43S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

The e-krona and the macroeconomyHanna Armelius, Paola Boel, Carl Andreas Claussen and Marianne Nessén*The authors work in the Payments Department and the Monetary Policy Department of the Riksbank

In this article, we discuss potential implications of an e-krona for the conduct of monetary policy and for macroeconomic developments in general. We argue that a universally accessible, non-interest-bearing e-krona supplied according to demand would establish a zero interest-rate floor for the policy rate and possibly all other interest rates in the economy. The effect of quantitative easing can thereby also be reduced. Furthermore, it is unlikely that the monetary policy transmission mechanism would be strengthened by an e-krona. We also note that international financial flows may increase and induce more exchange rate volatility. Finally, an e-krona could have long-run level effects on economic activity. The effects would be positive if an e-krona improves the efficiency and the resilience of the payment system and negative if an e-krona impinges on the supply of credit and financial stability.

1 IntroductionTheRiksbankiscurrentlyconductingareviewintowhethertoissueadigitalcomplementto physical cash, the so-called e-krona.1Inthisarticle,weanalysepossibleconsequencesofane-kronafortheconductofmonetarypolicyandformacroeconomicdevelopmentsingeneral.

Thediscussionregardingadigitalcentralbankcurrency(CBDC)isnewandaresultoftheongoingdigitalizationofmodernsociety.Butfromatheoreticalpointofview,thequestionsthatarisewhenthinkingaboutthepossibleconsequencesofCBDCsoftenturnouttobeclassictopicsthathavebeeninvestigatedinmacroeconomicsinthepastcenturyormore.Forexample,issuessuchastheliquiditytrap,thelowerboundtomonetarypolicy,insideversusoutsidemoney,andevenmonetarypolicyautonomyandtheclassictrilemmaarise.Thisarticleshedslightonsomeofthesematters.Insomecaseswearriveatfirmresults(conditionalonourassumptions),inothercaseswepresentonlyanoverviewoftheissuesinvolved.ManyofourcolleaguesatothercentralbankshavewrittenaboutCBDCsandtheirpossibleconsequences.Thefocusinthisarticleisonmonetarypolicyandmacroeconomicissues that are important in a Swedish context.

Thearticleisorganizedasfollows.Thenextsubsectiondescribesthekeypropertiesofthetypeofe-kronaanalysedinthisarticle.Section2studiestheimplicationsofsuchane-kronafortheeffectivelowerboundofthemonetarypolicyrateandotherinterestrates.Section3analyses how the transmission of monetary policy to the rest of the economy may be affected.Section4discussesothereffectsofane-kronaontheeconomy.Section5concludes.AppendixAcontainsthetheoreticalmodelthatunderliestheanalysisinsection3.

1 Seethetworeportsonthee-kronareviewpublishedsofar,SverigesRiksbank(2017)and(2018).

* WethankJanAlsterlind,RafaelB.deRezende,MeredithBeecheyÖsterholm,HenrikErikson,JesperHansson,StefanLaséen,JesperLindé,UlfSöderströmandDavidVestinforcommentsandusefuldiscussions.TheviewsexpressedinthisarticlearethoseoftheauthorsanddonotnecessarilycoincidewiththeviewsoftheExecutiveBoardofSverigesRiksbank.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y44

Key characteristics of the e-krona analysed in this articleThereareseveraldesignchoicesforane-krona,includingwhetheritshouldbemeantonlyforsmallpayments,bearinterest,beuniversallyaccessibleandinunlimitedquantities,andsoon.However,thetechnicaldesign,forinstancewhetherornotitshoulduseadistributedledgertechnology,mattersonlytothedegreethatitaffectsane-krona’sfunctionalfeatures.We therefore abstract from technical issues.

Thee-kronaanalysedinthisarticlehasthepropertiesoutlinedintheRiksbank’sfirste-kronareport(SverigesRiksbank2017):2

1. ItisadirectclaimontheRiksbankandspecifiedinSwedishkronor.

2. Itisuniversallyaccessible:bythiswemeanthatitcanbeheldbyfinancialinstitutions,firmsandmembersofthegeneralpublic,bothforeignanddomesticresidents,andwithoutrestrictions.

3. Itissuppliedaccordingtodemand:theRiksbankwillsupplyasmuche-kronaasisdemanded.

Withthesepropertiesane-kronawillbesimilartocashinthesensethatitisuniversallyaccessible(withoutrestrictions)andsuppliedaccordingtodemand.Thepropertiesmayalsobenecessaryconditionsforparitybetweenane-kronaandotherformsoftheSwedishkrona.Furthermore,andimportantly,theyalsoimplythatane-kronawouldconstituteasafeandliquidassetwithessentiallyzerotransactioncoststhatcouldbeheldbyall(includingprofessionalinvestors)andinunlimitedquantities.Thisis,asweshallseeinlatersections,important since some of our key conclusions apply only to such an e-krona. If, instead, an e-kronawerenotuniversallyaccessibleorprovidedonlyinlimitedquantitiesitseffectswould be much smaller.

Inouranalysis,wedistinguishbetweentwocases,whichinturnhaveverydifferentconsequencesfortheconductofmonetarypolicy:

a) The e-krona does not carry interest.Inthiscasethepolicyratecontinuestobethemonetarypolicyinstrumentandtheimplementationofmonetarypolicycanbeconductedlargelyasitistoday(seeNessénetal.2018).However,negativepolicyrates will, as we explain below, most likely not be implementable.

b) The e-krona carries interest.Inthiscase,theinterestrateonane-krona–positiveornegative–couldbecomeamonetarypolicytool,anditwouldhavetobesetinlinewiththeoverallstanceofmonetarypolicy.

2 Impact on the lower bound of the policy rateWebeginbyexaminingtheconsequencesofane-krona–withthecharacteristicsoutlinedabove–onkeyinstrumentsofmonetarypolicy.Normally,wethinkofmonetarypolicyasaimingtoaffectinflationandtherealeconomybyinfluencingmarketinterestrates,theexchangerateandexpectationsaboutfuturepolicyandtheeconomy.Traditionally,theprincipaltoolforinfluencingshort-termmarketratesisthepolicyrateatwhichmonetarypolicycounterparties(typicallybanks)canborrowordeposittheirreservesatthecentralbank.3

Sincetheonsetoftheglobalfinancialcrisistenyearsago,centralbanksinseveraladvancedeconomies,includingSweden,havealsousedotherpoliciestospureconomicactivity.Oneexampleisquantitativeeasing(QE)whichconsistsofpurchasinglargequantitiesoflong-termsecuritieswiththeobjectiveofreducinglong-terminterestrates.

2 Ane-kronathatfulfils1–3andisinterestbearingwillbequitereminiscentofwhatissometimescalled‘reservesforall’(seeNiepelt2018).Apossibledifferencecouldbethatweallowforaspreadbetweenthepolicyrateandthee-kronarate.3 Inacorridorsystem,otherinterestratesthanthepolicyratemaymattertoo.Forexample,inSwedenbankscurrentlydepositreservesattheRiksbankpartlyatthepolicyrate,partlyatthereporateminusafixedspread,seeNessénetal.(2018)fordetails.

45S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

Inthissection,weanalysetheimplicationsofane-kronaforthelowerboundofthepolicyrateandotherinterestratesintheeconomy.Specifically,wedemonstratewhyauniversallyaccessiblezero-intereste-kronathatissuppliedaccordingtodemandmostlikelywillraisethelower bound not only for the policy rate but also for market interest rates. This, in turn, may alsohaveimplicationsfortheefficacyofQE.

2.1 ThecurrentlowerboundtointerestratescomesfromcashThepolicyrateofthecentralbankwastraditionallythoughttobesubjecttoazerolowerbound(ZLB),meaningthatanycutsinthepolicyratebelowzerowouldhavenoeffect.Thereasoningwasthatwiththeoptiontoholdcash,yieldingazerorateofinterest,bankswouldexchangetheirreservesforcashifthepolicyratewassetbelowzero.Likewise,firmsandhouseholds,itwasthought,wouldquicklysubstituteintocashifinterestratesbecamenegative.However,holdingandhandlingcashisriskyandcostlyforfirmsandhouseholdsandforbanks(seee.g.Alsterlindetal.2015).Itiscostlytoacquiresafeandsecuretransportation,storageandinsurance,forinstance.Forbanks,itiscertainlylessexpensivetokeepreservesatthecentralbankthantoholdlargeamountsofcash.Asaconsequence,theRiksbankandsomeothercentralbankshaveinrecentyearsbeenabletosuccessfullyimplementnegativepolicyrates.However,thereisstillalimittohowlowthepolicyratecangoandstillhaveaneffectonmarketrates.Thislimitisdeterminedbytherisksandassociatedcostsofholdingcash.Thispointissomewherebelowzero,mayvaryovertimeandisoftencalled‘theeffectivelowerbound’(ELB)forthepolicyrate(seeNessén2016).

Whilenegativenominalpolicyratesarearelativelynewphenomenon,thelowerboundconstraintanditsimplicationshavelongbeenanalysed.TheconceptwasfirstintroducedbyKeynes(1936)whodiscusseditintermsofa‘liquiditytrap’.Inmodernmacroeconomics,aliquiditytraphascometoidentifysituationsinwhichthelowerboundforthepolicyrateisstrictlybinding,inthatitpreventsthecentralbankfromsettingtherealinterestrateatitsdesiredlevel.Therefore,theproblemwithaliquiditytrapisthateventhoughthepolicyrateiszero(orsomewhereslightlybelow),thereal(short-term)interestrateistoohighandeconomicactivityand/orinflationistoolow.Thecentralbankwouldthereforepreferamoreexpansionarymonetary policy in the form of a lower real interest rate, if that were at all possible.

