Monetary Theory and the Trade Cycle

FIRST PUBLISHED 1933

LIBRARY OF CONGRESS CATALOGUE CARD NUMBER

66 - 22629

PRINTED IN THE UNITED 8TATB8 OP AMERICA

by SENTRY PRES8, NEW YORK, N. V. IOOIO,

I N T R O D U C T I O N T O T H E S E R I E S

THE science of Economics, like all other branchesof knowledge, recognizes no limitation by nationalboundaries. Contiguity of residence may give acertain unity to the speculation of particulargroups of economists, a tradition of good teachingmay give a presumption of excellence to theproducts of particular seminars; and in this senseit is not foolish to speak of local schools of econo-mic thought, or to attach geographical labels toparticular theories. But to speak of Economics,as distinct from economists, in terms of nationalor municipal classifications, to distinguish anEnglish Economics from a Continental Economics,and so on, has no more sense than to speak ofEnglish Arithmetic. The criteria of scientificvalidity take no account of origins, and theeconomist who refused to avail himself of aparticular set of propositions because they wereforeign would be acting no less unscientificallythan the chemist or physicist who acted on

5

I N T R O D U C T I O N

similar principles. It has been well said thatthere are only two kinds of Economics — goodEconomics and bad Economics. All otherclassifications are misleading.

Unfortunately, the economist, far more thanthe practitioner of the natural sciences, is victimto the curse of Babel. The chemist and the phy-sicist—dealing as they do with tangible andquantitative relations between easily definablethings — can converse in what to all intents andpurposes is an international language. A verymodest linguistic equipment is sufficient to enableone to follow all the chief contributions in suchbranches of science. In Economics this is not so.The complicated social relationships which areits chief preoccupation lend themselves much lessto merely symbolic analysis. No doubt, even here,mathematical methods can be of considerable useboth in assisting thought and in securing com-plete precision. But the description of what thesymbols represent, the delimitation and inter-pretation of concepts, these are matters which, inthe social sciences, require a wealth of qualitativeterminology and a subtlety of expression calculated

6

INTRODUCTION

to strain to the full the resources of any language.Small wonder, then, that the professional econo-mist, with limited time and limited powers at hisdisposal, will often conclude that depth is to bepreferred to breadth and that a more intensiveexploitation of the resources available to him inhis own language is likely to yield more than theattempt to assimilate material only available inforeign tongues. Small wonder too that, inconsequence of these conditions, there is probablymore overlapping and wasteful duplication ofeffort in Economics than in any other branch ofscientific knowledge. I know of no naturalscience in which it would be possible for a man todevote years to the discovery of propositions whichare already commonplaces in language areas otherthan his own. It is notorious that in Economicsthis frequently happens.

It follows, therefore, that, in Economics, evenmore than elsewhere, there is an urgent need fora continuous series of translations which shallmake available to economists in different coun-tries the results of investigations in languagesother than their own. In England before the war

7

I N T R O D U C T I O N

this need was beginning to be met. The Brucetranslation of Pantaleoni, the Wotzel translationof Pierson, the Smart translations • of Bohm-Bawerk, are examples of work of a high order ofscholarship and literary excellence which keptus in touch with the best products of contem-porary thought abroad. But the war (which upsetso many good things) interrupted this process ofinternationalization, and since the war it has notbeen completely resumed. This series is anattempt to make good the gap — to make availableto English and American readers the chief recentcontributions in foreign languages to the advance-ment of Economic Science.

It is unnecessary to introduce at length theauthor of the present volume, Professor vonHayek, until recently Director of the AustrianInstitut fur Konjunkturforschung, now TookeProfessor of Economic Science and Statistics inthe University of London. Professor Hayek'sPrices and Production, and his various contribu-tions to Economica and the Economic Journal willalready have made him familiar to most English

INTRODUCTION

readers interested in recent developments in thetheory of money and credit. The present volume,Monetary Theory and the Trade Cycle, is a trans-lation of a work which was published in Austriabefore Prices and Production, and which deals withthe problem of trade fluctuation from a moregeneral point of view. The preface makes clearthe precise relationship between this work and theauthor's other investigations in the same field.

