Chapter 7 OTHER APPROACHES

All influential economic theories surely pass through four stages. Initially, the new idea is treated with skepticism, perhaps even with scorn. It strikes many theorists as being outlandish or irrelevant. A few, however, think they perceive something of value. They are challenged and excited by the prospect of exploring new territory. Thus the second stage is born. A variety of researchers begin to investigate, interpret, and modify the idea. It becomes, in short, a viable research topic. Such research leads to either the conclusion that the idea is irremediably defective and must be discarded, or the refinement of the idea and its eventual acceptance into the received canon of the discipline.¹ Finally, a theory that has survived and gained acceptance will find expression in some institutional structure or policy.

The theory of free banking has advanced well into stage two. A growing number of economists are energetically exploring its characteristics. Moreover, a variety of approaches to (or models of) free banking have emerged. This is healthy and natural, but it does breed controversy. The purpose of this chapter is to review several of the alternative approaches to free banking.

BASIC DIFFERENCES

There have been developed, as the author sees it, three basic models of free banking. Within each general approach, variations can be found. Therefore, the reader must not infer that every individual named below necessarily advocates precisely those attributes that are listed for each respective basic model.²

The issues regarding which the differentiation has occurred seem to be (1) direct convertibility versus either indirect convertibility or inconvertibility, (2) money as a debt instrument versus money as an equity instrument, (3) whether or not the medium of exchange (MOE) should also be the medium of account (MOA) or the medium of redemption (MOR),³ (4) a stable price-level rule versus a productivity norm, (5) whether or not the holding of reserves is relevant to the performance of free banks, and (6) if reserves are important, whether or not the reserve ratio should be 100 percent.

The White-Selgin Model

As has been made abundantly clear, this model is the focus of the present book. It has been formalized, extended, and slightly modified by the author, but it is essentially the White-Selgin (WS) model whose virtues have been extolled throughout this work. The WS model concerns a commodity-backed currency that is directly convertible into that commodity. The media of exchange, for example, private banknotes, are denominated in the unit of account, for example, dollars. The medium of account (gold) is also the medium of redemption. Inside moneys—notes or deposit credits issued by private banks—are liabilities of the issuing banks. The holding of a reserve of outside money is essential for meeting redemption demands by customers as well as for settling adverse clearing balances with other banks. However, it is argued that both withdrawals and deposits are randomly distributed in a mature free-banking system. Therefore, fractional reserves are adequate to meet all plausible conditions. This would be reinforced by the likely presence of option clauses. Finally, the price level would be stable, except when there occur pervasive changes in per-capita productivity. The price level would fall when such productivity rose, and rise when such productivity fell.

The advantages of the WS model over the alternatives are said to be quite significant. First of all, it is the one model of free banking that is most like banking as it exists now. It would strike most consumers as the least alien of the free-banking proposals.⁴ Related to that point is the fact that all the historical experiments with free banking looked more or less like the WS model. Inside money redeemable in specie was issued by private banks, which held fractional reserves. The other major advantage is that it adheres to the productivity norm rather than a stable price-level rule. As such, it is less likely to suffer business cycles.⁵

The Rothbard-Mises Model

For presentations of the Rothbard-Mises model (RM), one should see Murray Rothbard (1983, 87–124; 1984b; 1985; 1992) or Ludwig von Mises (1966, 441–48; 1971, 395–99). Other economists who adopt the RM model include Joseph Salerno, Hans-Hermann Hoppe, Mark Skousen, and Gary North (Rothbard 1992, 99).

This approach is very similar to the WS model in several respects. It proposes that banks issue inside money directly convertible into gold coin on demand (but without delay; i.e., option clauses would be illegal). Banknotes would be bank liabilities, the MOE, MOR, and MOA would be the same, and the price level would be inversely related to aggregate production. The most dramatic difference between the RM and WS models is that in the former, banks would be required by law to hold 100 percent reserves at all times. This would, of course, force banks to operate like modern consumer finance companies rather than like conventional commercial banks. Loans could be made only out of the bank’s capital instead of out of deposits. Both the supply of credit and the profits of banks would clearly be sharply reduced.

