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

What Bank Intermediation Means

As part of my relentless (some will say obsessive) quest to stamp-out fallacies perpetrated by the 100-percent reserve bunch, I found myself engaged in a discussion with some of them in the comments section of my last post. As the discussion took place some days after that post was published, I hope I may be pardoned for reproducing parts of it, in the hope that doing so might further my overarching objective.

The discussion was prompted by a remark from 100-percenter Paul Marks, who insisted (with his usual emphasis) that "Total borrowing (of all types) must never be greater than total REAL savings of PYSICAL [sic] money. …I repeat that I am NOT making a legal point – I am making a moral and logical one." In reply I wrote as follows:

Paul, what you are saying makes no sense at all. It is the very nature of lending and borrowing of "physical" assets through intermediaries that the value of financial assets or IOUs tends to exceed that of the physical assets involved. I lend a cow to A, an intermediary, in return for A's promise to return the cow to me with interest; A lends the same cow to B, in return for a like promise from B. So: one cow, two promises, no harm, no foul.

I then added,

Just to be clear, Paul, in case the "morality" of intermediated lending should not be sufficiently evident: In the example above, I understand that A is acting as my agent; because I am not in a position to expend resources to discover a worthy borrower to whom I may lend my cow, with reasonable assurance of having it returned with interest, or because I am otherwise unable or unwilling to execute the necessary contracts myself, I allow A to take on these tasks for me, in return for his own commitment to repay my principle with interest.

Where loans of "physical" money are involved, the fungibility of that money allows a bank–which is just a name for an intermediary of money loans–to assemble loans from numerous creditors, and to lend funds so assembled to an equally diverse set of borrowers, all of which serves to reduce, ceteris paribus, the banker's prospects of being unable to meet his various commitments, lowering in turn the credit risk borne by individual bank creditors.

For centuries persons with idle base money balances have found it convenient to relinquish them to bankers as a means for earning interest on such balances with less risk than they would incur by lending them directly, while also (in cases in which deposits are made in exchange for a bank's demandable debt instruments) having access to means of payment often far more convenient than physical (narrow) money itself.

Of course, as with all forms of lending, lending through banks is not risk free. But that hardly makes such lending either unethical or imprudent. Those who, rather than wishing merely to oppose such regulatory interventions as serve to augment artificially the prospects of bank failures and financial crises, plead instead for banning bank-intermediated lending altogether, though they affect to argue as proponents of freedom and morality, in fact seek to arbitrarily limit the scope of freedom of contract, and by doing so make themselves far more deserving of the charge of immorality than the bankers whom they so loftily–and so uncomprehendingly–criticize.

Reacting to my first remark, perhaps without having read the second, Mike Sproul wrote:

Except that if B doesn't pay A, and A doesn't pay you, there is both harm and foul. If the only security for A's IOU is A's possession of B's IOU, then you would insist that B's IOU be signed over to you when A lent the cow to B. Either that, or you would have placed 1 cow's worth of lien on A's other property before accepting A's IOU in the first place. Try it with a house sometime, and see if you can get lenders to carry $200,000 worth of IOU's based only upon a $100,000 house.

To which I observed:

Like I said, all lending is risky. And of course (in the absence of government bailouts) intermediaries don't survive if they continue to make excessively risky or insufficiently secured loans. The tendency, when it comes to banking, for some to hold the industry to be either inherently untenable or immoral or both because banks will occasionally fail is frankly silly. Applied to industry in general, this tendency would have it that we should put an end to all business activity, on the grounds that some people are bound to lose their shirts otherwise!

No one, in any event, is "forced" to transact with a fractional reserve bank. No law, so far as I am aware, has ever prohibited the establishment of 100-percent warehouse alternatives. (Please don't bring up deposit insurance: what I am saying goes for the long history predating both that and TBTF.) No law prevents anyone from keeping cash in a safe or safety deposit box. To the extent that the law has ever had any say regarding bank's [sic] reserve-holding decisions, that say has ever been one commanding banks to maintain some minimum positive reserve ratio–never a "maximum" ratio! And of the few important 100-percent "banks" ever established, almost all have been government sponsored arrangements, usually subsidized or otherwise propped up by laws banning would-be fractional reserve rivals.

I do sympathize with those younger students of economics, and of Austrian economics especially, who, having fallen under the sway of anti-fractional reserve propaganda disseminated by Rothbardians and their fellow travelers, have been tempted to jump on the anti-FRB bandwagon. But for the grown-ups responsible for so tempting them, I confess I have nothing but contempt. The a-priori grounds upon which they condemn FRB are utterly without merit, while a superabundance of empirical evidence flatly contradicts their positions. They are to Austrian economics what the Flat Earth Society is to geology, which is to say (to employ Leland Yeager's expression): an embarrassing excrescence.

I urge readers of who agree with me, and who know some of the misinformed students to whom I refer, to share this exchange with them, in the hope that it may contribute toward their eventual, successful deprogramming. We can, of course, never hope to purge Austrian economics entirely of the 100-percent-reserve bacilli by which it has become infected in recent years. But we can at least hope to build up such a core of well-informed antibodies as may eventually prevent those bacilli from doing any more harm to the main body of Austrian thought than the occasional e-coli does to an otherwise robust digestive tract.

  • Mike Sproul

    I am not in the 100% reserve camp. I am in the 100% ASSET camp. A bank that issues $100 of newly-created money will only issue that money to someone who offers $100 worth of stuff in exchange. George had said "one cow, two promises". That's incorrect. It's one cow, one promise, or two cows, two promises.

    • George Selgin

      When a bank makes a loan, it doesn't exchange its (spendable) credits, whether these consist of its own notes or of a deposit balance credited to the borrower, for "stuff." It exchanges those things for a promise to repay the loan, which may or may not be secured by some specific collateral. Your view confuses a spot transaction with a loan.

      In any case I stand by my method of bovine accounting: in the example I give, one and the same cow is lent, first to the intermediary, which offers its IOU in exchange, and then by the intermediary to a second borrower in return for the latter's IOU. If your theory is such as makes this a transaction involving two cows, I daresay that I shall not be alone in thinking that that is all the worse for the theory.

      • Mike Sproul

        "$100 worth of stuff" includes things like "a promise to repay the loan". That promise can be backed up by specific collateral (a lien on a house) or by a general claim on the borrower's assets (e.g., future wages).

        I am not making one cow into two. You are making one promise into two promises. Calling it a "method of bovine accounting" is demeaning to the bovine community. You lent your cow to A, who lent it to B. You hold A's IOU while A holds B's IOU. Assuming A and B have not pledged any of their other property as collateral, those two IOUs together are worth 1 cow, not 2.

        But you would not have accepted A's IOU in the first place unless A pledged 1 cow's worth of his other property to cover his IOU. A would do the same when accepting B's IOU. In this scenario we have 2 IOU's worth 2 cows, but only because 2 cows' worth of A's and B's property stand behind those IOU's.

        • Michael Miller

          Lots of lenders make loans and accept IOUs without requiring pledges of assets. Plus, the fact that both loans are unsecured does not mean that the value of the two IOUs totals one cow. Their combined value could be two cows, or 1.5 cows, or half a cow, or nothing.

    • Mathieu Bédard

      There are three individuals in the example. The lender (George, the "I"), the bank ("A"), and the borrower ("B"). One cow, two promises to repay.

      Look at it another way. If you rent out your home and the tenant subleases it, isn't there two leases for a single apartment? Is there foul play?

      • Mike Sproul

        See answer 1.a.1.

        Repeat your tenant scenario 1000 times and there will be 1000 leases, but still just 1 landlord, 1 house, and 1 tenant.

        • George Selgin

          True Mike. But what's the point? It surely can't be that of disproving what Mathieu has said.

          • Mike Sproul

            I suppose I could elaborate about rental contracts and cows, but the fundamental issue is that you, Mathieu, and 90% of the economics profession have fallen for Henry Thornton's 'False Wealth' argument:

            " "Real notes," it is sometimes said, "represent actual property. There are actual goods in existence, which are the counterpart to every real note. Notes which are not drawn, in consequence of a sale of goods, are a species of false wealth, by which a nation is deceived. These supply only an imaginary capital; the others indicate one that is real."
            In answer to this statement it may be observed, first, that the notes given in consequence of a real sale of goods cannot be considered as, on that account, certainly representing any actual property. Suppose that A sells one hundred pounds worth of goods to B at six months credit, and takes a bill at six months for it; and that B, within a month after, sells the same goods, at a like credit, to C, taking a bill; and again, that C, after another month, sells them to D, taking a like bill, and so on. There may then, at the end of six months, be six bills of 100 pounds each existing at the same time; and every one of these may possibly have been discounted. Of all these bills, then, only one represents any actual property. (Thornton, 1802, p. 86.)"

            The trouble with this argument is that A would only sell the goods to B in exchange for a bill that is worth 100 pounds. One way for B's bill to be worth 100 pounds is if B pledges 100 pounds of his property as backing for the bill. Even if no specific property of B's is liened by the bill, the bill still gives A a generalized claim on B's assets, and A must believe that claim to be worth 100 pounds, or he would not have traded.

            After that same process repeats six times, there will be six bills, which are collectively backed by 600 pounds worth of assets. All six bills are backed by assets of adequate value, whether or not they are backed by formal liens on actual property.

        • Michael Miller

          Actually, then there would be 1001 leases and 1001 tenants, but still just one house and one resident there.

      • ludwigvandenhauwe

        No foul play. But the question is a different one: How should one handle hybrid constructions? Are there any logical constraints on the kinds of contractual arrangements into which individuals can engage besides their voluntary agreement? There are different doctrines on this issue, but there is at least some logic in the position of the 100 percenters…. It's possible that tis issue is not quite decidable for now, and that it would be better to reconceive the entire free-banking debate and redirect it along different lines… Too much intellectual energy has already been wasted on this…

        • Mike Sproul


          Murray Rothbard gave a seminar at CSUN, about 1992 if I remember right. He made the usual claims that fractional reserve banks are counterfeiters, that the Fed is the head counterfeiter, etc. I asked him if he thought it was OK for a bank to issue 100 of its bank notes for 100 oz. of gold, and he said yes. I then asked if it was OK for that bank to trade its gold for an equal value of silver. He said no, that would be fraud. I asked what if the customers agreed to the swap in advance, is that fraud? He said yes, because they don't know what they are agreeing to. Then he got mad, asked "Why can't you understand this?", and the discussion ended.

          That's the level of understanding behind 100% reserve proposals. He thought that the expansion of the money supply caused by lending caused inflation. He didn't understand that as long as new money is adequately backed by the assets of the issuing bank, then there will be no inflation.

          • ludwigvandenhauwe

            There are two points: (1) I agree that the counterfeit argument is too simplistic. "Fraud" or "counterfeit" would be only one possible legal ground to invalidate a contractual construction, according to the circumstances…. There is a series of further legal techniques that can be used to invalidate certain contractual constructions…. Whetther such arguments should be used, a fortiori whether they will be convincing, when arguing with economists, is still a different issue…
            (2) Can you elaborate on your second point namely that "as long as new money is adequately backed by the assets of the issuing bank, then there will be no inflation"? I am sure this is not yet generally understood. That a contractual arrangement between market participants should entail a variation of the quantity of money in circulation will seem anomalous to some but that it will somehow have no effects at all, will seem counterintuitive to many…..

          • Mike Sproul


            My assertion, that the issuance of adequately backed money does not cause inflation, is known as the backing theory of money, and historically it has been closely associated with the real bills doctrine. I have several papers on my website, which you can find by clicking my name, or googling "real bills doctrine".

            By way of explanation:

            If a bank accepts 100 oz of silver on deposit, and issues 100 paper receipts ('dollars'), then those dollars will be worth 1 oz. each. If the bank receives another 100 oz on deposit, and issues another $100 of paper notes in exchange, then those dollars are still worth 1 oz., even though there are twice as many dollars. If the bank had instead issued those new dollars in exchange for bonds, land, or anything worth 100 oz. of silver, then the bank still has 200 oz worth of stuff backing $200, so $1=1 oz, even though there are twice as many dollars.

          • ludwigvandenhauwe

            In the first case the quantity of money in circulation does not change, OK? Only the form of the money changes. In the second case the quantity of money in circulation changes since bonds are not money, OK?

        • Mike Sproul


          Whether the quantity of money changes or not depends only on someone's arbitrary choice of how to define money. It's the value of money that matters, and whether a bank has issued 100 of its paper dollars backed by 100 oz worth of assets, or 200 of its paper dollars backed by 200 oz worth of assets, the dollar will be worth 1 oz.

          • If a dollar is worth 1oz because it is defined/backed by 1 oz, and a dollar is redeemable for that oz, it will maintain a market value (we need actual market mechanisms or transactions to establish market prices) of 1 oz because (and only because) of the exploitation of any difference between market and official (backing) price by arbitrage, which adjusts the supply of dollars in the market to match the demand for them at 1oz per dollar. I have described this mechanism in detail in my Hard Anchor note (it is called the quantity theory of money).

          • Mike Sproul


            Supply and demand is a good model for real goods that are really produced using scarce resources, and really consumed. It is not a good model for things (like modern bank money) which can consist of computer blips that can be instantly produced or retired in infinite amounts with no resource expenditure. In the case we are discussing, the only reasonable way to draw the supply and demand curves would be to make them both horizontal at a price of 1 oz./$. But once you try to do that it becomes clear that value is determined by backing, and not by supply and demand. After all, if backing was 2 oz./$, then supply and demand would both be horizontal at 2 oz/$.

  • Kevin L

    I've always assumed, from what I've heard from 100%-reserve proponents, that the issue is simultaneous demand claims on the same assets. Modern banks claim that deposits are as risk-free as a 100% depository, while at the same time lending those deposits in somewhat risky ways. Under a 100% reserve system, demand funds would be demand funds, and loanable funds would be loanable funds. The risk to depositor-lenders on interest-bearing savings would be explicit and accepted, rather than pretended away by federal guarantees. Demand deposit services would not be free, or would have to be offset by depositing loanable funds at the same institution. I don't see how a free banking system could operate otherwise without some sort of underwriting or governmental guarantee (the former being more amenable to me as a libertarian-minded person).

    • George Selgin

      "Modern" banks can claim that their deposits are risk free only to the extent that those deposits are insured by the FDIC or its foreign equivalent. Before the advent of insurance, no bank dared to make any such claim, though some could and did claim that their deposits were extremely safe because they were either backed by very large capital cushions or invested solely in low-risk bonds or both.

