Front Page Titles (by Subject) 2: The Evolution of a Free Banking System - The Theory of Free Banking: Money Supply under Competitive Note Issue
The Online Library of Liberty
A project of Liberty Fund, Inc.
Search this Title:
2: The Evolution of a Free Banking System - George A. Selgin, The Theory of Free Banking: Money Supply under Competitive Note Issue 
The Theory of Free Banking: Money Supply under Competitive Note Issue (Lanham, MD.: Rowman & Littlefield, 1988).
About Liberty Fund:
This title has been put online with the permission of the author.
Fair use statement:
The Evolution of a Free Banking System
In recent years several studies have been made of the properties of hypothetical, unregulated payment systems.1 The value of these studies is limited, however, by their authors’ use of ad hoc assumptions, ranging from the proliferation of competing fiat currencies at one extreme to the complete absence of money at the other. To be really useful in interpreting the effects of regulation in the past, or in predicting the consequences of deregulation in the future, a theory of unregulated banking should be based on realistic assumptions drawn, if possible, from actual experience.
Unfortunately, there have been few free banking episodes in the past, none of which realized it in a pure form. Thus history furnishes an inadequate basis for drawing theoretical conclusions about free banking. To rely exclusively on it would invite generalizing from features unique to a single episode or from features attributable to regulation.
Another approach, which also helps in interpreting historical evidence, is to base assumptions on a logical (but also conjectural) story of the evolution of a “typical” free banking system, as it might occur in an imaginary, unregulated society called Ruritania. The story can be supported along the way by illustrations from actual history. But it only accounts for features of past banking systems that were predictable (though perhaps unintended) consequences of self interested, individual acts, uninfluenced by legislation.
Our story involves four stages: First, the warehousing or bailment of idle commodity money; second, the transition of money custodians from bailees to investors of deposited funds (and the corresponding change in the function of banks from bailment to intermediation); third, the development of assignable and negotiable instruments of credit (inside money); and fourth, the development of arrangements for the routine exchange (clearing) of inside monies of rival banks. The historical time that separates these stages is not crucial. Also, innovation, rather than consisting of steady progress as pictured here, is likely to involve many dead ends and much “creative destruction.” What matters is that each stage is a logical outgrowth of the preceding stage. Moreover, though every step is a result of individuals finding new ways to promote their self-interest, the final outcome is a set of institutions far more complex and important than any individual could have contemplated, and one which was not consciously aimed at by anyone.
Since the use of money logically precedes the emergence of banks, we begin by considering how money evolves in Ruritania. Carl Menger (1892, 250) showed that money, rather than being invented or adopted by legislative act, emerges as “the spontaneous outcome . . . of particular, individual efforts of members of society.”2 Menger’s method serves as a prototype for our look at events in Ruritania, which will show how complex banking procedures, devices, and institutions—including some still present in regulated and centralized banking systems—can also evolve spontaneously.
Early in Ruritania’s history persons relying upon barter offer goods in exchange only for other goods directly useful to them. The number of bargains struck this way is very small, owing to the infrequency of what Jevons (1882, 3) called a “double coincidence” of wants. In time, some frustrated Ruritanian barterer realizes that, by trading his goods or services for some more saleable good regardless of its use value to him, he can increase his chances for profitable exchange. Ruritania’s earliest media of exchange are therefore simply its most barterable goods. Other traders, noticing the gains achieved by persons using indirect exchange, emulate them, despite their lack of awareness of all the advantages from using a small number of exchange media (such as the fact that it may eventually lead to the emergence of a unified price system). This further enhances the saleability of the most widely accepted media. In time, the chasm separating these more saleable goods from all others grows wider and wider. This snowballing of saleability results in the spontaneous appearance of generally accepted media of exchange. Eventually, traders throughout Ruritania converge on some single good as their most generally accepted medium of exchange, i.e., money.
Historically cattle were often the most frequently exchanged commodity. Yet, owing to their lack of transportability and their nonuniformity, cows left much to be desired as a general medium of exchange (Burns 1927a, 286-88). Their chief contribution to the evolution of money seems to have been as a unit of account (Menger 1981, 263-66; Ridgeway 1892, 6-11). It was the discovery of methods for working metals which finally allowed money to displace barter on a widespread basis.3 So we assume that Ruritania’s first money is some metallic medium.
Like that of money itself the idea of coinage does not “flash out upon” Ruritania or in the mind of one of its rulers (Burns 1927a, 285). Rather, it is the unplanned result of Ruritanian merchants’ attempts to minimize the necessity for weighing and assessing amounts of crude commodity money (e.g., gold or silver) received in exchange. The earliest surviving historical coins—the famous electrum coins of Lydia—were, according to Ridgeway (1892, 203ff) probably in use throughout the Aegean before the reign of Pheidon of Argus, who is often credited as the inventor of coinage.4 The merchants of Lydia at first marked shapeless electrum blobs after assessing their value. A merchant recognizing his own mark or the mark of a fellow merchant could thereby avoid the trouble and cost of reassessment. Marking led to stamping or punching, which eventually gave way to specialists making coins in their modern form (Burns 1927a, 297ff). Techniques for covering coins entirely with type safeguarded them against clipping and sweating, assuring their weight as well as their quality (Burns 1927b, 59).
States monopolized their coinage early in history. But this does not mean that they were the best makers of coin or that coinage is a natural monopoly.5 Rather, state coinage monopolies were established by force. Once rulers had set up their own mints they prohibited private issues, making their coins both a symbol of their rule and a source of profits from shaving, clipping, and seignorage. By the end of the 7th century such motives had caused coinage to become a state function throughout the Greek world (Burns 1927b, chaps. 3 and 4; and 1927a, 308).
In Ruritania, however, since state interference is absent, coinage is entirely private. It includes various competing brands, with less reliable, more “diluted” coins first circulating at discount and eventually forced out of circulation entirely. This appears to contradict Gresham’s Law, which states that “bad money drives good money out of circulation.” Yet, properly understood, Gresham’s Law applies only where legal tender laws force the par acceptance of inferior coins.6 In contrast, Ruritania’s free market promotes the emergence of coins of standard weights and fineness, valued according to their bullion content plus a premium equal to the marginal cost of mintage.
