The Online Library of Liberty

A project of Liberty Fund, Inc.

• Preface
• Foreword
• Part One: Setting the Stage
• 1: Overview
• 2: The Evolution of a Free Banking System
• Part Two: Free Banking and Monetary Equilibrium
• 3: Credit Expansion With Constant Money Demand
• 4: Monetary Equilibrium
• 5: Changes In the Demand For Inside Money
• 6: Economic Reserve Requirements
• Part Three: Free Banking Versus Central Banking
• 7: The Dilemma of Central Banking
• 8: The Supply of Currency
• 9: Stability and Efficiency
• 10: Miscellaneous Criticisms of Free Banking
• 11: Free Banking and Monetary Reform
• Conclusion
• Notes

PART TWO

Free Banking and Monetary Equilibrium

3

Credit Expansion with Constant Money Demand

The previous chapter ended with a description of long-run equilibrium in a free banking industry. It assumed that note issue and deposit granting under free banking involve increasing marginal costs, especially liquidity costs. This chapter attempts to justify this assumption, showing that it applies only to free banking. Throughout the chapter it is assumed that the public’s total demand for inside-money balances, and the division of this demand between the two forms of inside money (currency and demand deposits) are constant. The effects of changes in the demand for inside money balances will be examined later, in chapter 5; those of changes in the public’s desired currency-deposit mixture will be the subject of chapter 8.

The Rule of Excess Reserves

A well-known principle of money and banking is that, in a system with monopolized currency supply, an individual bank cannot increase its loans and investments unless it has reserves in excess of what it needs to meet currency demands of its customers as well as clearing balances it owes to other banks.1 If a bank has reserves in excess of its liquidity needs and there are no statutory reserve requirements it can expand credit by the amount of the excess, but no more. This “rule of excess reserves” assumes that bank borrowers generally secure loans from banks only when they have purchases to make, implying that the demand for inside money balances does not increase. When a bank makes a loan to a client, the client writes checks against the new balance made out to people who are mainly clients of other banks. The checks are soon deposited, and the banks that acquire the checks waste no time returning them for payment to the bank on which they were drawn. Consequently, the bank that expands credit suffers a clearing drain practically equal to the amount of new credit it grants: its reserves are reduced “just as effectively as if each individual borrower had elected to take cash in the first place” (Rodkey 1928, 41). If the bank lacks excess reserves at least equal to this drain, it has to borrow emergency reserves. Otherwise it has to liquidate some of its investments or contract its loans to restore its reserve/liability ratio to a sustainable level.

The rule of excess reserves is usually applied to the circumstance of a small bank with many, equally small competitors. Nonetheless the rule also applies to a small group of banks so long as the demand for inside money is stationary. Consider a system of five banks, each with an equal amount of fully employed reserves and having equal shares of the market for deposit balances. Suppose Bank A grants a loan of $50,000, a net increase in its outstanding liabilities, to John Smith. On average, four-fifths of the checks drawn by Smith on the new balance will be paid to persons having deposits at Banks B through E. This causes Bank A to suffer an immediate clearing drain of $40,000. The other checks end up with persons who are depositors at Bank A, so that the bank is temporarily spared from a further reserve drain of $10,000. However, the persons who temporarily add $10,000 to their accounts at Bank A also, by assumption, do not intend to increase their average money holdings. Therefore, in the next round of spending they write checks for $10,000 (beyond what they would usually spend) to keep their balances at their original levels. Once again Bank A suffers an immediate clearing drain equal to four-fifths of the amount withdrawn, or $8,000, and so on. The schedule of clearing losses would therefore be as follows:

After five rounds of spending Bank A suffers clearing drains equal to all save $16 of its initial expansion. Assuming that each round of spending and clearing takes five days, this loss is suffered during a twenty-five day period, which is almost certainly less than the time required for the maturation of the loan on which the credit was issued. In any event, the greater part of any reserve drain occurs in the first few spending rounds and hence well before an expanding bank can expect to recover any part of its latest loans or investments.2

The rule of excess reserves also assumes that banks face a determinate demand schedule for their deposit balances (a schedule of quantity demanded for various deposit rates of interest), so that they cannot profit from expansionary policies that increase their loans while simultaneously increasing the public’s demand to hold deposit liabilities. The assumption is justified because, although an expanding bank might offset its reserve losses by attracting an equal sum of new deposits, it can only do so (other things being equal) by raising its marginal deposit rate of interest. On the other hand, the bank can only expand credit by marginally lowering the interest rates it charges to borrowers (or by buying assets with lower risk-adjusted interest yields). It follows that, in its attempts to generate deposit demand sufficient to cover reserve drains from its new loans, an expanding bank, operating in a competitive environment, would incur interest and operating expenses in excess of what the new loans would themselves bring in. Then, although the bank does not lose reserves to its rivals, it still suffers a reduction of net revenues, so that its expansion is unprofitable.3

Finally, the rule of excess reserves requires that, although one or a few banks may try to expand credit beyond their customers’ willingness to hold deposit credits, other banks are not induced by this to engage in sympathetic overexpansion: if all banks expand in unison their behavior will not be checked by individual banks’ suffering net clearing debits. It is easily shown, though, that profit-maximizing banks are not likely to overexpand in sympathy with their overexpanding rivals. If overexpansion by one bank or group of banks has any effect at all on conditions influencing other banks it is the reduction of the rate of interest on new loans, which makes further lending by them seem less desirable.4

The rule of excess reserves demonstrates clearly that, if all deposit banks in a system are fully loaned up, aggregate credit expansion can only take place if the banks get additional reserves from outside their own ranks:

If new reserves are added to the system each bank is able to expand, in the first instance, by the amount of new reserves it receives. Loan expansion for the system as a whole, however, will be several times greater than the combined first-round expansion of the first banks to receive the new reserves. “The new reserve, split into small fragments, becomes dispersed among the banks of the system. Through the process of dispersion it comes to constitute the basis of a manifold loan expansion” (Phillips 1920, 40). Expansion continues until all of the new reserve media is absorbed in the economically required balances of the banks (see below, chapter 6).5

The Principle of Adverse Clearings

There is nothing controversial about the rule of excess reserves, applied to a conventional system of deposit banks. What is controversial is whether a similar rule restricts the granting of credit in the form of competitively-issued bank notes. The generalization of the rule of excess reserves involved here will be referred to as the “principle of adverse clearings.” Early upholders of this principle include Sir Henry Parnell (1827), Lord Peter King (1804), and G. Poulett Scrope (1832). More recently, Ludwig von Mises has stated the principle as follows:

Still more recent is Hayek’s allusion to the principle in his Denationalisation of Money (1978, 59):

According to Alex N. McLeod (1984, 202), adverse clearings “not only limit the total note issue” in a system of competing banks of issue but also “limit the share [of circulation] accruing to each individual bank. A disproportionate expansion of one bank’s issues will deplete its reserves just as surely as if its competitors returned them directly, though perhaps somewhat more sluggishly.”

To prove that the principle of adverse clearings is valid, one must demonstrate that the rule of excess reserves applies to lending and investment of competitively issued notes. The rule assumes that banks extend credit in the shape of checkable deposits. It also assumes that there is a determinate demand for deposit balances, that checks are drawn against unwanted balances, and that these checks rapidly find their way to other banks and thence to their issuers for collection, either directly or through a clearinghouse. The principle of adverse clearings rests on similar assumptions. These are (a) that the total demand for bank notes is determinate, and that surplus notes are parted with; (b) that most of these surplus notes end up in possession of banks other than their own issuers; (c) that banks return their rivals’ notes for redemption; and (d) that a preponderance of the notes that enter the clearing mechanism following overexpansion by any bank will be notes of the overexpanding bank. This implies that the division of the public’s demand for note balances among the issues of various banks is also determinate.

Let us examine these assumptions one by one. The first is straightforward. A note-issuing bank faces a market where the total demand for the combined note balances of issuing banks is given in the short run. This means that any note issue not in response to an increased demand for note balances causes an aggregate excess supply of notes. That surplus notes when spent eventually find their way into possession of rival banks (assumption b) is less obvious because, unlike checks, notes are not necessarily cashed or deposited immediately upon their receipt. They might be used to make additional purchases and so remain in circulation for an extended period. Still the assumption is justified. Consider the typical fate of a bank note not wanted in anyone’s average holdings. If the note is not immediately deposited at some bank, it will probably be used to purchase goods from a retail merchant.7 Merchants receive much more money in the form of notes in the course of a typical business day or week than they require for making change. They bring the excess notes to their banks in exchange for checkable deposits, a form of inside money more useful than notes for paying rent and insurance and making wholesale purchases, and one that can earn interest as well. Thus the majority of notes received by retail merchants is deposited: the deposit accounts of merchants act as “note filtering” devices, helping surplus notes find their way rapidly into banks that are (for the same reason as in the case of checks) mainly not the banks that created the note surplus to begin with. The farther a note travels up the ladder of production (from retail tradesmen to distributors to manufacturers of higher-order goods), the greater its chances of being exchanged for a deposit credit, and the greater its chances of entering the clearing system.

Assumption (c) is also not difficult to defend. Notes received by banks other than those responsible for their issue are immediately sent for redemption directly or through a clearinghouse. This is the strategy most consistent with the maximum profits and safety of the recipient banks. Some reasons for this were given in the previous chapter. By returning its rivals’ notes a bank gives up only assets (notes of other banks) that do not earn interest, in return for which it receives either its own notes held by other banks (which reduces its outstanding liabilities that could be employed to drain off its reserves) or, alternatively, more liquid and risk-free commodity money.

Note-Brand Discrimination

Hence, if, starting in a situation where no excess demand for bank-note balances exists, some bank adds $100,000 of notes to the total quantity of notes in circulation, the result would be an excess supply of notes of $100,000, which amount (if not redeemed or deposited directly) would be added to the stream of expenditure and income. As a consequence $100,000 of surplus notes would eventually enter the clearing mechanism, to be sent to their issuers for redemption. This brings us to assumption (d). Is it reasonable to assume that, of the $100,000 of notes returned for redemption, a preponderance will belong to the bank that caused the note surplus to begin with? If the banks initially had (on average) zero net clearing balances with one another, how would net balances due be altered by the additional $100,000 of clearing debits and credits? Would the expanding bank suffer adverse clearings after creating a note surplus, or would it emerge unscathed, even victorious, because other banks are forced to share the burden of a note reflux?

The answer depends on which notes are spent off when consumers first find themselves holding more notes than they want, and this in turn depends on how strong the note-brand preferences of consumers are. If consumers practice 100 percent note-brand discrimination—that is, if besides knowing the total quantity of notes they desire they have strict preferences about what brands or mixture of brands they will hold—then any bank that causes an aggregate note surplus in the face of unchanged preferences cannot expect consumers to substitute its notes for notes of other banks in their holdings. Such being the case, the expanding bank would bear the full burden of clearing debits caused by the note surplus it creates, just as if it had created a surplus of checkable deposits.

On the other hand, if consumers do not at all discriminate among note brands they are not likely to select for spending the notes of a bank that causes an excess aggregate supply of notes. Then the principle of adverse clearings would not be valid and note issue, from the perspective of the individual bank, would not be subject to increasing marginal liquidity costs. This means that there might be a tendency for banks to try to out-issue their competitors. Such “predatory overexpansion” could continually push the system as a whole beyond sustainable limits of expansion as determined by the supply of reserves. (On the demand for precautionary reserves as an ultimate check on expansion see below, chapter 6.)

Notice that there is no reason to doubt that consumers are discriminating when it comes to checkable deposit accounts. Were this not so, someone who receives a check written on any bank would be as likely to deposit it into that bank as into any other. If his total accounts became excessive, he might withdraw some part of each (or withdraw from one account chosen at random). The result would be that people would on average have accounts at scores of banks. Obviously such is not the case.

That consumers would discriminate among note brands is not so obvious, though. The inconvenience of holding an everchanging mixture of bank notes is not so apparent as is the inconvenience of holding deposit accounts at numerous, continually changing, banks. If all note brands are equally acceptable to merchants and to banks, why should consumers care which brand they use?

The answer is that, although many note brands may be equally receivable, people may still prefer to hold on to particular brands. Banks, like manufacturers of light bulbs, razor blades, and gasoline, have every reason to establish brandname reputations for their products so that individuals do not, at least in their holding behavior, treat all brands as equals.8 A bank profits if note holders discriminate in favor of its issues since their holdings of its notes constitute an interest-free loan to it. Since, however, some may doubt that this is so, the reasonableness of assuming note-brand discrimination can be more decisively demonstrated by showing the counterfactual implications of nondiscrimination among bank notes. Under nondiscrimination, consumers return excess notes in proportion to the banks’ shares of total circulation after the excess is created.9 This would be tantamount to consumers shedding excess notes while blindfolded. Thus, for example, suppose there are two banks, A and B, each with a starting note circulation of $1 million and commodity-money reserves of $100,000. Their combined circulation of $2 million is just what consumers want to hold. Suppose A decides to issue an additional $100,000, which consumers do not want to hold. The consumers shed the excess by depositing it in the proportion $11 of A’s notes for every $10 of B’s notes, or $52,381 of A’s notes and $47,619 of B’s notes.10 Taking into account assumption b (discussed above), that all returned notes fall into the hands of banks other than their issuer, A loses $4,762 of reserves to B at their next clearing. This is much less than the entire amount of A’s overissue. Although A would be deterred from overissuing if it lacked excess reserves, if A did have excess reserves it could expand credit (in the form of notes) in amounts greater than the excess, suffering only a partial reserve drain.

