This archived content originally appeared at, the predecessor site to, and does not carry the sponsorship of the Cato Institute.

Drawing the line

Where in practice does one draw the line between money and other assets? Here, even though I prefer to reserve the term “money” for the monetary base, I will bend to popular usage and use “money” to refer to “money in the broad sense,” which consists mainly of credit.

The economists most concerned with where to draw the line have been, naturally, the monetarists. They were interested both in what assets people use most commonly in payment and what definition of money has changes that correspond well statistically with changes in prices and output. The monetarists settled on M2 as the most appropriate definition in the United States and most other countries. In the United States, M2 comprises

  • M1, consisting of currency held by the public + traveler’s checks of nonbank institutions + demand deposits + other checkable deposits, such as NOW accounts, and
  • these additional components: savings deposits, time deposits less than $100,000, and balances in retail money market funds.

The monetarists’ confidence in M2 was such that the Federal Reserve Bank of St. Louis, a monetarist stronghold that tracked many definitions of money, stopped calculating M3 in 2006, although it has recently restarted calculating a version of it, as I will explain below.

The Austrians did not think about the question as a practical matter until the monetarists had already worked on it for some years. One reason was that at the time, the number of active Austrians could be counted on the fingers of one hand. Murray Rothbard made an attempt in America’s Great Depression, published in 1963, the same year as Milton Friedman and Anna Schwartz’s Monetary History of the United States, 1867-1960.  (See pages 87-91 of the hyperlinked edition. Rothbard signaled his hostility to the monetarists by citing them little.)

Although some other Austrians have followed Rothbard’s lead, overall the Austrians have shied away from strongly advocating any particular measure of the money supply, even with the caveats Rothbard had in applying it.

Another strand of Austrian thought offers a different approach. Rather than establishing a sharp dividing line between money and other assets, it  recognizes a spectrum of “moneyness.” W.H. Hutt was to my knowledge the first Austrian, or Austrian fellow traveler, to tease out the implications of the idea in his 1956 essay “The Yield from Money Held.” Hutt cited as his partial inspiration an obscure Dutchman, Tjardus Greidanus, whose book The Value of Money (1932, 1950), remains a neglected gem. Hutt in turn influenced George Selgin, through whom I and others got led back to Greidanus.

It has been quite a while since I read the second edition of Greidanus’s book, and I do not have it handy, but judging from the first edition, Greidanus was unaware of the work of a contemporary, the French economist François Divisia, who devised a basis for measuring the money supply that corresponded to Greidanus’s ideas. Divisia likewise seems to have been unaware of Greidanus.

Divisia monetary aggregates do not just add up components of the money supply; they give different components different weights based on their degree of “moneyness.” (My phrasing here is quite imprecise, but good enough for a blog post.) The leader in measuring Divisia aggregates is William Barnett, formerly of the Federal Reserve Bank of St. Louis and the Federal Reserve Board of Governors, now at the University of Oklahoma. Barnett is the author of a recent book on Divisia aggregates and a monthly calculation of them for the United States.

After Barnett began publishing his calculations, the Federal Reserve Bank of St. Louis revived its calculations of Divisia M3, although not of plain M3.

Long-time monetarists continue to consider plain M2 the most useful monetary aggregate for practical purposes of assessing monetary policy. The Austrians, though, would do well to look at the Divisia aggregates, which are more consistent with characteristics of goods that Austrians are wont to stress in other contexts. Many types of financial assets are substitutable for one another, but usually they are not perfectly substitutable, and the degree of substitutability may vary over time, for instance becoming less during recessions, when appetite for risk is lower.

  • Nicolas Cachanosky

    Very interesting post. It made me remember the following passage, which I thought could be of interest given the similar point of view to that of the post:

    “It also means that, although we usually assume there is a sharp line of distinction between what is money and what is not-and the law generally tries to make such a distinction-so far as the causal effects of monetary events are concerned, there is no such clear difference. What we find is rather a continuum in which objects of various degrees of liquidity, or with values which can fluctuate independently of each other, shade into each other in the degree to which they function as money.