The Development of Banks
Under Ruritania’s pure commodity-money regime traders who frequently undertake large or distant exchanges find it convenient to keep some of their coin (and bullion) with foreign-exchange brokers who can then settle debts by means of less costly ledger-account transfers.7 Money-transfer services also develop in connection with deposits initially made, not for the purpose of trade, but for safekeeping. Wealthy Ruritanians who are not active in commerce begin placing temporarily idle sums of commodity money in the strongboxes of bill brokers, moneychangers, scriveners, goldsmiths, mintmasters, and other tradespeople accustomed to having and protecting valuable property and with a reputation for trustworthiness.8 Coin and bullion thus lodged for safekeeping must at first be physically withdrawn by its owners for making payments. These payments may sometimes result in the redeposit of coin in the same vault from which it was withdrawn. This is especially likely in exchanges involving money changers and bill brokers. Such being the case, it is possible for more payments to be arranged, without any actual withdrawal of money, at the sight of the vault, or better still by simply notifying the vault’s custodian to make a transfer in his books. Such transfer banking was first practiced by English goldsmiths during the 17th century, when they began to keep a “running cash” account for the convenience of merchants and country gentlemen (De Roover, 183-84).
In transfer banking of this kind money on deposit is meant to be “warehoused” only. The custodian is not supposed to lend deposited money at interest, and receipts given by the “banker” for it are regular warehouse dockets.9 Thus, primitive Ruritanian bankers are bailees rather than debtors to their depositors, and their compensation comes in the form of depositors’ service payments.
Of course, Ruritania’s primitive bankers may also engage in lending; but their loans are originally made out of their personal wealth and revenues. The lending of depositors’ balances is a significant innovation: it taps a vast new source of loanable funds and fundamentally alters the relationship between Ruritanian bankers and their depositors. “The . . . bailee develops into the debtor of the depositor; and the depositor becomes an investor who loans his money . . . for a consideration” (Richards 1965, 223). Money “warehouse receipts” or bailee notes become IOUs or promissory notes, representing sums still called deposits but placed at the disposal of the banker to be reclaimed upon demand (ibid., 225).
W. R. Bisschop reports (1910, 50fn) that during the time of Charles II goldsmiths had already begun to function as savings-investment intermediaries: “they made no charge for their services . . . but any deposit made in any other shape than ornament was looked upon by them as a free loan. The cash left in their hands remained at call.” Soon afterwards, however, the practice of paying interest on deposits began. This led to a substantial increase in business: by 1672 the custom of depositing had become widespread among all ranks of people.10 It is significant that throughout this entire period, despite wars and internal disturbances, no goldsmith banker suspended payment (ibid.).
The ability of bankers, in Ruritania and elsewhere, to lend out depositors’ balances rests upon two important facts. The first is the fungibility of money, which makes it possible for depositors to be repaid in coin or bullion other than the coin or bullion that they originally handed to the banker. The second is the law of large numbers, which ensures a continuing (though perhaps volatile) supply of loanable funds even though single accounts may be withdrawn without advance notice. In his 1691 Discourse upon Trade, Sir Dudley North, an observer of developments in English banking, commented on the ability of the law of large numbers to render England’s effective money supply greater than what the existing stock of specie could alone account for:
The Merchants and Gentlemen keep their Money for the most part with Goldsmiths and Scriveners, and they, instead of having Ten Thousand Pounds in cash by them as their Accounts show they should have of other Mens Money, to be paid at sight, have seldom One Thousand in Specie; but depend upon a course of trade whereby Money comes in as fast as it is taken out.11
Assignability and Negotiability
Up to this point Ruritania’s most important banking procedures and devices have yet to emerge. Since purchases must still be made with actual coin, substantial savings remain locked up in circulation. Bank depositors, in order to satisfy changing needs for transactions balances, have to make frequent cash withdrawals from their balances. Though the withdrawals might largely offset one another, they still contribute to the banks’ need for precautionary commodity-money reserves. What is lacking is some negotiable alternative to standard IOUs which can pass easily in exchange from one person to another, replacing coin in transactions balances. Lacking also are efficient means for reassigning deposit credits represented, not by IOUs, but by ledger entries, which could also reduce the need for coin in circulation.
The development of such assignable and negotiable bank instruments proceeds through several steps. At first, deposited money (whether warehoused or entrusted to the banker for lending at interest) is assigned by the depositor to another party by oral transfer. This requires the presence of all three parties to the exchange or their attorneys,12 as standard money IOUs or promissory notes cannot be assigned by their owner without the banker acting as witness. The first important innovation, therefore, is a bank-issued promissory note transferable by endorsement; a parallel development is that of nonnegotiable checks for use in transferring deposits represented in book entries rather than by outstanding promissory notes. Nonnegotiable checks open the way to negotiable ones, while assignable promissory notes open the way to negotiable bank notes.13 What distinguishes the latter is that they are not assigned to any one in particular, but are instead made payable to the bearer on demand.
Thus Ruritania would evolve the presently known forms of inside money—redeemable bank notes and checkable deposits. With these forms of inside money at hand all that remains for Ruritania’s bankers to conceive is what Hartley Withers (1930, 24) called “the epoch-making notion” of giving inside money, not only to depositors of coin or bullion, but also to those who come to borrow it.
The use of inside money is not just convenient to bank customers. It also makes for greater banking profits, so that only the reluctance of Ruritania’s courts to enforce obligations represented by assigned or negotiable paper stands in its way.14 In England bearer notes were first recognized by the courts during the reign of Charles II, about the time when warehouse banking was beginning to give way to true banking (Richards 1966, 225). At first, the courts grudgingly approved the growing practice of repeated endorsement of promissory notes (ibid., 46). Then, after some controversy, fully negotiable notes were recognized by an Act of Parliament. In France, Holland, and Italy during the 16th century, merchants’ checks drawn in blank and circulated within limited groups may have cleared the way for the appearance of bank notes (Usher 1943, 189).