For an expanding bank, expansion is more costly the greater the number of other banks in the system, assuming the bank overissues by the same amount and that all banks begin with equal shares of the original ($2 million) circulation. Thus, if there are four banks beginning with $500,000 circulation each, and one overexpands $100,000, it will lose approximately $5,720 of reserves to one or several of the other banks, instead of losing $4,762 as in the previous example.11 Furthermore, if an expanding bank begins with a greater than average share of total circulation, this also causes it to suffer greater clearing debits relative to the total note surplus, other things being equal. Even so, the principle of adverse clearings does not apply with full force.

Even more troublesome is a situation where there are few banks and the expanding bank is smaller than the rest. Suppose there are three banks, where Bank A has $1 million of notes in circulation and $100,000 of commodity-money reserves, Bank B has $2 million of circulation and $200,000 of reserves, and Bank C has $3 million of circulation and $300,000 of reserves. If Bank A overissues $100,000 of notes again, consumers return $11 of A’s notes for every $20 of B’s notes and every $30 of C’s notes, or $18,033 of A’s notes, $32,787 of B’s notes, and $49,180 of C’s notes. Bank A does not lose any reserves at all; instead, Banks B and C, which did not overissue, lose reserves to it; furthermore, Bank C loses reserves to Bank B! Since the proportions in which consumers return the notes of one bank to the other banks makes a difference in the amounts of the net clearing balances, let us assume in accordance with the previous example that consumers return the notes of each bank to its rivals in proportion to the rivals’ relative shares of total circulation.12 They return the $18,033 of A’s notes in the proportion of $20 to Bank B for every $30 to Bank C, or $7,213 to B and $10,820 to C. They return the $32,787 of B’s notes in the proportion $11 to A for every $30 to C, or $8,797 to A and $23,990 to C. They return the $49,180 of C’s notes in the proportion $11 to A for every $20 to B, or $17,429 to A and $31,729 to B. The net clearing balances are:

Bank C suffers a net reserve loss of $14,348, whereas banks A and B gain $8,193 and $6,155 of reserves, respectively. These results are summarized in Table 3.1 below. They imply that Bank A can continue to engage in predatory overexpansion without losing any reserves until its circulation is as big as Bank B’s; Bank C will continue to lose reserves to Banks A and B until Bank A’s circulation is as big as its own.

Thus the absence of note-brand loyalty would make it especially profitable for “undersized” banks to overissue, while forcing larger than average banks to suffer the consequences. Notice, however, that this circumstance would not encourage any general overexpansion, since larger banks would not have any profitable opportunities to expand, much less to overexpand; their best strategy would be to stand pat while their smaller rivals whittle away their circulation. The assumption of note-brand indiscrimination is, in other words, equivalent to an assumption of diseconomies of scale in note issue.

Here, then, is the rub, for the assumption of diseconomies of scale in competitive note issue has empirical implications which are clearly counterfactual, namely, that a banking system with competitive note issue should tend toward large numbers (if not an infinite number) of banks each having a minuscule share of total note circulation. Nothing of the kind appears to have been the case anywhere where competitive note issue took place. In Scottish, Canadian, and Swedish experience there was considerable diversity of market shares of circulation of various banks, with no evidence that banks with larger shares were at a disadvantage.13 Indeed, in the various historical cases of free banking there seems to have been nothing at all that corresponds to the sequence of events one would expect were consumers indiscriminate among note brands. The Scottish, Canadian, Swedish, and Suffolk systems were all remarkably stable; prudent banks did not appear to suffer at all from overexpansion of their rivals, including smaller ones. During the months leading up to the failure of the Ayr Bank, for instance, the larger Scottish banks did not experience abnormal reserve drains. In fact, their reserves increased because they consistently enjoyed favorable net clearing balances with the Ayr Bank thanks to the latter’s overissue. So historical experience confirms that consumers do discriminate between note brands and, hence, that competitive note issue, like deposit creation, is subject to increasing marginal costs.

To reject the hypothesis of note-brand indiscrimination is, however, not to say that all persons have rigid note-brand preferences (as the 100 percent note-brand discrimination hypothesis would suggest). What can be said is that a significant number of money holders do exhibit brand loyalty. It is these people’s choices that determine the relative market shares of the various note issuers. Ultimately, the presence of a stratum of nondiscriminating individuals, if it has any influence whatsoever, merely increases the amount of precautionary reserves each bank has to hold relative to its total outstanding note liabilities. Banks may find that they do not always suffer reserve drains equal to the full amount of the expansion of their note liabilities. But this temporary reprieve will be offset at other times when the lack of note-brand discrimination works in favor of other banks. On the whole, fluctuations in the flow of clearing balances are random, so that no systematic opportunities for overexpansion result from the presence of nondiscriminating note holders. In other words, as long as note-brand discrimination is not rare, the principle of adverse clearings can be assumed to govern credit expansion in a system with competing banks of issue. It follows that there is no reason to expect competitively issued note liabilities to be more prone to overexpansion than competitively issued checkable deposit liabilities.

This result has other important implications. It means that a solitary bank in a free banking system cannot pursue an independent loan-pricing policy. A “cheap-money” policy in particular would only cause it to lose reserves to rival banks. Also, no bank would be able, by overissuing, to influence the level of prices or nominal income to any significant degree, since the clearing mechanism rapidly absorbs issues in excess of aggregate demand, punishing the responsible bank. Consequently, the structure of nominal prices would not be indeterminate. Assuming stationary conditions of production, free banks face a determinate schedule of nominal money demand which strictly limits the extent of their issues.

These conclusions are precisely opposite those reached by John G. Gurley and Edward S. Shaw in Money in a Theory of Finance (1960, 253ff). Gurley and Shaw claim that a laissez-faire banking system will lead to an indeterminate price level, with the nominal supply of money “subject to no rational rule and . . . free of guidance by any hand, visible or invisible” (ibid., 256). A (minimal) solution to this problem of price-level indeterminacy, according to these authors, requires the presence of a central bank able to create nominal private-bank reserve balances in the form of claims on itself (ibid., 257). The central bank must limit its issues of such reserves while also paying a fixed rate of interest on them:

Gurley and Shaw note that this solution does not depend on a legally-specified, minimum ratio of reserves to private banks’ liabilities.

What Gurley and Shaw overlook in their analysis is that private banks in an unregulated setting already have a well-defined demand for reserves (determinants of which are discussed below, in chapter 6). This demand does not rest on the existence of a central bank capable of augmenting or otherwise influencing the supply of reserve media. Rather, it stems directly from the fact, noted by Gurley and Shaw (ibid., 256) that “the individual bank [in a laissez-faire system] is not permitted to run up an indefinite amount of clearinghouse debt to its competitors.” Thus a market for reserves will exist under laissez faire, where reserves consist of some asset which, though not issued by a central bank, is acceptable to private banks in the settlement of clearinghouse balances. If the reserve asset is an outside commodity-money such as gold, which is costly to produce and which does not bear interest (or, in other words, bears interest at a fixed rate of zero percent), its use satisfies the conditions given by Gurley and Shaw for a determinate price level and nominal money supply. Hence their rejection of laissez-faire banking in favor of central banking is unwarranted. Indeed, as we shall see in the next section, their belief that a central bank is helpful for tying down the price level and nominal supply of money is in a sense the opposite of the truth.

Monopolized Note Issue

Now let us consider a centralized system, based on commodity money, where deposit banking is entirely free but where one bank has sole right of note issue. It has been shown that the rule of excess reserves limits the ability of deposit banks in such a system to expand credit. A monopoly bank of issue is under no similar restraint. This bank is the sole source of currency, apart from commodity money, for the entire system. When holders of deposit accounts want to convert parts of their balances into a form useful in hand-to-hand payments where checks are less acceptable, they will demand conversion of their deposits into the notes of the monopoly issuer (or, perhaps, into commodity money). If public confidence in the notes of the monopoly issuer is high, its notes will be preferred to commodity money, which is more cumbersome. For their part the deposit banks, stripped of the ability to issue their own notes to supply their depositors with currency, rely upon notes of the monopoly bank of issue, or upon deposit credits at the monopoly bank (which they can convert into its notes). As all deposit banks share a common motive for holding liabilities of the monopoly bank of issue, a general demand for these liabilities develops. As Charles Rist notes in his History of Monetary and Credit Theory ([1940] 1966, 208), the liabilities of the monopoly bank of issue come to be treated by the deposit banks much as commodity-money is treated by banks in a system with competitive note issue:

Commodity money, instead of being held by the deposit banks, may (to a large extent) be deposited with the bank of issue, possibly at interest, in exchange for liabilities of that bank which, besides being useful in settling clearing balances, are at least as useful as commodity money for supplying the public’s currency needs.14

Thus as a consequence (perhaps unintended) of monopolized note issue, the liabilities of the privileged bank acquire a special status in the banking system; they become a kind of reserve media, supplementing and even superseding reserves of commodity money. Unlike deposit liabilities of non-note-issuing banks and unlike any of the bank liabilities in a system with competing note-issuers, the liabilities of a monopoly bank of issue are a form of high-powered money. Issues of such liabilities add to the base money of the system. This means, in effect, that a monopoly bank of issue is, in the short run at least, exempt from the principle of adverse clearings. The liabilities it issues not employed as currency in circulation become lodged in the reserves of deposit banks, where they cause a multiplicative expansion of credit. In general (assuming a closed economy) these liabilities will not be returned to their issuer for redemption even though their issue, and the multiplicative expansion of credit caused by it, is not justified by any prior excess demand for inside money. In other words, in a closed system a monopoly bank of issue can cause an inflationary increase in the money supply—raising the level of nominal income and prices—without suffering any negative consequences.15 Unless some external short-run control is imposed on it, a monopoly bank of issue even when its issues are convertible into commodity money can for some time at least pursue any loan-pricing policy it desires, arbitrarily expanding or contracting the money supply and causing widespread changes in nominal income and prices.

Of course, in an open system, an increase in the level of domestic prices (due to monetary overexpansion) eventually reduces exports relative to imports, causing an outflow of commodity-money reserves to foreign banking systems. This belatedly checks overexpansion by the domestic bank of issue. Also, inflation usually involves a decline in the relative price of commodity money, encouraging its withdrawal, for industrial and other nonmonetary uses, from the monopoly bank. This last effect checks overexpansion by a monopoly bank of issue even in a closed economy. Still it is a long-term corrective, which takes place after overexpansion has done the larger part of its damage. (On the effects of short-run monetary disequilibrium see below, chapter 4.) Long-term checks on overissue differ significantly from the short-run adjustments that come into play when the principle of adverse clearings operates.

Because of the special character of its liabilities, a monopoly bank of issue is able to influence the money supply like a central bank. Indeed, it is no coincidence that all central banks—that is, banks responsible for carrying out some government-directed monetary policy—have either a monopoly or a virtual monopoly of currency supply. Such a monopoly “gives them the power to dictate terms to banks which are in need of notes for deposit conversion” (Whitney 1934, 17). By controlling the issue of currency, a central bank also controls deposit expansion by non note-issuing banks:

From this one may be tempted to conclude that legislators, realizing some of the consequences of monopolized note issue, saw fit to impose it as a means for rational monetary control. In reality, however, monopolized note issue is much older than the idea of centralized, “rational” money management. The Bank of England has had a monopoly or quasi monopoly of the London note issue since its establishment in 1694. Yet the idea that it should be held responsible for preventing undesirable fluctuations in England’s money supply was not adopted as public policy until the latter half of the 19th century. Before this, the Bank of England was essentially a profit-maximizing firm the directors of which vigorously denied any responsibility for fluctuations in the money supply. In English experience, which served as the model for all subsequent central-banking reforms, it was not a demand for rational, centralized monetary control that caused note issue to be monopolized. Rather, it was the existence of a partial note-issue monopoly that inspired demands for more rational, centralized control. When monopolization of note issue awards, to a single bank, the power to “control” the money supply, it also gives that bank power to over- and underexpand credit that it would not possess were it one of a system of competing note issuers.

Illustration: The Post-1910 Australian Inflation

Some of the conclusions reached in this chapter are strikingly illustrated by the experience of Australia in the years surrounding World War I.16 Before 1910 Australia had several note-issuing banks all adhering to a gold standard. The banks settled clearings with one another in specie, since this was the only form of high-powered money in the system at the time. Under this arrangement prices were fairly stable, and the principle of adverse clearings insured that no single bank could step out of line with its competitors. If by chance the entire system went out of line, adjustment would come as a consequence of gold losses abroad.