Ithasbeensuggestedbyanumberofresearchers(seeforexampleBordoandLevin2017andGoodfriend2016)thataninterest-ratebearingCBDCcouldrelaxcurrentlowerboundconstraintsonnominalinterestrates.Intheirview,theabilityofpayinginterestonCBDCswouldconstituteaclearadvantagecomparedtophysicalcash.However,asnotedbyCamera(2017),thecurrentcash-relatedlowerboundwillnotdisappearaslongascashisaviablemean of payment.4

2.2Anon-interest-bearinge-kronaraisesthelowerboundWeturnnowtotheeffectsofanon-interest-bearinge-kronaonthelowerboundfortheinterestratesintheeconomy.Theeffectsofsuchane-kronawilldependonhowattractiveanassetitisrelativetootherones.Inordertoanalysethis,wesetupasimplerelationshipthatbuildsonthebasicprinciplethatanassetwillbepreferredifitprovidesnetbenefitsthatareatleastashighasthosethatcanbeobtainedfromanalternativeone.

WecanstartbynotingthattheyieldofanassetAmaybedividedintotwocomponents:theaverageofexpectedshort(riskfree)rates(i)overthematurityoftheasset(n)andapremium(Pt

A,n),

4 Iftheintroductionofane-kronaweretobeaccompaniedbythephasingoutofcash,thiswayofescapingnegativerateswoulddisappear.This,however,hasbeenneitheragoaloftheRiksbank,norapartofthee-kronaproject.ItisalsoworthnotingthatinacountrylikeSweden,wheretheuseofcashisdecliningrapidly,cashmayceaseasamediumofpaymentsanyway.Forexample,Bigonietal.(2018)showthatifmoneyisacceptedinfrequently,itsvaluedecreases,whichinturnhasafeedbackeffectonitsuse.Simplyput,thevalueofcashdeclinesifitisacceptedlessfrequently.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y46

(1) itA,n = 1nΣn

1E[it + i]+ PtA,n.

Thepremiumrepresentsthenetofcompensationforilliquidity,risketc.and‘discounts’forservicesthattheassetmayprovide(forinstanceifitcanbeusedascollateral,forpayments,etc.).

Inspiredby(1)wedefineasimilarexpressionwherethepremiumrepresentsthedifferencebetweentheinterestonane-kronaandthealternativeasset.Letiekr and iA be the nominalinterestrateonane-kronaandanalternativeassetArespectively,overanarbitrarytimehorizon.Anagentjwillbeindifferentbetweenholdingane-kronaandanalternativeasset if

iA = iekr + Pj

where Pj isapremiumoverthesamearbitrarytimehorizon.5 Letϕj

ekr and ϕjArepresentthebenefitsthatane-kronaandanalternativeassetAprovide

respectivelyforagent j.6Moreover,letσjekr and σj

Arepresentthecostofholdingane-kronaand an asset A,respectively,includingthecostoftheperceivedriskforagentj. We can then definethepremiumas

Pj =(ϕjekr −ϕj

A)+(σ jA −σ j

ekr ).

BycombiningtheexpressionforiA and the one for Pj,wederivethefollowingrelationship,whereweabstractfromtheagentsubscriptjsincetheargumentisthesameforallagents:

(2) iekr + ϕ ekr −σ ekr = iA + ϕ A −σ A.

Relationship(2)describesaconditionthathastoholdinorderforanagenttobeindifferentbetweenholdingane-kronaandanalternativeasset.Ifiekr + ϕ ekr −σ ekr > iA + ϕ A −σ A, then theagentwillprefertoholdane-krona,andviceversaifiekr + ϕ ekr −σ ekr < iA + ϕ A −σ A.

Intheremainderofthissection,wewillusevariationsofequation(2)toanalysetheeffectofanon-interest-bearinge-kronaonthelowerboundofreturnsofdifferenttypesofassets.7

Central bank reservesTostudyhowane-kronawillaffecttheeffectivelowerboundforthepolicyratewecanthinkofthealternativeassetinequation(2)ascentralbankreserves.Then,i A denotes the policy rate,whichistheinterestrateonbankreserves.

Givenourassumptions,ane-kronaandbankreservescanbeseenasinvestmentswithveryshortmaturitiesandveryclosesubstitutes.Infact,theyarebothclaimsonthecentralbankandtheriskshouldbethesameforboth.Thus,(σ ekr −σ reserves)=0.Adifferencebetween the two is that an e-krona could be used as a broader means of payment and thus mightprovidesomemoreservicesandismoreliquidthanreserves.Wethereforehavethat(ϕ ekr −ϕ reserves)≥0.Usingthistogetherwithequation(2)weget

iekr +(ϕ ekr −ϕ reserves)=i reserves,

5 Ingeneralequilibrium,the(endogenous)marketrateiAmaychangewiththeintroductionofane-krona.However,fortheargumentinthissectionwecantakethemarketrateiAasgiven.Meaningetal.(2018)provideaframeworkforanalysinghowtheendogenous(market)premiumswilldependontheintroductionofaCBDC.6 Thevalueoftheserviceϕ ekrislikelytodependonhowmuche-kronatheindividualhas.However,evenifthemarginalutilityofholdingane-kronaisdecreasingitdoesnotaffectourresults.7 Asimilarasset-by-assetcomparisonisfoundinMeaningetal.(2018)althoughtherethefocusisnotonthelowerbound.

47S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

thatis,theinterestrateonreservescannotbelowerthantheoneonane-krona.Thus,ifane-kronaisuniversallyaccessiblewithoutlimitations,doesnotcarryinterestandissuppliedaccordingtodemand,thentherateonreservescannotfallbelowzero.Comparedtothesituationtoday,thismeansthattheeffectivelowerboundforthepolicyratewouldrisetozero,orevenslightlyaboveitif(ϕ ekr −ϕ reserves)>0withanon-interestbearinge-krona.

Interbank ratesLetnowthealternativeassetbeinterbankdebt,whichprovidesfewerservicescomparedto an e-krona. For instance, it cannot be used as a broad means of payment. Thus (ϕ ekr −ϕ interbank)>0.Furthermore,lendingtoaprivatebankistypicallymoreriskythantothecentralbank,sothat(σ interbank −σ ekr)≥0.Thus,inthepresenceofanon-interest-bearinge-kronaavailablewithoutlimitations,weget

iekr +(ϕ ekr −ϕ interbank)+(σ interbank −σ ekr)=i interbank,

thatis,interbankratesareunlikelytofallbelowzero(thetwotermsinparenthesesare(weakly)non-negative).LookingatFigure1,whichshowsthepolicyrateandinterbankmarketratesofdifferentmaturitiesfrom2008until2018,wecanseethatthiswouldconstituteachangefromthecurrentsituationinSweden,whereinterbankinterestrateshavebeennegativeforthepastthree years.

-1

0

1

2

3

4

5

6

08 09 10 11 12 13 14 15 16 17 18

Interbank rate, 3 month Interbank rate, 1 monthInterbank rate, T/N Policy rate

Figure 1. Interbank rates and the policy rate in Sweden

Sources: Macrobond and the Riksbank

Commercial bank deposit ratesIncomparingcommercialbankdepositrateswithane-krona,wecanfirstnoticethatdepositscoveredbydepositinsurancecanbeviewedasbeingasriskfreeasane-krona.Depositsthatarenotcoveredbydepositinsurancearemorerisky.Thus,(σ bankdep −σ ekr)≥0.Atthesametime,bank-depositaccountsareoftenbundledtogetherwithotherservices,e.gcreditlines,sothatwemayhave(ϕbankdep −ϕ ekr)>0.Inthatcase,bankdepositratesmaybebelowthereturnonane-krona:

iekr −(ϕbankdep −ϕ ekr)+(σ bankdep −σ ekr)=i bankdep.

Thus, if ϕbankdep issufficientlyhigh,thentheinterestratepayedondepositscouldpossiblybelowerthantheoneonane-krona,thatis,itcouldbenegativeincaseofanon-interestbearinge-krona,atleastforsomecustomers.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y48

Itshouldbenotedherethattheremaybeotherfactorsinfluencinghowcommercialbankssetdepositrates,ineffectpreventingthemfromdroppingbelowzero.ThishasbeenthecaseintherecentperiodwithanegativepolicyrateinSweden,wherebankshavenotpassedthisontohousehold’sdepositaccounts.Indeed,asseeninFigure2,suchrateshaveremainedatzeroduringthelastthreeyears.8

-1

0

1

2

3

4

5

6

08 09 10 11 12 13 14 15 16 17 18

Figure 2. The policy rate and average deposit rate to households

Policy rate Deposit rate

Note. MFIs’ (monetary and financial institutes) average deposit rate is a weighted average of all interest rates on deposits with different maturities.Sources: Statistics Sweden and the Riksbank

Government bondsNext,comparinggovernmentbondswithane-kronaweuse(2)andletgovernmentdebtbethealternativeasset.Wethenget

iekr +(ϕ ekr −ϕgov)+(σ gov −σ ekr)=i gov.

Weseethatgovernmentbondyieldscanbebelowtheinterestrateonane-kronaifgovernmentbondsprovidemoreservices((ϕ ekr −ϕgov)<0)and/orareassociatedwithlessrisk((σ gov −σ ekr)<0).However,ane-kronaisjustanotherformofgovernmentdebtanditscreditriskshouldthereforenotbehigherthanforgovernmentbonds.Furthermore,ane-kronaismoreliquidthanagovernmentbond.Thus,(ϕ ekr −ϕgov)≥0and(σ gov −σ ekr)≥0.Consequently,governmentbondyieldswouldnotfallbelowtheinterestrateonane-krona.LookingatFigure3,whichshowsthepolicyrateandgovernmentbondratesofdifferentmaturities,wecanseethatthiswouldconstituteachangefromthesituationinSweden,wheremediumtermgovernmentbondrateshavebeennegativeforthepastthreeyearsorpartsofthesethreeyears.However,andimportantly,ifvariousformsofregulationweretofavourgovernmentbondsoverane-krona,itispossiblethat(ϕ ekr −ϕgov)<0andgovernmentbondratescouldgobelowzeroeveninthecaseofazero-yieldinge-krona.WereturnbrieflytothisimportantissueinSection2.3.

