The recent contributions from Brunnermeier and Niepelt (2019) and Andolfatto (2019) offer useful analytical frameworks to organise one’s thinking on how the introduction of a central bank digital currency (CBDC) is likely to impact the macroeconomy. The former derives an equivalence result whereby the CBDC need not alter resource allocation and the price system, while the latter shows that the introduction of the CBDC leads to an expansion of commercial bank demand deposits (CBDD). Both suggest that many of the concerns that have been raised in the recent literature on digital cash are in fact misplaced.
Below I point to a key criterion that presides over the relevance of the CBDC for the macroeconomy and discuss some more general results, also in light of the position of the international central banking community on the issue (BIS, 2018).
Whether or not the issuance and circulation of the CBDC in the economy would have relevant macroeconomic implications rests on the effect that the new instrument would have on agents’portfolio preferences, and specifically on whether or not the agents considered CBDC and CBDD to be perfect substitutes of one another –an issue on which definitive observations can only be ex post, once CBDCs will actually be in circulation. If substitutability were less than perfect, no equivalence result could be derived between the two types of money and a nonzero cost element would always be associated with any attempts to make them equivalent.
At times of high uncertainties, or crisis, the very existence of the CBDC could facilitate the switch from CBDD to it, as the international central banking community recognizes, since «a CBDC could allow for “digital runs”towards the central bank with unprecedented speed and scale. Even in the presence of deposit insurance, the stability of retail funding could weaken because a risk-free CBDC provides a very safe alternative» (BIS, 2018, p.16). This is because, in a fiat money system where money issuance is unconstrained by a convertibility rule to another asset or currency, such as under a fixed exchange rate or currency board arrangement, central banks (unlike commercial banks) can always cover their obligations by issuing their own currency.
Thus, the availability to the general public of a CBDC as the most safe and liquid asset in the economy, and its immediate accessibility on demand (literally, at the click of a mouse), could make it highly preferable to CBDD, especially at times of market stress. If this were the case, any compensation mechanism that would be adopted to establish the equivalence of public preferences between the two types of money would impose a cost either on the banks or on the public holding the CBDC, since banks would have to raise the remuneration on CBDD or, as Brunnermeier and Niepelt consider, the central bank would have to set an unattractive (possibly negative) interest rate on CBDC accounts in order for banks to retain their customers.
In the event of strong public preferences for the CBDC, commercial banks would have to liquidate part of their assets (e.g., securities, loans) in exchange for cash (central bank reserves), possibly depressing the value of the assets and incurring balance-sheet losses. Alternatively, the central bank might intervene and purchase those assets at par in exchange for cash; however, this would require the central bank to absorb the attendant risk. No equivalence result would follow from either option.
Bruennermeier and Niepelt argue that compensatory central bank’s pass-through funding would re-establish the equivalence. However, understanding the nature of such funding would be critical, in particular to identify the terms and conditions under which the central bank would extend funds to commercial banks and the cost associated with them. The central bank could lend pass-through funds at an interest or against collateral to protect itself from credit risk. In both cases, the banks would face a cost. Moreover, under collateralized lending and in the event of runs on banks, situations may arise where banks could find themselves unable to borrow due to inadequate collateral. All such cases would violate the equivalence result.
Alternatively, the central bank could transfer funds free of charge, but this would cause losses to the central bank balance sheet (detracting from its capital), it would have the nature of a fiscal operation, and might bear implications both for the fiscal budget and the relationship between the central bank and government.
A further issue is the use that commercial banks would (be authorized to) make of the new reserves funded by the central bank. Would they be held as excess reserves with the central bank or could they be invested in securities or other assets? And, in the former case, would the extra reserves be remunerated so as to compensate banks for the related opportunity cost? Note that if commercial banks were precluded from receiving compensation, the use of CBDC would cause them to face losses. On the other hand, reserve remuneration or profits generated by the reinvestment of reserves would derive from the fiscal subsidy originally extended to the banking sector in the form of central bank reserves made available to them at a lower or zero cost.
Importantly, any intervention of the central bank in support of the liquidity needs of commercial banks would imply the former filling in a hole in the latter’s balance sheet. All the above violations of the equivalence result would be the consequence of the public sector seeking to compensate commercial banks for the shift of agent portfolio preferences from CBDD to the CBDC.
