WP/09/251 Monetary and Macroprudential Policy Rules in a Model with House Price Booms Prakash Kannan, Pau Rabanal, and Alasdair Scott © 2009 International Monetary Fund WP/09/251 IMF Working Paper Research Department Monetary and Macroprudential Policy Rules in a Model with House Price Booms Prepared by Prakash Kannan, Pau Rabanal, and Alasdair Scott1 Authorized for distribution by Jörg Decressin November 2009 Abstract This Working Paper should not be reported as representing the views of the IMF. The views expressed in this Working Paper are those of the author(s) and do not necessarily represent those of the IMF or IMF policy. Working Papers describe research in progress by the author(s) and are published to elicit comments and to further debate. We argue that a stronger emphasis on macrofinancial risk could provide stabilization benefits. Simulations results suggest that strong monetary reactions to accelerator mechanisms that push up credit growth and asset prices could help macroeconomic stability. In addition, using a macroprudential instrument designed specifically to dampen credit market cycles would also be useful. But invariant and rigid policy responses raise the risk of policy errors that could lower, not raise, macroeconomic stability. Hence, discretion would be required. JEL Classification Numbers: E52, E61 Keywords: house prices, monetary policy, macroprudential Author’s E-Mail Address: [email protected], [email protected], [email protected], 1 The authors would like to thank Olivier Blanchard, Charles Collyns, Jörg Decressin, Antonio Fatás, Jordi Galí, David Romer, and seminar participants at the Bank of Canada, Bank of England, Board of Governors of the Federal Reserve, IMF, and Her Majesty’s Treasury for thoughtful discussions and comments. The remaining errors are our own. 2 Contents Page I. Introduction ..................................................................................................................... 4 II. A Model for Analyzing House Price Booms .................................................................. 5 A. Households ............................................................................................................. 6 A.1 Savers ............................................................................................................ 6 A.2 Borrowers ...................................................................................................... 8 B. Financial Intermediaries ....................................................................................... 10 C. Producers.............................................................................................................. 10 C.1 Final goods producers ................................................................................. 11 C.2 Intermediate goods producers ..................................................................... 11 D. Closing the Model: Market Clearing Conditions ................................................. 12 III. Policy Regimes ............................................................................................................. 13 IV. Calibration..................................................................................................................... 14 V. Simulation Results ........................................................................................................ 15 A. The Performance of Policy Rules in Reaction to Financial Shocks .................... 15 B. The Performance of Policy Rules in Reaction to Productivity Shocks ............... 