Note: This post has been revised on the basis of Nick Rowe’s excellent comments.
- Sovereign control of the unit of account
- Aside: the metallist approach to money
- Cartalism: government puts money into circulation by accepting it in payment of taxes
- Cash settlement of debt
- Sovereign money gives value its objective existence
A basic premise of monetarism is that money defines a relationship between the government and the market and that nothing essential about the economy is lost by assuming that the banking system’s monetary role is wholly determined by the government. That is, monetarist models allow for banks, but then limit their function to that of “multiplying” deterministically the government supply of money. Because the role played by banks in providing liquidity to the economy is not modeled, monetarism effectively assumes away any meaningful role for the banking system. On the basis of this framework, monetarism proposes that the government by controlling the money supply can promote optimal economic activity in the market. In short, while there is long tradition in economics of arguing that the banking system exists to create liquidity (Thornton 1802; Hawtrey 1919: 9-16), monetarism explicitly rejects this view. Indeed, rejection of banking theory was a primary motivation driving the development of monetarism (Mints 1945, Friedman & Schwartz 1963: 169, 192, 253, 266). By abstracting from bank-provided liquidity, the monetarist framework implicitly posits that there are only two fundamental sources of liquidity in an economy: liquidity is either generated by markets spontaneously when buyers and sellers get together or liquidity is generated by government guarantees of payment.
Unfortunately the other major macroeconomic school that developed over the course of the 20th century, Keynesianism (particularly as espoused by New Keynesians, but also by many Post-Keynesians) has embraced monetarism’s rejection of the traditional view that banking plays an essential role in generating liquidity within an economy (see e.g. Tobin 1963). While Keynesians place more emphasis on the role played by government in the economy and less on markets, they are generally in agreement with the monetarists that the fundamental sources of liquidity in an economy are either markets or the government. In short, the more nuanced banking school view that liquidity is generated by the structured interaction of banks, markets, and government has been almost entirely lost in the modern macro-economic discourse.
Because the agenda of limiting the view of money’s role in the economy to that of defining a relationship between markets and the government can be attributed clearly to Milton Friedman, and cannot be attributed to the more diverse writings of J.M. Keynes, I consider this limited vision to be part of the “monetarist” agenda, and describe modern Keynesians as having bought into the monetarist agenda when they embrace this approach. Here I discuss how a variety of strains of heterodox monetary analysis embrace this basic premise of monetarism, even as they reject the specific policy proposals associated with monetarism. Interspersed throughout this discussion is an explanation of banking theory’s approach to the same issues.
I will address in particular four themes in the heterodox literature on money that emphasize the role of government in the money supply and are revisited by many different authors:
- The government establishes the unit of account and retains the sovereign right to revalue it.
- The government puts money into circulation by accepting in payment of taxes a token that it also spends. For cartalists this is the essential means by which money is created.
- When credit is used to trade it must be settled in cash.
- Markets and prices are derived from government forms of money.
That the government establishes the unit of account and retains the sovereign right to revalue it is a claim associated with Ingham (2006: 265, 271). It is entirely consistent with the monetarist framework, although heterodox theorists correctly observe that its importance is rarely acknowledged by mainstream monetary theorists. Of the four claims, this has the strongest support historically. For example, the U.S. Constitution arrogates to the government the power to “regulate the value” of the coinage. And medieval and early modern history has abundant examples of sovereigns that devalue or revalue the unit of account. Furthermore, modern examples of countries of which we can state clearly that this claim does not apply are few and generally consist only of countries that have so abused their authority over the currency via inflation that the government itself has lost a great deal of legitimacy and as a result the populace has chosen to denominate transactions in a more stable foreign unit of account.
On the other hand, I have argued elsewhere that an essential aspect of the transition to a modern economy involved the creation of a unit of account based on a convertible paper money supply that was managed by private sector banks. Given modern bank-based monetary systems, the authority exercised by the sovereign over the nominal value of the unit of account is much weaker than in those days when it could simply “cry up” the value of a coin with a fixed weight in gold. And, subsequent to the 2008 crisis with its origins in excessive money issue by shadow banks, the measure of control exercised by the government over the value of the unit of account is surely a little less clear than it was in the past.
