No. 8601 MONEY, DEREGULATION AND THE BUSINESS CYCLE

by Gerald P. O'Driscoll, Jr.* Research Department Federal Reserve Bank of Dallas

January 1986

Research Paper

Federal Reserve Bank of Dallas

This publication was digitized and made available by the Federal Reserve Bank of Dallas' Historical Library ([email protected]) t{o. 8601 }IONEY,DEREGULATIOI{ AI{DIHE BUSI}IESSCYCTE by

Gerald P. 0'Driscoll, Jr.* ResearchDepartnent Federal ReserveBank of Dallas

January 1986

* Paper pnepanedfor the cato Institute conferenceon lloney, Politics' and the Bulini:ssCycle, tlashington,D.C., January1986.

The views expressed'positionsar€ those of the author and do not necessarily reflect the of the Federal ReserveBank of Dallas or the Federal Reserve System. In this paper, I examjnea burgeon'ingliterature on the behavior of unregulatedbanking systems. Its analysis of bankingand moneyhas been dubbedthe legal restrictions theory. lt4anyof the theory's conclusionsare startling, as, for example,the proposition that it is unnecessaryto control the quantjty of depository liabilities in a competitivebanking system. Sjmilarly, the theoryrs modeof analysis is unconventional;for example,its benchmarkfor examiningthe nature of bankingservices is a nonmonetaryeconomy. It is precisely its unconventionalanalysis and startling conclusions,however, which makethe 1ega1restrictions theory both stimulating and worth further cons'ideration.

In what fol1ows, I fjrst explicate the newview on banking; I next consider implications of the newview fon controlling economjc fluctuations; I then present a critique and, finally, I suggesthow some of the vaiuable insights of the 1ega1restrjctions theory mjght be integrated with impol'tanttenets of moretraditional approachesto moneyand bank'ing.1

The Legal Restric!ions Theoryof Money

The legal restrictions theory examinesthe seemingparadox that individuals simultaneouslyhold governmentcurrency and governmentbonds

The currency is noninterest bearing, while governmentbonds bear interest.

The paradoxicalaspect of this behavionderives from the fact that both obligations are default-free iiabilitjes of the sameissuen. Assuming rational behavior by transactors, we would expect the interest-bearing securities to dominatecurrency. Accordingly,Wallace (1983, p,1) L

investjgates the features of interest-bearing governmentsecuritjes t,that prevent them from playing the samel"ole in transactions as Federal Reserve notes. For if they could play that ro1e, then it is hard to see \rhy anyone would hold non-jnterest-bearingcurrency instead of the interest-bearing securities.'r The newview identifies 1egal restrictions as the sourceof the sjmultaneousdemand for both currencyand bondsand contrasts the current environmentwith an unregulatedor rrlaissez-faire system.'r lr/allace

(1983, p. 4) states the view forceful ly and conciseiy.

...Laissez-fajre meansthe absenceof 1ega1 restrictions that tend, amongother things, to enhancethe demandfor a governmentrscurrency, Thus, the impositjon of laissez-faire would almost certainly reducethe demandfor government currency. It could even reduce it to zero. A zero demandfor a governmentrscur"rency should be interpreted as the abandonmentof one monetary unit in favor of another -- for example.the abandonmentof lhe dollar in favor of one ounceof go1d. Thus, my prediction of the effects of imposinglaissez-fajre takes the form of an either/or statement:either nominalinterest rates go Lo zero or existing governmentcurrency becomes worthI ess.

Wallace (1983, p. 1) identifies two conditions, the presenceof one of which is necessaryin order that governmentbonds not be substitutable fo" .u""ency.2 Either the bondmust be nonnegotjable(as is tnue of U.S. savingsbonds) or not issued in small denomjnatjons(as is true of Treasurybi 11 s) . As WalI ace ( 1983, pp. 2-3) further observes, neither of these two restrjctions by themse'lvescould prevent arbitrage by 3 fjnancial intermediarjes. Theseintermediaries could purchaselarge denomination,negotiable bonds (i.e., Treasurybilis in multiples of

$10,000)and jssue bearer notes in smalI denominations. By matching maturities of these notes and those of the Treasurybills, the intermediarywould be perfectly hedged. Since its assets are default-free by assumption,its bearer notes would also be default-free (fraud aside).

Wallacethus identifies a crucial leqal restriction that is sufficien! for the coexistenceof currencyand bonds: governmentis a monopolistjc provider of curre ncy.

Absent 1ega1restrictions, arbitrage would drive downthe yield djfferential betweenbonds and currency to the costs of intermediating betweenthem. l/lallace(1983, pp. 3-4) estimatesthat these might be less than one percent. As an approximation,one could ignore the difference.

Accordingly, Wallaceconcludes that either interest rates on bondsare driven to zero or currencydisappears.

Anotherway of statjng the conclusion is that moneywould not exist as a distinct financial asset.3 Thjs restatementbrings into sharper rel ief the clear connectionbetlreen Wal lace's statementof the legal restrictions theory and Fischer Black's earlier analysis of howan unregulatedfinancial systemwould operate.4 Black assunesthat depository institutions havecomplete freedom to create ljabjlities and to purchase financial assets as they see fit.5 Banks' incomederives from the spread betweentheir borrowjngcosts -- chiefiy interest on deposit liabilities -- and their revenue-- chief1y interest on loans. Black envis'ionsthat loans wi'lI take the form of negative bank balances,or, in other words, + overdrafts on deposit accounts. Indeed, h'is description of the hypothetical systemof positjve and negative bankbalances r eads like a virtual foretel ling of the moderncash managementaccount at brokerage houses(B1ack [1970], pp. 10-11).

B'lackpresents an evolutionarymodel of financial innovation, which beginswith a commoditymoney and ends in a money'lesswor1d. Early in the evolutionary process real goods, as well as the commoditymoney, becomepriced in lerms of an abstract unjt of account. Biack hypothesizes, however,that the meansof paymentwill likely be a portfol io of common stocks. He thereby invokes an assumptionthat characterizes subsequent presentatjonsof the newvjew: the separationof the meansof paymentand the unit of account.

