中央银行数字货币与货币政策传导.pdf

Bank of Canada staff working papers provide a forum for staff to publish work-in-progress research independently from the Banks Governing Council.This research may support or challenge prevailing policy orthodoxy.Therefore,the views expressed in this paper are solely those of the authors and may differ from official Bank of Canada views.No responsibility for them should be attributed to the Bank.DOI:https:/doi.org/10.34989/swp-2024-27|ISSN 1701-9397 2024 Bank of Canada Staff Working Paper/Document de travail du personnel2024-27 Last updated:July 25,2024 Central Bank Digital Currency and Transmission of Monetary Policy by Saroj Bhattarai,1 Mohammad Davoodalhosseini2 and Zhenning Zhao1 1 University of Texas at Austin saroj.bhattaraiaustin.utexas.edu znzhaoutexas.edu 2 Banking and Payments Department Bank of Canada mdavoodalhosseinibank-banque-canada.ca i Acknowledgements The views expressed here are those of the authors,and no responsibility should be attributed to the Bank of Canada.We would like to thank Pierpaolo Benigno,Martin Eichenbaum,Jos-Vctor Ros-Rull,Martin Schneider,the participants of the Society for Economic Measurement conference and many colleagues at the Bank of Canada for their comments.ii Abstract How does the transmission of monetary policy change when a central bank digital currency(CBDC)is introduced in the economy?Do aspects of CBDC design,such as how substitutable it is with bank deposits and whether it is interest bearing,matter?We study these questions in a general equilibrium model with nominal rigidities,liquidity frictions,and a banking sector where commercial banks face a leverage constraint.In the model,CBDC and commercial bank deposits can be used as a means of payments,and they provide liquidity services to households.Banks issue deposits and extend loans to firms,and bank deposits are backed by loans and central bank reserves.We find that the effects of a canonical monetary policy shock,a shock to the Taylor rule that governs interest on central bank reserves,is magnified with the introduction of a fixed-interest-rate CBDC.More generally,whether CBDC magnifies or abates the response of the economy depends on the type of shock(e.g.,interest rate or quantity of reserves shock).We also find that the response of the economy depends on the monetary policy frameworkwhether the central bank implements monetary policy through reserves or through CBDCas well as central bank balance sheet rules that govern the quantity of CBDC and reserves.Topics:Digital currencies and fintech;Monetary policy;Monetary policy framework;Monetary policy transmission;Interest rates JEL codes:E31,E4,E50,E58,G21,G51 Rsum Comment lintroduction dune monnaie numrique de banque centrale(MNBC)dans lconomie influence-t-elle la transmission de la politique montaire?Les caractristiques de cette MNBC,notamment quel point elle est substituable aux dpts bancaires ou si elle porte intrt,ont-elles de limportance?Nous analysons ces questions au moyen dun modle dquilibre gnral comportant des rigidits nominales,des frictions relatives la liquidit et un secteur bancaire dans lequel les banques commerciales sont assujetties une contrainte de levier dendettement.Dans ce modle,la MNBC et les dpts dans les banques commerciales peuvent tre utiliss comme modes de paiement et constituent un vecteur de liquidit pour les mnages.Les banques offrent des services de dpt et octroient des prts aux entreprises,et les dpts bancaires sont garantis par des prts et les rserves de la banque centrale.Nous constatons que les effets dun choc de politique montaire standard cest-dire un choc touchant le taux dintrt des rserves de la banque centrale,rgi par la rgle de Taylor sont amplifis par lintroduction dune MNBC taux dintrt fixe.De faon plus gnrale,cest le type de choc(p.ex.,choc au niveau des taux dintrt ou de la quantit de rserves)qui dtermine si la MNBC amplifie ou attnue la raction de lconomie.Nous notons aussi que cette raction dpend du cadre de politique montaire,soit le moyen utilis par la banque iii centrale pour mettre en uvre la politique montaire(les rserves ou la MNBC),et des rgles qui encadrent le bilan de la banque centrale et rgissent la quantit de MNBC et de rserves.Sujets:Monnaies numriques et technologies financires;Politique montaire;Cadre de la politique montaire;Transmission de la politique montaire;Taux dintrt Codes JEL:E31,E4,E50,E58,G21,G51 1IntroductionMany central banks are contemplating issuing a central bank digital currency(CBDC)andare concerned about the implications.As a means of payment,a CBDC would compete withbank deposits and thereby have implications for the banking and financial system.As a storeof value,CBDC would be used along with bank deposits,government bonds,and other safeassets,which would have macroeconomics implications.A growing body of literature hasstudied such implications recently.However,most studies have focused on the steady-stateor long-run effects.Fewer papers have studied the