LIONEL ROBBINS

The London School of EconomicsSeptember 1932

CONTENTS

PREFACE 15

I THE PROBLEM OF THE TRADE CYCLE 27

II NON-MONETARY THEORIES OF THE

TRADE CYCLE 5 I

III MONETARY THEORIES OF THE TRADE

CYCLE IOI

IV THE FUNDAMENTAL CAUSE OF CYCLICAL

FLUCTUATIONS 139

V UNSETTLED PROBLEMS OF TRADE CYCLE

THEORY I95

INDEX 243

11

ANALYSIS OF CONTENTS

CHAPTER I - THE PROBLEM OF THE TRADE CYCLE

1. The relation between empirical observation and theore-tical explanation

2. The use of statistics in the verification of theory3. The task of statistics is to give accurate information

about events which fall within the province of theory4. The main division in Trade Cycle theory is the division

between monetary and non-monetary theories5. The purpose of the following essay

CHAPTER II - NON-MONETARY THEORIES OF THE TRADE CYCLE

1. A general refutation of such theories is difficult onaccount of the absence of an adequate classification

2. The selection of subject matter for the present enquiry3. Theories which explain the Cycle in terms of the

technical conditions of production4. These all overlook the essential function of the mechan-

ism of relative prices5. The way in which prices react to such changes in

'data' as are assumed in these theories further examined6. Theories which explain fluctuations with reference to

phenomena arising out of the process of saving andinvestment

7. Psychological theories8. All these theories assume the variability of credit9. The main difference between monetary and other

changes in economic 'data'10. The existence of monetary disturbances having been

assumed, the first task of theory is to examine all thedeviations from the system contemplated by the puretheory of equilibrium which must necessarily followfrom this assumption

12


CHAPTER III - MONETARY THEORIES OF THE TRADE CYCLE

1. The main tasks of a monetary theory of the TradeCycle

2. Monetary explanations should not proceed from changesin the general level of prices

3. The essential point is the effect on the structure ofproduction of changes in the volume of money

4. Such changes in the structure of production are funda-mentally independent of changes in the value of money

5. Most criticisms of the monetary explanation are dueto a misunderstanding of this point

6. Changes in individual price relationships must be themain focus of attention, for such changes cause shiftsin the direction of production

7. Phenomena (such as deviations between the naturaland the money rates of interest) which are funda-mentally independent of changes in the value ofmoney

8. Advantages of the proposed method of approach; thepossibility of a rapprochement between monetary andnon-monetary theories

CHAPTER I V - T H E FUNDAMENTAL CAUSE OF CYCLICAL FLUC-TUATIONS

1. The problem is to discover why certain changes in'data' are able to lead the system away from equilibrium

2. Exogenous and endogenous monetary theories3. Of three possible ways of increasing the volume of

money, the most important, from the present point ofview, is the creation of credit by the banks

4. Confusion on this point is due largely to a failure todistinguish between the possibilities open to a singlebank and those open to the banking system as a whole

5. The actual origin of additional credits


6. The possibility of bankers arbitrarily creating creditsis ruled out because, in individual cases, it is impossibleto distinguish between those deposits which arosethrough cash payments and those which find theirorigin in credit

7. The reaction of the banks to an increased demand forcredit

8. The process of credit expansion and its cessation9. Elasticity in the volume of circulating media is a suffi-

cient cause of cyclical fluctuation10. The significance of these monetary influences from the

point of view of Trade Cycle theoryn . Their bearing on policy

CHAPTER V - UNSETTLED PROBLEMS OF TRADE CYCLE THEORY

1. All variations in the volume of effective circulationnecessarily involve movements other than those whichcan be immediately deduced from the propositions ofstatic theory

2. The most important effects are via movements in thestructure of the money rates of interest - which thuspresent a series of important problems

3. The significance of fluctuations in the natural rate ofinterest

4. The concept of a natural or equilibrium rate of interest5. The relation between the equilibrium rate and the

actual (money) rate of interest6. Forced saving as a cause of economic crises7. Interest rates in the money and capital markets re-

spectively8. Problems for statistical investigation

PREFACE

THE German essay*, of which the following is atranslation, represents an expanded version of apaper]* prepared for the meeting of the Vereinfiir Sozialpolitik,% held in Zurich in September1928, and of some remarks contributed to thediscussion at that meeting. Although, in revisingthe translation, I have made numerous minoralterations and additions (mainly confined to thefootnotes), the general course of the argument hasbeen left unchanged. The book, therefore, stillshows signs of the particular aim with which itwas written. In submitting it to a public differentfrom that for which it was originally intended, afew words of explanation are, perhaps, required.

* Geldtheorie und Konjunkturtheorie (Beitrage zur Konjunktur-forschung, herausgegeben vom Osterreichisches Institut fiir Kon-junkturforschung, No. 1). Vienna 1929.

t Einige Bemerkungen tiber das Verhdltnis der Geldtheorie zurKonjunkturtheorie in vol. 173, part 2 of the 'Schriften des Vereinsfur Sozialpolitik,* Miinchen 1928.

% Schriften des Vtreins fiir Sozialpolitik, vol. 175, pp. 369-374,Miinchen 1928.

