3 Should central banks issue their own digital currencies?
3.1 What are CBDCs?
There is no universally agreed definition of what constitutes a central bank digital currency (CBDC), but the term has become commonly used to designate any form of central bank digital fiat liability that is accessible to all economic agents.
As our taxonomy shows (Figure 1), CBDCs could take various forms. First, a CBDC could be issued in a centralised fashion though accounts at the central banks. This actually already exists in the form of reserves held by commercial banks at the central bank. The novelty of a CBDC would be that it would allow the general public to maintain directly deposit accounts at central banks, as they do at commercial banks. Second, a CBDC could also be issued in a decentralised fashion as a crypto-CBDC based on distributed ledger technology (DLT), which, once issued by the central bank, could then be transferred from one individual to another independently from the central bank. Finally, a CBDC could also take a hybrid physical-digital form with pre-paid cards or mobile wallets containing digital currencies.
We consider the last two versions of CBDCs less relevant for different reasons: 1) for crypto-CBDCs, DLT is an immature technology that is currently less efficient, slower and much more energy-intensive than a centralised system. The only advantage it offers is anonymity (which is valuable for individuals, but would probably be considered highly undesirable by monetary authorities, as it could facilitate illegal activities); 2) cards/wallets meanwhile would have properties very similar to cash and are thus not worth analysing in more detail.
We therefore focus on the more promising version of CBDC: deposit accounts at the central bank available to all.
3.2 What would be the purpose of CBDCs?
Interest in CBDCs on the part of authorities is partly motivated by the popularity of private digital currencies that could challenge the role of official currencies. Providing digital currencies issued by the central bank could possibly make private digital currencies less attractive and slow down their adoption. Allowing households and companies to open accounts at the central bank would give them direct access to efficient and instantaneous retail payment systems – such as the Target Instant Payment Settlement service that European banks can already use to exchange reserves. This would remove one reason for switching to a private digital currency with a better payment system. But there are other reasons why introducing CBDCs could be useful.
One important reason for central banks to provide CBDCs to citizens is that if cash is scarce or even disappears, citizens will lose direct access to sovereign money, the ultimate safe asset (as long as the central bank implements the necessary policies to maintain the value of the currency, ie low inflation and stable foreign exchange rates). Should cash disappear, citizens would only have access to bank deposits, which are not as safe, to store value. Deposits above a certain threshold (€100,000 in the euro area) are uninsured, and even below this threshold, there is the possibility of losing access to savings even for a few days or weeks.
In addition, the lack of direct access to the central bank currency could lead to a moral-hazard problem (Brunnermeier et al, 2019). If banks do not ‘fear’ convertibility of their deposits into central bank currency, they could lose some of the incentives (even though regulation would still be a major disciplining device) to manage well their solvency and liquidity risks. In extremis, if deposits do not have to be converted into a common currency, deposits from different commercial banks could at some point become imperfect substitutes for one another. In this case, there would be ‘exchange rates’ between them, depending on the trustworthiness of the particular issuer, as it was the case during the US free banking era in the nineteenth century. CBDCs would solve this problem by allowing households to access central bank currency in a new form, and thus restore the convertibility threat for banks.
The introduction of CBDCs could also strengthen monetary policy by transmitting it directly to the general public. Changes in policy rates would be transmitted directly to CBDC depositors, in contrast to today’s situation, in which interest paid by commercial banks on deposits are relatively sticky. This also means that CBDCs would make unconventional policies easier to implement.
First, as long as the CBDC is interest-bearing, it could help relax further the zero lower bound constraint because interest rates applied to the CBDC could be negative (unlike for banknotes). The abolition of cash would make this effect stronger. However, abolishing cash might be not desirable, because cash could still be useful at least as a back-up for a CBDC in case of a technical failure or cyberattack, and for privacy reasons. But even if cash continues to exist, as long as its use is inconvenient (which would be even more the case if CBDC were introduced) and its storage costly, implementing negative rates on CBDC holdings would be possible.
Second, CBDCs could reduce one of the potential side effects of quantitative easing (QE), especially when asset purchases are coupled with negative rates. Currently, central bank bond purchases from non-bank institutions create additional reserves that are inevitably held by the commercial banks that host the accounts of the non-bank sellers in the deposit facility of the central bank, because non-banks cannot hold reserves directly. On aggregate, this means that banks cannot control fully the quantity of reserves they want to hold. When rates are negative, as they are at time of writing, this becomes costly for banks and might result in potential side effects such as increased rates for lending to the real economy. If non-banks could hold CBDCs directly, QE would not affect the banking sector negatively.
