What do we know about Currency Competition?

Hellwig, Martin F. Article What do we know about Currency Competition? Zeitschrift für Wirtschaftsund Sozialwissenschaften (ZWS) - Vierteljahresschrift der Gesellschaft für Wirtschaftsund Sozialwissenschaften, Verein für Socialpolitik Provided in Cooperation with: Duncker & Humblot, Berlin Suggested Citation: Hellwig, Martin F. (1985) : What do we know about Currency Competition?, Zeitschrift für Wirtschaftsund Sozialwissenschaften (ZWS) - Vierteljahresschrift der Gesellschaft für Wirtschaftsund Sozialwissenschaften, Verein für Socialpolitik, ISSN 0342-1783, Duncker & Humblot, Berlin, Vol. 105, Iss. 5, pp. 565-588, https://doi.org/10.3790/schm.105.5.565 This Version is available at: https://hdl.handle.net/10419/291619 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ What do we know about Currency Competition? By Martin F. Hellwig* The paper presents a critical analysis of the proposals of Hayek and Vaubel for unregulated competition among private money suppliers. Because of externalities, time inconsistencies and moral hazard, these proposals are detrimental for outside money and, at best, dubious for inside money. I am skeptical about the price theoretical foundations of Vaubel's1 policy recommendations. This skepticism extends to the work of von Hayek (1976, 1977), which initiated our current concern with the optimal monetary constitution. Specifically, I have the following problems with the Hayek-Vaubel analysis. A. Both, Hayek and Vaubel, neglect the distinction between inside and outside money. Their discussion of competition among outside monies is based on an invalid premise. B. There are Pareto-relevant externalities in money demand decisions which justify the use of lumpsum taxation to create a real return on money. C. In contrast to the market for an inside money, the market for an outside money is destroyed by the coexistence of more than one firm in the market. D. Vaubel and Hayek fail to distinguish between the dynamic problem of time inconsistency and the static problem of monopoly power. In the absence of binding money supply announcements, the time inconsistency of profit maximizing policies rules out any unregulated private organization of the market for outside money. E. The analysis of competing inside monies pays too little attention to the problems of uncertainty, information asymmetries, and time inconsistency that are endemic to debtor-creditor relations. In the following, I shall discuss these points one by one. * This paper was presented as a comment on Vaubel (1985) at the May 1984 meeting of the Ausschuß für Geldtheorie und Geldpolitik of the Verein für Socialpolitik. I thank Michael Rey for research assistance and the Deutsche Forschungsgemeinschaft for financial support through Sonderforschungsbereiche 21 and 303. 1 Vaubel's analysis of external effects in the money market is presented in Vaubel (1984). My comments here cover that analysis as well because at the May 1984 meeting, it was presented as an integral part of the argument. 37* OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 566 Martin F. Hell wig 1. Inside versus Outside Money We must distinguish between inside money, which gives its bearer a legal claim against the issuer, and outside money, which entails no such claim. Suppliers of inside money are constrained by the need to fulfill their obligations or else go bankrupt. Suppliers of outside money are under no such constraint. Presumably then, the behaviour of a money supplier will depend on whether he issues inside or outside money.2 The distinction between inside and outside money will also affect the demand for money. My willingness to hold paper outside money depends only on the prospect of selling this paper outside money to somebody else, who in turn is willing to pay a positive price only because he hopes to resell it to a third agent, who ... In contrast, the decision to pay a positive price for inside money is at least partly motivated by the prospect of calling the claim on the issuer. The shopkeeper accepts my check — not because he expects to resell it to his wholesaler, but because he will present it to my bank. Whereas the real value of outside money is exclusively determined by resale considerations, i.e. by expectations of its real value in future transactions, the real value of inside money will depend on the "fundamentals" of the underlying claim against the issuer. Any positive or normative analysis of the monetary constitution must take account of this difference. Hayek and Vaubel neglect it, apparently because they believe that the existence of outside money itself is merely a consequence of government interference with the monetary system. According to this view, unregulated currency competition would lead to the disappearance of outside money and its replacement by inside money as a superior asset.3 However, as far as I can see, this point remains to be proved. Moreover, it is not clear that such an outcome is in fact desirable. For the economy as a whole, the use of paper money may be advantageous because it economizes on the holding of real assets. In a pure inside money economy, this advantage is partly lost because the supplier of money must hold a reserve against the claims on him. In the absence of outside money, this reserve will consist of real assets such as gold, 2 In stressing this distinction, I do not take issue with Johnson's (1969) proposition that "the real theoretical difference to be drawn is between interest-bearing and noninterest-bearing money". Johnson was concerned with the determination of the economy's net wealth rather than the behaviour of money suppliers. Even in the context of his analysis, the distinction between inside and outside money is apparent in the inside money suppliers' need to hold reserves against withdrawels. 3 This proposition was explicity asserted by Vaubel in the oral discussion following the presentation of his paper and my comments. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 567 machines, shares, etc. Quite possibly then, the pure inside money economy may be inefficient because it involves an oueraccumulation of real assets. To make this point precise, consider the precautionary money holding model of Bewley (1980, 1982) and myself (1980, 1982). In this model, agents with uncertain commodity endowments save and hold assets for self-insurance against future endowment fluctuations. These agents' portfolio choices between money and commodity inventories depend on their expectations about real rates of return. Agents will be indifferent between money and inventories if both assets have a zero own rate of return and if prices are nonrandom and constant over time. A monetary rational expectations equilibrium with nonrandom, constant prices does in fact exist if the endowment risks of different individuals cancel out so that all economic aggregates are nonrandom. In this equilibrium, the use of paper money as a store of value enables the system to economize on commodity inventories. This substitution of money for inventories is useful because commodity inventories require a deferral of consumption and thereby involve a real cost. Up to this point, the argument does not depend on whether we are dealing with an outside money which happens to have a nonrandom, constant purchasing power or an inside money whose purchasing power is supported by an instant repurchase promise of the issuer. In the absence of aggregate fluctuations, the inside money issuer need not hold any inventories because in each period, the public's demand for his money just balances the available supply.4 However, if money does not bear interest, the private incentives for holding money are too small for Pareto optimality.5 In the present context, the use of non-interest-bearing money as a buffer stock does not yield perfect insurance of individual risks even though such perfect insurance would be feasible. If we are dealing with a government supplied outside money, then the allocation can be improved by using the government's power to tax in order to create a positive real return on money.6 This device is not available for a privately supplied inside 4 The substitution of a money without backing for a money with backing is perhaps best illustrated in the story, told by Peter Kenen, of the island which used sardine cans for money. A tourist once opened a can, found the sardines inedible, and complained to the person who had given him the can. The answer was: "There is nothing to complain about. I gave you money sardines. If you wanted food sardines, you should have gone to the supermarket!" s Friedman (1969), Johnson (1969), (1970). e Hellwig (1982). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 568 Martin F. Hellwig money. Such a non-interest-bearing inside money would in fact be displaced by any government supplied outside money which has a positive real return.7 Furthermore, even without interest payments on money, the equivalence of inside and outside money disappears if the underlying uncertainty involves collective as well as individual risks. In this case, the real quantity of money that the economy wants to hold will fluctuate with the collective endowment realization. Therefore, the supplier of an inside money must expect that with positive probability the claims on him will be called. If the inside money represents a claim to real commodities, he must be prepared to actually deliver these commodities. If he wants to be sure to avoid bankruptcy — i.e. a default on his promise — then he typically needs to hold a 100 %> reserve against his money issue.8 In equilibrium then, inside money will have the same return structure as the underlying real asset to which it is a claim. Whether such an inside money can coexist with a non-interest-bearing outside money depends on (i) the extent of collective risk and (ii) the own rate of return and the carrying cost of inventories. The non-interest-bearing outside money is displaced by inventories or an inside money backed by inventories if the own rate of return on inventories is zero; it is not so displaced if inventories have a high carrying cost so that their own rate of return is close to — 100 %>.9 Even in those instances in which a non-interest-bearing outside money is displaced by an inside money with a value guarantee or repurchase clause, this result is socially undesirable. In this case, the portfolio choice between outside money and inventory-backed inside money is biased against the former despite the potential allocational role of paper outside money as a socially costless buffer stock against individual, i.e. insurable endowment risks. As in the case of purely individual risks, welfare would be increased by the introduction of a tax-financed real return on outside money. 7 The argument in the text is based on the assumption that commodity inventories with a zero own rate of return are the only asset in the model. If we introduce real capital with a neoclassical production function, the argument must be modified along the following lines (due to Truman Bewley in private correspondence): Any non-interest-bearing money is displaced by real capital or by an inside money which promises the same real return as real capital and is backed by real capital. However, because of the precautionary demand for saving, there would be an over accumulation of capital to a point where the net marginal product of capital is less than the rate of time preference in consumption. Again, welfare can be increased by the introduction of an interest-bearing outside money which is backed by the government's power to tax. s The problem of default by an issuer of inside money, which is neglected by Hayek and Vaubel will be taken up in Section 6 below. » Hellwig (1980). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 569 In summary, if we are interested in the impact of competition on the monetary system, we must consider both, outside and inside monies, and we must be careful to draw the distinction between them. 2. Competition among Outside Monies Both Hayek (1976) and Vaubel address the problem of competition among outside monies when they consider "the case for free currency competition among central banks". They claim that such competition "encourages less inflationary monetary policies" because the currency with the lowest inflation rate will be the most attractive to a public that is free to choose among the different currencies. This claim rests on the implicit assumption that each issuing bank can at least partially control the inflation rate of its own currency through its supply behaviour. This premise is invalid. To illustrate the basic problem, I consider an example of competition among two "currencies" that is taken from the recent past. The first Table 1 Annual Growth Rates over Preceding Year Relative Price "Currency 1" "Currency 2" 1970 6,6 0,6 2 1971 9,5 1,8 2 1972 13,6 6,9 2 1973 5,3 - 1,2 2 1974 9,1 2,0 2 1975 9,1 3,5 2 1976 6,2 2,7 2 1977 12,0 6,0 2 1978 11,7 7,2 2 1979 5,0 2,0 2 1980 5,0 2,3 2 1981 0,1 - 0,7 2 Cumulative Growth 1970 - 1981 128 °/o 38 o/o Source: Statistisches Jahrbuch für die Bundesrepublik Deutschland. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 570 Martin F. Hellwig two columns of Table 1 show the evolution of the outstanding quantities of the two "currencies". Over the period 1970 to 1981, "currency 1" grew by more than three times as much as "currency 2". Yet the last column shows that the relative price of the two "currencies" did not change. In each period, two units of "currency 2" were treated as a perfect substitute for one unit of "currency 1". In accordance with the Hicks-Leontief aggregation theorem, both "currencies" were in fact treated as parts of a single composite currency: Inflation concerned this composite currency as a whole rather than its individual components. The above average growth rate of "currency 1" probably raised the inflation rate of the composite currency and hence the common inflation rates of all individual components; it did not induce an above average inflation for "currency 1". In this example, "currency 1" are blue pieces of paper on which the Bundesbank has printed the number "100"; "currency 2" are brown pieces of paper on which the Bundesbank has printed the number "50". We must now see to what extent the analysis of this example can also be applied to outside monies that are issued by different agencies which compete with each other. First we need to note that the relative price of two DM 50,— bills for one DM 100,— bill is in fact a market price. One might object that DM 100,— is twice DM 50,— merely as a matter of arithmetic. However, here we are not concerned with arithmetic but with the price at which blue and brown pieces of paper, i.e. different physical objects, are exchanged in actual transactions. If you need to make an emergency phone call from a public phone booth at night, you may find yourself willing to part with a DM 10,— bill for much less than the ten DM 1,— coins that would be indicated by arithmetic.10 Similarly, the scarcity of Italian 5 and 10 lire pieces is due to the fact that the central bank's arithmetic stands in no relation to the value of the nickel contained in these coins. Once the relative price of DM 100,— and DM 50,— bills is seen as a market price, we must ask why the market seems to conform so well to the Bundesbank's arithmetic. The Bundesbank's readiness to intervene in support of the exchange rate of two DM 50,— bills for one DM 100,— bill provides only part of the explanation for this phenomenon. After all, there was a time when the Bundesbank also stood ready to support an exchange rate of DM 4,— for 1,— US It was significantly less sucio At the time of the opening of the then ultramodern Dallas-Fort Worth airport, the airport administration tried to cash in on this observation by installing money-change machines that returned 90 c in coins for a 1 $ bill. After considerable public protest, they had to abandon the idea. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 571 cessfull then. Moreover, we observe that the Bundesbank hardly even needs to intervene in the market of DM 50,— for DM 100,— bills. What would happen, if it did not intervene at all? I submit that even in the absence of current or expected future interventions by the Bundesbank, the market might clear at the constant relative price of two DM 50,— bills for one DM 100,— bill. For consider any agent who expects two paper outside monies to have a relative price x in all future transactions. This agent will regard x units of one currency as a prefect substitute for one unit of the other currency. If the current relative price of the two currencies is also x, he will be indifferent between them; if the current relative price differs from xf the agent will have a strict preference for one of the two currencies. If the expected price x is the same for all agents, the current equilibrium price must also be x since otherwise no agent would be willing to hold the currency that is too expensive. If at all times all agents expect two DM 50,— bills to exchange for one DM 100,— bill in all future transactions, then two DM 50,— bills will exchange for one DM 100,— bill in all actual transactions even if the Bundesbank does not intervene at all. Thus the fixed exchange rate of two DM 50,— bills for one DM 100,— bill will correspond to a rational expectations equilibrium. However, the same argument shows that if there is no Bundesbank intervention, then any other fixed exchange rate beween DM 50,— and DM 100,— bills will also correspond to a rational expectations equilibrium. More generally, any economy with multiple fiat monies whose use in different transactions is not subject to exogenous constraints will have multiple rational expectations equilibria (provided it has any rational expectations equilibrium at all). For any vector of exchange rates x > 0, such an economy will have an equilibrium in which all transactions take place at the fixed exchange rates x.11 This analysis is directly applicable to the Hayek-Vaubel discussion of currency competition among central banks. At present of course, a dollar bill and a DM coin cannot be used side by side in all transactions. However, the Hayek-Vaubel proposal aims precisely at lifting those restrictions which limit the use of dollars in Germany and of marks in the United States. If this proposal is realized, then the two currencies will circulate side by side and their relative acceptability in an individual transaction will depend only on the participants' expectations about their relative resale value in future transactions. In such a world, there can be no systematic differences in inflation rates between outside monies: Any anticipation that the mark will be relatively worthless in the future must make it relatively worthless today already. 