Chapter 8 CRITICISMS

The body of conventional literature on monetary matters has, for many years, included several propositions that explicitly challenge the viability and/or desirability of free banking. Chapters 2, 3, and 4 presented a detailed theoretical case for free banking on the grounds that it would maintain nominal national income, reflect consumer preferences in the time and money markets, keep the market rate of interest equal to the natural rate, achieve continuous monetary equilibrium, and avoid business cycles. Despite those arguments, in order to complete the discussion, one must address the standard criticisms. That is, one must examine the key points in the case against free banking. The issues so often raised in opposition to private money and laissez-faire banking include the following claims: (1) that money is a “public good” and thus cannot be profitably produced in a free market; (2) that there exist significant “external effects” such that private money balances would be suboptimal; (3) that money production is a natural monopoly, that is, that marginal and average costs decline over the relevant range of output; (4) that competition in money will lead to massive inflation; (5) that free banking is inefficient in that it represents a waste of resources; (6) that privately produced moneys suffer from serious counterfeiting problems; (7) that central banking arose in response to true consumer needs, that is, that it is central banking rather than free banking that represents the natural evolution of money markets; and (8) that a lender of last resort (a central bank) is necessary in order to prevent or mitigate financial crises.

MONEY AS A PUBLIC GOOD

The precise meaning of “public good” (sometimes called “collective good”) is still the subject of controversy.¹ No doubt, the most common interpretation of the concept is that any good that possesses both “nonrivalrousness” and “nonexcludability” is a public good. Nonrivalrousness means “that a given quantity of a public good can be enjoyed by more than one consumer without decreasing the amounts enjoyed by rival consumers” (Hyman 1987, 114). Nonexcludability “implies that it is too costly to develop a means of excluding those who refuse to pay from enjoying the benefit of a given quantity of a public good” (Hyman 1987, 115). If something is a public good, the implication is that “free riders” will exist, that is, that some people will be able to benefit from the good without having to bear the cost of its acquisition. If money is a public good, then free banking may be decidedly inferior to central banking.

First of all, one must realize that, despite the pervasive use of the term both within and without the economics profession, the concept of public goods may be questionable. At the most fundamental level, one may ask: Where is the “public,” the “collective,” which demands the good in question? “Only an individual exists . . . there is no existential referent of the ‘collective’ that supposedly wants and then receives goods” (Rothbard 1970, 884).² Two of the most frequently cited examples of public goods—national defense and police protection—are problematic. Murray Rothbard brings up a thorny problem that besets these two. He points out that neither can really be “collectively” demanded, since pacifists do not want national defense and criminals do not want police protection (1970, 884). Some would, of course, reply that these groups, especially the pacifists, represent small minorities within society; the vast majority may have intense desires for defensive agencies. Be that as it may, one cannot conclude that there is an unambiguous “collective” demand. To do so is to indulge in the fallacious technique of making interpersonal comparisons of utility.

Rothbard furthermore questions the notion of public goods as having an indivisible supply, from the benefits of which no one can be excluded. “ ‘[N]ational defense’ is surely not an absolute good with only one unit of supply. It consists of specific resources committed in certain definite and concrete ways. . . . A ring of defense bases around New York, for example, cuts down the amount possibly available around San Francisco” (1970, 885).

An additional facet of the issue should be mentioned. The whole discussion of public goods, external effects, and free riders presupposes knowledge of the “optimal” production level for the good in question. This cannot be known outside the artificial world of a perfectly competitive general equilibrium. Moreover, any inefficiencies that do arise may be the result of the failure (not of the market, but of the legal authorities) to either define or enforce property rights (Cordato 1992, 16–23).

For the sake of argument, one may grant that public goods exist. Some economists conclude that money is a public good, because “any one agent, holding cash balances of a given average size, is less likely to incur the costs of temporarily running out of cash, the larger are the average balances of those with whom he trades” (Laidler 1977, 321fn). If that were sufficient to establish money as a public good, however, would not all goods be public goods? Surely, the greater the stocks held by bookstores, for example, the less likely every purchaser of books is to “run out of books.” Does that mean that books are public goods? Obviously not. David Laidler, among others, simply overlooks the essential question regarding public goods: Can person A benefit from a particular quantity of good X possessed by person B, even when B might prefer that A not receive that benefit? Regarding money, the answer is no.

[A] particular sum of inside money renders its service—increased purchasing opportunities—only to those who actually possess it. Those who refuse to do without other forms of wealth or who do not abstain from consumption (by holding inside money instead of consuming a flow of services from goods) cannot take advantage of the benefits associated with inside money. (Selgin 1988a, 154)

In short, since only the owner of any given sum of money can benefit from its liquidity services, then money is neither nonrivalrous nor nonexcludable (Vaubel 1986, 934). Thus, it cannot be a public good on those grounds.

Others perceive money as a public good, because the various specific moneys serve as units of account—the dollar, the pound-sterling, the yen, and so forth. That is, the unit of account is also the unit of the medium of exchange. The argument here is that using a particular unit of account creates a positive externality by way of conveying valuable information that all may use. “Uniformity in quoting prices, maintaining accounts, and stating forward contracts has the same virtues as uniform standards for switching and transmitting telephone calls” (Hall 1984, 306), but this does not constitute a serious challenge to free banking of the sort (WS) modeled and defended in this book. At best, all it demonstrates is that “government should suggest a unit of account and publish a price index for it, but not that it should supply money” (Vaubel 1986, 935).