In 1910 the Australian government passed a law authorizing a limited issue of legal-tender Australian government notes, and a year later the Commonwealth Bank was set up as an agent for issuing these notes. Soon afterwards a prohibitive 10 percent tax was imposed on all private bank note issues, and restrictions on the issue of legal-tender notes were relaxed. This gave the Australian government and its agent, the Commonwealth Bank, a virtual monopoly in note issue. The result was that the Australian notes became a new kind of high-powered money. Almost immediately the government increased its issues, and a general expansion of credit followed. So as to thwart the corrective influence of the international price-specie-flow mechanism, the government declared a gold export embargo. In the space of two or three years what had been an open system with plural note issuers was transformed into a closed system with monopolized note issue.

More consequences, all in accord with the predictions of theory, followed. In September 1914, the private banks formally abandoned their regular procedure of settling clearings in specie, giving priority to acquisition of notes from the monopoly bank of issue. By this time Australian credit expansion was entirely unleashed from normal sources of control. The Treasury and the Commonwealth Bank were free to manipulate the money supply, by altering the supply of high-powered money, in any direction they desired. As it happened, the authorities took advantage of the new arrangement to finance wartime expenditures. As J. R. Butchart notes (1918, 29), monopolization of the Australian currency

The result was a dramatic rise in Australian prices that continued throughout the course of the war, and for some time thereafter.

4

Monetary Equilibrium

Having seen the limits to expansion by free banks when the demand for money is constant, it is logical to ask what happens to these banks when the demand for money changes. In preparation for this we must define concepts like the demand for money and monetary equilibrium. Not to do so would invite unnecessary misunderstanding, since those concepts have various meanings in different contexts. The particular definitions provided in this chapter, though hardly original, are the ones most useful for studying the implications of free banking. The chapter also draws attention to some consequences of monetary disequilibrium, showing that a banking system’s ability or inability to preserve monetary equilibrium is extremely important.

The Demand for Money

“The demand for money” is a very slippery expression. Financial writers, and sometimes economists as well, have a habit of using this expression as a synonym for the demand for bank credit or loanable funds. Consequently they refer to particular interest rates as “the price of money” and call the short-term credit market “the money market.” This use of terms is highly misleading. Bank borrowers generally acquire money balances only to spend them immediately on goods and services. The demand for money, properly understood, refers to the desire to hold money as part of a financial portfolio. A bank borrower contributes no more to the demand for money than a ticket agent contributes to the demand for plays and concerts; only holders of money or actual occupants of concert seats contribute to demand.

Thus to be useful the expression demand for money must refer to peoples’ desire to hold money balances and not just to the fact that they agree to receive money in exchange for other goods and services, including later-dated claims to money.1 It is only when people who receive money income elect to hold it rather than spend it on other assets or consumer goods that they may properly be said to have a demand for money. Edwin Cannan (1921) made this point forcefully years ago:

Following a suggestion by Alex McLeod (1984, 68), it should also be noted that, although transactions balances are less obviously “held” than are speculative and precautionary balances, they are still, strictly speaking, part of the demand for money-to-hold. The demand for them is distinct from demand for money-to-spend insofar as the latter kind of demand, if accommodated by an increase in the nominal quantity of money, would lead to an increase in total spending and nominal income. It follows that, although all money in existence is at every moment held by someone, this does not mean that the demand for it is necessarily equal to the existing stock given the existing purchasing power of the money unit. When an excess supply of money exists, people will spend their surplus holdings. Money payments will increase, and so will the flow of money income. If the nominal supply of money and the extent of real output do not independently change, the increased spending will cause prices to rise in the long run. This will reduce the real value of the existing money stock, bringing it in line with the real demand for money balances. If the nominal supply of money is deficient the opposite adjustments occur. Therefore, although long-run changes in the value of money equate the demand for money with its supply, when considering the short run it is entirely valid to speak of an excess demand for or an excess supply of money. Moreover, since changes in the value of money fully eliminate excess supply or demand only in the long run (because it takes time for changes in spending to influence prices in a general way), short-run corrections in the real money supply require changes in the nominal quantity of money.

A demand may exist for either of two kinds of money: “base” or commodity money—the ultimate money of redemption—and inside money (bank notes and demand deposits) redeemable in base money. In a mature free banking system, commodity money does not circulate, its place being taken entirely by inside money. Such being the case, the unqualified expression “demand for money” used in this study will henceforth mean demand for inside money. For example, an increase in the public’s demand for money means an increase in the aggregate demand to hold bank liabilities. Unless otherwise stated, a change in demand will refer to an autonomous change in both real and nominal demand, meaning a change not itself induced by any exogenous change in aggregate nominal income.2

The Market for Inside Money and the Market for Loanable Funds

As used here “monetary equilibrium” will mean the state of affairs that prevails when there is neither an excess demand for money nor an excess supply of it at the existing level of prices. When a change in the (nominal) supply of money is demand accommodating—that is, when it corrects what would otherwise be a short-run excess demand or excess supply—the change will be called “warranted” because it maintains monetary equilibrium.

This view of monetary equilibrium is appropriate so long as matters are considered from the perspective of the market for money balances. But it is also possible to define monetary equilibrium in terms of conditions in the market for bank credit or loanable funds. Though these two views of monetary equilibrium differ, they do not conflict. One defines equilibrium in terms of a stock, the other in terms of the flow from which the stock is derived. When a change in the demand for (inside) money warrants a change in its supply (in order to prevent excess demand or excess supply in the short run), the adjustment must occur by means of a change in the amount of funds lent by the banking system.

An important question, one particularly controversial among monetary economists in the middle of this century, arises at this point. Are adjustments in the supply of loanable funds, meant to preserve monetary equilibrium, also consistent with the equality of voluntary savings and investment? The answer is yes, they are. The aggregate demand to hold balances of inside money is a reflection of the public’s willingness to supply loanable funds through the banks whose liabilities are held. To hold inside money is to engage in voluntary saving. As George Clayton notes, whoever elects to hold bank liabilities received in exchange for goods or services “is abstaining from the consumption of goods and services to which he is entitled. Such saving by holding money embraces not merely the hoarding of money for fairly long periods by particular individuals but also the collective effect of the holding of money for quite short periods by a succession of individuals.”3

Whenever a bank expands its liabilities in the process of making new loans and investments, it is the holders of the liabilities who are the ultimate lenders of credit, and what they lend are the real resources they could acquire if, instead of holding money, they spent it.4 When the expansion or contraction of bank liabilities proceeds in such a way as to be at all times in agreement with changing demands for inside money, the quantity of real capital funds supplied to borrowers by the banks is equal to the quantity voluntarily offered to the banks by the public. Under these conditions, banks are simply intermediaries of loanable funds.

Thus a direct connection exists between the conditions for equilibrium in the market for balances of inside money and those for equilibrium in the market for loanable funds. An increase in the demand for money warrants an increase in bank loans and investments. A decrease in the demand for money warrants a reduction in bank loans and investments. To put the matter in Wicksell’s terms, changes in the supply of loanable funds that accord with changes in the demand for inside money also ensure that the money rate of interest is kept equal to the “natural rate.”

Any departure from monetary equilibrium has disruptive consequences. Consider what happens when the supply of money fails to increase in response to an increase in demand for money on the part of wage earners. The wage earners attempt to increase their money balances by reducing their purchases of consumer products, but there is no offsetting increase in demand for products due to increased, bank-financed expenditures. Therefore, the reduction in demand leads to an accumulation of goods inventories. Businesses’ nominal revenues become deficient relative to outlays for factors of production—the difference representing money that wage earners have withdrawn from circulation. Since each entrepreneur notices a deficiency of his own revenues only, without perceiving it as a mere prelude to a general fall in prices including factor prices, he views the falling off of demand for his product as symbolizing (at least in part) a lasting decline in the profitability of his particular line of business. If all entrepreneurs reduce their output, the result is a general downturn, which ends only once a general fall in prices raises the real supply of money to its desired level.5

As was said previously, such a crisis can occur only if banks fail to respond adequately to a general increase in the demand for inside money. The crisis involves a deflationary Wicksellian process during which bank rates of interest are temporarily above their natural level. This is opposite the inflationary Wicksellian process, with bank rates below the equilibrium or natural rate of interest, that economists of the Austrian school traditionally emphasize.6 Nevertheless deflation (resulting from unaccommodated excess demand for inside money) has been an important factor in historical business cycles, and a banking system that promotes deflation disrupts economic activity just as surely as one that promotes inflation, although the exact nature of the disruption differs in each case.

Opinions of Other Writers

The view of monetary equilibrium presented here should not be controversial, and has been upheld by many economists. To cite but a few examples, it has been put forth by J. G. Koopmans (1933), Gottfried Haberler (1931), Fritz Machlup (1940), Jacques Reuff (1953), W. Zawadski (1937), and (in a qualified way) Friedrich A. Hayek (1935, [1933] 1975b and 1939b, 164ff), among continental European theorists. Most of these writers link the concept of monetary equilibrium to that of “neutral” money.7 According to Koopmans (1933, 257), who has developed this approach most thoroughly, monetary policy should have the goal of “compensating for any deflation, due to hoarding, by creating a corresponding amount of new money, or of compensating for any inflation, due to dishoarding, by destroying money in like measure.” When this goal is achieved “the money outlay stream should remain constant.” In other words, money is neutral as long as Say’s Law remains valid (that is, as long as excess demand for money is zero). Conversely, monetary disequilibrium occurs and money is non-neutral whenever Say’s Law is violated:

Machlup has the same view in mind when he writes (1940, 291 and 184-89) that “credit inflation is ‘healthy’ if it compensates for deflation through current net hoarding, or for an increase in the number of cash balances or in the number of ‘stopping stations’ in the money flow” and that credit contraction is healthy if it compensates for dishoarding (“a decrease in idle balances”).

Hayek is more equivocal in his suggestions concerning an ideal monetary policy. At one point in Prices and Production he recommends that the money supply be kept constant.8 Yet he follows this with a statement acknowledging the need to make adjustments in the money supply in response to changes in the “co-efficient of money transactions.” In still another passage he mentions the need to accommodate changes in the “average velocity of circulation” of money, noting that “any change in the velocity of circulation would have to be compensated by a reciprocal change in the amount of money in circulation if money is to remain neutral towards prices.”9 Finally, in his most explicit statement concerning the importance of adjusting the nominal supply of money, he says (1939b, 165) that banks must “create additional credits for investment purposes to the same extent that holders of deposits have ceased to use them for current expenditure.” This serves to avoid the “undesirable deflationary consequences” of unaccommodated saving.

Hayek’s equivocation is due, on one hand, to his view that desirable adjustments in the money supply cannot be formulated into a “language of practice” (1935, 108) which, of course, does not argue for rejecting them as a theoretical ideal, and on the other to his notion that equilibrium in the market for “real capital” can be preserved only if banks do not issue unbacked (fiduciary) media (Hayek 1935, 23). The latter view contradicts our claim in the previous section: it ignores the fact that changes in the desire to hold inside money reflect the public’s willingness to lend “real capital” to and through the banking system. More will be said in defense of this criticism later on.

Allowing for the ambivalent views of Hayek, all of the continental writers cited have notions of monetary equilibrium similar to the one adopted in the present work. The same may be said concerning views on monetary equilibrium entertained by many well-known British theorists, including Dennis Robertson (1926, and [1957] 1964, chap. 5, sect. 4), E. F. M. Durbin (1933), J. E. Meade (1933), A. G. B. Fisher (1935), Ralph Hawtrey (1951), and A. C. Pigou (1933). Typically these writers express the notion in question in the form of the rule that the supply of money multiplied by its income velocity of circulation should remain constant. According to Durbin (1933, 187) such a policy would “avoid income deflation on the one hand and a profit inflation on the other.”10

J. E. Meade (1933, 8) argues along the same lines that the total increase in the supply of money in a given period of time should equal the net increase in the demand for money during the same period, with bank investments adjusted correspondingly. Besides preventing changes in final (nominal) income this policy would assure an equilibrium interest rate.

Robertson, in Banking Policy and the Price Level, states his views on the requirements for monetary equilibrium in very idiosyncratic language.11 Nevertheless he also believes that an increased general desire to hoard should be offset through additional bank lending:

Most of these authors explicitly distinguish the goal of accommodating changes in the demand for money through changes in nominal supply from that of stabilizing an index of prices. The two goals differ because general price movements may be caused by changes in productive efficiency, and not just by changes in the demand for money balances relative to nominal income. Offsetting price changes due to changes in productive efficiency would not preserve monetary equilibrium.13 The reasons for this will be discussed in chapter 7. For the time being it is only necessary to note that the procedures for maintaining monetary equilibrium discussed here should not be viewed as leading to price-level stabilization.

