The cryptocurrency economy

‘Cryptocurrencies’ or, less commonly but more accurately, crypto-assets are, for the time being at least, not as widely used as their media coverage might suggest. As a result, for most of us our view of them is dominated by a perception of their high level of technological sophistication and remains fairly vague. Although some professionals and fans of new technologies are very enthusiastic, an economist examining the topic might be more circumspect.

These contrasting positions led us to draw up an initial taxonomy in an attempt to define what they are and, more importantly, what they are not (genuine currencies). This economic definition serves as a preamble to a section on the state of the science. This is still relatively sketchy, but pathways can be drawn with the debate opened up by Friedrich August von Hayek’s arguments for competing currencies at the end of the 1970s.

Our thought process then leads us naturally on towards ‘central bank digital currencies’. Often presented by central banks themselves as substitutes for private ‘cryptocurrencies’, in reality there are significant differences in terms of the consequences they could have for the financing of the economy and for financial stability. By virtue of their similarities with ‘narrow banking’ or the Swiss ‘sovereign money’ proposals (convincingly rejected by Swiss voters in 2018), ‘central bank digital currencies’ could change the money creation process as we know it today and affect the cost and volume of financing. They would also create the risk of a run on ‘digital’ banks. Their possible adoption, and their characteristics when adopted, will need to be carefully considered in order to reduce these risks.

Are ‘cryptocurrencies’ currencies at all? (no)

Three shared characteristics allow us to define ‘cryptocurrencies’. Two criteria are universal, the virtual nature and the cryptographic technique of these assets, whilst one – decentralisation – is common but optional. The first stage is to develop a taxonomy of ‘private’ crypto-assets, according to their main characteristics, and consider the extent to which they are currencies.

Shared characteristics

The origins of ‘cryptocurrencies’ date back to the aftermath of the financial crisis of 2008[1]. They were initially supported by an upswelling of libertarian current, in turn underpinned by a desire to enable the settlement of transactions in a way that avoided commercial, and to a lesser extent, central banks. They were also encouraged by a desire to avoid major established currencies like the dollar and euro. In such circumstances, it is natural that the main innovation of the original ‘cryptocurrencies’ lies in the removal of the trusted third party, a role hitherto played by commercial banks for transactions denominated in official currencies, and the option of conducting direct ‘peer-to-peer’ transactions.

‘Cryptocurrencies’ are virtual

‘Cryptocurrencies’ are first and foremost virtual, with no material reality. Unlike other forms of digital money (electronic money in digital wallets, script money in bank accounts), they are not regulated.

Use of cryptography

Decentralisation is a common, but not systematic, feature of ‘cryptocurrencies’. Whilst real currencies are managed centrally by a central bank, each participant (associated to a ‘node’, in reference, among other things, to a computer on a network) can offer or approve transactions in a distributed ledger (see Diagram 1). In the absence of a trusted third party (financial intermediary or bank), the security of transactions is provided by cryptography, that is to say by encryption algorithms.

For example, the Bitcoin cryptocurrency, which has been in existence since 2009, has so far proved itself extremely resistant to attacks and falsification. The whole community of developers has succeeded, thanks to the blockchain, in collectively ensuring the security of transactions. The procedure of validating and authenticating transactions is known as ‘mining’. This involves making computer processing power available to the network to solve complex mathematical problems. Blocks of transactions are recorded in a public distributed ledger (which can be read by all members of the network) listing all Bitcoin transactions since this digital currency was launched.

The use of cryptography seeks to secure transactions made over the internet by allowing access to information only to members of the distributed ledger (including new entrants). In most cases ‘cryptocurrencies’ operate as a distributed register from which all the information is simultaneously accessible to all the participants. Transactions are thus validated by ‘consensus’.
Some networks are known as ‘permissioned’ which is to say that access to networks is limited to authorised members, who have been pre-designated or who meet certain criteria. Almost all ‘cryptocurrencies’ are based on blockchain technology (see Box).

The dynamic of the amount of currency issued is determined by protocols that vary from one cryptocurrency to another. Thus, in the case of Bitcoin, the flow of new issuance (through mining) is halved every four years, with a ceiling on issuance capped at 21 million units, a limit that is set to be reached in 2140.

Taxonomy of crypto-assets

As demonstrated by the collapse in the value of Bitcoin after the peak reached in December 2017 (see Chart 1), the highly volatile nature of the early crypto-assets crystallised the criticisms that crypto-assets are unable to function as a ‘store of value’ and to offer a low-risk asset to its users. The puzzlement of authorities and central banks led to the development of characteristics of a new generation of crypto-assets, which have received a somewhat warmer welcome: stablecoins. These digital assets are typically backed by a basket of stable assets (Libra) or a guarantee of convertibility (JPM coin), giving them an intrinsic value.