One might say that quoting prices in, and transacting with, a single money eliminates some formidable accounting and transaction costs. . . . But just to avoid the inconvenience of multiple monetary units, money does not have to be supplied monopolistically. If many suppliers of money were making their moneys convertible into the same asset, all the moneys would be exchangeable at a one-for-one rate. The benefits of a single monetary unit could be secured without monopoly. (Glasner 1989, 28)

It must be admitted that the unit of account argument would be germane in the context of a proposal such as Friedrich Hayek’s (1978). There banks issued their own distinctive banknotes, and each bank might use a different unit of account. This is yet another reason why a Hayekian system is unlikely to exist (see Chapter 1). A single unit of account is superior to a multiplicity of such units. Throughout this study, free banking has implied multiple currencies all convertible into the same commodity and all denominated in the same unit of account.

Karl Brunner and Allan Meltzer (1971) offer a variation on this theme when they argue that money is itself a substitute for information because it (potentially) reduces transaction costs. Since information is presumed to be a public good, then money must be, too. This argument fails also. It is clear that “to show that X is a substitute for a public good is not sufficient to prove that X is a public good . . . what has to be shown is not that money is a substitute for information but that it provides the public good of information” (Vaubel 1986, 935).

Even if money does, however, provide information (or knowledge), a problem remains: Is economic knowledge a public good?³ Superficially, it appears to be, since it seems nonrivalrous and nonexcludable. Nevertheless, it is not. The common supposition that all “true” knowledge, that is, correct factual statements, must be a public good arises from a strong positivistic turn of mind. That is to say, whosoever argues that economic knowledge is public in nature forgets that what is important (economically, at least) is not if the data are “correct” or not, but in what manner they are perceived to be useful by some specified economic agent. How else is it possible for two different market participants with identical endowments to examine the same price data series (for some commodity such as copper, say) and yet reach opposite conclusions (one buys copper; one sells copper)?

Gerald O’Driscoll and Mario Rizzo argue similarly that economic knowledge is fundamentally subjective in nature. What is perceived as “knowledge” depends upon the circumstances, values, and goals of the perceiver (1985, 35–50). As they suggest, there can be no such thing as public knowledge—all useful economic knowledge is private by its very nature.

SUBOPTIMAL MONEY BALANCES

Related to the public good argument is the oft-repeated assertion that a free market in money will lead to suboptimal money holdings by the public.⁴ This has been most forcefully proposed by Milton Friedman (1969, 16–48). Friedman argues that optimal real money balances will be held only if it is true that the marginal nonpecuniary yield from money (MNPS_m) is equal to the nominal rate of return on interest-bearing assets (r_b), with that rate equal to zero. That is, optimality follows from MNPS_m = r_b = (Friedman 1969, 36). If MNPS_m = r_b>, then suboptimal money balances result. Friedman’s concern stems from his contention that the marginal private cost of increasing one’s cash balances (assumed equal to the rate of return on nonmoney assets) will be greater than the marginal social cost of producing greater money balances (assumed equal to zero). It is clear that Friedman further assumes a fiat currency rather than a commodity-backed currency. He offers two solutions: (1) a fully anticipated, steady rate of price-level deflation that provides a pecuniary yield on money balances or (2) the payment of explicit interest on money balances. Both are intended to offset the difference between private and social marginal costs.

There are problems with Friedman’s analysis. First of all, the private cost of acquiring money is not the forgone interest that one would have gained from some alternative financial asset. The cost of acquiring money (its price) is the goods and services one forgoes. That is—as Friedman himself elsewhere declares (1972, 201)—the price of money is its purchasing power in terms of real goods (see Chapter 2). It is the cost of holding money that is correctly identified as the forgone interest income from alternative assets (Selgin 1988a, 52–53). Indeed, to forgo income from interest-bearing assets is the cost of holding any non-interest-bearing asset, but most assets are never used directly to acquire goods and services. Money is the intermediary; the cost of acquiring money is measured in real goods or services.

Also, one might question whether the concept of marginal “social” cost is even meaningful. Friedman assumes that his use of the concept is noncontroversial and, furthermore, that such costs are objective and quantifiable. Others disagree. Steven Cheung argues that most applications of the concept have been inappropriate because of either invalid specifications of the constraints involved or incorrect observations of events (1980, 51). He further suggests that the frequent misuse of “social cost” has “hindered the advancement of economics as a behavioral science” (1980, 51). James Buchanan reminds one that “cost is subjective; it exists only in the mind of the decision-maker or chooser . . . the opportunity cost involved in choice cannot be observed and objectified” (Buchanan and Thirlby 1981, 15). Stephen Littlechild insists that “social cost and production are not objective concepts” (1978, 88). O’Driscoll and Rizzo reinforce this position: “Cost, just like utility, is defined over projected want satisfaction and not directly over the commodities themselves. . . . Commodities are only way-stations to the ultimate satisfaction of basic wants” (1985, 47). As long as social cost is neither objective nor quantifiable, Friedman’s optimum quantity of money rule remains questionable.