PREFACE

In Germany, somewhat in contrast to thesituation in English-speaking countries, monetaryexplanations of the Trade Cycle were always, or atleast until quite recently, regarded with somemistrust. One of the aims of this study — oneon which an English reader may feel that I havewasted unnecessary energy — was to justify themonetary approach to these problems. But I hopethat this more explicit statement of the role of themonetary factor will not be found quite useless,for it is not only a justification of the monetaryapproach but also a refutation of some over-simplified monetary explanations which are widelyaccepted. In order to save the sound elementsin the monetary theories of the Trade Cycle, I hadto attempt, in particular, to refute certain theorieswhich have led to the belief that, by stabilizingthe general price level, all the disturbing monetarycauses would be eliminated. Although, since thisbook was written, this belief has been somewhatrudely shaken by the crisis of 1929, I hope that asystematic examination of its foundations will stillbe found useful. The critique of the programmeof the Stabilizers', which is in many ways the

16

PREFACE

central theme of this book, has now occupiedme for many years, and since I deal here onlywith some special problems which have grownmainly out of these studies, I may perhaps bepermitted to refer below to other publi-cations, in which I have partly dealt withcertain further theoretical problems and partlyattempted to use these considerations for theelucidation of contemporary phenomena.* Inparticular, my Prices and Production, originallypublished in England, should be considered asan essential complement to the present publication.While I have here emphasized the monetarycauses which start the cyclical fluctuations, Ihave, in that later publication, concentrated on thesuccessive changes in the real structure of production,

which constitute those fluctuations. This essentialcomplement of my theory seems to me to be the

• Die Wdhrungspolitik der Vereinigten Staaten seit der Uberwindungder Krise von 1920, Zeitschrift fur Volkswirtschaft und Sozialpolitik,N.F., vol. v, 1925. Das intertemporale Gleichgezvichtssystem derPreise und die Bewegungen des Geldwertes, 'WeltwirtschaftlichesArchiv,' vol. 28, 1928, The 'Paradox' of Saving, 'Economica' No. 32,May 1931. Prices and Production, London, 1931. Reflections on thePure Theory of Money of Mr. J. M. Keynes, 'Economica' Nos. 33-35,1931-32. Das Schicksal der Goldwdhrung, Der Deutsche Volkswirt,1932. Kapitalaufzehrung, 'Weltwirtschaftliches Archiv/ vol. 36,1932.

B 17

PREFACE

more important since, in consequence of actualeconomic developments, the over-simplified mone-tary explanations have gained undeserved pro-minence in recent times. And since, in all myEnglish publications, I have purposely refrainedfrom combining purely theoretical considerationswith discussions of current events, it may be usefulto add here one or two remarks on the bearingof those considerations on the problems ofto-day.

It is a curious fact that the general disinclinationto explain the past boom by monetary factorshas been quickly replaced by an even greaterreadiness to hold the present working of ourmonetary organization exclusively responsible forour present plight. And the same stabilizers whobelieved that nothing was wrong with the boomand that it might last indefinitely because pricesdid not rise, now believe that everything could beset right again if only we would use the weaponsof monetary policy to prevent prices from falling.The same superficial view which sees no otherharmful effect of a credit expansion but the riseof the price level, now believes that our only

18

PREFACE

difficulty is a fall in the price level, caused bycredit contraction.

There can, of course, be little doubt that, atthe present time, a deflationary process is goingon and that an indefinite continuation of thatdeflation would do inestimable harm. But thisdoes not, by any means, necessarily mean that thedeflation is the original cause of our difficultiesor that we could overcome these difficulties bycompensating for the deflationary tendencies, atpresent operative in our economic system, byforcing more money into circulation. Thereis no reason to assume that the crisis was startedby a deliberate deflationary action on the part ofthe monetary authorities, or that the deflationitself is anything but a secondary phenomenon,a process induced by the maladjustments ofindustry left over from the boom. If, however,the deflation is not a cause but an effect of theunprofitableness of industry, then it is surely vainto hope that, by reversing the deflationary process,we can regain lasting prosperity. Far from follow-ing a deflationary policy, Central Banks, particularlyin the United States, have been making earlier

PREFACE

and more far-reaching efforts than have everbeen undertaken before to combat the depressionby a policy of credit expansion — with the resultthat the depression has lasted longer and hasbecome more severe than any preceding one.What we need is a readjustment of those elementsin the structure of production and of priceswhich existed before the deflation began andwhich then made it unprofitable for industryto borrow. But, instead of furthering the inevit-able liquidation of the maladjustments broughtabout by the boom during the last three years, allconceivable means have been used to preventthat readjustment from taking place; and one ofthese means, which has been repeatedly triedthough without success, from the earliest to themost recent stages of depression, has been thisdeliberate policy of credit expansion.

