歐洲央行-央行長期資產(chǎn)負債表：宏觀視角（英）

EUROPEANCENTRALBANK

EUROSYSTEM

DiscussionPaperSeries

PeterKaradi,DominikThaler,OresteTristaniThecentralbank’sbalancesheetinthelongrun:amacroperspective

No25

Disclaimer:ThispapershouldnotbereportedasrepresentingtheviewsoftheEuropeanCentralBank(ECB).TheviewsexpressedarethoseoftheauthorsanddonotnecessarilyreflectthoseoftheECB.

1

ECBDiscussionPaperSeriesNo25

Discussionpapers

Discussionpapersareresearch-basedpapersonpolicyrelevanttopics,offeringabroaderandmorebalancedperspective.Whilebeingpartlybasedonoriginalresearch,theyplacetheanalysisinthewidercontextoftheliteratureonthetopic.Theyalsoconsiderexplicitlythepolicyperspective,withaviewtodevelopanumberofkeypolicymessages.Theirformatofferstheadvantagethatalternative

analysesandperspectivescanbecombined,includingtheoreticalandempiricalwork.TheselectionanddistributionofdiscussionpapersaresubjecttotheapprovaloftheDirectorGeneraloftheDirectorateGeneralResearch.

2

ECBDiscussionPaperSeriesNo25

Abstract:Westudywhetheritisdesirableforthecentralbanktosupplyreservesabundantly,i.e.beyondthelevelthatsatisfiesfinancialinstitutions’aggregate

liquidityneeds.Usingatheoreticalframework,wedemonstratethatabundant

reserveswouldhelpfulfiltheprivatesector’sdemandforsafeandliquidassets,

becausereservesaffectfinancialinstitutions’leverageconstraints.Morespecifically,

systematiccentralbankpurchasesofmedium-termgovernmentbondsfromfinancialinstitutionswouldrelaxthoseinstitutions’leverageconstraintsandallowthemto

expandtheirbalancesheetsandissuemoreprivateliquidity,intheformofdeposits.However,averylargeincreaseintheaveragesizeofitsbalancesheetwould

exposethecentralbanktotheriskoflargefinanciallosses.Onbalance,onlya

moderatelylargersupplyofreservesthanthelevelthatsatisfiesfinancialinstitutions’aggregateliquidityneedsappearsdesirable.

JEL:E41,E44,E58.

Keywords:CentralBankReserves,LiquidityPolicy,FriedmanRule

3

ECBDiscussionPaperSeriesNo25

Non-TechnicalSummary

Westudywhetheritisdesirableforthecentralbanktosupplyreservesabundantly,i.e.beyondthelevelthatisnecessaryforfinancialinstitutionstosatisfytheir

aggregateliquidityneeds.

Thefirstsectionofthepapersurveystheevidencedemonstratingthatsafeand

liquidassetswereundersuppliedbeforetheGreatfinancialcrisisbutbecamemoreabundantwiththelargeincreaseincentralbankreservesaccompanyingthe

implementationofunconventionalmonetarypolicymeasures.Thisevidence

suggeststhatapermanentincreaseinthesupplyofcentralbankreserves,throughwhatwedefineas“extendedliquiditypolicy”,couldbebeneficial.However,centralbankreservescanonlybeheldbyfinancialinstitutions.Wouldtheincreasedbankliquiditycontributetosatisfytheliquiditydemandoftherestofthenon-financial

privatesector?Whatwouldbethesideeffectsofanextendedliquiditypolicyonthefinancialrisksofthecentralbank?

Inthesecondsectionofthepaper,weanswerthefirstquestionrelyingona

theoreticalmodel.Morespecifically,westudywhetheralargestockofreserves

systematicallyavailableforfinancialinstitutionswouldinducethemtocreatemoreliquidityforthenon-financialsector,intheformofbankdeposits.Wedemonstrate

thatthisshouldbethecasebecausereservesaffectbanks’leverageconstraints.