Kevin Dowd—himself an advocate of a stable price-level rule—makes two interesting comments on Friedman’s analysis. First of all, he estimates “the welfare loss under a zero-inflation regime implied by an inability to pay interest on currency holdings” (199le, 7). This he finds to range from .00007 percent to slightly over 0.1 percent of 1991 Gross Domestic Product (GDP) for the United Kingdom (1991e, 9). Since this might be taken as a measure of the potential gain from achieving optimal money balances, it appears that such gains are trivially small. Second, Friedman seems to have realized how relatively small the gains would be, for “his own results did not persuade him to abandon the goal of price stability for the further benefits of optimal deflation” (Dowd 1991c, 12).⁵ The argument for optimal money balances cannot be very compelling when its own advocate opts for a different approach.

Furthermore, it is interesting to notice that Friedman’s analysis—by his own admission—is based in part on the existence of “perfect capital markets” as that term is used in the pioneering work of Franco Modigliani and Merton Miller (Friedman 1969, 35fn). Modigliani and Miller (MM) assume that (1) all transactions are costless, (2) debt is riskless, (3) capital markets are perfectly competitive, (4) all individuals and firms are able to borrow and lend at the same rate of interest, and (5) all firms can be sorted by risk class, with all the firms in each class exhibiting identical return distributions (1958). From these assumptions, MM conclude that in the absence of taxes, “the market value of any firm is independent of its capital structure” (1958, 268). In other words, the method of firm financing is irrelevant. If corporate taxes exist (as they do) and firms are allowed to deduct interest payments on debt as expenses, then MM propose that firm value is maximized at the point of 100 percent debt financing.

Does this proposition accurately predict actual firm behavior? Charles Haley and Lawrence Schall point out that “firms do not increase debt to the upper limit possible even though interest is tax-deductible” (1979, 290). The flaw may lie in the extremely artificial assumptions of the MM model. In reality, transactions are not costless, capital markets are not perfectly competitive (information is not costless and extranormal profits can be made), not everyone can borrow and lend at the same interest rate, and firms cannot readily be sorted into risk classes (Haley and Schall, 1979, 289–90). This is unstable ground upon which Friedman builds his argument.

Finally, George Selgin, without questioning the concept of “social cost,” offers a reply to Friedman. He points out that under free banking, competition would, of course, compel the payment of interest on deposits.⁶ However, interest payments on banknotes might indeed prove impossible because of high transaction costs. Therefore, “there would be a suboptimal quantity of banknotes, but the loss from this would be partly offset by a supraoptimal quantity of deposits. The only net loss would be that stemming from any inelasticity of substitution between deposits and notes” (emphasis in original) (Selgin 1988a, 156).⁷ Selgin goes on to ask whether government intervention might perform better. He thinks not. A central bank “is less likely to attempt interest payments on currency than private note issuers, since its monopoly privilege places it under less pressure to do so” (1988a, 156). In short, free banking would come closer to an “optimal quantity” of money than would central banking.⁸

BANKING AS A NATURAL MONOPOLY

A standard proposition of modern microeconomic theory is that if there exist significant economies of scale, that is, if long-run average costs decline over the range of output demanded in the market, then only one firm will survive. Such a firm would be a “natural monopoly.” It has often been argued that banking may be a natural monopoly, and if so, the most efficient approach is to restrict competition and allow only a single issuer of banknotes, the central bank. In short, if there are large cost economies in banking, then free banking may not be an optimal solution.

The most basic rebuttal to the natural monopoly argument is to point out that costs cannot be known to the producer (much less to the economist) prior to the process of production and a firm’s costs are likely to change when the market structure changes. From this it follows that

a governmental producer of money is not an efficient natural monopolist unless he can prevail in conditions of free entry. . . . The only operational proof that a common money is more efficient than currency competition and that the government is the most efficient provider of the common money would be to permit free currency competition. (Vaubel 1986, 933, 935)

That is, free banking is the necessary precondition for discovering whether or not banking is a natural monopoly. In the absence of such competition, those who claim that banking is a natural monopoly are guilty of making an unsupportable assertion.

Are there any logical reasons for expecting economies of scale in banking? There are two: economies of scale in reserve holding and economies of scale resulting from diversification (Dowd 1991b). The first is based on the widely accepted principle that reserves change with the square root of a bank’s liabilities. This suggests that the optimal reserve ratio will decline as the bank expands, which further suggests that the bank’s average costs will fall. However, these economies are extremely small. Citing David Glasner, Dowd states that “a bank with only $10,000 of liabilities would exhaust 99 percent of the possible savings in holding reserves” (1991b, 4). Furthermore, this estimate may exaggerate the benefits if reserves are interest-bearing or small banks can pool their reserves instead of merging (Dowd 1991b, 4). A second source of scale economies could be asset diversification. A large bank might experience lower average transaction and delegation costs because of its large and well-diversified portfolio of assets. Once again, however, these gains appear far too small to lead to natural monopoly (Dowd 1991b, 7).