Black (1970, p. 9) is also responsiblefor first articulating another characteristjc proposition of the 1ega1restrictions theory: in a deregulatedfinancial environment,I'it would not be possib1eto give any neasonabledefinition of the quantity of money. The paymentsmechanism in such a world would be very efficient, but moneyin the usual sensewould not exist.rr" In other words, havingmerged money and other financial assets, Black cannot readily quantify the formenseparately.

Wallace(1983, p.a) takes a different tack and anaiyzes open-marketpurchases and sales of TreasurybiIls by a central bank in a laissez faire regime. He assumesthat there is a constant-costtechnoiogy for producingcurrency, whjch is sharedby private and government intermediaries (a situation of rrtechnologicalsymmetry'r). In other words, g0vernmentand private notes are perfect subst'itutesproduced under 5 identical cost condjtions. In l;lallacers example,there i s a given private-sector demandfor currency. Thus, an expansionin the production of one type of curr"encyresults jn the contraction of other types. An open-marketpurchase of bills by the central banksconstitutes just such a change. As the central bank increasesits assets (Treasuryb'i lls), it will issue moreliabilities (including currency). Sjnce indjviduals nowhold central bank currency, they will curtai I their demandfor commercialbank currency. In the process, resourcesare reallocated from pl-ivate- to public-sector pnoducersof currency. Wa'llace (1983, pp. 4-5) concludes that:

. . . UnderLai ssez-fa ire and technologi cal symmetry, the openmarket purchase does no morethan change the location from the private sector to the governmentof a given quant'ity of economic activity, the production of small-denomination notes. Nothingelse is affected, neither interest rates nor the price Ievel nor the Ievel of economicactjvity. ., A similar argumentapplies to ' openmarket, sa1es.

In a laissez faire system,then, there are no macroeconomic effects of banks' issuing their ownI iabilities to purchasefinancial assets. Th'is conc'lusjon,which holds for central banksand private i ssuers al i ke, i s i n starl j ng contrast to conventional wisdom and constilutes the most importantpolicy conclusionof the 1ega1restrjctjons theory. The contention will be lhe focus of most of the rest of the

DaDer. 0nce again, Black suggestshow a laissez faire bankingsystem might operate. The world js a far cry from a monetarist environment containing a well-defined transactjon money,h,hose Lota'l quantity js linked to an exogenousmonetary base by a stable moneymultiplier. In

Black's wor'ld, debits and credits would be created and extinguishedwith every transactjon. In terms of Wallacersexample, under competitiVe conditions expansionby one intermediarywould comeat the expenseof contraction by others. Both analysesconclude that, in a lajssez faire system,the provision of paymentservices by bankswould have no special effects on prices or output (Cf. Fama[1980], pp. 45-47). Fconomjststraditionally modelbanks as creators of money.

Certain liabilities of private banksare addedto those of central banks with the resulting magnitudeconstituting the moneystock. The money-creationfunction is the benchmarkfor analyzing banks; of course, 'i n creating moneybanks are also providing the paymentssenvices on which legal restrictions theori sts concentrate. In the latter view, however, banksas creators of moneyare peculiar to a regimeof legal restrictions.

Consequently,conventional monetary theories are applicable only to a specific set of institutions. The 1ega1restrictions theory lays claim to being a moregeneral theory of financjal intermediation. Moreover,by abstracting from banks' role as creators of moneyin a regulated system,

1ega1restriction theorjsts feel that they better understandthe nature of bankingservjces. 0r, as Fama(1980, p.42) phrasesit, "the banking systemis best understoodwithout the mischief introducedby the concept of money." Legal restriction theorists focus instead on the accounting and portfol io managemenlservices provided by banks. 7

It js no\./ possible to restate the 1egal restrjctions theory as a set of five interrelated propositions.S

(1) Moneywould not exist as a distinctive financial asset in the absenceof legal restrictjons;

(2) The unit of accountis separablefrom the meansof payment;

(3) Conventionalmonetary theories are applicable only to a specific set of financial 'instjtutions;

(4) In a laissez faire system,the provision of paymentservices by bankswould have no special effects on prices or macroeconom i c actjvjty;

(5) The provision of paymentservices -- not the production of money-- is the benchmarkfor analyzingbanks.

In the next section, I focus on an implication of the 1ega1 restriction theory, namely,that a laissez fajre systemwould be insulated from economicfluctuations causedby monetaryshocks. 8

Economic Fluctuations

Somewriters have suggestedthat the problemof economic fl uctuations would be attenuatedif not elimjnated in an unregulated bankingsystem. Greenfield and Yeager(1983, p.304) contendthat such a systemrroffers muchless scopethan an ordinary monetarysystem for destructive monetarydisequilibrium.rt Theyalso suggestthat runs on rtwould banks be less catastrophjc under Ithis] system,r'essential ly becausebanks would exchangeliabilities under a floating rathen than a fi xed-rate domesticexchange system.

Fama(1980, p. 40) offers the most explicit underpinningfor the position that economicfluctualions result from regulatjons compelling banksto play a special role "in the processby which a pure nominal commodityor unit of accountis madeto play the role of numerairein a real world monetarysystem. "9 He arguesthat money'scausal inetficacy in a laissez fajre regimecan best be understoodas an implicatjon of the llodigliani-Miller theonem:"...The portfol io managementactivitjes are the type of pure financing decisions coveredby the Modigliani-Mjller (1958) 10 theorem."

The core of Fama'sargument is as follows. Fjrst, if there is competition, then there are actual or potential substitutes for the portfol ios offered by any banks. Second,to attract depositors, banks must hold portfol ios against which depositors are willjng to hold claims.

Third, competitioninsures that depositors are paid a return equaling that earnedon the bankrs portfol io less a managementfee. Gjven that they are J pure profit maximizers,the last assumptionrenders banksjndifferent to the compositionof their ownportfol ios (Fama[1980], pp. 45-46)

Whatdetermines the portfol io compositionof banks? Both the financing and the nature of economicactivity are determinedby "the tastes and endowmentsof individual economicunits and the state of the economy'stechnology (Fama [1980], p. 46). In this sense,then, banksare passive agents, v/hoseportfol ios are determinedby the nonfinancial sect0r.

If Famarsargument is correct, however,then banksare passive in another important sense:they exert no independentforce on prices or real activjty (Fama[1980], p. 45). The quantity and compositionof banks assets and liabilities are entirely demanddeterm'i ned. If one bankwere autonomouslyto changeits assets and liabilitjes, competit'ionwould insure offsetting changesby other bankingfirms (Fama119801 , p. 46). In the aggregate,banks would thus play no causal role in the determination of equilibrium price and quantity vectors. This conciusionis a neutrality finding writ large.