17 C. Policy Rules with Multiple Shocks ...................................................................... 18 VI. Robustness of the Results ............................................................................................. 19 VII. Conclusions ................................................................................................................... 20 References ................................................................................................................................. 22 Tables 1. Parameter Values .......................................................................................................... 24 2. Parameters of Policy Rules in Reaction to Financial Shocks ....................................... 25 3. Performance of Policy Rules in Reaction to Financial Shocks ..................................... 26 4. Parameters of Policy Rules in Reaction to Productivity Shocks .................................. 27 5. Performance of Policy Rules in Reaction to Productivity Shocks ................................ 28 6. Sensitivity of Parameters of Policy Rules Optimized to Financial Shocks to Changes in Key Parameters ............................................................................. 29 7. Sensitivity of Parameters of Policy Rules Optimized to Productivity Shocks to Changes in Key Parameters ............................................................................. 30 3 Figures 1. Effect of a Financial Shock ........................................................................................... 31 2. Effect of a Productivity Shock ...................................................................................... 32 3. Optimal Weight on Nominal Credit in the Macroprudential Rule ............................... 33 Appendix: Linearized Conditions ............................................................................................. 34 4 I. Introduction Empiricalworkhasfoundanumberofvariablesthatareoftenassociatedwithfutureasset pricebusts,suchascredit,theshareofresidentialinvestmentinGDP,andcurrentaccount balances.2 Thosesamevariablesdoreasonablywellinexplainingthedifferencesacross countriesinhousepricerisesleadinguptothecurrentcrisis.3 Suchevidencesuggeststhatif centralbankerswishtomitigatedamagingassetpriceboom-bustcycles,theyshouldconsider reactingtosuchvariablesratherthanfocusingmainlyoninflationandoutput. However,there arestillanumberofimportantquestionstobeaddressedinassessingtheappropriatepolicy responses: Whatarethepotentialgainsfromreactingtosignsofemergingfinancialvulnerability? (cid:15) Ismonetarypolicytheappropriatetoolforreactingtosuchindicators,orshouldother (cid:15) policiesbeused? Whatarethetrade-offsbetweenfocusingpolicyonstabilizingoutputandCPIinflationand (cid:15) attemptingtoreducetheriskofassetpricecrashes? Thispaperaddressesthesequestionswithsimulationsconductedusingamodeleconomythat hassomeofthekeyfeaturesrelevantforexaminingthepotentialroleofmonetarypolicyin mitigatingtheeffectsofassetpricebooms. Itmakesthecasethatputtingmoreemphasison macrofinancialriskcouldprovidestabilizationbenefits. Theresultsfromthesimulations suggestthatastrongermonetaryreactiontosignsofoverheatingorofacreditorassetprice bubblecouldhelpcounteracceleratormechanismsthatpushupcreditgrowthandassetprices. Inaddition,usingamacroprudentialinstrumentdesignedspecificallytodampencreditmarket cycleswouldalsobeuseful. Butexpectationsshouldberealisticaboutwhatcanbeachieved withsuchanapproach. Inparticular,itisoftendifficulttoaccuratelyidentifythesourceofthe shockdrivingobservedhousepricebooms. Invariantandrigidpolicyresponsesraisetherisk ofpolicyerrorsthatcouldlower,notraise,macroeconomicstability. Hence,discretionwould needtobeappliedtotheuseofmonetaryandmacroprudentialpoliciesinattemptingto mitigatemacrofinancialrisks. Howdotheseconclusionscomparewiththosefromotherstudies? Thereisavastliteratureon monetarypolicyandassetprices. Thedebatepersistsoverwhethercentralbanksshouldreact directlytoassetprices. Twowell-knownexamplesareBernankeandGertler(2001),who concludethatthereisnoroleforassetpricesinmonetarypolicyrules,andCecchettietal. 