One approach to the natural monopoly argument is actually to estimate cost functions for commercial banks. A number of studies have tried to ascertain whether or not banking exhibits either economies of scale (cost savings resulting from a larger scale of operation) or economies of scope (cost savings resulting from producing multiple products that share common inputs). The obvious problems with all such contemporary studies are that they are not estimating costs in a free-banking environment and they base their calculations on accounting costs rather than economic costs. The ideal study would be one that estimated economic cost functions in a laissez-faire context, but since true free banking has never existed, those recent studies, although a “second-best” approach, may be helpful.

Loretta Mester explains that “early cost studies treated financial services as a single product . . . many of these studies found that the average cost of production falls as more is produced—that is, there are economies of scale” (1987, 15–16). However, banks are not single-product firms; they offer loans, investments, deposit accounts, and currency of a sort (traveler’s checks). That variety of products makes it possible for them to enjoy economies of scope as well as scale. More recent studies have often therefore modeled banks as multiproduct firms. The results have been consistently the same. Economies of scale seem very small, although some studies find evidence that they do exist. See George Benston, Allen Berger, Gerald Han week and David Humphrey (1983), Benston, Hanweck, and Humphrey (1982), Berger, Hanweck, and Humphrey (1986), Thomas Gilligan and Michael Smirlock (1984), or Gilligan, Smirlock, and William Marshall (1984). Mester concludes that “there is no evidence that larger firms have a cost advantage over smaller firms” (1987, 24).

For example, Jeffrey Clark (1984) estimates the output elasticity of cost, (∂C/∂Q) (Q/C), for a sample of 1,205 banks using both log linear and generalized functional form cost functions. If such elasticity is significantly less than 1, one can conclude that scale economies do exist. Clark finds that the measure of this elasticity is “never below 0.95. Thus economies of scale in the banking industry appear to be small and operating efficiency is unlikely to be substantially improved by an increase in bank size” (1984, 67). To this may be added the research of David Humphrey, who similarly estimates cost elasticities for 13,000 commercial banks, grouped by asset size. Based on his results, Humphrey asserts that “there appears to be no strong reason to constrain bank mergers or inhibit nationwide banking for fear of conferring important cost advantages on large banks . . . reliance on the cost or scale economy argument is not supported by the data developed here or in other recent studies” (1987, 37).

The economies of scale that do exist seem to disappear rather quickly. After surveying nine studies of banking costs, Clark declares that the economies are gone by the time a bank has acquired $100 million in total deposits (1988, 26). “The conclusion seems to be that increasing returns to scale exist but are limited, and it is surely significant that not a single study finds evidence that banking is a natural monopoly” (Dowd 1991b, 14). The natural monopoly argument as a critique of free banking appears to rest on a weak empirical foundation.

FREE BANKING AND INFLATION

Many economists have assumed that laissez-faire banking and inflation (perhaps even hyperinflation) are intimately bound together. In this context, one might mention Friedman (1959, 7), Boris Pesek and Thomas Saving (1968, 86–87), and Pesek (1968), among numerous others. This belief appears to be based on the premises that (1) consumers are willing to use any banknotes that are issued; that is, they are indiscriminate; (2) there are no mechanisms within free banking that would constrain the overissue of notes; (3) the relevant marginal costs of a note-issuing bank are its marginal production costs (which are admittedly small, though positive, for commodity-backed paper banknotes); (4) profit-maximizing free banks would supply a quantity of notes at which marginal production cost equals marginal revenue; (5) marginal revenue in this context is measured in terms of the purchasing power of money; and (6) therefore, free banking would produce a money supply such that the purchasing power per monetary unit was quite small; that is, there would be substantial inflation.

This sounds convincing, but it is fallacious nonetheless. First of all, to propose that consumers would voluntarily use any currency that had depreciated in value (purchasing power) is to suppose that consumers would not act in their own self-interest.⁹ One of the fundamental tenets of economics is that individuals seek to maximize their total utility. There is no reason to assume that money is the sole exception to that principle. The only way in which an inflated currency can remain in circulation is to have its use mandated by law—to pass legal tender laws. Inflation is far more likely under central banking than under free banking for that reason.

Second, there are strong forces within a free-banking system that operate in an anti-inflationary manner. These are the processes of reflux and adverse clearings (see Chapter 2). Both serve to curtail any attempt by a bank to expand its liabilities in excess of the demand for same. Furthermore, as also discussed in Chapter 2, the key cost to a free bank is not the cost of producing banknotes, but the cost of maintaining them in circulation. These marginal costs are clearly positive and rising. Thus, a free bank will indeed supply a quantity such that marginal cost equals marginal revenue (purchasing power), but that does not imply inflation.

Benjamin Klein offers a compelling reason why free banks will avoid an excessive production and circulation of money: “[I]f information about future performance is costly, information is a valuable product. The brand name of a firm is then not only an identification mark but also a capital asset” (1974, 432). Furthermore, the value of such a capital asset possessed by a free bank would “be related to the anticipated predictability of the future price level in terms of the money” (emphasis in original) (Klein 1974, 433). For a free bank to inflate would be to destroy one of its important assets. What of its other assets? Donald Wells and Leslie Scruggs suggest that “another major factor that would prevent a bank of issue from inflating its currency would be the fact that the great majority of its assets would be denominated in its own banknotes; thus the bank of issue could be a leading victim of its own overissuance” (1983, 84).