In Fama'sanalysis at least, a real goodfunctions as the numeraire. There is no price level as suchto be determined,but only an equiI ibrium price vector. ldhatwould be the question of price-1evel determinationreduces to the issue of the stabil ity of equilibrium in a barter, general equilibrium system(Fama [1980], p.44). Consequently, macroeconom'icphenomena constituted by or attendant uponprice-1eve1 fluctuations are absent by assumptionin the competitivebanking environmentposLulated by B1ack, Famaand Wallace. 10

In Fama(1980), the assumptionof a nonmonetaryeconony is a modelingstrategy to isolate the essential functions of a competitjve bank. By contrast, Greenfield and Yeager(1983) view the abolition of moneyas an essential feature of a reform (one hesitates to say "monetary reformrr)that they propose. In the process, however,they appearto have confusedan assumptionwith a substantiveproof.

Greenfield and Yeager(1983) rely on the analysis of monetary disequilibrium presentedin Yeager(1968). Moneyis unique in having no marketof its own. Accordingly, an excessdemand for moneymust be worked off in all other markets. St'icky prices result jn quantity responsesand pervasivereal effects of the initjal excessdemand for money(Cf.

Greenfield and Yeager[1983], p. 309). In their analysis, Greenfield and

Yeager(1983, p. 310) identify the inelasticity of the supply of moneyas the necessarycondition for macroeconomicdisequil ibrium to deve'lopout of an excessdemand foruon"y.11 The superiority of the proposedsystem, they assert, devolvesar ound the demanddeterminat'ion of the meansof payment .

Greenfjeld and Yeagerseem to haveconfused themselves, if not their readers, with thejr argumentabout the demanddeterminatjon of the meansof payment. Theypoint out that their rrsystemwould get rid of any distinct moneyexisting in a definite quantity.... A wrongquantity of moneycould no longer causeproblems because money would not existrl

(Greenfield and Yeager[19S3], p. 305).12 Simplyput, there is no monetarydisequilibrium in their systembecause there is not money! The argumentabout the demanddetermjnation of the meansof payment,which 11 appearsto be a substantjveproof, really reducesto a crude approxjmation of the kjnd of stability analysis suggestedby Fama. As will be seen, however,the Greenfield-Yeagersystem js still susceptible to economic disorders similar in effects to that of monetarydisequiljbrjum.

Idhat. then. of the substantive issue of the existence of economic fluctuations in unregulatedbank.ing system? As already mentioned,the problemfacing Greenfield and Yeageris not that of price-1eve1 determinationbut the attainmentof a general equilibrium prjce vector.

Actually, it would be jnstructive to focus on a morebasic question: can market prices be determinedin the Greenfield-Yeagersystem?

The latteli s the operative question because,as Greenfield and

Yeager(1983, p. 307) clearly state, they are proposinga barter system.

ltJithno moneyquantitatively exist'ing, people make paymentsby transferfling other property. To buy a bicycle priced at 100 value units on pay a debt of 100 units, one transfers property having that total vaiue. Althoughthe...system is banter in that sense, it is not crude barter. Peopleneed not haggle over the partic-ular goodsto be acceptedin eachtransaction. The profit motive wil l surely lead competingprivate firms to offer convenient methodsof payment.

First, it must be noted that there is no other sensein which the term barte|i s usedthan to cover situations in which goodstrade directly for goods.'" Second,I knowof no theory of 'rsophisticatedrr barteri

Greenfjeld and Yeager(1983) does not present a theory of sophisticated barter but dependson the (nonexistent) theory of howsuch a world operates. Onemust concludethat they are talking of barter, pure and s.imple. !{

It mjght well be appropriate to reconsiderthe standardana)ysis of barter. Absenta newtheory of barter, however,one must be pessimistic concerningthe workablenessof the Greenfield-Yeageruyra"r.14

The systemwould appearto suffer from the textbook problemsof barter.

AlthoughYeager and Greenfield (1983, p. 303) really only assert the contrary, the claim is worth analyz'ing. Theyadmit that the "systemwould indeed lack moneyas we knowjt,'r but they state that "it would not entail the textbook inconveniencesof barter. The advantagesof having a definite unit of accountand convenienLmethods of paymentwould be retained and enhanced." The implicit argumentis that it is capitalism's accountingsystem, not jts paymentssystem comprising a physical mediumof exchange,which overcomesthe calculational difficulties of barter.

A key elementjn the Greenfield-Yeagerproposal is the government'sdefining a unjt of va1ue,which would then form a basis of a social accountingsystem.l5 Ratherthan choosinga single good(as in

Famarsanalys'i s) or securities (as in Black's model), Greenfjeld and Yeager(1983, p. 305) suggesta compositebundle of commodities.l6

The orices of the individual commoditieswould not be fixed and would remainfree to vary in relation to one another. 0nly the bundleas a wholewou1d, by definition, havethe fixed price of 1 unit.... The bundlewould be composedof precisely gradable, competitjvely traded, and industrjal ly important commodities,and in amountsconresponding to their relatjve importance. Manywould be the materials used in the productionof a wide range of goodsso that the bundleas the value unit would comeclose to stabjlizing the general level of prices expre s sed in that unit. 1J

Greenfield and Yeager(1983, pp. 303, 306) emphasizethe differences bet\,reentheir proposaland those for a composi te-commodi ty or commodity-reservemonetary system. No reservesof the compositebundle would be maintainedby any agencyor private entity. There is no convertjbility but only a defined unit of value. The latter distinction is important to the authors as well as to the reader assessingtheir proposal.

There is a striking similarity betweenthe logic of the trading 'in process lhe Greenfiel d-Yeagerproposal and that in early Marxist schemes for allocating and distributing goods. It is instructiVe to draw the paralle1s, since doing so helps isolate a critical flaw in Greenfieldand

Yeager( 1983) .

Marxrsoverriding economicgoal was to replace capital ism's

"anarchicrrsystem of productionwith a systemof conscioussocial control of the meansof production (Lavoie t1985]). Marxwanted to avoid any reliance on marketpnices in allocating resourcesand djstribuling goods.