The justification given for this constraint on banking activity is the assertion that fractional reserves are inherently fraudulent.

It should be clear that modem fractional reserve banking is a shell game, a Ponzi scheme, a fraud in which fake warehouse receipts are issued and circulate as equivalent to the cash supposedly represented by the receipts . . . fractional reserve banking is at one and the same time fraudulent and inflationary; it generates an increase in the money supply by issuing fake warehouse receipts for money. . . . Commercial banks—that is, fractional reserve banks—create money out of thin air. Essentially they do it in the same way as counterfeiters. (Rothbard 1983, 97–98)

Rothbard adopts this position because he insists that banknotes and deposit credits are (or should be) warehouse receipts, both legally and economically. That is, he perceives the proper role of banks to be that of a warehouse rather than a financial intermediary. The act of depositing funds in a commercial bank and receiving in return either banknotes or a deposit credit is not, to Rothbard, a credit transaction. The depositor is not loaning his or her funds to the banker, but merely hiring the banker’s vault as a place of safekeeping for valuable assets. Rothbard declares that depositing money in a bank should be perfectly analogous to storing one’s furniture in a warehouse (1983, 88–89). In other words, the banker should be dealt with as a bailee rather than as a debtor (Rothbard 1992, 98fn). It is crucial to realize at this point that in the RM model, a change in inside money demand does not reflect a change in consumer time preferences (a change in the rate of saving), and it will not bring about a change in the supply of credit. This stands in sharp contrast to the WS model as it was presented in Chapter 3.

The bank run plays a special role according to Rothbard. Unlike the overwhelming majority of economists, he views bank runs favorably. They function as a constraint on inflationary monetary expansion under free banking (1983, 112). They also “instruct the public in the essential fraudulence of fractional reserve banking” (1983, 113). Rothbard seems literally to believe that if one observes a period during which there occur very few bank failures, then this is sure to be a time of inflationary monetary expansion. A case in point is that of Scottish free banking (1765–1845). In discussing Lawrence White’s work on Scotland, Rothbard contends that a low rate of bank failure “might indeed mean that the banks are doing better, but at the expense of society and the economy faring worse. Bank failures are a healthy weapon by which the market keeps bank credit inflation in check . . . a lower rate of bank failure can scarcely be accepted as any sort of evidence for the superiority of a banking system” (emphasis in original) (1988b, 230).

Those who embrace the RM model⁶ seem always possessed of an intense, almost obsessive concern with monetary expansion. The author is not here implying that inflationary expansions should be of no concern to monetary theorists. Far from it. Such phenomena have very grave, often disastrous, consequences. It is just that those who advocate the RM model seem to “throw the baby out with the bath water” in their quest for limitations on such expansions. They also, as a result, give rather too little attention to unjustified contractions.⁷

The sources of this focus on expansionary bank policies include the proposition that precious metal coins (or bullion) are the only “true” money, a concomitant suspicion regarding banknotes, and an unusual view of what constitutes the optimal supply of money.⁸ The core proposition of the RM model is that money, correctly understood, is a commodity such as gold or silver and that the names of the various monetary units (dollar, pound, peso, etc.) “invariably originated as names for units of weight of a money commodity” (Rothbard, 1984b, 9). According to such a perspective, the “dollar” is—or should be—analogous to common units of measure, such as the liter, the meter, and the kilogram. That is, the dollar should be defined—once and for all—as a specific fraction of an ounce of gold. Moreover, Rothbard argues that that was once the standard textbook approach (1984b, 10). Reasoning on the basis of that perspective, Rothbard defines inflation as any increase in inside money that is not matched by an equal increase in specie (1984b, 16).