In theory, then, one would expect free-banking regimes to exhibit no inflationary secular trends of any consequence. What of the historical evidence? If one examines the two best-known cases of multiple note issues—Scotland and the United States—one finds some ambiguity in the case of Scotland and a clear answer for the United States. With Scotland, there is both anecdotal and circumstantial evidence of some inflation (see Chapter 5). However, the lack of data series for Scotland separate from Great Britain as a whole makes any conclusion very hazardous. As for the United States, the data exist and lead one to conclude that very little inflation occurred. Specifically, from 1850 to 1860 (the heyday of free banking), wholesale prices rose by 10.74 percent, an average of only 1.025 percent per year (see Table 17). Consumer prices rose by 7.99 percent, an average of a mere 0.772 percent per year (see Table 19). By comparison, one finds from Tables 18 and 20 that from 1959 to 1989 wholesale prices rose 252.1 percent (4.29 percent per year), and consumer prices rose by 325.3 percent (4.94 percent per year). Kevin Dowd gives a succinct summary of the historical evidence when he declares that the “claim that competition among unregulated banks would lead to an explosive money supply and rapid inflation thus has no support in the historical record, and indeed, inverts the truth that rapid inflations have always been associated with government interventions to suppress competition” (1992, 3).

FREE-BANKING INEFFICIENCY

Milton Friedman is one of the best-known proponents of yet another criticism of free banking, at least of that based on a specie standard. He asserts that such a system is horribly inefficient, because it requires the use of significant amounts of real resources in order to supply the metallic outside money (1959, 4–6; 1962, 220–21). “The use of so large a volume of resources for this purpose establishes a strong social incentive in a growing economy to find cheaper ways to provide a medium of exchange” (Friedman 1959, 5). He goes on to estimate the annual cost of a gold standard as 2.5 percent of the American national product (1959, 5).

Lawrence White responds to this criticism in two ways. First of all, he points out that Friedman makes the unjustified assumption of a “pure commodity standard,” that is, a system of banknotes backed 100 percent by gold coin (1984a, 148). That is, Friedman assumes that free banking will follow the Rothbard-Mises (RM) model (see Chapter 7), but none of the approximate free-banking experiments, for example, those in Scotland, the United States, France, Canada, and Sweden, were systems of 100 percent reserves. To make the assumption Friedman makes is to bias the case against free banking. White recalculates the cost of a laissez-faire, specie-backed system using 1982 data on M1 and GNP and reserve ratios comparable to those actually observed in Scotland (about 2 percent). He arrives at “an estimate of annual resource costs of between 0.01 and 0.03 percent of gross national product” (1984a, 148). White further opines that “this figure is insignificant in comparison with plausible estimates of the GNP losses due to monetary instability” (1984a, 149fn). Even if, instead, one uses the much higher reserve ratios for American free banks (a mean of 14.9 percent per Table 5), one still gets an estimate of roughly 0.04 to 0.1 percent of GNP per annum. Either way, Friedman’s figures are badly overstated.

In addition, White is sensitive to an insight that seems to have escaped Friedman’s notice altogether:

The forced substitution of fiat for convertible currency, like the Ricardian forced substitution of convertible currency for coin, is by no means efficient when it contravenes consumer preference for what is considered a more trustworthy currency. Lowering production costs does not constitute efficiency when the resulting product is one of lower quality in consumers’ eyes. The actual forced movement in the twentieth century from a gold-convertible dollar to an inconvertible dollar must have represented a “negative social saving.” (1984b, 296)

In other words, utility losses may occur at the microlevel even when certain (apparent) monetary gains are achieved at the macrolevel. Friedman’s error lies in adopting what White has brilliantly termed a “macroinstrumental” approach as opposed to a “microsovereignty” approach (White 1985, 114–15). The former means letting the economic analyst determine what is “desirable”; the latter means letting the consumer decide.

FREE BANKING AND COUNTERFEITING

A very common belief regarding free banking has been that it is conducive to counterfeiting. Most economists would not question the assertion that, for example, during the American free-banking period, “counterfeiting was rampant” (Kemmerer and Jones 1959, 195). Counterfeiting was not unheard of in free-banking systems, but was it a severe and chronic problem? The facts seem to suggest otherwise.

“The Scottish free banking system . . . seems to have provided few opportunities for profit through fraudulent note issue. Counterfeiting was rare” (White 1984a, 140). Henry Meulen agrees that “experience has demonstrated to the Scottish people that forgeries of the paper of its private banks are rarer than counterfeit sovereigns: the issuers of the paper are on the spot, and are interested in preventing forgeries of their own notes; whereas the government acts slowly and ponderously from a distance” (1934, 69).