2See,interalia,BorioandLowe(2004)andKannanetal. (2009). 3SeeIMF(2009). 5 (2000),whoarguethatcentralbanksshouldreacttoassetprices. Increasingly,theliterature suggestsgainsfromincludingassetpricesinmonetarypolicyrules.4 Grayetal. (2009)finda roleforafinancialstabilityindicatorinthemonetaryrule. Twoaspectsdifferentiatethis paper. First,wedistinguishbetweengainsthatarisefromaddingindicatorsoffinancialstress suchasexcessivecreditgrowthtostandardbutsecondbestmonetarypolicyrulesandthose fromincludingthesameindicatorsinfullyoptimizedrules. Second,asfarasweknow,thisis thefirsttimethecoordinationofmonetaryandmacroprudentialruleshasbeenformally evaluatedusingamacroeconomicmodelofthistype.5 Thepaperisstructuredasfollows. Section2containsadescriptionofthemodelandSection3 adescriptionofthepolicyregimesevaluatedinthesimulations. Section4summarizesthe model’scalibration. TheresultsofsimulationexperimentsarepresentedinSection5. Inthe followingsection,sensitivityanalysisisperformed. Thefinalsectionconcludes. Anappendix detailsthelinearizedconditionsforthemodel. II. AModelforAnalyzingHousePriceBooms ThemodelusedinthispaperhasanumberofmodificationstothestandardNewKeynesian model(Galí,2009)withregardtothecharacterizationofhouseholdsandfinancialmarkets, whichcreateaspecialroleforassetprices. Becausehousingwealthisgenerallymore importantformosthouseholdsthanequities,andbecausehousepurchasestypicallyrequire debtfinancing,weconcentrateontheroleofhousing.6 First,householdsmakechoicesabout howmuchtoinvestinhousing,aswellashowmuchtoconsumeinnondurablegoods. Housingisanassetthatprovidesservicesandisthemainvehicleforaccumulatingwealthin thiseconomy. Second,wemakeadistinctionbetweenborrowersandlenders,creating conditionsforleverage. Third,thelendingrateismodeledasaspreadoverthepolicyratethat dependsonloan-to-valueratios,themarkupchargedoverfunding(policy)rates,and,insome casesdiscussedbelow,amacroprudentialinstrument. Hence,lendingratescanchangefora numberofreasons: forexample,ariseinhousepriceswillraisemarketvaluationsof borrowers’collateral,loweringtheaverageloan-to-valueratio,andwillthereforeleadtoafall inlendingratesevenifmonetarypolicyhasnoteased. Creditmarketconditionscan change—becauseof,say,changesinperceptionsofriskorcompetitivenessinlending—which 4RecentexamplesincludeChristianoetal. (2007). 5GrussandSgherri(2009)studythewelfareimplicationsofprocyclicalloan-to-valueratiosinatwo-country model. However, because the model does not have a nominal side, the reaction of monetary policy cannot be addressed. 6For a model that considers the monetary policy implications of stock price fluctuations, see Christiano and others(2007). 