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

Mike Sproul responds (guest post)

(Mike Sproul's response to points I raised in a comment on his first guest post.)

In reply to Kurt’s commentary on the backing theory/real bills view, let me start with a reasonable working statement of the real bills doctrine:

Money should be issued in exchange for

(1) short-term

(2) real bills

(3) of adequate value.

 Rule 3 is by far the most important. No bank should (or would) issue $100 to someone who offered securities worth only $90 in exchange. A bank that fails to follow rule 3 will soon become insolvent, while a bank that follows rule 3 will get a dollar’s worth of new assets for every new dollar that it issues, so that bank’s dollars will hold their value, even as more are issued. Furthermore, no law is needed to force banks to follow rule 3. This is why the real bills doctrine is the natural ideology of free banking. Sargent and Wallace also equate the real bills doctrine with the free banking view.

Rules 1 and 2 also play a role. Rule 1 prevents maturity mismatching, and assures that the assets backing the bank’s money will mature in 60 days or less, so that even if customers want to redeem all their money at once, the worst that can happen to customers (assuming the bank is solvent) is a 60-day wait to get their money. Rule 2 automatically matches the quantity of money to the needs of business. When farmers and factories are busy, they will generate many bills, some of which will find their way to local banks, to be exchanged for newly-issued money. Old-time note-issuing bankers usually found that if they issued new money based on the bills of farms and factories, then their notes would stay in circulation. But if they issued new notes to people not directly engaged in production, the new notes would return to the bank the next day. Here again, no law is needed to force banks to follow rules 1 and 2, though I should add that banks were not terribly strict in following rules 1 and 2, and they often found that new notes could safely and profitably be issued for government bonds or other “solid paper” with maturities over 60 days.

The backing view says that the real bills doctrine avoids inflation by assuring that money does not outrun the issuer’s assets. But quantity theorists mistakenly think that the real bills doctrine aims to prevent inflation by assuring that money does not outrun the quantity of goods produced in the economy. The difference between the “money outruns assets” view and the “money outruns goods” issue has been a prolific source of misunderstanding.

Adam Smith, in the passage referenced by Kurt, takes the view that the real bills doctrine will prevent money from outrunning goods, that is, that “(money) can never exceed the quantity which the circulation of the country can easily absorb and employ”. Henry Thornton (1801) and David Ricardo (1810) took a similar approach. Each of them contended that a bank that followed the real bills rule might still cause its money-issue to outrun the quantity of goods. This seemed to them a sufficient proof that the bank would cause inflation, but they failed to realize that as long as a bank only issued money in exchange for assets of adequate value, the bank’s money-issue would not outrun the bank’s assets, and the money would hold its value. (More on Thornton and Ricardo here)

Lloyd Mints (1945) attacked the real bills doctrine with his “money’s worth” argument, which can be explained with the following sequence of events:

  1. Banks lend dollars. Borrowers promise to repay the loan, not with a physical amount of their assets, but with a specified dollar’s worth of their assets.
  2. Lending of dollars creates new money.
  3. New money causes the value of existing money to fall,
  4. which reduces the real value of borrowers’ debts,
  5. which allows borrowers to borrow still more,
  6. which brings us back to #2, and a self-perpetuating cycle of more loans, more money, and more inflation.

Step 3 above, that new money causes inflation, assumes the correctness of the quantity theory. But on backing theory principles, the new money will be adequately backed by the borrower’s collateral, and so will not cause inflation. Thus Mints’ “self-perpetuating cycle” never gets off the ground. Only by assuming the incorrectness of the real bills doctrine to begin with was Mints able to conclude that the real bills doctrine was incorrect.

Now, to Kurt’s questions and comments:

  1. Question: Mike, to give me and other readers a better sense of where you fit in the long real bills tradition, please explain (a) whether the real bills doctrine as you define it is similar to the way Sargent and Wallace define it and (b) whether you think it was a misnomer for Sargent and Wallace to call their idea a version of the real bills doctrine.