If only specie is “money,” banknotes are not money, but merely money substitutes. How do they come to be used as a medium of exchange, then? Mises explains this via his “regression theorem” (1971, 97–123). This analysis concludes that all present moneys must have originated as some highly marketable commodity that initially had a significant nonmonetary use. “Money cannot originate as a new fiat name, either by government edict or by some form of social compact” (Rothbard 1984b, 10). However, this does not mean “that fiat money, once established on the ruins of gold, cannot then continue indefinitely on its own” (Rothbard 1984b, 11). Banknotes, whether convertible or inconvertible, whether produced privately or by the state, are all derivations from the one true money—gold.⁹

Mises suggests that private banknotes not only tend to be dangerously inflationary, but also may not be necessary. He favorably quotes Thomas Tooke’s claim that “free trade in banking is free trade in swindling” and comments on that “state of affairs under which everybody is free to issue banknotes and to cheat the public ad libitum” (1966, 446). His curious position is that banks should be allowed to issue notes in order that consumers’ use of banknotes will diminish, if not disappear altogether. Mises does not deny the obvious convenience offered by paper banknotes vis-à-vis precious metal coins. Nevertheless, he declares that “banknotes are not indispensable. All the economic achievements of capitalism would have been accomplished if they had never existed” (1966, 447).

The third reason for the RM focus upon monetary expansion is an idiosyncratic definition of the “optimal supply of money.” Some have argued that the optimal supply occurs when the rate of deflation is such that the nominal rate of interest is zero (Friedman 1969). Others might suggest that any supply is optimal that equals the demand for money at the existing price level.

Rothbard, on the other hand, declares that the optimal supply is that minimum necessary to establish the money commodity as the accepted medium of exchange (1988a, 180). Any increase in the money stock in excess of that minimum will exhibit zero “social” marginal utility. In fact, such an increase will do only harm. “There is never any social benefit to increasing the quantity of money. . . . Monetary calculations and contracts are distorted, and the early recipients of the new money, as well as debtors, gain income and wealth at the expense of later recipients and of creditors” (Rothbard 1988a, 180). Furthermore, it is unnecessary to increase the money supply to match an increase in money demand brought about by population growth. All that will occur then is that prices will fall so as to raise each dollar’s purchasing power (Rothbard 1983, 47). In other words, Rothbard assumes that, although the nominal money stock is constant, real money balances will rise because of perfectly flexible prices (1983, 34–41).¹⁰

Two final features of the RM model are the absence of option clauses and the implication that the price level will be inversely related to the level of production. Considering the model’s view of outside money as specie and inside money as redeemable in specie (for which redemption purpose, 100 percent reserves are maintained), it may not be surprising that option clauses become superfluous. In the RM model, they are neither permissible nor useful. The defining characteristic of (inside) money is “whether a certain claim is withdrawable instantly on demand” (emphasis in original) (Rothbard 1983, 255), and there will be “no provision for emergency suspensions of redeemability” (Rothbard 1983, 263).

Rothbard states that “changes in prices in general . . . are determined by changes in the supply of and demand for money . . . an increased supply of goods will, other things being equal, increase the demand for money and therefore tend to lower prices” (emphasis in original) (1975, 15). That is, Rothbard’s concern seems to be with aggregate production. One should note that this differs from the WS concern with changes in per-capita productivity that lower goods’ unit production costs.¹¹

Problems with the RM Model

In general, one may say that the RM model is, as the saying goes, an “ingenious solution to a nonexistent problem.” If free banking on a specie standard but with fractional reserves (WS) were chaotic and inflationary, then the RM approach might deserve serious consideration. However, Chapters 2, 3, and 4 have argued—one would hope persuasively—that the WS model is noninflationary, responsive to consumer preferences, consistent with microeconomic principles, and resistant to business cycles. The RM model is clearly inferior.

The best that can be said for the RM model is that in the long run, it is noninflationary. It does not, for example, allow for any ready response to consumer demand. If the income velocity of money falls (rises), the money supply will not increase (decrease) to match the increase (decrease) in money demand. In the absence of perfectly flexible prices in all markets, departures from the natural rate of unemployment would ensue. Furthermore, the market rate of interest might depart from the natural rate. Assume the velocity of money falls (the Cambridge k rises). That is, assume consumer time preferences shift toward greater saving (higher money balances). The increase in saving drives the natural rate of interest lower.¹² However, the money supply is fixed in the short run. No new loans or deposits will be created. The market rate of interest will exceed the natural rate.¹³ A recession or depression will be the result; notional demand will exceed effective demand; Say’s Law will be violated.

It would appear that the RM model tries to sever that link between the market for money and the market for time, which is crucial to the success of the WS model, as was examined in Chapter 3. This follows from Rothbard’s view of money balance decisions as distinct from decisions regarding saving versus consumption.