Why then not take the full step forward and move to a fully-fledged narrow banking regime, thus eradicating the risk of run altogether? This question leads us to considering the impact of the CBDC on the cost of money to the economy, as follows next.
Digital cash and private banks
Whenever a narrow-banking type of monetary regime is introduced into the economy (even in the form of a CBDC), the cost of money is bound to rise (Bossone, 2001). The expression “narrow banking” refers to a money liability that is fully backed by a safe asset. In the case of fiat money issued by the state, or by the central bank in its stead, full safety would be ensured by the irredeemability of the fiat money into other assets or currencies, by the high-quality assets in the balance sheet of the issuing institution(s) and, ultimately, by the taxation power of the state and by the state (implicit or explicit) protection of the central bank’s capital. The CBDC would be narrow-banking money par excellence.
Andolfatto (2019) shows formally that, as long as the CBDC interest rate is below a positive policy rate, the CBDC has the effect of compelling banks to increase the deposit rate they offer on CBDD, since banks will always find it profitable to at least match the CBDC interest rate, a point already noted by the international central banking community in their evaluation of the CBDC (BIS, 2018). Thus, even if the take-up of the CBDC is small, its mere existence may be enough to encourage more attractive terms for depositors and therefore to expand the banking sector’s depositor base.
The effect of the CBDC on CBDD rates, however, would reduce the commercial bank seigniorage (as recognized by both Brunnermeier-Niepelt and Andolfatto) and/or translate into a higher cost of credit to the economy as banks would pass on their customers the higher cost of funding. Only general equilibrium analysis would tell what the impact would be on the aggregate demand for credit by the economy and, hence, on the aggregate volume of CBDD in the economy.
In any case, the higher cost of money would reflect the public’s willingness to pay for being able to access a fully safe and liquid (narrow-banking type of) money such as digital cash, which more than nondigital cash would be easily available to the public, convenient to use, and inexpensive to hold and administer. The very existence of such an instrument would make explicit and transparent to the public the higher riskiness of CBDD versus CBDC and would induce markets to assign a nonzero price to it.
The introduction of digital cash in the economy in the form of a general purpose central bank digital currency may have macroeconomic implications that relate to the portfolio preferences that the public would express for such type of instrument vis-à-vis the demand deposit liabilities issued by commercial banks and conventionally used by the public as money.
The stronger such preferences, the more relevant the implications, especially at times of market stress, both in terms of commercial bank losses (and, obviously, the cost to compensate them for such losses) and in terms of the demand for credit and demand deposits from the economy.
While agent portfolio preferences offer a key criterion as theoretical guidance to make predictions on the potential macroeconomic implications of digital cash, effective implications can only be determined empirically ex post, once central banks will actually start issuing digital cash and the public response to it will be measured.
Andolfatto, D (2019), “Central bank digital currencies and private banks, Vox, 17 March.
BIS (2003), “The role of central bank money in payment systems,” Report by the Committee on Payment and Settlement Systems, Bank of International Settlements, Basel, August.
BIS (2018), “Central bank digital currencies,” Report by the Committee on Payments and Market Infrastructures and the Markets Committee, Bank for International Settlements, Basel, March.
Bossone, B (2001), “Should Banks be Narrowed?,” IMF Working Paper, WP/01/159, October.
Bossone, B, M Costa, A Cuccia, and G Valenza (2018), “Accounting Meets Economics: Towards an 'Accounting View' of Money,” d/SEAS Working Paper No. 18-5. October 22.
Brunnermeier, M K, and D Niepelt (2019), “Public versus private digital money: Macroeconomic (ir)relevance,” Vox, 20 March.
 BIS (2018) further argues that, «Depending on the context, the shift in deposits could be large in times of stress (…) If CBDC were available, the incentives to run could be sharper and more pervasive than today, as the CBDC would be the favored destination, especially if deposits were not insured in the first place or deposit insurance was (made more) limited (…) Whereas weaker banks could experience a run, even stronger banks could face withdrawals in the presence of CBDC» (p.16).
 As referred to money, the concept of “safety” essentially means its likelihood of retaining its value to the holder, and hence its acceptability as a means of payment. Central banks are more creditworthy institutions than commercial banks as issuers of money. They have explicit or implicit state support. In addition, central banks tend to be risk-averse institutions that seek, as far as possible, to engage only in low-risk financial activities (BIS, 2003).
 On the nature of fiat money as equity, see Bossone, Costa, Cuccia, and Valenza (2018).