It is very probable that the much discussedrigidities, which had already grown up in manyparts of the modern economic system before1929, would, in any case, have made the processof readjustment much slower and more painful.It is also probable that these very resistances to

20

PREFACE

readjustment would have set up a severe defla-tionary process which would finally have overcomethose rigidities. To what extent, under the givensituation of a relatively rigid price and wagesystem, this deflationary process is perhaps notonly inevitable but is even the quickest way ofbringing about the required result, is a verydifficult question, about which, on the basis ofour present knowledge, I should be afraid to makeany definite pronouncement.

It seems certain, however, that we shall merelymake matters worse if we aim at curing thedeflationary symptoms and, at the same time (bythe erection of trade barriers and other forms ofstate intervention) do our best to increase ratherthan to decrease the fundamental maladjustments.More than that: while the advantages of such acourse are, to say the least, uncertain, the newdangers which it creates are great. To combat thedepression by a forced credit expansion is toattempt to cure the evil by the very means whichbrought it about; because we are suffering froma misdirection of production, we want to createfurther misdirection — a procedure which can only

21

PREFACE

lead to a much more severe crisis as soon as thecredit expansion comes to an end. It would not bethe first experiment of this kind which has beenmade. We should merely be repeating, on a muchlarger scale, the course followed by the FederalReserve system in 1927, an experiment whichMr. A. C. Miller, the only economist on the FederalReserve Board and, at the same time, its oldestmember, has rightly characterized as 'the greatestand boldest operation ever undertaken by the Fede-ral reserve system', an operation which 'resultedin one of the most costly errors committed by itor any other banking system in the last 75 years\It is probably to this experiment, together withthe attempts to prevent liquidation once the crisishad come, that we owe the exceptional severityand duration of the depression. We must notforget that, for the last six or eight years, monetarypolicy all over the world has followed the adviceof the stabilizers. It is high time that their influ-ence, which has already done harm enough, shouldbe overthrown.

We cannot hope for the overthrow of thisalluringly simple theory until its theoretical basis

22

PREFACE

is definitely refuted and something better substi-tuted for it. The opponents of the stabilizationprogramme still labour — and probably alwayswill labour — under the disadvantage that theyhave no equally simple and clear-cut rule topropose; perhaps no rule at all which will satisfythe eagerness of those who hope to cure all evilsby authoritative action. But whatever may be ourhope for the future, the one thing of which we mustbe painfully aware at the present time —a factwhich no writer on these problems should fail toimpress upon his readers — is how little we reallyknow of the forces which we are trying to influenceby deliberate management; so little indeed that itmust remain an open question whether we wouldtry if we knew more.

FRIEDRICH A. HAYEK

The London School of EconomicsJune 1932

THE PROBLEM OF THE TRADECYCLE

CHAPTER I

THE PROBLEM OF THE TRADE

CYCLE

i

ANY attempt either to forecast the trend ofeconomic development, or to influence it bymeasures based on an examination of existingconditions, must presuppose certain quite definiteconceptions as to the necessary course of economicphenomena. Empirical studies, whether they areundertaken with such practical aims in view, orwhether they are confined merely to the amplifica-tion, with the aid of special statistical devices, ofour knowledge of the course of particular phasesof trade fluctuations, can, at best, afford merely averification of existing theories; they cannot, inthemselves, provide new insight into the causesor the necessity of the Trade Cycle.

This view has been stated very forcibly by

Professor A. L6we.#

* In his essay: Wie ist Konjunkturtheorie uberhaupt mcglich? Welt-wirtschaftliches Archiv, vol. 24. 1926 part 2, p. 166.

27

MONETARY THEORY AND THE TRADE CYCLE

'Our insight into the theoretical interconnectionsof economic cycles, and into the structural laws ofcirculation', he says, 'has not been enriched at allby descriptive work or calculations of correlations.'We can entirely agree with him, moreover, whenhe goes on to say that 'to expect an immediatefurtherance of theory from an increase in empiricalinsight is to misunderstand the logical relationshipbetween theory and empirical research'.