6 couldleadbankstoadjusttheirmarkupsandthereforealterthelendingspread. Bothofthese mechanismshelpaccelerateariseinresidentialinvestment,nondurableconsumption,and prices.7 Insomesimulations,policymakerscanaffectspreadsdirectly,usinga macroprudentialtool,inadditiontoinfluencinglendingratesviapolicyrates. Inotheraspects,themodelhasconventionalNewKeynesianfoundations. Thetheoretical frameworkconsistsofageneralequilibriumtwo-sectormodel,durablesandnon-durables, whereeachsectoroperatesundermonopolisticcompetitionandnominalrigidities. Pricesin bothsectorsarestickyintheshortrun,asinCalvo(1983). Consumptionandresidential investmentadjustslowlydue,respectively,tohabitformationandadjustmentcosts. Itis costlyforworkerstoshiftfromworkingontheproductionofconsumptiongoodstobuilding houses,andviceversa. Forsimplicity,thereisnocapitalusedinproductionandtheeconomy isclosed. A. Households Householdsobtainutilityfromconsumingthestockofdurablesandtheflowofnondurables. Therearetwotypesofhouseholdsinthiseconomy,borrowersandsavers. Borrowersare assumedtobemoreimpatientthansavers,byhavingasmallerdiscountfactor. Inequilibrium, saverswillprovidefinancingtoborrowers. Afraction(cid:21)ofhouseholdsareconsideredtobe savers,theremainingfraction1 (cid:21)areborrowers. (cid:0) A.1 Savers Eachsaverj [0;(cid:21)]maximizesthefollowingutilityfunction: 2 Lj 1+’ E 1 (cid:12)t (cid:13)log(Cj "C )+(1 (cid:13))log(Dj) t ; (1) 0(t=0 " t (cid:0) t(cid:0)1 (cid:0) t (cid:0) (cid:0)1+(cid:1) ’ #) X whereCj denotesconsumptionofnon-durablegoods,Dj denotesconsumptionofdurable t t goods,andLi;j denoteshoursworkedbyhouseholdj ineachsectori = C;D. (cid:12) isthe t discountfactor. Householdsformexternalhabitsinconsumption,asinSmetsandWouters (2003)andIacovielloandNeri(2009),with"denotingtheimportanceofthehabitstock, whichislastperiod’saggregateconsumption(C ). Finally,followingIacovielloandNeri t 1 (cid:0) 7ThesefeaturesdrawonelementsofmodelsbyAokietal. (2004), CurdiaandWoodford(2009), Iacoviello (2005), and Monacelli (2009). The accelerator mechanism goes back to Bernanke et al. (BGG, 1998); unlike BGG,theacceleratorinthismodelworksthroughhousingfinanceratherthanfirms’capital. 7 (2009),weassumethatthereisimperfectsubstitutabilityoflaborsupplyacrosssectors,such thatthelabordisutilityindexcanbewrittenas 1 Lj = (cid:11) (cid:19)L LC;j 1+(cid:19)L +(1 (cid:11)) (cid:19)L LD;j 1+(cid:19)L 1+(cid:19)L ; where(cid:19) > 0; (2) t (cid:0) t (cid:0) t L (cid:0) (cid:20) (cid:21) (cid:16) (cid:17) (cid:16) (cid:17) and(cid:11) istheeconomicsizeofeachsector. Thisimperfectsubstitutabilityimpliesthat reallocatinglaboracrosssectorsfollowingashockiscostly. Notethatwhen(cid:19) = 0the L aggregatorislinearinhoursworkedineachsector,sotherearenocostsofswitchingfrom workinginonesectortotheother. Thiscosthelpsthemodeltoexplainpositivecomovement ofrealvariablesinbothsectorsinresponsetoaggregateshocks. Thebudgetconstraintofthesavers,innominalterms,isgivenby PCCj +PDIj +Bj R Bj +WCLC;j +WDLD;j +(cid:5)j; (3) t t t t t (cid:20) t(cid:0)1 t(cid:0)1 t t t t t wherePC andPD arethepriceindicesofdurableandnon-durablegoods,respectively,Wi is t t t thenominalwageineachsectori = C;D,andBj denotessavinginstruments(suchasdebt t instrumentsordeposits)thatborrowersplaceinbanks,atagrossinterestrateofR . (cid:5)j t t denotesnominalprofitsfromintermediategoodsproducingfirmsandbanks,whichare ultimatelyownedbysavers. Ij denotesresidentialinvestment. Weassumethatthelawofmotionofthehousingstock t evolvesasfollows: Ij Dj = (1 (cid:14))Dj + 1 S t Ij; (4) t (cid:0) t(cid:0)1 " (cid:0) Ij !# t t 1 (cid:0) where(cid:14) denotestherateofdepreciationofthehousingstockand,followingChristiano, Eichenbaum,andEvans(2005),weintroduceanadjustmentcostfunction,S(:),whichis (cid:22) (cid:22) (cid:22) convex(i.e. S () > 0). InthesteadystateS = S(cid:19)= 0andS > 0:Theaimofintroducingthis 00 00 costistoallowforthepossibilitythatthemodelcangeneratehump-shapedresponsesof residentialinvestmenttoshocks. Thefirstorderconditionstothehouseholdmaximizationproblemaregivenbythefollowing expressions,where(cid:21) istheLagrangemultiplierassociatedwiththebudgetconstraint,and(cid:22) t t istheLagrangemultiplierassociatedwithequation(4):8 U = (cid:21) PC (5) Ct t t 8Sinceallsaversbehavethesameway,wedropthej subscriptsinwhatfollows. 