Benefits of Fiduciary Substitution
Aside from its immediate benefits to Ruritania’s bankers and their customers, the use of inside money has wider, social consequences. Obviously it reduces the demand for coin in circulation, while generating a much smaller increase in the demand for coin in bank reserves. The net fall in demand creates a surplus of coin and bullion, which Ruritania may export or employ in some nonmonetary use. The result is an increased fulfillment of Ruritania’s nonmonetary desires with no sacrifice of its monetary needs. This causes a fall in the value of money, which in turn “acts as a brake” on the production of commodity money and directs factors of production to more urgent purposes (Wicksell 1935, 124).
Of even greater significance than Ruritania’s one time savings from fiduciary substitution (the replacement of commodity money with unbacked inside money) is its continuing gain from using additional issues of fiduciary media to meet increased demands for money balances. By this means every increase in real money demand becomes a source of loanable funds to be invested by banks, whereas under a pure commodity-money regime an increase in money demand either leads to further investments in the production of commodity money, or, if the supply of commodity money is inelastic, to a permanent, general reduction in prices. The latter result involves the granting of a pure consumption loan by money holders to their contemporaries.15 Thus, fiduciary issues made in response to demands for increased money balances allow Ruritania to enjoy greater capitalistic production than it could under a pure commodity-money regime.16
According to Harry Miskimim (1979, 283-89), the growth of private credit made possible by fiduciary substitution in actual history began to aid economic progress as early as the 16th century. The benefits in more recent times from the use of fiduciary media in place of commodity money have been outlined by J. Carl Poindexter (1946, 137):
It is highly probable that the phenomenal industrial progress of the last century would have been greatly reduced had institutions not developed through which the monetary use of the limited supply of precious metals could be economized . . . The evolution of fractional reserve banking permitted the substitution of a highly convenient and more economical money medium whose supply could be expanded by a multiple of the available quantity of monetary gold . . . [H]istory and theory support the view that the real value of the expanding bank credit pyramid has actually borne a reasonably close secular relationship to the real value of voluntary savings of depositors. If this be the correct view, it also follows that the pyramiding of credit was the fortuitous factor by virtue of which secular deflation and involuntary dissaving were avoided or minimized.
Two questions must be asked in light of these potential benefits: First, what economic forces exist in Ruritania to sponsor more complete fiduciary substitution; and second, what forces exist to prevent excessive use of the power of fiduciary issue? The rest of this chapter concentrates on the first question, since it is relatively easy to show how bankers and other persons in Ruritania, left to pursue their own interests, are led to improve the acceptability of inside money and the efficiency of banking operations. This is the essence of the story of Ruritania’s organized note-exchange and clearinghouse institutions. The second question, concerning whether adequate forces exist to limit issues of bank money, is addressed at length in later chapters.
The progress of fiduciary substitution in Ruritania requires more complete use of inside money as well as more complete development of bank note and check clearing facilities that reduce the need for commodity money reserves. When inside money first emerges, although bank notes are less cumbersome than coin, and checkable deposits (in Ruritania) are both convenient for certain transactions and interest paying, some coin might still remain in circulation. Ruritanian consumers trust notes of local banks more than those of distant banks because they know more about the likelihood of local banks honoring their notes and also because they are more familiar with the appearance of these notes (and hence less prone to accept forgeries).17 It follows that the cost to a bank of building a reputation for its notes in some market is higher the further away the market is from where the notes are issued and redeemed. On the other hand, the building of a network of branch offices for more widespread note issue and redemption is limited by transportation and communication costs. Therefore, in the early stages of Ruritania’s free-banking development the par circulation of each bank’s notes is geographically relatively limited. Those holding the inside money of a local bank, but wishing to do business in distant towns, must either redeem some of their holdings for gold (and suffer the inconvenience of transporting coin), or suffer a loss in the value of their notes by taking them where they are accepted only at a discount, if at all.18 In general, every brand of inside money is at first used only for local transactions, with coin remaining in circulation alongside notes of like denomination. The continued use of coin for non-local exchange also forces banks to hold commodity-money reserves greater than those required by the transfer of inside money. This is because the withdrawal of commodity money for spending generates more volatile reserve outflows than does the spending of notes and deposits.
Given that Ruritania’s banks have limited resources to devote to confidence- or branch-building, how can their need for commodity money be further reduced? Can individuals benefit privately from their actions that promote more widespread use of inside money in place of coin, or is some form of collective action needed to achieve further economies by use of fiduciary media?
The answer is that profit opportunities exist to promote a more general use of particular inside monies. The discounting of notes outside the neighborhood of the issuing bank’s office provides an opportunity for arbitrage when the par value of notes exceeds the price at which they can be purchased for commodity money or local issues in a distant town, plus transaction and transportation costs. With the growth of interlocal trade, note brokers with specialized knowledge of distant banks can make a business, just as retail foreign currency brokers do today, of buying discounted non-local notes and transporting them to their par-circulation areas (or reselling them to travelers bound for those areas). Competition eventually reduces note discounts to the value of transaction and transportation costs, plus an amount reflecting redemption risk. In accepting the notes of unfamiliar banks at minimal commission rates, brokers unintentionally increase the general acceptability of all notes, promoting their use in place of commodity money.
So far it has been taken for granted that Ruritania’s free banks refuse to accept one another’s notes. This is not an unreasonable assumption, since the banks have as many reasons as individuals do to refuse unfamiliar and difficult-to-redeem notes. They also have a further reason, which is that by doing so they limit the acceptability of rival banks’ notes and enhance the demand for their own issues. To cite just one historical illustration of this, in Edinburgh in the 1770s the Bank of Scotland and the Royal Bank of Scotland, the two then existing chartered banks of issue, refused to accept the notes of unchartered, provincial banks of issue for a number of years (Checkland 1975, 126).
Nonetheless, note brokerage presents opportunities for profit to Ruritania’s bankers. Moreover, because they can issue their own notes (or deposit balances) to purchase “foreign” notes and therefore need not hold costly till money, banks can out-compete other brokers. Still another incentive exists for banks to accept rival notes: larger interest earnings. If a bank redeems notes it acquires sooner than other banks redeem the first bank’s notes issued in place of theirs, it can, in the interim, purchase and hold interest-earning assets. The resulting profit from “float” can be continually renewed. In other words, a bank’s earnings from replacing other notes with its own may be due, not just to profits from arbitrage, but also to enhanced loans and investments. If transaction and transportation costs and risk are low enough, competition for circulation reduces brokerage fees to zero, reflecting the elimination of profits from arbitrage. This leads Ruritanian banks to accept each other’s notes at par.