Sincereserveshavealowerriskweightthanotherassets,includinggovernment

bonds,centralbankpurchasesfrombanksofmedium-andlong-termgovernment

bondswouldrelaxbanks’leverageconstraints.Theycouldthereforeexpandtheir

balancesheetsandissuemoredeposits.Weconcludethatanextendedprovisionofcentralbankliquidityislikelytotranslateintoanincreasedavailabilityofliquidityalsoforthenon-financialprivatesector.

Wediscusstheeffectsofanextendedprovisionofliquidityontheriskexposureofthecentralbankinthelastsectionofthepaper.Alargerbalancesheetcouldleadtooccasionalmarked-to-marketlossesforthecentralbank.Whilecentralbanklossesarenotproblematicperse,theymay,iflarge,leadtoalossofcontroloverinflation.Theseconsiderationsrestrictthescopeforanextendedliquiditypolicy.Furthermore,theyfavouracompositionofthebalancesheettiltedtowardssaferassetsandlowermaturitytransformation.

Overall,anextendedliquiditypolicywouldproducesomebenefits,butalsoside

effects.Onbalance,amoderatelylargersupplyofreservesthanpriortotheglobalfinancialcrisisappearsdesirable.Itwouldhelpsatisfytheprivatesector’sliquiditydemandwhileminimisingtheensuingfinancialriskforthecentralbank.

4

ECBDiscussionPaperSeriesNo25

1Introduction

InmanyindustrialcountriesthestockofreservesincreasedconsiderablyaftertheGlobalfinancialcrisis,asaresultofthedeploymentofvariousformsof

unconventionalmonetarypolicymeasures.Asthosemeasuresareslowly

reabsorbed,itisnaturaltoaskwhetherthesupplyofreservesshouldbereduced

proportionally,orwhetheritshouldremainabundant,andtowhatextent.Acertain

increaseinthestructurallevelofreservesappearsnecessary,forexamplebecausefinancialinstitutionsarelikelytousereservestofulfilnewliquidityrequirements.

Limitingtheincreasetosuchminimallyhigherlevelwouldbeconsistentwiththepre-crisispracticeandwiththeprincipleofminimisingthemarketfootprintofthecentralbank.Alargerlevelofreserveswouldarguablyhelptosatisfytheprivatesector’s

demandforsafeandliquidassets.

Thefirstsectionofthepaperreviewstheevidenceonthedemandforsafeandliquidassets.TheempiricalliteraturebasedontheU.S.hasconsideredasliquidallshort-termassetsissuedbythepublicsector,includingcentralbankreservesandshort-

termTreasurybills.ThesteepslopethatusedtocharacterisetheTreasuries’yield

curveatveryshortmaturitiesbeforethefinancialcrisisisconsideredasanindicatorofaliquiditypremiumthatinvestorsarewillingtopaytoholdliquidassets.Despitetheshortavailablesampleperiod,weshowcomparableevidenceforAAAtreasurysecuritiesalsointheeuroarea.Thisevidencesuggeststhatshort-termsafeand

liquidassetswereundersuppliedbeforethelargeexpansionofcentralbankreservesthatoccurredwiththefinancialcrisis.

OneoptiontosatisfythisdemandisforTreasuriestoissuemoreshort-termbillsorswaplong-termbondsforshort-termbills.SinceTreasuriesaretradable,theywouldprovideliquiditytothewholeprivatesector.Asecondoptionistoincreasethesupplyofcentralbankreservesthroughwhatwedefineas“extendedliquiditypolicy”.

ComparedtoincreasingthesupplyofTreasuries,extendedliquiditypolicywould

onlyprovideliquiditytofinancialinstitutions.Sincebankswouldneedtoholdlargequantitiesofreserves,thispolicyalsoriskscrowdingoutthesupplyofbankcredit.Isanextendedliquiditypolicyatalldesirable?Wouldatleastpartoftheincreased

liquidityspillovertothenon-financialsector?