8 Customerrelationsandcompetitionbetweenbankshavebeenmentionedaspossiblecauses.SeeAlsterlindetal.(2015).

49S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

-2

-1

0

1

2

3

4

5

6

08 09 10 11 12 13 14 15 16 17 18

10 year 5 year 2 year Policy rate

Figure 3. The policy rate and 2, 5 and 10-year government bond yields in Sweden

Note. Implied zero-coupon yields from government bonds.Source: The Riksbank

Risky assetsToconcludeourcomparisonacrossvarioustypesofassets,wenowturntomoreriskyassets.Theseareassetswithmorecreditriskthangovernmentbondssuchase.g.corporatebonds,sothat(σ risky −σ ekr)>0.Furthermore,riskyassetsprovidefewerservicescomparedtoane-krona,sothat(ϕ ekr −ϕ risky)≥0.Thus,itfollowsfrom(2)

iekr +(ϕ ekr −ϕ risky)+(σ risky −σ ekr)=i risky,

thatis,therateofreturnonriskyassetswillbehigherthantheoneonane-krona,andassuchhigherthanzero.

Tosummarize,ane-kronathatisuniversallyaccessiblewithoutlimitations,doesnotcarryinterestandissuppliedaccordingtodemandislikelytoimposeazerolowerboundconstraintonallmarketrates.However,andimportantly,ifvariousformsofregulationfavourgovernmentbonds,returnsonotherassetscouldstillbebelowthereturnonane-krona.Ifthereisazerolowerboundongovernmentbondyields,thismayinturnalsoreducetheeffectivenessofQE.Inthenextsectionweexplainwhy.

2.3Quantitativeeasingwithanon-interestbearinge-kronaAsmentionedabove,QEhasbeenusedasanexpansionarymonetarypolicytoolwherebythecentralbankbuysassets,typicallygovernmentbonds,inthesecondarymarkets.OneoftheaimsofQEistolowerlonger-termmarketrates.9

9 Indeed,thereissubstantialempiricalevidenceshowingthatquantitativeeasingcanalterlong-terminterestrates,asshownforexamplebyKrishnamurthyandVissing-Jorgensen(2011),HamiltonandWu(2012),Gagnonetal.(2010)andWilliams(2014)amongothers.ThisiswhyQEisconsideredashavinghadbeneficialeffectsontheeconomy,inparticularattheELB.Theoretically,Woodford(2012)andBhattaraietal.(2013)havearguedthatQEmayhaverealeffectsbyreinforcingforwardguidance.Byincreasingthesizeofthecentralbankbalancesheetandexposingittocapitallossesifinterestratesrise,thecentralbankcommitstokeepinginterestrateslowerthanisoptimal.AuerbachandObstfeld(2005),instead,showthatopenmarketoperationsattheZLBcanbewelfare-improvingprovidedthatlong-terminterestratesarepositiveandshort-terminterestratesareexpectedtobepositiveatsomepointinthefuture.Williamson(2016)isamodelwhereQEisbeneficialbecausepurchasesoflong-maturitygovernmentdebtbythecentralbankwillalwaysincreasethevalueofthestockofcollateralizablewealth.However,Wallace(1981)showedthatModigliani-Millerappliestoacentralbank’sbalancesheet,andthusQE-typepoliciesshouldbeineffective.EggertssonandWoodford(2003)andCúrdiaandWoodford(2011)showasimilarresultinaNew-KeynesianmodeloncetheZLBisreached.ThereremainsatensioninthetheoreticalliteratureaboutwhetherQEisbeneficialornot.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y50

From(1)wehavethatgovernmentbondyields(i gov)maybedividedintotwocomponents,theaverageofexpectedshort(riskfree)rates(i )overthematurityofthebond(n)andaso-calledtermpremium(TP)

(3) itgov,n = 1nΣ1

nE[it + i]+TPtn.

TherearedifferentaccountsofhowQEaffectsgovernmentbondsyields.Someemphasizetheeffectonexpectedshortrates,whileothersfocusontheeffectsthatQEmayhaveontermpremiums.ApragmaticinterpretationoftheempiricalliteraturewouldsuggestthattheQEprogramsputinplacebyseveralcentralbanksinrecentyearshaveaffectedbothcomponents.

Fromequation(3)weseethattherearetwochannelsthroughwhichtheintroductionofane-kronacoulddampentheefficacyofQE.First,afloorforthepolicyrateaffectsexpectedfutureshortratesastheycannolongerbenegative.Sincethelonger-termmarketrateistheaverageofexpectedfutureshortrates,higher(expected)short-termratesmakethelonger-termrateshigher.Anotherwayofstatingthisisthatthelowerboundtruncatestheyieldcurvesothatyieldsoflongermaturitiesarealsoaffected(seeforinstanceSwansonandWilliams2014andDeRezende2017).

Second,QEisthoughttoworkbyloweringthetermpremium(TPtn).Asmentionedinthe

previoussection,governmentbondsprovidecertain‘services’thatmakethemattractive.Forexample,thereareleverageconstraints,needsforcollateral,andotherfeaturesandfrictionsinfinancialmarketsthatmakesomeinvestorswillingtopaymoreforgovernment-emitteddebtinstrumentsthanothertypesofassets.Aslongasane-kronaisnotconsideredaperfectsubstituteinthisregard,QEcouldstillworkbyloweringtermpremiums.However,ifane-kronacametobeperceivedasprovidingthesameservicesasgovernmentbondsandtherewerenoregulationsthatmadeinvestorsprefergovernmentbondsoverane-krona,theefficacyofQEcouldbediminished.

2.4 ImplicationsfortheconductofmonetarypolicyInthedecadesincetheonsetoftheGreatFinancialCrisis,severaladvanced-economycentralbankshaveengagedinvariousformsofunconventionalmonetarypolicy.Specifically,somecentralbankshaveconductedlargescaleassetpurchases(orQE),othershaveloweredpolicyratesbelowzero,andsomeothershaveemployedforwardguidance.Afewcentralbankshaveimplementedalloftheabove.

Beginningin2015theRiksbankloweredthepolicyrateinstepsintonegativeterritory.AtthesametimetheRiksbankbeganpurchasinggovernmentbonds,andcurrentholdingsamounttoabout40percentoftheoutstandingstockofgovernmentdebt.Asbrieflymentionedabove(andshowninFigure3),throughthesevariousmeasurestheRiksbankhasbeenabletolowergovernmentbondratesdownbelowzero,attimesevenbeenbelowthepolicyrate.Eventhoughdepositratesandmanyotherrateshavestayedabovezero,changesinthepolicyrateintonegativeterritoryhaveledtoreductionsinother(positive)rates.Forinstance,andascanbeseeninFigure4,lendingratestohouseholdshavedecreasedafternegativepolicyrateswereimplemented,althoughbylessthanthedecreaseinthepolicyrate.Itisalsoworthnotingthatcorporatelendingrateshavedecreasedbyatleastasmuchasthepolicyrate.Thesefiguresandmoreformalanalysesbye.g.DeRezendeandRistiniemi(2018)andLaséenandDeRezende(2018)indicatethattheunconventionalpoliciespursuedbytheRiksbankinrecentyearshaveindeedledtomoreexpansionaryfinancialconditions.

51S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

-0.9

-0.8

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0.0

0.1

Jan-15 July-15 Jan-16 July-16 Jan-17 July-17 Jan-18 July-18

Figure 4. Change of repo rate and lending rates to households and companies since 2015 Percentage points

Households mortgage loans Households and corporatesRepo rate Corporate lending rate

Note. The cumulative changes in each rate since the start of January 2015. Outcomes are monthly data and lending rates are value-weighted averages of the actual lending rates reported by the banks. Sources: Statistics Sweden and the Riksbank

Thediscussioninsections2.3and2.4howeversuggeststhattheimpactofthesetypesofunconventionalpoliciesonfinancialconditionswouldbesmallerifanon-interestbearinge-krona were to be introduced.

Arelevantquestiontheniswhetheritislikelythatsuchunconventionalpolicieswillbeneededinthefuture.Therootcauseofthelowlevelsofnominalinterestratesandthefactthatthelowerboundhasbecomeaconstraintontraditionalinterestratepolicyistheseculardeclineinglobalinterestratesinthepastdecades.Indeed,therearemanystudiesdocumentinghowglobalrealrateshavefalleninthepastdecades,andalsoindicatingthatrealrateswillremainlowinthefuture(seee.g.Armeliusetal.2014,andHolstonetal.2016).Thisdevelopment,togetherwithlowinflationrates,meansthatnominalinterestrateswillmostlikelyremainlowintheforeseeablefuture,thusimplyingthatcentralbankscouldinthefutureagainhitthezerolowerbound.Thisimpliesthatwithane-kronathatisuniversallyaccessiblewithoutlimitations,doesnotcarryinterestandissuppliedinunlimitedquantities,theroomformanoeuvreformonetarypolicybymeansofthekeypolicyrateandQEcouldbecurtailedinthefuture.

Other options QEandnegativeinterestratesarenottheonlytoolsavailabletoacentralbankifthereisaneedformoremonetarystimulus.Therearefurthermeasuresthatworkthroughotherchannels,suchasforinstanceproceduresthatimprovethetransmissionmechanismorthatworkthroughtheexchangeratechannel.Thesemeasureswillgenerallynotbeaffectedbyane-krona.Itisalsoworthmentioningthatlower-boundconstraintsforthepolicyratecanbealleviatediftheinflationtargetisraised.Thatisbecauseifinflationishigheronaverage,thenominalinterestratewillalsobehigheronaverage,thusreducingtheriskofthepolicyratebecomingtoolowandhittingthelowerbound.SeeApeletal.(2017)foradiscussion.