    I have always found it useful to explain to students that if has been rather a misfortune that we describe money by a noun, and that it would be more helpful for the explanation of monetary phenomena if ‘money’ were an adjective describing a property which different things could possess to varying degrees. ‘Currency’ is, for this reason, more appropriate, since objects can ‘have currency’ to varying degrees and through different regions or sectors of the population.”

    Hayek, F. (1976). Denationalisation of Money. (2007 ed.) IEA. p. 55

  • Mike Sproul

    First, there’s gold. Then there are paper claims to gold, then there are bookkeeping entries that are claims to those paper claims, and so on. All those claims have varying degrees of liquidity, so we draw arbitrary lines where nature did not put any lines, and start talking about M1, M2, etc. The whole exercise is a classic case of asking the wrong questions.

    Here’s the right question: Whose liability are those claims? Paper dollars issued by the Fed are the Fed’s liability, and their value is determined by the Fed’s assets. Checking account dollars issued by Chase bank are Chase’s liability, and their value is determined by Chase’s assets. Credit card dollars issues by Visa are Visa’s liability, and their value is determined by Visa’s assets. Those various kinds of dollars have varying degrees of liquidity, but when it comes to the question of what determines the value of those dollars, asset backing counts for almost everything, while liquidity counts for almost nothing. It is therefore a huge mistake to divide various kinds of money according to their liquidity, while paying no attention to asset backing.

    • Bill Stepp

      Credit card issuers such as Visa facilitate cash transactions by providing credit to payers. The underlying cash transaction is the exchange of dollars, but the payer is on the hook to pay Visa’s monthly bill in dollars (plus interest).

      Here is one economist’s take:


      Money is a financial asset that provides us with current or future purchasing power. Credit cards are not part of the money supply because they are just convenient ways to make a loan. You are actually borrowing money from the bank that issued the card, and payment is deferred until you make pmt. to the credit card issuer. Money is an asset that represents current or future purchasing power. Credit purchases do not technically represent purchasing power, they represent consumer credit, a way to defer final payment. Credit cards are not counted as part of the MS, but do affect the money supply, since having the convenience of plastic reduces our average cash balances, decreases our demand for money.

      As Kurt points out in an email:

      “…the assets Visa owns directly under its own corporate name are distinct from its billings to customers. In effect Visa’s major asset is a bunch of unsecured loans.”

      • Mike Sproul

        Back in 1840, when checking accounts were relatively new, people denied that checking account dollars were money, on the grounds that checking account dollars ultimately had to be paid in paper or coin. In 1710, people denied that the new-fangled paper notes were money, since paper notes were ultimately paid in coin. Today, people deny that credit card dollars are money, since they are ultimately paid with a check, a paper bill, or in coin. In every case, it has taken people awhile to realize that there is a permanent float of dollars (paper, checking account, or credit card) that is never paid down, and that permanent float should be counted as part of the money supply. Credit cards were introduced in 1950, so if past history is any guide, it will be about the year 2080 before economics textbooks count credit card dollars as part of the money supply. By then, of course, there will be some new kind of money, and economists will deny that it is money.

        The more important point is that it is irrelevant what we call money. The important thing is what backs the money. As I said, Federal Reserve Notes are backed by the Fed’s assets, Chase’s checking accounts by Chase’s assets, and Visa dollars by Visa assets. In every case, the value of the various kinds of dollars is determined by the assets and liabilities of the issuer.

        • Bill Stepp

          Credit cards fit none of the components of the generally accepted definition of money.
          A cc is not a generally accepted form of payment; your local street hot dog vendor won’t accept one. A Federal Reserve note for one dollar (a k a Bernanke buck) is money.

          There is no such thing as a Visa dollar.