Inthesecondsectionofthepaper,werelyonatheoreticalmodeltoanswerthis

question.Thepremiseofthemodelisthatthereturnonsafeandliquidassetswill

tendtoincorporateapremium,whichcanalsobeviewedastheopportunitycostofholdingtheseassetsinsteadofotherrisk-freeassetsthatdonotprovideliquidity

services.Wethenaskwhetherthepremiumonliquidassetsheldbythenon-

financialprivatesector,suchasbankdepositsorshort-termTreasuries,canalsobeexpectedtofall,asthecentralbankprovidesincreasinglylargeamountsofreserves.

Weanswerthequestionintwosteps.Inthefirststepweidentifyatheoretical

conditionwhichmaywarrantaboundedsupplyofcentralbankreserves,notablyasupplybelowthesatiationlevelrecommendedbytheFriedmanrule,or,equivalently,aremunerationofreservesbelowthemarketrate.Theconditionisthatfinancial

5

ECBDiscussionPaperSeriesNo25

institutionsbesubjecttoaleverageconstraintapplyingtotheirtotalassets.Duetotheconstraint,financialintermediariesneedtofinancetheirinternalcapital,and

thereforerequireaspreadbetweenthereturnontheirassetsandthecostoftheirliabilities(deposits).Theequilibriumreturnondepositsmustthereforebelowerthanthereturnontheassetsheldbyintermediaries,includingreserves.Aslongas

reserveshavealowerriskweightthanothermarketableshort-termassets,their

remunerationshouldalsoremainbelowtheshort-termmarketrate.Fromthe

perspectiveofthenon-financialprivatesector,theopportunitycostofholdingliquidbankdepositswillbenon-zero.

Inthesecondstep,wediscusswhethertheopportunitycostcanneverthelessbe

reducedthroughanextendedprovisionofcentralbankliquidityintheformofa

permanentprogrammeofmedium-andlong-termgovernmentbondpurchasesfromfinancialinstitutions.Weshowthat,aslongasreserveshavealowerriskweight

thangovernmentbonds,forexamplebecausetheyarenotsubjecttointerestrate

risk,swappingthemforgovernmentbondswouldrelaxfinancialinstitutions’leverageconstraints.This,inturn,wouldallowtheseinstitutionstoexpandtheirbalance

sheets,andissuemoredeposits.Therefore,anextendedprovisionofcentralbankliquidityislikelytoalsotranslateintoanincreasedavailabilityofliquidity–intheformofbankdeposits–forthenon-financialprivatesector.Thispolicywouldthereforebebeneficial.Abstractingfromotherconsiderations,thescaleoftheassetpurchases

shouldbelarge.

However,anextendedprovisionofliquiditywouldproduceimplicationsonthe

centralbank’sexposuretofinancialrisk.Whileincurringlossesisnotimmediatelyproblematicforcentralbanks,becausetheyhavetheuniqueabilitytoissuemoneyandlossesdonotautomaticallyposesolvencyproblems,verylargelossesmay

eventuallyimplylosingcontroloverinflationif,forexample,thegovernmentis

unwillingtorecapitalisethecentralbankthoughafiscaltransfer.These

considerationsrestrictthescopeforanextendedliquiditypolicy.Furthermore,theyfavouracompositionofthebalancesheettiltedtowardssaferassetsandlower

maturitytransformation.Apartialqualificationofthisconclusionisthat,asshownbytherecentexperience,centralbanklossesmayneverthelessbeincurredduringtherecoveryfollowingthedeploymentofQE-typeinterventions.Assumingthatsuch

lossesinbadtimesareunavoidable,theycouldbepartiallycompensatedby

averagegainsifthecentralbankheldanon-zeroassetportfolioalsoingoodtimes.

Overall,extendingcentralbankliquidityisnotwithoutsideeffects,thereforethe

completeeliminationoftheliquiditypremiumthroughabundantliquidityprovisionisundesirable.Inotherwords,theFriedmanruleisausefulbenchmark,butitneedstobequalified.Thesideeffectwefocusonpointsoutthatreservescanonlybeheldbybanks,whichfacebindingbalance-sheetconstraints.Reserves,therefore,cancrowdoutbeneficiallendingopportunitiesandanoverlylowliquiditypremiumcan

adverselyaffectbanks’profitabilityand,thus,limittheirlendingcapacity.

