The Role of Money and Monetary Policy · cerning the role of money and monetary policy. ... “The new approach argues ... tions originally published by Gurley-Shaw and

4

Karl Brunner

The following article is reprinted from the July 1968issue of the Federal Reserve Bank of St Louis Review.

The Role of Money andMonetary Policy

HE DEVELOPMENT of monetary analysis inthe past decade has intensified the debate con-cerning the role of money and monetary policy.Extensive research fostered critical examinationsof the Federal Reserve’s traditional descriptionsof policy and of the arrangements governingpolicymaking. Some academic economists andothers attribute the cyclical fluctuations ofmonetary growth and the persistent problemconcerning the proper interpretation ofmonetary policy to the established proceduresof monetary policy and the conceptions tradi-tionally guiding policymakers.

The critique of established policy procedures,which evolved from this research into questionsconcerning the monetary mechanism, is de-rived from a body of monetary theory referredto in this paper as the Monetarist position.Three major conclusions have emerged fromthe hypotheses put forth. First, monetary im-pulses are a major factor accounting for varia-tions in output, employment and prices. Second,movements in the money stock are the mostreliable measure of the thrust of monetary im-pulses. Third, the behavior of the monetaryauthorities dominates movements in the moneystock over business cycles.

A response to the criticisms of existing mone-tary policy methods was naturally to be ex-pected and is welcomed. Four articles which de-fend present policy procedures have appearedduring the past few years in various FederalReserve publications.1 ‘I’hese articles comprise acountercritique which argues that monetary im-pulses are neither properly measured nor ac-tually transmitted by the money stock. Theauthors reject the Monetarist thesis that mone-tary impulses are a chief factor determiningvariations in economic activity, and they con-tend that cyclical fluctuations of monetarygrowth cannot be attributed to the behavior ofthe Federal Reserve authorities. These fluctua-tion are claimed to result primarily from thebehavior of commercial banks and the public.

‘I’he ideas and arguments put forth in thesearticles deserve close attention. The controversydefined by the critique of policy in professionalstudies and the countercritique appearing inFederal Reserve publications bears on issues offundamental importance to public policy. Under-lying all the fashionable words and phrases isthe fundamental question: What is the role ofmonetary policy and what are the requirementsof rational policymaking?

‘Lyle Gramley and Samuel Chase, “Time Deposits inMonetary Analysis,” Federal Reserve Bulletin, October1965. John H. Kareken, ‘Commercial Banks and theSupply of Money: A Market Determined Demand DepositRate,” Federal Reserve Bulletin, October 1967. J. A. Cacy.

Alternative Approaches to the Analysis of the FinancialStructure,” Monthly Review, Federal Reserve Bank of Kan-sas City, March 1968. Richard 0. Davis, “The Role of theMoney Supply in Business Cycles,” Monthly Review,Federal Reserve Bank of New York, April 1968.

FEDERAL RESERVE BANK OF ST. LOWS

5

The following sections discuss the major as-pects of the countercritique. These rejoindersmay contribute to a better understanding of theissues, and the resulting clarification may re-move some unnecessary disputes- Even thoughthe central contentions of the controversy willremain, the continuous articulation of opposingpoints of view plays a vital role in the searchfor greater understanding of the monetaryprocess.

A SUMMARY OF THECOUNTERCRITIO,UE

The four articles relied on two radically dif-ferent groups of arguments- Gramley-Chase,Kareken and Cacy exploit the juxtaposition“New View versus Traditional View” as the cen-tral idea guiding their countercritique. The ana-lytical framework developed by the critique isnaturally subsumed for this purpose under the“Traditional View” label. On the other hand,Davis uses the analytical framework developedby the critique in order to organize hisarguments.

Gramley-Chase describe their general argu-ment in the following words:

“(New) developments have reaffirmed the bankers’point of view that deposits are attracted, notcreated, as textbooks suggest. In this new environ-ment, growth rates of deposits have become moresuspect than ever as indicators of the conduct ofmonetary policy A framework of analysis [isrequired] from which the significance of timedeposits and of changing time deposits can bededuced. Traditional methods of monetary analysisare not well suited to this task. The ‘New View’ inmonetary economics provides a nlote usefulanalytical framework. In the new view, banks—like other financial institutions—are considered assuppliers of financial claims for the public to hold,and the public is given a significant role in deter-mining the total amount of bank liabilities. - -

Traditional analysis. - - fails to recognize thatsubstitution between time deposits and securitiesmay be an important source of pro-cyclical varia-tions in the stock of money even in the face ofcountercyclical central bank policy.”

This general argument guided the construction

of an explicit model designed to emphasize the

role of the public’s and the banks’ behavior inthe determination of the money stock, bankcredit and interest rates.

Kareken’s paper supplements the Gramley-Chase arguments. He finds “the received moneysupply theory” quite inadequate. His paper isdesigned to improve monetary analysis by con-structing a theory of an individual bank as afirm. This theory is offered as an explanation ofa bank’s desired balance sheet position. It alsoappears to form the basis of a model describingthe interaction of the public’s and the banks’behavior in the joint determination of themoney stock, bank credit and interest rates.The whole development emphasizes somewhatsuggestively the importance of the public’s andbanks’ behavior in explanations of monetarygrowth. It is also designed to undermine theempirical hypotheses advanced by the Monetar-ist position. This is achieved by means of ex-plicit references to specific and “obviouslydesirable” features of the model presented.

Cacy’s article develops neither an explicitframework nor a direct critique of the basicpropositions advanced by the Monetarist thesis.However, he provides a useful summary of thegeneral position of the countercritique. TheMonetarist analysis is conveniently subsumed byCacy under a “Traditional View” which is jux-taposed to a “New View” of monetary mecha-nisms: “The new approach argues... that thereis no essential difference between the mannerin which the liabilities of banks and nonbankfinancial institutions are determined. Both typesof institutions are subject in the same way tothe portfolio decisions of the public.” The newapproach is contrasted with the TraditionalView, which “obscures the important roleplayed by the public and overstates the roleplayed by the central bank in the determinationof the volume of money balances.”~The generalcomparison developed by Cacy suggests quiteclearly to the reader that the Traditional Viewallegedly espoused by the Monetarist positioncannot match the “realistic sense” of the NewView advocated by the countercritique.

In the context of the framework developed bythe critique, Davis questions some basic proposi-tions of the Monetarist position:

2Gramley-Chase, pp. 1380, 1381, 1393.

‘Cacy, pp. 5 & 7.

~lbid.,p- 7.

SEPTEMBER/OCTOBER 1989

6

“In the past five to ten years, however, therehas come into increasing prominence a group ofeconomists who would like to go considerablybeyond the simple assertion that the behavior ofmoney is a significant factor influencing thebehavior of the economy. - - - In order to bring afew of the issues into sharper focus, this articlewill take a look at some evidence for the ‘moneysupply’ view. -

It confines itself to examining the historical rela-tionship between monetary cycles and cycles ingeneral business. The article concludes that therelationship between these two kinds of cyclesdoes not, in fact, provide any real support for theview that the behavior of money is the predomi-nant determinant of fluctuations in business activi-ty. Moreover, the historical relationship betweencycles in money and in business cannot be used todemonstrate that monetary policy is, in its effects,so long delayed and so uncertain as to be an un-satisfactory countercyclical weapon.”

AN EXAMINATION OF THE ISSUES

A careful survey of the countercritique yield-ed the following results. The Gramley-Chase,Kareken, and Cacy papers parade the New Viewin order to question the status of empiricaltheories used by the Monetarist critique in itsexamination of monetary policy. The Davispaper questions quite directly, on the otherhand, the existence and relevance of the evi-dence in support of the Monetarist position, andconstitutes a direct assault on the Monetaristcritique. The others constitute an indirectassault which attempts to devalue the critique’sanalysis, and thus to destroy its central proposi-tions concerning the role of money and mone-tary policy.

The indirect assault on the Monetarist positionby Gramley-Chase, Kareken and Cacy requires aclarification concerning the nature of the NewView. A program of analysis must be clearlydistinguished from a research strategy and anarray of specific conjectures.6 All three aspectsare usually mixed together in a general descrip-tion. It is important to understand, however,that neither research strategy nor specific em-

pirical conjectures are logical implications of thegeneral program. The explicit separation of thethree aspects is crucial for a proper assessmentof the New View.

Section A examines some general character-istics of the countercritique’s reliance on theNew View. It shows the New View to consist ofa program acceptable to all economists, a re-search strategy rejected by the Monetarist posi-tion, and an array of specific conjectures ad-vanced without analytical or empirical substanti-ation. Also, not a single paper of the counter-critique developed a relevant assessment of theMonetarist’s empirical theories or central prop-ositions.