Many past and present American monetarists would probably agree with the theoretical views of the European writers discussed above. Their preference for other policies—for price-level stabilization or a fixed money growth rate rule—stems, not from any theoretical disagreement, but from their view that these policies provide the best achievable approximation to the ideal of a truly demand-elastic money supply.14 Other American writers have explicitly defended the monetary equilibrium ideal, calling for the adjustment of nominal money supply to avoid monetary disequilibrium. The most important of these theorists was Clark Warburton (1981), who noted the popularity of what he termed the “monetary disequilibrium” approach in American writings of the early decades of the 20th century. In the 1960s the same approach was “rediscovered” by Robert Clower (1965, 1967) and Axel Leijonhufvud (1968), who also interpreted Keynes as a monetary disequilibrium theorist. Lately Leland Yeager, who has been heavily influenced by Warburton, has defended the monetary equilibrium-disequilibrium approach against the “equilibrium always” theorizing of the new-classical school (Yeager 1986).

Finally, some remarks should be made about Keynes and Keynesian theory. It is well known that consumers’ propensity to hoard and “liquidity preference,” in conjunction with downward inflexibility of money wages, play a crucial role in Keynesian explanations of depression and unemployment. In general, Keynes believed, an elastic supply of inside money should prevent hoarding and liquidity preference from having any negative influence on aggregate demand: increased investment (financed by the banking system) should follow every net increase in aggregate money demand.15 This view is quite consistent with the other views on monetary equilibrium cited here. Keynesian analysis, however, came to attach great importance to the possibility of a liquidity trap, a possibility which Keynes himself treated as an extreme, limiting case. The presence of a liquidity trap (which involves an infinitely interest-elastic demand for money balances) renders monetary expansion through conventional banking channels impotent as a spur to investment. It therefore necessitates resort to increased government spending in order to augment aggregate demand. Also, some Keynesians (and Keynes himself may be included here) suggest that employment should not be considered “full” so long as it can be increased by an expansionary policy, even if the policy leads to an increase in money wages.16 This view seems to attach overriding importance to short-run reductions in unemployment without acknowledging the undesirable consequences, both in the long and in the short run, of monetary disequilibrium.17

Despite these important differences between Keynesian analysis and the views of other monetary-equilibrium theorists, many Keynesians might accept the prescription for monetary equilibrium offered in this chapter. Those who do not regard the liquidity trap as an important factual possibility would probably accept it as entirely adequate. Some might wish to supplement it with government spending programs, of course. Those who accept the possibility of a permanent or semi-permanent inflation-unemployment trade off, or otherwise think that the benefits of inflation generally exceed the costs, will likely reject it in favor of outright inflation.

These views of other writers are not cited as evidence of the correctness of any theory. Their purpose is merely to show that the concept of monetary equilibrium adopted in this study is neither new nor controversial. The concept is applied here to the appraisal of free banking. Any originality lies in this appraisal, rather than in the criteria on which it is based.

Transfer Credit, Created Credit, and Forced Savings

The difference between warranted and unwarranted additions to the stock of inside money is usefully illuminated by a distinction between “transfer credit” and “created credit.”18 Transfer credit is credit granted by banks in recognition of people’s desire to abstain from spending by holding balances of inside money.19 In contrast, created credit is granted independently of any voluntary abstinence from spending by holders of money balances.20 When the demand for money falls, its nominal supply must also be reduced or else some transfer credit becomes created credit.

Obviously, created credit can exist only in the short run: a spurt of credit creation prompts an adjustment of prices which eventually restores monetary equilibrium, causing all outstanding credit to conform to the aggregate demand for money. In equilibrium all credit is transfer credit because, by our definition of monetary equilibrium, nobody holds inside money balances in excess of the balances he desires to hold. Thus any reference to created credit or to credit creation means a temporary excess supply of money due to excessive bank lending or investment.

Credit creation, Fritz Machlup notes (1940, 183), “places money at the disposal of the market . . . without any corresponding release of productive factors . . . due to voluntary refraining from consumption.” Unlike operations involving credit transfer it “makes it possible for investment to take place in the absence of voluntary savings.” Such investment “gives rise to the development of disproportionalities in the production process.”

Whereas voluntary savings support transfer credit, real resources invested by means of credit creation represent “forced savings.” The notion of forced savings, which Hayek (1939a, 183-97) traces back to Bentham, refers to the reduction of real income suffered by earners of fixed money incomes when goods they normally purchase are bid away by recipients of new income having its source in credit creation.21 Malthus’s discussion of this phenomenon, which appeared in an 1811 issue of the Edinburgh Review, assumes a case where credit is created exclusively in the form of bank notes:

This artificial diversion of resources to new industries does not continue once prices adjust to eliminate the excess supply of money:

What begins, in other words, as both a nominal and a real increase in loanable funds becomes, after a general adjustment of prices, a nominal increase only, which cannot support the needs of capital maintenance and project completion for both new (post-expansion) and old (pre-expansion) employers of credit. Marginal borrowers will be excluded from the market for loanable funds, and their investments may have to be liquidated, resulting in a slump.

The point of this discussion is to show how forced savings and its consequences are bound up with created credit. They arise whenever the granting of credit gives rise to bank liabilities in excess of the demand for balances of inside money. They do not arise insofar as credit offered by banks consists solely of transfer credit, i.e., of credit the granting of which gives rise to liabilities in amounts consistent with the demand for inside money. The distinction between created credit and transfer credit will be employed later in examining the response of a free banking system to changes in the demand for money.

A contrasting view of bank credit appears in the writings of several of the Austrian economists, especially Ludwig von Mises, who give the phenomenon of forced savings a prominent place in their elaborations of the monetary theory of the business cycle.24 According to these writers any credit expansion or increase in the supply of fiduciary media—inside money not backed 100 percent by reserves of commodity or base money—is unwarranted. “The notion of ‘normal’ credit expansion,” according to Mises, “is absurd”:

In other words, all net expansion of fiduciary credit is a cause of loan-market disequilibrium. It causes bank rates of interest to fall below their “natural” levels, leading to forced savings and other trade-cycle phenomena. This contrasts with the view defended here, which holds that no ill consequences result from the issue of fiduciary media in response to a greater demand for balances of inside money. According to the latter view it is perfectly possible that fiduciary media may arise from loans or investments involving transfer credit only. The expansion of bank liabilities may represent a response to greater abstinence by money holders and, hence, to a fall in the “natural” rate of interest. In this case the fiduciary issue conforms with the “golden rule” referred to by Mises ([1953] 1980, 295). According to this rule, “the credit that [a] bank grants must correspond . . . to the credit that it takes up.”

If some issue of fiduciary media does not involve credit creation, then Mises’s “commodity credit,” which is supposed to be credit not based on fiduciary media (and hence, in Mises’s view, not having trade-cycle consequences), must be a mere fraction of what we are calling transfer credit. In fact precisely what Mises means by commodity credit is not clear. If the phrase refers to bank issues backed 100 percent by reserves of commodity money (which would make it the complement of what Mises calls “fiduciary” or “circulation” credit) then it does not refer to a form of credit at all. A bank holding 100 percent reserves against all of its liabilities is not a credit-granting institution, but a warehouse.

Alternatively, it may be that by commodity credit Mises means credit granted by banks on the basis of time liabilities, as opposed to liabilities redeemable on demand but not backed by 100 percent reserves. But in this case Mises confuses a difference of degree with one of substance. Holders of demand liabilities are granters of credit just as are holders of time liabilities. The only difference is that in the former case the duration of individual loans is unspecified; they are “call loans” that may mature at any time. Bankers must rely upon their entrepreneurial judgment to avoid violating the “golden rule” of not lending more than what is offered to them.26

However one interprets it, Mises’s view that commodity credit is the only sort of credit consistent with loan market equilibrium causes him to be critical of fractional reserve banking.27 This puts him in a league with such writers as P. J. Geyer (1867) and J. L. Tellkampf (1867), who called for the abolition of fiduciary media in Germany, and with Henri Cernuschi (1865) and P. Modeste (1866), who lobbied for its suppression in France.28 Indeed, Mises’s support for free banking is based in part on his agreement with Cernuschi, who (along with Modeste) believed that freedom of note issue would automatically lead to 100 percent reserve banking.29

This difference of opinion has implications for the appraisal of free banking’s consistency with monetary and loan-market equilibrium. If the view defended in this chapter is correct, then it is desirable, not that free banks should prevent all issues of fiduciary media, but rather that they should only prevent issues that are inconsistent with changing demands for money balances. If free banks function this way they are merely transferring credit, not creating it.

Chapter 3 showed why individual free banks cannot create credit as long as the demand for money is constant. But it did not show that they preserve monetary equilibrium when the demand for money is increasing or decreasing. Nor did it show whether credit creation is possible under either static or changing conditions for free banks acting in unison. How free banks respond to changes in demand, and whether there are any adequate restraints on their collective behavior, will be the subjects of the next two chapters.

5

Changes in the Demand for Inside Money

Except in Chapter 2, where account was taken of the effects from a lowering of the exchange value of commodity money in response to fiduciary substitution, this study has assumed a demand for money balances fixed in both real and nominal terms. We must now consider how a free banking system responds to changes in the demand for money. Would a free banking system accommodate an increase in the demand for money? Would it automatically reduce the supply of money in response to a fall in the demand for it? Would it, in short, continue to maintain monetary equilibrium?

A change in the demand for money—meaning real demand to hold inside money—can be due to a change in the number of bank liability holders, a change in the holdings of the same individuals, or a combination of both. It does not necessarily involve any redistribution of existing demands among various banks. References to an increasing or decreasing demand in this chapter will mean changes in aggregate or total demand. Thus when the demand to hold the liabilities of one bank increases, it is assumed, unless otherwise stated, that there is no change in the demand to hold liabilities of other banks.

Increased Money Demand

Suppose then that there is a growth in the total demand for inside money that takes the form of a growth in demand for the liabilities (notes and deposit balances) of one bank. How can this increased demand be satisfied? We have seen that, in a mature system where there is no more outside money in circulation to be deposited, all fresh sums of inside money make their first appearance through new loans and investments. In general such newly issued liabilities do not at first come into the hands of those persons who happen to desire to hold more of them. There is an exception to this, though, which also provides the simplest example of how a bank may profitably expand its liabilities (with a fixed supply of reserves) in response to the increased money demands of its clients. This exception involves so-called compensating balances. These are balances held by bank borrower-customers as part of their loan agreement. A person or firm that holds compensating balances is simultaneously a borrower and a lender of the sum in question: in accepting a loan the person or firm becomes a debtor to the bank, but to the extent that borrowed funds are willingly held (rather than spent) the borrower is also a creditor. The bank has issued claims to commodity money, but these claims are not going to be redeemed by anyone. Some of the commodity money that has been “borrowed” is, in effect, never taken from the bank. Hence the bank can lend it to someone else.

When a bank borrower explicitly agrees to keep a compensating balance, the banker knows in advance that the balance will be held rather than spent and that it will therefore not become a source of adverse clearings like other loan-created deposits. A bank’s compensating-balance liabilities are one of its most reliable sources of credit.

The holders of compensating balances, in turn, also benefit by holding them so long as their borrowed holdings do not exceed their ordinary demands for money balances. The benefit is due to the fact that loan rates are generally lower for loans involving compensating balance agreements, which reflects the banker’s preference for having his liabilities outstanding in a form not likely to contribute to unanticipated clearing losses.1

That banks can accommodate the fluctuating money demands of their borrower customers by means of changes in compensating balances is obvious enough. But how might they satisfy the increased demands of would-be note and deposit holders who are not among their borrower customers? Recall that every day a certain number of a bank’s notes and checks enter the clearing apparatus and become a source of debits against it. If the bank is in equilibrium vis-à-vis its rivals, it will on average have clearing credits equal to its clearing debits, and so it can maintain its outstanding liabilities, replacing its earning assets as they mature. Now suppose that some of the bank’s depositor customers write fewer checks on their balances (without increasing the average size of their checks), or that more individuals who come into possession of the bank’s notes hold on to them instead of spending them. The result will be a reduced flow of the bank’s liabilities into the clearing mechanism—a reduction in adverse clearings against it—much like the reduction that would occur if borrower customers elected to increase the portion of their borrowings represented by compensating balances.

A metaphorical description of the process of demand expansion may be helpful. The outstanding supply of inside money may be thought of as flowing through the economy in a stream of money income and expenditures. Along this stream are pockets or “reservoirs” representing individuals’ and firms’ holdings of inside-money balances. When the demand for money increases, either the number of reservoirs increases (as when the number of firms or people that want to hold money increases), or existing reservoirs deepen (as when demands of existing money holders become more intense). Either case results in a withdrawal of funds from the income stream. When the increased demand is for the liabilities of one particular bank, then that bank’s liabilities are removed from the flow of spending and income. Instead of being cancelled by the clearing mechanism, they become lodged in the reservoirs of demand.

The withdrawn liabilities thus cease to contribute to their bank’s reserve demand: the reserves held by their issuer—formerly just adequate to sustain its liabilities—become excessive as positive net clearings accumulate. To maximize its profits, the bank whose liabilities are in greater demand expands credit. This newly expanded credit is transfer credit, because it is issued in response to the desire of certain people to hold more of the liabilities of the bank that grants it. Hence it does not lead to any forced savings or upward pressure on prices. It allows the bank to recover its equilibrium vis-à-vis other banks in the system.