Furthermore,assetsheldbythecentralbankasacounterpartofitsreservecreationcanleadtoundueexposuretofinancialrisk.Onbalance,evidenceindicatesthat

pre-financialcrisiscentral-bankliquidityprovisionmighthavebeenoverlyscarce.A

6

ECBDiscussionPaperSeriesNo25

somewhatlargerbalancesheetthatsatisfiesmoreliquiditydemand,whilekeepingsideeffectscontained,wouldthereforebejustified.

7

ECBDiscussionPaperSeriesNo25

2Thecentralbankbalancesheetasatoolforextendedliquiditypolicy

Figure1providesastylisedcharacterisationofthecentralbankbalancesheet.Ontheassetsside,thebalancesheetcanincludeholdingsofgovernmentbondsandlendingoperationstofinancialinstitutions.Ontheliabilitiesside,onefindscurrencycirculatinginthepublicandreservesheldbyfinancialinstitutions,plusthecapitalofthecentralbank.Currencyandreserves,orthemonetarybase,arethekey

monetarypolicy-relatedvariablesthataffecttheoverallsizeofthecentralbank

balancesheet.Sincetheamountofcurrencycirculatinginthepublicdoesnot

changetoomuchovertime,toafirstapproximationthesizeofthebalancesheetisdrivenbytheamountofreserves.1Wethereforeabstractfromcashinthepaper.

Figure1

Stylisedcentralbankbalancesheet

assets

liabilities

cash

reserves(Rm)

capital

gov'tbonds(Rb)

lendingtobanks

Note:RbandRmwilldenotethegrossratesofreturnongovernmentbondsandreserves,respectively.

Theamountofreservesisoftenleftunspecifiedinmoderntreatmentsofmonetary

policy.Thisapproachisbasedontheviewthattheobjectivesofmonetarypolicycanbeachieved,andareachievedinthepracticeofcentralbanks,atleastinnormal

times,throughtheappropriatesettingofashort-terminterestrate.Suchinterestratecanbesetthroughopenmarketoperationsor,asintheeuroareacase,through

collateralisedlendingoperationswithbanks.Theseoperationswillberecordedinthecentralbankbalancesheet,triggeringademandforcentralbankliquidityandthe

creationofreserves.Aslongasthedemandforliquidityissatisfied,theexact

quantityofreservesthatiscreatedisinconsequentialforthedeterminationofthemonetarypolicystance.Interestratepolicyandreservessupplypolicycanbesetindependentlyofeachother.2

1CurrencyincirculationhasfollowedanincreasingtrendsincethestartofEMU.Nevertheless,itsdynamichasbeencontainedcomparedtothechangesrecordedintheamountofreserves.

2See,forexample,CurdiaandWoodford(2011).

8

ECBDiscussionPaperSeriesNo25

2.1Financialinstitutions’liquidityneeds

Whileashort-terminterestratedeterminesthemonetarypolicystance,thesupplyofreservesensuresthattheappropriateamountofcentralbankliquidityisavailabletothefinancialsystem.Ademandforliquiditycouldarisefromtransactionneeds,ortofulfilthedesiretomaintainabufferofliquidsavings.Idiosyncraticneedsmaybe

satisfiedthroughinterbanktransactions,buttheresidualliquiditydemandoftheaggregatebankingsystemmustbesatisfiedbythecentralbank.

Intheory,liquiditysupplycouldbedeterminedbythecentralbankintwoways.Afirstapproachwouldbetodirectlyspecifyitsquantity.Thisapproachcouldbe

cumbersomeinpracticeifaggregateliquiditydemandissubjecttolarge,stochasticchangesovertime,forexampleinreactiontoliquidity-preferenceshocks.The

centralbankwouldneedtoassessatahighfrequencytheappropriatequantityof

reservestosupply,dependingondemandconditions.Asecondapproachistoaffectthequantityofliquidityindirectly,bysettingitsopportunitycost,andlettingthe

quantitybedemand-determined.