Answer: I agree with Sargent and Wallace that “there should be unrestricted discounting of real bills”. Banks acting in their own best interest will only issue a new dollar to someone who offers a dollar’s worth of assets in exchange, so the bank’s assets will automatically be sufficient to cover the money it has issued. I disagree with Sargent and Wallace’s claim that the real bills doctrine leads to an indeterminate price level. As long as the bank holds some real assets, these will anchor the currency. Sargent and Wallace also accept the idea that modern paper moneys are fiat moneys, in the sense that they are unbacked. I think that modern paper moneys are backed by the assets of the issuing central bank, but they are normally not convertible into metal. But ‘inconvertible’ is not the same thing as ‘unbacked’.

  1. Question: Is the real bills doctrine as you define it a theory that is as generally applicable as the quantity theory claims to be, or is it a theory that applies to some kinds of monetary institutions and circumstances and not to others? (See my comment just below for a clarification.)

Answer: The backing version of the real bills doctrine says that money is valued according to the assets and liabilities of its issuer, just like stocks, bonds, bills, notes, warrants, and any other financial securities. In this sense it is generally applicable.

Comment: In wartime many occupying armies have issued a kind of currency, usually forced into circulation at par with the existing currency on the official market. The aim of this currency was frankly to enable the occupier an easy means to commandeer goods. The currency was not readily exchangeable into any foreign currency, including the home currency of the occupying army, and there were no reserve assets of recognized international value held against it. Does the real bills doctrine as you define it apply to analysis of these currencies, or is it confined to more normal historical cases?

Answer: If, for example, Mexico has issued 100 pesos, backed by assets worth 100 oz of silver, then 1 peso=1 oz. If America then takes over, issues 200 “occupation pesos” and spends them, then there will be 300 pesos backed by only 100 oz of assets, so then 3 pesos=1 oz.

  1. Comment: To me, saying that acceptance of currency for tax payments provides a kind of backing is metaphorical rather than literal. When I think of backing I think of convertibility at a set rate of exchange. A currency that is supposedly backed by certain assets but cannot be exchanged for any of them at a set rate is not backed in the way I believe most people think of the term, or in the way that issuers backed their currencies under most types of gold standard.

Answer: In the American colonial period, colonial governments would collect taxes worth maybe 5 silver shillings (English coin) from each colonist every year. In 1690, the colony issued paper shillings and paid them to soldiers, declaring that those paper shillings would be acceptable in lieu of silver shillings at tax time. If those paper shillings had been convertible in the conventional sense, a colonist could have presented a paper shilling to the government and demanded a silver shilling in return. But these paper shillings were convertible in the sense that the colonist could present them to the tax man and be relieved of having to pay 1 silver shilling. Either way, the paper shilling is convertible. If the present value of the colony’s “taxes receivable” is 1000 shillings, then the colony could issue up to 1000 paper shillings against that asset, just like a banker could issue 1000 paper shillings against 1000 shillings worth of assets held in his vault.

  1. Comment: In your example of the playing card money, the retort from the quantity theory side is that inflation does not occur when the money supply triples because the velocity of money (inversely related to the demand to hold money) changes. Other things are not equal, in other words.

Answer: Nobody claims that the value of General Motors’ bonds or stock is affected by the velocity with which those stocks and bonds circulate. The values of GM’s stocks and bonds are determined by GM’s assets and liabilities (broadly defined). The backing theory says the same is true of money. In the case where each paper livre is convertible into 1 silver livre, velocity cannot affect that value. Once this is recognized, the next step is to recognize that convertibility can take many forms. A paper livre can be convertible into silver, into bonds, into taxes, land, loan repayments, etc. If metallic convertibility makes velocity irrelevant, then so do the other kinds of convertibility. (I should add that velocity is a notoriously slippery concept, and is easily used to allow quantity theorists’ models to always be right no matter what.)