It should be noted that, as far as the maintenance of monetary equilibrium is concerned, the specific point of the injection of new liabilities is not crucial: the bank can be expected to make new credits available to borrowers in a way that satisfies the principle of equimarginal returns.2 All that matters is that the bank recognizes the decision to save made by holders of inside money.

Now consider the consequences of an increased general demand for inside money, one that confronts all banks at once. No bank in this case witnesses any improvement in its circumstance relative to other banks, that is, any positive net average clearings. Nevertheless each bank will have fewer gross clearings than before, insofar as the banks considered as a group do not respond to the increased demand, and this will reduce each bank’s need for (precautionary) reserves relative to its actual reserve supply. Thus the banks will find it profitable to expand until their total gross clearings are such as to again raise the demand for reserves to the level of available supply. A more complete explanation of this requires a discussion of the economic determinants of reserve demand, which is deferred until the next chapter.

To summarize, a general increase in the demand for inside money is equivalent to a general decline in the rate of turnover of inside money. Bank notes change hands less frequently, and holders of demand deposits write fewer (or perhaps smaller) checks. As a result, bank liabilities pass less frequently into the hands of persons or rival issuers who return them to their points of origin for redemption. The reduction in turnover of liabilities leads directly to a fall in the volume of bank clearings. When this happens banks find they have excess reserves relative to the existing level of their liabilities, and so they are able to increase their holdings of interest-earning assets, which they do by expanding the supply of inside money in a manner that accommodates the growth in demand for it. In standard textbook terminology, there is an increase in the reserve multiplier. Liability expansion continues so long as there are unfilled reservoirs of demand. Any issues in excess of demand, however, will lead to additions to the stream of payments, causing an increase in bank clearings and reserve requirements.

Yet another way to put the argument is to speak in terms of velocity rather than turnover. Then free banks can be said to accommodate a fall in the velocity of inside money with an increase in its supply. Regardless of how one phrases it, the actions of the banks are precisely the ones required for the maintenance of monetary equilibrium. For reasons that will be made clear later on, this cannot generally be the case for central-banking systems or for otherwise unregulated systems lacking freedom of note issue.

An issue arises at this point concerning the relative maturity structures of bank assets and liabilities. Individual money-holders’ offerings of credit to their banks are, generally speaking, of much shorter (as well as more uncertain) duration than the banks’ resulting offerings to their borrower customers. Hence the bankers must be able to transform short-term credits of unknown duration into longer-term loans and investments. They do so by taking advantage of the law of large numbers in the same way as goldsmith bankers took advantage of it in making the first loans based on fiduciary media (see above, chapter 2). The actions of large numbers of independently motivated users of inside money will largely offset one another in the day to day course of things. One person spends his deposit balance previously held at Bank A, thereby ceasing to be one of its creditors. Another simultaneously adds to his deposit account at Bank A. The same kind of largely offsetting transactions occur at all other banks. As a result, the sum of short-lived, individual demands for inside money can be treated by the banks as a fund suitable for investing in a portfolio of earning assets the average maturity of which significantly exceeds the average turnover period of individual notes or deposit credits. The maturity structure and quantity of a bank’s loans and investments therefore depends on the expected behavior of the aggregate demand for its liabilities, and not on changes in the composition of individual demands from which the aggregate is derived.

Decreased Money Demand

Now consider what happens when the demand for inside money falls. Supposing once more that the fall affects the liabilities of one bank only, it means the shrinking or disappearance of reservoirs in which that bank’s liabilities were formerly lodged. The liabilities return to the stream of money income, where they pass into the clearing mechanism and become debits to their issuer. To adjust its position vis-à-vis its competitors the bank, experiencing reduced demand for its liabilities, calls back some of its loans and liquidates some investments or at least ceases to renew some existing loans as they mature. Thus, just as the supply of inside money is increased through the expansion of credit, it may be reduced through the absorption of credit, that is, by the retirement of loans and investments.3

It is even possible for unwanted liabilities to be directly returned to their points of origin by way of the repayment of loans. Suppose, for example, that the surplus liabilities no longer wanted in money balances are paid over to someone who happens to be indebted to their issuer, and that the individual in question uses them to repay his loan. What is the overall result of this transaction? The bank suffers no clearing loss and no change in the volume of clearings against it.4 Yet the sum of its outstanding liabilities has fallen, assuming that the bank does not make any new loans or investments. If, on the other hand, the bank attempts to restore its assets and liabilities to their previous level by new extensions of credit it will suffer adverse clearings approximately equal to the new issues. Presumably the bank had no excess reserves to begin with (i.e., before the demand for its liabilities fell and before one of its loans was repaid). Since nothing has happened since to supply it with excess reserves, the bank cannot sustain further adverse clearings, and so it must ultimately accept a reduction in its business.

Similar consequences follow if the liabilities of one bank, the demand for which has fallen, are used to repay loans at another bank. What happens here is best understood as a two-step process. First imagine that Y, who is indebted to Bank B, receives inside money from X, who in making the payment in question permanently reduces his holdings at Bank A. Suppose furthermore that Y at first deposits the sum received from X at Bank B. Then Bank A will, other things being equal, suffer an adverse clearing equal to the amount of the payment made by X to Y. In response to this loss Bank A has to contract its liabilities. On the other hand, Bank B has become a recipient of new, excess reserves, and so it can expand its liabilities. The overall result so far is a zero net change in the supply of inside money, since Y’s increased holdings precisely offset the reduction of demand on the part of X.

Now, however, we must consider the second step of the loan repayment process, in which Y withdraws his new deposit by writing a check to his bank in repayment of a loan that was previously granted to him. This is identical to what happens when unwanted liabilities are used directly to repay loans from their original issuer. It leads to the extinction of the liabilities, with no possibility for offsetting expansion by the bank to which they are repaid. Thus when unwanted liabilities from Bank A are used to repay loans from Bank B, the overall result is an extinction of Bank A’s liabilities to the extent of the fall in demand with no offsetting increase (of any significant duration) in the supply of liabilities from Bank B or from any other bank.

The processes described here can once again be generalized for the case of an all-around reduction in the demand for inside money. In this case there is an increase in the rate of turnover (or velocity) of bank liabilities, which means that bank clearing assets have to “turn over” more rapidly as well. In other words, the liquidity needs of the banking system increase, and an existing volume of reserves can no longer support the same amount of inside money. Therefore assets and liabilities must contract, and the reserve multiplier falls. Once again the explanation of reserve demands involved in this generalization must wait until the next chapter.

Thus the capacity of free banks to maintain equilibrium applies also to conditions where the demand for inside money is changing. This result is just an extension of the static rule of excess reserves since it rests on the demonstration, to be completed in the next chapter, that the overall availability of excess reserves (or conversely the overall excess demand for reserves) is a function of the aggregate demand for inside money. It does not apply to systems with monopolized currency issue: adjustments in the money supply in such systems have to be achieved, assuming they can be achieved at all, by deliberate policy. The possibilities for controlling the money supply by means of central banking will be critically examined in chapters 7 and 8. Prior to this, though, it is necessary to examine certain common ideas concerning reserve requirements that might cast doubt upon the conclusions just arrived at.

6

Economic Reserve Requirements

The finding that free banks will maintain monetary equilibrium even as the demand for money changes rests upon the claim that the reserve multiplier alters in sympathy with changes in the demand for inside money. Many conventional discussions contradict this claim. Some treat the reserve multiplier as an institutionally given constant, rather than as a variable that adjusts in response to changes in the public’s behavior and in response to bankers’ reactions to changes in the public’s behavior. Others go to the opposite extreme and hold the reserve multiplier to be, in some situations at least, indeterminate. The purpose of this chapter is to show that both of these alternative views are, for a free banking system, incorrect.

The Conservation Theory

The view that the reserve multiplier is a constant may be dubbed the “conservation theory” of bank money.1 It concurs that an individual bank may expand or contract credit as a result of persons holding greater or lesser amounts of its liabilities. But it denies that system-wide expansions or contractions of the money supply can happen in response to changes in aggregate demand: individual banks may gain or lose business relative to one another, but gains by some banks are always compensated by losses of other banks.2 As long as the reserve base is unchanged, the system as a whole will support only a certain amount of inside money—no more, no less.3

The conservation theory looks upon the volume of liabilities in a banking system as if it were like the volume of water in a waterbed. A little pressure on one part of the waterbed reduces the amount of water there but causes an equal increase in its amount somewhere else. Likewise, a fall in the demand for one bank’s liabilities is supposed to cause that bank to contract, but only with the accompanying effect of an offsetting shift of reserves and lending power to other banks. In contrast, an increase in supply of one bank’s liabilities is held to be possible only if its rivals contract.

In money and banking literature the conservation theory is generally expressed in terms of bank deposits only, since competitive note issue is rarely discussed. John Philip Wernette’s argument (1933, 32) that “an effective thrift campaign cannot increase the total deposits of an entire banking system” is typical:

Another example is George Clayton. Although he concedes that banks may respond positively to growing demands for inside money in the early stages of their development, Clayton denies that this is possible in the “developed stage” of banking. He states that the “immobilization” of part of a bank’s liabilities as the result of deposits “left inactive at the bank . . . does not put the bank in a position to lend or invest any more money than before.”4 Thus deflation, according to Clayton (1955, 98), “can only be overcome by a deliberate policy of credit expansion under the direction of the central bank, which would have to provide the extra cash reserves necessary to maintain the [reserve] ratio.” Other theorists have made similar statements denying the possibility of a general contraction of inside money given a fixed amount of bank reserves.

The problem with the conservation theory is that it miscomprehends the forces that determine banks’ need for reserves. It assumes that this need depends only on the amount of banks’ outstanding liabilities and not on the demand for these liabilities relative to income as it affects their turnover.5 This error may stem from conservation theorists’ identification of economic reserve requirements with statutory reserve requirements. These are legally prescribed, minimum ratios of reserves to total liabilities. Obviously statutory reserve requirements are institutionally “given,” and they set an upward limit to the reserve multiplier. But statutory reserve requirements do not exist in all central banking systems, and are absent from a free banking system. Under free banking, reserve requirements are determined by the optimizing decisions of bankers. They are economic, rather than statutory, requirements. Granting this, the critical questions become, what factors determine a free bank’s economic reserve demand?6 and will this reserve demand necessarily be a constant fraction of a bank’s total outstanding liabilities?

Determinants of Reserve Demand

A free bank’s economic reserve demand for any planning period can be thought of as having two components. These are, first, a component equal to what the bank, because of the structure of its assets and liabilities, anticipates will be the difference between its total clearing debits and its total clearing credits for the period—its “average net reserve demand”—and, second, a component to cover the bank against any adverse clearings it may face during the planning period that (singly or cumulatively) exceed its average net reserve demand. The latter component is the bank’s “precautionary reserve demand.”7 It protects the bank, not from such adverse clearings as might be predicted given a determinate structure of the demand for the bank’s liabilities, but from temporary, random fluctuations in these adverse clearings above their expected value. A bank that fails to hold precautionary reserves might, on average, have credit clearings equal to its debit clearings, so that its average net reserve demand would be zero. Yet the bank would stand a great risk (one chance in two in fact) of being unable to redeem all its debits at the clearinghouse during any particular clearing session if it held zero reserves. It follows that banks have to hold positive precautionary reserves so long as the exact incidence of clearing debits is unknowable or uncertain,8 and even though they may have no reason to doubt that their clearing debits and credits will be equal in the long run.9

Both of these components of a free bank’s reserve demand are related to the total clearing debits it faces, and not necessarily to the total of its outstanding liabilities. Moreover, the quantity of a bank’s liabilities returned to it through the clearing mechanism depends just as much on their average turnover as on the quantity of them outstanding. Thus, to take the limiting case, additional liabilities with zero turnover would not add to an expanding bank’s reserve demand, and contraction of zero turnover liabilities by a bank would not add to its excess reserves. On the other hand, a bank’s reserve demand may increase even though it has not expanded its liabilities, because turnover of its liabilities has increased. Finally, a bank’s reserve needs may fall although its liabilities are unchanged because the average period the public holds its liabilities has increased.

In long-run equilibrium the average net reserve demand for every bank in a system with a fixed supply of reserve media has to be zero. A bank cannot continue to suffer a positive average net reserve demand without eventually disappearing, and it cannot have a continuously negative average net reserve demand unless it fails to exploit fully the demand to hold its liabilities and hence its lending power. Profit-maximizing banks will strive to adjust their outstanding liabilities to compensate for demand-induced changes in their net clearing debits so as to keep their average net reserve demand equal to zero: a bank that expects to acquire more reserves than it expects to lose during a planning period (because the demand to hold its liabilities has increased) will expand its loans and investments to make up the difference; one that expects to lose more than it gains (because the demand to hold its liabilities has fallen) will contract. A bank that does not adjust its issues when faced with changes in the demand for them is in no less unprofitable a position in relation to its rivals than one that overexpands or underexpands relative to its rivals when faced with an unchanged demand for inside money.