He suggestedusing labor tjme as a measureof the cost (value) of each commodityand actually exchanginggoods according to theiI embodjedlabor t'ime. CompareGreenfield and Yeager(1983, p. 307), whoobserve that "to buy a bicycle priced at. 100 value units or pay a debt of 100 un.its, one transfers property having that total value.rl

Using labor time as a mechan'ismfor allocaLing resouncesfounders on the problemof laborrs heterogeneityand nonuniformity. Marxtfljed to reduceheterogeneousr skilled labor to homogeneous,unskilled labor time.

He did not, however, solve the valuation problem. A competitivemarket 14 evaluatesdifferent types of labor but Marxwanted to eschewthe use of anarchic marketvalues. This ieft hjm with analytjcally insoluble problem of evaluating heterogeneouslabor without an evaluation mechanism(Lavoie

[1e8s] , pp. 67-74).

Greenfield and Yeagerface the evenmore complex problem of homogenizingthe heterogeneouscommodities of their compositenumeraire.

Greenfield and Yeager(1983, pp. 313-14)mention but do not solve the calculationa l problem.

Supposethat the ... bundlewere defined as 1 apple + 1 banana+ I cherry. Prjces are to be paid and debts settled in bundles-worthof convenient paymentproperty. Nowapples are struck by a fungus. Whatmarket forces arise to accompli sh the appropriatechanges in relative prices while still enforcing the unit's defjnitjon?r/

Greenfjeldand Yeagen(1983, pp. 313-14)are, as it were, hoist on their ownpetard. They themselvesnote that jf a fungusatlacks apples, the bundle becomesrelatively scarcer; deflatjonary pressureis exerted on other commodjtjes. This is the evjl from which their nonmonetaryexchange systemwas to save us. They suggestthat bananasand cherrjes are among the commoditjeswhose relative price will fall. The needfor an adjustment of the prjces of other commoditieswithjn the bundle adds to the adjustment problemrather than (partiaily) offsetting iL. In general, there will be morenot fewer price changesnecessary because there are two additional compositegoods whose prices havechanged, 15

In taking accountof the effects of the fungusattack, Greenfield and Yeager(1985, p. 314) suggestwidening the definitjon of the bundle. indeed, they indicate that the wider the defjnition, the better the results. Consider,however, what wouid occulif the suggestionwere carrjed to its logical extreme. Everytrade would constitute an exchange against a representativebundle of all commodjties. Using a conventional mediumof exchange("money," as we nowknow it) avoids having to calculate n-1 relative prices in makingindividual exchangesThe methodof paymentjn the Greenfield-Yeagersystem would require just this exercjse for each and every transactjon. Their systemwould accordingly involve the calcul ati onal chaosof barter.

To give somehi st or i ca I - i ns ! i t ut 'jo n a I rel evanceto the argument, the authors haveobserved that changesin the relative scarcity of gold under a gold standardproduces familiar macroeconomicconsequences. They suggestnot a bimetalIic but a trjmetallic systemas an improvement, ignoring the additional problemsintroduced by the possibiljly of re'lative-price changesbetween goods in the compositebundle.18

Actual1y, the analytica'l problembeing discussedis inherent in any schemeto stabilize a price level or other constructedaverage price.

The appealof stabilizing a price level or subsetof prices is that doing so wjll somehowminimize or djmjnish the numberof re'lative-price changes necessaryin a market economy(cf. Friedman[1969], p. 106). To my knowledgeno one has ever demonstratedthis rigorously; Greenfield and

Yeagercertainly do not do ro.19 They in fact havedone us the service of inadvertently showingwhy stabilizing a price or subsetof prices would not Lb necessarjly diminish the costly market adjustmentsnecessary in a monetary economy. Greenfield and Yeagerhave surely faijed, however,to demonstrate their main practjcal point, that economicfluctuations would be eliminated in a nonmonetarysystem. In a sense, this is gratifying, since it would be counter-intuitive to hold that a nonmonetarysysLem is moreefficient than a monetaryeconomy.

Whethereconomic fluctuations would occur in an economywith unregulatedbanks remains an openquestion. Resolutionof the question wou'ldrequire both a fuller developmentof the legal restrict'ions theory and careful specification of the sourcesof cyclical disturbances. l'lodels of the businesscycle increasingly identify real factors as the causeof fluctuations. If these modelsare correct, lhen it is unciear what effect monetaryderegulation would haveon the timing, amplitude on frequencyof cyclical fl uctuations.

Suppose,however, that economicfluctuations are causedby monetaryshocks. It would still be unclear whetherwe could be confident that an unrestricted bankingsystem would eljmjnate lhese fluctuations.

The uncertainty devolveson the issue of bank reservesand interbank deposits. The I iterature on the legai restrictjons theory has little to say about settlementpractices for banks(financial intermediaries) in a deregulatedenvironment. Yet the jssue is crucjal, since two bankscan only settle their liabilities by transferfling a third asset, which is the

I iabil ity of neither bank.20 To facj I itate settlement, banksmay hold interbank deposits. Moregenerally, howeven,banks wjll hold reservesof someasset acceptableto all as fjnal settlement. Today,base money t/

(deposits at FederalReserve banks pius currency) constjtutes the reserve asset. Evenabsent 1ega1restrictions, there would be a finite demandfor a reserve asset; agajn, the sourceof the demandwould derjve from the requirementsof the interbank c'learing p"0."rr.21 Indeed, these considerationslead 0sborne(1985b) to concludethat bankswould hold reserveseven in a laissez fajre paymentssystem. The optimal reserve ratio would be muchcloser to zero, however,than to one, which exposesthe systemto the periodic crises inherent in a fractional-neserve banking system. 0sborne(1985b, pp. 22-23) concludesas follows.

It is hard to imaginethat such a systemcould producemost of the uncertainties and absurdities lhat drjve observersof our present systemto despair... . But the speculationsdo not suggest that it would be free of monetarydisturbances. The bankersof a free systemwould choosetheir reserve ratios as profit dictates. The optimal reserve ratio would be less than one. There would be furtive abundance,and it would vanish with the gusts of djscredit that would blow amonga free peopleas amongothers, even if less often. 18

Barren lloney

In this section, I concentrateon the assumptjonsof the legal restrictions theory. Bryant and ldallace(1980, p. 1) provide the most explicit statementof the underlying assumptions.

