Quarterly Journal of Austrian Economics
Free Banking and Fractional Reserves: A Comment
Volume 1, No. 3 (Fall 1998)
The article by Jörg Guido Hülsmann (1996), “Free Banking and the Free Bankers,” is an important contribution to a proper understanding of free-banking systems. He is perfectly correct in blaming some advocates of free banking who support arguments which are irrelevant or wrong, such as the argument according to which free banking makes possible a higher rate of investment, due to increased credits, the argument according to which a 100-percent-reserve system would imply higher costs, would impede financial intermediation, or would lead to money shortages, etc.
Thus, although I basically agree with Hülsmann when he blames both opponents of free banking and some free bankers who do not develop the right arguments, I cannot agree with his plea in favor of a 100-percent-reserve system. To be sure, one has the feeling that in such a system holders of money substitutes (notes or deposits) consider it a certainty that they can get the promised amount of money (gold) in exchange for their money substitutes, whenever they want. However, we also know that a situation of perfect certainty cannot exist. In any human system a certain degree of risk does exist and the perception of this risk by different individuals is normally different. Therefore, there is always a problem of the optimal distribution of risks between individuals with different risk perceptions and risk preferences. In a 100-percent-reserve system, for instance, there is not perfect certainty that money holders can get their money back, since the issuer of money substitutes may go bankrupt. Bankruptcy is caused by the fact that a firm is not profitable; and this may happen in a 100-percent-reserve system, since the issuing of money substitutes is costly (for instance printing the notes, replacing those which are used, investing against counterfeiting, protecting the gold deposited in vaults, etc.). One may even say that, in such a system, the issuer of money substitutes necessarily fails, except if it can find some means to make people pay for the services it offers them.
It may not be difficult to charge people for part of the real cost issuers of money substitutes have to bear: for instance, they may charge a higher gold price for notes when they sell them than when they buy them back to compensate for the cost of printing notes. But how to make people pay for the cost of gold storage? This cost is in fact proportional to the length of time which the gold-backed notes are held in the vaults of the issuer. Therefore, it would be rational for the issuer of bank notes to buy back its notes at a gold price which would decrease with time. But notes issued at different dates would have different gold prices and would not be perfect substitutes, so their liquidity would be lowered.
In a perfectly free banking system, everyone must be free to offer any type of notes and to charge customers for his services in any way he can imagine. And any customer must be free to choose the kind of notes and the system of payment for services he prefers. One possible way for the issuers of money substitutes to make people pay for the cost of holding gold is nothing else than a fractional-reserve system. In fact, in such a system, the return obtained by the issuer of money substitutes is proportional to the length of time during which people hold notes, since he receives interest on a fraction of the value of these notes, namely those which have credits and not gold as counterparts in his balance sheet. In a perfectly free banking system, different kinds of issuers, with different methods for charging customers, may (at least potentially) coexist on the market. But, we cannot decide from outside that a 100-percent-reserve system is optimal, since optimality cannot be defined independent of the wants of individual. By defending a 100-percent-reserve system against a fractional-reserve system, Hülsmann takes a constructivist position (which is contradictory to his anti-utilitarian orientation in other respects).
The very concept of a free-banking system—i.e., free competition between producers of money substitutes—is not compatible with the a priori idea according to which a 100-percent-reserve system must be the only acceptable system. Moreover, there are only two ways of selecting one or a couple of monetary systems (100-percent-reserve system or fractional-reserve system):
1. It may be imposed by state regulation, in which case, for instance, the state would forbid fractional reserves and would impose a 100-percent-reserve system. Apparently Hülsmann does not suggest such a selection process which would be totally incompatible with the concept of a free banking system.
2. The monetary system may be selected by the “market,” which means that producers and users of money substitutes select one or a couple of monetary systems as the best means to meet their wants. If ever a 100-percent-reserve system is optimal—which means that it better meets the needs of producers and users of money substitutes—it will be selected, and fractional systems will not survive. But we cannot decide a priori that a 100-percent-reserve system is to be preferred; we have to experiment. It may happen that some people prefer a 100-percent-reserve system and others fractional systems, so that both systems coexist in a perfectly free banking system.