The reason for this is clear. The means ofperception employed in statistics are not the sameas those employed in economic theory; and it istherefore impossible to fit regularities establishedby the former into the structure of economic lawsprescribed by the latter. We cannot superimposeupon the system of fundamental propositionscomprised in the theory of equilibrium, a TradeCycle theory resting on unrelated logical founda-tions. All the phenomena observed in cyclicalfluctuations, particularly price formation and itsinfluence on the direction and the volume ofproduction, have already been explained by thetheory of equilibrium; they can only be integratedas an explanation of the totality of economic

28

THE PROBLEM OF THE TRADE CYCLEevents by means of fundamentally similar con-structions. Trade Cycle theory itself is onlyexpected to explain how certain prices are deter-mined, and to state their influence on productionand consumption; and the determining conditionsof these phenomena are already given byelementary theory. Its special task arises fromthe fact that these phenomena show empiricallyobserved movements for the explanation of whichthe methods of equilibrium theory are as yetinadequate. One need not go so far as to say thata successful solution could be reached only inconjunction with a positive explanation ofelementary phenomena; but no further proof isneeded that such a solution can only be achievedin association with, or by means of, a theorywhich explains how certain prices or certainuses of given goods are determined at all. It isnot only that we lack theories which fulfil thiscondition and which fall outside the categorybest described as 'equilibrium theories' * —

*Cf. Lowe: Der gegenwdrtige Stand der Konjunkiurtheorie inDeutsch-land, Die Wirtschaftswissenschaft nach dem Kriege, Festgabe furLujo Brentano zum 80. Geburtstag, vol. ii, p. 360.

29


theories which are characterized by taking thelogic of economic action as their starting point;the point is rather that statistical method isfundamentally unsuited to this purpose. Just asno statistical investigation can prove that a givenchange in demand must necessarily be followed bya certain change in price, so no statistical methodcan explain why all economic phenomena presentthat regular wave-like appearance which weobserve in cyclical fluctuations. This can beexplained only by widening the assumptions onwhich our deductions are is based, so that cyclicalfluctuations would follow from these as a necessaryconsequence, just as the general propositions ofthe theory of price followed from the narrowerassumptions of equilibrium theory.

But even these new assumptions cannot beestablished by statistical investigation. The statis-tical approach, unlike deductive inference, leavesthe conditions under which established economicrelations hold good fundamentally undetermined;and similarly, the objects to which they relatecannot be determined as unequivocally as bytheory. Empirically established relations between

3°

THE PROBLEM OF THE TRADE CYCLE

various economic phenomena continue to presenta problem to theory until the necessity for theirinterconnections can be demonstrated indepen-dently of any statistical evidence.* The conceptson which such an explanation is based willbe quite different from those by which statisticalinterconnections are demonstrated; they canbe reached independently. Moreover, thecorroboration of statistical evidence provides,in itself, no proof of correctness. A priori wecannot expect from statistics anything more thanthe stimulus provided by the indication of newproblems.

In thus emphasizing the fact that Trade Cycletheory, while it may serve as a basis for statistical

• Cf. the excellent analysis given by E. Altschul in his well-knownessay Konjunkturtheorie und Konjunkturstatistik (Archiv fur Sozial-wissenschaft und Sozialpolitik, vol. 55, Tubingen, 1926). Altschulas a statistician deserves especial credit when, recognizing thelimitations of statistical methods, he writes (p. 85) 'In economicsespecially, the final decision about the significance of a certainphenomenon can never be left to mathematical and statistical analysis.The main approach to research must necessarily lie through theoretic-ally obtained knowledge.' Cf. also A. C. Pigou, Industrial Fluctua-tions, 2nd ed. (London, 1929) p. 37, 'The absence of statistical corre-lation between a given series of changes and industrial fluctuationsdoes not by itself disprove - and its presence does not prove - thatthese changes are causes of the fluctuations.'


research, can never itself be established by thelatter, it is by no means desired to deprecate thevalue of the empirical method. On the contrary,there can be no doubt that Trade Cycle theorycan only gain full practical importance throughexact measurement of the actual course of thephenomena which it describes. But before wecan examine the question of the true importanceof statistics to theory, it must be clearly recognizedthat the use of statistics can never consist in adeepening of our theoretical insight.

11

Even as a means of verification, the statisticalexamination of the cycles has only a very limitedvalue for Trade Cycle theory. For the latter —as for any other economic theory — there are onlytwo criteria of correctness. Firstly, it must bededuced with unexceptionable logic from the funda-mental notions of the theoretical system; andsecondly, it must explain by a purely deductivemethod those phenomena with all their peculiar-

32


ities which we observe in the actual cycles.* Sucha theory could only be 'false' either through aninadequacy in its logic or because the phenomenawhich it explains do not correspond with theobserved facts. If, however, the theory is logicallysound, and if it leads to an explanation of the givenphenomena as a necessary consequence of thesegeneral conditions of economic activity, then thebest that statistical investigation can do is to showthat there still remains an unexplained residue ofprocesses. It could never prove that the deter-mining relationships are of a different characterfrom those maintained by the theory.f

It might be shown, for instance, by statisticalinvestigation that a general rise in prices is followed

• Professor A. Lowe, in his report Uber den Einfluss monetarerFaktoren auf den Konjunkturzyklus Schriften des Vereins fur Sozial-politik, (vol. 173, part 2, p. 357) expresses his views in almost thesame words. The above sentences first appeared in another articlein the same volume.