In sections B and C detailed examinations ofspecific conjectures centered on rival explana-tions of cyclical fluctuations of monetarygrowth are presented. The direct assault on theMonetarist position by Davis is discussed insome detail in section D. This section also statesthe crucial propositions of the Monetarist thesisin order to clarify some aspects of this position.This reformulation reveals that the reservationsassembled by Davis are quite innocuous. Theyprovide no analytical or empirical case againstthe Monetarist thesis. Conjectures associatedwith the interpretation of monetary policy (the“indicator problem”) are presented in section E.

A. The New View

The countercritique has apparently beendecisively influenced by programmatic elabora-tions originally published by Gurley-Shaw andJames Tobin.’ The program is most faithfullyreproduced by Cacy, and it also shaped thearguments guiding the model construction byKareken and Gramley-Chase. The New View, asa program, is a sensible response to a highlyunsatisfactory state of monetary analysis in-herited in the late 1950’s. A money and bankingsyndrome perpetuated by textbooks obstructedthe application of economic analysis to thefinancial sector. At most, this inherited litera-ture contained only suggestive pieces of an-alysis. It lacked a meaningful theory capable of

5Davis, pp. 63-64.6These three aspects of the New View will subsequently beelaborated more fully. Their program of analysis refers tothe application of relative price theory to analysis of finan-cial markets and financial institutions. Their researchstrategy refers to a decision to initiate analysis in the con-text of a most general framework. Their specific conjec-tures refer to propositions concerning the causes of fluc-

tuation of monetary growth and propositions about properinterpretation of policy.‘John 0. Ourley and Edward F. Shaw, Money in a Theoryof Finance, (Washington: Brookings Institute, 1960). JamesTobin, ‘Commercial Banks as Creators of Money,” Bank-ing and Monetary Studies, ed. Deane Carson (R. D. Irwin,1963).

FEDERAL RESERVE BANK OF St LOUIS

7

explaining the responses of the monetary sys-tem to policy actions or to influences emanatingfrom the real sector. The New View proposed asystematic application of economic analysis, inparticular an application of relative price theory,to the array of financial intermediaries, theirassets and liabihties.

This program is most admirable and in-contestable, but it cannot explain the conflictrevealed by critique and countercritique. TheMonetarist approach accepted the general prin-ciple of applying relative price theory to theanalysis of monetary processes. In addition, thisapproach used the suggestions and analyticalpieces inherited from past efforts in order todevelop some specific hypotheses which do ex-plain portions of our observable environment.The New Viewers’ obvious failure to recognizethe limited content of their programmatic state-ments only contributes to maintenance of theconflict.

A subtle difference appears, however, in theresearch strategy. The New View was introduc-ed essentially as a generalized approach, in-cluding a quite formal exposition, but with littleattempt at specific structuring and empiricalcontent. The most impressive statements pro-pagated by the New View were crucially in-fluenced by the sheer formalism of its exposi-tion. In the context of the New View’s almostempty form, little remains to differentiate oneobject from another. For instance, in case oneonly admits the occurrence of marginal costsand marginal yields associated with the actionsof every household, firm, and financial interme-diary, one will necessarily conclude that banksand non-bank financial intermediaries arerestricted in size by the same economic forcesand circumstances. In such a context there istruly no essential difference between the deter-mination of bank and non-bank intermediaryliabilities, or between banks and non-bank in-

termediaries, or between money and other fi-nancial assets.

The strong impressions conveyed by the NewView thus result from the relative emptiness ofthe formulation which has been used to elabo-rate their position. In the context of the formalworld of the New View, “almost everything isalmost like everything else.” This undifferentia-ted state of affairs is not, however, a propertyof our observable world. It is only a property ofthe highly formal discussion designed by theNew View to overcome the unsatisfactory stateof monetary analysis still prevailing in the late1950’s or early 1960’s.8

Two sources of the conflict have been recog-nized thus far. The Monetarists’ research strat-egy was concerned quite directly with theconstruction of empirical theories about themonetary system, whereas the New View in-dulged, for a lengthy interval, in very generalprogrammatic excursions. Moreover, the NewViewers apparently misconstrued their programas being a meaningful theory about our obser.vable environment. This logical error con-tributed to a third source of the persistentconflict.

The latter source arises from the criticism ad-dressed by the New Viewers to the Monetarists’theories of money supply processes. Three ofthe papers exploit the logically dubious butpsychologically effective juxtaposition between a“New View” and a “Traditional View.” In doingthis they fail to distinguish between the inheri-ted state of monetary system analysis typicallyreflected by the money and banking textbooksyndrome and the research output ofeconomists advocating the Monetarist thesis.This distinction is quite fundamental. Some for-mal analogies misled the New Viewers and theydid not recognize the logical difference betweendetailed formulations of empirical theories on

6Adequate analysis of the medium of exchange function ofmoney, or of the conditions under which inside moneybecomes a component of wealth, was obstructed by theprogrammatic state of the New View. The useful analysisof the medium-of-exchange function depends on a decisiverejection of the assertion that “everything is almost likeeverything else.” This analysis requires proper recognitionthat the marginal cost of information concerning qualitiesand properties of assets differs substantially betweenassets, and that the marginal cost of readjusting assetpositions depends on the assets involved. The analysis ofthe wealth position of inside money requires recognition ofthe marginal productivity of inside money to the holder.Adequate attention to the relevant differences between

various cost or yield functions associated with differentassets or positions is required by both problems. Theblandness of the New View’s standard program cannotcope with these issues. The reader may consult apreliminary approach to the analysis of the medium of ex-change function in the paper by Karl Brunner and Allan H.Meltzer, in the Journal of Finance, 1964, listed in footnote9. He should also consult for both issues the importantbook by Boris Pesek and Thomas Saving. Money, Wealthand Economic Theory, The Macmillan Company, NewYork, 1967, or the paper by Harry Johnson, ‘InsideMoney, Outside Money, Income, Wealth and Welfare inMonetary Theory,” to be published in the Journal ofMoney, Credit and Banking, December 1968.


8

the one side and haphazard pieces of unfinishedanalysis on the other side.9

A related failure accompanies this logical er-ror. There is not the slightest attempt to assessalternative hypotheses or theories by systematicexposure to observations from the real world. Itfollows, therefore, that the countercritiquescarcely analyzed the empirical theories advanc-ed by the Monetarist critique and consequentlyfailed to understand the major implications ofthese theories.

For instance, they failed to recognize the roleassigned by the Monetarist view to banks’ be-havior and the public’s preferences in themonetary process. The objection raised by theNew View that “the formula [expressing a basicframework used to formulate the hypothesis]obscures the important role played by thepublic” has neither analytical basis nor meaning.In fact, the place of the public’s behavior wasdiscussed in the Monetarist hypotheses in somedetail. Moreover, the same analysis discussedthe conditions under which the public’sbehavior dominates movements of the moneystock and bank credit.b0 It also yielded informa-

tion about the response of bank credit, moneystock and time deposits to changes in ceilingrates, or to changes in the speed with whichbanks adjust their deposit-supply conditions toevolving market situations. Every single aspectof the banks’ or the public’s behavior emphasiz-ed by the countercritique has been analyzed bythe Monetarists hypotheses in terms whichrender the results empirically assessable. Littleremains, consequently, of the suggestivecountercritique assembled in the papers byGramley-Chase, Kareken and Cacy.11

B, A Monetarist Examination of theNew View’s Money Supply Theory

Three sources of the conflict have been dis-cussed thus far. Two sources were revealed aslogical misconstruals, involving inadequate con-struction and assessment of empirical theories.A third source pertains to legitimate differencesin research strategy. These three sources do notexplain all major aspects of the conflict. Beyondthe differences in research strategy and logicalmisconceptions, genuinely substantive issues re-main. Some comments of protagonists advocating

°Asexamples of the empirical work performed by theMonetarists, the reader should consult the following works:Milton Friedman and Anna Jacobson-Schwartz. AMonetary History of the United States, 1867-1960,(Princeton: Princeton University Press, 1963). PhilipCagan, Determinants and Effects of Changes in the Stockof Money, (Columbia: Columbia University Press, 1965).Karl Brunner and Allan H. Meltzer, “Some Further In-vestigations of Demand and Supply Functions for Money,”Journal of Finance, Volume XIX, May 1964. Karl Brunnerand Allan H. Meltzer, “A Credit-Market Theory of theMoney Supply and an Explanation of Two Puzzles in U.S.Monetary Policy,” Essays in Honor of Marco Fanno, 1966,Padova, Italy. Karl Brunner and Robert Crouch, “MoneySupply Theory and British Monetary Experience,” Methodsof Operations Research Ill—Essays in Honor of WilhelmKrelle, ed. Rudolf Henn (Published in Meisenheim, Ger-many, by Anton Ham, 1966). Karl Brunner, “A Schema forthe Supply Theory of Money,” International EconomicReview, 1961. Karl Brunner and Allan H. Meltzer, “AnAlternative Approach to the Monetary Mechanism,” Sub-committee on Domestic Finance, Committee on Bankingand Currency, House of Representatives, August 17, 1964.