Finally, provided the concept of helicopter money is an acceptable monetary policy tool, it would be easier to implement if all citizens had accounts at the central bank, because the central bank would be able to credit their accounts with CBDC units.
3.3 The potential risks of CBDCs
The introduction of CBDCs is sufficiently disruptive that it could pose a number of risks.
First, one the main fears of policymakers (see, for example, Coeuré, 2018) is that CBDCs will lead to cyclical bank runs. If households and companies have access to central-bank reserves, there is a risk of a flight-to-safety from commercial-bank deposits to CBDCs in each economic downturn. This type of run from banks to the central bank happened in the 1930s during the Great Depression in France, when it was possible for non-banks to maintain accounts at the Banque de France (Baubeau et al, 2018). Bank runs are already possible today by withdrawing cash or transferring deposits between banks, but the main concern is that digital bank runs towards CBDCs would be easier and happen more rapidly than traditional bank runs.
In addition to this cyclical financial stability risk, another serious, more structural, risk would be the reduction of financial intermediation. Banks would compete with the central bank to hold deposits. It is very difficult to predict what would happen, because it would depend on the particular properties of the CBDC introduced and on the behaviour of the central bank after its introduction, but this could lead to different outcomes (as explained, for instance, by Stevens, 2018).
A first possible outcome could be an evolution towards a financial system characterised by narrow(er) banks that are less reliant on deposits. Banks could either offer higher returns to depositors to try to retain their deposit base, or they could rely on other sources of financing. This would have profound implications, both potentially positive and negative. The extra competition from CBDCs would reduce the monopoly power of the banking sector and allow depositors to obtain higher returns from their deposits. For banks, by definition, the effect would be the opposite because they could be forced to rely on more expensive and potentially less stable sources of funding, such as the wholesale market. This new banking model would make banks look more like investment funds, which could be less stable thus requiring an adjustment of the financial safety net. The need for traditional deposit insurance would be reduced because deposits could be kept safely in the form of CBDCs. However, if we consider that the maturity transformation provided by banks is a valuable service, then it needs to be protected from liquidity risk. Either insurance cover for banks’ short-term liabilities would have to be broadened to include wholesale funding, with all the risks that this would entail (but the alternative would be frequent ‘wholesale runs’ such as those that happened during the last financial crisis), or regulation would have to be toughened significantly to avoid any maturity mismatch on banks’ balance sheets, for example by forcing them to be financed mainly through equity and long-term debt.
Another possibility would be a tightening of credit conditions by banks if they are unable to retain depositors or attract new sources of funding. This tightening would lead to less lending and/or at a higher price, which would, all else being equal, result in a significant drag on investment and ultimately on economic activity.
3.4 How could central banks minimise these risks?
Policymakers have several tools at their disposal, should bank runs become more frequent as a result of the introduction of CBDCs. First, deposit insurance offsets the risk of runs when deposits are within the guaranteed amount. Second, the central bank should play its crucial role of lender of last resort by providing liquidity through loans to the banks that suffer runs, as long as they are solvent. The financial instability episodes in France in the 1930s discussed in Baubeau et al (2018) showed that the main problem was not the bank runs (towards the central bank or towards other saving institutions) per se but rather the strong “gold standard mentality” prevailing at the Banque de France at the time. This mentality prevented the central bank from playing its role lender of last resort and from replacing the shortfall in deposits held at commercial banks with central bank loans to avoid a strong credit crunch.
Central banks would also have various instruments to counter the risk of structural financial disintermediation that could happen as a result of the introduction of CBDC, if it was considered an unfavourable evolution that could endanger price or financial stability. The central bank’s reaction would thus vary depending on the magnitude of the problem.
In moderate cases, such as if the quantity of credit provided by banks is not significantly affected, but banks ask for higher lending interest rates (for example because they need to increase the returns paid to depositors to retain them), the central bank would have to lower its policy rates structurally to offset this effect and maintain financial conditions at the same (presumably adequate) level, all else being equal. In normal times this should not be a particular problem, but at a time when the effective lower bound is binding, it might be problematic, and might involve the increased use of unconventional monetary policies.