It is important to see that the development of par acceptance of notes does not require that Ruritania’s banks get together to explicitly agree on such a policy. It only takes a single bank acting without the consent or cooperation of other banks to nudge the rest toward par acceptance as a defensive measure on their part to maintain their reserves and circulation. This has also been the case historically. In New England at the beginning of the 19th century it was the Boston banks that gave the nudge that put the whole region, with its multitude of “country” banks lacking branches and with offices far removed from the city, on a par-acceptance basis.19 In Scotland it was the Royal Bank of Scotland which, when it opened for business in 1727, immediately began accepting at par notes from the Bank of Scotland, at that time its only rival (Checkland 1975). In both New England and Scotland established banks that were accepting each other’s notes at par sometimes refused to take the notes of newly entering banks. But they soon had to change their policies, because new banks that accepted their notes were draining their reserves, whereas the established banks were not offsetting this by engaging in the same practice in reverse.
In the long run, banks that accept other banks’ notes at par improve the market both for their own notes and, unintentionally, for the notes that they accept. Statistics from Boston illustrate this dramatically: from 1824 to 1833 the note circulation of Boston banks increased 57 percent, but the Boston circulation of country banks increased 148 percent, despite the Boston banks’ intent to drive the country banks out of business.20
The rivalrous behavior of banks in Ruritania causes inside money to become even more attractive to hold relative to commodity money. Because notes from one town come to be accepted in a distant town at par, there is little reason to lug around commodity money any more. This, too, can be seen in history. As par note acceptance developed during the 19th century in Scotland, Canada, and New England—places where note issue was least restricted—gold virtually disappeared from circulation.21 In England and in the rest of the United States where banking (and note issue in particular) were less free, considerable amounts of gold remained in circulation.
Even complete displacement of commodity money in circulation by inside money does not, however, necessarily mean increased fiduciary substitution. Commodity money, formerly used in circulation to settle exchanges outside Ruritania’s banks, might now be used to settle clearings among them. To really economize on commodity money rival banks have to exchange notes frequently enough to allow their mutual obligations to be offset. Then only net clearings, rather than gross clearings, need to be settled in commodity money. Thus banks can take further advantage of the law of large numbers, and more commodity money becomes available for nonmonetary uses.
Initially there might not be much movement towards rationalization of note exchange. That Ruritania’s banks accept one another’s notes at par does not mean that they exchange notes regularly. In Scotland par acceptance without regular note exchange was present before 1771. During that period, banks’ sought to bankrupt their rivals by “note dueling”—aggressively buying large amounts of their rival’s notes and presenting them for redemption all at once.22 For a bank to stay solvent during such raids it has to keep substantial reserves, so that its contribution to the process of fiduciary substitution is small. Charles Munn reports that one Scottish provincial bank at one point kept reserves equal to 61.2 percent of its inside-money liabilities to protect itself against raids by its rivals. More typically, reserves during Scotland’s note-dueling era were in the neighborhood of 10 percent of total liabilities (Munn 1981, 23-24). Yet even this smaller figure contrasts greatly with reserve ratios of around 2 percent which were typical under the Scottish free banking system after note clearings became routine (ibid., 141).
Though it does not catch on immediately in Ruritania, regular note exchange has advantages that guarantee that it will eventually be adopted, as it was in every historical instance of relatively unregulated plural note issue. Note dueling ceases to be advantageous to any bank as all of them learn how to protect themselves in response to it by holding large reserves. Because of this, Ruritania’s banks soon find it more convenient to accept their rival’s notes only as they are brought to them for deposit or exchange. They do not continue actively to buy notes in the marketplace since this is both costly and unreliable as a means for expanding circulation. Also, instead of being accumulated in large sums, rivals’ notes are immediately returned at once to their issuers for redemption in commodity-money reserves, which can be profitably employed. Finally, as banks in Ruritania realize the savings to be had by offsetting note debits with one another, they may formally agree to engage in regular note exchange and to refrain from purchasing rivals’ notes except as they are brought to them for deposit or exchange.23
Suppose Ruritania has three banks, A, B, and C. A has $20,000 of B’s notes, B has $20,000 of C’s notes, and C has $10,000 of A’s notes.24 If they settle their obligations bilaterally, they need to have $20,000 to $40,000 of commodity-money reserves on hand among them, depending on the chronological sequence of their exchange.25 On the other hand, if they settle their balances multilaterally, they need only $10,000 of reserves among them: A’s net balance to B and C combined is +$10,000; B’s net balance to A and C combined is $0; and C’s net balance to A and B combined is -$10,000. Hence all three balances can be settled by a transfer of $10,000 from C to A. Apart from reducing reserve needs, multilateral clearing also allows savings in operating costs by allowing all debts to be settled in one place rather than in numerous, scattered places.
Such advantages impel Ruritania’s banks to establish clearinghouses—organizations devoted to the multilateral settlement of bank obligations—which help to further unify the banking system. The clearinghouses do not, however, have to spring into existence full blown. Instead, they may gradually evolve from simpler note-exchange arrangements. The history of the earliest and most well known clearinghouses, in London, Edinburgh, and New York, illustrates this. All of them were products of human action but not of human design or, as Adam Smith would say, they were instances of the invisible hand at work.
Geographical, economic, and legislative differences in each of those cities affected the shape their clearinghouses took. Nevertheless the clearinghouses shared a common pattern of development. The circumstances leading to the establishment of the New York Clearinghouse in 1853, as reported by Gibbons (1858, 292-93), were typical. The first improvements were due to note porters anxious to save shoe leather:
The porters crossed and re-crossed each other’s footsteps constantly; they often met in companies of five or six at the same counter, and retarded each other, and they were fortunate to reach their respective banks at the end of one or two hours. This threw the counting of the exchanges into the middle and after part of the day, when the other business of the bank was becoming urgent.