First-generation crypto-assets (Bitcoin-type)

This type of digital asset is not a claim held on its issuer, unlike official currencies which represent a claim on the central bank (commercial bank deposits with the central bank, fiat money) or the issuing credit institution (bank deposits). By virtue of this first characteristic a real currency has the de facto backing of balance sheet assets, albeit that these can be of variable quality or liquidity (the balance sheet of the central bank or issuing commercial bank). In the case of a commercial bank, assets are, on average, significantly less liquid than bank deposits are for their holders. Prudential regulations require banks to create reserves at the central bank equivalent to a certain percentage of customer deposits (1% in the euro zone since 2012).

Conversely first-generation ‘cryptocurrencies’ have no intrinsic value and thus nothing, other than the trust their users placed in them, serves to guarantee their value over time. Their relative scarcity is not a sufficient condition to ensure that their value remains within an acceptably narrow range to ensure a relative stability of their price. Thus, the fall in demand for Bitcoin in 2018 resulted in a brutal collapse in its price.

Stablecoins

Stablecoins, issued by entities which back them with stable assets in one way or another, were looked on more favourably by central banks. For example, the ‘JPM Coin’, from US bank JP Morgan, which completed its test phase in February 2020, falls into this category. This is a ‘wholesale’ crypto-asset (for use by financial institutions wishing to make use of a dedicated blockchain) which is tradeable at parity with the dollar and backed by the issuer’s guarantee. Facebook’s Libra project seems, however, to have dented the central banks’ more positive view of stablecoins. Like official currencies stablecoins represent a claim on their issuers, the quality of whose balance sheet is, more or less, that of a basket of more or less stable assets. They are thus similar to units in a fund. In any monetary analysis, it should be stressed that units issued by money multual funds in the eurozone are included in the M3 broad money aggregate. Their issuers are part of the Monetary Financial Institution (MFI) sector alongside credit institutions. Stablecoins differ in at least two respects. First, their issuers are not necessarily money market funds and are therefore not subject to the same regulations as the latter. Secondly, stablecoins are designed to be used (at least for those who accept them) as a means of settlement of a transaction or to extinguish a debt; to be used for these purposes, money market fund units need to be sold or redeemed for cash in the narrower definition (M1, consisting of sight deposits, notes and coins).

Facebook’s Libra project is perhaps the best known of these stablecoins. This is intended, over time, to become a virtual means of payment backed by a basket of stable assets denominated in the main global currencies. The ‘exchange rate’ with the basket will be, by construction, maintained. The issuance of any additional quantity of Libra will require the purchase of the same combination of stable assets in an amount determined by the exchange rate. This represents an initial limit for Libra: sellers of stable assets may accept Libra in settlement. But under such circumstances, all newly issued Libra would be issued to these sellers. This would not allow demand for new Libra in exchange for currency from new buyers to be satisfied, unless the issuing entity purchased these from a third party for cash.

The most likely approach would therefore be for sellers of stable assets to be paid in one of the major currencies. Thus, the issuance of a Libra would have as its counterparty the receipt of a quantity of currency determined by the Libra exchange rate, which would then be used to settle the purchase of stable assets. This situation would create an unbreakable link between the major currencies and Libra, which might seem paradoxical for a new instrument that aims to compete with precisely these currencies (see Diagram 4).

In the final analysis, a stablecoin is broadly comparable to a money market fund in which the units are digitised and can be traded on a blockchain. Unlike official currencies pegged to the dollar under the Bretton Woods agreements (1944 to 1971), stablecoins are not directly linked to a currency, but backed by ‘stable’ assets. As in the case of Libra, these can be in a range of currencies. This diversification naturally creates an exchange rate risk, not only with reference to any given user’s domestic currency but more importantly relative to any of the benchmark currencies making up the basket.

The particular case of tokens

Another innovation in crypto-assets lies in the digitisation of certain physical assets (such as works of art) or intangible assets (patents, copyright) in the form of tokens, which are digital assets that represent a right to a future service (native token) or existing item (non-native token). Under the same principles as ‘cryptocurrencies’, these tokens can be exchanged on the internet without the intervention of a third party. The ledger for each protocol can function independently of the tokens, whereas primary crypto-assets (bitcoin, ether, ripple, etc.) are indivisible from it. A specific type of transaction, the Initial Coin Offering (ICO), allows the raising of funding in ‘cryptocurrency’.