Furthermore,somearguethataCBDCopensupthepossibilityofanewformofunconventionalmonetarypolicy,asmoneytransferstohouseholdswouldbeeasiertoimplement,muchlikeadigitalhelicopterdrop.TheideabehindsuchmeasuresisnotnewanddatesbacktoFriedman(1969).Itinvolvesthecentralbanksupplyinglargeamountsofmoneytothepublic,asifthemoneywasbeingdistributedorscatteredfromahelicopter.Colourfulimagesaside,helicoptermoneyismeanttobemadedirectlyavailabletoconsumerstoincreasespendingintimesofweakdemand.FormerFederalReserveChairmanBenBernankepopularizedthisideain2002asamoney-financed(asopposedtodebt-

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y52

financed)taxcutpolicythattheoreticallygeneratesdemandandshouldthereforeideallybeusedinalow-interest-rateenvironmentwhenaneconomy’sgrowthremainsweak.10 However,inSwedenitisnotobviousthathelicopterdropswouldbeeasiertoimplementwithane-kronasincealmostalladultSwedesalreadyhaveaccountsatcommercialbanks(seeSverigesRiksbank2017).

Insum,wecanconcludethatraisingtheeffectivelowerboundforthepolicyratemeansthatthereisariskthattheprimarytoolformonetarypolicycannotbeusedoptimally.Intheabsenceofotherpolicies,thiscouldimpactnegativelyoneconomicactivity.Wediscusslongruneffectsofane-kronainSection4.

3 Effectsonthemonetarytransmissionmechanism

Wehaveshownthatanon-interestbearinge-kronacouldreducetheeffectivenessofmonetarypolicyifitraisestheeffectivelowerbound.BIS(2018)andMeaningetal.(2018)amongstothershavesuggestedthataninterest-bearingCBDCmaymakemonetarypolicymoreeffectivethroughimprovedpass-throughofpolicyratechanges.Inthissectionweanalyse if this is the case for an e-krona.

The monetary policy transmission mechanism normally describes the process by which changesinthepolicyrateinfluencetherealeconomyandinflation.Themechanismcanbedividedintotwoparts.Thefirstdescribeshowchangesinthepolicyratepassthroughtochangesindepositrates,lendingratesandothermarketinterestratesthatmatterforeconomicdecisions.Thesecondpartdescribeshowchangesintheseinterestratesinfluencetherealeconomyandinflation.Asexplainedabove,thepass-throughmaybehamperedwhentheeffectivelowerboundisincreasedtozero.Inouranalysisbelowwefocusonscenarioswithaninterest-bearinge-kronaandthusnobindingeffectivelowerboundinducedbyit.

3.1 Transmissionfromthepolicy-ratetomarketratesInordertokeeptheanalysisinthissubsectiontractable,weaddafewassumptions.First,wefocussolelyonthepass-throughtobanks’depositandlendingrates,whichareconsideredkeyinthetransmissionmechanism.Second,weonlyconsideranattractivee-krona,i.e.ane-kronathatpaysahighenoughinterestratetocreatecompetitionwithbankdeposits,sinceanunattractiveonewouldnotinfluencethebanks’behaviour.Third,weassumeafixedspread(whichcouldbezero)betweenane-kronaandthepolicyrate.Ifthespreadcouldvaryitwouldnotmakesensetotalkaboutthepass-throughfromthepolicyratetomarketrates.Furthermore,ifthespreadwereallowedtovary,thespreaditselfwouldbeaseparatepolicyinstrument.

Theleft-handpanelofFigure5containsascatterplotofaveragebankdepositratesandthepolicyrateinSwedenoverthepast25years.Itillustratesthatthepass-throughfromthepolicyratetodepositrateshasbeenlessthanonetooneinthisperiod.Specifically,aregressionbasedonthedatainthefiguresuggeststhatanincreaseinthepolicyratebyonepercentagepointleadstoanincreaseindepositratesbyonaverage0.6percentagepointsduringthesamequarter.11 Thus, historically when the policy rate has increased in Sweden, depositrateshavealsoincreasedbutbyasmalleramount.

10 SeeBernanke(2002).Helicoptermoneyisenjoyinganewrevivalasalast-resourceoptionwithinfluentialadvocatesincludingCaballero(2010)andGalí(2014)amongothers.Suchanunconventionalideahasitscriticstoo.Forexample,sincecentralbankspayinterestonreserves,Kocherlakota(2016)observesthatnewmoneycreatedwouldeventuallyhavethesamecostasifthefiscalauthorityborrowedit.Alongthosesamelines,Borioetal.(2016)findthathelicoptermoneybecomesmoreexpansionarythanadebt-financedprogrammeonlyifthecentralbankscrediblycommitstosettingpolicyatzeroonceandforall,thusimplyinggivingupmonetarypolicyforever.11 AnOLSestimationofthefollowingequationΔit

D = βΔitR + ϵ, where it

D and itRdenotethedepositandthepolicyraterespectively,

gives(p-valuesinbrackets)β=0.64(0.00),R2adj=0.86.Weexcludethemostrecentyearswhenthepolicyratehasbeennegative.

53S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

-2 0 2 4 6 8 10 12 14-2

0

2

4

6

8

10

12

14

-2 0 2 4 6 8 10 12 14-2

0

2

4

6

8

10

12

14

Figure 5. Pass-through from the policy-rate to banks deposit rates and lending rates for householdsa. Deposit rates b. Lending rates

Note. Quarterly data for the period 1993:Q1 to 2018:Q2. Deposit rates for new agreements. Lending rates for floating rate contracts.Sources: The Riksbank and Statistics Sweden

We assess that with an e-krona the pass-through to banks’ deposit rates is likely to increase andbecomeclosetoonetoone.Toseewhy,considerasanexamplearepresentativebank.Ifsuchabankwantstoretaindeposits,ithastomakethematleastasattractiveasane-krona.Informalterms,thisimplies:

(4) iekr + ϕ ekr −σ ekr ≤i D + ϕD −σ D,

where‘D’refersto‘bankdeposits’.Itfollowsdirectlythatforanygivenϕ ekr, σ ekr, ϕD and σ D,anincreaseinthee-kronarate(iekr)willhavetobefollowedbyasimilarincreaseintheinterestrateondepositaccounts(i D).Similarly,thebankcanfollowareductioninthee-kronaratewithacorrespondingreductionintheinterestrateondepositaccountswithoutfearoflosingdeposits.Thus,unlessthebankcompensatese-kronaratechangesbyalteringϕD and σ D,thepass-throughfromthepolicyratechangestothebank’sdepositrateswillbecomeonetoonewithane-kronaundertheassumptionsmadeatthebeginningofthecurrentsection.12

Thebank,however,mightnotfinditprofitabletocompetewithane-krona.Inthatcase,thatbank’sdepositswillflowintodepositaccountsatotherbanksthatcompetewithane-kronainsteadandwherethepass-throughisagainclosetoonetoone.Alternatively,depositsmightflowintoe-kronaaccounts.Wecanthereforeconjecturethatwithanattractivee-krona,pass-throughtodepositrateswillbeclosetoonetoone.Indeed,inAppendixAweprovethatthisconjectureholdsinaformalbankingmodel.

Onemightalsoarguethatane-kronawillspeedupthepass-throughasitwillbeaveryexplicitcompetitivealternativetobankdeposits.Meaningetal.(2018),however,suggestthatapotentialoffsettingeffectcouldbeforbankstorespondtotheincreasedcompetitionfromaCBDCbymakingitmorecostlytomovefundsoutofthebank.Sucheffectsmayalsoslowdownthepass-throughandwecannotexcludeapriorithatthiswouldhappeninSweden.

Insum,ouranalysissuggeststhataninterestbearinge-kronawithafixedspreadvis-à-visthepolicyratemayimprovepass-throughfromthepolicyratetodepositratesinSweden.13

Results for the pass-through to banks’ lending rates are less clear cut. There are two reasons for why this is the case.

12 Animportantcaveatisthatthesemechanismsmaynotcomeintoplaywhentheinterestrateonthee-kronaisclosetozero,see Appendix A. 13 However,if–differentlyfromtheassumptionsofthissubsection–thespreadbetweenthepolicyrateandane-kronawereallowedtovarythisconclusionwouldnotnecessarilyhold.Forexample,ifthee-kronaratewerekeptconstantwhilethepolicyrateincreased,themechanismsdescribedabovewouldnotcomeintoplay.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y54

First,thepass-throughtobanks’lendingratesisalreadyhigh,closetoonetoone,withoutane-krona.ThiscanbeseeninFigure5,Panelb.,whichcontainsascatterplotoftheaverageofSwedishbanks’lendingratesandthepolicyrate.Ascanbeseeninthefigure,thedotslieona45degree-line.Furthermore,aregressionbasedonthedatainFigure5suggeststhatanincreaseinthepolicyratebyonepercentagepointleadstoanincreaseinthelendingrateofonepercentagepoint.14

Second,theoreticallyitisnotobviousthatane-kronawouldinfluencethepass-throughfromthepolicyratetolendingrates.Thinkforexampleofbanksaspursuingbusinessintwoseparatemarkets:adepositmarketandalendingmarket(seeAppendixAforaformalmodel).15Underthisscenario,banksinthedepositmarketborrowfromdepositorsandinvestinthemoneymarket.Theprofitfromthisactivityarisesfromthedeposit intermediation margin, i.e. the spread between the money market rate and the deposit rate.Inthelendingmarket,banksborrowinthemoneymarkettoinvestinloans.Theprofitfromthisactivityarisesfromthelending intermediation margin,i.e.thedifferencebetweenthelendingrateandthemoneymarketrate.16Ane-kronawouldhavenodirecteffectonthelendingmarketinthisenvironment.Ifithadany,sucheffectswouldhavetocomefromchangesinthewaythepolicyrateaffectsmoneymarketrates,changesinloandemandrelations,alteredcompetitioninthelendingmarket,orchangesinbanks’costsforprovidingloans.Itisnotobviousthatanyofthesewouldbeaffectedbyane-krona.AformalandmorethoroughdiscussionofthesetheoreticalargumentscanbefoundinAppendixA.Noticethatsuchaconclusionmightdifferdependingontheinterconnectednessofthedepositandlendingmarkets.However,theassumptionofseparatedepositandlendingmarketsmakessenseinSwedenwherethebanksrelyheavilyonmarketfunding.