What about the United States? Is it not true that American free banking witnessed massive quantities of forged notes? Perhaps not. The widespread belief in the magnitude of such counterfeiting stems from a misunderstanding of the various “banknote reporters” published at the time. These were periodicals (usually weekly) that listed each free bank, the discount (if any) then prevailing on the bank’s notes and any known counterfeit notes. A perusal of such banknote reporters conveys the impression of an enormous number of counterfeit notes.

Such an impression is erroneous, however. Hugh Rockoff explains that “historians have not understood that the reporters listed all counterfeits and all banks which had failed, even if the notes had been removed from circulation years before. . . . In other words, the reporters contained cumulative indexes of bank failures and counterfeits, not annual indexes” (emphasis in original) (1975, 23). He mentions, for example, a banknote reporter from 1849 that still listed a bank in Savannah, Georgia that had closed ten years earlier (1975, 25–26). Robert King suggests that the unreliability of such data may make it impossible ever to construct an accurate series on American counterfeit notes (1983, 155). In short, the precise extent of counterfeiting experienced under free banking in the United States may never be known. Nevertheless, one thing appears to be true: the frequency with which counterfeiting occurred has in the past been exaggerated.

As for other episodes of free banking, one should see the detailed surveys in Dowd (1992). Generally, counterfeiting seems to have been a minor problem. In Scotland, banks sometimes even honored forged notes that were innocently offered for redemption or deposit (White 1984a, 40). In Canada, “since all banks had an interest in preventing forgeries, it was only natural that an organization like the Canadian Bankers Association (CBA) would emerge to look after this common interest” (Wells 1989, 16).

Does theory suggest that free banks ought to have a problem with counterfeit notes? No, since the likelihood of successful counterfeiting varies directly with the average period of circulation; that is, notes that return to the issuer very slowly are more likely to be forgeries. Central banking systems are inherently inferior in this regard. If there are both legal tender laws and a monopoly issuer of currency, then notes tend to circulate for long periods before they are returned to the issuer (Selgin 1988a, 149). This encourages attempts at counterfeiting. Free banks should, in contrast, have their notes returned frequently via the mechanisms of reflux and adverse clearings. Indeed, the more intense the competition, the more rapidly free banks would return their rivals’ notes for redemption. As a result, the average period of circulation for free-bank notes should be quite short, and the likelihood of counterfeiting should be small (White 1984a, 140).

CENTRAL BANKING AS A SPONTANEOUS EVOLUTION

Recently the view has been expressed that competitive banking is neither a natural development from market forces nor an efficient approach to the supposedly unique problems of banking. It has been proposed instead that central banking is both superior to free banking and, moreover, the result of an evolutionary process rather than a legal imposition that constrains the market. The most vigorous expositor of this idea is probably Charles Goodhart, a former economic advisor for the Bank of England. Goodhart’s argument boils down to the following propositions: (1) Centralized control of banking is necessary and desirable because of instability resulting from “contagion effects,” the impracticability of private deposit insurance, and the “moral hazard” problem; and (2) central banks came into existence as a natural reaction to the deficiencies of competition (1988, 6–11, 103–4).

Goodhart believes free banking necessarily suffers from panics that arise from “contagion effects.” Yet Arthur Rolnick and Warren Weber found very little evidence of contagion in American free banking between 1841 and 1861 (1986), and one would expect even less in a true laissez-faire system. If the common elements that link commercial banks together in a central banking scheme—centralized reserves, a lender of last resort, a uniform currency, and deposit insurance—were eliminated, then each bank would have to be judged on its own merits. The liquidity or solvency of one bank would have little or no bearing on the soundness of any other bank.¹⁰ “Contagion” may be largely a product of those restrictions on the free market that comprise central banking.

Goodhart further asserts that private deposit insurance is impractical. Eugenie Short and O’Driscoll (1983), Catherine England (1988), and Selgin (1988a, 135–36) flatly disagree. Glasner (1989, 195–200) and Dowd (1991d, 6–8) both argue that deposit insurance (private or governmental) is unnecessary. Glasner observes that private moneys in the form of mutual fund shares are immune to bank runs, because the holders of such moneys have equity claims, not debt claims. A bank run would simply diminish the wealth of those who possessed the mutual fund shares. Dowd makes a compelling case for believing that capital adequacy is a viable (and perhaps superior) alternative to deposit insurance. As Dowd puts it, “runs can be eliminated when equity-holders maintain a sufficient capital buffer to provide depositors with credible reassurance that their deposits are safe” (1991d, 8).¹¹

England insists that unregulated systems with private insurance would exhibit greater stability than our present system and suffer from little informational asymmetry. This conclusion follows from her beliefs that without federal regulation and insurance, the value to depositors of information about specific institutions would increase, leading to both greater demand for and greater supply of such information, and banks would become more heterogeneous, making easier the differentiation between safe and unsafe institutions (England 1988, 785–86).

Goodhart also feels that “the avoidance of moral hazard requires a degree of regulation and interventionist control over the freedom of bankers” (1988, 104). This is indeed a peculiar suggestion; regulation appears to be the chief source of moral hazard problems rather than the means for their elimination. As O’Driscoll has stated: “The FDIC’s pricing of deposit insurance creates a subsidy to risk taking, a subsidy that can only be captured insofar as banks actually make their asset portfolio riskier” (1988, 663). Can the FDIC solve the problem it has created? No. Since regulatory agencies are by nature noncompetitive, non-profit-maximizing entities (a fact Goodhart praises), they are incapable of discovering the appropriate risk premium.’¹² Only a market participant has access to such information because it is information generated by a trial-and-error, market process. Thus, it is that “a rational system of risk-based insurance premiums offered monopolistically by a public agency is simply impossible” (O’Driscoll 1988, 667).