  1. Comment: In the 1990s I heard this aphorism about central banks in poorer countries (which, remember, had had a terrible decade from 1982-1992): “The assets are garbage; the liabilities, everyone believes in.” For a floating currency during normal, noncrisis periods, as long as the central bank’s liabilities are limited in quantity, I don’t see what difference it makes whether the assets are Swiss government bonds or dodgy loans purchased from domestic banks. During a crisis it matters, because Swiss government bonds are easy to sell in quantity without having to offer fire-sale discounts and dodgy domestic loans are not. If backing matters so much, though, shouldn't we see less marked a difference between crisis and noncrisis periods in the value of the currency, since the assets on the day before the crisis and the assets the day the crisis begins are the same?

Answer: It makes no difference whether a central bank issues 100 pesos in exchange for Swiss bonds worth 100 oz of silver, or dodgy loans worth 100 oz. of silver. Either way, 1 peso=1 oz.  But if a crisis comes along, the Swiss bonds will stay at 100 oz, while the dodgy loans drop to 20 oz, and the pesos that are backed by the dodgy loans will fall from 1 oz/peso to 0.2 oz./peso.

  • Justin Merrill


    I'm on board with backing theory if there is redeemability and with the context of money demand, but your exposition of it and support for real bills I think is problematic. Most loans have something "backing" them, such as an auto, a car, or a business's assets. Only some are signature loans or open lines of credit. These items that have a down payment associated are often more collateralized than a bill of exchange. Your assertion that banks should discount real bills because they are of short duration is begging the question. If backing theory is correct, then duration wouldn't matter as long as the market value of the assets is above the liabilities of the issuing bank.

    Another thing I've noticed is that sometimes your explanation of backing theory implies that it is the income of the assets that back the currency, and sometimes it is the market value of the assets. The reason I point this out is because the two can differ. An enterprise's stock doesn't often trade at book value. A firm that trades below the marketable value of its assets is better off dead or is misvalued. So is it the NPV of the income or is it the marketable value of the assets that back the money? Does this depend, in your view, on whether or not there is redeemability?

  • Mike Sproul

    I think of redeemability in a broad sense. A checking account dollar can be redeemable for a paper dollar, for a bond, for loan repayments, or for points at my local grocery store. The backing theory is workable in all those cases.

    You're right about the importance of collateral. In fact, I should have re-worded rule #3 to say "of adequate value, backed by adequate collateral". A bank might reasonably prefer a 30-year mortgage backed by a house to a 60-day bill backed by the good credit of a farmer or factory owner. That's why banks pay far more attention to adequacy of value than they do to whether the bill is short-term or productive.

    I'm thinking of income as being equivalent to market value. For example, a piece of land that yields an income of $50/year forever will have a market value of $1000 if the interest rate is 5%.

    • Justin Merrill


      I just stumbled upon JP Koning's blog post which I fully agree with; it's what I've been saying. It is the liquidity premium from the demand for money that makes it possible for the liability of the issuer to trade above the backing assets, thereby making it profitable to issue money. I also don't think the QTM needs to be abandoned, just not taken so literally.

      • Mike Sproul


        JP and I have discussed that quite a lot. I can't pin him down on how big the liquidity premium might be, but I usually throw out a number of 3%. In that case, the backing theory explains 97% of the value of money and the quantity theory explains 3%. But I have lots of trouble swallowing that premium concept. For one thing, suppose that the economy is in a tight money condition and the premium is 3%. Doesn't that mean that if the economy is in a cash glut, or easy money condition, that the premium might be -3%? So isn't the premium just noise? And of course, if every dollar is backed by assets worth 1 oz, then any mispricing above or below 1 oz creates an arbitrage opportunity.

        I've seen grocery stores get low on pennies, and offer $1.03 to people who bring in 100 pennies, so I can live with the concept of a premium. But I think the backing theory itself is the elephant in the room, and the premium is, at best, a refinement.

  • McKinney

    "Nobody claims that the value of General Motors’ bonds or stock is affected by the velocity with which those stocks and bonds circulate."

    On the contrary; they do and should if they don't.

    " The values of GM’s stocks and bonds are determined by GM’s assets and liabilities (broadly defined)."