The porters finally hit upon the idea of meeting at a convenient spot, outside of any bank, to combine and reconcile their claims.
Approximately three-quarters of a century earlier nearly identical events took place in London, though here checks rather than notes were being exchanged, the Bank of England being the only note-issuing bank in the region. As later happened in New York, the porters first traveled among the banks to settle accounts bilaterally:
The majority of them belonged to offices which were situated in Lombard Street. . . . Soon . . . occasional encounters developed into daily meetings at a certain fixed place. At length the bankers themselves resolved to organize these meetings on a regular basis in a room specially reserved for this purpose [Bisschop 1910, 160].
And so the London clearinghouse came into being.26
Whenever a clearinghouse is set up in Ruritania, all banks within the region feel compelled to join it due to the advantages membership brings. However, banks suspected of being ill managed or unsound may be denied membership. Also, where there are several clearinghouses within a region that clear with each other, a bank would only need to join one of them to partake of the full advantages of multilateral clearing.27
The principal purpose of Ruritania’s clearinghouses is the economical exchange and settlement of banks’ obligations to each other. Once established for this purpose, however, the clearinghouses serve a variety of other uses, becoming “instruments for united action among the banks in ways that did not exist even in the imagination of those who were instrumental in [their] inception” (Cannon 1908, 97).
One of the more common tasks the clearinghouses take on is to serve as a credit information bureaus for their members. By pooling their records, Ruritania’s banks can discover whether people have had bad debts in the past or are presently overextended to other banks. This allows them to take appropriate precautions (Cannon 1900, 135).28 Through a clearinghouse banks can also share information concerning forgeries, bounced checks, and the like. Clearinghouses may also conduct independent audits of member banks to assure each member bank that the others are worthy clearing partners. For example, beginning in 1884 the New York Clearinghouse carried out comprehensive audits to determine its members’ financial condition (ibid.). Others, such as the Suffolk Bank and the Edinburgh clearinghouse, took their bearings mainly from the trends of members’ clearing balances and the traditional canons of sound banking practice. Those two clearinghouses enjoyed such high repute that to be taken off their lists of members in good standing was a black mark for the offending bank (Trivoli 1979, 20; Graham 1911, 59).
Another task Ruritania’s clearinghouses may undertake is to set common reserve ratios, interest rates, exchange rates, and fee schedules for their members. However, inasmuch as common rates on reserve ratios are not consistent with profit-maximizing strategies of individual banks, they tend to break down. Those that prevail are in most instances merely formally agreed upon confirmations of results that rivalrous competition would also have established informally but no less firmly. There is an excellent example of this from Scottish experience. In order to formalize certain banking practices, the Edinburgh banks set up a General Managers’ Committee in 1828. In 1836 the Glasgow banks joined the committee, which then represented the preponderance of Scottish banks in number as well as in total assets. Though not itself a clearinghouse association, the committee had much the same membership as the Edinburgh clearinghouse. If ever there was an opportunity to cartellize Scottish banking, this was it. Yet in spite of repeated attempts the banks could not agree on a common reserve ratio. Though they gave the appearance of having agreed on common interest rates, in reality the General Manager’s Committee was full of dissent even on this matter. The moment any course other than that recommended by the Committee appeared profitable to any of its members, interest-rate agreements would collapse (Checkland 1975, 391).
Perhaps the most interesting of all roles clearinghouses in Ruritania may perform is aiding members in times of crisis.29 If a bank or group of banks is temporarily unable to pay its clearing balances, or should it experience a run on its commodity-money reserves, a clearinghouse can serve as a medium through which the troubled bank borrows from more liquid banks. It provides the framework for an intermittent, short-term credit market similar to the continuous Federal Funds market to which reserve-deficient banks resort in the present U.S. banking system.
Related to their role in assisting illiquid members is clearinghouses’ role as note issuers. This function has been exercised by historical clearinghouses where member banks have had their own rights to issue notes artificially restricted, preventing them from independently filling all of their customers’ requests for currency. It is not, however, a function that clearinghouses would be likely to perform in Ruritania, where bank-note issue is unregulated. This is not the place to discuss the causes and consequences of currency shortages, which are treated at length later in chapter 8. It will suffice to note that currency shortages occurred frequently in the United States during the 19th century, and that clearinghouses helped to fill the void caused by deficient note issues of the National banks.30
So far the most important of the unintended effects of Ruritania’s clearinghouses has not been mentioned. This is their ability to regulate strictly the issues of their members through the automatic mechanism of adverse clearings. Together with free note issue, the existence of an efficient clearing arrangement gives Ruritania’s banking system special money-supply properties, to be examined in detail in later chapters.
The Mature Free Banking System
We have now described the mature stage of Ruritania’s free banking system, insofar as consideration of self-interested actions allows us to predict its development. Historical evidence on industry structure from Scotland, Canada, Sweden, China, and elsewhere suggests that Ruritania would possess, not a natural monopoly in currency supply, but an industry consisting of numerous competing note-issuing banks, most having widespread branches, all of which are joined through one or more clearinghouses. In Scotland’s final year of free entry, 1844, there were 19 banks of issue. The largest four banks supplied 46.7 percent of the note circulation. These banks also had 363 branch offices, 43.5 percent of which were owned by the largest (measured by note issue) four banks (L. White 1984d, 37).
Banks in Ruritania issue inside money in the shape of paper notes and demand deposit accounts (debited either by check or electronically) transferred routinely at par. Each bank’s notes and tokens bear distinct brand-name identification marks and are issued in denominations most desired by the public. Because of computational costs involved in each transfer, interest is generally not paid on commonly used denominations of bank notes or tokens.31 Checkable accounts, however, pay competitive interest rates reflecting rates available on interest-earning assets issued outside the banking system.
Outside of Ruritania the most familiar kind of checkable bank accounts are demand deposits which have a predetermined payoff payable on demand. One important reason for their popularity is that, historically, debt contracts have been easier for depositors to monitor and enforce than equity contracts, which tie the account’s payoff to the performance of costly-to-observe asset portfolios. This predetermined payoff feature of demand deposits does, however, raise the possibility of insolvency and, consequently, of a run: bank depositors may fear that, by being late in line, they will receive less than a full payoff.