      Your distinction between demand funds and loanable funds is untenable. The fact that a deposit may be transferred or withdrawn on demand does not make it any less a deposit, and hence a source of loanable funds, so far as the bank is concerned. It merely adds to the challenge the banker faces in seeing to it that the bank does not lend more or for longer duration than is warranted by the funds at its disposal. The challenge is aided by the pooling of many independent deposits; yet even so the banker must speculate about the total pool to be drawn from at different dates in the future, just as he must speculate regarding borrowers' credit worthiness. Mistakes, being unavoidable, will occur. But bank capital is there–when guarantees haven't rendered its presence otiose–to protect depositors against such errors in all but exceptional cases. Still, some banks–badly managed ones, presumably– will occasionally fail. So what? In free-market based banking as in free markets generally, the rule should be caveat emptor.

      To insist that banks should not be allowed to offer risky (that is, uninsured) demand deposits to their customers, in exchange for various services and interest (including resort to circulating banknotes) is to take part in a very un-libertarian proscription of freedom of contract; while to suggest that such deposits could not possibly survive in the absence of guarantees is to overlook centuries of banking experience.

    • Mike Sproul


      If a bank has issued 100 checking account dollars, while holding cash reserves of 20 paper dollars, then that bank will also have $80 of other assets (bonds, loans, buildings, etc.). So it's not a question of simultaneous demand claims on the same assets, it's just that assets can be in various forms, like $20 in cash and $80 in bonds. That's the natural that free banks have always operated. It's only under compulsion that banks would insist on holding assets only consisting of cash.

  • Prof. Selgin, I suggest your first paragraph is deceptive in that FDIC just doesn’t have the funds to deal with a systemic collapse, as was shown in the recent crisis. Ergo all claims by fractional reserve banks that depositor’s money is safe are fraudulent or at least debatable. Banks in Cyprus claimed depositors’ money was safe. Depositors then got a rude shock.

    I prefer the system advocated by Laurence Kotlikoff, Matthew Klein, John Cochrane and others. Depositors can have their money kept in a 100% safe fashion, but they get no interest. Alternatively, if they want interest, they buy into mutual funds. And if the fund makes silly loans, then depositors take a hair-cut.

    That way, banks as such cannot suddenly fail. So no more credit crunches. Plus there’s no need for any bank subsidies (something that Dodd-Frank and Basel III have failed to achieve).

    As you put it in your book “Theory of Free Banking”, “For a balance sheet without debt liabilities, insolvency is ruled out.”.

    • George Selgin

      I agree, Ralph, that there is some risk of the FDIC going bust. But the point is that no bank not relying on insurance or guarantees would dare to advertize its deposits as risk free.

  • Chad Duper

    Selgin would make his job of persuasion sufficiently easier if he could refrain from ad hominem: "moronic cult, 100-percenters, deprogramming." As someone who thinks highly of Rothbard, Block, Salerno, I dig my heels in to resist George when he assumes a Krugmanesque grandee tone.

    That said, George makes good points when criticizing full-reserve banking: paying for warehousing, small change (And to that I'd add large change: If warehouse receipts must be recorded, anonymity is completely lost), the notion fractional reserve-banking is "fraudulent."

    I'm a "self-styled" Austrian, and one who's open to sound arguments, but I'm more open to them when they are devoid of insults.

    • George Selgin

      Guilty perhaps on at least two counts, but "100-percenters" is merely a convenient shorthand. But keep in mind the fact that, having made the points and arguments to which you refer, and many others besides, more times than I can count, and over the course of many years, I have yet to notice any decided retreats of the senior 100 percenters, including the living ones you mention, from their original stand. Some, indeed (e.g. Block and Barnett) have only been inspired to embrace the still less reasonable view that even fractionally-backed time deposits are inherently dangerous, while others (Hoppe, Bagus, and Huerta de Soto come to mind here) would rather resort to flagrant casuistry than concede the unsoundness of their arguments. I make no secret of finding all this exceedingly exasperating. But that's not all: having spent plenty of time in the circles in question, and having myself once been among the "hardest of hard-core Austrians," I thoroughly believe it to be symptomatic of the presence of a Mises-Rothbard personality cult within the more broadly-defined and reason-based body of Austrian economics. And though I quite agree that pointing this out is not the same as offering substantive arguments about FRB, and certainly don't imagine doing so to be a substitute for such arguments, I nonetheless think it high time that someone should call that spade a spade.

      • zog9605

        I am newbie here so be gentle, how does the Dodd -Frank Bill/law play into the fractional reserve vs. 100% rreserve. As a libertarian, I am skeptical of government in general and admit i do not take as much time to study all this more. Bur I agree with what I have read to this point about Mises and Hayek. Free the markets and minimize the government and wealth will enjoy greater progress, yes? How exactly is fractional so much worse than 100% reserves? Am I even on the right track in asking? Sorry if this is a waste of time. I read Jude Wanniski's "the way the World works" and found it to be very enlightening on tax rates and the government wedge, but I have yet to delve into the banking debate and which side is more worthy.

        Seems both should be let free and see which one out competes the other.

        • George Selgin

          zog9605, I'm sure that both sides of the FRB debate will agree in holding that Dodd-Frank does nothing to address the fundamental shortcomings of the present arrangement. Among other things, it does not reduce the implicit and explicit guarantees that serve to subsidize bank risk taking, thereby adding to risks necessarily present in any FRB system. As for progress, FRB can, if properly regulated (or deregulated, as it were), contribute substantially to economic progress compared to the 100-percent reserve alternative. I have posts on the topic here and here. But the locus classicus is Book II, chapter II, of Smith's Wealth of Nations.

  • George,

    I have found, as you have before me, that making a comment here can be very costly time wise. I agree with your points about some "unusual" ideas (real notes, "backing theory") floating around here about money, which have little to do with what money exists for. But please do not put all of use who favor 100% reserve banking in that box. I favor it because I think defining the unit of account is properly a government function and that the government should control the rules of its issue (e.g. currency board rules that give determination of the supply of money to the market). By isolating "money" from other financial assets and making it 100% safe from credit or default risk (and having a good real anchor to protect its purchasing/exchange value), it becomes more feasible in my view to eliminate most regulations on other financial instruments. I think that 100% reserve requirements for demand deposits, and thus clearly separating the means of immediate payment from everything else, would reduce government regulation over all.

    • George Selgin

      We disagree about the essential role of government in money, Warren. I don't believe that government must either define the unit of account or manage the supply of the corresponding medium. (On the "natural" basis for establishing a metallic commodity unit of account see anthropologist William Ridgeway's Origin of Metallic Currency and Weight Standards.) Of course a fiat base must be artificially managed in some sense, and though I agree that a currency board approach is best for some nations, it of course presupposes some other system of management in the "host" currency country.

      But on the specific matter of 100-percent reserves, while I don't doubt that implementing such a solution would render many other sorts of bank regulations otiose, it would do so only by replacing them with a single one that is more Draconian (if not more harmful) than all the others combined. By throwing out the fractional-reserve baby with the nasty regulations bathwater, it would destroy an institution that has proven itself capable, as numerous economic historians have shown (and especially when not irresponsibly regulated), of making very substantial contributions to economic development, for reasons most eloquently set forth by Adam Smith in The Wealth of Nations (Book II, chapter II). For this reason I'm not for tossing in the towel when it comes to trying to rescue the institution of fractional reserve banking from unnecessary and often counterproductive government interference.

  • Jim

    This is great. You have stirred up the hornets nest. FWIW I support this 100% since it was your clear and cohesive web video talks that converted me. The Canada and Scotland examples help you do the hard thing, picture a free market in our current one.

    You are essentially going up against the right wing faction who stuff the mattress (will admit I'm doing this in our non free market FRB system) and maybe it's b/c most people come to free markets (eg libertarianism) from the right, not the left or center.

    So would you give us mortals some bunker portfolio advice for the world we live in? M0 cash in the mattress, bank vault cash in a safety deposit box, bullion, farm land, cows?

    • George Selgin

      Jim, the sad truth is that, to the extent that you want to hold money rather than non-monetary assets, you are best off putting your money in a bank, after dividing it (if need be) into separate accounts so as to have it all insured; should you seek a little extra safety, you might choose to put it only in banks likely to qualify as Too Big To Fail, but otherwise you can just go for the ones offering the highest rates, ignoring risk (of course) since that is likely ultimately to be borne by suckers who deal with more responsible bankers, or by taxpayers.

      Of course, I deplore the fact that ours is a system thanks to which advice of the sort I'm giving can actually be called prudent. But it isn't the advice itself that should be faulted, or those who might act according to it. It is the regulations and the bureaucrats responsible for them, which are behind the perverse structure of incentives, that deserve to be condemned, and that I do in fact condemn at every opportunity.

      • Jim

        This is good sane advice George and will probably follow. I too believe in TBTF but I really no longer believe in the FDIC when the SHTF. But you have to admit this statist monetary system ends with a bang. I think sooner than later, only reason for some cold hard cash in hand. Who knows when but my 3 doomsday scenarios:

        1) rapid inflation, and it comes from abroad, maybe when the USD is no longer traded for barrels of oil, which leads to price controls which leads to lines which leads to the TSA/NSA/DHS/INS 'bringing order'

        2) One or more of the commercial banks fails abroad (85% of the money supply within), like Creditanstalt in Austria during 30's, shrinking the money supply and deflation kicks, inevitably leading to negative interest rates, forcing/coercing M1+ money to be spent while cash withdraws are halted

        3) Cyprus bank replay. lock down all accounts. no withdraws. in the meantime here are your food stamps. only whole foods for the overweight peasants, gas on Tuesdays, movies on Wednesdays.

        In all these, some M0 cash would keep some liberty and dignity. I'll continue to preach Selgin's free banking, but I'll practice Rothbardian bunkarism. with nickels, bitcoins, and small denominations for all.

  • Michael Miller

    Part or maybe most of the problem here is due to the improper or ambiguous use of words, especially “security,” “backed by,” and the like. “Security” in the context of a debt refers to assets that are pledged to secure payment of the debt, which means that the creditor may sell them if the debt is not timely paid. The debtor’s promise to pay is not “security.” Similarly, when people speak of a debt or a note being “backed by” something, they mean something *other than* the debtor’s own promise (or the note containing it) itself.

    If we count the mere promise to pay the debt as “security” or “backing” for itself, if we count it as “stuff” given in exchange for a loan, then every debt is necessarily a secured debt and is “backed by” something, and there’s no such thing as an unsecured debt – indeed, “unsecured debt” would be a contradiction. In fact, however, the term “unsecured debt” has a real use, and most debts are unsecured, that is, the creditor is relying on the mere promise to pay. Of course, at least at the time of the transaction, the creditor believes the debtor will pay, but there need be nothing in existence (no stuff nor even any intangible) beyond the promise, beyond the “IOU.”

    Another part of the problem seems to be a confusion here among several issues:

    1. security for a debt;
    2. a bank’s reserves for its account obligations
    3. “backing” for money
    4. a debtor’s assets compared to its debts
    5. a debtor’s ability to pay compared to its debts
    6. the amount of money in a society compared to the total of all loan debts in the society

    Fractional reserve banks typically are solvent, i.e., their assets exceed their debts. They just don’t hold on hand an amount of money equal to the total of all their account obligations. They try to hold more than enough to satisfy their customers’ actual demands for the money. Everyone who wants to is free to hire a warehouse or to rent a safe deposit box to hold his/her money. Although banks offer safe deposit boxes for rent, they are not warehouses.

    Paul Marks was concerned that no one lend money unless they have it. Of course, no one can lend money unless they have it. That includes fractional reserve banks. There is no need for a rule to prevent anyone lending money they do not have. Unfortunately, he expressed his rule designed to prevent people from lending money they don’t have as: total borrowing (of all types) must never be greater than total real savings of physical money. It’s not clear what Paul meant by “real savings of physical money” and how that relates to the actual amount of money in existence. Most of what is called “savings” is not held in the form of money. The totality of “savings” can easily exceed the amount of money. Perhaps Paul meant that portion of savings actually held in the form of money (“money savings”). Of course, once such savings are lent, they are held in a form other than money (a bond, note, account, or other IOU). Now the borrower holds the money and that’s not savings in his hands. Thus, a loan of $100 by someone who held $100 in money savings simultaneously increases the amount of borrowing by $100 and decreases the amount of money savings by $100. Lending of real money savings tends to result in the amount of borrowing (debt) exceeding the amount of money savings. If everyone who has saved money lends it, there will be a lot of borrowing (debt) and no money savings, i.e, the amount of borrowing (debt) will infinitely exceed (expressed as a ratio) the amount of money savings.

    Perhaps Paul meant by “real savings of physical money” merely the actual amount of money in existence. At least that doesn’t decrease when the money is lent, so if everyone who has saved money lends it, total borrowing need not exceed the total amount of money in existence. However, it is very easy for total borrowing to exceed the total amount of money in existence, even if no one lends money he does not have (just as it’s very easy for total savings to exceed the total amount of money in existence) and there is nothing the matter with it, as George showed with his cow analogy. George was correct in his description: there are two cow debts, yet only one cow in existence. Analogously, A could lend $100 to B, and B in turn to C. Then, there would be $200 in debt, yet still only $100. Mike Sproul’s objection is misguided, as he changes the issue from how many cows there are (how much money there is) to how much the IOUs are worth (which might be $200 even if there is only $100 in existence).

    • Mike Sproul


      People lend money they don't have all the time. If I want to borrow $100 from my bank, they don't wait until someone else has deposited 100 paper dollars. They just credit $100 to my checking account, in exchange for my promise to repay the loan. My promise might be backed by land or just by my future earnings. My future earnings might not materialize, or my land might turn out to be worthless. That's not the point. The point is that the bank must have thought my promise was worth at least $100, or they would not have issued the money to me.

      When textbooks talk about how $100 of reserves can support $1000 of money, they talk as if some initial deposit of 100 paper dollars was lent and re-lent several times. That's one way to do it, but another way is that the bank with $100 reserves just issues $900 of checking account dollars and keeps them in circulation.

      • "If I want to borrow $100 from my bank, they don't wait until someone else has deposited 100 paper dollars. They just credit $100 to my checking account, in exchange for my promise to repay the loan."

        Private banks actually *do* wait until someone has deposited $100. Their balance sheets must, and I'm told do, show that.

        "When textbooks talk about how $100 of reserves can support $1000 of money, they talk as if some initial deposit of 100 paper dollars was lent and re-lent several times."