Does extending this conclusion to the banking system and hence to adjustments in aggregate liabilities involve a fallacy of composition? It does not, because expansion by any one bank in response to reduced clearing debits against it does not, in the case of an increased demand for (reduced turnover of) its liabilities, involve any reduction of the reserves or lending power of rival banks. Indeed, such expansion actually prevents the redistribution of reserves that would occur if the supply of inside money were not adjusted in response to demand. The same holds for credit contractions by individual banks when these contractions serve to maintain an equilibrium of supply and demand for their liabilities.

Uniform Changes in Money Demand

What has just been said refers only to actions brought about by banks’ desire to maintain zero long-run net average reserve demand, that is, by their need to remain in equilibrium in relation to one another. It leaves a very crucial issue unaddressed—an issue that is sometimes raised in connection with the conservation theory. Granted that particular banks may contract or expand the aggregate sum of bank liabilities to stay in line with other banks in the face of changing demands for their liabilities only, what incentive can there be for system-wide expansion or contraction when all banks are confronted by equal and simultaneous changes in the demand for their issues? For example, if there is a general fall in the demand for inside money that uniformly raises the gross clearings of all banks no single bank will suffer a deficiency of average net reserves. Each will have its debit and credit clearings increase in equal amounts, with no change in adverse clearings. Similarly, if all banks witness equal increases in the demand for their issues, none will feel a need to expand in so far as the only motivation to do so is to prevent excess reserves (due to positive clearings) from accumulating.

Obviously banks have no incentive to contract or to expand under such circumstances if their only motivation for doing so is to keep in step with one another; they are already in step, and a uniform increase in inside-money demand will not put any of them out of it. Does it follow, therefore, that under such conditions the banks do nothing, so that the conservation theory is correct?

The answer is an emphatic “no.” Forces operate in a free banking system to make the supply of inside money adjust to changes in demand even when such changes fall upon the banks simultaneously and uniformly. The reason for this has to do with the precautionary demand for reserves. Unlike the average net demand for reserves, the precautionary demand is affected by unaccommodated, uniform changes in the demand for inside money. The reasons for this are discussed in detail in the literature on precautionary reserve demand, beginning with Edgeworth’s pioneering article.10 The essential conclusion of this literature, based on the law of large numbers, is that the precautionary demand for reserves rises or falls along with changes in the total volume of gross bank clearings, though not necessarily in strict proportion to the change in gross clearings. Specifically, a uniform increase in the total volume of clearing debits due to an increase in the frequency of payments (such as would occur if there were an across-the-board fall in the demand for inside money with income constant) requires that precautionary reserves increase by a factor at least equal to the square root of the factor by which clearings have increased. A fall in the total volume of clearings will likewise lead to a fall in the demand for precautionary reserves.11

This result can be represented by a set of simple diagrams (Fig. 6.1) showing the frequency distribution of clearing debits at a representative bank before and after a doubling of the total volume of clearings. The smoothness of the diagrams implies a fairly long planning period with many clearing sessions; one might also interpret them as showing the statistical likelihood of particular net clearings based on a large number of trials. The doubling of gross clearings doubles the scale of the horizontal axis of the frequency-distribution diagram. Because of the law of large numbers, however, the distribution becomes more concentrated at its center and the variance increases, but by less than the increase in the scale of clearings.

The intuition behind the square-root result is fairly simple. As the volume of gross clearings increases, so do random fluctuations in their distribution among the banks—the source of variance of net clearings faced by individual banks—only less than in proportion. This comes directly from the laws of probability. Since precautionary reserves are held against deviations of average net demand from its mean or expected value, it follows that precautionary reserve demand rises by the same factor as the variance of net clearings. Since gross bank clearings increase whenever there is an uncompensated, general decline in the demand for inside money (income constant), and gross clearings fall when there is an uncompensated, general increase in the demand for inside money, it follows that bank reserve needs are affected by changes in the demand for inside money even when these changes affect all banks simultaneously and uniformly.

If a banking system has a fixed supply of reserves, the square-root law of precautionary reserve demand implies (a) that banks contract their issues in response to a uniform fall in the demand for inside money to prevent their need for precautionary reserves from exceeding the available supply of such reserves (so that they do not come up short more frequently at the clearinghouse); and (b) that banks expand their issues in response to a uniform increase in the demand for inside money so that the aggregate demand for precautionary reserves does not fall short of the available supply.12

Algebraically, with f′ and h′ > 0, where R^d is the optimum (minimum) precautionary reserve demand for the banking system, and is a measure of the variance of net clearings for a representative bank around their mean of zero for some level of gross bank clearings, G. In a state of monetary equilibrium, for some given price level, income and gross bank clearings are still positive. Let this monetary-equilibrium value of gross clearings be equal to N. Then, more generally,

G = N + λ (IM^s - IM^d)

where λ >1,13 IM^s is the nominal supply of inside money, IM^d is the nominal demand for inside money, and IM^s - IM^d is the excess supply of (or negative excess demand for) inside money. Thus G increases when there is excess supply of inside money, and falls when there is excess demand for inside money. Since

R^d = f {h [N + λ (IM^s - IM^d)]}

it follows that R^d also increases whenever there is excess supply of inside money, and that R^d falls whenever there is excess demand for inside money. If available reserves, R^s, are fixed, then any change in R^d causing it to differ for R^s must be offset by an appropriate adjustment of IM^s (assuming IM^d is exogenous).

The square-root law of precautionary reserve demand assumes that bank clearings rise or fall due to changes in the frequency of payments. The total volume of clearings may also rise or fall because of an increase or decrease in the average size of individual payments where the frequency of payments is constant. This results in an increase in precautionary reserve demand proportional to the increase in bank clearings.14 Though this possibility gives further strength to most of the conclusions just arrived at, it also points to a potential cause of monetary disequilibrium under free banking. Consider a situation where the volume of gross bank clearings per week is $1 million, consisting of 100,000 checks with an average value of ten dollars. Now suppose that bank customers alter their spending habits by writing only 50,000 checks per week with an average value of twenty dollars. The weekly volume of gross bank clearings is still $1 million, but the smaller number of larger, “lumpier” payments leads to an increased precautionary demand for bank reserves. The tendency (given a fixed volume of reserves) is, therefore, for the supply of inside money to fall. Yet the change in the public’s spending habits reflects, not a smaller, but a greater demand for money balances. So the money supply, rather than adjusting in the same direction as the demand for money (as it does when average payment size is unchanging and the volume of clearings moves inversely with the demand for money) adjusts in the opposite direction.

That this is a potential defect of free banking cannot be denied. But it is unlikely to be of great practical importance. This becomes apparent if one considers that changes in the average size of payments are usually accompanied by changes in frequency in the same direction, in which case their effect is to reinforce demand-accommodating changes in money supply. The exceptional case, where the frequency and average size of payments move in opposite directions, is only likely to occur in response to a change in the general level of prices which is not itself a consequence of monetary disequilibrium. In this case a real balance effect might lead to a change in frequency of payments opposite the change in average payment size. The scope for this kind of price-level change under free banking is rather limited. Suppose though, for the sake of argument, that such a price-level change did occur, causing a disequilibrating change in the supply of inside money. The disequilibrium would be short lived, because its effect would be to reverse the movement in prices that set it in motion to begin with.

Thus the potential damage from disequilibrium money-supply changes under free banking is likely to be very slight. To the extent that such changes could occur, their effect would be to counter somewhat the already limited potential for changes in the general price structure under free banking (such as when there is a general change in productive efficiency per capita). This should be kept in mind in later chapters where the special possibility considered here is ignored and it is assumed that the structure of prices adjusts fully under free banking to reflect changes in productive efficiency.

Variability of the Reserve Multiplier

The variability of reserve demand implies that, under free banking, the reserve ratio (the ratio of reserves to demand liabilities) would vary considerably from bank to bank and also within individual banks viewed at different points of time. Other things being equal, a bank would operate with a lower reserve ratio when the demand for its liabilities is greater and vice versa. For the banking system, in turn, the reserve multiplier (the number of units of inside money supported, in the aggregate, by a unit of outside money) would increase with increases in aggregate demand for inside money balances, and would decrease when aggregate demand for inside money balances decreases (holding the number of banks and their market shares constant).

These results show the conservation theory to be invalid for a free banking system. Moreover, they suggest that it is invalid even for a system with monopolized currency supply (which is less able to accommodate changes in demand) and even where statutory reserve requirements exist. Empirical research supports this. In the United States figures for excess reserves held by banks constantly change, and some economists have even recommended that statutory requirements be modified to reflect the diverse and continually changing turnover rates of liabilities of various banks—with higher requirements for banks with greater deposit turnover. Thus Neil Jacoby (1963, 218-19) recommends that “the legal reserve requirement of an individual bank should be proportional to the contribution of its depositors to the aggregate demand for the total national product.” In this way “banks whose deposits turned over rapidly would be required to carry a higher reserve per dollar of deposit balances than banks with a low deposit turnover.”15 The effect of such proposals, not acknowledged by their authors, would be to have statutory reserve requirements mimic their economic or voluntary counterparts in an unregulated system. Examples of this already exist in the United States. Lower statutory reserve requirements are imposed on certain classes of time deposits than on demand deposits, presumably because time liabilities turn over less rapidly. This arrangement approximates the actual liquidity needs for different kinds of deposits, which implies that, in a deregulated system, it would be at best superfluous.

In a comparative survey of 12 nations Joachim Ahrensdorf and S. Kanesthasan also observed money multipliers that varied substantially over time and from country to country.16 They claim that variations in money multipliers were responses to changes in the behavior of the public, and not simply to changes in the demand for currency relative to total money demand which (under systems with monopolized currency supply) would alter the supply of bank reserves.17

Despite all this there is some justification for accepting the conservation theory as an approximate, though flawed, description of conditions under monopolized currency issue, even in systems without statutory reserve requirements. The reason is that, with a limited supply of high-powered money available to them, the deposit banks are limited in their ability to accommodate increases in the demand for money that involve increases in the demand for currency. The reserve multiplier in a monopolized system, given a constant supply of base money, has an imposed upper limit. Accommodative credit expansion depends on additions to the supply of base money to meet increased needs for currency in circulation. It cannot be accomplished by deposit banks acting alone except when increases in the demand for deposit balances are unassociated with any increase in currency demand.18

For these and other reasons traditional banking studies devote very little attention to the possibility of demand-induced changes in the supply of inside money. Most ignore this possibility entirely.19 Their focus is on supply-side driven changes in the quantity of inside money: changes caused by the injection or withdrawal of sums of high-powered money to and from deposit bank reserves. The chain of causation they consider runs from (a) expansion or contraction of the issues of the monopoly bank to (b) multiple expansion or contraction of deposit bank liabilities (via an institutionally fixed reserve multiplier) to (c) increased or decreased nominal income and prices to (d) increased or decreased nominal demand for inside money and, finally, to (e) monetary equilibrium with nominal variables scaled up or down in the same proportion as the quantity of high-powered money.

However appropriate this approach may be for describing monopolized or central banking, it is unsuitable for describing credit expansion under free banking. Here demand-responsive changes in the supply of inside money are the rule rather than the exception. The relevant chain of causation generally runs from (a) changes in demand for inside money to (b) expansion or contraction to (c) a de facto change in the reserve multiplier and to (d) monetary equilibrium with no change in nominal prices or income. For example, consider a hypothetical free banking system with commodity-money reserves of $1000. The supply of inside money is $50,000 and this is, initially, the amount desired by the public. The reserve multiplier has a value of 50. Now imagine that the demand for inside money falls $10,000 to $40,000. As a result, bank demand liabilities contract $10,000, and the reserve multiplier falls to 40; that is, the final, aggregate ratio of reserves to demand liabilities rises from 2 to 2.5 percent.20

Of course there may also be supply-side driven changes in the quantity of inside money under free banking stemming, for example, from increases in the quantity of outside (commodity) money. The historical significance of such outside-money supply changes will be discussed later in chapter 9. For now it will suffice to note that commodity-money supply shocks have been of only minor historical significance as compared with shocks due to fluctuations in supplies of base monies caused by central banks.

Credit Expansion “in Concert”

The arguments used here to criticize the theory that the reserve multiplier is rigidly fixed also refute the view that the reserve multiplier is, under certain circumstances, indeterminate. That view is implicit in the argument that, with a fixed reserve supply, if banks expand in concert none suffer negative effects even if the expansion is not warranted by any increase in the demand for money.

Eugene A. Agger provides a very clear statement of the indeterminate-multiplier idea:

Thus the system as a whole is supposedly able to expand, on the basis of a fixed supply of reserves, not merely in response to a general increase in the demand for inside money (as was argued in the previous section) but also if demand does not increase.