Finally, Goodhart proposes that central banks are a natural evolution, an “obvious solution” to the flaws of unregulated banking. What are the facts? Has the legal creation of central banks been motivated by a desire to increase the public welfare (Goodhart’s position) or by a desire to extract privileges for special-interest groups?

“The Federal Reserve Act was the result of a movement led by bankers . . . hoping to offset the decentralization of banking toward small banks and state banks. The expansion and domination of banking by big city bankers was possible only with the aid of the federal government” (Kolko 1963, 243). Furthermore, “the Federal Reserve was designed to act as a government-sponsored and -enforced cartel promoting the income of banks by preventing free competition from doing its constructive work on behalf of the consumer” (Rothbard 1984a, 135). Roland Vaubel states that central banks have not attained their privileged status as a result of out-competing all other banks: “The Bank of England, for example, was granted its monopoly not because it was gaining ground in the market but because it was losing out to the other joint-stock issuing banks which had emerged after the Bank’s joint-stock monopoly had been abolished in 1826” (1986, 933fn).

Glasner offers that “the state monopoly over money emerged not to increase the efficiency of the monetary system but to help the state protect its sovereignty” (1989, 38). One might add that central banks have consistently been utilized by the state as a source of credit.¹³ Even a cursory perusal of history reveals that nation-states, in order to finance wars or massive domestic projects, have found it “necessary” to reduce (or eliminate altogether) competition in banking. Central banks have been the principal tool of governmental expansion. It is no accident that the twentieth century—the “century of collectivism”—is also the century in which central banking became the norm. Richard Wagner sums these points up neatly when he argues that “the support for central banking seems more likely to be explained by the economic theory of rent-seeking than by the theories of market failure and public goods” (1986, 519).

A LENDER OF LAST RESORT

One of the objections to free banking that is most frequently cited is the assertion that banking is inherently “unstable” or “fragile,” and therefore, there must exist some agent outside the market process itself that will come to the aid of individual banks in times of crisis. The problem is said to stem from the fact that banks—unique among financial institutions—perform two different roles. They are both (1) financial intermediaries and portfolio managers, and (2) suppliers of transaction services (Goodhart 1988, 86–87). This allegedly leads to a fatal flaw. Namely, bank illiquidity can bring about bank insolvency, and the failure of one bank may very well precipitate a run on all banks. In other words, bank runs are inescapable (without a lender of last resort) and potentially disastrous.

A central bank as the ultimate source of liquidity, as the lender of last resort (LOLR), is proposed as the solution to the weaknesses of banking. It would indeed be foolish to claim that commercial banking in the United States, for example, is fundamentally sound, but what is the source of the problem? Are regulation and centralized control the cure—or the disease?

Dowd points the way to an answer when he says that “the experience of relatively free banking . . . appears to lend no support to the claim that systems without a state-sponsored LOLR are more prone to crises . . . the banking system is quite capable of protecting its liquidity provided it is left free to evolve the means to do so” (1989, 39). The necessary means to achieve such protection most fundamentally requires that each bank be free to supply notes to the market as the demand arises. As seen in Chapter 2, a bank that can issue notes is likely to be able to keep a currency run from becoming a redemption run. When private banknotes are forbidden, every increase in the demand for currency relative to deposits drains reserves from the system and brings on large decreases in the money supply.

If freedom of note issue were accompanied by direct—but delayed—convertibility in the form of option clauses, bank panics might be altogether unknown. “Option clauses give note-issuing banks that operate on fractional reserves an effective means of protecting themselves against bank runs” (Dowd 199lh, 761). The reason is disarmingly simple. In the absence of option clauses, a depositor who had the slightest doubt about the bank’s ability to redeem would be motivated to go ahead and demand redemption for fear of not being first at the redemption window. The sudden “panic” behavior on the part of depositors (each of whom is concerned about his or her place in the redemption line) exhausts the liquidity of the bank. This may force the bank to absorb “firesale” losses on assets that are sold in order to accommodate the redemption demands. In short, insolvency may result.¹⁴

On the other hand, the presence of an option clause forestalls the entire process. Its effect is to reward the depositor—via explicit interest payments—for delaying the demand for redemption. Option clauses “convert speculative demands for redemption from the destabilizing force they are under full convertibility to a stabilizing force that protects the banks’ reserves when they are run down” (Dowd 1991h, 764).