ThissecondapproachhassimilaritieswiththepracticeoftheECBafterthefinancialcrisis.3Overthisperiod,centralbankliquiditywasabundantandtheremunerationofexcessreservesbecameveryclosetotheovernightrate.Itisexpectedthatthiswillalsobethecaseinthefuture(seeSchnabel,2024,foradiscussionofthenewECBoperationalframework).4

Inpractice,financialinstitutions’demandforreserveswillbeatleastaslargeas

necessarytosatisfyminimumliquidityneeds.Attheaggregatelevel,andabstracting

fromperiodsoffinancialdisruption,minimumliquidityneedsaretraditionally

determinedbyreserverequirementsandbyso-calledautonomousfactors,suchasgovernmentdeposits.

Forgivenreserverequirements,thequantityofreservesincreaseswiththestockofbankdeposits.TheleftpanelofFigure2presentsevidenceofasecularincreaseinbankdepositsintheeuroareabothinnominaltermsandasashareofGDP.The

increaseindepositsraisesthestructuralliquiditydemandofbanks,which,therefore,canbeexpectedtobehigherthanbeforetheGlobalFinancialCrisis.

Afterincreasingmarkedlyduringtheperiodofnegativeinterestrates,autonomousfactorsareexpectedtobecomesmallerinthefuture,buttheyalsoappeartohavebecomemoreunpredictable(Baldoetal.,2017;KinseleandLizarazo,2022).A

somewhatlargerliquiditysupplymaythusbedesirableforprecautionaryreasons,thatistoavoidthatunexpectedincreasesinautonomousfactorsubtractliquidityfromfinancialinstitutions.

3Beforethefinancialcrisis,theECBusedtoremunerate“excessreserves”,i.e.reservesinexcessoftheamountrequiredtosatisfythereserverequirement,ataratelowerthanthe(overnight)marketrate.Thisapproachimpliedthatexcessreservescarriedanopportunitycost.However,thequantityof

excessliquidityprovidedtobankswascappedbytheECB,itwasnotdemand-determined.

4Requiredreservesareanexception.On27July2023,theECBhadalreadydecidedthatrequiredreserveswouldnolongerberemunerated.

9

ECBDiscussionPaperSeriesNo25

Afinalfactorthatislikelytohavecausedastructuralchangeinthedemandfor

centralbankreservesarethenewbankingregulationmeasuresintroducedaftertheGlobalFinancialCrisis.Thesewereeitheraimedtoensuresufficientcapitalagainstanybankexposures,theleverageratio,ortolimitliquidityrisk,theliquiditycoverageratio(LCR)andthenetstablefundingratio(NSFR).Whiletheleverageratiolimitsthedemandforreserves,theLCRandNSFRarelikelytoraiseit.

Figure2

Evolutionofbankdepositsandbankleverageintheeuroarea

Bankdeposits(left).Euroareabankleverage(right).

18

16

14

12

10

8

6

4

2

0

Deposits,shareofannualGDP,rightaxisDeposits,trillioneuro

18

17

16

15

14

13

12

11

10

160%

140%

120%

100%

80%

60%

40%

20%

1997

1999

2001

2003

2005

2007

2009

2011

2013

2015

2017

2019

2021

2023

0%

EAbanks:Leverage

leverageratioreliefperiod

2000

2001

2002

2004

2005

2006

2008

2009

2010

2012

2013

2014

2016

2017

2018

2020

2021

2022

Source:EuropeanCentralBank,Eurostat.

Note:TheleftpanelshowstheevolutionofthebankdepositsintheeuroareabothinnominaltermsandrelativetoannualGDP.Inparticular,itshowsthedepositsofthegeneralgovernmentandtheprivatesectorintheconsolidatedeuroareaMFIbalancesheet.Therightpanelshowstheevolutionoftheratiobetweentotalassetsoftheaggregatebankingsectorandtheirequity(capitalplusreserves).