Such nuances are in fact encompassed in the broader picture of Ingham’s Nature of Money. Ingham ties his analysis of the sovereign unit of account into an analysis of the role of money in markets and in capitalism itself. He argues that modern capitalism is founded on a “memorable alliance” between currency defined by the state and the issuance of credit-money by the banking sector. Indeed, Ingham acknowledges that networks of traders can form their own money of account (Ingham 2006: 271). The distinction in our approaches lies in the importance that we attribute to such private moneys of account. For Ingham they are fundamentally unstable, and not worthy of too much emphasis. By contrast, I argue that it was the integration in Britain of the sovereign unit of account with the specie-linked private money of account supported by a network of elite international merchants that laid the foundations for modern capitalism. Thus, in my view even at the level of the unit of account a private-public partnership played an essential role in the development of capitalism.
Most of the literature on money is not focused on its role as a unit of account, but as a transactional medium that circulates, serving both as a means of payment and a store of value. There are two basic approaches to the question: What makes it possible for something to enter into circulation as money? (Observe that this is not a question regarding how something, once it gets its start as money, can become a dominant form of money. Menger described what in modern terms are called the network effects that entrench the use of a given money after its initial entry into circulation.) The Metallist approach claims that money circulates because of its intrinsic value, and treats gold coin as a quintessential form of money. Menger clearly falls within the Metallist category as his focus is commodity money. Cartalists challenge this view, arguing that all that is necessary for money to enter into circulation is that an authority – for example a government or a major bank – treats the money as a liability and accepts it in payment for amounts due as taxes or on loans. In this section, I argue that Metallism, as a school of thought, originated alongside and in support of bank money, and thus that our understanding of Metallism should go well beyond the simplistic appearance that the approach presents. (Cartalism will be discussed in the following section.)
The idea that intrinsic value can serve as a foundation for the circulation of money proved through the European experience of money in the Middle Ages to be an extremely flawed approach to money. As the appellation Metallist implies almost all proponents of the intrinsic value approach view metals as having characteristics that make them the most suitable commodities to serve as money. In the Middle Ages it became clear that the problem with a metal coinage that is meant to circulate based on intrinsic value is that the process of circulation itself reduces the metal in the coins through both wear and tear, and deliberate clipping. Thus it became clear that the ideal form of Metallist money isn’t very good at holding its value. For this reason it is extremely costly to maintain a money supply based on fixed weight metal coins as the principal medium of exchange. The typical compromise that medieval sovereigns opted for was to slowly reduce the metal weight of their coins so that existing low-weight coins did not need to be collected and recoined and could continue to circulate on a par with new coins. In short, medieval experience demonstrated that as practical matter a metallist currency can only operate if inflation is built into the system.
On the other hand, the philosophy of metallism – that is, the claim that coins should have a fixed and stable value in terms of specie – was presented and argued most forcefully by John Locke at the end of the 17th century. While arguing against a reduction in the metal content of the coinage and in favor of the demonetization of below-weight coins he wrote: Metals have “as money no other value but as pledges to procure what one wants or desires, and they procur[e] what we want or desire only by their quantity” (Locke 1824: 22). “Men in their bargains contract not for denominations … but for the intrinsic value; which is the quantity of silver by public authority warranted to be in pieces of such denominations. … it is only the quantity of the silver in [the coin] that is, and will eternally be, the measure of its value.” (Locke 1824: 144-45, quoted in Desan 2014: 347).
I argue here that this claim that coins should have a fixed and stable metal content could only develop in an environment where coins were more of a reference point than an important medium of exchange. I have argued elsewhere that it was abstract “bank money” not coinage that had demonstrated the capacity to maintain a fixed relation to a specific weight of gold or silver: this was the advantage of being an abstract unit of account that was convertible into, but not tied to, the steadily deteriorating coinage. It is, therefore, no surprise that John Locke, a founding philosopher of metallism, was writing in the late 17th century when “bank money” was well-established in European money markets and indeed, when his political allies were advocating for the establishment of the Bank of England (Desan 2014: 345) in hopes of playing an important role in the European money market. In other words, when monetary scholars refer to coin as the means by which bank money was settled, they are getting the relationship backwards. Coin was able to serve as a stable reference point for the value of bank money precisely because the vast majority of transactions could take place using bank money alone and coin was used very rarely.