8 U = (cid:22) (cid:12)(1 (cid:14))E (cid:22) (6) Dt t (cid:0) (cid:0) t t+1 I I I I I 2 (cid:21) PD = (cid:22) 1 S t S t t +(cid:12)E (cid:22) S t+1 t+1 (7) t t t (cid:0) I (cid:0) 0 I I t t+1 0 I I (cid:26) (cid:18) t 1(cid:19) (cid:18) t 1(cid:19) t 1(cid:27) " (cid:18) t (cid:19)(cid:18) t (cid:19) # (cid:0) (cid:0) (cid:0) Absentadjustmentcoststoresidentialinvestment(i.e. S = 0),thesethreeequationscanbe reducedtothefollowingcondition: PD 1 (cid:13) (cid:24)D(C "C ) C "C PD PtC = (cid:0)(cid:13) t tD(cid:0) t(cid:0)1 +(cid:12)(1(cid:0)(cid:14))Et Ct (cid:0) "tC(cid:0)1 PtC+1 t t (cid:20)(cid:18) t+1 (cid:0) t(cid:19) t+1(cid:21) Notethatifthedurablegoodisinfactcompletelynon-durable(i.e. (cid:14) = 1),thiscondition simplyequatesthemarginalutilitiesofconsumptiontorelativeprices. Sincethedurablegood hasaresidualvaluethefollowingperiod,thisinducestheextra-termofholdinganadditional unitofthedurablegood. TheEulerequationfortheconsumptionofnon-durablegoodswithhabitsisstandard: PC C "C 1 = (cid:12)RtEt PCt Ct (cid:0) "tC(cid:0)1 ; (8) (cid:20) t+1 (cid:18) t+1 (cid:0) t(cid:19)(cid:21) andthelaborsupplyconditionstobothsectorsaregivenby (cid:13) WC L’t(cid:0)(cid:19)L(cid:11)(cid:0)(cid:19)L LCt (cid:19)L = C "C PCt (cid:18) t (cid:0) t(cid:0)1(cid:19) t (cid:0) (cid:1) and (cid:13) WD L’t(cid:0)(cid:19)L(1(cid:0)(cid:11))(cid:0)(cid:19)L LDt (cid:19)L = C "C PtC : (9) (cid:18) t (cid:0) t(cid:0)1(cid:19) t (cid:0) (cid:1) A.2 Borrowers Eachborrowerj [(cid:21);1]maximizesthefollowingutilityfunction: 2 1+’ LB;j 1 t E (cid:12)B;t (cid:13)log(CB;j "CB )+(1 (cid:13))(cid:24)Dlog(DB;j) ; (10) 08 2 t (cid:0) t 1 (cid:0) t t (cid:0) (cid:16) 1+(cid:17)’ 39 (cid:0) ><Xt=0 >= 6 7 4 5 > > : ; 9 inwhichallvariableswithaB superscriptaretheborrowers’analogtothesavers’variables above. (cid:12)B < (cid:12) isthediscountfactoroftheborrowers;itisassumedthatborrowersaremore impatientthansavers. Theirbudgetconstraintinnominaltermsisgivenby PCCB;j +PDIB;j +RL BB;j BB;j +WCLC;B;j +WDLD;B;j: (11) t t t t t 1 t 1 (cid:20) t t t t t (cid:0) (cid:0) BorrowerscanobtainloansfrombanksatalendingrateofRL: t Weassumethatthefunctionalformsforaggregatelaborsupply,LB;j,andforthelawof t motionofhousingstock,DB;j;arethesameasinthecaseofsavers. Hence,thefirstorder t conditionsfortheborrowerhouseholdsaregivenby U = (cid:21)BPC; (12) CB t t t U = (cid:22)B (cid:12)B(1 (cid:14))E (cid:22)B ; (13) DtB t (cid:0) (cid:0) t t+1 and IB IB IB IB IB 2 (cid:21)BPD = (cid:22)B 1 S t S t t +(cid:12)BE (cid:22)B S t+1 t+1 : t t t (cid:0) IB (cid:0) 0 IB IB t t+1 0 IB IB (cid:26) (cid:18) t 1(cid:19) (cid:18) t 1(cid:19) t 1(cid:27) " (cid:18) t (cid:19)(cid:18) t (cid:19) # (cid:0) (cid:0) (cid:0) (14) TheEulerequationforborrowersis (cid:21)B = (cid:12)BRLE (cid:21)B : (15) t t t t+1 SubstitutingfortheexpressionofmarginalutilityweobtaintheconsumptionEulerequation forborrowers: PC CB "CB 1 = (cid:12)BRLE t t (cid:0) t 1 ; (16) t t PC CB "C(cid:0)B (cid:20) t+1 (cid:18) t+1 (cid:0) t (cid:19)(cid:21) whilethecorrespondinglaborsupplyconditionstobothsectorsaregivenby LBt ’(cid:0)(cid:19)L(cid:11)(cid:0)(cid:19)L LBt ;C (cid:19)L = CB (cid:13)"CB WPCtC (cid:0) (cid:1) (cid:16) (cid:17) (cid:18) t (cid:0) t(cid:0)1(cid:19) t and LBt ’(cid:0)(cid:19)L(1(cid:0)(cid:11))(cid:0)(cid:19)L LBt ;D (cid:19)L = CB (cid:13)"CB WPtCD: (17) (cid:0) (cid:1) (cid:16) (cid:17) (cid:18) t (cid:0) t(cid:0)1(cid:19) t