    No. All value is subjective. Mainstream finance values stock by NPV, but investors arrive at a very wide range of NPV depending on revenue forecasts and discount rates.

    As I have written, the main problem I have with Mike's theory is 1) it ignores the theory of value by claiming it doesn't apply to money. Mike claims money is unique so value theory doesn't apply, but he doesn't show why it's uniqueness justifies that. All goods are unique in some respects and similar in some respects. 2) His theory tries to drag economics back to the days of objective value before the marginal/subjective revolution.

    Still, I think the economy would be much better if banks followed the three rules he specifies for the real bills doctrine. Those rules would limit the supply of new money dramatically.

    I do have a problem with #3, however. Investment banks loaned less than the value of mortgage-backed securities before the recent financial crisis. They violated rule #1 by borrowing short and loaning long term, but they followed #3 very closely and they still got burned because the value of the assets backing the loans collapsed when housing prices plummeted. Of course, that is a reason for rule #1. If the loans are all short term then you won't be holding assets long enough for them to change in value much.

    But you can't get rid of the practice of borrowing short to lend long. It's too lucrative, just like fractional banking. If banks don't do it, other companies will spring up to do it, such as insurance companies, and cause the same problems. The rules only remove the problem to another level. Free banking may not be the perfect system, but it seems to have fewer flaws than all the rest.

  • Mike Sproul


    1. If a piece of paper reliably promises 1 oz of silver, then that piece of paper will be worth 1 oz regardless of whether it changes hands once a year or 50 times a year.

    2. Mainstream finance has all kinds of ways to arrive at risk-adjusted values for securities. The proposition that a firm's liabilities are valued according to its risk-adjusted assets (including income streams) is not controversial.

    3. The theory of value applies to commodities like gold, apples, land, labor, salt, etc. It does not apply to pieces of paper that are the liability of their issuer. Commodities are valued according to supply/demand principles. Liabilities (bonds, bills, notes, warrants, options, etc) are valued according to the assets and liabilities of their issuer.

    4. The real bills doctrine is about banks covering the money that they issue. It's not about limiting the amount of money that they issue. If the bank issues $100, backed by assets worth 100 oz, or if the bank issues $300, backed by assets worth 300 oz, then either way, $1=1 oz.

    5. A bank might issue $100 and lend it to a guy who offers a $120 house as collateral. Initially the $100 is adequately backed by the house. But if the house falls to $90, the $100 is no longer adequately backed, and the money might lose 10% or so of its value. The real bills doctrine says there will be inflation in this case, but it does not imply that this kind of loan should be prohibited, as long as the bank and its customers went into it with their eyes open. (The money would lose its value whether or not the bank followed rule #1.)

    6. Banks and their customers must balance the costs and benefits of borrowing short and lending long. It's in their own best interest to do so, so legal restrictions would only hinder.

    • McKinney

      "The proposition that a firm's liabilities are valued according to its risk-adjusted assets (including income streams) is not controversial."

      Yes it is controversial. Take a class in finance or read a finance textbook. BTW, income streams are never considered assets.

      " The theory of value applies to commodities like gold, apples, land, labor, salt, etc. It does not apply to pieces of paper…"

      It would be nice if you would stop repeating yourself and offer some kind of evidence or rationale. The theory of value applies to human beings. If humans don't set the value of paper but angels do, then I might agree with you.

  • Martin Brock

    I've expressed my sympathies with Mike in the past, so I won't belabor the point now.

    I will reiterate a pet peeve on semantics, because it bears repeating. "Fiat" means "by decree" or "forced", not "unbacked". Bitcoin for example is unbacked, while the dollar is backed by entitlement to tax revenue, but Bitcoin is anything but fiat money, while the modern dollar is quintessentially fiat money. I prefer to use "fiat money" only in the context of Chartalism or Modern Monetary Theory.

    I don't fully understand the core disagreement between Kurt and Mike, so I'll explore with reference to instruments like Bitcoin.

    I suppose an instrument like Bitcoin (though not Bitcoin 1.0 itself) could become money. Bitcoin itself does not and could not have a sufficiently stable value, but alternatives with a more stable value are conceivable and arguably already exist. Peercoin is a possible example.