[A consumer] allocates between the various categories on the basis of two embracing utilities: his time preferences decide his allocation between consumption and investment (between spending on present vs. future consumption); his utility of money decides how much he will keep in his cash balance. In order to invest resources in the future, he must restrict his consumption and save funds . . . saving and investment are always equivalent. . . . The demand for money is completely unrelated to the time-preference proportions people might adopt, (emphasis in original) (Rothbard 1975, 40)

For Rothbard, there are three separate allocative channels for one’s income: consumption, cash balances, and savings-investment (these two are always equal). As suggested above, this reasoning leads to a model in which discoordination can arise between the money and credit markets, and it raises a troublesome question. How can one deny that to refrain from consumption expenditures in order to increase one’s money balances is to engage in voluntary saving? Yet this is what Rothbard apparently does deny.

Paradoxically, the RM model does not honor consumer preferences. Choices as to reserve ratios, redemption contracts, and the nature of the base money surely should be left to the unconstrained¹⁴ interaction between banks and their customers if one claims to be proposing a truly free market approach to banking.¹⁵The RM model declares that to have a nonfraudulent banking system, one must have a specie base, 100 percent reserves, and immediate redemption upon demand. It is furthermore important to realize that much of Rothbard’s argument on behalf of such a system is legalistic rather than economic in nature.

The British Court decisions cited and criticized by Rothbard, to the effect that bank notes do not contractually bind their issuers to holding 100 percent reserves, seem eminently reasonable given the inscription actually found on the face of a typical British bank note. . . . There is no promise made about reserve-holding behavior. There is nothing to indicate that the note constitutes a warehouse receipt or establishes a bailment contract. . . . Nothing in a free banking system prevents an individual who desires 100 percent reserve banking from explicitly contracting for it. . . . Fractional reserves do not constitute breach of contract. (White 1985, 120–21)

Rothbard might be well advised in this case to heed his own statement to the effect that “[free market action] is optimal, not from the standpoint of the personal ethical views of an economist, but from the standpoint of the free, voluntary actions of all participants and in satisfying the freely expressed needs of the consumers” (1970, 887).

Finally, the RM model may imply a waste of resources as well as being unnecessarily restrictive insofar as economic growth is concerned. Many have argued that 100 percent specie reserves require an inordinate use of real resources. Of course, if the RM model led to monetary equilibrium and the avoidance of business cycles, the benefits might justify the expenditure, but it does not. The supply of money is entirely too inflexible to maintain monetary equilibrium. Indeed, the RM structure reminds one somewhat of the National Banking System in which notes were issued in strict proportion to bank holdings of government bonds. That rigidity made it impossible for banks to respond appropriately to seasonal fluctuations in the demand for currency. The result was a series of financial crises (Sprague 1910).

Also, what of economic growth? If the RM model were implemented, then the long-run growth in money and credit would mirror changes in the stock of monetary gold. Averaged over a considerable span of time, the rate of change in that gold stock might approximate the rate of change in the demand for money, but that will not suffice. First of all, monetary equilibrium cannot be maintained or business cycles avoided, if the money supply is incapable of rising or falling in response to short-run conditions. RM theorists usually keep in mind the need for macroconclusions to rest on a bedrock of microprinciples. What they tend to forget is that long-run results are merely a summation of short-run actions.

Second, economic growth will not be sustained without parallel growth in both money and credit (in the aggregate if not per capita). Yet this basic insight is rejected by the RM model. “The notion of ‘normal’ credit expansion is absurd. Issuance of additional fiduciary media, no matter what its quantity may be, always sets in motion those changes in the price structure the description of which is the task of the theory of the trade cycle” (Mises 1966, 442fn). The RM model does not aim for monetary equilibrium or achieve it.