11 Formal analyses of this principle are presented by Girton / Roper (1981). Hellwig (1976), Kareken and Wallace (1981). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 572 Martin F. Hellwig 3. External Effects in Money Demand12 I now consider the question whether money gives rise to Paretorelevant external effects. It will be convenient to distinguish between the external effects in money demand and the external effects in the decision to use money rather than barter. First, I must take issue with both Friedman's (1969) formulation and Vaubel's criticism of the price-level externality of money demand. Both authors use the language and the tools of static equilibrium theory for what is essentially a non-static, sequential problem. Unlike a refrigerator, money is not an asset that one buys once in order to hold it forever and to enjoy the "liquidity services" that it yields. Instead, money is traded back and forth: one acquires it, then resells it, acquires it again, etc. "Liquidity services" do not arise from the possession of money as such, but from the ease with which money can be resold. If one expects that with probability one, one will never actually use this resale opportunity, then one has no reason to hold money in the first place. The use of money must therefore be analysed in a framework of sequential, incomplete markets rather than the usual set of simultaneous, complete markets. In the sequential framework, the different periods and occasions in which agents trade money must be tied together by the concept of a rational expectations equilibrium, i.e. an "equilibrium of plans, prices, and price expectations".n In such a non-static setting, there is no presumption that pecuniary externalities are Pareto-irrelevant.14»15 For consider the effects of an increase in some individual's demand for money in period t. From the perspective of period t, this demand increase raises the value of money in that period, thereby conferring a positive pecuniary externality on all net sellers and a negative pecuniary externality on all net buyers of money in money in period t. Ceteris paribus, Vaubel is right in observing — against the formulation of Friedman (1969) — that these pecuniary externalities are Pareto-irrelevant. However, from the perspective of period t — 1, the increase in the (real) value of (nominal) money in period t serves to raise the indirect expected utility associated with money holdings at the end of period t — 1. In terms of the temporary equilibrium of period t — 1, this effect is a non-pecuniary externality and is definitely Pareto-relevant. It tends to raise the equilibrium value of money in period t — 1, which in 12 See Footnote 1. 13 Radner (1972). 14 See, e. g., Scitovsky (1954), 184 f. is Contrary to footnote 14 in Vaubel (1984), a main point of Scitovsky's classic paper is the Pareto-relevance of pecuniary externalities outside the narrow framework of static equilibrium theory. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 579 The question is how the franchise is to be awarded. If there is no conflict about the ranking of any two bids {M],T}} and {M^, T?}, then the regulatory commission should simply award the franchise to the bidder whose bid is (unanimously) ranked highest. The bidders are thus involved in a type of Bertrand game in which they make "contract" offers and "the public" chooses whichever contract maximizes its utility. The usual Bertrand argument shows that in an equilibrium of this game, the winning bid must be the one that is most highly ranked among all those that are technically feasible and do not impose a net loss on the bidder. In contrast to the unregulated monopoly, the approach would always lead to the second best monetary policy, i.e. the one that is best among all policies that do not rely on the government's power to tax. Moreover, in this approach even the seigniorage would not stay with the winning bidder but would be channelled back to the economy through the seigniorage taxes Tt. The preceding analysis has its weak spot in the assumption of unanimity among the users of money. In general, there is no reason to expect such unanimity. Different people with different tastes will have different views about monetary policies and the distribution of seigniorage. In this case, the criteria of the regulatory commission become problematic. However, the distributional conflicts that arise are no different than the distributional conflicts arising in any other area of collective choice, or, more narrowly, in any other regulatory problem. The problem of regulating the supply of outside money no longer is a problem sui generis, but it has been brought into the confines of traditional public choice and welfare analysis. Even if one is pessimistic about the possibility of resolving the distributional issues that arise, one may still expect that a commission of the sort that is suggested will pay rather more attention to the public's aversion to inflation and less attention to the money suppliers' desire for seigniorage than an entirely unregulated monopoly. 