10The reader will find this analysis in the following papers:Karl Brunner and Allan H. Meltzer, “Liquidity Traps forMoney, Bank Credit, and Interest Rates,” Journal ofPolitical Economy, April 1968. Karl Brunner and Allan H.Meltzer, “A Credit-Market Theory of the Money Supplyand an Explanation of Two Puzzles in U.S. MonetaryPolicy,” Essays in Honor of Marco Fanno, Padova, Italy,1966.

“The reader is, of course, aware that these assertions re-quire analytic substantiation. Such substantiation cannotbe supplied within the confines of this article. But thereader could check for himself, If he finds, in the contextof the countercritique, an analysis of the monetarists’ ma-jor hypotheses, an examination of implication, and ex-

posure to observations, I would have to withdraw mystatements. A detailed analysis of the banks’ and thepublic’s role in the money supply, based on two differenthypotheses previously reported in our papers will bedeveloped in our forthcoming books. This analysis, by itsvery existence, falsifies some major objections made byCacy or Gramley-Chase. Much of their criticism is eitherinnocuous or fatuous. Gramley-Chase indulge, for in-stance, in modality statements, i.e. statements obtainedfrom other statements by prefixing a modality qualifier like“maybe” or “possibly.” The result of qualifying an em-pirical statement always yields a statement which isnecessarily true but also quite uninformative. The modalitygame thus yields logically pointless but psychologically ef-fective sentences. Cacy manages, on the other hand,some astonishing assertions. The New View is creditedwith the discovery that excess reserves vary over time. Hetotally disregards the major contributions to the analysis ofexcess reserves emanating from the Monetarists’research. A detailed analysis of excess reserves wasdeveloped by Milton Friedman and Anna Schwartz in thebook mentioned in footnote 9. The reader should also notethe work by George Morrison, Liquidity Preferences ofCommercial Banks, (Chicago: University of Chicago Press,1966), and the study by Peter Frost, “Banks’ Demand forExcess Reserves,” an unpublished dissertation submittedto the University of California at Los Angeles, 1966. Theclassic example of an innocuous achievement was sup-plied by Cacy with the assertion: “. . - the actual volumeof money balances determined by competitive marketforces may or may not be equal to the upper limitestablished by the central bank” (p. 8). Indeed, we knewthis before the New View or Any View, just as we alwaysknew that “it may or may not rain tomorrow.” The readershould note that similar statements were produced byother authors with all the appearances of meaningfulelaborations.


9

the New View should probably be interpretedas conjectures about hypotheses to be expectedfrom their research strategy. It should be clear-ly understood that such conjectures are not logi-cal implications of the guiding framework. tn-stead, they are pragmatic responses to the gen-eral emphasis associated with this approach.

A first conjecture suggests that the moneystock and bank credit are dominated by thepublic’s and the Banks’ behavior. It is suggested,therefore, that cyclical fluctuations of monetarygrowth result primarily from the responses ofbanks and the public to changing business con-ditions. A second conjecture naturally supple-ments the above assertions. It is contended thatthe money stock is a thoroughly “untrustworthyguide to monetary policy.’

Articles by Gramley-Chase and Kareken at-tempt to support these conjectures with the aidof more explicit analytical formulations allegedlyexpressing the general program of the NewView. The paper contributed by Gramley-Chasehas been critically examined in detail onanother occasion, and only some crucial aspectsrelevant for our present purposes will be con-sidered at this point.’2 Various aspects of thefirst conjecture are examined in this and thenext section. The second conjecture is examinedin sections D and E.

A detailed analysis of the Gramley-Chasemodel demonstrates that it implies the followingreduced form equations:

M = g(B’, Y, c)E = h(B°,Y, c) h>o>h,, and h1 >g,’3

explaining the money stock (M) and bank credit(E) in terms of the extended monetary base (B’),the level of economic activity expressed by na-tional income at current prices (Y), and the ceil-ing rate on time deposits (c).’~

The Gramley-Chase model implies that mone-tary policy does affect the money stock andbank credit. It also implies that the money stockresponds positively and bank credit negatively toeconomic activity. The model thus differs fromthe Monetarist hypotheses which imply that

both bank credit and the money stock respondpositively to economic activity. The Gramley-Chase model also implies that the responses ofboth the money stock and bank credit tomonetary actions are independent of thegeneral scale of the public’s and the banks’ in-terest elasticities. Uniformly large or small in-terest elasticities yield the same response in themoney stock or bank credit to a change in themonetary base.

A detailed discussion of the implicationsderivable from a meaningfully supplementedGramley-Chase model is not necessary at thispoint. We are foremost interested in the rela-tion between this model and the propositionsmentioned in the previous paragraph. The firstproposition can be interpreted in two differentways. According to one interpretation, it couldmean that the marginal multipliers g, and h(i = 1, 2) are functions of the banks’ and thepublic’s response patterns expressing varioustypes of substitution relationships between dif-ferent assets. This interpretation is, however,quite innocuous and yields no differentiationrelative to the questioned hypotheses of theMonetarist position.

A second interpretation suggests that thegrowth rate of the money stock is dominated bythe second component (changes in income) ofthe differential expression:

AM = g, AB’ + g3 AY

This result is not actually implied by theGramley-Chase model, but it is certainly consis-tent with the model. However, in order toderive the desired result, their model must besupplemented with special assumptions aboutthe relative magnitude of g, and g2, and alsoabout the comparative cyclical variability of AB’and AY. This information has not been providedby the authors.

Most interesting is another aspect of themodel which was not clarified by the authors.Their model implies that policymakers couldeasily avoid procyclical movements in AM. Thismodel exemplifying the New View thus yields

‘2The reader may consult my chapter “Federal ReservePolicy and Monetary Analysis” in Indicators and Targets ofMonetary Policy, ed., by Karl Brunner, to be published byChandler House Publishing Co., San Francisco. This bookalso contains the original article by Gramley-Chase. Fur-ther contributions by Patric H. Hendershott and RobertWemntraub survey critically the issues raised by theGramley-Chase paper.

131n the Gramley-Chase model, 93 and h, are indeterminant.‘~Thisimplication was demonstrated in my paper listed in

footnote 12. The monetary base is adjusted for the ac-cumulated sum of reserves liberated from or impounded inrequired reserves by changes in requirement ratios.


10

little justification for the conjectures of itsproponents.

A central property of the Gramley-Chasemodel must be considered in the light of theprogrammatic statements characterizing theNew View. Gramley-Chase do not differentiatebetween the public’s asset supply to banks andthe public’s demand for money. This procedureviolates the basic program of the New View,namely, to apply economic analysis to an arrayof financial assets and financial institutions.Economic analysis implies that the public’s assetsupply and money demand are distinct, and notidentical behavior patterns. This difference inbehavior patterns is clearly revealed by dif-ferent responses of desired money balances anddesired asset supply to specific stimuli in the en-vironment. For instance, an increase in the ex-pected real yield on real capital raises thepublic’s asset supply but lowers the public’smoney demand. It follows thus that a centralanalytical feature of the Gramley-Chase modelviolates the basic and quite relevant program ofthe New View.

Karenken’s construction shares this fundamen-tal analytical flaw with the Gramley~Chasemodel, but this is not the only problem faced byhis analysis. The Karenken analysis proceeds ontwo levels. First, he derives a representativebank’s desired balance sheet position. For thispurpose he postulates wealth maximization sub-ject to the bank’s balance sheet relation be-tween assets and liabilities, and subject toreserve requirements on deposits. On closer ex-amination, this analysis is only applicable to amonopoly bank with no conversion of depositsinto currency or reserve flows to other banks.In order to render the analysis relevant for arepresentative bank in the world of reality, ad-ditional constraints would have to be introducedwhich modify the results quite substantially. Itis also noteworthy that the structural propertiesassigned by Karenken to the system of marketrelations are logically inconsistent with the im-plications one can derive from the author’s

analysis of firm behavior developed on the firstlevel of his investigation.

This disregard for the construction of aneconomic theory relevant for the real world iscarried into the second level of analysis wherethe author formulates a system of relationsdescribing the joint determination of interestrates, bank credit, and money stock. A remark-able feature of the Karenken model is that ityields no implications whatsoever about theresponse of the monetary system to actions ofthe Federal Reserve. It can say nothing, as itstands, about either open market operations orabout discount rate and reserve requirement ac-tions. This model literally implies, for instance,that the money stock and the banking system’sdeposit liabilities do not change as a result ofany change in reserve requirements ratios.

None of the conjectures advanced by thecountercritique concerning the behavior of themoney stock and the role of monetary policyfind analytical support in Karenken’s analysis.To the extent that anything is implied, it wouldimply that monetary policy operating directly onbank reserves or a mysterious rate of return onreserves dominates the volume of deposits—apractically subversive position for a follower ofthe New View.”