In France, the Autorité des Marchés Financiers (AMF) defines such deals as “a fundraising transaction carried out through a distributed register system (or "blockchain") and resulting in a token issue. These tokens can then be used to obtain goods or services, as the case may be.” In common with share-based funding rounds, these transactions allow companies to raise funds at an early stage of their development.

However, they differ from Initial Public Offerings (IPOs), with which they are often compared, to the extent that, unlike shares, tokens do not give their holders rights over the company’s share capital but over the products or services that will be provided by the company in future. The can thus be considered as pre-sales (or pre-financing), allowing the company to receive cash in advance of the completion of the project and the subscriber – demonstrating their confidence in the company – to receive, on relatively attractive terms, rights over the products or services offered by the company. Although it has been used primarily by start-up companies so far, this solution is a potential option for any company planning to sell a new product or service in the future. Tokens thus span a wide range from digitised assets to pre-financed projects, making the market for them more narrow and less liquid than that for - coins[2].

In France, Law 2019-486 of 22 May 2019 (the ‘PACTE Act’[3]) introduced a specific regime for ICOs, establishing the principle of approval by the AMF. This new regime, designed to encourage the development of ICOs, does not apply to the issue of tokens which can be considered as financial securities (Security Token Offerings, or STOs) but only to so-called utility token issues. Article 26 thus created a legal framework for ICOs, with the possibility of the AMF providing approval for proposals that it believes to be serious. It is worth noting that this approval is not required, and issuers are free to seek it or not. However, those who have not received approval may not solicit investment from the general public. The AMF issued its first approval to French-ICO, a financing platform for cryptocurrency projects, in December 2019. The approval will be valid until the end of the subscription period, which is expected to be on 1 June 2020.

What does the economic theory say?

‘Cryptocurrencies’ fulfil the functions of a currency only very imperfectly. ‘cryptocurrencies’ promoters sometimes lay claim to Hayek’s proposals, but it is clear that one of the main arguments on which these were based (the inflation associated with official currencies), whilst powerful in the 1970s is much weaker in current conditions (see Chart 2). Indeed, the European Central Bank is working hard to bring the annual inflation rate back to its target level (below, but close to 2%) without much success so far (inflation was an estimated +1.2% in February 2020).

The three traditional functions of a currency

Money has taken many forms down the ages: shells (cowrie or porcelain money, the first traces of use of which date back to the Chang Dynasty in China [1600-1046BC]), stone money on the Micronesian island of Yap (large circular carved aragonite stone disks with pierced centres brought from the island of Palau some 400 kilometers away)between the end of the 17^th century and the 1970’s, shell bead necklaces (or wampum) in northeast America between the early 17^th century and mid 18^th century, or cocoa beans in Mesoamerica (covering the modern countries of Central America and Mexico), first used by the Mayans in the first millennium and still in use by the Aztecs in the 16th century. Money in the form of coins, still in use today, was probably invented in the 7^th century BC by the Greeks of Asia Minor (Byzantium).

Looking beyond this non-exhaustive list, all different forms of money shared the fact that they fulfilled, to a greater or lesser extent, the three major functions that the economics textbooks attribute to money:

• Store of value function. This implies that money will retain most of its purchasing power over time and that inflation (which results in erosion of monetary value) remains under control. The in-built control over the quantity of bitcoin issued (via algorithms) and, eventually, its upper limit, have not so far demonstrated their ability to stabilise its value (even in relative terms), as shown by the volatility of the price of bitcoin (see Chart 1). ‘Stablecoins’ will no doubt have little difficulty in demonstrating their superiority on this criterion, which will allow them to stake a more legitimate claim to be a ‘store of value’.
• Means of exchange function. A currency must be recognised and accepted as a means of payment. This acceptance can also be imposed by positive law. Thus, in France, the ‘legal tender’ nature of cash forbids shopkeepers from refusing payment in this form for sales below a certain amount[4] (article R. 642-3 of the criminal code).
• Unit of account function. Money must serve as a yardstick for the comparison of value of the objects being transacted for, which assumes that it is sufficiently widely used and its value is sufficiently stable for it to be a unit of measurement that is accepted by a large community if not universally. The extremely modest take-up of ‘cryptocurrencies’ and the small number of companies and merchants who accept them as a means of payment make it impossible to claim that they fulfil the unit of account function.