        That is the very definition and practice of fractional reserve banking.

        "That's one way to do it, but another way is that the bank with $100 reserves just issues $900 of checking account dollars and keeps them in circulation."

        A central bank might, being able to do pretty much whatever it wants. But otherwise, private banks do not. At any rate, such a practice is not fractional reserve banking.

        • Mike Sproul


          "Private banks actually *do* wait until someone has deposited $100. "

          When someone borrows $500,000 from a bank in order to buy a house, the bank (which probably has only $50,000 cash on hand) will credit $500,000 to the borrower's account. The borrower writes a $500,000 check to the home seller, who deposits it in his account. Sometimes the buyer and seller bank at the same place, and sometimes the seller actually keeps the $500,000 in his account for years. In this simplest of cases, the bank has issued 10x as much money as it originally had on deposit. It has, in effect, coined the house into money.

          But of course the seller might deposit the $500,000 check in another bank, and that bank might demand payment from the buyer's bank. The buyer's bank has only $50,000 on hand, but don't forget it also has the buyer's $500,000 IOU. The buyer's bank has only to transfer the $500,000 IOU to the seller's bank, and the deal is done. The buyer's bank is back where it was before, with $50,000 cash on hand (and no IOU from the home buyer), while the seller's bank has (say) 60,000 cash on hand, plus the $500,000 IOU, while it also has just issued $500,000 worth of checking account dollars to the home seller. The seller might now keep that $500,000 for years or he might spend it again, repeating the same process in various ways. One way or another, the house was coined into money, without anyone having to deposit $500,000 in the first place.

          • "When someone borrows $500,000 from a bank in order to buy a house, the bank (which probably has only $50,000 cash on hand) will credit $500,000 to the borrower's account."

            That is not correct. The bank (e.g. Bank of America or Alameda Regional, etc.) will not loan the borrower $500,000 unless that bank has or expects to readily obtain for the sake of redemption, $500,000 in deposits. That is to say, within a workable narrow time window, the bank's balance sheet (including all its branches) will always show that total deposits (plus other liquid assets) exceed total loans, with the excess being its reserves. I'd be surprised if regulations didn't require it. At any rate, to whatever small extent in practice this does not occur, the practice is not fractional reserve banking.

            Perhaps the practice you describe could be called "negative reserve banking" or "rothbard banking". But without a central bank handy to bail out such a practice with its own deposits, it is hard to see how such a bank could either get started or long survive.

          • Michael Miller

            Mike, I've never seen or heard of a bank making a home loan by crediting funds to the borrower's account. The bank pays the money directly to the seller in a simultaneous closing in which the borrower receives the deed and the bank receives the mortgage. In your example, the bank needs $500,000. Banks do not accept another bank's customer's IOU as money. The money will be the bank's account at the Fed (that actually is money, oddly enough), not Federal Reserve Notes.

      • George Selgin

        I do wish, Mike, that you would at least acknowledge here, if only to explain away, the point I insisted on in my last post, to which this one is a mere addendum, to wit: that, whether a bank waits for deposits before writing up a loan or not, and no matter what sort of assets "back" a borrower's promise to eventually repay the loan, the fact that borrowers borrow money from banks in order to spend it means (1) that the borrowers' account will be drawn upon by the borrower before the loan comes due (for otherwise what is the point of borrowing?); (2) that (also before the loan comes due) a check or check for sums drawn will be presented to the lending bank for collection; and (3) that the bank had better have the necessary cash with which to settle or it will fail. It is a trivial point only that banks today often lend in anticipation of acquiring the necessary funds. The important point is that they cannot lend unless they either have the funds already, or secure them within a short time of making the loan. The security behind a loan makes not the least difference so far as such considerations of liquidity are concerned, as it cannot possibly come into play until the loan period has expired.

        As for the statement, "another way is that the bank with $100 reserves just issues $900 of checking account dollars and keeps them in circulation," it is (as I have also pointed out on this forum several times), for reasons that should be evident given what I have just explained, utterly mistaken. No competitive bank could get away with doing what you describe, which seriously misconstrues the manner in which liabilities in a fractional reserve banking system come to exceed by some multiple the value of cash reserves held by the system. Every decent undergraduate money and banking textbook explains this and every middling or better undergraduate student gets it.

        But I suppose you know better. So: kindly explain to me (and to anyone else who is reading) how your imaginary banker will manage to make do with his mere $100 in cash reserves, once a rival bank presents to him for collection the $900 check that its new borrower writes shortly after acquiring his loan.

        • Mike Sproul


          (1) Acknowledged
          (2) Acknowledged, but of course the payee might re-deposit the check in the same bank, not to mention that cross-clearings between banks can leave this bank not having to deliver the cash demanded by the payee.
          (3) No. The bank could deliver bonds and such rather than actual cash, or the bank could ask to delay payment to the demanding bank, thus borrowing from the demanding bank; or as before, cross-clearings could erase the debt to the demanding bank.

          "But I suppose you know better. So: kindly explain to me…"

          Yes. Yes I do. If a single check for $900 is presented, the bank can sell $900 worth of its securities, or borrow $900, or offer the demander the $900 of securities in lieu of $900 cash. But more likely, the $900 will have been lent in dribs and drabs of, say 90 loans of $10 each. As those demand of $10 come in, the bank's cash reserves will fall by $10 at a time. But of course the bank will simultaneously be getting cash deposits, and most banks find that $100 cash on hand is enough to satisfy the cash demands of $900 worth of account-holders.

          • George Selgin

            No, Mike: a bank might borrow to cover reserve losses as a temporary expedient only; it could not do so, and banks in fact do not do so, routinely except to cover random shortfalls that tend to be offset by as many random net surpluses over time. When a bank has routinely to go into the funds market to cover its loans, it looses money.

            Banks sometimes can settle with "secondary" reserves, as Scottish banks agreed to do with Exchequer bills. But that isn't the case in most clearing systems, which require settlement in cash or "good funds." In the U.S. today, for instance, settlement takes place on the books of the Fed, using banks reserve balances there, and nothing else.

            Yes, a bank might sell assets in its portfolio to acquire the extra $800 it needs to settle; but in that case the bank doesn't in fact end up having succeeded in increasing its lending by $900 on the basis of a mere $100 increment to its reserves. Instead it increases its (net) lending by $900-$800 = $100 only. The bank might have more expeditiously accomplished the same outcome by having limited its original loan to $100 in the first instance. In any event all we have here is a bank deciding to play around with its portfolio, not one deciding to increase its total outstanding loans and investments by $900. (In practice, of course, the bank couldn't even lend $100, assuming that the equilibrium fractional reserve ratio is in fact > 0.)

            Nor, finally, can a bank expect another bank to accept the first bank's loan, or security for it, in settlement! That you can imagine some such thing happening–something that would, in this particular case, mean having banks lose all control over what sorts of loans they have in their portfolios!–is not the same as it's being something actually practical, let alone something actually practiced! I am happy to grant that you have a very lively imagination, but that's not what I meant to establish by asking my question.

          • Mike Sproul

            Vikingvista and George:

            This might be clearer if I speak only of bank notes, rather than checkable deposits.

            A bank has 100 silver dollar coins on deposit. For every coin deposited, the banker issued a $1 bank note.

            Ninety borrowers come in asking to borrow $10 each, and each one offers collateral worth at least $10.

            The banker prints $900 in notes and hands them out $10 at a time to those borrowers. The bank's assets now consist of $100 in coins plus $900 of IOU's from the borrowers. The bank's liabilities consist of $1000 of its notes.

            Maybe those notes will stay in circulation or maybe they will reflux to the bank. If they reflux, the bank either cancels the notes against the IOU's it holds, or hands out coins. In this case there is no loan expansion and nothing to explain.

            But we are discussing the case where the notes stay in circulation, presumably because the economy is booming and people need those notes to handle all their trades. Can those notes stay in circulation? Of course they can. Was the bank enabled to make those loans because someone deposited $900 of coins? Of course not. The thing people 'deposited' in the bank was their IOU's. The confusion seems to stem from not recognizing that 'deposits' can take the form of coins or of IOU's. You keep saying that a bank that lent $900 must have first gotten $900 of deposits. Yes, $900 WORTH of deposits. It is irrelevant whether those deposits consisted of coins or of borrowers' IOU's.

          • "Ninety borrowers come in asking to borrow $10 each, and each one offers collateral worth at least $10."

            The bank will choose the NINE most favorable out of the 90 to make loans to, as that is all the loanable funds it has (if it desires a 10% reserve ratio).

            Now, each of those 9 borrowers will likely deposit the borrowed funds into the same bank, providing the bank with additional loanable funds to cover some of the remaining 81 requests, but deposits will always exceed loans.

            "The banker prints $900 in notes and hands them out $10 at a time to those borrowers."

            "Banker" is not a good term here. Better terms would be "soon-to-be-bankrupt banker" or "rothbardian banker" or "banker looking to spend time in a 9 x 9 cage". At any rate, if this ever happened it was anecdotal, probably during some extraordinary time under a central bank that either authorized it and propped it up, or sent in the regulators to shut it down.

          • George Selgin

            No, referring to banknotes doesn't make your position "clearer," Mike. What it does is to make it easier for you to try and evade the fundamental issue, by imagining that notes given out to a borrower, unlike a deposit credited to him, will remain in circulation after the borrower spends them, instead of being rapidly returned for payment just as a check written by him would have been.

            In fact the evasion doesn't succeed if notes are issued competitively–and the debate here is solely one concerning the lending power of competitive banks. In fact in competitive note issue arrangements like those of Scotland, Canada, and New England under the Suffolk System, notes placed into circulation tended to return to their sources within a week or two, or about as quickly as it took checks to be written and cleared on borrowed accounts.

            I know you too well now, Mike, to not realize that you will have come up with another angle before you have finished reading these words!

          • Mike Sproul

            The banker can easily maintain 10% reserves by keeping his $100 of coins, printing $900 of his notes, and lending them to people who give him $900 worth of IOU's in exchange. I'm not saying whether this is legal or not. I'm saying that it is a valid, ethical, non-fraudulent process. If the bank had followed the textbook loan expansion model of lending coins, getting the coins back, relending them, etc, until the bank had $100 of coins in reserve, (plus $900 of IOU's) backing $1000 of its notes, then the bank's balance sheet would look exactly the same.

            It doesn't matter if a dollar exists as a printed piece of paper carried in your pocket, or as a checkable bookkeeping entry. The dollar is 'in circulation' either way. So a bank that is capable of keeping $1000 of its notes in circulation is just as capable of keeping $1000 of deposits in circulation (which is to say, held in checking accounts). What I said about dollar notes applies equally to checking account dollars.

            It also doesn't matter if notes return to the issuing bank within a week or two. As one note is being paid into the bank, another will be paid out of the bank, and there will be a permanent float of bank notes that is constantly being renewed. This will be the case for all note-issuing banks, competitive or monopolistic.

          • Mike,

            “the bank's balance sheet would look exactly the same.”

            Only because of an unbelievable coincidence. It only works, because 90 independent borrowers somehow all came in to the bank simultaneously and coincidentally. Only then is sequentially loaning $10 upon $100 reserves (where the bank physically has the money for each loan) equivalent to a single step of setting up 90 $10 deposit accounts in exchange for 90 $10 borrower IOUs. But what bank (or any human) would believe such a coincidence? A banker depending upon those kinds of odds will not be a banker for long.

            In the realm of reasonable possibility, with the above scenario you might as well just have one guy repeatedly take $10 loans from the same bank until he amasses a deposit of $900. But no bank is going to set its reserve ratio at 10% if it has only one borrower. Regardless of how good a risk the one borrower appears to be, if the bank is going to grant its entire lending capacity to him, the bank is going to set its own reserve ratio at 100%, lest it turn out that that seemingly perfect borrower is nothing but a front for a competing bank, or not such a perfect borrower after all.

            People seem to think the reserve ratio is handed down by Odin, probably because America’s Odin (the Federal Reserve Bank) famously dictates an arbitrary nearly rigid 10% ratio. In reality, in fracres banking, the reserve ratio is not exogenous–it is not independent of the bank’s business profile. Instead, it is statistically determined by the bank, in the bank's best interest, to maintain a prudent reserve for the redemption risk that it is taking. To have a low ratio, the loans must be numerous, reasonably independent, and diversified or natural selection will weed it out.

            “I'm not saying whether this is legal or not. I'm saying that it is a valid, ethical, non-fraudulent process.“

            I confess, I don’t know if it is legal either, but I would be surprised if regulations, paternalistic as they often are, permitted a bank to loan money it did not have. Governments frequently follow-up market realization of bad behavior with unnecessary prohibitions against such behavior. It isn’t so much that it is fraud, but rather that it is the pinnacle of imprudence. One could argue that an exceedingly imprudent banker is unethical. I suspect most bankers would argue that.

            There’s something absurd about me arguing in parallel with Prof. Selgin, so unless you find a flaw in my argument that I can’t resist, I’m just going to sit back in my desk and take notes from here on out.

          • Mike Sproul


            The coincidence you mentioned is irrelevant to the point in question: Can a bank, or can it not, lend money without first having received money on deposit? That imaginary bank of mine started with $100 coins on deposit, against which it issued $100 of its own bank notes. The bank then printed up another $900 of its own notes and lent them to people with good collateral. It doesn't matter if the $900 was lent all at once, to one borrower or to several. In this whole process there is not a single step that is even remarkable, let alone implausible. Once the process is complete, the bank's balance sheet will show assets of $100 in coin plus $900 of IOU's, while the bank's liabilities will consist of $1000 of notes issued by the bank.

            If it happened that the $900 in notes were spent at local shops, and if the shopkeepers deposited all those notes at the bank, then the bank's balance sheet would show assets consisting of $100 in coins plus $900 of IOU's, while the bank's liabilities would consist of $900 of deposits plus $100 in notes. (The notes are retired and replaced with deposits.) You will notice that it was the loans that came before the deposits, not the other way around.

          • Mike,

            To reiterate, a loan portfolio worthy of a 10% reserve ratio is simply not going to present itself requesting a bank’s entire lending capacity on a negative reserve basis.

            Even a competing bank wishing to cause bankruptcy would at the very least order its henchmen to scatter their borrowing requests on a fractional reserve basis. Given the number and diversity of loans banks make to justify a low reserve ratio, even this would almost certainly be a fool’s errand by the conspiring bank.