C. Alternative Explanations ofCyclical Fluctuations in MonetaryGrowth

The examination thus far in this article hasshown that even the most explicit formulation(Gramley-Chase) of the countercritique, allegedlyrepresenting the New View with respect tomonetary system analysis, does assign a signifi-cant role to monetary policy. This examinationalso argued that the general emphasis given bythe New View to the public’s and the banks’behavior in determination of the money stockand bank credit does not differentiate its pro-duct from analytical developments arising fromthe Monetarist approach. It was also shown that

‘5Two direct objections made to the Brunner-Meltzeranalysis by Kareken should be noted. He finds that thequestioned hypotheses do not contain “a genuine supplyfunction” of deposits. Accepting Karenken’s terminology,this is true, but neither does the Gramley-Chase modelcontain such a supply function. But the objection has noevidential value anyway. If a hypothesis were judged un-satisfactory because some aspects are omitted, allhypotheses are “unsatisfactory.” Moreover, the cognitivestatus of a empirical hypothesis does not improve simply

because an “analytical underpinning” has been provided.Karenken also finds fault with our use of the term “moneysupply function.” Whether or not one agrees with this ter-minological preferences surely does not affect the relationbetween observations and statements supplied by thehypothesis. And it should be clear that the status ofhypothesis depends only on this relation, and not onnames attached to statements.


11

the only explicit formulation advanced by theNew Viewers does not provide a sufficient basisfor their central conjectures. It is impossible toderive the proposition from the Gramley-Chasemodel that the behavior of the public andbanks, rather than Federal Reserve actions,dominated movements in the money supply. Butthe declaration of innocence by the countercriti-que on behalf of the monetary authorities withrespect to cyclical fluctuations of monetarygrowth still requires further assessment.

The detailed arguments advanced to explainthe observed cyclical fluctuations of monetarygrowth differ substantially among the con-tributors to the countercritique. Gramley-Chasemaintain that changing business conditionsmodify relative interest rates, and thus inducecountercyclical movements in the time depositratio. These movements in demand and timedeposits generate cyclical fluctuations inmonetary growth. On the other hand, Cacydevelops an argument used many years ago byWicksell and Keynes, but attributes it to theNew View. He recognizes a pronounced sen-sitivity of the money stock to variations in thepublic’s money demand or asset supply. Thesevariations induce changes in credit market con-ditions. Banks, in turn, respond with suitableadjustments in the reserve and borrowingratios. The money stock and bank credit conse-quently change in response to this mechanism.

Davis actually advances two radically differentconjectures about causes of cyclical fluctuationsof monetary growth. The first conjecture at-tributes fluctuations of monetary growth to thepublic’s and banks’ responses. Changing busi-ness conditions modify the currency ratio, thebanks’ borrowing ratio, and the reserve ratio.The resulting changes generate the observedmovements in money. His other conjecture at-tributes fluctuations in monetary growth toFederal Reserve actions: “the state of businessinfluences decisions by the monetary authoritiesto supply reserves and to take other actionslikely to affect the money supply.””

The various conjectures advanced by Gramley-Chase, Cacy, and Davis in regard to causes ofmovements in money and bank credit can beclassified into two groups. One set of conjec-tures traces the mechanism generating cyclicalfluctuations of monetary growth to the re-sponses of banks and the public; the behaviorof monetary authorities is assigned a com-paratively minor role. The other group of con-jectures recognizes the predominant role of thebehavior of monetary authorities.

In the following analysis the framework pro-vided by the Monetarist view will be used toassess these conflicting conjectures. The em-phasis concerning the nature of the causalmechanisms may differ between the variousconjectures regarding sources of variations inmoney, but the following examination will beapplied to an aspect common to all conjecturesemphasizing the role of public and bankbehavior.

In the context of the Monetarist framework,the money stock (M) is exhibited as a product ofa multiplier (m) and the monetary base (B),(such that M = mB). This framework, withoutthe supplementary set of hypotheses and theo-ries bearing on the proximate determinants ofmoney summarized by the multiplier and thebase, is completely neutral with respect to therival conjectures; it is compatible with any setof observations. This neutrality assures us thatits use does not prejudge the issue under con-sideration. The Monetarist framework operatesin the manner of a language system, able to ex-press the implications of the competing conjec-tures in a uniform manner.

The first group of conjectures advanced bythe countercritique (behavior of the public andbanks dominates movements in money) impliesthat variations in monetary growth betweenupswings and downswings in business activityare dominated by the variations in the mone-tary multiplier. The second group (behavior ofmonetary authorities dominates movements in

‘6Davis, p. 66. One argument about monetary policy in thesame paper requires clarification. Davis asserts on p. 68that the money supply need not be the objective of policy,and “given this fact, the behavior of the rate of growth ofthe money supply during the period cannot be assumed tobe simply and directly the result of monetary policy deci-sions alone.” This quote asserts that the money supply is“simply and directly the result of policy alone” wheneverpolicy uses the money supply as a target. This is in a

sense correct. But the quote could easily be misinter-preted due to the ambiguity of the term “policy.” Thisterm is frequently used to designate a strategy guiding theadjustment of policy variables. Is is also frequently used torefer to the behavior of the policy variables or directly tothe variables as such. The quote is quite acceptable in thefirst sense of ‘policy,” but thoroughly unacceptable in thesecond sense.


12

money) implies that, in periods with unchangedreserve requirement ratios and ceiling rates ontime deposits, variations in the monetary basedominate cyclical changes in monetary growth.The movements of the monetary multiplierwhich are strictly attributable to the changingof requirement ratios can be separated from thetotal contribution of the multiplier and com-bined with the monetary base. With this adjust-ment, the second group of conjectures impliesthat the monetary base, supplemented by thecontribution of reserve requirement changes tothe multiplier, dominates variations in themoney stock.

In this examination of contrasting explanationsof monetary fluctuations, values of the moneystock (M), the multiplier (m), and the monetarybase adjusted for member bank borrowing (B)are measured at the initial and terminal monthof each half business cycle (i.e., expansions andcontractions) located by the National Bureau ofEconomic Research. We form the ratios of thesevalues and write:

M0

mB~ ~~2;or l~= a/3

m0 B0

The subscript I refers to values of the terminalmonth and the subscript 0 to values of the in-itial month. ‘I’hese ratios were measured foreach half cycle in the period March 1919 to De-cember 1966. They were computed for two def-initions of the money stock, inclusive andexclusive of time deposits, with correspondingmonetary multipliers.

Kendall’s rank correlation coefficients betweenthe money stock ratios (j4 and the multiplierratios (a), and between (p) and the monetarybase ratio (/3) were computed. We denote thesecorrelation coefficients with ~ (hi, a) and ~ (i-i, /3).The implications of the two rival conjecturescan now be restated in terms of the two coeffi-cients. The first group of conjectures impliesthat Q(p, a) > Q(J1, /3); while the second groupimplies that in periods of unchanged reserve re-quirements ratios and ceiling rates on timedeposits, the coefficient q(pt~, /3) exceeds the coef-ficient Q(p, a). The second group implies nothingabout the relation of the two coefficients inperiods of changing reserve requirements andceiling rates on time deposits. It follows,therefore, that observations yielding the ine-quality Q(~1,/3 > Q(p, a) disconfirm the firstgroup and confirm the second group.

The correlations obtained are quite unam-biguous. The value of ~ (p. /3) is .537 for the

whole sample period, whereas ~ (p. a) is only.084. The half-cycle from 1929 to 1933 wasomitted in the computations, because move-ments in the money stock and the multiplierwere dominated by forces which do not dis-criminate between the rival conjectures underconsideration. The sample period, including1929 to 1933, still yields a substantially largervalue for Q(p, /3). The same pattern also holdsfor subperiods. In particular, computations bas-ed on observations for 1949 to 1966 confirmthe pattern observed for the whole sampleperiod. The results thus support the secondgroup of conjectures but not the first group.These results also suggest, however, that forcesoperating through the multiplier are not quitenegligible. The surprisingly small correlation Q(p,a) does not adequately reveal the operation ofthese forces. Their effective operation is revealedby the correlation Q(p, /3), which is far fromperfect, even in subperiods with constant re-serve requirement ratios. This circumstancesuggests that the behavior of the public andbanks contributes to the cyclical movements ofmonetary growth. The main result at this stage,however, the clear discrimination between thetwo groups of conjectures. The results are quiteunambiguous on this score.