Theregulatoryleverageratiocanbeexpectedtolimitbanks’demandforcentral

bankreserves,becauselargerreservesholdingsmustbeaccompaniedbyan

increaseininternalfunds.TheBaselCommitteeonBankingSupervisionintroducedtheleverageratiointhe2010BaselIIIpackageofreforms.Asof2018theleverageratiobecameabindingrequirement.TherightpanelofFigure2showsthatthis

measurestartedproducingeffectsveryquickly.Aggregateleverageoftheeuroareabankingsector,representedinthefigurebytheratiobetweentotalassetsand

internalequity,fellsignificantlybetween2010and2013.DespitethelaunchoftheECB’sAssetPurchaseProgrammein2015,bankleverageremainedstableuntil

2019andonlyincreasedagainwiththeoutbreakoftheCOVID-19pandemic,whentheleverageratiowastemporarilysuspended.

Bycontrast,theLCRrequirementcanstructurallyincreasebanks’demandfor

reserves.Itrequiresbankstoholdastockofhigh-qualityliquidassetsabovea

certainreferencelevel,andcentralbankreservesabovethereserverequirementqualifyashigh-qualityliquidassets.Similarly,liquidityprovidingoperationswitharesidualmaturityaboveoneyearcanbeusedtoimprovebanks’NSFR,whichaimsatlimitingbanks’liquidityandmaturitytransformation(SugoandVergote,2020).

10

ECBDiscussionPaperSeriesNo25

Insum,thedemandforliquidityofthebankingsectorappearstohaveincreased.

Centralbankreserveswillthereforehavetoincreaseaccordingly,toensurethatthebankingsectorcansatisfyitsliquidityneeds.

Intherestofthissection,wereviewtheliteraturecallingforanincreaseinreservesbeyondsuchminimumlevel.Wedefinethisapproachas“extendedliquiditypolicy”.Wereviewavailableevidenceonthebroaderprivatesector’sdemandforsafeandliquidassetsandsomeconceptualargumentsonthedesirabilityofanextended

liquiditypolicy.Byandlarge,weconcludethatthispolicyhassomebenefits,butitalsoentailscosts.

2.2Scarcityofliquidandsafeassets

Therewasanexcessdemandforinterest-bearing,money-likeinstrumentsthatarebothliquidandsafebeforetheGreatFinancialCrisis.Greenwood,Hanson,and

Stein(2016)showthattheUSTreasuryyieldcurvehadanexcesssteepnessatveryshortmaturities–seetheleftpanelofFigure3.Thisissuggestiveevidenceontheexistenceofaliquidityor‘moneyness’premiumthatinvestorsarewillingtopayto

holdtheseshort-termassets,overandabovetheirpecuniaryreturn.TheevidencealsosuggeststhattheamountofreservessuppliedbytheFedatthetimewasnotsufficienttofullyeliminatethisliquiditypremium.

TherightpanelofFigure3presentsrelatedevidencefortheeuroarea.Inparticular,itshowstheshapeoftheeuroareaAAAyieldcurvebeforethefinancialcrisis.InlinewiththeUSevidence,theslopeoftheyieldcurvebelow1-yearmaturityisinexcessofitsslopeabove1-yearmaturity.Thepositivepremiumisconsistentwiththe

assumptionthattheseshort-termsafeassetswereundersuppliedalsointhisregion.

Figure3

MoneynesspremiumintheUS(Greenwoodetal.,2016)andEuroArea

YieldspreadsonUSTreasuryBills(left).YieldsonAAAEAgovernmentbonds(right).

(Left:basispoints.Right:%.)