            “Can a bank, or can it not, lend money without first having received money on deposit?”

            No, if by “can” you mean actually produced by reality. If by “can” you mean humanly imaginable, then yes, it is imaginable that the end result of a negative reserve bank can mimic the end result of a fractional reserve bank. It is also imaginable that a homeless person could break into a Kinkos, print up $1 million of purple “Shorty” promissory notes and use them to buy $1 million worth of gold coins. The imaginable is far to broad and unrealistic to be interesting.

            The issue is what free markets produce. They produce fractional reserve banks, not negative reserve banks, because the practice of negative reserve banking will not be sustained in a competitive environment. If you look at the balance sheets of a bank you will (almost certainly) see that at no time did it practice negative reserve banking, only fractional reserve banking.

            Is what you describe a fractional reserve bank? No, by definition. Is what you describe the actual practice of banking in a free market? No, unless perhaps by some strange odds-defying anecdotal fluke.

          • Michael Miller

            Mike, this is to your 2.a.2. When a bank's note is presented to the bank after some period of circulating as money, the bank must pay money. It is unable to cancel the note against an IOU it holds from one of its borrowers, unless it just so happens that it is that very borrower who presents the note.

          • Mike Sproul

            Vikingvista (4:20)

            There is nothing fanciful about a bank with $100 of coins on deposit crediting $200 to the checking account of a borrower, at least if the borrower has offered good security for the loan. If the IOU is really worth only $199 then yes, that's a problem, but the whole idea of being a banker is to make sure that you get IOU's that are worth more than the money lent.

            I don't know what you mean when you say 'negative reserves'. That bank is operating on 100/300=33% reserves.

          • Mike Sproul

            Michael (6:03)

            If a note was issued on loan, then repayment of the loan will normally involve the borrower presenting a note to the bank, to be cancelled against the borrower's IOU.

          • “There is nothing fanciful about a bank with $100 of coins on deposit crediting $200 to the checking account of a borrower”

            Other than the fact that it doesn’t happen.

            “I don't know what you mean when you say 'negative reserves'. That bank is operating on 100/300=33% reserves.”

            For a fractional reserve bank, reserves = deposits – loans > 0. If deposits < loans, then reserves < 0. Like you said, the loans are made before the deposits.

            If someone else has coined a term for the fictional non-fractional reserve entity that you describe, then I’m open to it. Until then, “negative reserve bank” is reasonably descriptive. “Rothbard bank” is also reasonable since he is the one who popularized this fiction in the minds of so many Austrians.

          • Mike Sproul


            That terminology is uniquely yours. When economists speak of a bank operating on 10% reserves, they mean $10 of reserves is held for every $100 that the bank carries in checking accounts. That $10 will virtually always be less than the amount of loans made by the bank.

          • Well Mike,

            That is a peculiar response, given that I just TOLD you that I came up with the terms “negative reserve banking” and “rothbard banking”. The problem here, is that when economists talk about banking, they are talking about fractional reserve banking, and not this fiction that you keep describing. I wish I could say that fiction is uniquely yours. Nevertheless, I am required to come up with some nonstandard terms just to entertain your mythical notions. As I said, if you have better terms, I am open to them.

            But one thing we know for sure, what you are describing is most definitely not fractional reserve banking. As any economist (not of a particular Rothbardian variety) will tell you, fractional reserve banks hold A FRACTION OF DEPOSITS as reserves, and loan the rest. DEPOSITS = RESERVES + LOANS. You might try cracking open an economics text. Or at least a dictionary…


          • Michael Miller

            Mike, I agree with your 11. However, you were replying to my 9, in which I was speaking generally about anyone presenting a bank's note to the bank, because that replied to your 2, and you were there speaking generally about notes that circulate in the community and so could be brought to the bank by anyone. In 9, I said that the bank has to pay money when the note is presented, and can't just cancel a borrower's IOU — unless the presenter is a borrower. Your 11 says only that the bank can cancel a borrower's IOU when the borrower presents a note. That's the exception I noted in 9.

        • George Selgin

          Truly, Mike, you are a remarkable fellow. You never run out of, or shy away from making, false arguments.

          (1) The "equivalence" you remind me of is merely the one I just finished pointing out to you; (2) I had established already that the equivalence, far from implying that it is easy for banks to keep either deposits or notes that it has lent "in circulation" for long, means that both will give rise quickly to a settlement requests equal in value to the amount lent; (3) of course it does a bank no good at all to issue still more notes as others are returned to it for payment, unless it does so in exchange for fresh base money deposits, rather than by making new loans; and while a bank may control its loans, it cannot dictate what fresh base-money deposits come its way. Saying that a bank might cover reserve requests using funds brought to it (whether in exchange for notes or for deposit credits) shortly after the loan was made is equivalent to saying that the bank might rely on sheer luck to cover the settlement debit.

          On the other hand, if the new funds were ones the bank could anticipate receiving, then of course the argument amounts to saying no more than that banks can lend an amount not exceeding the sum of reserves on hand plus those it expects to receive in short order–a claim consistent with what I and your other critics here have insisted on all along, and one far removed from your original claim that a bank having only X dollars in on hand might generally be able to lend several times that amount.

          • Mike Sproul

            Obviously, a bank is capable of printing $900 of its own notes and lending them to people who offer $900 worth of their IOU's. The closest you have come to refuting this (other than your usual blather about my character defects) is your claim that the $900 of notes lent will immediately be returned to the bank by people demanding payment in coins. They MIGHT be returned to the bank, or they MIGHT stay in circulation. If the notes return, then there is no loan expansion and nothing for us to talk about. But there are times when the notes will stay in circulation: harvest time, Christmas shopping season, economic booms, etc. During busy times the bank will be able to maintain $1000 of its notes in circulation. During slow times the bank might only be able to maintain $600 in circulation. In any case, bankers are able to maintain SOME notes in circulation, and those notes can originate from a loan that was not funded by deposits held at the bank, or by new funds that the bank might anticipate receiving.

            You mentioned the banks of Scotland in support of your claim that loans must be funded by deposits. If you would spend some time studying the paper money issued by the American colonies, you will learn that the colonial governments printed paper notes and lent them to farmers in exchange for liens on the farmers' land. These loans of notes were not funded by deposits, since there was hardly any cash in the colonies to deposit. The notes would steadily reflux to the government as loan repayments, but as one loan was repaid, another would be originated, and there was a permanent float of bank notes that stayed in circulation.

            As before, everything I've said about bank notes applies as well to checkable bank deposits.

          • Justin Merrill


            might I interject because it appears that you are talking past each other on an issue you actually agree on given Mike's last comment regarding seasonal demand for currency in a free banking system and how banks extend credit.

            Bank's would only be able to expand credit to the extent that the public is willing to hold their liabilities (notes or deposits). Yes, banks will face almost certain clearing redemption after making a loan, which they will have to attract depositors to prevent losing reserves and performing equity lending.

            To quote George's ToFB, "When banks are unrestricted in their ability to issue bank notes each can meet increases in its clients’ demands for currency without difficulty and without affecting its liquidity or solvency. Under such free-banking conditions the “transformation of deposits into notes will respond to demand,” and banks will be able to supply credit in the form that borrowers require (Agger 1918, 154). The supply of currency is flexible under unrestricted note issue because bank note liabilities are, for a bank capable of issuing them, not significantly different from deposit liabilities."

            Banks respond to the liquidity preference of the public. Mike might have been a little confusing in his articulation, but I believe this is what he is trying to say. If so, then on this particular matter I see no disagreement.

          • George Selgin

            I appreciate your intervention, Justin. The issue at hand has, so far as I'm aware, always been that concerning how much a competing bank can lend in light of its having a certain value of reserves at its disposal. I quite agree that free banks can lend more when the demand to hold their IOUs increases, and that they can freely swap notes for deposits to any extent needed without sacrificing liquidity. But the fact that a bank has $X on hand certainly does not alone suffice to enable the bank to lend $AX, where A is a number greater than 1, no matter how low the bank's desired or optimal reserve ratio may be.

          • Mike Sproul


            We agree that the amount of money circulated by banks is determined by the public's desire to hold that money. The argument is about whether the bank can lend more than is has previously received as deposits. George says no. I say yes.

        • George Selgin

          If I carry on about your character, Mike, it's because you don't take arguments seriously. Here, for example, you overlook the fact that I did not merely declare that notes lent by a bank will be returned for payment (not by "people demanding coin," by the way, but by rival banks); I pointed out the empirical evidence to the effect that competitively-issued notes generally come back withing a matter of one to two weeks.

          Deal with the facts with which you are confronted, instead of ignoring them, and you need not fear my criticizing the way in which you comport yourself in a discussion.

          • Mike Sproul

            Of course notes come back in 1-2 weeks. Notes are constantly coming out of banks and constantly going back in. I already answered that in 2.a.5, and you didn't reply. You also didn't answer my point about American colonial currency in 2.b.1.

            As for clearing redemptions by rival banks: Bank A will have claims against bank B, and B will have claims against A. Settlement can be made by a small payment at the clearinghouse. A small competing bank that has issued its own notes (or checkable deposits) to an amount larger than its reserves, has nothing to fear from the clearing process as long as it is solvent. A bank that has issued $900 of notes (or checkable deposits) and lent them in exchange for IOU's that are themselves worth at least $900, has done nothing to endanger its solvency.

        • Mike Sproul

          Probably a good time for me to sum up:

          1. Since George's post was primarily a defense of fractional reserve banking against the Rothbardian 100% reserve camp, I should point out that I am in full agreement with George on this point.

          2. In the standard textbook treatment of fractional reserve banking, competitive banks will receive deposits of currency, which they lend. The borrowed currency is spent and re-deposited in other banks, lent again, and so on until the banks are collectively holding (say) $100 of currency reserves (plus $900 of their borrowers' IOU's) against $1000 of checkable deposits. Naturally, the quantity of checkable deposits will rise and fall according to the public's desire to hold those deposits.

          3. I have claimed (and this is where my disagreement with George begins) that the banking system can arrive at this same end result by a different route. Instead of banks limiting their loans to the amount of currency on deposit, the banks can lend by crediting their checkable deposits for amounts greater than their currency reserves. For example, Banks A and B might each have $10 of currency available to lend, but instead they choose to credit $30 into the checking accounts of their borrowers. Banks A's customers spend their $30, and their checks are deposited into bank B. Meanwhile, bank B's customers spend their $30, and their checks are deposited into bank A. When banks A and B go to the clearinghouse, their accounts can be settled without any cash payment, but they have just expanded the amount of checking account dollars held in the banking system. This process can be carried on until the banks are collectively holding $100 of currency reserves (plus $900 of their borrowers' IOU's) against $1000 of checkable deposits. This is exactly the same result as before, only reached by different means.

          4. George claims that checkable deposits issue in the way just described would reflux to the issuing bank through the check clearing process, so that the loan expansion process would not get off the ground. But the clearing process described in (3) above allows the loan expansion process to proceed, so that the banking system ends up with the same $1000 total of deposits, just arrived at by a different path. Any problems with liquidity and check clearing would be the same for either system.

          5. Two examples of money being lent in excess of deposits on hand are (a) credit card dollars, since the credit card company typically has no deposits, and (b) American colonial currency, which was lent by institutions whose deposits were negligible.

          • You do know, don't you, that it is banks and not credit card companies that extend credit. I fail to see how this counts as an example of your negative reserve banking.

          • Justin Merrill

            George and Mike,

            Might I offer some thoughts on the ongoing discussion since I think there is only a disagreement maybe over the timeframe? I agree with Mike about the endogeneity of bank money and that in the very short run (even absent a central bank) a bank can create more credit than they have in deposits (equity lending). I believe there is a loose-jointedness in the production of inside money, and that finding the optimal amount of credit creation is a discovery process. While not a perfect analogy, a firm can produce too many widgets in the short run but they'll see inventories accumulate. Even with JIT production there can still be miscalculations.

            Going back to banking, a bank can use their net interest margins as a guide to how much credit to extend, and if they overshoot, they'll borrow reserves, raise interest on deposits and rein in lending a bit.

            So is George adhering a "cloak room" theory of bank credit? I mean, I agree that in the medium run a bank will only be able to lend what they have in deposits but in the short run its different. Imagine a bank starting up and trying to take market share. I think it is apparent that they make loans & investments first and look for reserves and deposits after while using interest rates as guide.

          • George Selgin

            Justin, let us please not loose track of the nature of the original propositions that elicited my criticisms. These were to the effect that a bank that received $100 in fresh deposits/reserves was thereby able to lend $900. (The specific numbers vary, the multiple being all that is pertinent.) Such claims are simply wrong.

            I have recognized in several places already that a bank might lend in anticipation of the imminent receipt of extra reserves, or of plans to liquidate some of its other assets. Those possibilities, though they have since been used in an effort to rescue the original claims from the charge of being erroneous, do so only by shifting the ground of the original point of contention. The statement "a bank with $100 in new reserves can lend $900 if it also knows that it will shortly receive at least another $800 in reserves" is an entirely different one from that originally put forth.

          • Mike Sproul

            1. It is not the $100 of fresh deposits that enables the bank to lend $900. It is the borrowers' offer of $900 worth of IOU's that enables the bank to lend $900. If you think that a bank that receives a deposit of $900 of currency is thereby enabled to lend $900 of its notes or deposits, then it is clear that if the 'deposit' took the form of $900 worth of IOUs, the bank is also thereby enabled to lend $900 of its notes or deposits.

            2. The loan might consist of a single issue of $900 of the bank's own notes to a local farmer. The farmer might pay $10 of notes to each of 90 workmen, who pay them to grocers, landlords, etc. If the economy is busy enough that people have use for those notes, then they will stay in circulation. If the economy slows, then they will reflux to the bank and the banker will have to buy back the refluxing notes with his various assets. But this would be the case regardless of whether the $900 was issued according to the textbook process that George favors, or according to the 'depositless' method that I explained above.

            3. The process would be the same if the loan consisted of $900 of checkable deposits. Instead of the 90 workmen each holding $10 of notes, they would each hold $10 of checkable deposits, or else transfer them to grocers, landlords, etc.

          • "I have recognized in several places already that a bank might lend in anticipation of the imminent receipt of extra reserves, or of plans to liquidate some of its other assets."

            Even then, is the bank not going to assess the risk of imminent receipt failing? And how are funds in transit not already properly the bank's funds, so that this doesn't really even count as a trivial exception, but is rather a different storage mode for reserves? Even then, I doubt banks are going to bring themselves so close to this "negative reserve" practice, if ever, for it to qualify as characteristic of banks, let alone in anyway contributory to business cycles.