Additional information is supplied by table I.For each postwar cycle beginning with thedownswing of 1948-49, the average annualgrowth rate of the money stock was computed.The expression M = mb was then used to com-pute the contribution to the average growth


13

/ ~ / ~ ~ ~ ~ “ c~ç C K~4 ~ ~ ~< /

~/ //~ ~ ~/< c ~ ~ \,‘ / ~%/~//~ ~ //\~ /fl ‘ ~_r/~ ~// ~/, ;~ ~ ~c ~ ~ ~N ~ ~/~‘~/‘ ~ ~/4 /~~/\ ~ ~ / ~‘~ / / 7~ r ~ ~: ~~ ~ //

/~ /~~ \ / ~ ~ <~ ~/tc/~ ~ ~/> / C’~~/~//~c ~, ~/ t~/ % ~ // ~, )r~ ~ ~ /~/ /// A~ ~ ‘/ ~ <~/,/~c ~/~/ ~ ~Ac/~ ~ ~7 ,// / ~ ~ \>/

// / ~ ~ / ,c~ ~ ~ / ~ ~ /~~v t ~ / ~\ /~ \ N

// ~/ ///~/// ~ ~ ~ ~ c~/& ~ -\~ ~// ~/~/ /~7//~\/ ~ // ,/ ~/\~~// ~ ~ ~/ ~/ -/ ~* 4 /: ~ // ~‘t

~ ~ ~ ~ ~ ~ _~/&/~/~\~ ~ ~/ “~ ~ ~ ~ ~ ~ //~ /// /~/~// / /\ ~\~/~ / // / g~ ,~ ~ ~//~A \\ ~ v~ ~ ~ ~ ~ / ‘~~, -\ ~ ~ < / < / \ /~ / ‘~/~/~‘/ ~rk~/’ ~ ~ ~ \ ~*

;~t 4~ ~ ~ ~// ~ ~ ~ 4~/ t~~r ~ ~,t ~ ~~/// ~ ~ ~ // // ~ / \,/ N

///~/ / ~ ~ ~/ / 4 \‘ / ~ / 7/ ~/ /~/~ 7~/~\ ~ ~ —\ /~ ~~//\~/77~ ‘m~7~/, / /~ , /7/~~ N

?; 1~:~4~:T~~ N ~ ~ / ~/ // ~ ~ 7/

rate of money from three distinct sources: (i)the behavior of monetary authorities (i.e., themonetary base and reserve requirement ratios),and the public’s currency behavior, (ii) the timedeposits substitution process, and, (iii) the varia-tions in the excess reserve and borrowing ratiosof commercial banks (Wicksell-Keynes mecha-nism).

The rank correlations between each contribu-tion, and the average growth rate of the moneystock over all postwar half-cycles clearly sup-port the conclusion of the previous analysis thatcyclical movements in the money stock aredominated by Federal Reserve actions.

Table I also presents the results of a similarexamination bearing on causes of movements inbank credit. The reader should note the radicaldifference in the observed patterns of correla-tion coefficients. The behavior of monetaryauthorities, supplemented by the public’s cur-rency behavior, does not appear to dominatethe behavior of bank credit. The three sourcescontributing to the growth rate of money all cx-erted influences of similar order on bank credit.

It appears that bank credit is comparatively lessexposed to the push of Federal Reserve actionsthan was the money stock. On the other hand,the money stock is less sensitive than bankcredit to the time-deposit substitutionmechanism emphasized by Gramley-Chase, andthe Wicksell-Keynes mechanism suggested byCacy. Most astonishing, however, is the negativeassociation between the average growth i-ate ofbank credit and the Wicksell-Keynes mechanismemphasized by Cacy.

It should also be noted that the averagegrowth rate of money conforms very clearly tothe business cycle. Such conformity does nothold for bank credit over the postwar half-cycles. This blurring occurred particularly inperiods when the ceiling rate on time depositswas increased. These periods exhibit relativelylarge contributions to the growth rate of bankcredit emanating from the time deposit substitu-tion mechanism.

A regression analysis (table II) of the reducedform equations derived from the Gramley-Chasemodel confirms the central role of the monetary

//~ ~ ~ /~ âIaMitasc ~ ~a*taaii~i: ~7 ~ ‘~aC~4~2 ~

;~~z ~ :: /~/ ~ ~:J~T:~d7 //4A~/ ~ / ~/ :~:/ ~//NN~ // ~ ~ / ~ / 7~ / ~ ;:~~> /7\ ~ /


14

base in the money supply process. Estimates ofthe regression coefficient relating money to in-come are highly unstable among different sam-ple periods, relative to the coefficient relatingmoney to the monetary base. Furthermore,estimates of regression coefficients relatingmoney to income occur in some periods withsigns which contradict the proposition ofGramley-Chase and Cacy, or exhibit a very smallstatistical significance. These diverse patterns ofcoefficients do not occur for the estimates ofcoefficients relating money and the monetarybase. It is also noteworthy that the averagegrowth rate of the monetary base (adjusted forchanges in reserve requirement ratios), over theupswings, exceeds without exception theaverage growth rate of adjacent downswings.This observation is not compatible with the con-tention made by Gramley-Chase that policy iscountercyclical.

Additional information is supplied by table III,which presents some results of a spectral analy-sis bearing on the monetary base and its sources.Spectral analysis is a statistical procedure fordecomposing a time series into seasonal,cyclical, and trend movements. After such ananalysis was conducted on the monetary baseand its sources, a form of correlation analysiswas run between movements in the monetarybase and movements in its various sources. Theresults of this procedure (table III) indicate thatmovements in Federal Reserve credit dominateseasonal and cyclical movements in the mone-tary base.

In summary, preliminary investigations yieldno support for the contention that the behaviorof banks and the public dominates cyclical move-ments in the money stock. The conjectures ad-vanced by Gramley-Chase or Cacy are thusdisconfirmed, whereas Davis’ second conjecturethat fluctuations in monetary growth may be at-tributed to Federal Reserve actions seems sub-stantially more appropriate. However, furtherinvestigations are certainly useful.

EL Relevance of Money andMonetary Actions with Respect toEconomic Activity

At present, a broad consensus accepts therelevance of money and monetary policy withrespect to economic activity. But this consensusconcerning the relevance of money emergesfrom two substantially different views about thenature of the transmission mechanism. One

Table Ill

.24617145

(n is lhe lCevnesrun conreplion (not Lii lu run-I LNCd \\ tb Kt’_~IWS ~1fl\ I enshrined in standardhorinuIatuon.~ot the inrome~expenditure(name—n ork. In We view, the interest rate h the mainlink heRs een nrone\ and economic acli’ it) - 1 hetither Vii’n rejects the traditional neparatuon olei’iir~(ini~it[lie )I\ iliti.) ft1115 rthtilrthl income,ir~iI~’.is(nNi(l—l) (ci)E~oi~iicSJ;iniii j)r’ite l111’liI\(rtiitio (~l()nl)ll)j(—’,). \ccorcling li~Ihirs r.itlu’i ~lntIll)lrl :inrij t’tril)Il)_~rn~’nit~iii’ {‘\hIIaIried h~a —~trit—able applirarinn of relati~ u~~’theory. Withr-eLfar-l.l to clisc-ii,siniis 1)1 tirt inripart ol morie~anici rrii,nretzir~aclior~sor~‘rl)Irl)riri(- arti~it~ thislatter den ha’~been termed the \lunt’larisl

po’itioui. Ilus ~05ilh0t1 rims he dWided into then eak \lnneL i-RE thesis anti the ~Lrong\honetarist lhe~j~.In a ‘,ense, both the Nm\ ren and the \honetarkt extension ol he tradi—tional en are represenled in tire n ak\honetarisl po~ilion).

ibm lohlii~xiO~4 (ii~r1Is~i1,r~¼clix 1101) [lii’ xvuakand We ~l rung \lrrnetarist thesk Fbi xx enkthi’’,i’, is cr)nrhiarecl xx ith ‘~rjn3ieaspecI~ol lireillrlirnr(’.e\llenlirtnrre ~rpr)rI)arl~Ii’ th’ lintrrrrrir1a—turn cit nratioiral ernn~omniraclidtx. tl1r ~trninrg

FEDERAL RESERVE BANK OF ST. LOUIS

15

thesis supplements the weak thesis with specialassumptions about our environment, in order toestablish the role of monetary forces in thebusiness cycle.

1, The Weak Monetansi Thesis

According to the %veak Monetarist thesis,monetary impulses are transmitted to theeconomy by a relative price process whichoperates on money, financial assets (and liabili-ties), real assets, yields on assets and the produc-tion of new assets, liabilities and consumables.The general nature of this process has beendescribed on numerous occasions and may beinterpreted as evolving from ideas developed byKnut, Wicksell, Irving Fisher and John MaynardKeynes.17

The operation of relative prices betweenmoney, financial assets and real assets may beequivalently interpreted as the working of aninterest rate mechanism (prices and yields ofassets are inversely related). Monetary impulsesare thus transmitted by the play of interestrates over a vast array of assets. Variations ininterest rates change relative prices of existingassets, relative to both yields and the supplyprices of new production. Acceleration ordeceleration of monetary impulses are thus con-verted by the variation of relative prices, or in-terest rates, into increased or reduced produc-tion, and subsequent revisions in the supplyprices of current output.