f Cf. the analysis concerning 'Argument der Wirklichkeitswidrig-keit' in the recent book of E. Carell, Sozialokonomische Theorieund Konjunkturproblem\ Munchen and Leipzig, 1929, for a veryacute methodological argument. He opposes the thesis of L6we (whichremains, however, despite his analysis, the basis of my own work)that the incorporation of cyclical phenomena into the system ofeconomic equilibrium theory, with which they are in apparentcontradiction, remains the crucial problem of Trade Cycle theory.

c 33


by an expansion of production, and a general fall inprices by a diminution of production; but thiswould not necessarily mean that theory shouldregard the movement of price as an independentcause of movements of production. So long as atheory could explain the regular occurrence ofthis parallelism in any other way, it could not bedisproved by statistics, even if it maintainedthat the connection between the two phenomenawas of a precisely opposite nature.* It istherefore only in a negative sense that it ispossible to verify theory by statistics. Either sta-tistics can demonstrate that there are phenomenawhich the theory does not sufficiently explain,or it is unable to discover such phenomena. Itcannot be expected to confirm the theory in apositive sense. The possibility is completely ruledout by what has been said above, since it wouldpresuppose an assertion of necessary interconnec-

• A well-known instance of such an apparent contradiction betweena correct theoretical assertion and experience is the connectionbetween the level of interest rates and the movement of prices. Cf.Wicksell, Vorlesungen, vol. ii, 'Geld und Kredit,' Jena 1922. See alsomy essay, 'Das intertemporale Gleichgettnchts-svstem der Preiseund die Bewegungen des Geldtveries, (Weltwirtschaftiches Archiv>vol. 28 Jena 1928, p. 63 et seq.).

34


tions, such as statistics cannot make. There is noreason to be surprised, therefore, that althoughnearly all modern Trade Cycle theories use statis-tical material as corroboration, it is only where agiven theory fails to explain all the observedphenomena that this statistical evidence can beused to judge its merits.

I l l

Thus it is not by enriching or by checking theo-retical analysis that economic statistics gain theirreal importance. This lies elsewhere. The propertask of statistics is to give us accurate informationabout the events which fall within the provinceof theory, and so to enable us not only to connecttwo consecutive events as cause and effect, a pos-teriori, but to grasp existing conditions completelyenough for forecasts of the future and, eventually,appropriate action, to become possible. It isonly through this possibility of forecasts of sys-tematic action that theory gains practical impor-

35


tance.* A theory might, for instance, enable us toinfer from the comparative movements of certainprices and quantities an imminent change in thedirection of those movements: but we should havelittle use for such a theory if we were unable toascertain the actual movements of the phenomenain question. With regard to certain phenomena

• It should be noted that the idea of forecasting is by no means anew one, although it is often regarded as such. Every economictheory, and indeed all theory of whatever sort, aims exclusively atforetelling the necessary consequences of a given situation, event ormeasure. The subject-matter of trade cycle theory being what it is,it follows that ideally it should result in a collective forecast showingthe total development resulting from a given situation under givenconditions. In practice, such forecasts are attempted in too uncondi-tional a form, and on an inadmissibly over-simplified basis; and,consequently, the very possibility of scientific judgments about futureeconomic trends to-day appears problematical, and cautious thinkersare apt to disparage any attempt at such forecasting. In contrast tothis view, we have to emphasize very strongly that statistical researchin this field is meaningless except in so far as it leads to a forecast,however much that forecast may have to be hedged about withqualifications. In particular any measures aimed at alleviating theTrade Cycle (and necessarily based on statistical research) must beconceived in the light of certain assumptions as to the future trend tobe expected in the absence of such measures. Statistical research,therefore, serves only to furnish the bases for the utilization of existingtheoretical principles. Dr. O. Morgenstern's recent categorical denial(Wirtschqftsprognose, Untersuchung ihrer Voranssetzungen und MGglich-keiten, Wien, 1928) of the possibility of forecasting seems to bedue only to the fact that he demands more from forecasting thanis justifiable. Even the ability to forecast a hailstorm would notbe useless - but, on the contrary, very valuable - if the latter couldthereupon be averted by firing rockets at the clouds 1

36


having an important bearing on the Trade Cycle,our position is a peculiar one. We can deducefrom general insight how the majority of peoplewill behave under certain conditions; but theactual behaviour of these masses at a givenmoment, and therefore the conditions to which ourtheoretical conclusions must be applied, can onlybe ascertained by the use of complicated statisticalmethods. This is especially true when a phe-nomenon is influenced by a number of partly knowncircumstances, such as, e.g., seasonal changes.Here very complicated statistical investigationsare needed to ascertain whether these circum-stances whose presence indicates the applicabilityof theoretical conclusions were in fact operative.Often statistical analysis may detect phenomenawhich have, as yet, no theoretical explanation,and which therefore necessitate either an extensionof theoretical speculation or a search for newdetermining conditions. But the explanation ofthe phenomena thus detected, if it is to serve as abasis for forecasts of the future, must in every caseutilize other methods than statistically observedregularities; and the observed phenomena will

37


have to be deduced from the theoretical system,independently of empirical detection.