transitory and short-run effects of shockson both banking and the macroeconomy in the presence of CBDC.1In this paper,we propose a framework that can help understand the effects of shocks onoutput,consumption,inflation,and investment through various channels in the presence ofa CBDC.The framework is built on a New Keynesian(NK)model with financial frictionsthat includes two new key features.Firstly,CBDC competes with deposits in providingliquidity services to households,2with the elasticity of substitution between the CBDC anddeposits a design feature of the CBDC.Secondly,the model considers a general equilibriumchannel in that,on the one hand,the production of final goods relies on loans made bybanks.On the other hand,consumers use deposits issued by banks as a source of liquidityneeded to buy the final goods.As the demand or supply of CBDC changes,so does thedemand for bank deposits.This changes banks cost of funding,which in turn alters the costof loans for firms.Subsequently,the supply side of the economy is affected.Overall,ourframework offers a flexible approach to studying the impact of CBDC on monetary policytransmission by considering its effects on both the demand and supply sides of the economy.To understand the basic economic forces in the model,we first study the impact of a canonicalmonetary policy shock,an increase in the interest rate on central bank reserves,in theabsence of a CBDC.There are three main channels through which the shock transmits to theeconomy.Firstly,a standard NK channel,where an increase in bond interest rates leads tolower current consumption and thus reduced aggregate demand and output,which impactsinflation via the NK Phillips curve.Secondly,a New Monetarist(NM)channel,where anarrower spread between illiquid bond and bank deposit rates reduces the opportunity costof holding deposits(cost of liquidity),thereby boosting deposit demand,consumption,laborsupply,and ultimately output.This offsets some effects of the initial shock.Lastly,asupply channel,where higher capital costs arising from higher costs of bank funding reducesinvestment,leading to a decline in output.These three channels together illustrate theinterplay of economic forces following a monetary policy shock,showing both contractionaryand expansionary effects on the economy.1For steady-state effects,see for example,Barrdear and Kumhof(2022);Davoodalhosseini(2022);andChiu and Davoodalhosseini(2023).For transitory effects,see for example,Minesso et al.(2022),who examinehow a CDBC affects the international transmission of monetary policy and technology shocks.2We extend the model in the appendix to include cash.In the main text,we do not include cash to makethe exposition simpler.2We then turn to studying the effects of shocks in the presence of a CBDC.Given thatmany central banks are considering a zero-interest CBDC,we focus on such a CBDC inthis exercise.The first takeaway from our paper is that the introduction of a zero-interestCBDC magnifies the effects of a traditional monetary policy shock.This amplificationremains regardless of the substitutability between deposits and the CBDC.An increase inthe reserves interest rate means that the leverage constraint becomes less costly for banks.Banks are therefore willing to offer more deposits.To attract depositors to hold moredeposits,the opportunity cost of holding deposits,i.e.,the spread between illiquid bondsand bank deposits(deposit spread),should fall.At the same time,the illiquid bonds interestrate tends to rise in response to the increase in the reserves rate.In the absence of a CBDC,the decline in the deposit spread suggests that the effective real wage increases,so therewill be a subsequent boost in labor supply,consumption,and output.This is the same NMchannel stated above.However,when a zero-interest CBDC is introduced,a fixed interestrate on the CBDC implies that the rise in the illiquid bond rate elevates the opportunitycost of holding CBDC.Consequently,the presence of a CBDC tends to heighten the overallcost of liquidity in the economy,thereby attenuating the NM channel.As the NM channelcounteracts the NK channel,its attenuation means amplification of the decrease in output.Whether a CBDC amplifies or abates the effects of a shock compared with the benchmarkmodel without a CBDC also depends on the type of monetary policy shock.In contrast tothe response to a standard monetary policy shock that is a shock to the interest on reserves,we find that a CBDC abates the effects on consumption,output,and inflation of a reservequantity shock.A positive reserve quantity shock is expansionary with or without a CBDC.Without a CBDC,investment declines considerably initially and starts recovering only aftera few quarters.With a CBDC,the decline in investment is much smaller.This explains whythe overall effect on output and consequently consumption is dampened with a CBDC.Another takeaway from our paper is that the response of the economy depends on themonetary policy