The problem, however, goes beyond prohibitions of private banknotes and option clauses. The justification for central banks has dwelt on the “financial weaknesses” of commercial banks. Those weaknesses, however, prove on close inspection to be largely, if not entirely, the result of a maze of regulations and restrictions.¹⁵ For example, in the United States, banks have been prevented from reducing risk by means of the diversification of their assets and/or liabilities. The restrictions on diversification have taken the form of antibranching laws (the National Banking Act of 1864 and the McFadden Act of 1927), as well as the stipulation that commercial banks could not underwrite corporate stock and bond issues (the Glass-Steagall Act of 1933). Furthermore, the flat-rate insurance premia of the FDIC’s deposit insurance have even encouraged banks to undertake ever-riskier portfolio strategies. Indeed, the very existence of an LOLR itself encourages more risk-taking, especially when loans are offered by that LOLR at subsidy rates of interest rather than penalty rates. “In doing so they in effect act as lenders, not of last, but of first resort” (Selgin 1989, 442). Ceilings on deposit rates of interest (the Glass-Steagall Act and the Banking Act of 1935) and limitations on bank mergers (the Bank Merger Act of 1966) further harmed banks by reducing their responsiveness to changes in market conditions. Regulation is and has been the problem, not the solution.

SUMMARY

It has long been thought that there exist compelling reasons why private money production by unregulated banks “simply cannot work.” These include the beliefs that money is a public good, private money will be held in suboptimal quantities, banking is a natural monopoly, laissez-faire banking is unavoidably inflationary, specie-based currencies imply an inefficient use of resources, multiple currencies encourage counterfeiting, and only a central bank can prevent periodic financial crises. To these may be added the more recent argument that central banks are a natural development necessitated by the instability of decentralized, competitive banking. As was seen above, all such criticisms of free banking are, despite their evident popularity, of questionable merit.

NOTES

1. Some economists emphasize the “joint supply” aspect. This is the proposition that if the good is produced at all, it will be produced in a quantity sufficient to benefit all persons. National defense is supposedly the premier example of such production. Others focus on the proposition that, once it is produced, the marginal cost of a public good is zero.

2. This perspective is totally foreign to many economists. That is, many adhere—implicitly if not explicitly—to a methodological holism rather than to methodological individualism.

3. See Chapter 4 for a discussion of this issue. The core of the problem seems to lie with the unthinking assumption that data are the same as knowledge (or information) in economics. Only those who embrace an individualistic method will apprehend the distinction.

4. The relationship between the two arguments consists of the fact that both involve certain supposed external effects.

5. This is true despite the fact that Friedman’s estimate of the potential gains from achieving optimal money balances is considerably higher than Dowd’s (Dowd 1991e, 10).

6. Such payments on demand deposits were ended in the United States only when they were prohibited by the Banking Act of 1933 (perhaps better known as the Glass-Steagall Act).

7. One might see both as suboptimal quantities from a utility perspective, but Selgin’s point is well taken nevertheless.

8. See Glasner (1989, 238) and Selgin (1990, 281) for further discussion.

9. That self-interest can, of course, follow either an ethic of rational selfishness or an altruist ethic. In other words, “selfishness” and “self-interest” are not synonyms. Noneconomists frequently fail to make the distinction. However, in neither context would a depreciated currency seem desirable.

10. This would remain true even if there were a private clearinghouse. Clearinghouses are organizations that exist primarily in order to facilitate the settling of interbank debts. They are likely to arise spontaneously (and did in actual historical cases). Clearinghouses would probably evolve certain functions in addition to interbank clearings. They might serve as credit information bureaus, monitor questionable practices on the part of the member banks, and assist members by supplying short-term credit (Selgin 1988a, 28–29). Despite a superficial similarity, clearinghouses are not central banks. They are, however, capable of performing all the beneficial functions of a central bank.

11. Dowd has long argued that bank reserves are less important to bank safety than is bank capital. This would depend, at least in part, on whether convertibility was direct or indirect.

12. See Chapter 4 for discussion of a similar point.

13. There is a kind of symbiosis at work here. The central bank is granted certain legal privileges—such as establishing its currency as legal tender—in exchange for which it assists the state in achieving certain policy goals.

14. Under free banking, failure would still be unlikely, because banks would have an incentive (in the absence of government deposit insurance) to maintain adequate capital. The topic of capital adequacy is a familiar theme in Dowd’s work and one that deserves more attention. Dowd develops his ideas along these lines most effectively when he critiques the Diamond-Dybvig model of bank instability (1991c, 1991d). The Diamond-Dybvig (DD) model (1983) has proven extremely influential and is often taken to be the definitive proof of the proposition that banking is inherently unstable. Briefly, the DD model states that (1) depositors’ demands for liquidity are unpredictable, (2) a bank’s liabilities (demand deposits) are more liquid (possess a shorter time to maturity) than its assets (its loans outstanding), and (3) therefore, bank runs will occasionally occur unless there exists government deposit insurance and/or a lender of last resort in the form of a central bank. Dowd’s very effective rebuttal makes four key points. First, the DD model assumes that all investors exhibit the same attitude toward risk, which ignores the fact that some investors prefer debt and some prefer equity. Second, the “banks” in the DD approach do not resemble real-world banks at all; they are a kind of mutual fund. Third, DD assume that government has access to both certain technology and resources that are unavailable to the private sector. Finally, DD overlook the fact that banks (even with fractional reserves) can forestall runs and panics by holding an adequate capital buffer.

15. See Selgin (1989) for an excellent review of such arguments.