The Black-Fama-Hall Model

As W. William Woolsey and Leland B. Yeager (1991, 2) explain, the Black-Fama-Hall (BFH) system is so named to “acknowledge ideas borrowed, altered, and recombined from writings of Fischer Black, Eugene Fama, and Robert Hall.” Leland Yeager has been one of the principal exponents of this approach, along with his long-time collaborator Robert Greenfield and, more recently, W. William Woolsey (Greenfield and Yeager 1983; Yeager 1985; Yeager 1986; Yeager and Greenfield 1989; Woolsey and Yeager 1991). Two other well-known proponents of a laissez-faire approach to banking—David Glasner (1989) and Kevin Dowd (1989)—embrace significant aspects of the model. One might also point out that Friedrich Hayek’s seminal work on free banking (1978) anticipates certain features of the BFH approach.

In the BFH model (because of his extensive work in this area, one is tempted to call it the Yeager model, but here the conventional terminology will be retained), distinctive inside moneys are issued by private banks. These are either indirectly convertible into some outside asset(s) or are inconvertible. The holding of reserves is largely irrelevant to bank success (Woolsey and Yeager 1991, 8). The MOE is not the same as the MOA. The MOR may be different from both the MOA and the MOE. These inside moneys might be either equity instruments in the form of checkable mutual fund shares, or debt instruments in the form of notes, coins, or demand deposits. An early misunderstanding arose in this regard as a result of the statement that “the BFH system would indeed lack money as we now know it” (Greenfield and Yeager 1983, 303). What would be lacking would be base money as we now know it. That is, the government “would be forbidden to issue money” (Yeager 1985, 104). It is, however, fair to say that early work on the BFH system tended to concentrate on the model as one of an advanced and sophisticated payments system in which most, if not all, transactions were accomplished via checks or electronic transfers.

More recently, BFH theorists have granted the importance of hand-to-hand currency. Yeager allows that “some institutions would presumably issue notes and even coins denominated in the BFH unit” (1985, 104). Indeed, under a BFH structure, “apart from crucial differences in the unit and in the media for redeeming their obligations, financial institutions would be practicing something similar to free banking under a gold standard” (Yeager 1985, 105).

The differences between the BFH and WS models ought not to be minimized. The differences Yeager refers to are certainly crucial. Paramount among them is the fact that in the BFH approach, the “unit of account would no longer coincide with the unit of the medium of exchange. . . . The government would define the new unit . . . in terms of a bundle of commodities so comprehensive as to have a nearly stable value against goods and services in general” (Yeager 1985, 104). This “value unit,” in which all prices would be stated, would consist of the market values of specific quantities of specific widely traded commodities. For example, 1 unit might equal the market values of 5 ounces of commodity A plus 9 square feet of commodity B plus 6 gallons of commodity C. This would remain the definition of the value unit forever.¹⁶

Arbitrage would alter the relative prices of those individual commodities that comprise the bundle, so that the bundle’s value would always equal 1 unit. Media of exchange would take whatever form was found to be mutually acceptable to both banks and their customers. However, these exchange media would all be denominated in the one common unit of account. Redemption could be made by using any asset, the quantity of which possessed a market value equal to the required number of standard units (or commodity bundles). It would be unlikely that consumers would actually seek redemption in terms of the components of the standard bundle, since most of those would be substances of no use to the typical consumer. More likely would be indirect convertibility in terms of some familiar asset, such as gold.

Kevin Dowd and David Glasner adopt somewhat different positions regarding the issue of a stable unit of account. Although not objecting to the idea of a commodity bundle, Dowd suggests that the definition of the MOA be revised periodically. This he justifies on the grounds that “a commodity index that seemed ideal ex ante might turn out to be less satisfactory than it was anticipated to be. Furthermore, there is always the possibility that certain goods might cease to exist at all” (1989, 99). Glasner, on the other hand, is concerned that a stable price level may lead to sharp fluctuations in employment. Therefore, he opts for stabilizing an index of wages. “A dollar would always be convertible into the amount of purchasing power that would buy a stipulated amount of labor power . . . a labor standard would imply a gently falling output-price level and very low nominal interest rates” (Glasner 1989, 240). It is obvious that Glasner’s position bears some strong similarities to George Selgin’s espousal of a “productivity norm,” “neutral money” policy.