Hayek and ‘competing currencies’

It seems highly likely that the designer of one of the early ‘cryptocurrencies’ was aware of the work of the Austrian School. Nick Szabo (who some suspect of being the creator of bitcoin, known only by the pseudonym Satoshi Nakamoto) invented a decentralised, digital currency, ‘Bit Gold’, as a theoretical exercise. He has referred to the work of Carl Menger, the economist and founder of the Austrian School.

A number of cryptography professionals freely refer to the 1976 work The Denationalization of Money[5] by Friedrich von Hayek (Nobel Prize, 1974) as a theoretical basis for their innovations. Hayek argues for the creation of deregulated monetary conditions, under which private issuers would issue competing currencies.

According to Hayek, such a solution would help protect economies against inflation and monetary erosion, which he identified as the root of the problems of modern societies. Issuers, who would need to ensure that their currency attracted users, would be encouraged to protect its value and restrict its issuance. Hayek stressed that his proposal was not incompatible with Gresham’s Law, summarised pithily by W.S. Jevons as “Bad money drives out good”. This economic law applies in circumstances where there are two distinct currencies in circulation. An increase in the value of the underlying precious metal will lead to the good money (whose metal value is higher than its nominal value) being hoarded or used for other purposes, whilst the circulation of the bad money (with a lower metal value) is encouraged. “What Jevons, as so many others, seems to have overlooked, or regarded as irrelevant,” wrote Hayek, “is that Gresham’s Law will apply only to different kinds of money between which a fixed rate of exchange is enforced by law.”

And yet, with all due respect to the cypherpunk[6]community, it is hard to consider Hayek’s work as providing a theoretical base for ‘cryptocurrencies’. First, and in the most general terms, Hayek did not share their libertarian worldview – far from it. Indeed, he recognised the legitimate role of government in many areas (other than money): social protection, education and the support of certain business activities[7]. Meanwhile, the Austrian School was the source of a number of sometimes contradictory opinions on the subject of money. Hayek’s proposals, for example, suffered a chilly reception, to the extent of being considered naive, even within the Austrian School. Some economists have suggested even more radical changes, which perhaps provide an interesting theoretical framework when considering the future of ‘cryptocurrencies’. According to Murray Rothbard and Hans Hoppe, the most important function of money is as a medium of exchange, and it is only natural that economic actors will spontaneously choose the currencies that they believe will be used by other economic actors[8]. They therefore viewed Hayek’s proposal as transitional in nature. These authors believed that were these proposals to be put in place, they would lead to a trend of unification towards a single global currency: gold.

Impacts on monetary policy

Notwithstanding their highly speculative nature and the substantial risk to which investors are exposed, the ECB does not believe that crypto-assets pose any threat to the financial stability of the euro zone[9]. Their relative value remains modest compared to standard economic aggregates, and the cumulative indirect exposure of financial institutions, particularly banks, to these instruments is vanishingly small (EUR 20,000 in the third quarter of 2018), with ownership reserved almost exclusively to individuals (for a total of just over one billion euros). The IMF also judges that the development of ‘cryptocurrencies’ and the exposure of economic actors to them remain modest given the absence of any impact on financial stability or monetary policy[10].

However, none of this means that wider use of these new instruments will not have effects in the longer term.

Even a perfect mastery of the technology of crypto-assets does not give a full understanding of their economic and social function. In the current context of monetary creation and fractional-reserve banking, the quantity of money is influenced rather than fully controlled by the central bank (despite the powerful tools available to the latter). Moreover, the supply of money, its velocity of circulation and levels of output inter-react, such that perfect price stability is something of a pipe dream (hence central banks giving themselves a safety margin in their inflation targets).

Even if we suppose that use of crypto-assets will become more widespread, it is hard to see this happening with issuance rules set by ‘protocols’ alone. According to Danielson (2019)[11], a crypto-asset whose protocol sets out a slow mining process, which converges to zero (such as that used by bitcoin), would sow the seeds of persistent deflation. If the growth, through mining, of the quantity of a crypto-asset in circulation is lower than economic growth over the long term, then unless there is a steady increase in the velocity of its circulation there will be a fall in prices, which in turn will depress activity. Growth in the supply of money (or a substitute) must at least match economic growth if deflation is to be prevented.