There is, therefore, some contradiction in the position held by Hülsmann. He is supporting the idea according to which a 100-percent-reserve system is optimal, but, apparently, he is not supporting a compulsory system by which fractional systems would be forbidden by the state. But if he was right about the optimal character of a 100-percent-reserve system, it is difficult to understand why the market would not select the optimal system. Moreover, according to Hülsmann, there would be a paradoxical situation in which competition would induce people to choose a fractional system and in which such a system would be ineluctably unstable and self-destructing. It is difficult to understand why, in the specific case of money, and only in this case, there would be any sort of market failure.
Hülsmann has another line of defense. He maintains that a bank developing a fractional system is fraudulent, since it is promising customers that it will redeem money substitutes for a fixed quantity of gold, and it is unable to do it. Some may deduce from this interpretation that the state is entitled to impose 100-percent-reserves, insofar as it would be in charge of preventing frauds. But is there really any element of fraud in a fractional-reserve system?
Let us in fact assume that, as a potential customer in a perfectly competitive monetary setting, I am offered the choice between two possibilities, namely holding notes issued by a producer under a 100-percent-reserve system or holding notes issued by a producer under a fractional-reserve system. Let us also assume that I have perfect information about the ways both producers operate. In the first case, I may have to pay for monetary services offered to me by the producer of money substitutes, or I may have to bear the already mentioned uncertainty costs implied by the low profitability of this producer. In the second case, I know that it may happen that I cannot redeem my money substitutes for money (gold). In both cases, the contracts I am signing with the producers of money substitutes are perfectly clear, and I have all the available information. If I decide that it is better for me to contract with the second producer and to bear the risk of illiquidity, I am totally responsible for this choice, and there is not the slightest element of fraud in the behavior of the banker and in our contract. I just prefer a certain structure of risks and charges over others and I decide that, from my own point of view—the only one to be considered—a fractional system is optimal. There is no justification for the claim of Hülsmann (or Murray Rothbard, or any other respectable economist) to decide on my behalf.
Even if we cannot demonstrate that a fractional system is preferable to a 100-percent-reserve system—anymore than Hülsmann can demonstrate the superiority of a 100-percent-reserve system—we get some information from history. In fact, the first notes were backed by 100-percent-reserves, but fractional systems emerged little by little, and there was certainly a period during which both systems were competing. And fractional systems were selected not as a consequence of some state regulation forbidding 100-percent-reserve systems, but because they better met the needs of producers and holders of money substitutes.
Thus, let us assume that initially all monetary systems are based on 100-percent-reserves and that, little by little, they are transformed into fractional-reserve systems, just because these systems are preferred by the public and the money producers. It is quite true that, during the whole process of adjustment from one system to the other, there is a multiple creation of money substitutes, with all related effects (inflation, excess credits, over-investment, etc.). These effects are costly, but they may be viewed as a type of investment costs, those which have to be borne in order to shift from one given system to another preferred system. Now, these costs are only temporary costs (as any investment cost is). In fact, each issuer of money substitutes in a fractional-reserve system decides upon the reserve ratio which is viewed as optimal by him. By expanding the quantity of money substitutes beyond that point, the producer would increase unit gains, but he would lose the confidence of his clients and, probably, he would lose clients, which would imply a negative marginal gain. In a fractional-reserve system, there is thus a long-run equilibrium structure of reserve ratios. As soon as all issuers of money substitutes have reached their optimal positions, there is no more multiple expansion of money substitutes. (Or, to put it differently, the rate of growth of money substitutes is perfectly equal to the rate of growth of money, i.e., gold.) It means that, contrary to what Hülsmann seems to suggest, there is no possibility for an indefinite expansion of money substitutes. It is constrained by the amount of money (gold). If the possibility of an unlimited creation of money substitutes existed, it would mean that the free decisions of individuals, in a free-banking system with fractional reserves, would lead to a “socially” non-optimal situation, characterized by an excessive issue of money substitutes. By failing to distinguish between the long-run working of a fractional system and the (short-run) adjustment from a 100-percent-reserve system to a fractional system, Hülsmann suggests that a fractional system is costly and does not provide an optimal quantity of money. But, as we stressed above, there is an over-expansion of money substitutes only during the intermediate period and the effects of this over-expansion can be interpreted as investment costs.