5. The Problem of Time Inconsistency I now turn to what is probably the most important problem for any monetary constitution. Up to now, I assumed that the sequence of money supplies is announced before the first period and that this announcement cannot be revoked in any later period. In practice there is no reason why such initial announcements should be binding at later dates. The question then is how the money market behaves in the absence of binding announcements of future money supplies. 38 Zeitschrift fttr Wirtschaftsund Sozialwissenschaften 1985/5 OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 580 Martin F. Hellwig If initial announcements are not binding, the optimal policy precommitment {Mt} of an unregulated private monopoly that was discussed in Section 4.1 no longer is an equilibrium. This conclusion is obvious in those cases in which the money supplies Mt are constant and ¡ut = 0 for all t. With zero money growth, the monopolist earns the revenue tzq Mo in period zero and nothing thereafter. From the perspective of period 0, this may be a good policy because zero money growth and zero inflation in later periods enhance the real revenue m Mo in period 0. From the perspective of period 1, the revenue rcoMo of period 0 is forever bygone and does not enter into the monopolist's considerations any more. From the perspective of period 1 the zero money growth policy continuation with zero revenues in all periods t = 1 is dominated by a policy of positive money growth and positive revenues in some periods. More generally, let {Mt} be the optimal policy precommitment of the private monopolist, and recall that this policy yields the real revenues So in period 0 and-^—St in period t ¡> 1, where pn = (Mt — Mt-i)/Mt-i 1 + fa is the money growth rate and So, Si, ... are the economy's real money demands in periods 0,1, ... at the given rates of return. For a given quantity of money Mo in period 0 and given money growth rates pn for t ^ 2, the quantity of money Mi in period 1 affects only the real money Uf demand So in period 0 and the seigniorage income — Si in period 1. 1 + iut The seigniorage income is an increasing function of ¡x\ and hence, for given M0, of Mi. Nevertheless, from the perspective of period zero, it is not desirable to set ju\ = oo because a large value of Mi entails a low real value of money 7it = Si/Mi at date 1, and under rational expectations, a low real value of money no at date 0. (If inventories have no carrying costs, arbitrage between inventories and money ensures tto = ; if arbitrage considerations impose no bound on the rate of inflation, intertemporal substitution will reduce So.) However, from the perspective of period 1, this consideration plays no role, and the revenue "maximizing" policy requires an infinite money growth rate. The basic rationale of the argument is very simple: At any date f, a fixed pattern of money growth rates jLit + i, pit• • • induces a certain pattern of expected inflation rates, which determines the real resources St that the economy is willing to spend on its money holdings at the end of period t. These real resources St are shared between the monopolist LLt 1 and the previous holders of money in proportions and . By 1 + ft 1 + t*t making pn indefinitely Zarge, the monopolist can dispropriate the previous money holders and raise his portion of the quantity St that goes into money. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 581 In general then, the monopolist's revenue-maximizing policy is timeinconsistent25 because in later periods, the monopolist wants to deviate from this policy. It follows that the monopolist's initial announcement will not actually be credible unless he can devise an institution that makes this announcement binding. Moreover, it is now easy to see that in the absence of binding precommitments about future money supplies, there cannot be any equilibrium in which the value of money is positive. The preceding arguments show that no matter what situation we are considering, as of period t, the monopolist has an incentive to make Mt and JLit arbitrarily large (and hence nt arbitrarily close to zero). Under rational expectations, this future behaviour of the monopolist is anticipated by the market. With this anticipation, the market sets St = nx = 0 for t < t because nobody wants to spend real resources on an asset that will be made worthless by the monopolist's future behaviour. In general St = nt = 0 for all t is the only possible equilibrium. In summary, an unregulated private monopoly without binding commitments destroys the use of outside paper money just as surely (and by almost the same argument) as the coexistence of several competing outside money supplies.26 I conjecture that this conclusion does in fact hold for any unregulated private organization of a market for outside paper money. I also believe that the problem of time inconsistency bedevils any government run or regulated monetary system. This is obvious if the government itself behaves like a revenue-maximizing monopolist. Most of von Hayek's (1977) historical overview illustrates the very conflict between the monetary stability that is promised to make money acceptable and the money growth that is generated later to raise revenues. However, time inconsistency would probably be a problem even if the government were run by welfare economists who do not try to maximize seigniorage revenue per se. For consider again the model of Bertrand competition among potential money suppliers that was discussed in Section 4.3. Suppose that the regulatory commission has awarded the franchise to a firm with a bid {Mt, . In period 0, 25 This time-inconsistency was first discussed by Calvo (1978). For the general problem of time-inconsistency of monopoly in a durable goods market, see Coase (1972) and Stokey (1981). 20 See Coase (1972) on the analogy between the durable goods monopoly and perfect competition. Because of the peculiarity of money, the conclusion here is even stronger than Coase's, which holds only if the time span between subsequent market dates is small, see Stokey (1981). 38» OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 582 Martin F. Hellwig the initial money supply Mo and seigniorage tax To are implemented. Now in period 1, the winning firm (or some other firm) presents a new bid {Mt, T*)~=0 to the regulatory commission, which it finds preferable because the net seigniorage revenues nt (Mt — Tt) under the new policy are higher than the net seigniorage revenues nt (Mt — Tt) under the old policy. How will the regulatory commission react to this new bid? At this point, there is going to be an important distributional conflict in the economy: Those who have held money from period 0 will object to any increase in the quantity of money Mi because it reduces the value of their own money holdings; those who do not hold any money will not object unless the new proposal also raises the inflation rate from period 1 to period 2 and thereby worsens the intertemporal price ratio with which they are faced. The latter agents will actually favour the proposed policy change if they can share in the spoils by obtaining a large portion of the seigniorage tax Ti. The regulatory commission's reaction to the proposed change in monetary policy will therefore depend on the commission's composition and on its rules of procedure. Specifically the question is how the commission's rules of procedure adjudicate the distributional conflict between money-holders and non-money-holders.27 The most conservative rule would require unanimity for any changes in policy and would thereby give either group an effective veto. A unanimity requirement would probably eliminate the problem of time inconsistency by making it impossible to change monetary policy. On the other hand, a unanimity requirement will make it hard to agree on a winning bid in the first place. Moreover, it .might be desirable to discipline the firm that has been awarded the money supply franchise by threatening to give the franchise to another bank if it fails to comply with the terms of the contract. If such a move requires an unanimous agreement by all members of the commission, then the threat is not very effective, and the existing supplier of outside money can try to violate the terms of the winning bid without much fear of repercussions. However, for any voting rule that does not require unanimity of decisions about monetary policy, time inconsistency is likely to be a problem. I suspect that time inconsistency is indeed the deepest and least solvable problem for the monetary constitution. 27 In Section 4.3, this distributional conflict played no rule because prior to the determination of the winning bid there were not yet any money holders. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 583 6. Competition among Inside Monies To conclude the discussion, I briefly consider competition among inside monies, i.e. among monies whose issuers give their clients a claim of some sort or other. In the simplest case, the holder of an inside money has the right to obtain a certain specified quantity of a real good, an asset, or another money upon demand or at some prespecified date. The Hayek-Vaubel notion of a "value guarantee" is a bit more complicated because it does not seem to involve a legal claim of the money holder on the money issuer. Instead, the value guarantee is publicly announced as the guiding principle for future policy.28 However, as long as the money issuer fulfils his obligation, it does not matter whether the obligation arises from a policy announcement or from a legal claim. The distinction matters only when the money issuer defaults on his promise and the question is how one can make him pay. If we accept the usual treatment of demand deposits as "money", we see that most countries already have some competition among inside monies. However, this competition is rigidly regulated by the government. The Hayek-Vaubel proposal amounts to an outright abolition of all government regulation of this sector. In particular, they want to abolish the following regulations: a) The ban on the issue of private bank notes that can circulate as money. b) The requirement to hold (minimum) reserves in central bank money. c) The requirement to denominate the private money issuer's obligation in units of central bank money. I should wholeheartedly support these proposals if we lived in a world in which all agents are completely informed about everything and all contracts and promises are always honoured. Unfortunately, the very use of money has to do with the fact that we do not live in such a world. Given the imperfections and uncertainties of actual markets, I see no conclusive evidence either against or for government regulation of the banking system and the market for inside money. Economic theory simply has too little to say on these matters to warrant any firm conclusions of the sort Hayek and Vaubel want to draw. The relation between the holder and the issuer of an inside money is akin to that between a creditor and a debtor. This relation is problematic because when the contract is made the creditor surrenders a real asset and gets no more than a piece of paper with a repayment promise for the future. The whole creditor-debtor relation hinges on 28 Hayek (1977), 31, OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 584 Martin F. Hellwig the question what this promise is going to be worth. In considering this question one must deal with a whole spectrum of difficulties arising from uncertainty, moral hazard, asymmetric information, and again time inconsistency. These difficulties which beset the theory of credit markets are just as important in the market for inside monies. 