            I know you said that you concede this point as trivial. But I suspect it is less than trivial. It may be purely theoretical. I question whether it ever happens beyond perhaps the anecdotal.

            Are you aware of any data regarding the frequency of this type of occurrence?

          • Justin Merrill

            Aha! It is clear to me now. It's a jungle when you have free bankers, Rothbardians, Greenbacker/positive money types, Real Bills Doctrine types, and Warren Coats, in one comment thread.

            So for the sake of clarity, I can set aside the 100% camp and synthesize Mike Sproul's argument and see where there is agreement and inevitable disagreement. Now that we are talking inside baseball, we can actually have more precise technical language.

            Real Bills D. focuses on an elastic currency for "the needs of trade" aka "transaction demand" for money. But, there is also the portfolio demand for money and thus we have interest bearing deposits(Tobin). Then there is a debate over the corollary transmission mechanism for the creation of inside money. Is it loan demand, like Mike Sproul is saying, or is it the portfolio motive? This question, I believe sums up the disagreement.

            Personally, I think of it as a Marshallian scissors effect that expresses the Supply-Demand intersection in the NIMs. Increased loan demand can increase the inside money supply because banks have unique roles as intermediaries that competitors can't play, but they don't have the "widow's cruse" either. Ultimately it comes down to the public's liquidity preference.

          • George Selgin

            Justin, demand for loans or credit isn't demand for money balances; on the contrary, as I've stressed several times, people who borrow spend what they borrow in short order, which means that on the margin a net debit balance is created at settlement time against a lending bank for whatever amount it lends. Despite what Mike keeps suggesting, banks cannot settle clearing balances using the IOUs or promissory notes that they collect from borrowers, no matter how "good" those notes are; nor can they use the repayment those IOUs promise eventually to yield to finance the loan supplied in exchange for them (were an IOU itself equivalent to immediate cash, what would be the point of borrowing on it be in the first place?). In Mike's statement above about clearings, he first ignores the crucial difference between marginal and average redemption demands*, and then switches from speaking of liquidity to speaking of solvency as if there were no difference. But there is a difference, which, for good 30-day paper, might be up to 30 days!

            The discussion has in any event gotten very tiresome, and that for all concerned, I'm sure. It won't lead to an agreement, so I'm taking myself out of it at this point. I need to have some time for new posts.

            I see that you are anxious to "reconcile: the disparate views being argued here. That's worthwhile if the views are indeed reconcilable. But I'm quite certain they are not.

            *Simply put, the difference is as follows: when one asks, "What can a bank do if it receives $100 in fresh deposits?" the question is usually taken to mean, what can the bank do assuming that it and the other system banks were previously in a state of reserve supply-demand equilibrium. That is, it is a question about change at the margin. Of course, in a FRB system it is trivially true that on average each dollar of reserves a bank has on hand supports several dollars of loans, and then (in stationary equilibrium) gross settlement credits for any bank tend to equal settlement debits. But these average relationships don't give the right answer to the posed question.

          • Justin Merrill


            I'm in total agreement as I side with the "portfolio view" that it is the holder of money who demands it, not the borrower. I'm also just trying to cut through the fog of war and clarify both sides' points. Now that the fog is lifted (at least in my own mind), it is clear that a "pure loan demand" view and a "portfolio" view won't be reconciled here.

          • Mike Sproul


            Let's broaden the example a little. Farmers want to borrow a total of $20,000. Banks print up $20,000 of new notes and lend them. Desired cash balances in the market total up to $1000. People already have $100 in bank notes from having deposited their silver dollar coins, so they want to hold $900 more than they currently have. So when the bank issues $20,000 of new notes, $19,100 of those notes will reflux to the bank and $900 will be held by the public as currency.

            How will the banks cope with the $19,100 of refluxing notes? One way is if they had $19,100 cash on hand before issuing the notes. Another way is to use $19,100 of their IOU's (out of a total of $20,000) to buy back the refluxing notes. The end result is that the people who had received the $19,100 of notes in payment for goods sold are the new owners of $19,100 worth of IOU's. All the bank has done is broker a loan of $19,100 from the sellers of goods to the farmers, using its bank notes as a short-lived expedient.

            Those notes might or might not bear interest. If they do, then the public might want to hold $1200 in notes instead of just $1000, but that doesn't change the story.

          • This is a reply to Mike Sproul's Aug 23 7:19 reply to Justin:

            You have not been willing to agree that a deposit (or currency) and loans are different things. My local Safeway will not accept an IOU from you representing my bank's loan to you and thus it is not money for me. George can explain what banks must redeem for the currency they issue for with free banking. But with the currency system where bank deposits must be redeemed for Fed currency, I agree with you that in your point #3 the multiplier is the same whether you describe the expansion your way or the conventional way. Either way, however, banks cannot lend more than the deposits the public is prepared to hold (less any reserve, i.e. assets (loans + deposits at the Fed + t-bills and other assets) = liabilities (deposits and funds borrowed from other banks + capital)) plus any reserve prudence or reserve requirements require.

            Your example with Justin of banks getting the public to hold IOUs in place of currency they don't want just doesn't work. The bank will have to sell its loans to another bank that wants loans. But other banks will only want loans if they have (or can reasonably expect to get or buy) reserves with which to pay for them. If your bank has over lent (money it does have or can't get) it will come out of capital as it pays out cash as deposits are withdrawn.

          • Mike Sproul

            Warren (8:10)

            Loans=$20,000. Currency held=$1000, so they can be different things.

            "My local…with free banking" It seems like you garbled some words there. Please clarify.

            "banks cannot lend more than the deposits " Banks lent $20,000 (though $19,100 of that was later bought by the public), while the public only held $1000 of bank notes and the bank held only $100 of coins on deposit.

            "other banks will only want loans if they have (or can reasonably expect to get or buy) reserves with which to pay for them."

            This doesn't make sense to me, and I don't see how it answers this:

            "use $19,100 of their IOU's (out of a total of $20,000) to buy back the refluxing notes. The end result is that the people who had received the $19,100 of notes in payment for goods sold are the new owners of $19,100 worth of IOU's. All the bank has done is broker a loan of $19,100 from the sellers of goods to the farmers, using its bank notes as a short-lived expedient.".

          • Mike,

            You: 1. "Loans=$20,000. Currency held=$1000, so they can be different things".
            2. "use $19,100 of their IOU's (out of a total of $20,000) to buy back the refluxing notes. The end result is that the people who had received the $19,100 of notes in payment for goods sold are the new owners of $19,100 worth of IOU's.

            me: 1 and 2 above are contradictory. In 2 you are saying that loans (IOUs) can be used as money. But they can't. I cannot pay my bills with IOUs of strangers to some bank–not even with US t-bills which have a good secondary market.

            You: "This doesn't make sense to me," You are referring to my brief statement of how banks decide how much to lend. It is the essence of the back and forth debate over the answer to the question "do banks lend deposits or create them by lending". Banks are in the business to make a profit. They will only lend when they think it will be profitable to do so (i.e. when they think the return on the loan will be greater than their cost of funds). I am sure that you can figure it out from there.

          • Mike Sproul


            Change this: "use $19,100 of their IOU's (out of a total of $20,000) to buy back the refluxing notes."

            to this: "use $19,100 of their IOU's (out of a total of $20,000) to buy {$19,100 of silver coins, bonds, or other stuff of comparable liquidity, and use that stuff to buy} back the refluxing notes."

            We all know that there is a broad spectrum of things that are usable as money in various situations, (e.g., the Fed's district banks settle with each other just once a year, using bonds, not 'money'), and there is no point trying to draw a dividing line between money and not-money.

            People (and banks) make loans all the time without having the reserves with which to pay for them. I can sell my $100,000 house and carry $80,000 of paper on it, with no reserves involved. I trade $80,000 worth of house for $80,000 worth of the buyer's IOU.

      • Michael Miller

        Mike, I know that many say that a checking account is "money," but it's not. A checking account is only a debt owed by the bank to the customer, plus the bank's convenient service of paying that debt when and to whom the customer directs by checks.

        A check written on the account is also not money. When A hands a check to B to pay a debt, A has not paid money and B has not received money. B or B's bank must present the check to A's bank to receive money. The check is only an order to a bank (the bank whose name is on the check) to pay the stated amount. If B or B's bank presents the check to A's bank and, for whatever reason, A's bank does not pay the money, B will go back to A and demand payment of the debt because the debt still has not been paid. A check is part of a convenient way of being able to pay money without having to carry a lot around, but the payment is not actually made until the drawee bank pays money.

        Sometimes a bank's own notes (IOUs) would circulate as an actual medium of exchange within a certain community. Even though the bank note is just a promise to pay money to the holder, if the folks in the community had enough faith in the financial strength of the bank, they might start accepting the bank notes as actual payment. In such cases, if A were to tender a bank note to B as payment, and B were to accept it as payment, then the debt is paid upon transfer of the bank note to B, and whether or not B or some later holder of the bank note gets paid when presenting it to the bank does not affect that.

        Nowadays, when we talk about someone accepting checks as "payment," we mean only that they will let you walk away with the goods upon handing over a check (and giving them your ID, etc.), but there remains a debt until the check is paid upon presentation to the drawee bank. If your check were money, there would be no reason for them to want to see your ID.

        When bank notes were treated as money by a community, they would circulate. Checking accounts do not circulate.

        When a bank merely credits $100 to my checking account, it merely acknowledges a debt to me, plus the convenient checking service of promising to pay the debt to whomever I order it to pay. It has not yet given me any money. We have to be careful in thinking about what counts as a loan and when a loan has taken place. If a bank lends me money by handing me money, the loan has been made upon handing over the money. It’s a little trickier when a bank makes a loan by giving me a checking account. All that the bank has done so far is to agree to pay money (up to the loan amount) to people of my choice. It’s a negotiable matter whether or not I pay interest on the entire amount beginning on the day my checking account become usable, or pay interest only on amounts the bank has actually paid, from the dates they paid. The fact that borrowers nearly always pay interest on the whole amount available from the first day is really only a fact about the typical structure of payments for loans. If I write checks and the bank does not pay them, I think it’s fair to say that the bank has not made me a loan and that I owe it nothing. On the other hand, if I never write checks for a year, just keep my account as is, it’s fair to say that the bank made me a loan and that I owe repayment plus a year’s interest. But we have to be careful to recognize what actually happens in each situation. The loan by checking account is not necessarily a loan of money, if by the latter we mean that the bank actually gives me or someone else (at my order) money. We can call it a “loan of money,” but that doesn’t mean that the borrower gets any money at the beginning; it gets only the right to money at the beginning. The loan by checking account is really a promise to pay money. It is nothing more until the bank actually pays money upon my order. They have every right to charge me for that promise even if I never use it, but that’s another matter.

        When a bank’s notes circulated as money, and the bank made a loan by handing the borrower some of its bank notes, the bank completed the loan by handing over its notes, and clearly the bank would have its bank notes before lending them.

        • Mike Sproul


          The fact that something is debt, or based on debt, does not mean that it is not money.

          Around 1710, people denied that the new-fangled paper bank notes were money, since those notes had ultimately to be paid off in coin, so called 'real' money.
          Around 1840, people denied that checking account were money, since those checking account dollars had ultimately to be paid off in either coins of notes (which had by then been recognized as money)
          Since 1950, people have denied that credit card dollars are money, since credit card dollars must ultimately be paid off with a coin, a note, or a checking account (which nowadays is recognized as money)

          In every case, it takes people several decades to realize that there is a permanent float of the new-fangled money that is constantly being paid off and then re-issued.

          • Michael Miller

            Mike, I agree with your first sentence. I said so myself in my comment to which you replied:

            "Sometimes a bank's own notes (IOUs) would circulate as an actual medium of exchange within a certain community."
            "When bank notes were treated as money by a community, they would circulate. Checking accounts do not circulate."

            Even though an IOU is itself a promise to pay something else, it could be (and sometimes has been) accepted as money in some communities, being transferred from one person to another as actual payment. As I said, this happens when folks in the community have enough faith in the financial strength of the bank that they treat the IOU as "as good as gold" so that tendering the bank note actually pays a debt. Contrast that with a check, the delivery of which is not actual payment, but only the first step in a payment process that is completed only when the check is presented to the bank and paid in money.

            So, bank notes have at times been money, but usually not. The issue is not whether the bank note itself must be paid off in something else, e.g., gold. It is the very nature of a bank note or any IOU that it is to be paid in something else. But sometimes they were accepted as money, as actual final payment, among folks in the community.

            So, with checking accounts, the issue is not that they are to be paid in money, it is that they are not final payment, they are not what counts as payment. Incidentally, you said that checking account dollars had ultimately to be paid off in either coins or notes. Paying in bank notes was/is generally not acceptable, since in most cases, bank notes were/are not money. Even when they were (where people in the community around the bank would accept them as final payment), the bank itself could not use them as final payment for its own debts, whether its debt was an account or a note.

            I don't know what you mean by "credit card dollars" or what counts as paying them off, or who pays them off to whom. Do I, as a credit card holder, own these dollars? How many do I have: the combined credit limits on my cards? That would make it analogous to the checking account. Or, is it the amount I spent with the card? If I want someone to pay me off for these, who will do so? Or does someone other than I, the credit card holder, own these dollars? Is it the merchant, who will be paid off by the bank behind the credit card? Is it the bank, who will be paid off by me? Do these dollars circulate in any way? If so, how exactly?

          • With great reluctance, I poke my figure in briefly.

            Michael Miller,

            1. You seem to have your own special definition of "money". The economics profession and most of the public call money whatever they can settle financial obligations with generally (leaving aside special barter arrangements with specific people), I.e. means of payment, store of value, and unit of account.
            2. Demand deposits with banks circulate as money every bit as much as paper currency. A payment is the transfer of ownership of money (currency or deposit). Cash "circulates" when you pay it to someone else (not when you walk around). Deposits circulate in the same way.
            3. Payment by transfer of a deposit is final settlement when the settlement rules governing it have been met. Delivery a check (which is an order to transfer a deposit) is not final settlement until it has been cleared and collected. That does not make the actual transfer of the deposit's ownership any less final. Most of us pay these days by pushing our deposit money to the payee via wire transfers rather than the other way around via checks.
            None of this is rocket science.