The dependence of statistical research on pre-existing theoretical explanation hardly needs fur-ther emphasis. This holds good not only as regardsthe practical utilization of its results, but also in thecourse of its working, in which it must look totheory for guidance in selecting and delimiting thephenomena to be investigated. The oft-repeatedassertion that statistical examination of the TradeCycle should be undertaken without any theoreticalprejudice is therefore always based on self-deception.*

• Prof. Bullock, the Principal of the Harvard Economic Service(now the Harvard Economic Society) constantly emphasizes the com-plete absence of theoretical prepossession with which the work of the In-stitute is carried out. Sincere as this belief unquestionably is, however,one may doubt its validity when one reads, for instance, the followingaccount given by Prof. Bullock's chief collaborator, Mr. W. M.Persons, the inventor of the famous Harvard Barometer. Here heattempts the following popular explanation of the latter:—'Thisaccount of the business cycle, based upon our statistical analysis,revolves about the fluctuation of short-time interest rates, speculation,and business. We may think of interest rates as varying inverselywith the amount of the bank reserves in the credit reservoir. The flowin the supply pipe to this reservoir depends upon the volume of goldimports, gold production, and the volume of paper currency. Thereare two outlets from this reservoir of credit: one pipe furnishes creditfor speculation in securities; the other pipe is for the flow of credit

38

THE PROBLEM OF THE TRADE CYCLEOn the whole, one can say without exaggeration

that the practical value of statistical researchdepends primarily upon the soundness of thetheoretical conceptions on which it is based.To decide upon the most important problems ofthe Trade Cycle remains the task of theory; andwhether the money and labour so freely expendedon statistical research in late years will be repaid bythe expected success depends primarily on whether /the development of theoretical understanding keepspace with the exploration of the facts. For wemust not deceive ourselves: not only do we nowlack a theory which is generally accepted by

into business. When the level of credit in the reservoir is high, andperhaps the outlet to business is partially clogged, the flow of fundsinto speculation begins. After this flow goes on for some time, how-ever, and the flow into business increases, the level of credit in thereservoir falls. Obstruction is offered to the flow into speculativemarkets by the devices of higher interest rates and direct discrimina-tion against speculation and in favour of business. The outlet intospeculation therefore becomes clogged but the flow into business goeson. The level in the reservoir becomes still lower until the time isreached when bankers consider it dangerous to allow the outflowto continue. We then have a halt in further credit expansion, or to useour illustration, both outlets are clogged for a time and bank reservesare brought back to normal by allowing the supply to again fill thetank.' ('A non-technical explanation of the index of general businessconditions,' The Review of Economic Statistics, prel. vol. ii, Cam-bridge, 1920. p. 47).

39

MONETARY THEORY AND THE TRADE CYCLEeconomists, but we do not even possess onewhich could be formulated in such an unexcep-tionable way, and worked out in such detail, aseventually to command such acceptance. A seriesof important interconnections have been estab-lished and some principles of the greatest signifi-cance expounded; but no one has yet undertakenthe decisive step which creates a complete theoryby using one of these principles to incorporate allthe known phenomena into the existing systemin a satisfactory way. To realize this, of course,does not hinder us from pursuing either economicresearch or economic policy; but then we mustalways remember that we are acting on certaintheoretical assumptions whose correctness has notyet been satisfactorily established. The 'practicalman* habitually acts on theories which he does notconsciously realize; and in most cases this meansthat his theories are fallacious. Using a theoryconsciously, on the other hand, always results insome new attempt to clear up the interrelationswhich it assumes, and to bring it into harmony withwhich theoretical assumptions; that is, it results inthe pursuit of theory for its own sake.