Insum,ouranalysissuggeststhatthepass-throughfromthepolicyratetobankinterestratesisalreadyhighinSwedenandanymarginalimprovementwouldmostlikelyoccuronthedepositside.Twothingsareimportanttonoticeinconnectionwiththis.First,theimprovedpass-throughmightnotbeofmuchhelpastheimprovementinthepass-throughmightonlytakeplaceforhighlevelsofthepolicyrate(seeAppendixA).However,itisprimarilywhenthepolicyrateislowandclosetothelowerboundthatastrongerpass-throughisuseful.Athigherlevels,weakpass-throughcanbefullycompensatedforbylargerchangesinthepolicyrate.Second,animprovedpass-throughtodepositratescoupledwithanunchangedpass-throughtolendingratesmightbeproblematic,sincetheaggregatedemandeffectsofachangeinthedepositrateareambiguous.17

3.2 Transmissionfrommarketratestotherealeconomyandinflation

We now turn to how an e-krona may impact the second part of the transmission mechanism. Thatis,weanalysewhetherane-kronawouldchangethetransmissionfromdeposit-,lending-andothermarketratestothewidereconomy.Wefinditusefultoformulatethediscussionalongthefollowingchannelsofthetransmissionmechanism:theinterestratechannel,theexchangeratechannel,thecreditchannelandtherisk-takingchannel.

The interest rate channelreferstotheeffectofinterestratechangesonhouseholds’savingsandconsumption,aswellasfirms’investment.Ifpricesandinflationexpectationsaresticky,areductioninnominalmarketrateswillalsoreducetherealinterestrateinthe

14 OLSestimationofΔitL = βΔit

R + ϵgives(p-valuesinbrackets)β =1.00(0.00),R2adj =0.795.Weshortenthesampletoexcludetherecentyearswithapolicyratebelowzero.15 Result1inAppendixAshowsthatthisseparationderivesfromdisjointvariablecostsofmanagingloansanddeposits.16 This(theoretical)separationdoesnotmeanthatallbankdepositsareliteraryinvestedinthemoneymarketandalllendingisliteraryfundedbythemoneymarket.Thebanksusedepositstofinancelending.Onlythegapbetweendepositsandlendingisactuallyfinancedorinvestedinthemoneymarket.17 Theincomeeffectofalowerdepositratereducesthe‘income’fromdepositsandleadstoareductionindemand.Thepriceeffect(substitutioneffect)ofalowerdepositratereducesthepriceofconsumptiontodayrelativetotomorrowandleadstoanincrease in demand today.

55S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

economy.Lowerrealinterestratesmakeitmorebeneficialforhouseholdstoconsumeandborrowandlessbeneficialtosave.Similarly,firmswillprefertoborrowandinvest.Theincreaseddemandintheeconomygraduallyresultsinpricesandwagesstartingtoincreasemorequickly.Theeffectswillbethesamebutofoppositesignwhentheinterestrateincreases.

We,asotherauthors,assessthatane-kronaisunlikelytoaffecthowchangesinrealmarketratesaffectagents’consumption,savingsandinvestmentsdecisions.Theserelationsaredeterminedbyunderlyingpreferenceswhicharenotexpectedtobeinfluencedbytheintroductionofane-krona.

The exchange rate channelreferstothemechanismthroughwhichmonetarypolicyinfluencesinflationandtherealeconomybyaffectingtheexchangerate.Areductioninthepolicyratenormallyleadstoanexchangeratedepreciation.Ifpricesaresticky,theexchangeratealsoweakensinrealterms,whichinturnmakesdomestically-producedgoodscheapercomparedtoforeignones.Thisleadstoanincreaseinthedemandforexportsandforproductsthatcompetewithimportedgoods,whichgraduallyresultininflationrisingaswell.Theexchangeratechannelalsohasamoredirecteffectoninflation.Thatisbecausethedomesticpriceofimportedgoods,whichareincludedintheconsumerpriceindex,riseswhentheexchangerateweakens.

Theparityconditionsdeterminingtheexchangerateareunchangedbytheintroductionofane-krona.However,auniversallyavailablee-kronawouldconstituteanew,liquidandsafe deposit where to hold money balances in Swedish Krona. To the extent that this leads to moreactivecurrencymanagementbydifferentactors,ane-kronamightinducetheexchangeratetobecomemoresensitivetochangesinmarketrates.This,inturn,wouldimplystrongerand/orfasterexchangeratemovementsforagivenchangeinthemarketratesinSwedenand abroad.18However,wearenotawareofanyformaltheoryofthiseffect.

The credit channel referstothemechanismthroughwhichinterestratechangesaffectthecreditmarketandtherebythemacroeconomy.Alowerinterestrategenerallyleadstoanincreaseinthepriceofvariouskindsofassets.Forexample,itleadstoanincreaseinthenetpresentvalueofthefuturecashflowsthatafinancialassetcanbeexpectedtogenerate.Thismeansthatthepriceofthefinancialassetincreases.Whentheinterestrateislow,thedemand for and prices of real assets such as houses also increase. As these assets are used ascollateralforloansandthecollateralincreasesinvalue,banksbecomemorewillingtolendmoney.Inaddition,futurewagesofhouseholdsandfutureprofitsofcompaniestendtorisewhendemandincreasesasaresultofthelowerinterestratelevels.Onthewhole,thecreditchannelisamechanismbywhichtheeffectofchangestothepolicyrateisenhancedthroughlendingfromthebanks.

Themainreasonforwhytheintroductionofane-kronawouldmatterforthecreditchannelisthereducedsupplyofcreditifbanksweretocutdownontheirlendingduetolowerrevenuesonthedepositside.Inthiscase,thecreditchannelcouldbecomeweaker.Theoretically,whetherthiswilloccurdepends–amongotherthings–ontheinterconnectednessofthelendinganddepositmarkets.Ifthetwoareindependentofeachother,thenitmaybelesslikelythatbankswilldecreaselendingasaresponsetolowerprofitsfromthedepositmarket(seeAppendixA).ItisalsoworthpointingoutthataCBDCmayenablegreatercompetitionintheprovisionofcreditforinstancethroughimprovedpossibilitiesforpeer-to-peerlending(Meaningetal.2018).19

Another, and much discussed, channel in the transmission mechanism is the so-called risk-taking channel.Itsuggeststhatlowpolicyratesleadbanksandotherfinancial

18 ThischangeisdifferentfromtheonedescribedinMeaningetal.(2018).Theysuggestthattheexchange-ratechannelmightbecomestrongerbecausemarketratesbecomemoresensitivetochangesinthepolicyrates.19 ‘Forinstance,peer-to-peerlenderswouldnolongerhavetoclearsettlementsthroughtheircompetitorsinthebankingsector,asiscurrentlynecessaryintheexistingsystemoftieredaccesstocentralbankmoney.ThisprocessincursacostwhichCBDCcouldpotentiallyeliminate,puttingnon-bankcreditprovidersonamoreequalfootingwiththeirbankingsectorcounterpartsandwouldlimittheextenttowhichbankscouldvarymarginsinlightofchangesinfundingcosts.’Meaningetal.(2018),p.21.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y56

institutionstotakegreaterrisks.Thisisnotaspecific,well-definedmonetarypolicychannel,butacollectivetermusedtodenotedifferentkindsofmechanisms,wherebymonetarypolicycanaffecttherisk-takingofbanks,financialinstitutionsandtheeconomyasawhole.Onemechanismisduetolowinterestratesresultinginasocalledsearchforyield,wherebybanksstarttosearchforriskierinvestmentswithahigherexpectedreturn(Rajan,2005).Onereasonfordoingthiscouldbethatbankshaveaspecificnominalrateofreturnthattheyhavetoachieve.Anothermechanismmightbeduetotheeconomyexperiencinglowriskandlowinterestratesoveralongperiodoftime,thusleadingeconomicactorstobecometoocomplacentandplacingadisproportionallylowweightonriskfactors.20Again,weconsiderittobeunlikelythattherelationshipbetweenmarketinterestratesandrisk-takingintheeconomywouldchangewiththeintroductionofane-krona.

Insum,weassessthattheexchange-ratechannelandpossiblyalsothecreditchannelaretheonlychannelsthatmaybealteredinasignificantwaybytheintroductionofane-krona.

4 Othereffectsontheeconomy

4.1 Smallopeneconomyaspectsofane-kronaAsdiscussedintheintroduction,thee-kronawestudyisuniversallyavailableandsuppliedaccordingtodemand.Thisopensupnewquestions,sinceinvestmentinane-kronabyinternationalinvestorscouldgiverisetolargecapitalflows,thusamplifyingthepotentialvolatilityofthebalancesheetofthecentralbankandpossiblycreatinggreaterexchangeratevolatility.21

Butitisveryhardtoanticipatemorepreciselywhateffectsane-kronamighthaveontheexchangerate.Aslongasane-kronaisprimarilyusedfordomesticpaymentsitwillmostlikelynotinfluencetheexchangerateatall.However,thereisanimportantdifferencebetweenane-kronaandcash,andthatisthatane-kronacanbeagoodsubstitutetootherformsofsavingvehiclessuchasgovernmentbondsorsavingsaccounts.Thereisalso the added factor that an e-krona can be purchased and sold much faster than cash, thusincreasingtheriskofvolatility.Ifane-kronabecameanattractiveassetamongforeigninstitutionalinvestorsthenitcouldinfluencetheexchangerate,bothitslevelanditsvolatility.