(1) Assets are valued only in terms of their payoff di str i buti ons.

(2) Anticipated payoff dj stributions are the same as actual payoff di stributions.

(3) Underlaissez-faire, no transaction costs 'inhibit Lhe operation of marketsand, in particular, the law of one price.

Simply stated, the 1ega1restrictions theory assumesaway the exi stenceof any nonpecuniaryyield from holding noney.22 Since at least currencyyields no explicit return, this quickly leavesus with no reason for ratjonal economicagents to demandthe asset. Any neoclassical economistworth his salt shouldbe unsatisfied with this situation and quickly stnive to identify the intervent'iongenerating this otherwiseodd situation. In terms of their ownassumptions, Bryant and VJallace, et a1., havedone a goodjob of modelingthe problem. The assumptionsmust not go unchal'lenged,howeve r .

The denial of a nonpecuniaryyield to moneyis really anotherway of stating the old view that moneyis "barren." In an undeservedly neglectedessay, Hutt (1956) surveyedthe histony of monetaryeconomics and could find only one orthodoxmonetary theorj st (Greidanus)who was not, to one degreeor another, underthe swayof the doctrine that moneyis barren.

Thoughmany economists have had all the elementsof a correct theory -- 19 clearly perceiving that moneyprovides convenjencesservices and cost savings-- virtually ali continuedto assumeexplicitly that moneyrsyield is, in Keynesrwords, nil" (Keynes[1936], p. 226).

The view that moneyyields no return is as old as Arjstotle. Ir enteredmodern economics through the schoolmen,thence via Lockeand Adam

Smjth. Not surprisingly, Hutt traces the idea thnoughthe classical economists. l,rlhatis surprising, however,are the illustrious neoclassical economistswho haveechoed the point downto the present. ldhereasLocke said that "moneyis a banrenthing" (Hutt [1956], p. 199), Bohm-Bawerk rrMoney assuredus that: is by nature incapableof bearing fruit" (Hutt ,,sterilerr [1955], p. 203). \rlickselldescribed money as (Hutt [1956], p. 204).

Perhapsthe most puzzling of all is Keynes. I've already quoted him as denyingthat moneyhas a yie'ld. Thjs statementjs the more remarkable,since it appearsin the section of the GeneralTheory in which

Keynesanalyzes the liquidity premiumon money. If we take him 1itera11y, then economicagents exhibit a preferencefor an asset tr,ith no yield.23

The confusionis even clearer jn MarshalI than in Keynes.

Marshall explicitly recognizedthat somecapital assets yield an implicit or nonpecun'iaryreturn but denied that moneyis one of these assets. He averred that holding resourcesin the form of moneyrrlocks up in a barrer, form resourcesthat mjght yieid an incomeof gratification if invested, say, in extra furniture; or a moneyincome jf invested in extra machinery or cattlerr (Hutt [1956], pp. 205-06). 20

MarshalI was quite modern-- more so even lhan Keynes-- in noting that the yield on an asset can be either nonpecuniaryor pecuniary. He simply denied that moneyhas a yield of either kind. I submjt that modern treatmentsof the demandfor moneymake essentially the samemistake. The modernI iteralure is quite clear in treating foregoneinterest as the cost of holding money,bul is moreambiguous by far on the benefits derived from cash holdings. FoliowingBaumol, one traditjon focuseson brokeragecosts of movingjn and out of interest-bearing assets. This explanationrings hollow as we return to a financial systemwith sophisticatedfinancial instrumentsand cash managementtechniques. FollowingTobin, a second ljterature focuseson f iquidity preferenceas behavior tov{ardrjsk. The latter tradition perhapsadheres more closely to Keynes,but, in so doing, perpetuateshis error on the yield from holding money.

Hutt contendsthat modernmonetary theory perpetuatesan 18th century view of productivity. The 18th century vjew treats productivity in entjrely physical terms: an asset is productive if it yields a return in kind, i.e., if it bears frujt. If it yie'lds no fruit, the asset is barren.

Since moneytraditional1y yielded no interest, 18th century economists viewed it as barren. Moderncapital lheory has movedbeyond that view by acceptingthat assets can yield an implicit return. This insight explains, for example,the holding of so-called idle land.24

Whenit comestorridle balances,rrhowever, the 18th century vjev{ hoids sway. As suggestedabove, the neoclassical spirit is restive when confrontedw'ith a demandfor an asset apparently having no yield. The restive spirit has yielded the 1egal restrictions theory. Indeed, so long L! as economistsadhere to the 18th century Viel/{on money,the legal restrictions theory maybe the only consistent resolutjon of the conundrum.

Moneyyields a nonpecuniaryreturn, just as does furniture, a pajnting or wine coilection. In deciding whether-to hold moreor less money,an indivjdual comparesthe advantagesof holding the moneybalances with the advantagesof holding othen assets. In doing this, the indjvidual is comparingdifferent expectedyields; he is not companing an asset yielding a return with oneyielding no return. The latter wou1d,indeed, be a paradoxical sjtuation.

0nce we accept that moneyyields a nonpecuniaryreturn, the paradoxidentified by the 1ega1restrictions theory is seento be apparent rather than real . In other words, the paradoxis resolved by denyingthe thesis. Along the way, we also manageto jettjson a gooddeal of phiiosophical baggagethat \^/ecan do well without.25

WhatI am identifying js a property of moneythat is the property neithen of legal restrictions nor of hi storical accident, but which reflects a preferenceexhibited by jndjvjduals over tjme and in radically different trading environments. The property or characteristic is money's liquidity.

Hicks (1974, pp. 38-39) has succinctly characterizedthe demand for liquidity as a desire for flexjbility: rrLiquidity is not a property of a single choice; it is a matter of a sequenceof choices, a related sequence. It is concernedwjth the passagefrom the knownto the unknown

-- with the knowledgethat if we wait we can havemore knowledge.rr In contrast, Hicks (1974, pp. 43-44) pojnts out that rrbyholding the 22 imperfectly liquid asset the holder has narrowedthe trend of opportunities which maybe opento hjm.... He has rlocked hjmself jn.r" In so characterizing the value of liquidity to t,ransactors,Hicks clearly links 'l the demandfor money(and other iquid assets) to uncertainty. In this sense,money can only be analyzedwith a theory incorporating uncertainty.26

Moneyis not merely highly liquid, but that asset which is perfectly liquid. It trades in every marketand neednever be sold at a ,7 discount.'' Evenhighly liquid, nonmonetaryassets are subject to price ri sk. For this reason! people are w'illing to forego substantial pecuniary returns in order to hold moneybaiances. In highly regulated and substantial 1y unregulated monetarysystems al i ke, i ndividual s havedemanded absolutely I i quid assets.