6.1 Uncertainty and Product Heterogeneity The returns that the bank earns on its own investments are typically uncertain. Therefore, its own ability to fulfil its obligations to the holders of its money is uncertain. The returns on inside money will generally be uncertain and will depend on the bank's own investment policy. In consequence the inside monies issued by different banks will be less than perfect substitutes for each other. Inside monies must be regarded as a set of differentiated products rather than a single homogeneous product. Competition among inside monies then must be analysed as monopolistic competition in the sense of Chamberlin rather than perfect competition. In a world of Chamberlinian monopolistic competition, there is no presumption that the market outcome has any nice welfare properties.29 Both Vaubel and Hayek are aware that the market for inside monies must be analysed in terms of differentiated rather than homogeneous products. They do not seem to be aware that the welfare properties of monopolistic competition in a differentiated products market are quite unclear. 6.2 Moral Hazard and Bankruptcy The return that an inside money holder eventually gets depends on the behaviour of the issuer of the money. If the issuer selects poor investments, the holder of the inside money gets a poor return — like those depositors who suffered f rom Herstatt's bad currency speculations. If the issuer embezzles the company's funds, the holder of the inside money gets a poor return — like those IOS certificate holders who suffered from Mr. Vesco's depleting the fund's assets. In principle, the contract between the issuer and holder of an inside money might prescribe the most careful behaviour on the side of the bank; in practice, the holder has no way of enforcing such a clause. The recent literature on credit rationing and credit contracts shows that many institutional peculiarities in capital markets may be interpreted as devices that eliminate or reduce such instances of moral 29 See, e. g. Hart (1983). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 585 hazard. Thus, the standard debt contract with a fixed repayment obligation and bankruptcy if and only if the repayment obligation cannot be met may be interpreted as the market's response to the moral hazard that arises if the debtor (here the bank), but not the creditor (here the money holder) can costlessly observe the realized return on the debtor's investment.30 In the same setting, banks as intermediaries may serve to reduce the agency costs of financing final real investment.31 Credit rationing with bounds on both loan sizes and interest rates may serve to induce less risky investment policies by debtors and banks.32 Such devices reduce, but do not eliminate the problem of moral hazard in financial relations. The central issue is that the behaviour of a debtor, in particular the issuer of an inside money, exerts an external effect on the creditor, in particular the holder of the inside money. Because of this external effect, market allocations will generally not be more than n-th best, and it is unclear whether government intervention is harmful or useful. To some extent, Vaubel seems to see that moral hazard might be a problem. He suggests that the danger of "profit snatching" can be eliminated through value guarantees ((1985), 554).33 However, he does not see that such guarantees themselves might be subject to moral hazard and therefore might not be credible. A value guarantee, debt obligation and the like may eliminate moral hazard if the penalties for non-compliance with one's obligation are very large. If as in the Vesco-IOS case, the penalties are small in comparison to the gains from noncompliance, then such obligations may simply be irrelevant. Morever, even if the penalties are large enough to eliminate outright dishonesty, they may still not be large enough to ensure an appropriate investment policy ex ante.34 Could it be the case that minimum reserve requirements for banks or investment regulations for insurance companies are just one admittedly coarse way to "internalize" the effects that these companies' decisions have on their financiers through the risk of bankruptcy? 30 Gale / Hellwig (1983). 31 Diamond (1984). 32 Jaffee / Russell (1976), Stiglitz / Weiss (1981). 33 Vaubel himself dismisses the Klein-Tullock argument that moral hazard is less of a problem in a repeated-game setting in which banks care about their long run prospects. This argument requires that agents do not discount the future so that no matter how large the short turn gains from dishonest or negligent behaviour may be, they are always outweighed by the infinite tail of continued future relations with one's creditors. At least Mr. Vesco does seem to have had a positive discount rate. 34 See, e. g. Stiglitz / Weiss (1981). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 586 Martin F. Hellwig 6.3 Moral Hazard and Time Inconsistency If a debtor tells his creditor that he cannot pay, does the creditor call a bankruptcy or does he wait in the hope of sharing in the debtor's better luck in the future? Given the moral hazard problems discussed above, it seems desirable ab initio to threaten bankruptcy fairly quickly in order to induce the debtor to take care to avoid bankruptcy. However, after it has been determined that the debtor cannot pay, at least at present, the creditor may prefer to keep him alive. If bankruptcy is called immediately, the creditor has to write off his claims on the debtor. If bankruptcy is not called, there might be a time in the future when these claims could be collected. The decision to call a bankruptcy is thus subject to time inconsistency just like the optimal supply of outside money.35 Concrete examples of these considerations have been observed in recent proceedings concerning the City of New York as well as the so-called "International Debt Crisis". In those cases where the debtor is a large bank and the debts are inside money held by the public, the reluctance to call a bankruptcy seems to be especially great. In the case of the Continental Illinois Bank, it was made clear that because of adverse effects on the monetary system, a large bank would not be allowed to go bankrupt no matter how many bad loans it might have made. The problem is, of course, that if the banks