Similarly, we cannot accept the idea expressed by Hülsmann according to which a fractional system is necessarily unstable and creates a systemic risk because of a contagion effect. According to him, whenever a sufficiently large bank goes bankrupt, the whole system breaks down. Such an effect may happen, but, contrary to Hülsmann’s presentation, it cannot be considered as an inherent feature of every fractional system. In fact, the systemic risk depends on the precise institutional characteristics of the system. For instance, if the convertibility contract between a bank and its customer includes an option clause according to which the bank can suspend convertibility for a while, the holder of money substitutes does know in advance that there is a risk and that he may be obliged to remain in the position of a creditor of his bank longer than he would have expected. But an illiquidity crisis does not necessarily imply the failure of the bank (for instance if it has a high net worth and it can sell a huge amount of assets (Nataf 1993, pp. 97–106)). And if it goes bankrupt, it may well happen that all holders of money substitutes finally get back the gold counterpart of their holdings.
Let us also imagine the case of a banking cartel in which all participants give convertibility guarantees for their own notes against the notes of others. These guarantees may be conditional, which means that whenever one bank is expanding money creation too much and there is a potential or actual risk of illiquidity, the other banks may suspend their convertibility guarantees. In such a case, there is a clear signal given to the market that all banks are safe except the one which, in fact, has been evicted from the cartel. Therefore, there is no reason for the holders of money substitutes issued by these banks to be anxious about their ability to meet their contractual obligations and there will be no run on their gold reserves. There may be a great number of institutional devices which may make the probability of systemic risk unimportant. As the owners of banks do not wish to go bankrupt, in a free banking system they would experiment with different institutional systems which might be able to avoid systemic risk. If there are several monetary cartels issuing various money substitutes and competing one with one another, they may develop different institutional systems and, little by little, the best ones will be selected. Therefore, there is no logical basis for claiming that an indefinite expansion of money creation is unavoidable in a fractional system. It all depends on the specific characteristics of the monetary system.
Hülsmann, Jörg Guido. 1996. “Free Banking and the Free Bankers.” Review of Austrian Economics 9, no. 1: 3–53.
Mises, Ludwig von. 1949. Human Action. New Haven, Conn.: Yale University Press.
Nataf. Philippe. 1993. “New England’s Depression-Proof Banking System.” In Perspectives on the History of Economic Thought. Vol. 9. Robert F. Hébert, ed. London: Edward Elgar.
Rothbard, Murray N. 1996. Classical Economics. Vol. 2. An Austrian Perspective on the History of Economic Thought. Brookfield, Vt.: Edward Elgar.
White, Lawrence. 1995. Free Banking in Britain. 2nd ed. London: Institute of Economic Affairs.
Pascal Salin is professor of economics at the Université Paris-Dauphine. I wish to thank Philippe Nataf for several suggestions made from a previous version of this comment.
 It seems that, before 1650, scarce commercial banks issued “notes” for deposits of metallic money. Later on, British goldsmiths and the Bank of Stockholm issued notes for commercial bills, thus opening the way for fractional reserves.
 Later on, as everyone knows, the state monopolized the production of money substitutes and made possible an indefinite expansion of money substitutes via a steady decrease in the reserve ratio. But this is a completely different story. In that case, fractional systems—and their actual workings—would have been adopted not because of their ability to meet the needs of individuals, but because of the possibility for the state to extract monopoly profits.
 Lawrence White (1995, pp. 141–42) perfectly explains why banks are not inclined to exceed this optimal reserve ratio.
 This was stressed by Ludwig von Mises, according to whom “Free banking keeps credit expansion within narrow limits” and, even, “within very narrow limits” (Mises 1949, p. 443). According to Philippe Nataf, statistics of the Massachusetts House and Senate show that bank credit expansion under free competition has been kept below 1 percent a year for more than five decades. One may also quote Murray Rothbard, who wrote that “if these banks are not to be based on 100-percent specie reserve, which Say indicates would be the best system, competition would keep (the banks) investing in sound, very short term credit which could easily be used to redeem their bank notes” (Rothbard 1996, p. 40).
 We certainly agree with him when he stresses that any quantity of money meets the needs of money users.