The aboveaddresses the demandfor liquidity. The 1ega1 restrictions theorists maybe interpreted as emphasizinga supply issue: why cannot intermediariespurchase interest-bearing assets and issue circulating notes ("currencyrr)backed by these assets? It is certainly true that the willingness of people to forego a pecunianyneturn does not imply that they needdo so. As Bryant and lrjallace(1980, p. 11) insist, we must investigate the rrtransactiontechnology" in a moderneconomy.

Bryant and lllallace(1980, pp. 14-15) and V/allace(1983, p. 3) estimate the costs of intermediatingby observingthe spreadbetween the rates of return earnedand paid by mutua)funds. |.iallace (1983, pp. 3-4) asserts that rr there js no reasonto expect that the cost of intermediating securities like Treasurybills into bearer notes would be muchdifferent 23 from the cost of operating these i ntermedjarj es,'r 0bservation suggests, however,that there is goodreason to supposea great deal of difference betweenthe costs of suppling low-turnoverdeposits (moneymarket mutual fund shares) and high-turnovercurrency. ldhite (1985) examinesthe t r an s a c t i on - co s t structure and concludes that the'i ntermediat ion costs for currencyare of an entirely different or-derof magnitudethan for deposits.

He offers three types of evidence: (1) historical evidenceon currency issued in the Scottish free-bankingsystem; (2) evidencecurrent practice with respect to travelers checks; and (3) an a priori estimate.

With respect to the fjrst type of evidence,White (1985, pp. 3-4) observesthat rrthe legal restrictions theory provides us wjth a clear anc falsifiable prediction: non-interest yielding currency shouldnot coexist w'ith positive-interest-yjelding securities in the absenceof 1ega1 restrjctions against the sort of intermediation that could produce interest-yielding bearer bondsbacked by those samesecurjties.r' In the free-bankingera (i.e., before 1844)Scottish banks had completefreedom to pay interest on bank notes and the bankingenvironment was competitjve,

Yet noninterest-bearingcurrency flourished, falsifying the predjctjon of the 1ega1restrj cti ons theory.

Second,h/hite (1985, p. 4) notes the nonpaymentof interest on travelers checkstoday. Moreover,it would surely be computationally easier to pay interest on travelers checksthan on currency. There appear to be no restrictions on paying interest on travelers checks.

White's third piece of evidenceis perhapsthe most jnteresting.

He adducesarguments v{hy interest-bearing currencywould inherently be rnore 24 costly to transact with than noninteresting-beaflingcurrency. He then makesa reasonablecalcuiation of the costs of coilecting the jnterest accruedon a note and concludesthat jt would be prohibitive (|'/hite 11985], pp. 7-10)

Both theoretical argumentsand obsenvationalevidence suggest that there was never a paradoxto explain. It is certainly tnue that the existing financial systemis replete with reguiations. Someof these negulationswould even serve to restrajn an issuer from circulating 'interest-bearing currency if he wantedto do so. The evidenceindicates, however,that the restraints are irrelevant. Interest-bearing currency would not plausibly evolve wjth reasonableassumptions made about costs and benefits. It has not existed whenbanks were free to issue it. it will probably not exist whenbanks are free to issue it aqajn in the future. Conclus ion

In the previous section, White's analys'is addressedthe supply-sideor cost considerationsadduced by Bryant and Wallace. At least for argument'ssake, the analysis acceptsthe plausibility of currency's yielding jnterest. At minimum,however, the interest earnedon moneymust alwaysbe less than that earnedon nonmoneyassets. For if moneywere to yield both a nonpecuniaryneturn of liquidity services and an explicit market rate of interest, then the return on holding moneywould be supra-normal. Osborne(1984 and 1985a)argued that basemoney alone correspondsto t.hemoney of economictheory. It wou'ldbe plausible to supposethen that currencywould be the most liquid transactions money.

Its lack of an explicjt yie.ld scarcely seemstroublesome in that light.

Onecan, of course, deny (as Bryant and Wa1lace [1980] did) that there 'is a distinctive asset called money. In their case, the denial rea11y is an impljcation of a methodologicalargument about the form that economicreasoning ought take. it ciear'ly is beyondthe scopeof thjs paper to deal directly with that debate.28 It would be unfortunate, however,if the debateover bankingderegulation became entangled in a modernmethodenstreit. Moreconcretely, commitmentto (or against) banking deregulationdoes not presumecommitment to the equiI ibrium theorizing advocatedby the 1ega1restrictions theorists. Indeed,historically unregulatedbanking has born Iittle resemblanceto the hypothetical

"laissez-faire" systemspostulated in various models derived from the legal restrjct'ions theory. In that sense,the theory is a detour jn the debate over bankinq derequlati on. 26

Froma different perspective, however,the legal restrictions theory has done a great service by challengingeconomists to rethjnk their commitmentto monetaryregulation. 0n their ownterms, conventional macroeconomicmodel s makeno sense. lriallace (1983, p. 6) correctly identifies that, on convenLionalgrounds, the one remainingjustification for 1ega1restrictions on moneyjs revenuecoilection. If economists pursuethe suggestionof modelinglegal restrictions on moneyas a species of fiscal policy, then the legal restrictions theorists will havemade a

I asti ng contrj buti on. 27 NOTES

*SeniorEconomjst and Policy Advisor,Federal Reserve Bank of Dallas. The viewsexpressed in thjs paperare the authorrsand shouldnot be construed as representingt.he official positjon of any part of the FederalReserve System. I wouldIjke to thankDale Osborne for hjs commentson an earlier draft.

1. Although I r"efer to the 1ega1restrictions theory as the rrnewview,r' Cowenand Krosner(1985) argue that theory is anything but new. They contendthat it has a long histony, which begins in the eighteenth centuny.

2. A "bondrrrefers to a dated i nterest-bearing ob1i gati on, whi1e "currency" refers to a noninterest-bearjngnote callable on demand.