Keynes takes a similar position in his Treatise on Money:

Two more examples of this view are especially interesting since they refer specifically to free-banking arrangements. The first concerns the note-clearing system supervised by the Suffolk Bank. The author writes that arrangements of this sort “keep the various banks more or less in step with one another in their emission of notes, but would not [prevent] them from preceding too fast (or too slowly) as a whole”:

The second example is taken from Lawrence H. White’s analysis of the Scottish free banking system: “Supposing [a] group of banks expand by a common factor, no consequent adverse clearings will arise among members of the group. Adverse clearings will not arise among a group of banks in consequence of whatever degree of expansion is common to all.”23 If the group comprises a closed system, White continues (1984d, 18), then only an internal drain of reserves to meet the public’s desired commodity-money holdings acts as a check on expansion. Later in his book White notes that “no important theorist of the Free Banking School” explicitly denied the theoretical possibility of in-concert overexpansion by banks in an unregulated system.24 Nevertheless most of them “found the scenario of coordinated expansion implausible as a description of actual events.”25

Even if one grants, following members of the Free Banking School, that in-concert overexpansion is improbable, one might still question whether the supply of inside money under free banking adjusts properly to changes in demand for it. The previous section showed why banks as a group tend to respond positively to a uniform change in demand—refuting the fixed-multiplier view. But we must also confront the possibility that in such a situation the banks can not only respond but can respond to any extent they desire, so long as they act in common. What is to prevent them, furthermore, from “responding” when there is no change in demand at all?

Once again the difficulty is resolved by considering the determinants of precautionary reserve demand. Under in-concert expansion no member of a system of banks expanding in unison (and in the face of an unchanged demand for money) will experience any increase in its average net reserve demand; the change in expected value of its clearing credits will be exactly equal to the change in expected value of its clearing debits. But the growth in total clearings will bring about a growth (though perhaps less than proportionate) in the variance of clearing debits and credits, which increases the precautionary reserve needs of every bank. Thus, given the quantity of reserve media, the demand for and turnover of inside money, and the desire of banks to protect themselves against all but a very small risk of default at the clearinghouse at any clearing session, there will be a unique equilibrium supply of inside money at any moment. It follows that spontaneous in-concert expansions will be self-correcting even without any “internal drain” of commodity money from bank reserves.

Banks as Pure Intermediaries

This and the preceding chapter have attempted to show that, even in the face of changes in the demand for inside money, free banks help to maintain monetary equilibrium. They passively adjust the supply of inside money to changes in the demand for it. They are credit transferers or intermediaries, and not credit creators.

In light of this, and granting appropriate assumptions (namely, a fixed stock of commodity money, with no demand for its use in balances of the public), what can be said about banks in a system with monopolized note issue? A monopoly bank of issue is clearly not a pure intermediary, since the principle of adverse clearings does not apply to it. The position of deposit banks in a system where the supply of currency is monopolized is more complicated. They can respond by a multiplicative expansion to any issues by the monopoly bank that exceed the public’s pre-existing demand for currency. Since such expansion is a response to the exogenous actions of the monopoly bank and not to any change in the money-holding behavior of the public, it involves “created” credit and is disequilibrating. Similarly, if the monopoly bank of issue contracts its issue in excess of any fall in the public’s demand for currency, a multiplicative, disequilibrating reduction in the supply of deposit money will result.

But what role do deposit banks play in the absence of any expansion or contraction by the monopoly bank of issue? In this context deposit banks are more like free banks and other “pure intermediaries”: they cannot, generally speaking, engage in disequilibrating expansion or contraction of the money supply. There are two exceptions to this: First, insofar as the public wish to save in part by holding greater balances of currency, deposit banks are, beyond a certain point, powerless to accommodate their wants without assistance from the monopoly bank. They can issue currency from the monopoly bank held in their reserves only by sacrificing liquidity. Second, changes in the public’s relative demand for currency (i.e., shifts from deposit holding to currency holding and vice-versa) are disequilibrating: they alter the supply of high-powered money available in deposit-bank reserves, and so affect lending power and the total supply of deposit money even though the overall demand for inside money (though not its division between notes and deposits) is unchanged.

Chapter 8 will discuss problems caused by changes in the demand for currency under monopolized issue in some detail. But first, given the conclusions just reached, let us compare our view—that deposit banks are intermediaries of credit—with views of other writers on this subject. J. Carl Poindexter (1946) and James Tobin (1963) have held similar beliefs, as did Edwin Cannan in his much-derided “cloakroom” theory (1921). Cannan denied that bankers are any more capable of lending more than is offered to them than cloakroom clerks are capable of “creating” hats and umbrellas. “The banker,” he wrote, “is able to lend X, Y, and Z more than his own capital because A, B, and C are allowing him the temporary use of some of theirs on condition that he will let them have what they want when they ask for it” (ibid., 32).

Cannan seemed unaware, however, that monopoly banks of issue can create credit by creating new reserves, which throw deposit banks out of equilibrium in their holdings of monopoly-bank liabilities. Tobin, in contrast, recognizes a difference between possibilities for overexpansion of “bank-created” money and those for overexpansion of “government” money. “The community,” he writes (1963, 415), “cannot get rid of” an excess supply of the latter. Therefore “the ‘hot-potato’ analogy truly applies.” On the other hand, “for bank-created money . . . there is an economic mechanism of extinction as well as creation, contraction as well as expansion. . . . The burden of adaptation is not placed entirely on the rest of the community.” Furthermore, for deposit banks acting alone the possibility of credit expansion “depends on whether somewhere in the chain of transactions initiated by the borrowers’ outlays are found depositors who wish to hold new deposits equal in amount to the new loan” (ibid., 413). This is very close to our own view, allowing for the two provisos with regard to currency supply, except that Tobin’s category of “government” money should really include all money issued by any bank with a monopoly in currency supply, whether the bank is officially a government bank or nominally a private one.

Poindexter’s analysis of the role of deposit banks, although less well known than those of Cannan and Tobin, is in some ways superior. Unlike Cannan, Poindexter is fully aware of the credit-creating powers of central banks of issue. But regarding deposit banks he writes (1946, 142): “It is merely the fact that they are at the institutional center of the credit-creating and credit-destroying process of the community that gives their role the apparently unique character which is commonly imputed to them.” In fact, Poindexter argues, deposit banks, like other private competitive financial institutions, cannot lend beyond what their depositors desire unless the central bank that operates alongside them alters the “data” of the system, to which they respond (ibid., 143-44). Otherwise deposit banks are merely “the institutional media through which the public determines the volume of bank deposit currency which will be created at any given moment” (ibid., 142).

Controversy has surrounded the views of J.G. Gurley and E. S. Shaw, who first argued, in a series of articles,26 that banks always function as pure intermediaries—responding through profit signals to the wants of the public—and who later modified their position and claimed that banks and non-bank financial firms alike are equally capable of active credit creation.27

The error of these authors’ earlier writings lay in their use of an ex post definition of savings. This approach failed to distinguish properly individuals’ voluntary abstinence from purchasing from their involuntary abstinence due to forced saving. “Pure intermediation” should refer to credit operations based on voluntary savings only.28

The early Gurley and Shaw view does not really differ from Cannan’s “cloakroom” approach, which also failed to recognize that certain kinds of banks, namely those having a monopoly or quasi-monopoly in the issue of currency, can indeed create credit, and that deposit banks also could contribute to this credit creation by responding to changes in their holdings of high-powered money having its source in unwarranted issues by a privileged bank.

In their later work, on Banking in a Theory of Finance, Gurley and Shaw commit the even more serious error of claiming that all financial institutions are equally capable of actively creating credit. Because of its failure to recognize the role of monopoly banks of issue as the ultimate source of created credit this view has served as a rationale for maintaining legal restrictions on credit expansion by deposit banks and for imposing similar restrictions on savings institutions and other non-bank financial intermediaries.29 Such restrictions not only interfere with efficient intermediation, but reinforce the erroneous notion that competitive financial firms are independent sources of inflation, which the central bank has to “control.”

Students of banking theory often get the impression that central banks are uniquely capable of preventing monetary disequilibrium: they are not inclined to think of them as throwing a wrench in the works. Yet, in contrast to deposit banks and to banks in a free banking system, central banks (or any bank with a monopoly or quasi-monopoly in currency supply) have a unique capacity for generating monetary disturbances. The question that has to be asked, therefore, is whether the disturbances central banks perhaps prevent outweigh the disturbances they cause that would otherwise not occur.

[1.] Good statements of the rule of excess reserves are in Chester Arthur Phillips (1920, 33-34), Robert G. Rodkey (1928, 178-85), and Alex N. McLeod (1984, 201).

[2.] See Rodkey (1928, 42). The above argument assumes that, after the first clearing round, Bank A must reduce its loans $40,000 to restore its reserves. The simplest means by which this might be accomplished is if those persons holding the $40,000 of new money balances at Banks B through E use them to buy goods from a debtor of Bank A, who in turn pays back a $40,000 loan. The same sort of transactions can be imagined to occur after each clearing round. By this means one can most readily see that Bank A must soon contract by the full amount of its overexpansion.

[3.] Banks might also try to attract more depositors by offering them higher rates while earning the interest through riskier loans and investments. This of course is just a definition of bad banking: the high-risk loans reduce bank revenues in the long run, as default occurs. But the bank may grow inordinately in the short run. Such banking has been behind many of the growing number of bank failures in recent times. It would not, however, be common in unregulated circumstances, where bank owners bear the full costs of failure. Its frequency today must be blamed on regulatory arrangements, including Federal deposit insurance, bank bailouts, and the promise of Federal Reserve support, that subsidize excessive risk-taking and artificially limit losses to bank stockholders.

[4.] The argument proves only that spontaneous in-concert absolute overexpansion is unlikely; it does not show why, facing a uniform fall in the demand for money, banks as a whole should contract their liabilities, or why a planned in-concert overexpansion would not be sustainable. A more general treatment of the problem of in-concert overexpansion appears in chapter 6.

[5.] If currency is used by the public some of the new reserve media may pass into circulation, lessening the potential expansion. In an open economy reserve media useful in international trade might also be lost to foreign banks. Here we assume that the public’s demand for currency does not increase, and that the economy is a closed one.

[6.] Ludwig von Mises ([1928] 1978, 138). “Circulation credit” is Mises’s term for bank liabilities not backed by commodity-money reserves. Elsewhere Mises uses the term “fiduciary credit” (see chapter 4 below).

[7.] It is reasonable to assume that most non-retail purchases are made using checks rather than notes.

[8.] According to Checkland (1975), the Scottish banks encouraged their customers (depositors and borrowers) to show note brand loyalty by paying in the notes of rival banks, while making their advances to others with their own bank’s notes. Pressure to “push” a bank’s notes was most effectively placed on those in debt to it.

[9.] Such behavior is assumed by Eugene E. Agger (1918, 103).

[10.] This example, as well as one to follow, assumes that new notes are initially issued in a proportional loan to everybody. This assumption is the strongest that can be made in favor of a “predatory overexpansion” scenario. If it is assumed, more realistically, that new issues are first made available to one person or group only, more of them will enter the clearing mechanism, ceteris paribus.

[11.] Compare L. White (1984d, 97fn), who employs somewhat different assumptions and concludes that an overexpanding bank’s reserve losses will fall as the number of its rivals increases.

[12.] We have already seen why it makes little difference if some notes return directly to their issuers.

[13.] For statistics on the note circulation of various banks during the Scottish free banking era, see L. White (1984d) and Munn (1981). For statistics on note issues of Canadian banks in the 19th century, see the Dominion Bureau of Statistics Canada Yearbook. For evidence from the Swedish experience, see Jonung (1985).

[14.] The existence of legal-tender status for the notes of a privileged bank of issue, although it encourages their general acceptance, is not essential to their being held for reissue by deposit banks.

[15.] To recall, we are assuming that there is no public demand to hold coin as a proportion of total money balances. The conclusion also ignores the possible presence of industrial (nonmonetary) demands for the money commodity.

[16.] Except where otherwise noted the facts cited are from Copland (1920).

[1.] See Gilbert (1953, 144).

[2.] In terms of the Cambridge equation of exchange, this means a change in the value of “k.”

[3.] See Clayton (1955, 96) and Brown (1910)

[4.] “In extending any particular individual a loan of a certain sum of money, the lending bank is in effect conferring upon the borrower a claim to a corresponding fraction of the wealth of the community whose real value is matched by the real value of goods which some anonymous depositor [or note holder] has refrained from exercising the right to consume” (Poindexter 1946, 135). Compare Hutt (1952, 237ff).

[5.] Gottfried Haberler (1931, 19) notes that such downturns may be reinforced by the aggravation of debt burdens as a result of the unanticipated fall in prices. This effect may for some time remain “an obstacle to recovery, unless relief is found in the shape of a crop of bankruptcies.”

[6.] See Friedrich A. Hayek ([1933] 1975b and 1935). Also see Ludwig von Mises ([1928] 1978, 59-171). The Austrian economists’ relative lack of attention to the problems of deflation is due in part to their views on the requirements for credit market equilibrium. These are discussed below in this chapter.