What the founders of the Bank of England were seeking was not a metallist money supply, but a unit of account that had a fixed value in terms of metal. (This is made clear in Paterson’s 1694 proposal for the Bank.) The elite bankers had been able to establish such a measure of value for their own transactions since the mid-16th century. By founding the Bank of England and simultaneously reforming the British coinage on metallist principles they sought a stable sovereign unit of account to which their bank money could be tied.
Thus, the philosophy of metallism, even as it espoused the view that money circulates because of its intrinsic value, was developed in order to support a system of paper-based bank money. Not only that, but the paper-based bank money in question was designed from the start to be backed by a combination of specie, sovereign debt, and private debt (in the form of commercial bills). That is, the Bank of England was founded as a partial reserve bank (in contrast to the Dutch Wisselbank), and it was understood from the start that the Bank of England’s circulating liabilities would not be fully backed by specie. On the other hand, the founders of the Bank were undoubtedly confident that the sovereign, private and metal assets of the Bank had an “intrinsic value” sufficient to support the Bank’s liabilities.
The metallists who succeeded Locke would argue that money circulates not because of its intrinsic value, but because it is backed by something with intrinsic value. The term “backed” itself already implies that there exists an entity with a balance sheet and the liability-side of the balance sheet is “backed” on the asset-side by specie or its equivalent.
In summary, Locke’s original metallist argument was actually an argument in favor of stabilizing the sovereign unit of account by tying it to specie. And even though this argument was carefully framed by Locke as a matter of “natural law,” the actual goal appears to have been to lay the foundations for a new type of monetary system, one where a bank credit-based money supply was anchored by a specie-based sovereign unit of account. Thus, the metallist argument morphed quickly into an argument in favor of bank money that is backed by specie or its equivalent and that can be redeemed in specie.
Cartalism, by contrast, explains how a “token” with no intrinsic value can serve as unit of account, means of payment, and store of value. All that is necessary for token-money to circulate is that an authority to whom money is owed stands ready to accept the token in payment. The feasibility of this means of putting money into circulation was demonstrated empirically by the many colonial powers that used tax policy to put their preferred money into circulation (h/t Nick). Note, however, that the authority which puts money into circulation can also be a bank that accepts the notes that it issues in payment of loans.
One can think of cartalism as a theory of synthetic value rather intrinsic value. Even though the tokens that circulate are nominally valueless, the authority makes them valuable by accepting them in payment of obligations. In the case of taxes the obligations are imposed by force (or social contract); in the case of (voluntary) loans the obligations are generated by the authority’s wealth or lending capacity. In both cases, the tokens have value — it is just value created by the promise to accept them in payment rather than intrinsic value.
In fact, Knapp’s original statement of the cartalist view is that it is necessary for the government to accept a token in order for it to circulate as money. This was, however, true as a matter of definition: Knapp defined “circulating as money” to require that all parties accept the token including the government. Knapp also explained that banks could put notes into circulation by the same mechanism which governments use to put tokens into circulation. Thus, “when a bank note … [can] be used unrestrictedly for making payments to the bank,” it becomes “a chartal means of payment – issued privately,” and the bank’s “customers and the bank form, so to speak, a private pay community.” These bank notes cannot, however, be “currency” until the State announces that it will accept them (Knapp 1924: 133-35). Thus, the original statement of cartalism is a “state theory of money,” only by definition. The analysis explicitly allows for a cartal means of payment to be private or public (see also Minsky 1985: 4).
On the other hand, because the state can demand payment of taxes from all citizens and residents, governments are understood to be particularly well-placed to put cartal means of payment into circulation. Thus, the “cartalist” theory of money is sometimes viewed simply as the claim that money has value because the government accepts it in payment of taxes. Like the role played by the sovereign in defining the unit of account, this too is entirely consistent with the monetarist framework, where formal models include a government budget constraint which assumes that the government is empowered to withdraw money through taxes or distribute it through cash grants. On the other hand, heterodox theorists are correct that mainstream monetary theorists rarely ask how money gets into circulation and thus rarely mention the cartalist approach.
The validity of the claim is somewhat more difficult to evaluate. It is certainly true that in many – or even most – environments that which is accepted as money is also accepted in payment of taxes, so there is abundant evidence that is consistent with the cartalist view. This view includes, however, a causal statement and, as always, correlation is much easier to demonstrate than causation.