          • Mike Sproul

            One day, I write a check to my grocer that instructs my bank to transfer $1 out of my checking account (most of which is borrowed money) to the grocer’s checking account. Next day, I swipe a card that instructs Visa to transfer $1 (borrowed) from my Visa account to the grocer’s account. At the end of the month I’ll pay off those Visa dollars, but then I’ll immediately start charging up new Visa dollars, so on an average day of the month, there will be a permanent float of about $1000 Visa dollars on my account alone. Worldwide, the permanent float is in the trillions. Just like checking account dollars (and paper dollars) there is a permanent float of dollars that is officially counted if it was created by scribbling numbers on a piece of paper called a check, or printing them on another piece of paper called a bank note. But let those numbers be imprinted on a credit card slip, and officialdom will deny that they are money.
            My local hot dog vendor will indeed accept a Visa dollar. He has a card reader on his cell phone.

        • hamblin

          But credit card issuers cannot spend their cardholders’ debts to them, so how can those debts be money? For that matter, savings deposits cannot be spent either, so how can they be money?

        • harrydavid1

          Wouldn’t the credit card balances have to have the potential to increase prices, for them to have the properties such that it fits into the equation of exchange or monetary theory in general — as is the case with deposits and cash? Or do they have that potential? Or does it not matter?

  • The Arthurian

    I find useful the PDF by Joseph Salerno on TMS, the “True Money Supply”:

    Also, your post reminds me of a footnote from Keynes: “we can draw the line between “money” and “debts” at whatever point is most convenient for handling a particular problem. For example, we can treat as money and command over general purchasing power which the owner has not parted with for a period in excess of three months, and as debt what cannot be recovered for a longer period than this; or we can substitute for “three months” one month or three days or three hours or any other period; or we can exclude from money whatever is not legal tender on the spot.”


    • hamblin

      I looked again at that paper, and got as far as the excerpt from Palyi regarding the money-status of savings deposits. With all due respect to Professor Salerno, I just don’t get it. Here’s how it seems to me:

      If I lend you money, you spend it into circulation, and I don’t. When you give it back, you withdraw it from circulation so that I can spend it. We can’t both spend it at the same time, so counting it as my money when it has been lent to you amounts to double-counting it.

      When I lend it to a savings bank by depositing it into a savngs account, the bank lends it into circulation, and I can’t spend it. When I withdraw it to spend, the bank reduces its outstanding loans by the same amount, minus the fraction that was held in reserve. (Beause of that fraction, depositing money into a savings account reduces the money stock, and withdrawing it correspondingly increases it.) If we count savings deposits as the depositor’s money, we are double-counting it.

      What am I missing?

  • MichaelM

    If asset backing is what counts, then what asset backs up gold?

    • Mike Sproul

      Gold is a commodity and is its own backing. A paper claim to gold, on the other hand, is the liability of its issuer, and is backed by the issuer’s assets.

      • MichaelM

        Isn’t a liability backed more by the liable’s ability to cover it, rather than strictly their existing assets?

        • Mike Sproul

          If, by “ability to cover it” you mean future earnings, then those future earnings are an asset to the firm, and would determine the value of the firm’s liabilities.

  • McKinney

    Would be interest in your thoughts on Frank Shostak’s Austrian Money Supply series.

  • RalphMusgrave

    Let’s just face it: there is no sharp dividing line between money and non money. Plus I don’t even agree with Kurt Schuler’s suggestion at the start of his post that there is something wrong with counting commercial bank created “credit/money” as money.

    Money is defined as anything widely accepted in payment for goods and services or in settlement of debts. And commercial bank “credit/money” is widely accepted. Ergo it is money.

    But that is not to say that ALL commercial bank money is money. As I think Selgin points out in his book, money which the relevant depositor intends lodging in a bank for an extended period is not money, or has far less “money characteristics” than money in a checking account. And that is recognised in nearly every country’s money supply measures. That is, money in deposit accounts is only counted as money where a very broad definition of the word is intended.