    Does Mike disagree? I would only add that an unbacked money similar to Bitcoin, but with an elastic supply responding to demands for exchange media, would tend to behave like backed money even though it is unbacked, i.e. the supply of unbacked money with a stable value approximates the supply of backed money.

    Similarly, the supply of backed money responds to demands for money, as increasing demands for money lead people to circulate more notes backed by more valuable assets. If the supply of backed money does not respond to demands this way, then I expect the value of backed money to be unstable, even if this means that the value of assets backing the money are similarly unstable. Does Mike disagree?

  • Martin Brock

    In 5, when discussing the stability of backed money, I assume that backed money is the only money, so when I say that an inelastic supply of backed money destabilizes the value of the money (and the value of assets backing it), I assume that people are not free to increase the money supply by adopting an unbacked exchange media instead.

    I don't assume that only one thing can be money, and this assumption seems increasingly archaic. A variety of competing monies, even within a small region, is more possible than ever before, and I expect this variety if states do not effectively rule it out.

    In a global market, already accepts many different currencies, and each Amazon consumer may imagine that Amazon deals only in the consumer's currency. The consumer's currency typically is a national, fiat currency, rather than a currency of choice, but Amazon could as easily deal with many currencies if consumer choice determined, rather than nation-states, determined the variety, and I could as easily imagine everyone dealing with Amazon in my currency of choice even if my neighbor across the street uses a different currency, even if the street does not mark a border between nation-states.

  • Martin Brock

    In other words, money is not a natural monopoly now, even if it was more so in the past.

    • Mike Sproul


      You make a good point about the meaning of "fiat money". I suppose confusion is unavoidable, given the ambiguity of the underlying concept. For example, I'd say that bitcoin has value for the same reason as gold. There is some initial kernal of "curiosity value", and once this value is established, people start to use it as money. This creates additional demand, and adds a certain premium to value. For gold, this historical monetary premium might have been 30%, while for bitcoin, the monetary premium looks like 99%+. Nobody says that gold is unbacked, but most people say that bitcoin is unbacked. I'd say that "backed/unbacked" are not quite the right words for gold and bitcoin. Much better to think like an accountant, and recognize that neither gold nor bitcoin appear as a liability of their issuer, not in the way that paper dollars appear as the Fed's liability.

      Or suppose I'm a landlord and I issue a piece of paper that says "good for 1 oz of rent on my property". That's a decree on my part, so maybe I've just issued fiat money? But it's not "forced" and it is backed. Governments do the same kind of thing all the time, and words like "decree", "forced", and "backed" apply in varying degrees. The closer we look, the fuzzier things get. The macro textbooks don't help, because in one sentence they define fiat money as money created by decree, and in the next they say it's value is determined not by decree, but by money supply and money demand.

  • Mike Sproul


    "If the supply of backed money does not respond to demands this way, then I expect the value of backed money to be unstable, even if this means that the value of assets backing the money are similarly unstable. Does Mike disagree?"

    Suppose a town has $100 paper in circulation, backed by assets worth 100 oz of silver, so $1=1 oz. Then desired real balances rise from 100 oz to 120 oz. If the extra $20 is issued in exchange for another 20 oz, then it's still true that $1=1 oz, and in addition, the town has as much money as it desires. But what if desired real balances rose to 110 oz, and politics prevented the extra $20 from being issued? Maybe the unsatisfied money demand will drive the dollar up to $1=1.2 oz, or maybe the fact that $100 is backed by 100 oz will keep the value of the dollar at $1=1 oz. I'm much more inclined to the $1=1 oz view, though maybe dollars might sell for a small premium like $1=1.03 oz. I'd also expect a recession because of the shortfall of real balances.

    • Mike Sproul

      Correction: 110 oz=>120 oz.

  • David Perrins

    Hi Mike. Haven't read this yet but delighted to see you posting. The reception in here seems to be a tad more "academic", shall we say, than over at