HerewecanreturntothesimpleframeworkintroducedinSection2,expressingitintermsofforeigncurrency:

(5) iekr* + ϕ ekr* −σ ekr* = i A + ϕA −σ A,

wherealltermsnowaredenominatedinforeigncurrency,e.g.iekr* is the return on an e-kronainforeigncurrency.Thetermσ ekr*includesexchangerateriskfromthepointofviewoftheinternationalinvestor.Theinterpretationofequation(5)isthattherewillbeinflowstothedomesticeconomyiftheleft-handsideexceedstheright-handside,e.g.iftheinterestrateonane-kronaishigh,ifitprovidesusefulservices,etc.Itispossiblethatfinancialstressabroad(hererepresentedbyanincreaseinσ A)couldtriggerlargeinflowstoane-krona,forinstance.Conversely,therecouldbelargeflowsoutofe-kronaholdingsiffinancialconditionschange.

Insum,forasmallopeneconomy,aCBDCthatisuniversallyaccessiblewithoutrestrictionsandlimitationscouldfacilitatelargecapitalflowsthatmightinturnleadtovolatilityintheexchangerateandinthesizeofthecentralbank’sbalancesheet.22

20 SeeApelandClaussen(2012)foradetaileddiscussionoftherisk-takingchannel.21 SeealsoNessénetal.(2018),DanmarksNationalbank(2017)andBIS(2018).22 SeetheappendixofNessénetal.(2018)foraverysimpleillustrationusinghighlysimplifiedbalancesheets.

57S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

4.2 FinancialstabilityJuks(2018)analyzestheeffectsane-kronamighthaveonSwedishbanks.Inwhatfollows,instead,wesummarizethecurrentliteratureontheconsequencesCBDCsmighthaveonfinancialstability.EngertandFung(2017),forexample,suggestthatifaCBDCisnon-interestbearing,thenitisunlikelythatitwouldleadtoasignificantshiftawayfromtraditionalinstrumentssuchasdepositaccounts.ThatisbecausefinancialinstitutionscaneffectivelycompetewithCBDCasastoreofvaluesincetheycanofferenhancedfinancialservicessuchaswealthmanagementorengageincost-cuttingmeasures.Nonetheless,intimesofeconomicstress,theremaybeanincreaseindemandforCBDC,whichwouldbeviewedasriskfree.Theshiftawayfromtraditionaldepositswouldbelikelytodisruptthefinancialsystemandincreasevolatility,asdiscussedbyCamera(2017).

Inthisregard,theanalysisinKumhofandNoone(2018)distinguishesbetweenrunsonindividualbanksandsystemicruns.Inthefirstcase,theyclaimthatthepresenceofCBDCscouldpotentiallymakeiteasierandfastertoresolveanindividualtroubledinstitution,bygivingtheauthoritiestheoptionofrepayingitsdepositorsinsafeCBDCatanearlystageandthusreducingthepotentialforcontagion.Sincebankdepositorswouldknowthisex-ante,this may in fact reduce the probability of a bank run compared to a world without CBDCs. Theydofindthatsystemicbankrunswouldbemoredifficulttosolveinstead,eveninaworldwithCBDCs.Indeed,insuchacasetheruntoCBDCscouldpotentiallybesolargeatthecurrentCBDCinterestrate,thatCDBCholderswouldnotbewillingtosellsufficientquantitiesofCBDCtosatisfythedemandforit.ThehighdemandcouldbeaddressedbythecentralbankwithadecreaseintheinterestrateonCBDC,ifanywerepaid.However,therewouldbepotentiallimitstosuchapolicyifitrequiredahighlynegativeinterestrate,whichcouldbecomepoliticallyuntenable.

4.3 EconomicactivityInthestandardmodelsusedinpolicyanalysis,monetarypolicyeffectsontherealeconomyareusuallyduetonominalfrictionsthatlimitthespeedoftheadjustmentofthegenerallevelofprices.Suchfrictionsareshort-termphenomenaandtheirempiricalsignificanceisamatterofongoingresearch.Thereisageneralconsensusamongeconomiststhatlong-termeconomicgrowth,instead,isdrivenbyfactorssuchastechnologicalchange,populationgrowth,andhumancapitalaccumulation,thusimplyingmonetarypolicy’seffectsonrealeconomicactivityaresmallinthelongterm.Weshouldthusexpectane-kronatohavenosignificanteffectonlong-termgrowthviamonetarypolicy.

However,ane-kronacouldpotentiallyleadtosignificantleveleffectsoneconomicactivitybecauseofitsinteractionwiththepaymentsystemandthebankingsector.Indeed,ithasbeenshownthatawell-functioningpaymentinfrastructureenhancestheefficiencyoffinancialmarketsandthefinancialsystemasawhole,boostsconsumerconfidenceandfacilitateseconomicinteractionandtradebothingoodsandservices(seeECB2010).Atthesametime,unsafeandinefficientpaymentsystemsmayhampertheefficienttransferoffundsamongindividualsandeconomicactors(Humphreyetal.2006).Hasanetal.(2013)evenconfirmthatmoreefficientelectronicretailpaymentsstimulatetheoveralleconomy,consumptionandtrade.Indeed,theyfindthatdevelopmentsintheuseofelectronicpaymentsystemsarerelatedtonotableimprovementsinbankingperformance,duetobothadecreaseincostsandanincreaseinrevenues.Moreover,asshownbyBerger(2003),switchingtoelectronicpaymentinstrumentshassignificanteffectsintermsofbanks’gainsinproductivityandeconomiesofscale.So,totheextentthatane-kronawouldenhancetheresilienceandtheefficiencyoftheSwedishpaymentsystem,wecouldexpectittohavemeaningfulpositiveeffectsontherealeconomy(seeSverigesRiksbank2017and2018).

Moreover,ane-kronamayraisetheseignioragerevenueofcentralbanks(seee.g.BIS,2018).Ifsuchincreaseswerelargeandtransferredtothegovernment,theywouldallowfor

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y58

lessdistortionarytaxationandmightthereforeevenhaveGDPeffects.BarrdearandKumhof(2016)arguethattherecouldbesuchpositiveconsequencesforthelevelofGDP.

However,aswealreadydiscussed,ane-kronacouldalsohavenegativeimplicationsforfinancialstability.Thiscouldinturnhavedetrimentaleffectsoneconomicactivityeveninthelongrun.Forexample,EnnisandKeister(2003)useanendogenousgrowthmodeltoshowthatbankrunscanhavepermanenteffectsonthelevelsofthecapitalstockandofoutput.That is because as the probability of a run increases, it becomes more likely that a bank willhavetoliquidateinvestmentsearly.Sincetheliquidationvalueofilliquidinvestmentsisrelativelylow,thebankpreferstoholdmoreliquidassetstodealwitharunifitoccurs,thusleadingtosubstantiallylessinvestmentinnewcapital.Moreover,ifbanks’fundingcostsweretoincreaseinameaningfulwayandifsuchcostswerepassedontoconsumers,wewouldexpecttherealeconomytobenegativelyaffected.

5 ConcludingremarksWehaveanalysedpossibleimplicationsofintroducingane-kronaformonetarypolicyandoverallmacroeconomicactivity.Sinceane-kronathatisuniversallyaccessibleandsuppliedaccordingtodemandwouldbeaperfectsubstituteforbankreserves,anon-interestbearinge-kronawouldintroduceazerointerestratefloorforthepolicyrateandplausiblyallotherinterestratesintheeconomy.Thisresultarisesasane-kronaislessriskyandoffersalevelofotherbenefitsorpaymentservicesthatareofequalmagnitude(orhigher)thanotherassets.Theinabilitytoimplementnegativeinterestratesineconomicdownturnscouldpossiblybecompensatedforbytheuseofothermonetarypolicytools.However,thezerointerestratefloorwouldalsomostlikelyapplytogovernmentbonds,whichwouldreducetheeffectivenessofQEduringtimesofabindinglowerboundconstraint.Wealsoarguethattheeffectsonthetransmissionmechanismarelikelytobesmallinnormaltimes.

Itispossiblethatane-kronacouldhaveconsequencesforboththelevelandthevolatilityoftheexchangerateoftheSwedishkronaandthebalancesheetoftheRiksbankifitweretobecomeattractiveforforeigninvestors.Itisalsoplausiblethatane-kronacouldaffectthefinancialsystemandincreaseitsvolatilityintimesofeconomicstress.Moreover,whileane-kronacouldbehelpfulindealingwithrunsonindividualinstitutions,systemicrunswouldbemoredifficulttosolveasthatmightrequirehighlynegativeinterestrates.

Finally,wearguedthatane-kronacouldpotentiallyhavelong-runleveleffectsoneconomicactivitybecauseofitsinteractionwiththepaymentsystemandthebankingsector.Ontheonehand,itcouldimprovetheefficiencyandresilienceofthepaymentsystemthusstimulatingeconomicactivity.Ontheotherhand,wewouldexpectdetrimentallong-runeffectsifane-kronaimpingesonfinancialstability.

Insum,thereseemstobean’impossiblequaternity’or‘quadrilemma’forthetypeofCBDCenvisionedintheRiksbank’sfirste-kronareport.23Ifane-kronaisdesignedwithsimilarcharacteristicstocash–i.e.non-interest bearing, in perfectly elastic supply and attractive touse–thenitwillmostlikelynotbecompatiblewithunchanged macroeconomic risks. Consequently,aCBDCcannothavethesefourfeaturesatthesametime.

ItisworthnotingthatthenegativeeffectswehaveidentifiedcouldbemitigatedbygivinguponeormoreofthefourfeaturesinthequaternitywhichwouldgivetheRiksbankamechanismtoinfluencethedemandforane-krona.Oneobviousexampleistoletthee-kronabeinterestbearing,butthereareotheralternativessuchasfeesorotherfrictionsthatwouldlimittheattractivenessofane-kronainrelationtootherassets.However,addinglimitstotheamountofe-kronathatcanbeheldrisksbreakingtheparityagainstotherforms

23 Bjerg(2017)discussesaCBDC‘trilemma’.HearguesthatinthepresenceofaCBDCacentralbankthattriestoupholdfreeconvertibilitybetweenprivatemoneyandCBDC,andparitybetweenallformsofmoney,wouldhavetogiveupmonetarysovereignty.

59S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

ofkrona,suchasmoneyheldinprivatebankaccountsorbankreservesattheRiksbank.Othertypesoffrictions,suchasfeesonwithdrawalsmightthereforebepreferable,butwouldhavetobecarefullycalibratedsothatane-kronawouldstillfunctionasaviablepayment instrument.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y60

ReferencesAlsterlind,Jan,HannaArmelius,DavidForsman,BjörnJönssonandAnna-LenaWretman(2015),‘Howfarcanthereporatebecut?’,Economic Commentariesno.11,SverigesRiksbank.

Apel,MikaelandCarlAndreasClaussen(2012),‘Monetarypolicy,interestratesandrisk-taking’,Sveriges Riksbank Economic Review,no.2,pp.68–83.

Apel,Mikael,HannaArmeliusandCarlAndreasClaussen(2017),‘Theleveloftheinflationtarget–areviewoftheissues’,Sveriges Riksbank Economic Review,no.2,pp.36–56.

Armelius,Hanna,PaoloBonomolo,MagnusLindskog,JuliaRådahl,IngvarStridhandKarlWalentin(2014),‘LowerneutralinterestrateinSweden?’,Economic Commentariesno.8,SverigesRiksbank.

Auerbach,AlanJ.andMauriceObstfeld(2005),‘Thecaseforopen-marketpurchasesinaliquiditytrap’,American Economic Review,vol.95,no.1,pp.110–137.

Barrdear,JohnandMichaelKumhof(2016),‘Themacroeconomicsofcentralbankissueddigitalcurrencies’,StaffWorkingPaperno.605,BankofEngland.

Berger,AllenN.(2003),‘Theeconomiceffectsoftechnologicalprogress:evidencefromthebankingindustry’,Journal of Money Credit and Banking,vol.35,pp.141–176.

Bernanke,BenS.(2002),‘Deflation:makingsure“it”doesn’thappenhere’,RemarksbyGovernor BenS.BernankebeforetheNationalEconomistsClub,Washington(November21).

Bhattarai,Saroj,GautiB.EggertssonandBulatGafarov(2015),‘Timeconsistencyandthedurationofgovernmentdebt:asignallingtheoryofquantitativeeasing’,WorkingPaperno.21336,NationalBureauofEconomicResearch.

Bigoni,Maria,GabrieleCameraandMarcoCasari(2018),‘Partnersorstrangers?Cooperation,monetarytrade,andthechoiceofscaleofinteraction’,American Economic Journal: Micro (forthcoming).

BIS(2018),‘Centralbankdigitalcurrencies’,CommitteeonPaymentsandMarketInfrastructure,BankforInternationalSettlements.

Bjerg,Ole(2017),‘Designingnewmoney–thepolicytrilemmaofcentralbankdigitalcurrency’,WorkingPaper,June,CopenhagenBusinessSchool.

Bordo,MichaelD.andAndrewT.Levin(2017),‘Centralbankdigitalcurrencyandthefutureofmonetarypolicy’,WorkingPaperno.23711,NationalBureauofEconomicResearch.

Borio,Claudio,PitiDisyatatandAnnaZabai(2016),‘Helicoptermoney:theillusionofafreelunch’,Vox,CEPR’sPolicyPortal(May24).

Caballero,Ricardo(2010),‘AhelicopterdropfortheU.S.Treasury’,Vox,CEPR’sPolicyPortal(August30).

Camera,Gabriele(2017),‘Aperspectiveonelectronicalternativestotraditionalcurrencies’,Sveriges Riksbank Economic Review, no.1,pp.126–148.

Cúrdia,VascoandMichaelWoodford(2011),‘Thecentral-bankbalancesheetasaninstrumentofmonetarypolicy’,Journal of Monetary Economics,vol.58,no.1,pp.54–79.

DanmarksNationalbank(2017),‘CentralbankdigitalcurrencyinDenmark?’,Analysis,DanmarksNationalbank.

DeRezende,RafaelB.(2017),‘Theinterestrateeffectsofgovernmentbondpurchasesawayfromthelowerbound’,Journal of International Money and Finance,vol.74,pp.165–186.

DeRezende,RafaelB.andAnnukkaRistiniemi(2018),‘Ashadowratewithoutalowerboundconstraint’,WorkingPaperno.355,SverigesRiksbank.

ECB(2010),The Payment System. Payments, Securities and Derivatives, and the Role of the Eurosystem, ed.byKokkola,Tom,EuropeanCentralBank.

61S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

Eggertsson,GautiB.andMichaelWoodford(2003),‘Thezeroboundoninterestratesandoptimalmonetarypolicy’,Brookings Papers on Economic Activity,TheBrookingsInstitution,vol.34,no.1, pp.139–235.

Engert,WalterandBenS.Fung(2017),‘Centralbankdigitalcurrency:motivationsandimplications’,StaffDiscussionPapersno.16,BankofCanada.

Ennis,HubertoM.andToddKeister(2003),‘Economicgrowth,liquidityandbankruns’,Journal of Economic Theory,vol.109,no.2,pp.220–245.

Freixas,XavierandJean-CharlesRochet(2008),Microeconomics of Banking,MITPress:Cambridge.

Friedman,Milton(1969),‘Theoptimumquantityofmoney’,InMiltonFriedman,The Optimum Quantity of Money and Other Essays,1–50.Chicago:Aldine.

Gagnon,Joseph,MatthewRaskin,JulieRemacheandBrianSack(2010),‘Large-scaleassetpurchasesbytheFederalReserve:didtheywork?’,StaffPolicyReportno.441,FederalReserveBankofNewYork.

Galí,Jordi(2014),‘Theeffectsofamoney-financedfiscalstimulus’,DiscussionPaperno.10165,CentreforEconomicPolicyResearch.

Goodfriend,Marvin(2016),‘Thecaseforunencumberinginterestratepolicyatthezerobound’,inDesigning Resilient Monetary Policy Frameworks for the Future,FederalReserveBankofKansasCityEconomicSymposiumatJacksonHole2016.

Hamilton,JamesD.andJingC.Wu(2012),‘Theeffectivenessofalternativemonetarypolicytoolsinazerolowerboundenvironment’,Journal of Money, Credit and Banking,vol.44,no.1,pp.3–46.

Hasan,Iftekhar,TaniaDeRenzisandHeikoSchmiedel(2013),‘Retailpaymentsandtherealeconomy’,WorkingPaperno.1572,EuropeanCentralBank.

Holston,Kathryn,ThomasLaubachandJohnC.Williams(2016),‘Measuringthenaturalrateofinterest:internationaltrendsanddeterminants’,WorkingPaperno.11,FederalReserveBankofSanFrancisco.

Humphrey,David,MagnusWillesson,GöranBergendahlandTedLindblom(2006),‘BenefitsfromachangingpaymenttechnologyinEuropeanbanking’,Journal of Banking and Finance,vol.30, pp.1631–1652.

Juks,Reimo(2018)‘Whenacentralbankdigitalcurrencymeetsprivatemoney:theeffectsofane-kronaonbanks’,Sveriges Riksbank Economic Review,no.3,pp.79–99.

Keynes,JohnM.(1936),The General Theory of Employment, Interest and Money, Harcourt, Brace and Co.

Klein,MichaelA.(1971),‘Atheoryofthebankingfirm’,Journal of Money, Credit, and Banking,vol.3,no.2,pp.205–218.

Kocherlakota,Narayana(2016),‘Helicoptermoneywon’tprovidemuchextralift’,fromBloombergView,https://www.bloomberg.com/view/articles/2016-03-24/-helicopter-money-won-t-provide-much-extra-lift,retrievedonSeptember17,2018.

Krishnamurthy,ArvindandAnnetteVissing-Jorgensen(2011),‘Theeffectsofquantitativeeasingoninterestrates:channelsandimplicationsforpolicy’,Brookings Papers on Economic Activity,vol.2,pp.215–287.

Krishnamurthy,ArvindandAnnetteVissing-Jorgensen(2012),‘Theaggregatedemandfortreasurydebt’,Journal of Political Economy,vol120,no.2,pp.233–267.

Kumhof,MichaelandClareNoone(2018),‘Centralbankdigitalcurrencies–designprinciplesandbalancesheetimplications’,StaffWorkingPaperno.725,BankofEngland.

Laséen,StefanandRafaelB.deRezende(2018),‘Monetarypolicytransmissionandspilloversinanopeneconomyduringnormalandnegativeinterestrateperiods’,Manuscript,SverigesRiksbank.

Matthews,KentandJohnThompson(2014),The Economics of Banking, Wiley.

Meaning,Jack,BenDyson,JamesBarkerandEmilyClayton(2018),‘Broadeningnarrowmoney:monetarypolicywithacentralbankdigitalcurrency’,StaffWorkingPaperno.724,BankofEngland.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y62

Monti,Mario(1972),‘Deposit,creditandinterestratedeterminationunderalternativebankobjectivefunctions’,InKarlShell,GiorgioP.Szegö(Eds.),Mathematical Methods in Investment and Finance, 431–454.North-Holland:Amsterdam.

Nessén,Marianne(2016),‘CommentsonMarvinGoodfriend’,inDesigning Resilient Monetary Policy Frameworks for the Future,FederalReserveBankofKansasCityEconomicSymposiumatJacksonHole2016.