          • Mike Sproul


            The concept of a permanent float is key. As one bank note is paid off with a coin, another new bank note is issued (usually on loan), so at any given moment there is a permanent float of bank notes that never goes away. Same with checking account dollars. Most of them were created by a loan and extinguished when the loan is repaid, but there is a permanent float of checking account dollars that never goes away.

            You asked specifically about credit card dollars, so I'll elaborate. I create credit card dollars every time I swipe my card, just as if I had a printing press in my wallet that prints a bank note whenever needed. So when you ask "How many do I have?", the answer is "As many as you want to create, keeping in mind that you have to pay them back."

            So when I buy $50 worth of gas, I hand 50 credit card dollars to the gas station. In the old days these credit card dollars existed as a physical piece of paper with my card number printed on it, along with my signature promising payment of $50. Nowadays it's all just computer blips. Anyway, the gas station owner has his 50 credit card dollars (Easier to think of them as paper rather than blips.). In rare situations he could use that $50 paper slip to buy his groceries, and the grocer could use it to pay his rent, and the slip could circulate just like a bank note. But of course the gas man usually presents the $50 slip to the credit card company at the end of the month, and he gets paid shortly after I pay the credit card company. The credit card dollars are borrowed money, but so are most of the dollars in my checking account, and so are most of my bank notes. Just like with notes and checks, as one credit card dollar is paid off, another is issued, and there is a permanent float of credit card dollars that never goes away.

            No doubt George would disagree, since he insists that banks (and by extension credit card companies) can only lend money that they have on deposit.

          • Michael Miller

            Warren Coats (my numbers follow yours):

            1. I don’t think I have my own special definition of "money." I take it to mean roughly a generally accepted medium of exchange, and I agree with you that it must be something by which people settle financial obligations. It’s for that very reason that I say that checking accounts are not money. A checking account is nothing but a debt of a bank. It does not involve the bank holding something that belongs to the customer. People do not generally settle their obligations by (unpaid) debts owed to them. If I give you a check, and you deposit it with your bank, your bank will seek to settle with my bank. Your bank wants money. Your bank will not be amused if my bank takes the presented check and says to your bank “Thanks, I owe ya.”

            2. Checking accounts do not circulate. No one thinks that “circulate” means that people walk around with them. It means that they are transferred around the relevant economy and generally accepted to pay debts and make purchases. A checking account is nothing but a debt of a specific bank. That debt is not transferred around the economy and it is not generally accepted to pay debts. There have been times in which paper instruments that evidence the debt of a specific bank (the bank’s notes) did circulate, at least to a limited extent. However, the fact that those notes were issued by and signed by the bank facilitated their acceptance and circulation.

            It’s much more cumbersome to transfer the bank’s debt without such notes, and those notes are not involved in checking accounts. I could write on a piece of paper a direction to my bank to transfer $100 of the debt it owes me to a specific person or even to the holder of the paper, whoever that might be. I might get someone (if anyone, likely another customer of my bank) to accept that as actual payment for something. But even if someone did accept it, he would not likely be able to get someone else to accept that piece of paper. That paper would not circulate. However, that’s not needed, as the question is whether or not *the bank’s debt* will circulate. The person could take the paper to my bank and have the bank transfer the debt on its books from me to him (not something a banker is likely to do, but let’s assume). That person could then try to do what I did, write a direction to the bank and try to get someone else to accept that as actual payment. If one person after another were successful in getting that transfer of bank debt accepted as payment, then we could say that the bank’s debt circulates. However, that does not actually happen, and it is extremely unlikely that it ever would happen, except among customers of that specific bank for a rather limited period of time. Keep in mind that debt of a specific bank circulates only so long as each person in the chain is satisfied with holding debt of that bank and does not ask the bank to pay the debt. As soon as one person in the chain asks my bank for money and my banks pays, the debt ceases to circulate, in fact, ceases to exist. Customers of other banks and the other banks themselves are not generally interested in becoming creditors of that bank, and will not accept that bank’s debt as actual payment/settlement. If offered debt of that bank as payment/settlement, they will refuse, or they will provisionally accept it, not as actual payment/settlement, but as a first step in obtaining actual payment of money.

            3. A check is not an order to transfer a deposit. First, let’s clarify what is meant by “deposit.” I may “deposit” money with my bank. When I do, I no longer own that money; the bank owns it. There is nothing inside the bank that I own. My account is simply the bank’s acknowledgment that it owes me a debt. I have lent the bank the money. We can say that I have a “deposit” with my bank, but that is nothing but the bank’s debt to me. A check is not an order merely to transfer the bank’s debt from me to someone else. The pieces of paper that I described in 2. above were such orders. A check is an order to pay money. For my bank merely to transfer its debt from me to someone else, the latter would end up with an account at my bank (or an increased account, if he already has one). In the vast majority of cases, the final payment/settlement of a check involves paying money, not merely the bank acknowledging that now it owes the presenter.

          • Michael Miller

            Mike, your “float” is credit, not money, and it’s not permanent. Credit expands and contracts.

            You seem to be saying that the credit card holder never has any credit card dollars. He doesn’t have any before he creates them, and in creating them, he makes the merchant the holder. In fact, what the merchant gets is the credit card bank’s promise to pay money within a few days (not the end of the month). The merchant actually extends credit, expecting to be paid within a few days. What the merchant gets upon the sale is not money; it is not generally accepted as payment. The merchant might be able to sell it or use it as collateral to get a loan, but that doesn’t make it money (it’s not generally accepted as a medium of exchange by folks throughout the economy). In fact, it implies that it is not money (why get a loan or sell it if it’s already money?). The merchant just waits the few days and receives payment.

          • Mike Sproul


            That's what they used to say about checking account dollars. Spending a credit card dollar is analogous to calling your bank, asking for a $1 loan to be credited to your checking account, and writing a $1 check to a merchant. The merchant then presents the check to the bank and gets paid in a few days. I suppose that if you're going to deny that credit card dollars are money, then you'll also deny that checking accounts (and bank notes) are money.

            Where do you draw your lines between 'money' and 'not money'? For example, people could pay for a house with a $100,000 T-bond, and that same bond could circulate among the buyers and sellers of houses for some time.

          • Michael Miller

            Mike, I already explained why I think checking accounts are not money (in my comments to you, 7.a.3. and 7.a.3.a.1., and elsewhere in this whole page). I agree that credit cards are similar. In fact, my first sentence in my 7.a.3.a.5. immediately above ("your “float” is credit, not money, and it’s not permanent. Credit expands and contracts.") was about both checking accounts and credit cards. As I also explained, bank notes, though just debt like checking accounts, have at times been accepted as money.

            Checking accounts and credit cards differ, though. A checking account is an asset of the holder, and the debt owed to the holder is a definite amount of dollars. The holder uses the checking account to buy things, not by transferring the debt but by causing his bank to pay the debt, although pay it to another person. A credit card is not an asset of the holder. It's a means of getting loans. The holder uses the credit card by having his lender bank itself promise to pay the seller. The seller likes that much better than a check, which includes no commitment by the bank to pay. That's the part missing from your analogy. Other than that, having and using a credit card is like having and using a checking account that was created as part of a loan. In both cases, and unlike the checking account created by the customer's own deposits, the holder really has no (net) asset. Thus, the credit card holder never has any of these "credit card dollars." As soon as he creates them, they are held by the merchants. But, again, what the merchant holds is not really money, but only a promise by the bank to pay money.

            I explained where I draw the line as to what is money in my reply to Warren Coats in 7.a.3.a.4. above ("roughly a generally accepted medium of exchange, and I agree with you that it must be something by which people settle financial obligations"). That explains why checking accounts and "credit card dollars" (the rights held for a few days by merchants who accept credit cards) are not money. I agree that a $100,000 T-bond *could* circulate among the buyers and sellers of houses for some time. However, it actually does not, so the T-bond is not money. Moreover, even if it did, it still would not be a *generally accepted* medium of exchange, and therefore would not be money, even though in your hypothetical, it would be used to settle (not merely be something, like a check, that triggers and authorizes a later settlement).

          • Mike Sproul


            Your definition of money leaves some very fuzzy edges (overdrafts? Disney dollars? gold bars?), so I'll ask instead: What use is this definition? Money is spent just the same whether you call it money or not. Personally, I'd say that anything of value is usable as money. Some are used as money frequently and some seldom, but there is no line that can be drawn between money and not money, and there would be no point in drawing such a line anyway.

          • Michael Miller

            Mike could you say why you think my definition of money leaves some very fuzzy edges?
            You give three examples, presumably of things that my definition leaves unclear as to whether or not they are money, but it seems clear that none fit within the definition and none are money in our economy.

            Gold bars are not generally accepted media of exchange. The fact that they *could* be is beside the point. Many different things have been used as money in various times and places throughout human history, but that doesn’t mean that any specific one actually *is* money in any specific economy.

            Tokens that can be used to buy things at a specific theme park (Disney dollars) are not generally accepted media of exchange. They are accepted by the theme park owner only. That’s the same as a store credit at a specific store. I think it’s theoretically possible for such things to *become* generally accepted media of exchange in an economy. To reach such a scenario, the store or theme park likely would have to become extremely popular and it would have to issue a lot more credit than would be used to buy things at the store or theme park, since the credits would be circulating and used in all sorts of transactions outside the store or theme park as actual final payment. [Query: Would those who would have the government require banks to hold in stock 100% of the money that they would need if 100% of their outstanding credits were presented at once also require such a store or theme park to keep in stock 100% of the goods that they would need if 100% of their outstanding credits were presented at once?]

            I don’t see how overdrafts even *could* be money. They certainly are not generally accepted media of exchange, but they aren’t even things that *could* be exchanged. An overdraft is simply the act of someone writing a check that exceeds his credits. That’s nothing that can be exchanged. If by “overdraft” you meant the money with which the bank pays that check (if it does), of course, that is money, but that’s not called an “overdraft.”

            “Money is spent just the same whether you call it money or not.” Of course. One can say the same thing about roses, tables and everything else. There is a level at which all language is merely conventional. But, when actually speaking, it does matter which words one uses, because of the actual conventions in use. We have developed the concept of “money.” Any word could have been chosen for that concept, but “money” is the one we use in English.

            But you are making an unwarranted conclusion from the fact that money could be called by a different word and still be used the same. It seems that has led you to lose sight of the difference between money and all other things of value. But we have developed reasons to distinguish and refer to those things that are generally accepted as media of exchange, and those that are not (even if they could be in other circumstances). We also have developed reasons to distinguish and refer to those things that are accepted to pay/settle and those that are not. That distinction is extremely important, and very obviously so during a credit contraction, even for those who do not see the distinction otherwise.

            There can be borderline cases for nearly all concepts. There may be some things about which it is uncertain whether it is an “X” or not, yet most things are clearly X or not and “X” has a good use. Useful distinctions do not require lines.

  • DMXRoid

    The validity of the 100% reserve argument depends on the circumstances you look at it in. From the perspective of banking as it exists right now, with all of its history (at least in the US), I don't think it's unfair to call FRB fraudulent given that it only exists in the form that it does because of explicit and implicit guarantees from government that insolvency on the part of any given bank a.) won't be enforced (via suspension as in various bank crises during the 1800's), b.) won't actually impact depositors (via FDIC, transferring the risk/cost to taxpayers at large), c.) won't be allowed to happen at all (trucks of money from the Fed, bailouts, buyouts orchestrated/massaged by the Treasury), and d.) won't significantly impact the value of the dollar (one bank going down isn't going to put that big of a dent in the total money supply for dollars). Within the context of these conditions, banks are more or less encouraged to engage in behavior that would be considered fraudulent in any other industry. It's not just a question of making the timeframe for inflows and outflows match up, it's a matter of claims to dollars being straight up oversold, without telling their customers, making any of the information about how much they're overselling available, etc…

    In a world of free banking, the 100% reserve args are significantly less good. If all banks are issuing their own notes, go under if they can't pony up the cash, aren't privileged by the power of the state when facing liquidation, etc, then all the 100% reserve position becomes is a preference. I might WANT to only do demand deposits at 100% reserve banks/warehouses, and time deposits at FRB institutions, but there's no good economic or ethical reason why firms _shouldn't_ offer whatever array of services to whomever will buy them.

    That's not to say that I don't understand the basically emotional reasons Rothbard had for hating fractional reserve. For someone who spent so much time on the history of banking, it was probably hard to look at the pre-1900 or so US experience and not want to throw your hands up at the whole mess, and this became one of the handful of cases where Rothbard's personal preferences were more injected into his polemics than they should have been.

    • Mike Sproul

      Rothbard hated FRB's because he thought they had the same effects as counterfeiting new money. He failed to see that FRB's put their name on their money, recognize their money as their liability, hold assets against that money, and stand ready to use those assets to buy back that money.

      • DMXRoid

        Well, that's kind of my point, with Rothbard looking at the question through the lens of the history of banking, that's simply never been categorically true. Even when note issue was permitted to state/federally chartered banks, they _didn't_ behave that way as a whole, because every time a wave of bank failures happened, there was always significant pressure on the state and federal governments to do something to prevent it (usually, as is the case now, in order to bail out creditors more than depositors). Banks knew that, and behaved accordingly given their incentives. What's the point of _not_ setting up some fly by night wildcat bank in the boonies and over-issuing notes when you know that you, your depositors, and your creditors will all be absolved of both responsibility and consequence when things inevitably turn sour? And how is it _not_ the same as counterfeiting new money when the incentive structure more or less guaranteed that there, in fact, would not be assets backing up note issue? Even without note issue today, the same problems exist simply because there's good reason for banks to continue to engage in insanely risky behavior since they have things like the Fed/FDIC/a whole swath of institutions that more or less exist almost solely to insulate them from the consequences of unchecked money creation.

        That's why I think that the division here at least started as a matter of perspective. Rothbard _wasn't_ looking at the question through the lens of free banking, as he should have been if he wanted to make the kind of universal pronouncements about the validity of fractional reserve, but if you only apply his arguments to the historical reality of FRB in the US, they make a lot more sense. In a world where the price of money is continually pushed downwards by non-market action, how could fractional reserve banking not bear significant responsibility for business cycles given that there's little to no effective pressure to make the price of money rise as demand increases?

        • George Selgin

          The argument about bailouts, though it applies well-enough to the post-1980 situation in the U.S., doesn't generally apply before. As for fly-by-night or "wildcat" banking, which refers to the pre-Civil War era, it has been shown to have been relatively uncommon.