Over and above the rate of money creation in the medium term, there is also the question of its adjustment to circumstances. Within a distributed ledger, the rules for issuance of crypto-assets would, by their nature, be unable to reproduce the pragmatic approach taken by monetary authorities in response to exogenous shocks. This lack of flexibility can exacerbate the situation, as was the case in the aftermath of the 1929 crisis. Friedman and Schwartz (1963)[12] demonstrated how the economic crisis of the 1930s was preceded, in the USA, by a fall in broad measures of money supply (M2, M3), whilst the M0 and M1 measures continued to rise. The central bank would have provided insufficient liquidity to the banking system to tackle the fall in deposits, limiting its open market operations at the beginning of the crisis, in late 1929, and then again briefly in the summer of 1932. The resulting squeeze on the supply of bank lending would have amplified, in its turn, the economic slowdown. They argue that the Federal Reserve’s inability to respond effectively to this shock in demand for money was a powerful factor in aggravating the recession. More recent works support this analysis (ECB, 2004[13]).

However, this reasoning would only hold in circumstances where there was only a single crypto-asset. As Bofinger (2018)[14] highlights, whilst the total number of units issued by a private issuer may be capped in a bid to protect the value of its crypto-asset, the principle of free competition between private issuers does not limit the number of issuers. It follows that there is no limit in the total quantity of crypto-assets taken across all issuers. Over and above the risk of loss inherent in holding any given crypto-asset, the overall quantity of crypto-assets would quickly become uncontrollable. In such circumstances, the opposite problem – that of inflation – would be the threat.

This transposition of the quantitative theory of money to crypto-assets nevertheless relates to an imaginary scenario in which the use of ‘cryptocurrencies’ as a means of payment has expanded considerably. It is first necessary to consider the possibility of a long-lasting coexistence of several competing private currencies. This topic was widely debated as long ago as the late 1970s within the Austrian School (see above). In addition, there is the question of the effectiveness of monetary policy in a situation where an official currency exists alongside one or more crypto-assets. Benigno (2019)[15], amongst others, have shown, through an analysis of different models of the coexistence of an official currency and currencies issued by private issuers, that competitive ‘currencies’ could reduce the central bank’s ability to use the interest rate tool, and make it more difficult to achieve a balanced level of inflation. The entrance into the market of multiple private issuers, whose aim would be to maximise their profit, would in fact strip the central bank of any control over interest rates and the inflation rate, which would become dependent only on exogenous factors (time preference, market entrance and exit costs, etc.)

The specific issue of the possible impact of stablecoins (see below) on monetary policy is perhaps more acute, as these instruments clearly have the greater potential for growth. In the event that ‘deposits’ in the form of stablecoins earned interest (the Calibra Association has indicated that Libra deposits will not earn interest), the possible consequences for the transmission of monetary policy would depend on the level at which the interest rate was set (G7 Working Group on Stable Coins, 2019[16]).

Let us suppose that this rate reflects the yield on the basket of assets used to back the coin. If the assets in question are denominated only in domestic currency, there will be little or no effect on monetary policy. However, if the basket of assets includes assets in other currencies (as is the case with Libra), the link between the central bank policy interest rate and the stablecoin interest rate will be all the looser as the share of assets in the domestic currency falls, perhaps even to zero. Moreover, the rate of return on holdings of stablecoins would have an impact on the outstanding amount of deposits, and thus on deposit and lending rates in the economy as a whole. The G7 working party noted that this effect would be fairly similar to that currently seen in countries affected by high levels of dollarization, but that it would extend to countries with lower dollarization levels.

In addition, a substitution of bank deposits by stablecoins would increase the dependence of commercial banks on market resources. As the costs of such resources are more elastic than bank deposits to money market conditions, adjustments to monetary policy would certainly be accurately transmitted by the vector of bank lending, but this would play a smaller role. At the same time, and on a more structural level, coupled with greater volatility in client deposits, the increased dependence of banks on market resources could incite them either to cut lending volumes, or to increase the risk and extend the maturities of lending in response to the increase in the average cost of resources. The first response would affect the financing of the economy; the second, financial stability.

So far we have considered that the stablecoin’ is an alternative form of savings, but that banking and financial intermediation would continue in the domestic currency. Let us assume instead that financial intermediaries would emerge that would lend and borrow in ‘stablecoins’. This new form of intermediation would again weaken the transmission of monetary policy because the rate of return on holdings and the interest rate on such loans would be more clearly uncoupled from monetary policy.

Some orders of magnitude

Depending on the source used, estimates of the number of crypto-assets vary between 1,600 and over 3,000. Their total capitalisation saw exponential growth in 2017, taking it to more than USD 800 billion at the beginning of 2018[17] (Chart 3). By 26 January 2019, it had fallen back to USD 237.5 billion. At this point, bitcoin accounted for USD 156.3 billion, or 66%, of the total. However, its share of total capitalisation is itself highly volatile (Chart 4). It peaked at over 80% in early 2017, before the development of rival crypto-assets later that year, and then fell to less than 40% of the total when the value of bitcoin collapsed in early 2018. Since then it has seen a recovery marked by significant fluctuations, oscillating around 65% in January and February 2020. Since 2016, the top six crypto-assets have accounted for between 70% and 100% of total capitalisation. Their share is, however, trending downwards and fluctuating with increasing amplitude as competing crypto-assets take off.