The advantages claimed for the BFH system are truly impressive (Greenfield and Yeager 1983). It would provide a stable unit in which prices would be quoted. This would foster more reliable economic calculation, reduce uncertainty in credit markets, and encourage long-term contracts. Since inflation would be a thing of the past, implicit taxation by means of monetary expansion would be avoided. Fiscal restraint would be increased. Competition in financial services would both stimulate innovations and curb the waste of resources brought on by attempts to escape regulatory restrictions. Bank panics would be unknown. Since there would exist no outside money that was held as a reserve asset by banks, then there would also exist no redemption runs. An increased demand for currency by consumers would not bring about a multiplicative decrease in the total money supply. Banks would be “run-proof.” Also, “since media of exchange would bear interest or dividends at competitive rates,” the optimal quantity of money—as defined by Milton Friedman (1969)—would be held (Yeager 1985, 105).¹⁷ Finally, “monetary disequilibrium as we have known it could no longer occur . . . painful macroeconomic disorders would be practically forestalled” (Yeager 1985, 105). The supply of money would automatically adjust to the demand for money. Effective demand would tend always to equal notional demand.

Before discussing problems with the BFH, it might be helpful to digress briefly and mention those premises that inspire the model. First of all, BFH advocates think it “absurd” for the value of the unit of account (e.g., the purchasing power of the dollar) to depend on the supply of and demand for the medium of exchange. They seek a stable unit of account. Second, they insist that monetary equilibrium has proven so elusive because money has no market, and no single price, of its own.¹⁸ Money trades against all other goods and services. As a result, a change in either money supply or money demand may not be reflected in the value of the unit of account. Prices are pervasively “sticky,” and therefore, monetary disequilibrium may persist for long periods because its elimination requires changes in a multitude of specific prices. Finally, the persistence of monetary disequilibrium brings about major macroeconomic disruptions that inflict substantial suffering on the populace. The only viable solution seems (to the proponents of BFH) to be a system that stabilizes purchasing power by separating the MOA from the MOE. This stability is in part dependent upon the fact that the components of the MOA are distinctly nonmonetary in nature, whereas the MOE may take the traditional forms of checks, banknotes, and coins.

Problems with the BFH Model

Despite the care and elegance with which the BFH model has been developed, there remain serious questions about its desirability. Some of these have been addressed—with at least some success—by defenders of BFH. Others have not been adequately dealt with at all.

One of the early criticisms of the BFH system was that it must suffer from the well-known inefficiencies of barter because of (1) the abolition of base money and (2) the separation of the unit of account from the medium of exchange (O’Driscoll 1986; White 1984c). Allegedly, however, the BFH is not really barter. There is, indeed, a separation of the sort described. However, this does not imply that “people would be making and receiving payments, awkwardly, in miscellaneous commodities and securities with fluctuating values that would have to be transferred into numbers of Units on each occasion. Instead, people would be using coins, banknotes, and checking accounts furnished by banks and denominated in Units” (Yeager and Greenfield 1989, 414).

Closely related to the above is the suggestion that the BFH model must be defective, because it violates the evolutionary process by which money came into being in the first place (Menger 1892). The historical origin of money as the single most marketable commodity in society has, to many, implied the corollary proposition that there will be no separation of the unit of account from the medium of exchange.

[A] unit of account emerges wedded to a general medium of exchange. Prices are universally posted in the characteristic units of a medium or set of media that sellers are routinely prepared to accept in exchange. This process is self-reinforcing: a buyer or seller who communicated bid or ask offers in nonstandard units would impose calculation costs on potential trading partners. For this reason the unit of account remains wedded to the medium of exchange. (White 1984c, 711)

Greenfield and Yeager point out that all existing monetary systems are the products of legislation and regulation to a large extent (1983, 303). Some major nonevolutionary, legislative change will be required if a laissez-faire system is to come into existence. Therefore, they see nothing contradictory about constructing an “ideal” system—even though nothing like it has ever actually existed.