          But the discussion here isn't supposed to be about the perverse effects of regulation–which (I believe) everyone recognizes as a problem. It is about the inherent shortcomings of FRB. If you start mixing the bad effects of bad regulation up with problems inherent to an industry's existence, you will end up arguing that every industry on the planet must be inherently bad.

  • ludwigvandenhauwe

    I don't believe the problem with advocates of 100% reserve banking is that they are misinformed, or that they do not understand what bank intermediation means or what the issues are or are otherwise somehow stupid. They just disagree. The importance of the debate is perhaps overestimated. Look at Mises. He didn't explicitly defend 100% reserve banking. But why? My suspicion is it was because he just didn't want to question the positive law, and become embroiled in natural law considerations…. But of course libertarians should not fear that prospect……

    • MichaelM

      Ludwig here is really on to something. Ultimately, the arguments about fraud are ancillary to the central point of the 100% reserve camp: They believe that any expansion of the money/credit supply beyond some certain, pre-defined point drives business cycles. Everything else is epi-phenomenal as an argument, either trying to overcome the inherent contradiction between their (usually) doctrinaire libertarianism and their hostility to fractional reserve banking, or some explanatory or necessary appendage to the initial idea.

      Much more important to this debate, I think, would be a long, detailed post explaining the differences between monetary disequilibrium theory as it would be understood by Hayek or even Yeager and Rothbard's version of Mises' business cycle theory. Basically, if there can be a unified version of the general macroeconomic tools used in monetary disequilibrium analysis and the capital theory of Bohm-Bawerk, then there can be peace between the Rothbardians and the Hayekians.

    • Michael Miller

      Ludwig, von Mises had no problem questioning the positive law when he thought it wrong. It was not from any desire to avoid questioning the positive law that he did not support or defend proposals to prohibit fractional reserve banking. In fact, he *opposed* and explicitly *criticized* such proposals. See, for example, his section on "The Case Against the Issue of Fiduciary Media" in "The Theory of Money and Credit." ("Fiduciary media" is an unfortunate translation of von Mises's "Umlaufsmittel," itself an unfortunate choice of words by von Mises for what he meant ("notes and bank balances not covered by money"). The same word is translated as "Credit" in the book's title.)

  • Paul Marks

    I have stated many times that I am not making a legal point (I am not talking about statutes and court judgements), I am making a logical and moral point. Logical because money is not just credit (money has to be a store-of-value not just a medium-of-exchange) and moral because people who know basic logical truths have a moral duty to express them – not to obscure them or deny them.

    What George Selgin calls money is not money – this "broad money" is actually bank credit (a credit bubble), it is not a real store-of-value.

    Again this is not a demand that all loans be in gold (or silver) – this is simply pointing out that a lot of bank credit (so called "broad money") is not even government fiat (fiat – order-command) money (token notes and coins – backed by government legal tender laws and tax demands, i.e. the threat of violence).

    This "broad money" is a credit bubble – and bubbles (unless rescued by the government printing press) must burst.

    It is not real savings of physical money – not even of fiat notes and coins.

    Presently gold is flooding out of Britain (I do not know if it is flooding out of the United States) – Swiss factories are running round the clock (on a three shift system) taking gold bars from Britain and melting them down for customers (mostly in Asia).

    The establishment elite claim this is a wonderful "export trade" (as if the gold were coming from vast gold mines in Britain – which do not, in fact, exist). It is certainly not mercantilism to be concerned about the collapse of finance (with people who thought they would never have to honour their paper promises when they sold gold – suddenly finding that the people who hold the paper promises want physical gold), for the manufacture of goods and services in the West is also fundamentally unstable – consumption is being financed by a drain of money (disguised by the establishment elite pretending that the numbers on their computer screens and in their accounting ledgers are money). I am not a mercantilist – but I know that one can not (for ever) consume more than one produces, and borrow more than one saves.

    However, George Selgin (and his allies) go beyond even this.

    To them "money" does not have to be gold or silver (or any freely chosen commodity) – "money" does not even have to be government fiat notes and coins (backed by the threat of violence) "money" (to George Selgin and his allies") is also the credit bubbles of bankers.

    Even though these bankers not only do not have the gold (or silver, or any commodity) to back up their credit bubbles – they (the bankers) do not even have the government fiat notes and coins to back up their credit bubbles (their so called "broad money").

    It is not just a logical error to pretend that a credit bubble is "money" (i.e. a store-of-value) it is also, I repeat, an immoral thing to do.

    To claim that people and governments can consume as much as they like by printing more money is not just a logical error – it is an immoral claim.

    As for claiming that banker credit bubbles (loans that do not even represent government fiat notes and coins) is "money"……..

    Such a claim is false – obviously false (obvious even to an ordinary person such as me) – so if an highly intelligent person (far more intelligent than I am) makes obviously false claims it is clear that they are not just making a logical error.

    And it is the moral duty of anyone who sees wildly false and misleading claims (such as that banker credit bubbles, "broad money", are money) being made, to point out that the claims are wildly false.

    Total loans (of all types) must not be greater than real savings of physical money.

    Total loans (of all types) must not be greater than real savings of physical money.

    Total loans (of all types) must not be greater than real savings of physical money.

    I repeat – this is not a legal point (nothing to do with statutes and court judgements) it is a logical and moral point.

    People who pretend to the public that banker credit bubbles are money are not just making a logical error – it is immoral to mislead the public in this way.

    Bubbles must burst – the "Gods of the Copybook Headings" will not be denied, people can not (for ever) borrow more than they save (really save – not book keeping trick "savings" that do not, in fact, exist).

    Bankers who claim to have vast amounts of money when, in fact, their vaults are semi empty (without even a vast amount of government fiat notes and coins in them) are as immoral as people who pretend that a panic flight of gold is a wonderful "export boom".

    • Given the many strange things that have been said here I assume that someone probably said that bank credit (an asset of the bank) is (broad) money (much of which is a liability of the bank), but it certainly was not George.

      • George Selgin

        Thanks for that point, Warren.

    • Michael Miller

      Paul, I happen to agree with you that bank accounts are not money. However, I have a couple questions.

      1. Why do you say that money must be a "store-of-value"? What is it to be a store of value other than to have some value, even if the value changes? You don't mean that the value must remain constant, do you? The value of money can change, yet it's still money.

      2. You imply that the reason bank accounts (and credit generally) are not money is that they are not stores of value. But credit can be a store of value. Many people hold bonds, notes, bank accounts, etc. as stores of value.

      3. You keep saying "Total loans (of all types) must not be greater than real savings of physical money." What do you mean by "real savings of physical money"? In my first comment on this thread (#7 above on the first level of the outline), I wondered what you meant and gave a few possibilities. It's better to hear it from you.

      4. You say that this is a logical point. If so, it's necessarily true, at all times, in all situations, by virtue of the meanings of the words. Is that really what you mean about "Total loans (of all types) must not be greater than real savings of physical money"? In context, it sounds like you are prescribing (making a moral point, as you say), but if it's merely a logical point, then you're saying that total loans are necessarily no greater than real savings of physical money. If so, you'd be saying that now, and at all times, and in all places, total loans cannot exceed that amount, by definition. Is that what you're saying? If so, then I'll really need to know what you mean by "real savings of physical money."

  • Paul Marks

    As for George Selgin's specific claim.

    A promise of a cow is not a cow.

    No more than the promise of a gold bar is a gold bar.

    Credit is not money (it is not a store-of-value).

    And if your promises of cows are larger (wildly higher) than the number of cows than exist in the entire world (not just than you have – than cows in the whole world) you are not just making an intellectual error – you are acting in an immoral way.

    Bankers are not only promising money than they have – they promise (with their credit bubbles) more money (wildly more money) than actually exists.

    • I think if I ever saw you actually address a critique to one of your claims rather than simply repeat them ad nauseum, I should invest my life savings into pig wing futures.

  • A bank's credit department is always evaluating demands for its loans and lends when it is satisfied that the borrower will repay at an interest rate sufficiently above the bank's cost of funds. Its expected cost of funds depends on is actual and expected deposits and the cost of interbank borrowing should its deposits not be sufficient. It only lends, because it only makes business sense to lend, when it has the money to lend or is confident that it can get it at a cost less than the return on its loan.

  • Prof Selgin,

    I’m not sure about your claim that full reserve banks have nearly always existed only thanks to subsidies or by being “propped up by laws banning would-be fractional reserve rivals.”

    In the UK, the Trustees Savings bank movement existed for a long time without artificial support.

    For more details, see:

    • George Selgin

      No, Ralph: According to the article, "[S}avings banks were allowed either to keep deposits only at the Bank of England or to use them to buy UK government bonds that could be redeemed or sold in time to pay out on any permitted withdrawal." The author of this article, Andrew Lilico, apparently doesn't understand the difference between a back that holds cash reserves only and one that invests some part of its asset portfolio in interest-earning government bonds, which are certainly not cash–that is, base money–by any definition.

      It also appears, to judge from an early Parliamentary Report, that (1) depositors had to give advance notice for withdrawals beyond small amounts, making TSB deposits "time" rather than "demand" deposits; and that, consistent with this, (2) the deposits in question did not carry check-writing privileges.

  • Paul Marks

    Some time ago (on another thread) I asked (in response to attacks – even when I lose my temper, which I have not done here recently, it is in response to false [although perhaps honestly made] attacks) to be shown the money – the money that bankers (and banker-friends) claim to have.

    I have not been shown this money, I have been shown words instead (lots of words) – but I did not ask to be shown words, I asked to be shown the money. Not ink in ledger books or figures on a computer screen – the actual money.

    I keep repeating the same point – because I am not been shown the money.

    I want to see the money – and I want to see it now. I do not ask to be shown words, I ask to be shown the money (all of it).

    And I am not alone – lots of other people (for example in the gold market) want physical delivery of what they were promised, and they want it now.

    If the money is not produced I am forced to the conclusion that the "broad money" (or whatever form of words is used) is a cover for a semi empty vault.

    I remember Tim Congdon (I could not care less about how he spells his name – so save the academic "sic", after all even the Oxford English Dictionary dies not care about the rules of the language any more – see their treatment of the word "literally") a British banker-friend.

    He declared (repeatedly) that Northern Rock was a "sound bank" ("well managed", "honest" and so on) even as it collapsed.

    I remembered T.C. from many years before – when he claimed that my support for the restoration of commodity money by the Ming in China was support for the ban on international trade by the later Ming ("poor Paul" destroyed, in his youth, by false [although honestly made?] attacks – time for tears and violin music… no? oh well let us get on….).

    Anyway…. I told people that Northern Rock was about to collapse, I knew that because Tim Congdon was making a special point of saying it was sound (I am not saying that Tim Congdon is dishonest – only that the things he says tend not to be true, perhaps he honestly believes them).

    Now even though I am dirt poor person in the gutter (still no tears and violin music? "No Paul – get on with it") I think the people who took my advice (not the advice of the noble academics) are glad they did.

    And my advice now?

    Do not trust the property market and the stock market – they are bubbles produced by the credit expansion.

    And do not trust money in the bank either – the people in Cyprus made that mistake (the banking system is a example of credit expansion also).

    Get something real (such as physical gold) and do not let the government know you have it.

    In a good economy this would be BAD advice – as people should invest in productive industry (and gold is NOT productive)

    But this is not a good economy – it is an illusion.

    I do not know whether banker-friends really believe what you say or not – I have no "window into your soul".

    I only know that what you say (although, perhaps, honestly meant) is wrong.

    True money lending should be based on (amongst other things) a clear understanding of what is money (a store-of-value) and what is credit (which is not a store-of-value).

    This has principle is not understood (or at least not practiced) in the current system.

    I am NOT asking for the government to enforce it (I do not trust the government to do anything right) I asking bankers (and banker-friends) to show some Common Sense.

    But I do not expect my appeal to be answered in a constructive way.

    After all Common Sense was defeated long ago.

    Princeton became the home of Woodrow Wilson (not James McCosh) and, in psychology, a rejection of the ideas of Noah Porter and the accepting of the ideas of people such as William James (these days the only jobs Noah Porter could get at Yale are looking after the car park or cleaning the toilets). There has been a rejection of objective reality and objective truth.

    I can not live in this culture, and I do not want to (again – I am not asking for tears and violin music, which I will not get anyway).

    However, this culture (this culture of illusions) is coming to an end.

    And good riddance to it.

    Even though the collapse will mean my own death.

    • George Selgin

      I actually believe, based on my understanding of what happened to it, itself informed by discussion with Tim Congdon as well as with others intimately familiar with the details, that Northern Rock was sound when it failed–that is, that it was solvent. It failed because it was unable to maintain its liquidity when the commercial paper market froze up. It saw the danger coming in advance, warned the B of E of it, and was told, more or less, not to worry! Had the B of E taken the warning seriously, it could easily have arranged the bridge loan–really not a loan at all but a guarantee of a private-market loan–needed to avert disaster. In the end, by not having done precisely what central banks are supposed to do in their LOLR capacity, it ended up having to take part in a far more costly rescue of Northern Rock's creditors.

      This isn't to completely exonerate Northern Rock. But it is to suggest that its spectacular failure was as much proof of the Bank of England's incompetence as it was of Northern Rock's own poor management.

      • ludwigvandenhauwe

        It would seem that Northern Rock failed due to its liquidity risk and that, in retrospect, its funding strategy was too risky. It is less clear that that was obvious before the fact.

  • ludwigvandenhauwe

    It is sometimes suggested that most of the debates work out to arguments about what kind of banking we would end up with in a free banking environment, considering in particular that competition is a discovery procedure. I am not so sure that this is an adequate characterization of the debate. Look at it this way. The market interactions that allow for all these marvelous discoveries presuppose the definition of rules that constrain these interactions, that define what anyone has the right to do or not to do. When the rules are not well construed, for instance when they allow for external effects, we know that the extent to which the interactions will contribute to socially optimal results will be diminished. We know this from welfare economics. So market interaction is interaction within pre-defined rules…. We would not normally contend that the discovery processes of the market lead us to discover the very rules the existence of which is pre-supposed by the market interactions (although they may to some degree contribute to this….) Conceptually we would thus distinguish two levels: the level of the rules constraining market interaction (call this level 1) and the level of the patterns and discoveries resulting from interactions within those rules (call this level 2). Is the debate between 100 percenters and fractional-reserve free bankers a debate about level 1 or level 2? Arguably it is about level 1 and not about level 2. One reason why 100 percenters reject fractional-reserve banking, for instance, is that they believe credit expansion generates external effects…. The rules are bad and must be corrected, they say….