3. 0r, as Hall (1982b,p. 1554)puts it, rrmoneyis exactly a creature of regu'lation.rr Thefollowing discussion of Fischer Black's views draws from0rDriscoll (1985a,pp. 6-7).

4. ldallace (1983, pp. 1n) refers the reader to Fama(1980) and Hall (1982) I'for other discussionsof the 1ega1restrictions theory." He also cites six other articles as app'lications of the theory, but does not refer to Black (Wa11ace p. 3). Black is clearly the 'i [1983], ntellectual predecessor,however, of Fama,Ha11, biallace, et a1. q Black does not adherestrictly to a laissez faire assumption. For instance, he specifies that 'reverybank wjl'l be required to have capjtal equal to a certain fraction of its loans..." (Black 11970],p. 12). His support of capita1 requjrements i s parti cularly odd, gi ven his opposition to reserve requirements.

6. Black later adds that rrneitherthe quantity theory of moneynor the liquidity preferencetheory of moneywould be applicable.rr And he further states that: rrTraditionalmonetary theoflies will be inapplicable; in fact, it will not be possible to define the quantjty of moneyi n meaningful terms ( B1ack[ 1970], pp. 9-10). 7. In a footnote to thjs passage,lliallace adds that: I'Theresult that central bank intermediationdoes not matter under laissez-faire aiso holds for central bank exchangeof Federal Reservenotes for other assets -- risky mortgages,risky commercialloans, or commonstock. It is a straightforward extensionof a well-known finding in corporate finance called the Modigliani-Miller theorem." 0n the latter point, see Fama( 1980, pp. 45-47).

8. Cowenand Krosner(1985, pp. 2-3) adduce7 propositions characterizing the theory. 28

9. Yeagerand Greenfield (1983) offer their ownanalysis of the problem, which I examinebelow.

10. Fama(1980, p. a0). Fama(i980, pp. 45-47)offers rwo varjanrsof the theorem. Woodand hiood(1985, pp. 477-82)offer a textbook presentation of the theorem.

11. 0n the crucial role of the supply elasticjty of money,Cf. Keynes(1964 [1e36], pp. 234-36).

12. 0r see Yeagerand Greenfield (1980, p. 303), where they state that their systemrrwould indeed lack moneyas we knowit..."

13. SeeClower (1970), pp. 202-277. Clower(pp. 207-08) srares rhe foliowing as "the central themeof the theory of a moneyeconomy": "Moneybuys goodsand goodsbuy money;but qoodsdo not buy goods.,' By contrast, in Greenfield and Yeager(1983), goodsbuy goods.

14. OrDriscoll (1983) offers a moredetailed critique of an earlier presentationof the 1ega1restfictions theory by Yeager(1983).

15. Governmentplays an ironic role in manyof the I'laj ssez-fairerrmodel s of the paymentsmechanism. In Greenfield and Yeager(i983), government 4efines the unit of value. In llallace (1983), governmentjmposes la'issezfaire. In Hall (1982a), governmentreplaces the existing rnonetarvstandard by fiai and in jnterest-rate targetingi- The use of "laissez-fairefi in this "ngigesciais of models aDDearsto be a neo'logism.

16. Greenfield and Yeager(1983, p. 305) cite RobertHallrs suggestionof a bundle of 50 ki lognamsof ammoniumnitrate plus 40 kilogramsof copper plus 35 ki'iogramsof aluminumplus 80 squaremeters of plywood(of specified grade), but indicate a preferencefor an evenmore encompassing composi te bundle,

17. The authors invjte misunderstandingby such phrasesas t'enforcingthe unitrs definition.rr Greenfield and Yeager(1983, p. 303) haveassured us that the I'unit of accountdoes not require I i mpI ementati on I through convertibility of any famjl iar sort, anymorethan does maintenanceof the defined length of the meter.rr What, then, is to be enforced?

18. Greenfield and Yeagerhave, of course, designatedtheir systemas nonmonetary.I am not arguing, therefore, that it would be similar jn all respectsto a trimetallic sysLem,but amonly suggest'ing that it would involve the theoretical and practical prob'lemsdi scussedhere.

19. Uallace ( 1983, p. 6) observesthat:

...Ther"eexist no completeargument,s leading to the conclusionthat people are on averagebeLter off the morestable the price 1eve1,qiven the steps 29

that have to be taken to attain greater stability of the price level . 0n the contrary, as Sargent and |rlallace (1982) argue, the restrictions that makegreater price 1eve1stabiljty possible hurt somepeople and benefit others, while on average, 'in a certain sense,making all worseoff."

He concludesthat, without 1ega1r"estrictions, rrit is no easier to achieveprice 1eve1stabil ity than it is to achieve stability of some rel ati ve pri ce."

20. 0rDriscoll (1985, pp. 7-9) examinesthis issue jn moredetaii; cf. 0sborne( 1985, pp. 18-23).

21. Recenthistoriography on the clearinghousefunction in a free-banking systemincludes Gorton(1985) and Timberlake(1984). White (1984a, pp. 1-22) presentsa modelof free banking in which banks demandreserves.

22. Cf. Whit.e(1985, p. 5). The first assumptjonexplicitly precludesa nonpecuniaryyield on money. But the secondand third assumptions separateiyexclude the possibility, since the eljminate the reasonfor moneyrsyield. (See the textual discussion, infra.)

23. Keynes'point was precisely that moneyyjelds a nonpecuniaryyield. That he felt compelled to say that moneyrsyield is "df" indicates, however,that the old view of barren moneystill held swayover him even as he was engagedin trying to overturn jt. As Keynessajd in the Prefaceto the GeneralTheory, rrthedifficulty 1ies, not in the new 'in ideas, but esc'apingtrom tfre o1d ones, which, ramify, for those brought up as most of us have been, into every corner of our minds."

24. Andit can serve to explain the holding of idle resourcesgenerally. For an insightful analysis along these lines , see Hutt (1939). 25.One also avoids having to adopt the troublesomemodel ing strategy adoptedin Bryant and ltJallace(1980). Bryant and Wallace(1980, p. 6) defendthe strategy by arguing that "the reader js not giving up much by entertaining fthe three] postulates as a potential basjs for a theory of financial systems. By not gjving up much,we meanthat exist'ing alternative models of financial systemshave taught us very ljttle.rr I am inclined to agree that we would not be giving up muchby jettison'ing the macroeconomicmodel s examinedby Bryant and lrial lace (1980, pp. 6-10). I try to indicate this on p. 21, supra. 0'Driscoll (1985b)discusses the origins of the tradition presentedhere. A1so, see OrDriscoll and Rizzo (1985, pp. 191-98).