    And the latter point applies to monetary base as much as it does to commercial bank money. That is, if I get a check from the central bank in exchange for government debt I’ve sold, and I deposit the money for an extended period in some bank, it is debatable as to whether that should be counted as money.

    • Mike Sproul

      But there are sharp dividing lines between money that is the Fed’s liability, or money that is Chase’s liability, or B of A’s liability, and this is true regardless of whether the money exists as checking account dollars, savings account dollars, or paper dollars. It is pointless to try to decide whether or not something is money based upon whether it sits in the bank for a long time or a short time. Anything of value is potentially usable as money, and whether we call it money or not, its value is determined either by supply and demand (if it’s a commodity) or by the assets backing it (if the money is someone’s liability).

      • RalphMusgrave

        I agree there is a big distinction between Fed and Chase produced money: the former is a net asset as viewed by the private sector, and the latter is not. (I’m assuming one counts the Fed as part of the government machine.) Plus I agree that that distinction obtains regardless of “whether the money exists as checking account dollars, savings account dollars, or paper dollars..”.

        I also agree that “Anything of value is potentially usable as money…”. Examples include platinum or diamonds. But you aren’t seriously suggesting that platinum or diamonds should be counted as money are you?

        Re deciding whether something is money on the basis of how long that money is lodged in a bank, it is certainly normal practice in most countries to take that into account in working out the quantity of money – depending on the definition (M0, M1, M2 etc etc). And that strikes me as a valid procedure.

        To illustrate, if I plonk $X in a deposit or “term” account, where I am denied access to the money for six months, that money is just not part of the money supply. There is no way I can spend it. I can of course use the money in the account as collateral to get a loan, and spend the money that way. But then I can do the same with a car or house. And cars and houses are not normally counted as money.

        • Mike Sproul

          But why should we care what is counted as money? Monetarists care because they think the quantity of money determines the price of groceries. But if you think that the value of money is determined by the assets backing it, then the overall quantity of money doesn’t matter, and still less does it matter where we put our arbitrary lines between what is money and what is not.

          For example, people might issue a bunch of IOU’s promising to deliver a share of GM stock on demand. There might be 1 million shares of genuine GM stock (which is GM’s liability), and 2 million IOU’s (which are the liabilities of their issuers). Furthermore, some of those IOU’s might be instantly redeemable, some in 6 months, and some in 2 years. GM’s assets and liabilities are unaffected by those IOU’s, so the IOU’s do not affect GM’s stock price. It matters even less that the IOU’s might be due now or in 2 years. It would thus be a pointless exercise for someone to try to measure the supply of GM stock by counting only the IOU’s that are instantly redeemable.

  • Rick

    Money is strictly a legal convention, i.e., it is only what a country’s constitution declares it to be. Therefore, as absurd and simplistic as it sounds, under the Coinage Clause (Article 1, Section 8, Clause 5) only current metallic coin is money, and only current metallic U.S. coinage can discharge debt. Every other kind of currency, other than current metallic U.S. coin, though it may pass as money, is a money substitute that must be regulated by income taxation (where the legal tender privilege is the non-property source from which the income is derived, and incidentally, this kind of income tax needs no support from the 16th Amendment).

    Henry George (though he wrote when the temporary 1862-1872 Greenback-regulating income tax had ended and a proposed 1894 income tax had been struck down as unconstitutional) is the only economist I know that made this distinction between coin and non-coin currencies, which caused him to reject that idea money substitutes could be labeled “capital.” Only the Treasury’s coined money in the possession of the businessperson would qualify as capital. Debt notes, whether issued directly by Congress or the Fed, are not capital, no matter how commonly we refer to them as such.

  • RalphMusgrave

    It probably wouldn’t be a disaster if we abandoned trying to define money, and if governments ignored the monetary aggregates, and just concentrated on boosting demand in a recession and damping down demand during a boom.

    But my hunch is that it’s worth keeping an eye on ALL THE AGGREGATES, including a country’s total exports, imports, total amount being loaned to mortgagors and so on. That gives us an idea as to what is going on.