  • Many business and people do things similar to FBR (over booking etc.), should that be illegal? Never the less I think that we might be better off if money more like preferred bank stock than can flat off par when the bank extends bad credit.

    • In what way do fractional reserve banks overbook?

  • Paul Marks

    Northern Rock was "sound" – and if quote George Selgin as saying that, it will be the "banks only lend out real savings" thing again.

    I have had enough Selgin – you are prepared to say anything (anything at all).

    • George Selgin

      I'm inclined to agree, Paul, that you are not likely to learn anything useful from me. Nevertheless, for the sake of others who might wonder whether only a nutcase could claim that NR was "sound" (that is, solvent) before the run that led to its failure, here and here are just a few of many examples of opinions similar to mine from persons whom only a nutcase would label nutcases.

  • Paul Marks

    Professor George Selgin – a person may use technical language for three reasons.

    To enlighten – because ordinary language just "does not do the job".

    Out of habit – because he is so used to using technical language he does not even know he is doing it.

    Or to OBSCURE – to make something that is obviously false (if expressed in ordinary language) sound true (indeed "scholarly") if expressed in technical language.

    To say that the Bank of England should have provided more "liquidity" to Northern Rock is actually to say that the Bank of England should have provided more SUBSIDIES (sweet heart loans and other such) to Northern Rock.

    This is not "Free Banking" – it is nothing to do with Free Banking.

    • Are there three reasons why a person might ignore counterpoint?

    • George Selgin

      Were you ever to descend from that high horse of yours, Paul, you would find that, here at ground level, (1) the words "liquidity" and "solvency" have perfectly distinct meanings, long recognized by economists, by accountants, and by most educated but otherwise ordinary human folk; that (2), as I have said already (though it is like banging one's head against a brick wall), far from being alone in claiming that Northern Rock was solvent on the eve of the run that led to it being nationalized, I am merely inclined to concur with the opinion of most experts on this subject (all, presumably, bent on obscuring the truth); and, (3) that in saying that the B of E ought to have supported, by its loan guarantee, one of the proposed private sector rescues of NR, I meant simply that doing so was precisely the sort of thing that Bank had assumed responsibility for doing, and that by meeting this assumed responsibility it might have achieved a resolution far less costly than what actually transpired. Of course I never called either the actual or the better solution free banking. But an ability to resist putting words in other people's mouths is evidently not among the fine scruples you are so keen on trumpeting from astride that lofty mount of yours!

  • Paul Marks

    When a person says "We need to increase liquidity" they mean "I want to pick your pocket".

    As for Professor George Selgin.

    As for putting words in your mouth?

    So it was me who said that Northern Rock was "sound"?

    And it was me who said that one could increase the number of cows by book keeping tricks? Have you tried milking one of your paper cows yet?

    And me who (in a previous thread) made the absurd claim that banks only lend out real savings?

    Is it me who claims that one can lend out more money than EXISTS?

    Is it me who treats ninety (not nine – ninety) tenths (and other such) as if it was a normal "fraction"? Nothing to be worried about?

    "Put words in your mouth" indeed.

    You, repeatedly, say things that are false (indeed utterly absurd) and then you have the gall to pretend that it is my fault – that you never really said these things.

    "Scruples" – you have none Sir.

    "Free society" – you would not the moral responsibility of freedom if you tripped over it.

    Good day to you.

    • George Selgin

      Oh stop, Paul. All this melodrama is killing me. What next: a gauntlet tossed, and pistols at 20 paces?

      But since I am being unusually patient, and my honor evidently is a stake, my original words, in full, were these: "I actually believe, based on my understanding of what happened to it, itself informed by discussion with Tim Congdon as well as with others intimately familiar with the details, that Northern Rock was sound when it failed–that is, that it was solvent." That by "sound" I mean merely "solvent" I make perfectly clear. Nor is the meaning unusual: the two words are routinely used synonymously in banking literature.

      And I never referred to the NR situation as having been one of "free banking." Those two conjoined words are the ones I accuse you of attempting to "put in my mouth." And the accusation still stands.

  • Vinay Kolhatkar

    Hello Prof Selgin,

    First of all, let me start by saying that I agree entirely that any notion of FRB as a definitional fraud can indeed be overcome by appropriate disclosure. However, I do believe that FRB will not survive the true free market in which central banking disappears, at least not to the extent of current levels of like-for-like fractions (around 3%, ignoring bond and central bank deposit reserves; cash for cash averaging 3%). In a real free market, banks will not be treated as different from mining companies, or software corporations, or conglomerates for that matter. In other words, there would be one law governing corporations, and not banking law for banks and something else for everyone else. In a free market, there would be no reason to tolerate extreme illiquidity, it's simply too big a risk for stockholders, especially when collusive (cooperative really, but securities law makes no distinction) covered short-selling would be perfectly legal. Which brings me to my fundamental point, which is that financial intermediation can work just fine without maturity mismatching to render one's own institution illiquid. Investment companies and hedge funds work just fine without fractional reserving. Banks can easily overcome illiquidity by issuing long-term CDs and listing them on the stock market. The incremental cost of listing is quite low when you have equity and hybrid stock already listed. Depositors would effectively buy 10-15 year bonds, and sell them on the secondary market when they need funds. That solves the "problem". Unless, of course, you don't regard illiquidity as a problem. It seems to me that the free banking school at times simply asserts that maturity transformation is a good thing, without an argumentative basis (not evidentiary, because that is open to debate) for it. At times the argumentative basis supplied is that the aggregate money supply will adjust to demand, and that is a good thing. But that requires us to compare a static, frozen money supply versus such assumed adjustments, to judge the better alternative. Perhaps I am mistaken, but is there an argumentative basis for why the free market needs to have institutions that are blatantly illiquid (effectively insolvent by a legal definition, unable to pay debts as & when they fall due)? Otherwise we could say "let the free market decide what do with FRB banks, indeed, the market may force them into higher fractions, or none at all." Please keep in mind the secondary market liquidity solution that I proposed here, which I have maintained for over two years as the real solution. In other words, I suspect banks will respond to market pressures by amending their liability structures, rather than offering to charge interest for warehousing. What is wrong with that?

    Kind Regards,

    Vinay Kolhatkar
    Author, The Frankenstein Candidate

    • Paul Marks

      I do not want to sound like Bill Clinton, but….. it all depends what one means by….. in this case what is meant by "Fractional Reserve Banking".

      If a bank says "look you are not depositors, this is not a grain silo – stuff is not really "deposited" here, you are investors – we take the money you hand over to us, and lend it out". That is fine – as far as it goes. A bank can then lend out 100% (ten tenths) of these NOT "deposits", with the clear CONTRACTRUAL understanding that investors (not "depositors") can not have the money back, till when (and IF) the borrowers pay it back.

      Investors (not "depositors") would then understand they do NOT have "money in the bank" and can not draw it out (because it is not there).

      People who really did want "money in the bank" would have to PAY THE BANK TO LOOK AFTER IT (after all if the bank is not going to lend the money out….). Many people already pay for "current accounts" (I am one of these people).

      However, by the complex interactions between banks (and the fundamental confusion between money and credit – which are different things, money has to be a store-of-value not just a medium-of-exchange and banker credit bubbles are NOT a store-of-value they are NOT money) sometimes bankers (and others) get it into their heads that there is more money than actually physically EXITS.

      This will not work – this "credit expansion" is a credit BUBBLE.

      100% (ten tenths) lent out can (possibly work).

      200% (twenty tenths) lent out – is a bubble.

      The latter is better described as "Credit Bubble Banking" or "Pyramid Scheme Banking" NOT "Fractional Reserve Banking".

      Twenty tenths (or a hundred tenths, or a thousand tenths) is not what most people regard as a normal "fraction".

      Should there be a law against it?

      Should there be a law against people sincerely thinking that 1+1=27?

      Or against people throwing themselves into active volcanos sincerely thinking it is good for their health?

      A society that is so far gone that such things are common beliefs – will not be saved by any government laws.

  • Brett N.

    Most of the debate on this thread stems from a single place. Neither side is accurately identifying the assumptions that support its arguments.

    Mike Sprouls' argument that a bank can lend in excess of reserves is absolutely correct, IF…

    1) The bank has at least a very large market share within it's region.
    2) The issued notes themselves become "transactable", if that were actually a word.

    These two assumptions together provide a situation where the notes will be able to remain in circulation for extended periods without the holders needing to redeem in base money. They may or may not realize that they all have equal claim to the dramatically scarce base money that satisfies note redemption. Redemption in base money is the only threat to the bank. The greater the ability of the bank to convince the public to use the new notes in transaction, the greater it's ability to loan in excess of base money, because if the first assumption holds, a significant number of base money redemption is going to re-deposit into the same bank, resulting in no actual redemption threat to the bank. Additionally, the newly issued notes actually become money. If they become money, are they now part of the base? If they are part of the base, then the bank carries even less risk of a run on base money because in theory it could now begin to loan against the new notes, so that no new loans would be redeemable in the original base, but instead in the new base notes.

    This situation isn't too terribly different from the history of our current fiat system. Mr. Selgin and others point out that the current FRB mechanism doesn't work like that. This is absolutely true, but only because of the assumptions that they are relying on. How did we get our current fiat notes? Were they originally backed by some other base money? Was the public somehow convinced to begin using the notes themselves as money, thus creating a "new base money"? I think the answers are obvious, but it is all about assumptions.

    My opinion (and hopefully those of everyone else) of free banking and the allowance of FRB within such, are entirely predicated on a full understanding of the associated assumptions.

    To go back to the cow example. Is it fraudulent for the "intermediary" to re-loan the cow? In George's example, AFTER CLARIFYING THE ASSUMPTIONS, no, it is not fraudulent. The assumption (later clarified) was that he was approaching party A specifically as an intermediary with the understanding that he would otherwise be loaning the cow himself to party B if he had the resources and skills to do so.

    Yes there are two contracts for one cow, but there is an enormous problem. Another assumption has gone neglected. Namely, do all parties involved have equal claim to the same cow? Modern finance revolves around debt seniority. In the cow example, both parties do not have equal claim to the cow. This is very different than in the above note printing example, where all note holders seemingly have equal claim to the scarce base money.

    Which all gets to my opinion on FRB in free banking. It has already be demonstrated ad nauseum that the legal system maintains that deposit banks are not actually warehouses, but instead lending intermediaries. The assumption in allowing free banks to practice FRB is that customers are adequately informed of the banking practices. Interestingly, in the investment world, there are strict regulations regarding private placement memorandums, prospectuses, etc. while current "deposit banking" provides for almost none. If I am to support FRB as a component of free banking, I have to assume that the depositor will be adequately informed as to how the intermediary will be employing his money. IOW, simply knowing that the bank has the ability to lend out the so called demand deposit is not adequate. Proper due diligence as defined by modern investment regulation provides a fairly precise description of both the instruments utilized and the selection criteria as well as a fairly exhaustive list of the various risks of each. Repeated bold verbatim statements regarding the risk of loss. Additionally, there is always an extensive management bio, and insight into the financial soundness of the financial intermediary. Further, any time that the holding and/or strategy of the fund change, the clients must be notified and allowed to redeem if they so desire.

    If the assumption is that there would be a similar mechanism for Joe America to conduct proper due diligence, then I don't have any more problem with lending out deposits than I would with lending out any other committed investment. Conversely, to simply assume that Joe America realizes that the bank isn't a free warehouse, is not realistic. If I recall correctly, Mr. Selgin, you even paused for a moment in your debunking of the goldsmith fraudsters myth to note that gold depositors were not like Joe America, but were rather more sophisticated. It's probably fairly obvious to most that our current populace lacks financial sophistication.

  • JoeD

    Hi Dr. Selgin,

    I enjoyed your paper on synthetic commodity currency. On the issue of fractional reserve banking, I've found Austrians to be incomprehensible. I had to backtrack and lay out the definition of fraud, and invite them to explain how customers who are fully informed that they are dealing with a fractional reserve bank (i.e. a bank) are victims of fraud. At that point, they claimed that fractional reserve banking harms third parties — an externality, basically. I don't remember what they said exactly, but it wasn't just about inflation.

    Someone needs to explain to the Austrians that risk is a continuous variable and is priced accordingly. Only the most risk-averse people at the tail of the bell curve would demand full reserve banking, which I guess would be equivalent to renting a safe deposit box and putting all your money in it. Full reserve seems like a very unfavorable arrangement all around, with currency that would presumably lose value over time due to storage costs (and might be variable in value, based on age/storage, which would be inefficient.) I stumbled across the proposed Terra currency, an international commodity basket-backed concept. They're baking in a 3 or 4% annual fee or shave for storage costs. Ouch. (

    A think a key element with the Austrians is their novel view of prediction and predictability. At graduate seminars, they'd said things like "You can't predict human behavior." and "Economics cannot make valid predictions because it assumes ceteris paribus, but cp never holds." I'm a social psychologist, and research psychologists have been successfully predicting human behavior for a hundred years (although many times with small effect sizes.) It was pointed out that Apple routinely predicts human behavior, as did the organizers of the graduate seminar in allocating space and resources based on predicted attendance. It took me a couple of years, but I eventually discovered that they think nothing can be predicted "because people have free will." Then I learned that what they really mean is that we cannot predict the behavior of a specific individual. (Social science methods are not about individuals, but about population averages, tendencies, group differences, like predicting that people are more likely to buy something if they think it's scarce, or more likely to reuse their hotel linens if the room placard says 70% of guests in that very room chose to reuse their linens, as opposed to a placard that preaches conservation…) They think the brain is a black box, and they call this black box "free will", and we can never predict anything about human behavior as a result.

    Of course, people who know you well can probably predict your behavior at far better than chance across many domains. It has nothing to do with free will, and they desperately need to upgrade their understanding of what they're talking about, or decide to not bake so much extraneous philosophical content into an economics theory. The consequence of all this is I think Austrians attach a lot more unpredictability in certain domains, even to the point of mysticism. They're shockingly physicalist about property and value, for example, with their claims that ideas literally do not exist, and that no IP can ever be valid, that a movie isn't really a movie — it's just non-rivalrous ones and zeroes on a media disk. They get really tripped up over anything that creates something beyond what's physically present — like banking, "but if the depositor wants their gold, it won't be there!", where they don't seem to understand that the math of multi-use is well-understood. Hotels and airlines overbook all the time, and they mostly get it right. I don't quite understand the whole picture yet, but Austrians really struggle with the transition from physical to non-physical property, claims, valuation, etc.