The strongest version of the cartalist claim, that it is sufficient for the government to accept a token in payment of taxes in order for it to circulate as money, is easy to disprove, since a single example will do. In particular, when a government seeks to put something into circulation as money by accepting it in payment of taxes, but there is a better alternative form of money already available, the effort has in some cases failed. An example is the Exchequer bill in late 17th century England (discussed here).
Modern proponents of cartalism typically focus their attention on the role of the state and not on the way in which banks play a very similar role in putting money into circulation (Bell-Kelton 2001, Desan 2014, Glasner 2017). Thus, the weakness of the modern cartalist approach lies in the way that the modern literature edits out the earlier, more sophisticated, approach to the banking system.
Indeed, one can interpret banking theory as a means of uniting the metallist and the cartalist approaches to money. Metallists claim that to circulate bank money must be backed by “intrinsic value,” and banking theorists would reply that good bills, denominated in a stable unit of account, have intrinsic value. Thus, through banking theory metallism is converted into two requirements: (i) that the unit of account be anchored, for example, to gold or by a central bank, and (ii) that the origination practices of the banks must produce assets that have almost no risk of loss. (I refer those who reflexively assume that private sector assets are risky and never as safe as government debt to Sissoko 2016.) Cartalists claim that to circulate money must be accepted by an authority in payment of obligations to the authority, and banking theorists explain that this is precisely what banks do: they accept bank liabilities in payment of the good bills that back their balance sheets. For banking theorists, bank money circulates both because it is a cartal means of payment and because it is backed by assets that are of indisputable quality and denominated in a stable unit of account.
Another common heterodox claim is that credit-based forms of money must be settled in cash. Indeed, this is the premise underlying the “hierarchy of money” framework. Bell-Kelton (2001: 160) argues that “bank money is converted into bank reserves so that (ultimately)” payment is in government liabilities. Similarly, the tiers of the hierarchy as Mehrling (2012) describes it are based on settlement: securities are at the bottom and settled in bank money, bank money is settled in currency (including central bank reserves), which in turn must be settled in some form of international means of payment (e.g. gold or SDRs).
Mehrling observes that in the design of his hierarchy he is building on the distinction traditionally made by economists between money and credit. He fails to note, however, that the importance of drawing this distinction is a fundamental monetarist principle (Friedman and Schwartz 1963; Meltzer 2003: 27). Thus, when Mehrling states that he seeks to avoid “sterile debates about what is money and what is credit” and instead to focus “on the point that the system is hierarchical in character” (p. 4), he is indicating that his goal is simply to refine the monetarist approach to money. This is made especially clear when he explains that the “money from the level above serves as a disciplinary constraint that prevents expansion” (p. 8). This echoes the claim in Friedman and Schwartz that the quantity of deposits is determined by the underlying supply of reserves.
Banking theorists, by contrast, view bank money – backed by high-quality private sector assets – as the form of money that sits at the top of any hierarchy. Gold or central bank reserves play an important role in anchoring the value of the unit of account and thus in defining the environment in which bankers function, but are ultimately less important to determination of the quantity of money available than the quality of the debt origination practices followed by the banks when issuing bank money. (Indeed, it was only the robustness of these practices that constrained the British money supply – and inflation – when the original debate between the Currency and Banking Schools took place, since money was not convertible into gold and Bank Rate was set at its legal maximum from 1797 to 1822. As Nick observes, these practices were almost certainly influenced by the expectation that there would be a return to the old gold standard.)
Banking theorists focus on the fact that bank money transactions are settled by clearing. While it is, of course, true that some balances remain after clearing, the bank clearing system provides for the balances to be settled by borrowing. As a result of the structure of the clearing process, reserves (or gold) play a minimal role in the settlement process. (Note that I am not claiming that they play no role in the settlement process, just that it is remarkably small.) Their principal function is instead to anchor the value of the unit of account.