0

-20

-40

-60

-80

%

Yieldspreadto26-WeekT-billZ-spread

0510152025

Weekstomaturity

4.0

3.8

3.6

3.4

3.2

3.0

YieldsonAAAEAgovernmentbonds

012345678910

Yearstomaturity

Source:Greenwood-Hanson-Stein,2016,Chart3A(leftfigure),EuropeanCentralBank(rightfigure)

11

ECBDiscussionPaperSeriesNo25

Note:ThefiguresshowaverageyieldsandyieldspreadsofUSandeuroareagovernmentbonds.Theleftpanelplotstheaverage

spread,overtheperiod1983to2009,betweenT-billyieldsofmaturitiesfromoneto26weeksandtheyieldonthe26-weekbill.Italsoplotsthez-spread,definedasthedifferencebetweenT-billyieldsandfittedyields,wherefittedyieldsarebasedontheflexible

extrapolationoftheTreasuryyieldcurvefromGürkaynak,Sack,andWright(2007).Therightpanelplotstheaverageyield,overtheperiod2004-2008,offittedAAAeuroareagovernmentbondyieldsofmaturities3monthsand10yearsrelyingonthedailyECByieldcurveestimates.

Whycannottheprivatesectorsatisfytheexcessdemandforliquidandsafeassets?Stein(2012)arguesthatapotentialreasonisawell-understoodinnatetension

betweenliquidityandsafety.Liquidityrequiresthattheassetcanbesoldatminimallossatveryshortnoticeincaseasuddenliquidityneedarises.Bankscansatisfy

suchneedsbytransformingilliquid,long-terminvestmentintoliquiddeposits

(Diamond-Dybvig,1983).Theycandothisbecauseliquidityneedsofdifferent

depositorsnormallyariseinastaggeredfashion.Thesafetyissuerelatedtosuch

depositsarisesbecauseofacoordinationproblem:thebankcannotsatisfythe

liquidityneedofallitsdepositorsatthesametimeincaseofabankrun.Thereasonisthattheilliquidassetsitholdscannotbesoldatshortnoticewithoutfire-sale

losses.Inotherwords,thedeposits,similarlytootherprivatelycreatedliquidassets(likemutualfundshares)are‘runnable’andthereforerisky.

Anargumentforextendedliquiditypolicyisthatalargestockofreservescanhelpmitigatethescarcityofliquidassetsandpotentiallyeliminatethemoneyness

premium.Thereasonisthatreservesaresafepublicliquidassets,sothecentral

bankisinagoodpositiontoaddressthismarketfailure(Greenwood,Hanson,Stein,2016).Thecentralbankcanachievethisbymaintainingastructurallongrun

balancesheetthatreplaceslessliquidlong-termgovernmentbondsandbankloanswithliquidreserves.

2.3Scarcityofreservesversusscarcityofconvenientassets

Lopez-SalidoandVissing-Jorgensen(2023)provideadditionalevidencefromtheUS,whichshowsthathigherreservesupplyindeedincreasestheliquidityofbanksandreducesliquiditypremiaintheinterbankmarket.Inparticular,theydocumentastablenegativerelationshipbetweenreservesandtheinterbank-ratespreadandapositiverelationshipbetweenbankdepositsandtheinterbankspread.Vissing-

Jorgensen(2023)showsthatasimilarstablerelationshipexistsintheeuroareaaswell.Bothpapersemphasisethatthegainsfromthisliquidityprovisionneedtobebalancedwiththeriskscomingfromreservespartiallycrowdingoutthesupplyofbankcredittotheprivatesector.

Anadverseeffectofincreasingthesupplyofreservesthroughpurchasesof

governmentbondsisthecorrespondingreductioninthestockofgovernmentbondsavailabletotheprivatesector.Krishnamurthy-Vissing-Jorgensen(2012)andVissing-Jorgensen(2023)showthatgovernmentbondsissuedbysafesovereigns,liketheUSorGermanycarrya‘convenienceyield’,asinvestorsderiveconvenience(utility)fromtheirhighliquidityandsafety.FocusingontheAAAbonds,Table1showsan

indicatoroftheconvenienceyieldintheeuroarea,namelythespreadbetweenbondyieldsandtheOISrateforacorrespondingmaturity(wheretheOISrateisusedasaproxyforapure,risk-freerate).Thenegativevaluesrevealaconvenienceyieldof

safegovernmentbonds.Lowvaluesattheshortestsegmentsareconsistentwiththe

12

ECBDiscussionPaperSeriesNo25

moneynesspremiumofshortandsafegovernmentbondsdiscussedintheprevioussection,butthetableshowsthattheconveni