One way Ruritania’s banks can prevent runs is to advertise having large equity cushions, either on their books or off them, in the form of extended liability for bank shareholders.32 A second solution is to link checkability to equity or mutual-fund type accounts with post-determined, rather than predetermined, payoffs. Former obstacles to such accounts in actual history—such as asset-monitoring and enforcement costs—have eroded over time due to the emergence of easy to observe assets (namely: publicly traded securities) and may be presumed absent in Ruritania also. For a balance sheet without debt liabilities, insolvency is ruled out and the incentive to redeem ahead of other account holders is eliminated.
A Ruritanian bank that linked checkability to equity accounts would operate like a contemporary money-market mutual fund, except that it would be directly tied to the clearing system instead of having to clear through a deposit bank. Its optimal reserve holdings would be determined in the same way as those of a standard bank (see chapter 6).
The assets of Ruritania’s banks include short-term commercial paper, corporate and government bonds and loans on various types of collateral. The structure of asset portfolios cannot be predicted in detail without particular information on the assets available in Ruritania as a whole, except to say that Ruritania’s banks, like banks elsewhere, strive to maximize the present value of their interest earnings, net of operating and liquidity costs, discounted at risk-adjusted rates. The declining probability of larger liquidity needs, and the trade-off at the margin between liquidity and interest yield, suggest that the banks hold a spectrum of assets ranging from perfectly liquid reserves to highly liquid interest-earning investments (which serve as a “secondary reserve”), to less liquid, higher-earning assets. Because the focus in this book is on monetary arrangements, the only bank liabilities discussed are notes and checking accounts. Nonetheless, free banks would almost certainly diversify on the liability side by offering several kinds of time deposits as well as traveler’s checks. Some banks would probably also get involved in the production of bullion and token fractional coins, the issue of credit cards, and management of mutual funds. Such banks would fulfill the contemporary ideal of the “financial supermarket,” but with the additional feature of issuing bank notes.
In a mature free-banking system, such as Ruritania’s, commodity money seldom if ever appears in circulation, most of it (outside numismatic collections) having been offered to the banks in exchange for inside money. Some commodity money continues to be held by individual banks or clearinghouses so long as it remains the ultimate settlement asset among them. Since no statutory reserve requirements exist, reserves are held only to meet banks’ profit-maximizing liquidity needs, which vary according to the average size and variability of clearing balances to be settled after routine (e.g. daily) note and check exchanges.
The holding of reserve accounts at one or more clearinghouses results in significant savings in the use of commodity money.33 In the limit, if inter-clearinghouse settlements are made entirely with other assets (perhaps claims on a super-clearinghouse which itself holds negligible amounts of commodity money), and if the public is weaned completely from holding commodity money, the only active demand for the old-fashioned money commodity is nonmonetary: the flow supply formerly sent to the mints is devoted to industrial and other uses. Markets for these uses then determine the relative price of the money commodity. Nonetheless, the purchasing power of monetary instruments continues to be fixed by the holders’ contractual right (even if never exercised) to redeem them for physically specified quantities of commodity money. The special difficulty of meeting any significant redemption request or run on a bank in such a system can be contractually handled, as it was historically during note-dueling episodes, by invoking an “option clause” allowing the bank a specified amount of time to gather the needed commodity money while compensating the redeeming party for the delay. The clause need not (and, historically, did not) impair the par circulation of bank liabilities.
Our image of an unregulated banking in Ruritania differs significantly from visions presented in some recent literature on competitive payments systems. The Ruritanian system has assets fitting standard definitions of money. Its banks and clearinghouses are contractually obligated to provide at request high-powered reserve money. They also issue debt liabilities (inside money) with which payments are generally made. These features contrast sharply with the situation envisioned by Fischer Black and Eugene Fama, in which banks hold no reserve assets and the payments mechanism operates by transferring equities or mutual fund shares unlinked to any money.
In the evolution of Ruritania’s free banking system, bank reserves do not entirely disappear, since the existence of bank liabilities that are promises to pay continues to presuppose some more fundamental money that is the thing promised. Ruritanians forego actual redemption of promises, preferring to hold them instead of commodity money, so long as they believe that they will receive money if they ask for it. Banks, on the other hand, have a competitive incentive to redeem each other’s liabilities regularly. As long as net clearing balances are sometimes greater than zero, some kind of reserve, either commodity money itself or secondary reserves priced in terms of the commodity money unit of account, has to be held.
The scarcity of the money commodity, and the costliness of holding reserves, also serves to pin down Ruritania’s price level and to limit its stock of inside money. In a moneyless system, on the other hand, it is not clear what would be used to settle clearing balances. Hence, it is not clear what forces would limit the expansion of payment media or what would pin down the price level. Nor are these things clear, at the other extreme, in a model of multiple competing fiat monies.34
Our story suggests that a commodity-based money would persist in Ruritania because its supreme saleability is self reinforcing. This contradicts recent suppositions that complete deregulation would lead to the replacement of monetary exchange by a sophisticated form of barter.35 In a commodity-based money economy prices are stated in terms of a unit of the money commodity, so that the need to use an abstract unit of account does not arise as it does in a sophisticated barter setting.36 Even if actual commodity money disappears entirely from reserves and circulation, media of exchange are not divorced from the commodity unit of account. Rather, they continue to be linked to it by redeemability contracts. Nor is renunciation of commodity-redemption obligations compelled by economization of reserves, as Warren Woolsey predicts (1985). There is, therefore, no reason to expect deregulation to lead to the spontaneous emergence of a multi-commodity monetary standard or of any pure fiat monetary standard, as is suggested in works by Robert Hall, Warren Woolsey, Benjamin Klein, and F. A. Hayek.37 In few words, unregulated banking is likely to be far less radically unconventional, and much more like existing financial arrangements, than recent writings on the subject suggest.