26. The latter point is scarcely original . If accepted,however, it precludesthe strategy adoptedby Bryant and lriallace(1980). 0'Driscoll and Rizzo (1985) argue that uncertainty is the sourceof manyeconomic processes and jnstitutions, which can be analyzedonly by 30 'i ncorporatinguncertajnty. Moneyis, in fact, one of the best examples of a market institution that would not exjst in a world with perfect foresight and no transaction costs. At this levei of genenality, Bryant and l{allace (1980) had their chief result as soonas they wrote downthei r assumptions. The analysis of liquidity drawson 0'Driscoll (1985a,p. 11).

27. This characterizationtakes not namesbut properties seriously (see Bryant and l.Jallace[1980], pp. 8-9). Choosingthe empirical counterpart of the Lheoretical construct is not an easy task, as Osborne(1984 and 1985a)demonstrates.

28. The issue is taken up in great detaiI in 0rDriscoll and Rizzo (1985). 3i

BIBLiOGRAPHY

Black, Fischer. rrBankingand Interest Ratesin a Worldhlithout lloney." Journal of BankResearch 1 (Autumn1970): 9-20.

Bryant, John and Wa11ace,Nei 1 . I'TheInefficiency of Interest-bearjng Nationa'lDebt.rr Journal of Political Economy87 (Apri1 1979): 365-81.

Bryant, John and Wallace, Neil. "A Suggestionfor Further Simplifying the Theoryof Money." l'4inneapoljs: Photocopy,1980.

Clower, RobertW., ed. l4onetaryTheory: SelectedReadings. Baltimore: Penguin, 1969.

Cowen,Tyler and Krosner, Randy. "The Developmentof the Legal Restrjctjons Theoryof Money." (Cambridge,l'4A: Photocopy,1985). Fama,Eugene. rrBankingin the Theoryof Finance.'r Journal of Monetary Economics6 (January1980): 39-57.

Friedman,Mjlton. The OptimumQuantity of Money. Chicago: Aldine, 1969. Gorton, Gary. '' C 1 e a r i ng h o u s e s and the 0ri gi n of Central Bankingin the United States.rr Journal of EconomicHistory 45 (June 1985): 277'83.

Greenfield, Robert L. and Yeager, LeiandB. 'rA Laissez-FaireApproach to MonetaryStability." Journal of Money,Credjt and Bankjng15 (August 1983): 302-15. rrExplorations HalI , Robert E. jn the Gold Standardand RelatedPol icies for Stabilizing the Dollar." In Inflation: Causesand Effects, pp. 777-22.Chicago University of ChicagoPressJ982a.

"MonetaryTrends in the United States and the United Kingdom: A Revi ew from the Perspectiveof NewDevelopments in MonetaryEconomics." Journal of EconomicLiterature 20 (December1982b): 1552-56.

Hicks, J.R. The Crisis in KeynesianEconomjcs. New York: Basic Books, 1974.

Hutt, ld.H. The Theoryof Idle Resources.(1939).

"The Yield from MoneyHeld.rr In 0n Freedomand Free 796-2?3. Edited by MarySennholz. Princelon: D. Van Nostrand,1956.

Keynes,John l,laynard. The GeneralTheory of Employment,Interest, and Money. NewYork: Harcourt, Brace& |,'lo11d,Harbinger Books, 1964 ( 1e36). 32

Klei n, Benjamin. "The Competiti ve Suppiyof Floney." Journal of Money, Credjt, and Banking6 (November1974):

Lavoie, Don. Rivalry and Central Planning: The Socialist Calculation DebateReconsidered. Cambridge, U.K.: CambridgeUniversity Press, 1385:-

0'Driscoll, Gerald P. "A Free-MarketMoney: Commenton Yeager.rr Cato Journal 3 (Spring 1983): 327-33.

"Moneyin a DeregulatedFinancial System.rr Federal Reserve Bankof Oallas EconomicReview (May 1985a): I-72.

"l'loney: Menger'sEvolutionary Theory." Federal ReserveBank of Dallas ResearchPaper No. 8508 (1985b); forthcoming,History of Political Economy(1986).

0sbonne,Dale K. rrTenApproaches to the Definition of Money.rrFederal ReserveBank of Dallas EconomicReview (l4arch 1984): 1-23.

"hlhatis MoneyToday?" Federal ReserveBank of Dallas EconomicReview (January 1985a): 1-15.

--- . "0n the Theoryof Lajssez-Fai re PaymentsSystems.rr Dal I as: PhotoEopy,1985b.

Sargent, ThomasJ. and Wall ace, Neil . "The Real-bj l l s Doctr"ine versus the Quantity Theory: A Reconsideration.r'Journal of Political Economy90 ( December1982): 7272-36.

Timbenlake,Jr., RichardH. "The Central BankingRole of Clearinghouse Associations.r' Journal of Money,Credit and Banking16 (February 1984): 1-15. rrlYoney'r ila1lace, Ne'i1. "A Legal Restrictions Theoryof the Demandfor and the Role of MonetaryPolicy.rr Federal ReserveBank of Minneapolis Quarte r]y Revi ew ( Win te r 1983) : l-7 . l{hite, LawrenceH. Free Bankingjn Britain: Theory, Experience,and Debate,1800-1845. Cambridge: CambridgeUniversity Press, 1984a.

rrCompetitivePayments Systems and the Unit of Account. American Economic Revi ew 74 ( September1984b) : 699-712.

'rAccountingfor Non-'interest-bearingCurrency: A Critique of the 'Lega1Restrictionsr Theoryof Money.rrNew York: Photocopy,1985. l{ood,John H. and lrlood,Norma L. Fjnancjal lvlarkets. NewYork: Harcourt BraceJovanovi ch, 1985. J5

Yeager,Lel and B. rrEssentialProperties of the Medium of Exchange." Kyklos 21 ( 1968): 4s-69. Reprinted i n Clower ( 1969), pp. 37-60.