This general conception of the transmissionmechanism has important implications whichconflict sharply with the Keynesian interpreta-tion of monetary mechanisms expressed bystandard income-expenditure formulations.’~Inthe context of standard income-expenditureanalysis, fiscal actions are considered to have a“direct effect” on economic activity, whereasmonetary actions are considered to have onlyand “indirect effect.” Furthermore, a constantbudget deficit has no effect on interest rates in

a Keynesian framework, in spite of substantialaccumulation of outstanding government debtwhen a budget deficit continually occurs. Andlastly, the operation of interest rates on invest-ment decisions has usually been rationalizedwith the aid of considerations based on the ef-fects of borrowing costs.

These aspects of the income-expenditure ap-proach may be evaluated within the frameworkof the weak Monetarist thesis. The effects offiscal actions are also transmitted by the relativeprice mechanism. Fiscal impulses, i.e., govern-ment spending, taxing, and borrowing, operatejust as “indirectly” as monetary impulses, andthere is no a priori reason for believing thattheir speed of transmission is substantiallygreater than that of monetary impulses. Therelative price conception of the transmissionmechanism also implies that a constant budgetdeficit exerts a continuous influences on eco-nomic activity through persistent modificationsin relative prices of financial and real assets.Lastly, the transmission of monetary impulses isnot dominated by the relative importance ofborrowing costs. In the process, marginal costsof liability extension interact with marginal re-turns from acquisitions of financial and realassets. But interest rates on financial assets notonly affect the marginal cost of liability exten-sion, but also influence the substitution betweenfinancial and real assets. This substitution modi-fies prices of real assets relative to their supplyprices and forms a crucial linkage of the mone-tary mechanisms; this hnkage is usually omittedin standard income-expenditure analysis.

The description of monetary mechanisms inDavis’ article approaches quite closely the no-tion developed by the weak Monetarist thesis.This approximation permits a useful clarificationof pending issues. However, the criticisms andobjections advanced by Davis do not apply to the

“The reader may consult the following studies on thisaspect: Milton Friedman and David Meiselman, “TheRelative Stability of Monetary Velocity and the InvestmentMultiplier in the United States, 1897-1958,” in StabilizationPolicies, prepared by the Commission on Money andCredit, Englewood Cliffs, 1963. The paper listed in foot-note 21 by James Tobin should also be consulted. HarryJohnson, ‘Monetary Theory and Policy,” AmericanEconomic Review, June 1962. Karl Brunner ‘The Reportof the Commission on Money and Credit,” The Journal ofPolitical Economy, December 1961. Karl Brunner, “SomeMajor Problems of Monetary Theory,” Proceedings of theAmerican Economic Association, May 1961. Karl Brunnerand Allan H. Meltzer, ‘The Role of Financial Institutions in

the Transmission Mechanism,” Proceedings of theAmerican Economic Association, May 1963. Karl Brunner‘The Relative Price Theory of Money, Output, andEmployment,” unpublished manuscript based on a paperpresented at the Midwestern Economic AssociationMeetings, April 1967.

18The paper on “The Effect of Monetary Policy on Expen-ditures in Specific Sectors of the Economy,” presented byDr. Sherman Maisel at the meetings organized by theAmerican Bankers Association in September 1967, ex-emplifies very clearly the inherited Keynesian position. Thepaper will be published in a special issue of the Journal ofPolitical Economy.


18

weak Monetarist position. They are addressed toanother thesis, which might be usefully labeled thestrong Monetarist thesis.

Z The Strong Monetarist Thesis

If the theoretical framework of the weakMonetarist thesis is supplemented with addi-tional and special hypotheses, the strong Mone-tarist thesis is obtained. An outline of the strongthesis may be formulated in terms of three setsof forces operating simultaneously on the paceof economic activity. For convenience, they maybe grouped into monetary forces, fiscal forcesand other forces. The latter includetechnological and organizational innovation,revisions in supply prices induced by accruinginformation and expectation adjustments, capitalaccumulation, population changes and otherrelated factors or processes.

All three sets of forces are acknowledged bythe strong thesis to affect the pace of economicactivity via the relative price process previouslyoutlined. Moreover, the strong Monetarist pointof view advances the crucial thesis that the vari-ability of monetary forces (properly weightedwith respect to their effect on economic activi-ty) exceeds the variability of fiscal forces andother forces (properly weighted). It is arguedfurther that major variabilities occurring in asubset of the other forces (e.g., expectations andrevisions of supply prices induced by informa-tion arrival) are conditioned by the observedvariability of monetary forces. The conjecturethus involves a comparison of monetary vari-ability with the variability of fiscal forces andindependent “other forces.” According to thethesis under consideration, the variability ofmonetary of impulses is also large relative tothe speed at which the economy absorbs theimpact of environmental changes. This predomi-nance of variability in monetary impulses im-plies that pronounced accelerations in monetaryforces are followed subsequently by accelera-tions in the pace of economic activity, and thatpronounced decelerations in monetary forcesare followed later by retardations in economicactivity.

The analysis of the monetary dynamics, usingthe relative price process, is accepted by boththe weak and the strong Monetarist theses. Thisanalysis implies that the regularity of the ob-

served association between accelerations anddecelerations of monetary forces and economicactivity depends on the relative magnitude ofmonetary accelerations (or decelerations). Thesame analysis also reveals the crucial role ofchanges in the rate of change (second differ-ences) of the money stock in explanations offluctuations in output and employment. It im-plies that any pronounced deceleration, occurr-ing at any rate of monetary growth, retardstotal spending. It is thus impossible to statewhether any particular monetary growth, say a10 percent annual rate, is expansionary withrespect to economic activity, until one knowsthe previous growth rate. The monetary dynam-ics of the Monetarist thesis also explains thesimultaneous occurrence of permanent price-inflation and fluctuations in output and employ-ment observable in some countries.

The nature and the variability of the “Fried-man lag” may also be analyzed within theframework of the Monetarist thesis. This lagmeasures the interval between a change in signof the second difference in the money stock andthe subsequent turning point located by the Na-tional Bureau. In general, the lag at an upperturning point will be shorter, the greater theabsorption speed of the economy, and thesharper the deceleration of monetary impulsesrelative to the movement of fiscal forces andother forces. Variability in the relative accelera-tion or deceleration of monetary forces neces-sarily generates the variability observed in theFriedman lag.

What evidence may be cited on behalf of thestrong Monetarist thesis? Every major inflationprovides support for the thesis, particularly incases of substantial variations of monetarygrowth. The attempt at stabilization in the Con-federacy during the Civil War forms an im-pressive piece of evidence in this respect. Theassociation between monetary and economic ac-celerations or decelerations has also beenobserved by the Federal Reserve Bank of St.Louis.hhl Observations from periods with diver-gent movements of monetary and fiscal forcesprovide further evidence. For instance, suchperiods occurred immediately after terminationof World War II, from the end of 1947 to thefall of 1948, and again in the second half of1966. In all three cases, monetary forces pre-

‘~U.S.Financial Data, Federal Reserve Bank of St. Louis,week ending February 14, 1968. Also see “Money Supply

and Time Deposits, 1914-1964” in the September 1964issue of this Review.

FEDERAL REBERVE BANK OF St LOUIS

17

vailed over fiscal forces. The evidence adducedhere and on other occasions does not “prove”the strong Monetarist thesis, but does establishits merit for serious consideration.

Davis’ examination is therefore welcomed. Hisobjections are summarized by the followingpoints: (a) observations of the persistent associa-tion between money and income do not permitan inference of causal direction from money toincome; (b) the timing relation between moneyand economic activity expressed by the Fried-man lag yields no evidence in support of thecontention that variations in monetary growthcause fluctuations in economic activity; (c) thecorrelation found in cycles of moderate ampli-tude between magnitudes of monetary and eco-nomic changes was quite unimpressive; (d) thelength of the Friedman lag does not measurethe interval between emission of monetary im-pulse and its ultimate impact on economic ac-tivity. Furthermore, the variability of this lag isdue to the simultaneous operation and interac-tion of monetary and non-monetary forces.

Davis’ first comment (a) is of course quite trueand well known in the logic of science. It is im-possible to derive (logically) causal statements orany general hypotheses from observations. Butwe can use such observations to confirm ordisconfirm such statements and hypotheses.Davis particularly emphasizes that the persistentassociation between money and income couldbe attributed to a causal influence running fromeconomic activity to money.

Indeed it could, but our present state ofknowledge rejects the notion that the observedassociation is essentially due to a causal influ-ence from income on money. Evidence refutingsuch a notion was presented in section C. Theexistence of a mutual interaction over theshorter run between money and economic ac-tivity, however, must be fully acknowledged.Yet, this interaction results from the conceptionguiding policymakers which induces them to ac-celerate the monetary base whenever pressureson interest rates mount, and to decelerate themonetary base when these pressures wane. Ad-mission of a mutual interaction does not disposeof the strong Monetarist thesis. This interaction,inherent in the weak thesis, is quite consistentwith the strong position and has no disconfirm-ing value. To the contrary, it offers an explana-tion for the occurrence of the predominantvariability of monetary forces.