One important contemporary financial institution is nonetheless absent from the Ruritanian system: to wit, the central bank. This is because market forces at work in Ruritania do not lead to the natural emergence of a monopoly bank of issue capable of willfully manipulating the money supply.38 As was shown in chapter 1, the historical emergence of central banks was typically a consequence of monopoly privileges respecting note issue being conferred on some state-owned or state-chartered bank. Legal restrictions imposed on commercial banks directly or indirectly promoting unit (instead of branch) banking have also played an important part in the historical emergence of central banks. Where legislation did not inhibit the growth of plural note issue and branch banking, as in Scotland, Sweden and Canada in the 19th century, there was not any movement toward monopolized note issue or toward spontaneous emergence of a central bank.39
A free bank adds to its gross income by enlarging its holdings of interest-earning assets. But it can only do this by either attracting more depositors and note holders or by losing some of its reserves.40 Thus its costs include not only operating costs but also liquidity costs and costs that arise from its efforts to maintain a demand for its liabilities, such as interest payments to depositors. Assuming that these costs—liquidity costs in particular—are increasing at the margin, there is a limit to the bank’s accumulations of interest-earning assets and hence to its overall size. Individual free banks compete for shares in the market for checkable deposit accounts and currency just as commercial banks today compete for shares of deposits alone.41
Assuming that free-bank liability issues run up against increasing marginal costs (an assumption to be defended in the next chapter), the conditions for long-run equilibrium of a free banking industry can be stated. As the public holds only inside money, with commodity money used only in bank reserves to settle clearing balances, these conditions are as follows: First, the demand for reserves and the available stock of commodity money must be equal. Second, the real supply of inside money must be equal to the real demand for it. Once the first (reserve-equilibrium) condition is met, the tendency is for any disequilibrium in the money supply to be corrected by adjustments in the nominal supply of inside money. An excess supply increases, and an excess demand reduces, the liquidity requirements (reserve demand) of the system. This is shown in chapters 5 and 6 below. On the other hand, if the reserve-equilibrium condition is not satisfied, the system is still immature. An excess supply of reserves then causes an expansion of the supply of inside money. If this leads to an excess supply of inside money, it will promote an increase in both reserve demand and prices, causing both the nominal demand for money and the demand for reserves to rise.
There must be one price level at which both equilibrium conditions are met. When this price level is achieved, the system is in a long-run equilibrium. For the sake of simplicity, the analysis that follows starts with a free banking system (similar to Ruritania’s) in long-run equilibrium and assumes an unchanging supply of bank reserves. It may be thought of as involving a closed banking system in which production of commodity money is limited by rising average costs.
Thus we have our hypothetical free banking system, painted in bold brush strokes that permit us to regard it as typical. Though much detail is lacking in this picture, it will, in due course and with some filling in of additional details as we proceed, allow us to derive far-reaching theoretical results.
Free Banking and Monetary Equilibrium
[1.] See, for example, Black (1970); Fama (1980, 1983); Greenfield and Yeager (1983); Hall (1982); Hayek (1978); Wallace (1983); L. White (1984c); Woolsey (1984); and Yeager (1983, 1985).
[2.] The same view appears in Carlisle (1901, 5); and Ridgeway (1892, 47). Ridgeway opines that “the doctrine of primal convention with regard to the use of any one particular article as a medium of exchange is just as false as the old belief in an original convention at the first beginning of Language and Law.”
[3.] On some alleged, non-metallic monies of primitive peoples see Temple (1899). According to Jacques Melitz (1974, 95), one must be skeptical concerning accounts of primitive monies, including especially non-metallic monies (with the exception of cowries in China), since many of these are based on loose definitions. For example, the Yap stones of Melanesia could hardly be described as serving as a general medium of exchange. An exceptionally sedate review of anthropological findings in this area is Quiggin (1963).
The most successful of the metallic monies—gold—appeared in many places as the “direct successor” of the cow barter-unit (see Burns 1927a, 288). According to Ridgeway (1892, 133), the earliest gold coins, found throughout the ancient world from Central Asia to the Atlantic, contain approximately 130-135 grains of gold, which was the approximate value in gold of a cow. “This uniformity of the gold unit,” Ridgeway argues, “is due to the fact that in all the various countries where we have found it, it originally represented the value in gold of a cow, the universal unit of barter in the same regions.” Many early gold coins also bore the image of a cow or ox. Carlisle observes (1901, 30) that “In all such cases, for the transition of meaning [from the barter-unit to a money unit of account] to take place at all, there must have been a ‘bimetallic’ period . . . in which the old unit and the new, as far as value was concerned, meant one and the same thing.”
[4.] In contrast to Ridgeway, Historian J. B. Bury (1902, 116) gives credit for the invention of coinage to Gyges (687-652 bc)—another (Lydian) king who, according to A. R. Burns (1927b, 83), was responsible not for the invention of coinage, but for the establishment of a state coinage monopoly.
[5.] According to Burns (1927b, 78), “There is no evidence that, as a general rule, the officers of the royal or public treasury, by their greater honesty, made the royal or civic seal a mark more reliable than the seals of the [merchants].”
[6.] On Gresham’s Law and its proper interpretation see Rothbard (1970, 783-84).
[7.] Transactions of this kind marked the earliest beginnings of banking in 12th-century Genoa and also (on a smaller scale) at medieval trade fairs in Champagne. See Usher (1943); De Roover (1974); and Lopez (1979). In time there were “transfer banks,” as De Roover (1974, 184), calls them, in all the major European trading centers.
[8.] In England, scriveners were the earliest pioneers in the banking trade; in Stuart times they were almost entirely displaced by goldsmith-bankers. The confiscation by Charles I of gold deposited for safekeeping at the Royal Mint ended that institution’s participation in the process of banking development. Nevertheless, one may conjecture that, had they existed, private mints would have been logical sites for the undertaking of banking activities.
[9.] The same holds for the receipts given by bill-brokers to their trader-depositors, although ownership of these may be subject to more routine transfer than occurs at first at other “warehouse banks.”
[10.] “It is obvious that . . . the advantage of being able to place money at safe custody and at interest, while at the same time retaining it within comparatively easy reach . . . would be so palatable as to attract custom from many quarters” (Powell 1966, 56-57).
[11.] Quoted in Bisschop (1910, 18).