Nessén,Marianne,PeterSellinandPerÅsbergSommar(2018),‘Theimplicationsofane-kronafortheRiksbank’soperationalframeworkforimplementingmonetarypolicy’,Sveriges Riksbank Economic Review,no.3,pp.29–42.

Niepelt,Dirk(2018),‘Reservesforall?Centralbankdigitalcurrency,deposits,andtheir(non)-equivalence’,WorkingPaperno.2,SwissNationalBank,StudyCenterGerzensee.

Prisman,EliezerZ.,MyronB.SlovinandMarieE.Sushka(1986),‘Ageneralmodelofthebankingfirmunderconditionsofmonopoly,uncertainty,andrecourse’,Journal of Monetary Economics,vol.17,no.2,pp.293–304.

Rajan,Raghuram(2005),‘Hasfinancialdevelopmentmadetheworldriskier?’,inThe Greenspan Era: Lessons for the Future, FederalReserveBankofKansasCityEconomicSymposiumatJacksonHole2005.

SverigesRiksbank(2017),‘TheRiksbank’se-kronaproject’,Report1,SverigesRiksbank.

SverigesRiksbank(2018),‘TheRiksbank’se-kronaproject’,Report2,SverigesRiksbank.

Swanson,EricT.andJohnC.Williams(2014),‘Measuringtheeffectofthezerolowerboundonmedium-andlonger-terminterestrates’,American Economic Review,vol.104,no.10,pp.3154–85.

Wallace,Neil(1981),‘AModigliani-Millertheoremforopen-marketoperations’,American Economic Review,vol.71,no.3,pp.267–274.

Williams,JohnC.(2014),‘Monetarypolicyatthezerolowerbound:puttingtheoryintopractice’,HutchinsCenteronFiscalandMonetaryPolicyatBrookings.

Williamson,Stephen(2016),‘Scarcecollateral,thetermpremium,andquantitativeeasing’,Journal of Economic Theory,vol.164,pp.136–165.

Woodford,Michael(2012),‘Methodsofpolicyaccommodationattheinterest-ratelowerbound’,inThe Changing Policy Landscape,FederalReserveBankofKansasCityEconomicSymposiumatJacksonHole2012.

63S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

Appendix A. Theeffectsofane-kronaon pass-through:Abankingmodel

Inthisappendix,weuseanindustrialorganizationmodelofbankingtoanalysetheeffectsofane-kronaonthepass-throughfrompolicyratechangestobanks’depositandlendingrates.24Themodelhelpsstructuretheanalysisandprovidessomekeyinsights.Inparticular,itdemonstratesthattheconclusionsconjecturedinSection3holdinastandardbankingmodel.

ArepresentativebankTokeeptheanalysisassimpleaspossible,weconsideramonopolisticbankandourdiscussionfollowswhatisknownastheMonti-KleinmodelfromMonti(1972)andKlein(1971).Qualitatively,theresultswillbeanalogouswithanoligopolisticbankingsector.25

ConsiderarepresentativebankacceptingdepositsDandgivingoutloansL.26LetiL = iL(L)

denotethe(inverse) loan demand where i Listhelendingrate,andassume diL(L)dL <0.

Similarly, let iD = iD(D)denotethe(inverse)depositsupplywhere iD is the deposit rate, and

assume diD(D)dD >0.

The bank has also access to a money market, from which it can borrow and lend in unlimitedquantitiesatthepolicyrateiR.NotethatthisassumptionisreasonableinSwedenwheremonetarypolicyisimplementedinacorridorsystem,andwherethe(short-term)money-market rate is typically close to the policy rate.

Finally,thebankissubjecttomanagingcostsC(D,L)satisfyingCL(D,L)>0,CD(D,L)>0,CLL (D,L) ≥0andCDD (D,L)≥0.NotethatthesignofthemixedpartialderivativeCDL(D,L)isrelatedtothenotionofeconomiesofscope.IfCDL(D,L)<0,auniversalbankjointlyofferingloansanddepositsismoreefficientthantwoseparateentitiesspecializinginloansanddeposits. If CDL(D,L)>0,therearediseconomiesofscope.IfCDL(D,L)=0thereareneither.

Asamonopolist,thebanktakesintoaccountthatlendingdemandanddepositsupplydependontherespectiveinterestrates,whichareunderthecontrolofthemonopolist.Thebank’sprofitthereforeis:π = L[ iL (L)−iR ] + D[ iR −iD (D)]−C (L,D).

Thus,thebank’sprofitisthesumoftheintermediationmarginsonloansanddeposits,netofmanagingcosts.Thefirstorderconditionsforprofitmaximizationthenare:

(6) δπδL =0→(diL

dL L + iL ) = iR + CL (L,D)

(7) δπδD =0→iR = (diD

dD D + iD ) + CD (L,D)

Notethatmarginalrevenuesandmarginalcostsareontheleft-andright-hand-sideoftheequations,respectively.Thisimpliesthemonopolisticbanksetsthelendingandthedepositratessothatmarginalrevenuesinthetwomarketsequalmarginalcosts.

24 See,forexample,FreixasandRochet(2008)andMatthewsandThompson(2014)foratextbookpresentationofthemodel.25 SeeFreixasandRochet(2008),pp.79–80fordetails.26 Thissimplifiestheanalysis,butqualitativelytheresultsarethesameinmoreelaborateversionsofthemodelfeaturingseveralidenticalbanks.

T H E E - K R O N A A N D T H E M A C R O E C O N O M Y64

Thefollowingobservationisakeyresultinthemodel,andwerefertoitextensivelyinSection3:

Result 1: If there are no joint variable costs in the managing of loans and deposits, then the bank sets deposit and loan volumes separately.

Proof: Set CLD(L,D)=CDL(L,D)=0.Then,Result1followsfrom(6)and(7).■

Case1:Pass-throughwithoutane-kronaThefollowingresultholdsinthiscase.

Result 2: The pass-through from the repo rate to deposit and lending rates can be different from one to one.27

Proof:Ifthepass-throughisonetoone,thenthedepositintermediationmarginiR −iD mustbeconstant.From(7)itfollowsthatthiscanonlybethecaseif

(8) diD(D)dD D + CD (L,D)=k

where kisaconstant.Thus,equation(8)willonlyholdundersomeveryspecificassumptionsregardingthedepositsupplyandthemanagerialcostrelations.28 The proof for the pass-throughtolendingratesissimilar.■

Result2impliesthatwecanexpectthepass-throughtobetypicallydifferentfromonetoone without an e-krona.

Case2:Pass-throughwithane-kronaWeherebyexaminetheeffectsoftheintroductionofane-kronaonthepass-throughfromthepolicyratetolendinganddepositratesinthecaseofamonopolisticbank.LetiD

MON be the deposit rate that such a bank would set if there were no e-krona. If iEkr < iD

MON, an e-krona wouldbeunattractiveandthereforenotusedinequilibrium.29Inthatcase,theintroductionofane-kronawouldnotimpactthepass-through.

Beforelookingatthepass-throughwithanattractivee-krona,itisusefultonoticethatifthee-kronamarginiR −iEkrisfixed,theattractivenessofane-kronaandthereforealsothepass-throughmaydependonthelevelofthepolicyrate.Toseewhythatisthecase,noticethatitfollowsfrom(7)thattheoptimaldepositintermediationmarginiR −iD can beincreasinginD.Thatistrue,forinstance,ifthemarginalmanagerialcostisconstantorincreasinginD andthedepositsupplyfunctionislinearinD.Thus,ifthee-kronamarginiR −iEkrissufficientlylarge,theoptimalintermediationmarginiR −iD can be smaller than the marginiR −iEkr for DsmallerthanathresholdvalueD_. Thus, if D < D_,theprofit-maximizingmonopolistmayanyhowsetadepositratethatishigherthanthee-kronaratethusrenderingane-kronaunattractive.IfinsteadD > D_,thiswillnolongerbethecaseandane-kronawillbeattractive.

Thefollowingresultholdsforthecasewhenane-kronaisattractive.

Result 3: If the e-krona margin iR −iEkr is constant and an e-krona is attractive, then the pass-through from policy-rate changes to deposit rates will be one to one.

27 Pass-throughwillbeonetooneunderperfectcompetitionifCDD is constant as in that case the term diD

dD D disappears from expression(8).Similarly,thepass-throughtolendingrateswillbeonetooneunderperfectcompetitionandconstantmarginalmanagerialcosts.28 For example, this will be true if iD (D) = ln (D)andC(D,L)=ϒ DD + ϒ LL29 Notethatherewedisregardthegainsfromadditionalservicesandfromdifferencesinriskandsetϕt

ekr − σt

ekr = ϕtD − σt

D (seeequation(4)inSection2.2).

65S V E R I G E S R I K S B A N K E C O N O M I C R E V I E W 2018:3

Proof:IfiEkr ≤iDMON,weneedtoidentifytwoseparatecases:

(i) IfiEkr > iDBRE, where iD

BREisthebank’sbreak-evendepositinterestrate(i.e.iD

BRE D −C (D,L)=0),thebankwillceaseitsdeposit-takingactivitiesastheyareloss-making.Then,alldepositswillbee-krona.Moreover,thepass-throughwillbeonetooneaslongasthemarginbetweenthepolicy-rateandthee-kronarateisconstant.

(ii) IfinsteadiEkr ≤iDBRE the monopolist bank will set iD = i Ekr,andthepass-throughto

depositratesbecomesonetooneaslongasthemarginbetweenthepolicyrateandthe e-krona rate is constant. ■

Thefollowingresultalsoholds.

Result 4: If there are no joint variable costs in the managing of loans and deposits, the pass-through from policy-rate changes to lending rates will not be affected by the introduction of an e-krona.

Proof:ThisfollowsdirectlyfromResult1.■