know this, then they have no reason to avoid making bad loans. More generally, debtors who know that they will not be put into bankruptcy have only weak incentives to manage their means carefully so as to make sure that they can fulfil their obligations. In summary, I believe that the markets for inside monies and the larger set of capital markets of which they form part are so replete with market imperfections, information asymmetries and problems of moral hazard that we cannot make any firm assessment about the welfare properties of the outcomes in such markets. Whether government regulation in these markets is warranted at all, whether it should take the form it does take, is something that at present we do not know — unless of course we start from the axiom that everything would be for the best in the best of all possible worlds if only the government ceased interfering. If we do not accept this axiom, we must admit that we simply do not know very much about how competition among inside monies works. 35 Hellwig (1977). OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 What do we know about Currency Competition? 587 Summary The paper studies the proposals of Hayek and Vaubel for unregulated private competition in the money market. These proposals are shown to rest on an insufficient distinction between inside and outside money. The existence of an outside money without a backing is desirable on welfare grounds. However, any private supply of outside money in perfect competition, Cournot oligopoly or monopoly would actually destroy the use of outside money. The main problem is that of time inconsistency of the optimal money supply policy. Problems of time inconsistency and of moral hazard arise also in the market for inside money. Because of these problems, the appropriateness of unregulated competition in the market for inside money must also be doubted. Zusammenfassung Die Arbeit befaßt sich mit den Vorschlägen Hayeks und Vaubels zur Einführung des Wettbewerbs im Geldwesen. Es wird gezeigt, daß diese Vorschläge auf einer unzureichenden Unterscheidung zwischen Außengeld und Innengeld beruhen. Die Existenz eines Außengeldes ohne Deckung ist aus wohlfahrtstheoretischen Erwägungen wünschenswert. Ein privates Angebot an Außengeld im Wettbewerb, Cournot-Oligopol oder Monopol würde aber die Funktionsfähigkeit des Marktes für Außengeld zerstören. Zentrales Problem ist die Zeitinkonsistenz jeglicher Geldangebotspolitik. Zeitinkonsistenzprobleme in Verbindung mit „moral hazard" treten auch im Markt für Innengeld auf und lassen auch hier die Angemessenheit der Vorschläge von Hayek und Vaubel als zweifelhaft erscheinen. References Baumol, W., J. Panzar and R. Willig (1982), Contestable Markets and the Theory of Industry Structure. New York. Bewley, T. (1980), The Optimum Quantity of Money, in: J. Kareken / N. Wallace, Models of Monetary Economics. Minneapolis. — (1982), A Difficulty with the Optimum Quantity of Money. Econometrica 51, 1485 - 1504. Calvo, G. (1978), Optimal Seigniorage from Money Creation. Journal of Monetary Economics 4, 503 - 517. Coase, R. M. (1972), Monopoly and Durability. Journal of Law and Economics 15, 143 - 149. Diamond, D. (1984), Financial Intermediation and Delegated Monitoring. Review of Economic Studies 51, 393 - 414. Friedman, M. (1969), The Optimum Quantity of Money, in: M. Friedman, The Optimum Quantity of Money and Other Essays. Chicago. Gale, D. and M. Hellwig (1983), Incentive Compatible Debt Contracts: The One-Period Problem. Discussion Paper No. 124, Sonderforschungsbereich 21. Bonn. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54 588 Martin F. Hellwig Girton, L. and D. Roper (1981), Theory and Implications of Currency Substitution. Journal of Money, Credit, and Banking 13, 12 - 30. Hart, O. D. (1983), Monopolistic Competition in the Spirit of Chamberlin: A General Model. Discussion Paper No. 83/82. ICERD, London. Hayek, F. A. v. (1976), Choice in Currency: A Way to Stop Inflation. Institute of Economic Affairs, Occasional Paper 48. London. — (1977), Entnationalisierung des Geldes. Tübingen. Hellwig, M. F. (1976), A Model of Monetary Exchange. Discussion Paper No. 202, Econometric Research Program. Princeton. — (1977), A Model of Borrowing and Lending with Bankruptcy. Econometrica 45, 1879 - 1906. — (1980), Precautionary Asset Holding under Individual and Collective Risk. Paper presented at the World Congress of the Econometric Society. — (1982), Precautionary Money Holding and the Payment of Interest on Money. Discussion Paper No. 92, Sonderforschungsbereich 21. Bonn. Jaffee, D. W. and T. Russell (1976), Imperfect Information, Uncertainty and Credit Rationing. Quarterly Journal of Economics 90, 651 - 666. Johnson, H. G. (1969), Inside Money, Outside Money, Income Wealth, and Welfare in Monetary Theory. Journal of Money, Credit, and Banking 1, 30 - 45. — (1970), Is there an Optimal Money Supply? Journal of Finance 24, 435 - 442. Kareken, J. and N. Wallace (1981), On the Interminancy of Equilibrium Exchange Rates. Quarterly Journal of Economics 96, 202 - 222. Pesek, B.P. and T. Saving (1967), Money, Wealth and Economic Theory. London. Radner, R. (1972), Existence of Equilibrium of Plans, Prices, and Price Expectations in a Sequence of Markets. Econometrica 40, 289 - 303. Scitovsky, T. (1954), Two Concepts of Externalities. Journal of Political Economy 62, 143 - 151. Stiglitz, J. and A. Weiss (1981), Credit Rationing in Markets with Imperfect Information. American Economic Review 71, 393 - 410. Stokey, N. (1981), Rational Expectations and Durable Goods Pricing. Bell Journal of Economics 12 (1), 112 - 128. Vaubel, R. (1984), The Government's Money Monopoly: Externalities or Natural Monopoly? Kyklos 37, 27 - 58. — (1985), Competing Currencies: The Case for Free Entry. Zeitschrift für Wirtschaftsund Sozialwissenschaften 5, 547 - 564. OPEN ACCESS | Licensed under CC BY 4.0 | https://creativecommons.org/about/cclicenses/ DOI https://doi.org/10.3790/schm.105.5.565 | Generated on 2023-04-04 12:05:54