[7.] See Friedrich A. Lutz (1969).

[8.] Hayek (1935, 121). See also p. 91, where Hayek writes that “to be neutral . . . the supply of money should be invariable.” Such statements give credence to the view that Hayek advocated a “do nothing” policy for business cycles. See for example Lawrence R. Klein (1966, 51). Of course, given monetary institutions that exist today and those that existed in the 1930s, even a constant money supply is not really the same as a do-nothing policy.

[9.] Hayek (1935, 107). Hayek refers in particular to the need to accommodate changes in the demand for money due to the multiplication of stages of industrial production.

[10.] Durbin’s views were challenged by J. C. Gilbert (1934). This was followed by Durbin’s reply (1935) with a “Rejoinder” from Gilbert and a final note by Durbin (ibid., 223-26).

[11.] The discussion in Money is clearer but at the same time less complete.

[12.] Robertson (1926, 53-54). Robertson’s “Automatic Lacking” has the same meaning as the notion of “forced saving” that I discuss below in this chapter. The expression “cloak-room” banking is a reference to the views of Edwin Cannan, also discussed below.

[13.] Thus, for example, Allan G. B. Fisher writes (1935, 200) that “Apart from increases in population and from changes in the desire of individuals to hold money, economic development which takes the form of increased production per head . . . does not require any increase in the money supply.” See also Robertson (1964, 80-82 and 111-14).

[14.] See for example Lloyd Mints (1950, 129-30).

[15.] Keynes (1936, 167fn) refers specifically to bank deposits. In some respects the arguments in Keynes Treatise on Money (1930) have even more in common with those of other authors cited above.

[16.] See for example Keynes (1936, 16, 303).

[17.] This aspect of Keynesian analysis provoked Jacob Viner to remark prophetically in 1936 that “In a world organized in accordance with Keynes’ specifications there would be a constant race between the printing press and the business agents of the trade unions.” See Viner ([1935] 1960, 49).

[18.] These terms are taken from Machlup (1940, 231-32). A third kind of credit discussed by Machlup (232-37) is credit granted out of “surplus cash balances.” This results from reductions in the public’s demand for base money in circulation. Since base money is assumed not to circulate under free banking (where bank notes supply demands for currency) this type of credit expansion is not relevant to it. On the other hand, it is relevant for the case of central banking if central bank notes or fiat money are used as currency. In this case it may be regarded as a special type of created credit.

[19.] When, on the other hand, the demand for money increases but its supply does not increase correspondingly (i.e., when there is a failure on the part of banks to issue transfer credit) the effect, in Machlup’s terminology, is one of “credit destruction.” The contraction of bank credit in the face of an unchanged demand for money is also an example of this.

[20.] This way of putting it seems preferable to Machlup’s definition (171) of created credit as credit that provides “purchasing power . . . which has not been given up by anybody before hand.” This might be interpreted as including in “created” credit credit granted in response to an increased demand for inside money (which is actually transfer credit), since persons who add to their balances of inside money do not necessarily sacrifice “purchasing power.” What they sacrifice is actual purchasing, which is something different. The trouble lies with the expression “purchasing power,” which sometimes refers to a potential to purchase, and sometimes to the exercise of that potential. Because of its ambiguity I try to avoid using this expression.

[21.] Good, brief discussions of this are Pigou (1933, 227-31), Clayton (1955, 98-101), and Robertson (1964, chap. 5 sect. 3 and 173ff). Pigou uses the expression “forced levies” instead of forced savings. Keynes (1936, 183) called the doctrine of forced savings “one of the worst muddles” of neoclassical economics. But he also associated it with Hayek’s recommendation that the money supply should be kept constant. Keynes might have rejected Hayek’s constant money supply bathwater without throwing out the forced-savings baby. For a critical discussion of the role of forced savings in Keynes’s thought, see Victoria Chick (1983, 236-39).

[22.] Quoted in Hayek (1935, 20).

[23.] Clayton (1955, 99). Compare Machlup (1940, 171). Clayton’s last statement would be accurate only if capital goods were homogeneous, and capital “sunk” as a result of credit creation were not a cause of permanent changes in the structure of production.

[24.] In addition to works of Hayek and Mises cited in the text see Ludwig von Mises ([1953] 1980, 359-67; 1949, 545-73; 1966, chap. 20) and Murray N. Rothbard (1970, 850-63).

[25.] Mises (1949, 439fn). Compare Rothbard (1970, 862), and Hayek (1935, 23).

[26.] Compare Mises (1980, 300-301).

[27.] In the writings of Rothbard (1970, 850-60 and 1962, 115ff) this tendency is complete.

[28.] On these authors see Vera Smith (1936, 91-93 and 110-12).

[29.] See Mises (1966, 443).

[1.] See Richard G. Davis and Jack M. Guttentag (1962), who describe compensating balances as a way for banks to guarantee that their borrower-customers will hold their working balances with them rather than elsewhere. They also note (123fn) that in many instances the requirements, rather than being based on “hard and fast agreements,” take the form of informal understandings.

[2.] See Harold Barger (1935, 441). Barger, however, believes that monetary equilibrium can be maintained only if credits are restricted to producer (and not consumer) loans. His position seems to be based on the view that “non-productive” loans will not generate interest necessary for their repayment. This view overlooks the fact that consumer-borrowers can pay interest out of their future income even if this involves a reduction in wealth. Thus, banks’ granting of loans to consumers, although it does not necessarily contribute to capital accumulation, is still consistent with the preservation of monetary equilibrium.

[3.] The process of inside-money absorption is discussed in Shotaro Kojima (1943, 17-18).

[4.] It is assumed that before being paid-in the liabilities were strictly “idle,” so that they were not a cause of any bank clearings.

[1.] By analogy with the “law of conservation of energy” (the first law of thermodynamics), which states that energy can be moved from one place to another but cannot be created or destroyed.

[2.] “That any single person can make his own balance at [a] bank rise by paying in money . . . (whether in cash or in checks) and make it fall by withdrawing cash or paying away checks, everyone who has ever had a balance to his credit knows. No one denies this, but some theorists have denied that what is true of each lender taken separately is true of the whole body of lenders taken together.” Edwin Cannan, “Growth and Fluctuations of Bankers’ Liabilities to Customers” (1935, 8).

[3.] It is generally recognized that, in systems with monopolized currency supply, changes in the demand for currency relative to deposits will alter the base-money multiplier by causing base money to shift between bank reserves and circulation. This perverse adjustment is different from the accommodative adjustments considered here, which are adjustments in the supply of inside money in response to changes in demand when the supply of bank reserves is constant.

[4.] George Clayton (1955, 97-98). The only exception Clayton allows for is the case of a transfer of funds from demand accounts to time-deposit accounts. This reflects a presumption that different statutory reserve ratios apply to these types of accounts. On the reliance of the conservation theory on the assumption that there are binding statutory reserve requirements see below.

[5.] The turnover of liabilities will change temporarily if the demand for them changes (with constant supply) as consumers attempt to spend off excess balances or to add to their deficient balances.

[6.] Following J.H.G. Olivera (1971, 1096), “reserve demand” is used here to indicate needs arising in connection with bank clearing transactions that “make it necessary for the reserve holder to transfer some amount of the reserve asset.”

[7.] Ernst Baltensperger (1974, 205) defines precautionary reserve demand as the “excess of total holdings of the reserve asset . . . over the expected or average net demand for (‘loss’ of) reserves.” Olivera (1971) defines it as “the part of the total reserve which is held against possible deviations of net demand above its expected value.”

[8.] That is, so long as “reserve changes are known in probabalistic form only” (Baltensperger 1974, 205).

[9.] It is assumed throughout this analysis that the demand for reserves is uninfluenced by changes in interest rates on loans and investments. In defense of this assumption it may be noted that interest rates on overnight, “emergency” loans to cover reserve deficiencies will tend to rise along with other rates of interest, so that the penalty costs for insufficient reserve holding increase with the opportunity costs of keeping adequate reserves on hand. This suggests that high interest rates do not necessarily make it desirable for banks to skimp on reserves. For arguments and evidence in support of this view see Leijonhufvud (1968, 358), and Hancock (1983). Of course, if high rates are passed on to deposit holders, this might increase the quantity of money demanded and so reduce indirectly the demand for reserves.

[10.] F.Y. Edgeworth (1888). The best recent articles on this subject are the ones previously cited by Baltensperger and Olivera. In addition to these articles the discussion in Don Patinkin (1965, 82-88) is recommended.

[11.] As Baltensperger notes (1974, 205), the “square-root” law gives a conservative estimate of the relation between changes in gross clearings and precautionary reserve demand, in part because it assumes an increase in frequency, rather than in average size, of transactions. Edgeworth’s demonstration of the square-root law also relies on the assumption that individual clearing debits are stochastically independent and identically distributed. Olivera shows, however, that the law holds even if individual clearing debits (“the components of net average demand”) are serially correlated.

[12.] Whether adjustments in nominal supply of bank liabilities will entirely offset changes in nominal demand depends on the extent of note-brand discrimination. An increase in demand among nondiscriminating persons results in a smaller reduction in precautionary reserve needs than an equal increase in demand (affecting all banks uniformly) of discriminating persons. When there is 100 percent (marginal) note-brand discrimination, nominal supply adjustments will be complete.

For the sake of simplicity the argument assumes that banks are in equilibrium with respect to one another, that is, it assumes that the average net demand for reserves is zero. Then precautionary demand for reserves = total demand. Olivera (1971, 1100) notes that the square-root law is relevant to “decision units taken individually” and that “its possible use as a macroeconomic relationship involves a nonlinear aggregation problem.” He adds, however, that “the obvious ‘aggregation condition’ is that the number of reserve-holders, as well as their shares of the expected market demand, remain stationary when the latter grows.” But this simply means that it is necessary to abstract from changes in average net reserve demand, which is precisely the procedure I have adopted.

[13.] For the banking system, every dollar’s worth of excess money supply generates several dollars’ worth of additional bank clearings during any extended (but finite) planning period.

[14.] See Patinkin (1965, 87-88 and 576-77).

[15.] Jacoby (1963, 220). Similar proposals were made in the 1930s, following the lead of Winfield Riefler. See George Garvey and Martin R. Blyn (1969, 56-57).

[16.] Joachim Ahrensdorf and S. Kanesthasan (1960). For evidence on cross-sectional and temporal variation in Scottish free banks’ reserve rations, see Munn (1981).

[17.] See note 3 above, and also chapter 8.

[18.] Once again I am abstracting from problems arising due to changes in the demand for currency relative to total money demand. These problems are discussed at length in chapter 8.

[19.] See for example Albert E. Burger (1971), which is one of the more detailed, modern discussions of factors influencing the money supply. McLeod, whose analysis is also very detailed, merely hints at the possibility of a demand-elastic money supply when he notes (1984, 100) that the “credit multiplier” may rise to infinity, while the “total income multiplier” associated with a given increase in credit (bank loans and investments) may at the same time approach zero if increased lending is offset by increased holdings of inside-money balances. McLeod cites borrowings used to increase borrower’s liquidity—a case similar to the one of compensating balances discussed above—as a limiting case. Our claim is a much stronger one, viz, that under free banking changes in money supply generally do not influence total income and spending.

[20.] The example assumes 100 percent marginal note discrimination.

[21.] Eugene E. Agger (1918, 101). Were he writing about an unregulated system Agger might also have referred to an increase in the volume of notes.

[22.] Keynes (1930, 1: 27). Keynes’s two premises are incorrect: an individual bank may “move forward” on its own without weakening itself if the demand for its liabilities has increased, and overexpansion by one bank or set of banks generally will not inspire sympathetic overexpansion by remaining ones.

[23.] L. White (1984d, 17). White presumably meant to say “by the same amount” rather than “by a common factor.”

[24.] The more important members of the Free Banking School were Sir Henry Parnell, Samuel Bailey, and James William Gilbart. (For other names see ibid., 52.)

[25.] L. White (1984d, 98). Not surprisingly, opponents of free banking also accepted the in-concert overexpansion argument, and were in addition more willing to view it as a description of the likely course of events under unregulated banking. They included J. R. McCulloch and Samuel Jones Loyd (cited in ibid., 98-99), and G. W. Norman, a Director of the Bank of England (cited in V. Smith, 68).

[26.] See especially Gurley and Shaw (1955, 1956).

[27.] Gurley and Shaw (1960, 202, 218).

[28.] Joseph Aschheim, in response to Gurley and Shaw, argues (1959, 66) that commercial banks “can make ex post savings exceed ex ante savings,” i.e., can engage in credit creation that leads to forced savings. In contrast, Aschheim says, other financial institutions “can lend no more than they have received from depositors and, therefore, cannot create loanable funds.” What I have tried to show, in contrast, is that deposit commercial banks are mainly passive “credit creators.” The only active credit creators are those institutions having a monopoly or quasi-monopoly in the supply of currency.

[29.] See, for example, James M. Henderson (1960).