This function of anchoring the value of the unit of account is integral to the operation of the banking system. Without such an anchor it is far from clear that the decentralized credit-based structure of the banking system could over the long-term avoid generating so much price inflation that the banking system itself would be destabilized. Thus, from the perspective of banking theory liquidity crises are best viewed as part of the process by which the anchor constrains the credit growth generated by the banking system. As I argue below this constraint is different in nature from Mehrling’s “disciplinary constraint that prevents expansion.” To the degree that the anchor “prevents expansion,” it does so through its effect on each bank’s decision-making process with respect to loans, not through quantitative constraints. And the anchor’s effect in a liquidity crisis is not to “prevent expansion,” but on the contrary to prevent a sudden contraction of the money supply.
In a liquidity crisis, either the public that usually holds bank money or the banks that lend on a regular basis to each other are concerned about a bank default that could reduce the value of their bank-issued assets and leave them holding losses. From a metallist viewpoint, the intrinsic value of the assets backing the money supply is in doubt. From a cartalist viewpoint, the public fears that the authorities will demonetize some tokens. Such a crisis has the potential to throw into doubt the foundations upon which the monetary system is built. In this situation the authorities must draw a bright line distinguishing any bank money that will be demonetized from bank money that will not – or equivalently distinguishing those banks whose assets fail the intrinsic value test from those with adequate assets. Because of the doubts that have been generated by the (fear of) default, proof of the authorities’ support of a bank will generally be required and the classic means of indicating such approval is by providing the approved banks with a temporary but abundant supply of reserves in exchange for their assets. In a genuine liquidity crisis, no bank fails and the declaration that support is available for all the banks is sufficient to end the crisis. When a liquidity crisis is set off by an actual bank insolvency, the process can take longer. Because periodic bank failure is unavoidable in a decentralized banking system and indeed plays a part in constraining the growth of bank credit, this process of supporting the banking system through a liquidity crisis is integral to anchoring the value of the unit of account.
In short, banking theory indicates that reserves are not a “higher” form of money than bank money. Instead they are an integral part of the system of bank money. From the perspective of banking theory the “hierarchy of money” misrepresents the role of reserves or gold in the monetary system by implying that the value of this “higher” form of money exists independent of the system of bank money. This latter claim is precisely the claim made by the monetarists when they marginalized the role of the banking system in their models.
Ingham presents a very subtle heterodox point. Capitalist markets can only exist if (i) there is coordination on a unit of account and (ii) that unit is sufficiently stable that market participants have reasonable predictability of future values for purposes of business planning and entering into debt contracts. Given such a money of account which in general must be imposed by the state (Ingham 2013: 300), “money is the stable measure of value which makes it possible to establish the relative prices of all commodities” Ingham (2008: 68). The idea is further explained: “ ‘value only attains social existence by means of its monetary embodiment. It is money which makes value exist objectively for all’ (Orlean p. 52 this volume), which is accomplished by its capacity to gain general assent as the legitimate expression of value” (Ingham 2013: 313).
Let me restate the argument as I understand it. The communication that must take place in order for market transactions to be effected depends fundamentally on the existence of a language in which market participants can communicate. An essential element of this language is the unit of account. Furthermore, that unit must be stable enough for market participants to engage in the forward-looking analysis of benefits and costs that underlies rational price-based behavior. Only given these antecedents can market trade generate prices that both put a value on goods and have meaning for the traders. This is the sense in which “value attains social existence by means of its monetary embodiment.”
In this literature markets are sometimes defined with reference to the neo-classical model, that is, markets are viewed as venues where traders meet and generate a single price for each uniform good before trading (Ingham 2006: 260). Ingham’s argument then provides the antecedents that are necessary in order for “market” prices to be produced, where market prices refer to what is apparently a neo-classical price vector.
As was discussed in the first section, Ingham observes that such a unit of account is usually, but not always, provided by the state. I argued above that Ingham underestimates the role of bank money in determining and stabilizing the unit of account. I continue that argument here: banks play an essential role in “making it possible to establish the relative prices of all commodities.” Economic theory indicates that achieving this goal requires not only a stable unit of account, but also that there be very many participants on both sides of every market (Geanakoplos 1987). The crucial element that ensures the presence of many buyers is the absence of liquidity constraints, or in other words generalized access to credit lines supporting the payments system (Sissoko 2007). When such credit lines are not available, price variations will be common due to the ubiquitous liquidity constraints generated by the need for working capital in order to engage in production.