If disintermediation becomes a more significant issue and there is clear downward pressure on bank credit availability, the main way for the central bank to offset this trend would be to provide structurally more funding to the commercial banks to replace the lost deposits, so that they can maintain the same level of financing to the economy. This means the central bank balance sheet would have to become structurally much bigger and also more exposed to the banking sector than has traditionally been the case.
The debate on the optimal size of central banks’ balance sheets has not so far been settled. However, the two main risks for central banks in increasing massively their refinancing operations would be:
- First, the central bank would take more risks onto its balance sheet because it would be more exposed to the risks faced by banks: in a way, the central bank would become itself a financial intermediary between depositors that would hold CBDCs and the commercial banks.
- Second, this means that the central bank would be involved more directly in the credit allocation process. In order to be able to provide a much greater amount of refinancing to the banks, the central bank might have to adjust significantly its collateral framework so that banks are able to access its operations at a sufficient scale. Central banks’ decisions on collateral eligibility and haircuts are often perceived as purely technical decisions, but they are not always as neutral as they seem (Claeys and Goncalves Raposo, 2018). In particular, deciding to include new asset classes as eligible collateral (in order to increase the pool so that banks can obtain more refinancing) could have some powerful effects on credit allocation by the banks. The main advantage is that this would give the central bank greater control over the macroeconomic situation, but the drawback would be that it could potentially make the overall allocation of resources in the economy less efficient, and could also have some distributional effects (that should preferably be in the hands of citizens or elected officials).
To avoid the extreme situation in which deposit accounts held at the central bank would fully crowd out bank deposits, the central bank could also try to carefully calibrate the properties of CBDCs in order to reduce ex ante the incentive to use a CBDC as a main store of value. The simplest way to do this would be through its remuneration system. CBDC accounts should be remunerated at a rate below other policy rates (which could both reduce the structural disintermediation risk and the frequency of bank runs). However, the returns from CBDCs should not be so disadvantageous that their use as a medium of exchange becomes unattractive. In particular, when policy rates are negative, a portion of CBDC holdings could be exempted from the negative rates to avoid the negative impact on small savers (and also so that households are not given a reason to switch back to holding cash). Bindseil (2019) proposed a very practical system to put that in place with a two-tier remuneration system for CBDCs: below a threshold of €3,500, CBDC holdings would be remunerated at the maximum level between the deposit rate and 0, and above that threshold CBDC holdings would be remunerated at the deposit rate minus 200 basis points. These numbers are indicative and the central bank would need to experiment to find the right balance, so that it incentivises the use of CBDCs as a medium of exchange, and gives access to everyone to the ultimate safe asset when necessary (especially if cash disappears), but disincentivises use of these accounts as a main store of value in normal times.
3.5 Who else could provide an equivalent to CBDCs?
Finally, an alternative solution to give the general public access to digital central bank liabilities would be not to provide it directly through a CBDC, but to do it indirectly through what could be considered ‘full-reserve banks’ (sometimes also referred to as ‘narrow banks’).
The idea, as described for example by Adrian and Mancini-Griffoli (2019), would be to allow new entities to hold reserve balances at the central bank, subject to some specific conditions. These entities would have a very particular balance sheet with only central bank reserves as assets (they would not give credit, nor buy any other type of asset) and only simple deposits as liabilities (they probably would not need to hold much capital, if any at all, given the absence of risk from their portfolios). The entity would pass the remuneration of central bank reserves to depositors and earn a small fee for the service provided.
This system would allow households and companies to hold indirectly the central bank currency and would have two additional advantages. First, it would allow central banks to focus on their mandates and not use their resources to provide direct services to their new customers (which could also have some negative reputational consequences for central banks if not handled properly). If all households and companies in the euro area opened CBDC accounts at the ECB, the number of accounts in the Eurosystem would grow from around 10,000 to more than 500 million (Bindseil, 2019). Second, as argued by Bordo and Levin (2019), this would help prevent a conflict of interest for the central bank. Competition from a CBDC could be considered unfair by banks given the crucial role central banks play in the organisation of the banking sector (for instance as a supervisor, among other functions). For all these reasons, privately-managed alternatives to CBDCs should not be discarded by central banks. Rather, they should be considered as one way to provide CBDCs (which would represent an acceptable form of stablecoin labelled in the central bank unit of account) to the general public.