economy models, Journal of Inernaional Economics 61, Trebesch, C., J. Zeelmeyer (2014): ECB inervenions in disressed sovereign deb markes: The case of Greek bonds, CESifo Working Paper Vayanos, D., J.-L. Vila (2009): A preferred-habia model of he erm srucure of ineres raes. NBER Working Paper Vogel, L. (2014): Srucural reforms a he zero bound, European Economy Economic Papers 537. Woodford, M. (2001): Fiscal requiremens for price sabiliy, Journal of Money, Credi and Banking 33,