The defenders of the BFH model are obviously correct when they bring one’s attention to the constructivist nature of present monetary systems, and there is no a priori reason to deride any idea merely because it is unusual. However, the critics of BFH have a point. It has become a commonplace to observe that the study of monetary economics is (or should be) eminently practical. As ingenious as the BFH system may be, is it consistent with consumers’ revealed (or demonstrated)¹⁹ preference? Furthermore, considering the widespread confusion about and criticisms of BFH on the part of professional economists, are the advocates of such a system likely ever to convince either lawmakers or voters/consumers of its supposed virtues? One must answer both questions in the negative.²⁰

Bennett McCallum (1985) offers two criticisms of the BFH model. He first suggests that it is really only a version of the old idea of a composite-commodity standard. The composite-commodity standard retains a base money in units of which inside money is denominated. Price stability is supposedly attained by anchoring the base money to a basket of commodities via redemption in terms of those commodities. However, the BFH system claims to do away with base money altogether. Moreover, the BFH commodity bundle would only “define the unit of account, and define it independently of any particular medium of exchange” (Yeager and Greenfield 1989, 417).

Some still question whether the redemption unit might not itself become base money²¹ (Meltzer 1989, 427). One might also wonder whether the system will, as Yeager and Greenfield assume, converge on a single unit of account (Meltzer 1989, 426–27). There would be no legal barrier to competitive units of account, but convergence is assumed in BFH because “government would exert a nudge against the inertia of old practices by conducting its own transactions and accounting in the new unit” (Yeager 1985, 104). Is there not a danger inherent in letting government define the unit of account, even if it is legally applicable only to the government’s own records? Furthermore, how is this necessarily consistent with consumer preferences? How can the advocates of BFH be so sure that either consumers or bankers want an MOA that is neither the MOE nor the MOR?

McCallum also asserts that the price level would be indeterminate in a BFH system. Yeager and Greenfield reply that it “would provide determinacy by defining the unit of account with a commodity bundle. . . . This definition would be made operational by indirect convertibility” (1989, 417). The issue of indirect convertibility has proven controversial. Several writers have claimed that it “would cause severe monetary instability” (Woolsey and Yeager 1991, 1). The “paradox” of indirect convertibility asserts that extreme inflations and deflations will occur as a result of redemption in a medium different from the medium of account.

Assume the commodity bundle (MOA) is initially worth one noninflated dollar and the redemption medium (MOR) is gold. Imagine that an inflating bank’s notes have fallen in market value to $.90 per dollar of face value. An arbitrageur then buys a depreciated dollar, using $.90 issued by a noninflationary bank. He turns around and presents the $1.00 to the inflating bank and demands redemption. If the bank redeems the note with $1.00 worth of gold, then the arbitrageur may buy $1.11 face value of depreciated notes. These he may redeem for $1.11 worth of gold and buy $1.23 face value of depreciated notes, and so forth. The system quickly collapses.²²

There is, however, a flaw in such a scenario, which Woolsey and Yeager quite persuasively explain. “The paradox occurs, we emphasize . . . only if the bank calculates the amount of gold required in redemptions not at a genuine market price but at the very ‘price’ of gold implied by its own redemptions. Hence, the paradox is easily avoided” (1991, 11). One may—and should—have other reservations about the BFH model, but indirect convertibility does not seem to lead to the severe instability that has been hypothesized.

Last, but not least, one may point out that the basic objective of the BFH model—a money that possesses stable purchasing power—has been shown to be a suboptimal goal (see Chapter 4). The details of the argument will not be repeated here. However, the principal proposition was that a stable price-level policy would—in the face of pervasive productivity shocks—lead to departures of effective demand from notional demand, that is, to business cycles.

SUMMARY

The three basic models of free banking—WS, RM, and BFH—have been reviewed and compared. Their similarities and differences were noted in terms of (1) convertibility, (2) reserve holdings, (3) whether or not the MOA, MOE, and MOR were the same, (4) media of exchange as debt versus equity, and (5) a stable price level versus a productivity norm. Although all three are interesting attempts to apply laissez-faire principles to money and banking, only one (WS) seems superior on all counts. The WS model avoids both the confusing complications of the BFH approach and the reserve-holding rigidity of the RM. It is both simple and efficient. It also remains consistent with Say’s Law under all circumstances. Moreover, the WS model is the least alien form of free banking that has yet been proposed. If free banking is ever again to be a reality, the overwhelming probability is that it will be instituted along the lines of the White-Selgin model.