framework,whether the central bank implements monetary policy throughreserves or through a CBDC.To see this point,we compare the outcomes in response to thesame shock with two monetary policy rules:one in which the central bank uses a Taylorrule to set the CBDC interest rate and another one in which it uses a Taylor rule to set thereserves interest rate.Our analysis focuses on the responses to a positive CBDC interestrate shock under these two monetary policy rules.We find that when CBDC serves as theprimary tool for monetary policy,an increase in the CBDC interest rate is contractionary,resembling traditional monetary policy effects observed in most NK models.Alternatively,when reserves are the main tool,the shock to the CBDC interest rate is expansionary due tothe reduced opportunity costs of liquidity for households,leading to increased labor supplyand consumptiona mechanism akin to the NM channel.3This result underscores theimportance of monetary policy framework in shaping economic responses.Finally,we show that the response of the economy to a standard monetary policy shockdepends on balance sheet quantity rules that govern the evolution of the central bank balance3Here,the CBDC interest rate changes exogenously and does not follow a Taylor rule.3sheet variables.To best illustrate this point,we compare two scenarios,one in which thecentral bank fixes the CBDC rate and one in which it fixes the CBDC quantity.In responseto a standard monetary policy shock(i.e.,a shock in the Taylor rule that governs theinterest rate on reserves),fixing the CBDC interest rate leads to a significant fall in outputand consumption compared to fixing the CBDC quantity.When the CBDC interest rateis fixed,the opportunity cost of holding CBDC rises with an increase in the illiquid bondrate,making liquidity more expensive and reducing consumption and output further.Whenthe CBDC quantity is fixed and its interest rate is flexible,the CBDC rate increases inresponse to pressure from rising deposit rates,activating the NM channel and mitigatingcontractionary effects of the shock.This exercise thus highlights the importance of thebalance sheet quantity rules,which is in stark contrast to standard NK models where thequantity of the real balance has no independent effect on real variables once the short-termnominal rate is determined.Literature.The literature on CBDCs has grown significantly in recent years,covering awide range of topics.First,we compare our results with some closely related papers includingPiazzesi et al.(2019),regarding the transmission of monetary policy shocks,and Chiu et al.(2023),regarding the main economic forces at play in the steady state.Next,we discuss therelated literature more broadly.Piazzesi et al.(2019)study an NK model with money in the utility function and complemen-tarity between consumption and money.There are two key differences between our modeland theirs.First,in our model,the CBDC and bank deposits both provide liquidity tohouseholds and could be substitutes or complements depending on the design of the CBDC.In their paper,agents either use central bank money(Section 2)or bank deposits(Section3)and it does not consider the effects of a CBDC on the banking system.Second,banksin our model lend to firms to finance their capital expenditure.In Piazzesi et al.(2019),banks simply invest in some assets with an exogenous rate of return.Modeling the inter-actions between deposits and other means of payments provides insight into how changesin the interest rate of a CBDC or its design features affect demand for bank deposits(thefirst difference),which in turn changes the cost of funding for firms,eventually affecting thesupply side of the economy(the second difference).In Chiu and Davoodalhosseini(2023),banks issue deposits and extend loans to firms,andthe CBDC and bank deposits compete in the sense that both can be used in a fraction oftransactions.The similarity between their model and ours is that in both models,the supplyside of the economy is affected by the demand for means of payments.They find that anincrease in the CBDC interest rate can improve intermediation,consumption,and output inthe steady state because an increase in the CBDC rate increases aggregate demand,leadingfirms to demand more loans to finance production.Thus,intermediation,consumption,andoutput can increase.A similar channel works in our model.However,Chiu and Davoodal-hosseini(2023)do not study dynamic responses to shocks and focus only on the steady-stateanalysis.Moreover,our model allows for a wider range of design features in terms of comple-mentarity or substitutability between a CBDC and other payment methods.Finally,in our4model,banks face financial frictions,which gives rise to demand for central bank reserves.As a result,monetary policy affects this economy via a richer set of policy tools(i.e.,inflationrate,the CBDC interest rate and quantity,and the reserves rate and quantity)relative totheir paper,in which there is no demand for ce