The same logical property applies to Davis’second argument (b). ‘the timing relation ex-pressed by the Friedman lag, in particular thechronological precedence of turning points inmonetary growth over turning points in eco-nomic activity, can probably be explained bythe influence of business conditions on themoney supply. Studies in money supply theorystrongly suggest this thesis and yield evidenceon its behalf. ‘The cyclical pattern of the curren-cy ratio, and the strategy typically pursued bymonetary policymakers explain this lead ofmonetary growth. And again, such explanationof the timing relation does not bear negativelyon the strong conjecture.

The objection noted under Davis’ point (c) issimilarly irrelevant. His observations actuallyconfirm the strong thesis. The latter impliesthat the correlation between amplitudes ofmonetary and income changes is itself cor-related with the magnitude of monetary ac-celerations or decelerations. A poor correlationin cycles of moderate amplitude, therefore,yields no discriminating evidence on the validitydiscriminating evidence on the validity of thestrong thesis. Moreover, observations describingoccurrences are more appropriate relative tothe formulated thesis than correlation measures.For instance, observations tending to disconfirmthe strong Monetarist thesis would consist of oc-currences of pronounced monetary accelera-tions or decelerations which are not followed byaccelerated or retarded movements of economicactivity.

Point (d) still remains to be considered. Onceagain, his observation does not bear on thestrong Monetarist thesis. Davis properly cau-tions readers about the interpretation of theFriedman lag. The variability of this lag is pro-bably due to the interaction of monetary andnon-monetary forces, or to changes from cycleto cycle in the relative variability of monetarygrowth. But again, this does not affect thestrong thesis. The proper interpretation of theFriedman lag, as the interval between reversalsin the rate of monetary impulses and theirprevalence over all other factors simultaneouslyoperating on economic activity, usefully clarifiesa concept introduced into our discussions. Thisclarification provides, however, no relevantevidence bearing on the questioned hypotheses.

In summary, the arguments developed byDavis do not yield any substantive evidenceagainst the strong Monetarist thesis. Moreover,


18

the discussion omits major portions of theevidence assembled in support of this position.2°

K Gountercvclical Polkv and theInterpretation of Monetary Policy

The usual assertion of the New View, at-tributing fluctuations of monetary growth to thepublic’s and the banks’ behavior, assumed astrategic role in the countercritique. The coun-tercritique denied, furthermore, that monetaryactions have a major impact on economic activi-ty. With the crumbling of these two bastions,the monetary policymakers’ interpretation oftheir own behavior becomes quite vulnerable.In a previous section, the substantial contribu-tion of the monetary base to the fluctuations ofmonetary growth has been demonstrated. Thesefacts, combined with repeated assertions thatmonetary policy has been largely counter-cyclical, suggest the existence of a pronounceddiscrepancy between actual behavior of themonetary authorities and their interpretation ofthis behavior.

A crucial question bearing on this issue per-tains to the proper measure summarizing actualbehavior of the monetary authorities. Two ma-jor facts should be clearly recognized. First, themonetary base consists of “money” directlyissued by the authorities, and every issue ofbase money involves an action of the monetaryauthorities. This holds irrespective of theirknowledge about it, or their motivation andaims. Second, variations in the base, extendedby suitable adjustments to incorporate changingreserve requirement ratios, are the single mostimportant factor influencing the behavior of themoney stock. And this second point applies ir-respective of whether Federal Reserve author-ities are aware of it or wish it to be, orwhatever their motivations or aims are. Theiractual behavior, and not their motivations or

2QMilton Friedman’s summary of the evidence in the Forty-Fourth Annual Report of the National Bureau of EconomicResearch is important in the respect. Davis overlooks inparticular the evidence accumulated in studies of themoney supply mechanism which bears on the issue raisedby point (a) in the text. A persistent and uniform associa-tion between money and economic activity, in spite oflarge changes in the structure of money supply processes,yields evidence in support of the Monetarist theses.The reader should also consult Chapter 13 of the book byMilton Friedman and Anna Schwartz listed in footnote 9;Studies in the Quantity Theory of Money, edited by MiltonFriedman, University of Chicago Press, 1956; and a doc-toral dissertation by Michael W. Keran, Monetary Policyand the Business Cycle in Postwar Japan,” Ph.D. thesis atthe University of Minnesota, March 1966, to be published

aims, influences the monetary system and thepace of economic activity. Thus, actual changesin the monetary base are quite meaningful andappropriate measures of actual behavior ofmonetary authorities.~’

The information presented in table IV sup-ports the conjecture that monetary policy-makers’ interpretation of their own behaviorhas no systematic positive association with theiractual behavior. Table IV was constructed onthe basis of the scores assigned to changes inpolicies, according to the interpretation of theFederal Open Market Committee.22 Positivescores were associated with each session of theFOMC which decided to make policy easier,more expansionary, less restrictive, less tight,etc., and negative scores indicate decisions tofollow a tighter) less expansionary, more restric-tive course. The scores varied between plus andminus one, and expressed some broad orderingof the revealed magnitude of the changes.

An examination of the sequences of scoreseasily shows that the period covered can benaturally partitioned into subperiods exhibitingan overwhelming occurrence of scores with auniform sign. These subperiods are listed in thefirst column of table IV. The second columncumulated the scores over the subperiods listedin order to yield a very rough ranking of thepolicymakers’ posture according to their owninterpretation.

Table TV reveals that the FOMC interpretedthe subperiods from August 1957 to July 1958,and from July 1959 to December 1960 asamong the most expansionary policy periods.The period from November 1949 to May 1953appears in this account as a phase of persistent-ly tight or restrictive policy. The next two col-umns list the changes of two important vari-ables during each subperiod. The third column

as a chapter of a book edited by David Meiselman.21The reader may also be assured by the following

statement:monetary policy refers particularly to determination of

the supp1y of (the government’s) demand debt Thisdemand debt coincides with the monetary base. The quoteis by James Tobin, a leading architect of the New View,on p. 148 of his contribution to the Commission on Moneyand Credit, “An Essay on Principles of Debt Manage-ment,” in Fiscal and Debt Management Policies, PrenticeHall, Englewood Cliffs, 1963.

22The scores were published as Appendix It to “An Alter-native Approach to the Monetary Mechanism.” See foot-note 9.


19

~

N \N<NN 4~çN \NN~> ~N(~ N

tábIet\t\ N N / N

The$soclMtOftbthleen P ky~nakers’1ntetpátatièrot~oiiçy,Chanies in ths MosSrBS and CftMtget~f’reaR~ánt. N

N C ScØt’eS N N N N cttanØsN1n’~bp

0 Po~cymakevt N N NN lntøptestt*n tP* ~ettoØIn tt~ø~PeSt N

PerIods tev~hePe8bd , htS~4I*O4tN NN

N N N N N

eisattS4 4-US N N N~~N N + N

2184\CtM\ \\e~\N~ \~\ ~ 34~~N ~ -+-tn ~Ssz

S5~ N~1N N’~44 N ~7NNN~

N iisi ~ N kS$GNN~ NN4-fl N NN N N~8~bs N aN N

N NaN ~ N ~ N4N N

15 ~ flN ~~N4fl ~X N

~ N \~NN\N’N\/

N N NN NNN ~ N / ~N N/ /~/ ~ N N

describes changes in free reserves, and thefourth column notes changes in the monetarybase. A cursory examination of the columns im-mediately shows substantial differences in theirbroad association. The rank correlation betweenthe various columns is most informative for ourpurposes.

These rank correlations are listed in table V.The results expose the absence of any positiveassociation between the policymakers’ own in-terpretation or judgment of their stance andtheir actual behavior, as indicated by move-ments in the monetary base. The correlationcoefficient between the monetary base andcumulated scores has a negative value, sug-gesting that a systematic divergence betweenstated and actual policy (as measured by themonetary base) is probable. On the other hand,

Table VRank Correlation Between Changes inthe Monetary Base, Changes in FreeReserves and the Cumulated Scoresof Policymakers’ Interpretations

Cumulated scores and base -- 09~umL1latedscores and free reserves -- .70Free reserves and base — .26

the correlation between the policymakers’ de-scriptions of their posture, and the movementof free reserves, is impressively close. This cor-relation confirms once again that the FederalReserve authorities have traditionally used thevolume of free reserves as an indicator to gaugeand interpret prevailing monetary policies. Yetlittle evidence has been developed which estab-lishes a causal chain leading from changes infree reserves to the pace of economic activity.