Crypto-assets are in no sense currencies. However, given their shared ambition to fulfil the functions of a currency, as reflected in their inaccurate designation as ‘cryptocurrencies’, there is a significant temptation to compare their capitalisation with real currencies.

We calculate that in December 2018, the aggregate money supply (in the broadest sense of the term) of the OECD nations and China was more than USD 88,000 billion.

Even narrow measures of money aggregrates, consisting solely of assets available immediately to settle transactions or extinguish debts (sight deposits and fiduciary money) have a value of a completely different order of magnitude than that of crypto-assets. By way of illustration, on 31 January 2020, M1 money supply in the euro zone was EUR 8,975.5 billion whilst in the USA it was USD 3,968.6 billion[18]. Despite the significant (if uneven) growth in their capitalisation since 2016, crypto-assets have gone only a tiny fraction of the journey they will need to take to rival official currencies. Perhaps the realisation of the Libra project, due to its scale and the fact that it is a stablecoin, will be a more decisive step along this path. But this project is experiencing some vicissitudes.

What about central bank digital currencies?

The arguments

Against the background of the emergence of numerous crypto-assets and falling demand for fiat money (which is part of base money), several issuing institutions have begun to consider making central bank deposits available to non-banking actors.

In his comments of 4 December 2019[19], the Governor of the Banque de France identified three goals for the possible creation of a central bank digital currency (CBDC). The first is the preservation of the link between citizens and the official currency, something made necessary in societies, such as Sweden, where the use of cash is in decline. The second is a reduction in intermediation costs in the central currency. The third and “most important” purpose lies in “the confirmation of the sovereignty [of the political authorities] faced with private initiatives such as Libra.”

Other arguments have also been put forward. For Dyson and Hogson (2016)[20] and Rogoff (2017)[21], the substitution of a CBDC for cash would allow the removal of the ‘zero lower bound’ which limits the effectiveness of negative interest rate policies that these authors expect to persist.

The risks

A central bank digital currency should not be thought of only as an official alternative to private-issuer ‘cryptocurrencies’; the reality would be much more far-reaching. By making central bank money available to non-bank actors (non-bank financial intermediaries for so-called ‘wholesale’ CBDCs, or even individual consumers or companies for ‘retail’ CBDCs) in a form other than cash alone, the central bank digital currency would become an alternative to cash. At the same time, if it is held in the form of accounts in a centralised register, it would also become an alternative to script money held as deposits with commercial banks (broad money).

In order to assess the issues related to such a proposal, one needs to draw a distinction between two types of money:

- Central bank money (high-powered money), consisting of commercial bank deposits at the central bank and fiat money;

- broad money supply, consisting of the part of central bank money in the form of banknotes and coins and to a much larger extent the money created in script form by credit institutions (the amount recognised in ledgers of bank deposits). If we exclude non-conventional monetary policy (quantitative easing) or, to a lesser extent, open market operations, all creation of money in its broad sense (M3 in the euro zone) has as its counterparty the simultaneous creation of a debt: the bank pays out the loan by crediting the account of the borrower. The broad money thus created allows to pay for the purpose of the loan and then circulates in the economy.

Sovereign money proposals, such as the Sigurjonsson parliamentary report in Iceland in March 2015 and the Swiss Vollgeld proposals of December 2015, sought to strip commercial banks of their ability to create money, reserving that right to the central bank alone. Far from being novel, these solutions dug up ideas that first saw the light of day in the thinking of Chicago School economists in the 1930s.

To achieve their main objective, supporters of these proposals favoured splitting client deposits currently recorded as liabilities by lending establishments into two separate types of account. “Transaction accounts,” which can be used to settle transactions and make transfers, would be recorded as liabilities on the central bank balance sheet whilst term deposits (Investment Accounts) would remain on the balance sheet of commercial banks, as they are at present. Money created by the central bank would exclusively be paid into transaction accounts, from which transfers could then be made to Investment Accounts.

The banks?–?which would only provide an interface between account holders and the central bank, and would have their role limited to that of a financial intermediary in payments, in the style of PayPal. However, they would retain Investment Accounts on their balance sheets. These would continue to be intermediated by the banking system, which would use these resources to make medium- and long-term loans.