In short, the bargaining problem that Ingham associates with barter is equally present in any environment where traders enter the market subject to liquidity constraints that force them to sell – their labor if they have nothing else – before they can make purchases. The neo-classical economic model with its simple price vector assumes that traders who enter the market with knowledge of how to produce can seamlessly hire workers and rent capital to take advantage of that knowledge and create goods for sale in the market to those workers and capitalists. Implicit in the seamlessness with which the production process takes place in this model is easy access to the short-term credit needed to finance the wage and rent bills. This is, of course, precisely the role played by banks in the early years of their existence.
Before capitalist bank money, the sovereign provided a unit of account, but its value was limited because liquidity constraints frequently determined prices. After capitalist bank money – at least for those with the privilege of access to bank money – short-run liquidity constraints were largely eliminated, so that prices became more neo-classical in character. At this time it made sense to generate and publish price lists of commodities, because many “market” prices were no longer determined by a process of bilateral bargaining. The key innovation making such “value” possible was the banking system and the credit that it offered to merchants and tradesmen.
In short, Ingham should revise his list of the antecedents necessary in order for “market” prices to be produced. The relative prices imagined by the neo-classical market are founded not only on a stable unit of account, but also on a banking system that provides payments system credit to finance working capital. Given that historically such a banking system was successfully established only in a “memorable alliance” with the state, the banking system should be viewed as a complement to and not a substitute for the role of the state in stabilizing the unit of account.
Heterodox monetary theory is often influenced by the basic monetarist framework which envisions markets and the state as the only sources of liquidity and deliberately denies the importance of the banking system. By explaining how banking theory relates to several heterodox approaches I demonstrate both how heterodox theory allows itself to be constrained by mainstream theory and how its horizons can be expanded by combining it with banking theory. Although Ingham recognizes the essential role played by banks in the monetary system and in capitalism itself, he underestimates the role played by banks in determining the unit of account and in the process of price formation. Similarly, banking, properly understood, turns the hierarchy of money on its head, as it is the system of bank money that gives central bank reserves their value.
Banking theory explains what makes price-stable liquidity possible (cf. Holmstrom 2015). The banking system offers a fixed rate of exchange between bank money and the money of account while stability of the money of account is managed by the central bank. On this foundation banks expand the money supply, extending payments credit and ensuring that markets are not beset by episodic local liquidity events. This has the effect of stabilizing the price structure of these markets. In short, banks provide the liquidity that makes it possible for markets to approach the neo-classical ideal. Because of the essential role played by banks in the economy, the most important factor in financial stability is the “intrinsic value” of bank assets or, in other words, the origination practices of the banks.
 In the terminology of modern economic analysis it is network effects that support bank money as a cartal means of payment. For example, David Glasner reframes Knapp and Minsky’s basic intuition using network effects and concludes with “the following preliminary conjecture: the probability that a fiat currency that is not acceptable for discharging tax liabilities could retain a positive value would depend on two factors: a) the strength of network effects, and b) the proportion of users of the existing medium of exchange that have occasion to use an alternative medium of exchange in carrying out their routine transactions.” In short, modern analysis supports the view that the early cartalists were correct: the state’s power to tax may be an important means of getting money into circulation, but it is far from the only means by which a token money can be put into circulation.
 “Monetary policy ought to be concerned with the quantity of money and not with the credit market. The confusion between ‘money’ and ‘credit’ has a long history and has been a major source of difficulty in monetary management.” Milton Friedman, 1964 Congressional Testimony (discussion with Congressman Vanik) p. 1151 cited in Hetzel 2007.
 Inappositely, Mehrling cites Hawtrey’s Currency and Credit, which takes a banking theory approach, on the traditional distinction between money and credit. In fact, Hawtrey (1919: 377) “treat[s] credit as the primary means of payment and money as subsidiary.” Indeed, Hawtrey disputes the relevance of settlement itself: “Purchasing power is created and extinguished in the form of credit” (380). For Hawtrey “Credit possesses value, and it is more correct to say that the value of gold is due to its convertibility into credit than that the value of credit is due to its convertibility into gold” (371). In short, the thesis of Currency and Credit is that Mehrling’s hierarchy – from bank liabilities on up – is false, because bank liabilities are the ultimate form of money and it is bank liabilities that give value to currency, central bank reserves, and even gold.