Another observation contained in table IVbears on the issue of policymakers interpreta-tion of their own behavior. Changes in thecumulated scores and free reserves between theperiods listed always move together and areperfect in terms of direction. By comparison,the co-movement between cumulated scores andchanges in the monetary base is quite haphaz-ard; only three out of eight changes betweenperiods move together. This degree of co-move-ment between cumulated scores and themonetary base could have occurred by purechance with a probability greater than .2,whereas the probability of the perfect co-movement between cumulated scores and free

reserves occurring as a matter of pure chanceis less than .004. The traditional selection offree reserves or money market conditions as anindicator to interpret prevailing monetary policyand to gauge the relative thrust applied by

policy, forms the major reason for the negativeassociation (or at least random association) be-tween stated and actual policy.

551 >,;1~4~7 \~//

;,//NNNN~~\~~N/N/, ~N~/N, /


20

Attempts at rebuttal to the above analysisoften emphasize that policymakers are neitherinterested in the monetary base, nor do they at-tach any significance to it. This argument is ad-vanced to support the claim that the behaviorof the monetary base is irrelevant for a properexamination of policymakers’ intended behavior.‘this argument disregards, however, the factsstated earlier, namely, movements in the mone-tary base are under the direct control and arethe sole responsibility of the monetary author-ities. It also disregards the fact that actions mayyield consequences which are independent ofmotivations shaping the actions.

These considerations are sufficient toacknowledge the relevance of the monetarybase as a measure summarizing the actualbehavior of monetary authorities. However,they alone are not sufficient to determinewhether the base is the most reliable indicatorof monetary policy. Other magnitudes such asinterest rates, bank credit and free reserveshave been advanced with plausible argumentsto serve as indicators. A rational proceduremust be designed to deter-mine which of thepossible entities frequently used for scalingpolicy yields the most reliable results.

This indicator problem is still very poorlyunderstood, mainly because of ambiguous useof economic language in most discussions ofmonetary policy. The term “indicator” occurswith a variety of meanings in discussions, andso do the terms “target” and “guide.” The in-dicator problem, understood in its technicalsense, is the determination of an optimal scalejustifying interpretations of the authorities’ ac-tual behavior by means of comparative state-ments. A typical statement is that policy X ismore expansionary than policy Y, or that cur-rent policy has become more (or less) expan-sionary. Whenever we use a comparative con-cept, we implicitly rely on an ordering scale.

The indicator problem has not been givenadequate treatment in the literature, and therecognition of its logical structure is oftenobstructed by inadequate analysis. It is, for in-stance, not sufficient to emphasize the proposi-tion that the money supply can be a “misleadingguide to the proper interpretation of monetarypolicy.” This proposition can be easily demon-strated for a wide variety of models and hy-potheses. However, it establishes very little. Thesame theories usually demonstrate that the rateof interest, free reserves, or bank credit can

also be very misleading guides to monetarypolicy. Thus, we can obtain a series of proposi-tions about a vast array of entities, assertingthat each one can be a very misleading guide tothe interpretation of policy. We only reach auseless stalemate in this situation.

The usual solution to the indicator problem atthe present time is a decision based on mysticalinsight supplemented by some impressionisticarguments. The most frequently advanced argu-ments emphasize that central banks operatedirectly on credit markets where interest ratesare formed, or that the interest mechanismforms the centerpiece of the transmission pro-cess. Accordingly, in both cases market interestrates should “obviously” emerge as the relevantindicator of monetary policy.

These arguments on behalf of market interestrates are mostly supplied by economists. ‘themonetary authorities’ choice of money marketconditions as an indicator evolved from a dif-ferent background. But in recent years a subtlechange has occurred. One frequently encoun-ters arguments which essentially deny eitherthe existence of the indicator problem or its ra-tional solution. A favorite line asserts that “theworld is very complex” and consequently it isimpossible or inadmissible to use a single scaleto interpret policy. According to this view, onehas to consider and weigh many things in orderto obtain a “realistic” assessment in a compli-cated world.

‘this position has little merit. The objection toa “single scale” misconstrues the very nature ofthe problem. Once we decide to discuss mone-tary policy in term of comparative statements,an ordinal scale is required in order to providea logical basis for such statements. A multiplicityof scales effectively eliminates the use of coni-parative statements. Of course, a single scalemay be a function of multiple arguments, butsuch multiplicity of arguments should not beconfused with a multiplicity of scales. Policy-makers and economists should therefore realizethat one either provides a rational procedurewhich justifies interpretations of monetarypolicy by means of comparative statements, orthat one abandons any pretense of meaningfulor intellectually honest discussion of suchpolicy.

Solution of the indicator problem in thetechnical sense appears obstructed on occasionby a prevalent confusion with an entirely dif-ferent problem confronting the central banker—


21

the target problem. This problem results fromthe prevaihng uncertainty concerning the na-ture of the transmission mechanism and thesubstantial lags in the dynamics of monetaryprocesses.

In the context of perfect information, the in-dicator problem becomes trivial and the targetproblem vanishes. But perfect information is theprivilege of economists’ discourse on policy; cen-tral bankers cannot afford this luxury. The im-pact of their actions are both delayed anduncertain. Moreover, the ultimate goals ofmonetary policy (targets in the Tinbergen-Theilsense) appear remote to the manager executinggeneral policy directives. Policymakers will beinclined under these circumstances to insert amore immediate target between their ultimategoals and their- actions. These targets should bereliably observable with a minimal lag.

It is quite understandable that central bankerstraditionally use various measures of moneymarket conditions, with somewhat shiftingweights, as a target guiding the continuous ad-justment of their policy variables. This responseto the uncertainties and lags in the dynamics ofthe monetary mechanism is very rational in-deed. However, once we recognize the rationali-ty of such behavior, we should also considerthe rationality of using a particular target. Thechoice of a target still remains a problem, andthe very nature of this problem is inadequatelyunderstood at this state.

This is not the place to examine the indicatorand target problem in detail. A possible solutionto both problems has been developed on an-other occasion.23 The solutions apply decisiontheoretic procedures and concepts from controltheory to the determination of an optimalchoice of both indicator and target. Both pro-blems are in principle solvable, in spite of the“complexity of the world.” Consequently, thereis little excuse for failing to develop rationalmonetary policy procedures.

CONCLUSION

A program for applying economic analysis tofinancial markets and financial institutions iscertainly acceptable and worth pursuing. Thisprogram suggests that the public and banks in-

teract in the determination of bank credit, in-terest rates, and the money stock, in responseto the behavior of monetary authorities. But therecognition of such interaction implies nothingwith respect to the relative importance of thecausal forces generating cyclical fluctuations ofmonetary growth. Neither does it bear on thequality of alternative empirical hypotheses, orthe relative usefulness of various magnitudes orconditions which might be proposed as an in-dicator to judge the actual thrust applied bymonetary policy to the pace of economicactivity.

The Monetarist thesis has been put forth inthe form of well structured hypotheses whichare supported by empirical evidence. This ex-tensive research in the area of monetary policyhas established that: (i) Federal Reserve actionsdominate the movement of the monetary baseover time; (ii) movements of the monetary basedominate movements of the money supply overthe business cycle; and, (iii) accelerations ordecelerations of the money supply are closelyfollowed by accelerations or decelerations ineconomic activity. Therefore, the Monetaristthesis puts forth the proposition that actions ofthe Federal Reserve are transmitted to economicactivity via the resulting movements in themonetary base and money supply, which initiatethe adjustments in relative prices of assets,liabilities, and the production of new assets.

The New View, as put forth by the counter-critique, has offered thus far neither analysisnor evidence pertaining relevantly to an ex-planation of variations in monetary growth.Moreover, the countercritique has not devel-oped, on acceptable logical grounds, a system-atic justification for the abundant supply ofstatements characterizing policy in terms of itseffects on the economy. Nor has it developed asystematic justification for the choice of moneymarket conditions as an optimal target guidingthe execution of open market oper-ations.

But rational policy procedures require both areliable interpretation and an adequate deter-mination of the course of policy. The necessaryconditions for rational policy are certainly notsatisfied if policies actually retarding economicactivity are viewed to be expansionary, as in the

23The reader may consult the chapter by Karl Brunner andAllan H. Meltzer on “Targets and Indicators of MonetaryPolicy,” in the book of the same title, edited by Karl Brun-

ner. The book will be published by Chandler HousePublishing Co., Belmont, California.


22

case of the 1960-61 recession, or, if inflationaryactions are viewed as being restrictive, as in thefirst half of 1968.

The major questions addressed to our mone-tary policymakers, their advisors and consul-tants remain: How do you justify your inter-

pretation of policy, and how do you actually ex-plain the fluctuations of monetary growth? Themajor contentions of the academic critics of thepast performance of monetary authorities couldpossibly be quite false, but this should be dem-onstrated by appropriate analysis and relevantevidence.

FEOERAL RESERVE BANK OF St LOUIS

Documents

The Role of Money and Monetary Policy · cerning the role of money and monetary policy. ... “The new approach argues ... tions originally published by Gurley-Shaw and