Thus the creation of a ‘retail’ central bank digital currency displays significant similarities with these ‘narrow banking’ or ‘sovereign money’ concepts. Indeed, CBDCs are sometimes presented as a partial restriction on the banking system, albeit one that is less radical than the proposal that ‘narrow’ banks must back all customer deposits with their own deposits at the central bank (Gouveia et al. (2017)[22]).

The creation of a CBDC differs from a sovereign money solution in its intensity. So long as the proportion of clients who transform their bank deposits into CDBC remains limited, the money creation process will be unaffected, and loans will continue to ‘create deposits’. However, this supposes that customers are not encouraged to move their deposits into the central bank currency, which in turn would require that deposits with commercial banks earn an adequate rate of interest. An increase in the rate paid on deposits would result in higher average costs for bank resources. This would either put pressure on the supply of credit, or it would be passed through into lending rates, thus reducing demand. In any event, the new equilibrium would coincide in reduced money creation, which could complicate central bank efforts to tackle deflationary pressures. This risk of substitution would be exacerbated in the event of a “digital” run on the bank (conversion of customer bank deposits into the CBDC) and could heighten the threat to financial stability from a ‘normal’ bank run (flight of deposits of fiduciary money).

Some aspects remain unresolved

As CBDCs are still in the early stages of consideration, some of their core features have yet to be determined. In a document reporting the work of a Banque de France in-house working part[23] (which “expresses the views of the authors and not those of the Banque de France or Eurosystem”), the characteristics, benefits and risks of the two main categories of central bank digital currencies are discussed.

Wholesale CBDCs are defined as digital currencies accessible only to financial institutions, or perhaps only to some of them. Retail CBDCs are accessible to all. As identified by Bech and Garatt (2017)[24], the only distinguishing criterion between wholesale and retail CBDCs is accessibility (Chart 5). It is nonetheless important to analyse the very different issues raised by these two forms of CBDC.

Retail CBDC

The report’s authors believe that the model of accounts held directly with the central bank would be more beneficial for the CBDC. However, this model would carry the risk of disintermediation of banking system deposits (see below) and the authors highlight that a token-based model would make this retail CBDC a simple “virtual complement to cash”, which would be more in keeping with its philosophy.

Irrespective of the means of circulation (transfers from account to account or tokens), the report’s authors believe that the distribution of a CBDC could take place via intermediaries. Even in the event of a CBDC in the form of tokens, which would not, strictly speaking, involve CBDC deposits substituting for bank deposits, holdings of the latter for the Keynesian transaction motive would necessarily diminish. The effect on bank deposits would therefore be significantly less pronounced than for an account-based CBDC but would be far from neutral. Lastly, there is the question of interest. Some authors have argued for the benefits of a society in which cash become scarcer (Rogoff (2017)[25]), highlighting the more effective transmission of negative interest rates and the improved transmission of monetary policy that would result. But we believe that such an approach is incompatible with the spirit of a retail CBDC, which, in its roles as a digital equivalent to fiat money, would not by any logic earn interest.

Wholesale CBDC

The main social advantage expected from a wholesale CBDC lies in the benefits of a blockchain-type technology (traceability if desired, cost, speed). Lending establishments already have access to central bank money in electronic form (reserves).

A wholesale CBDC also raises the issue of scope (whether it would be accessible to banks only or also to other non-financial institutions). Historically, access to central bank money has been limited to registered deposit-taking establishments. In return they are obliged to retain reserves in central bank money as a certain proportion of their short-term client deposits (in the euro zone the required reserves ratio has, since 2012, been 1% of the deposits which form the base). This constraint saw a de facto tightening under Basel III and the introduction of the Liquidity Coverage Ratio, or LCR. This makes central bank money held by lending establishments a coercive tool for the transmission of monetary policy.

Some new actors, for example Fintech companies, seeking to invest in the payments market or take advantage of the opportunities created by the blockchain could have their activities facilitated by access to the central bank balance sheet. These conditions would require in-depth consideration given the redefinition of payment circuits and the modification of the respective roles of banks and payment intermediaries that could result (settlement in central bank money). The possible risks that these changes would create in terms of the operation of payment infrastructure, the transmission of monetary policy and financial stability would have to be taken into consideration.