NOTES

1. This process does not necessarily proceed along perfectly rational lines, nor is it always unaffected by the political and ethical beliefs of the researchers involved. A case in point might be the “Keynesian revolution” of the late 1930s.

2. Some may even disagree with the segmentation into three models.

3. “The MOE refers to the debt instruments which are transferred in the exchange process, the MOA refers to the commodities in terms of units of which prices are quoted” (Dowd 1989, 84). The unit of account means the unit used in stating prices. The MOR is that asset used by banks to redeem inside money. For example, in the United States, under the classical gold standard, the gold-backed paper dollar was the MOE, gold was both the MOA and the MOR, and the dollar was the unit of account. The MOA, MOR, and MOE were really indivisible or coextensive.

4. Practically speaking, this could prove to be overwhelmingly important. From a political standpoint, free banking (in any of its guises) is going to be a difficult idea to “sell” to the public. Furthermore, to dwell on the fact that the hostility toward free markets in banking is largely due to the economic ignorance of the populace and/or propagandistic pronouncements by various public agencies is not very helpful. The “strangeness” of free banking needs to be minimized if free banking is to become a reality.

5. See Chapters 3 and 4.

6. All who embrace the RM model are Austrian economists, but not all Austrians adopt the RM approach.

7. It is in regard to this that the monetarists have presented a somewhat more robust exposition (Horwitz 1990, 17–20).

8. Of course, an important additional source is the Mises-Hayek (or Austrian) trade cycle theory, which identifies monetary and credit expansions as the root cause of most, if not all cycles.

9. Actually, Rothbard has proposed a “parallel standard” in which both gold and silver would be used and there would be freely fluctuating exchange rates between the two forms of base money (1984b, 8). This should be distinguished from historical bimetallism, in which the gold/silver exchange rate was set by law. Under bimetallism if, for example, the demand for gold rose relative to the demand for silver, then the market price of gold in terms of silver would exceed the official price. Gold would be officially undervalued, and silver officially overvalued. Gold coins would disappear from circulation, but silver coins would remain in circulation. This is the phenomenon with which Gresham’s Law deals. It is an application of the theory of price controls to the realm of money.

10. Rothbard seems never to make explicit his assumption of price flexibility, but more or less perfectly flexible prices are clearly what he has in mind.

11. See Chapters 2 and 3.

12. Of course, at this point, Rothbard would reject the scenario, since he does not take an increase in money demand to represent increased saving.

13. This is literally the defining characteristic of a recession that is used by Austrians, such as Rothbard (Mises 1971, 346–66; Rothbard 1975, 12–25).

14. Here “unconstrained” is taken to signify a market context in which only acts of force or fraud are outlawed.

15. Please note that the WS model does not impose gold-backed inside money with fractional reserves on the society. It merely argues that both history and logic suggest that such are likely to be the characteristics of future free banking.

16. Both Glasner and Dowd suggest departures from this. See later discussion.

17. This seems to presuppose that all inside money would be equity-based.

18. This writer would suggest that the monetary disequilibria of recent decades is the product of the institutional structure (central banking) rather than the inescapable result of something inherent in money itself. That is, price “stickiness” (as well as price “rigidity”) may be largely the product of decades of discretionary monetary policy. Thinking in terms of price indexes instead of specific prices also may help foster the belief that the market for money is “just a figure of speech, not a reality” (Yeager 1985, 104).

19. The distinction is made by Rothbard (1977), for example.

20. It will be difficult enough to convince the average consumer of the benefits to be gained from free banking generally. It will be enormously more difficult to persuade consumers that they should (1) redeem the media of exchange in assets, such as gold or “investment-grade securities,” but denominate their accounts in units of “standard commodity bundles” (Yeager and Greenfield 1989, 413).

21. In fact, it is likely that, if gold is used as the MOR, the system will soon evolve into one of direct convertibility. There is a wealth of logical and historical evidence that suggests that a single commodity will dominate a basket of commodities. It would be ironic indeed if the BFH model were tried and it transmuted into the WS model.

22. The process works in the opposite direction, too. However, the effects are “asymmetrical because there is a limit, namely zero, to how low reserves of the redemption medium can fall but no definite limit to how high they can rise” (Woolsey and Yeager 1991, 4).