[12.] According to Usher (1943, 7-8) oral-transfers continued to be the norm in Europe “long after the bill of exchange was well established” whereas “the use of checks and orders of payment seems to have been incidental and sporadic before 1500.”
[13.] The use of nonnegotiable checks, although it allows customers to avoid some visits to the bank, does not allow banks to reduce their reserve needs by extending the scope of clearing transactions. “Each check would have to become the basis of specific journal and ledger entries so that to the bookkeeper it made no difference whether the entry originated in an oral order or a written order. [Negotiable checks] made it possible to establish clearinghouses in which all transactions among a group of banks were liquidated by drawing up general debit and credit balances between each bank and the clearinghouse” (ibid., 23).
[14.] “There can be no true note until the law is willing to recognize the right of the bearer to sue in his own name without any supplementary proof of bona fide possession” (ibid., 177).
[15.] “Anyone who saves a part of his income [by hoarding] to that extent exercises a depressing influence on prices, even though it may be infinitesimal as regards each individual. Other individuals thereby obtain more for their money; in other words they divide among themselves that part of consumption which is renounced by those who save” (Wicksell 1935, 11-12).
[16.] Ibid. See also Marget (1926, 275-76) and (on the importance of free note issue in particular) Conant (1905, chap. 6).
An author who denies that fractional-reserve banking can be beneficial is Lloyd W. Mints, in Monetary Policy for a Competitive Society (1950). Mints writes (p. 5) that fractional-reserve banking institutions are “at the best unnecessary” and that (p. 7) “they constitute no positively desirable element in the economy.”
[17.] Much of what is said here about notes applies also to checkable deposit credits. For expository convenience, and also because unrestricted note issue is a defining feature of free banking, I will sometimes refer explicitly to bank notes only.
[18.] To keep notes from banks of each locality dealt with on hand would probably be prohibitively costly in terms of foregone interest.
[19.] See Trivoli (1979).
[20.] See Lake (1947, 188), and also Trivoli (1979, 10-12, 17).
[21.] Small amounts of gold coin continued to be used in these places because of restrictions upon the issue of “token” coin and of small denomination notes. In an entirely free system such restrictions would not exist.
[22.] See Leslie (1950, 8-9). On note-dueling in the Suffolk system see Magee (1923, 440-44).
[23.] This result is contrary to Eugene Fama’s suggestion (1983, 19) that note-dueling will persist indefinitely. It is an example of the “tit for tat” strategy, as discussed by Robert Axelrod (1984), proving dominant in a repeated-game setting.
[24.] The “$” symbol is used for expository convenience only. $1 here is some fixed amount of commodity money.
[25.] Even $40,000 is a savings compared to the $50,000 of reserves that would be needed if even bilateral offsetting of debts did not take place.
[26.] For an account of the origins of the Suffolk clearing system, which emerged under rather special circumstances, see Trivoli (1979).
[27.] There is an example of this in Boston banking history: in 1858 the Bank of Mutual Redemption was set up to compete with the Suffolk Bank as a note-clearing agency. The Suffolk Bank first resorted to note-dueling to quash its rival, but later became discouraged and largely gave up the business of clearing country bank notes. In response, the other Boston banks established a new organization, the New England Sorting House, to take the Suffolk’s place. The Sorting House accepted the notes of Bank of Mutual Redemption members, which the Suffolk Bank had refused, and this allowed it and the Bank of Mutual Redemption to clear with each other.
[28.] A later edition of Cannon’s book was published in 1910. This is the most comprehensive work on the subject of clearinghouses.
[29.] “The clearinghouse, which was begun simply as a labor-saving device, has united the banking interests in various communities in closer bands of sympathy and union and has developed into a marvelous instrumentality for the protection of the community from the evil effects of panics and of bad banking” Cannon (1910, 24).
[30.] See Cannon (1908, 99-112); Andrew (1908); R. H. Timberlake, Jr. (1984); and Gary Gorton (1985c).
[31.] This contradicts Neil Wallace’s hypothesis that all currency would bear interest under laissez-faire. See L. White (1984d, 8-9 and 1987).
[32.] The possibility of bank runs and ways free banks might deal with them is discussed further in the 4th section of chapter 9.
[33.] Though it would probably be more economical for banks to hold reserve accounts with their clearinghouse, in later chapters it will sometimes be assumed that each bank holds its own commodity-money reserves. Though less realistic, this assumption offers advantages in conceptual clarity that will aid readers’ understanding of free-banking money supply processes, while leading to the same conclusions as would be reached using more realistic assumptions.
[34.] Bart Taub (1985, 195-208) has shown that a dynamic inconsistency facing issuers in Klein’s (1974) model will lead them to hyperinflate.
[35.] See for example Black (1970); Fama (1980); Greenfield and Yeager (1983); and Yeager (1985). To be sure, Greenfield and Yeager recognize that their system would be unlikely to emerge without deliberate action by government, especially given a government-dominated monetary system as the starting point.
[36.] This point is emphasized in L. White (1984b). See also O’Driscoll (1985); Kevin D. Hoover (1985); and Bennett T. McCallum (1984).
[37.] See their works cited in note 1, this chapter.
[38.] In contradiction to Goodhart (1984). On the question of natural monopoly see below, chapter 10.
[39.] We have seen that reserves do tend to centralize, on the other hand, in the clearinghouses. And clearinghouses may take on functions that are today associated with national central banks: holding reserves for clearing purposes, establishing and policing safety and soundness standards for member banks, and managing panics should they arise. But these functional similarities should not be taken to indicate that clearinghouses have (or would have) freely evolved into central banks. The similarities instead reflect the pre-emption of clearinghouse functions by legally privileged banks or, particularly in the founding of the Federal Reserve System, the deliberate nationalization of clearinghouse functions. Central banks have emerged from legislation contravening, not complementing, spontaneous market developments. See Gorton (1985c, 277 and 283; and 1985b, 274); and Timberlake (1984).
[40.] To simplify discussion, the possibility of a bank’s adding to its capital or net worth is not considered.
[41.] A fuller discussion of conditions determining a free bank’s optimal size is presented in L. White (1984d, 3-9).