***

The extent to which crypto-assets fulfil the functions of a real currency leaves considerable scope for improvement. They should not therefore be viewed as such. One desirable development would be to legislate to prevent issuing networks from promoting such assets to users as a ‘currency’. This would mean that users would still be free to use the assets as a means of exchange, but would be less likely to misunderstand their true nature. From this point of view, stablecoins, which are backed by baskets of stable assets, clearly offer greater certainty as to their value. Their design makes them structurally dependent on official currencies, in a similar fashion to currencies linked to the dollar under the Bretton Woods agreements (1944-1971), distancing them from the libertarian approach that motivated the first generation of crypto-assets. The Libra stablecoin initiative and the perception of a possible threat to monetary sovereignty have accelerated the consideration of the creation of central bank digital currencies. However, these would not just be official alternatives to privately-issued ‘cryptocurrencies’. Some of their features, notably those giving access to individuals (retail CBDC) in the form of accounts, or opening access to non-bank intermediaries (wholesale CBDC), would bring structural changes in the operation of the banking system and have structural consequences for the process of money creation (in the real rather than virtual sense) and the vectors of transmission of monetary policy. In particular, a retail CBDC would introduce the risk of a ‘digital bank run’, which would have deleterious effects on financial stability. For these reasons alone further reflection is called for, and central banks would do well to “hurry up slowly”. Nor should the process of reflection concentrate solely on central bank digital currencies. It would benefit from being extended to the digital forms of script money issued by lending establishments that could provide users with the same services as a CBDC but without the same disadvantages.

[1] Nakamoto S. (2008), Bitcoin: A Peer-to-Peer Electronic Cash System, November 1^st.

[2] The liquidity of coins should nevertheless be seen in context, as demonstrated by the significant volatility in the price of Bitcoin.

[3] PACTE – Action Plan for Business Growth and Transformation

[4] They must, however, refuse cash payments of over EUR 1,000 where the customer has their tax residence in France or is making the purchase for professional or business purposes; this limit is EUR 15,000 when the customer can prove they are not tax resident in France and is not making the purchase for business purposes.

[5] Hayek F. (1976), The Denationalization of Money, Institute for Economic Affairs, London

[6] A crypto-anarchist or libertarian capitalist movement of the 1980s in the USA. The originators of the first crypto-currencies promoted freedom of expression, free trade and privacy (enabled by cryptography) as means of overturning the social model based on a system of centralised power.

[7] Hayek F. (1960), The Constitution of Liberty

[8] Hoppe H.-H. How Is Fiat Money Possible?--or, the Devolution of Money and Credit, The Review of Austrian Economics, 7, (2), 49–74, 1994. Hoppe quotes Ludwig Von Mises p. 51 “(…) there would be an inevitable tendency for the less marketable of the series of goods used as media of exchange to be one by one rejected until at last only a single commodity remained, which was universally employed as a medium of exchange; in a word, money”

[9] European Central Bank, Crypto-Assets: Implications for financial stability, monetary policy, and payments and market infrastructures, Occasional Paper Series, ECB Crypto-Assets Task Force, n° 223, May 2019

[10] Franks J., Crypto-currencies and monetary policies, International Monetary Fund, Europe Office, 22 January 2019

[11] Danielson, Cryptocurrencies: Policy, economics and fairness, London School of Economics, July 2019

[12] Friedman M. and Schwartz A., A monetary history of the United States 1867-1960. Princeton University Press, 1963

[13] Christiano L., Motto R., Rostagno M., The Great Depression and the Friedman-Schwartz Hypothesis, ECB Working Paper 326, March 2004

[14] Bofinger, Digitalisation of money and the future of monetary policy, VOX EU, CEPR Policy Portal, 12 June 2018

[15] Benigno P., Monetary policy in a World of crypto-currencies, EIEF working Paper 19/05, April 2019

[16] G7 Working Group on Stable Coins, Investigating the Global Impact of Global Stable Coins, G7, IMF, BIS, October 2019

[17] Source: Coinmarketcap

[18] 27 January 2020, Federal Reserve

[19] Speech by François Villeroy de Galhau, A central bank digital currency and innovative payment solutions, Paris, 4 December 2019

[20] Dyson B. and Hodgson G., Digital cash : why central banks should start issuing electronic money, Positive Money, 2016

[21] Rogoff, Dealing with Monetary Paralysis at the Zero Bound, Journal of Economic Perspective, September 2017

[22] Gouveia, Olga Cerquiera et al., Central Bank Digital Currencies: assessing implementation possibilities and impacts, BBVA, Working Paper n° 17/04, March 2017

[23] Internal working group at Banque de France, La monnaie Digitale de Banque centrale, 8 January 2020

[24] Bech M.L., Garatt R., Central bank cryptocurrencies, BIS Quarterly Review, 17 September 2017

[25] Rogoff, Dealing with Monetary Paralysis at the Zero Bound, Journal of Economic Perspective, September 2017