40


IV

The value of business forecasting depends uponcorrect theoretical concepts; hence there can, atthe present time, be no more important task inthis field than the bridging of the gulf whichdivides monetary from non-monetary theories.*This gulf leads to differences of opinion in thefront rank of economists; and is also the charac-teristic line of division between Trade Cycletheory in Germany and in America — wherebusiness forecasting originated. Such an analysisof the relation between these two maintrends seems to me especially important because

* Since the publication of the German edition of this book, I havebecome less convinced that the difference between monetary andnon-monetary explanations is the most important point of disagreementbetween the various Trade Cycle theories. On the one hand, itseems to me that within the monetary group of explanations thedifference between those theorists who regard the superficial pheno-mena of changes in the value of money as decisive factors in determin-ing cyclical fluctuations, and those who lay emphasis on the realchanges in the structure of production brought about by monetarycauses, is much greater than the difference between the latter groupand such so-called non-monetary theorists as Prof. Spiethoff and Prof.Cassel. On the other hand, it seems to me that the difference betweenthese explanations, which seek the cause of the crisis in thescarcity of capital, and the so-called 'under-consumption' theories, istheoretically as well as practically of much more far-reaching import-ance than the difference between monetary and non-monetary theories.

41

MONETARY THEORY AND THE TRADE CYCLEof the peculiar position of the monetary theories.Largely through the fault of some of their best-known advocates in Germany, monetary explana-tions became discredited, and their essentialshave, moreover, been much misunderstood; while,on the other hand, the reaction against them formsthe main reason for the prevailing scepticism asto the possibility of any economic theory of theTrade Cycle — a scepticism which may seriouslyretard the development of theoretical research.*

There is a fundamental difficulty inherent inall Trade Cycle theories which take as their start-ing point an empirically ascertained disturbanceof the equilibrium of the various branches of pro-duction. This difficulty arises because, in statingthe effects of that disturbance, they have to makeuse of the logic of equilibrium theory .f Yet thislogic, properly followed through, can do no more

• Cf. the above-mentioned essay of A. Lowe in the 'Weltwirt-schaftliches Archiv.'

f By 'equilibrium theory' we here primarily understand themodern theory of the general interdependence of all economicquantities, which has been most perfectly expressed by the LausanneSchool of theoretical economics. The significant basic concept of thistheory was contained in James Mill's and J. B. Say's Theorie desDebouches. Cf. L. Miksch, Gibt es eine allgemeine Uberproduktion?Jena, 1929.

42


than demonstrate that such disturbances of equi-librium can come only from outside — i.e. thatthey represent a change in the economic data— and that the economic system always reactsto such changes by its well-known methodsof adaptation, i.e. by the formation of a newequilibrium. No tendency towards the specialexpansion of certain branches of production, how-ever plausibly adduced, no chance shift in demand,in distribution or in productivity, could adequatelyexplain, within the framework of this theoreticalsystem, why a general 'disproportionality' betweensupply and demand should arise. For the essentialmeans of explanation in static theory, which is, atthe same time, the indispensable assumption forthe explanation of particular price variations, isthe assumption that prices supply an automaticmechanism for equilibrating supply and demand.

The next section will deal with these difficultiesin more detail: a mere hint should therefore besufficient at this point. At the moment we haveonly to draw attention to the fact that the problembefore us cannot be solved by examining the effectof a certain cause within the framework, and by

43


the methods, of equilibrium theory. Any theorywhich limits itself to the explanation of empiricallyobserved interconnections by the methods ofelementary theory necessarily contains a self-contradiction. For Trade Cycle theory cannotaim at the adaptation of the adjusting mechanismof static theory to a special case; this scheme ofexplanation must itself be extended so as to explainhow such discrepancies between supply and de-mand can ever arise. The obvious, and (to mymind) the only possible way out of this dilemma, isto explain the difference between the course ofevents described by static theory (which onlypermits movements towards an equilibrium, andwhich is deduced by directly contrasting the supplyof and the demand for goods) and the actual courseof events, by the fact that, with the introductionof money (or strictly speaking with the introduc-tion of indirect exchange), a new determiningcause is introduced. Money being a commoditywhich, unlike all others, is incapable of finallysatisfying demand, its introduction does awaywith the rigid interdependence and self-sufficiencyof the 'closed* system of equilibrium, and makes

44


possible movements which would be excludedfrom the latter. Here we have a starting-pointwhich fulfils the essential conditions for any satis-factory theory of the Trade Cycle. It shows, in apurely deductive way, the possibility and thenecessity of movements which do not at any givenmoment tend towards a situation which, in theabsence of changes in the economic 'data', couldcontinue indefinitely. It shows that, on the con-trary, these movements lead to such a 'dispro-portionality' between certain parts of the systemthat the given situation cannot continue.

But while it seems that it was a sound instinctwhich led economists to begin by looking on themonetary side for an explanation of cyclicalfluctuations, it also seems probable that the one-sided development of the theory of money has, asyet, prevented any satisfactory solution to theproblem being found. Monetary theories of theTrade Cycle succeeded in giving prominence tothe right questions and, in many cases, madeimportant contributions towards their solution;but the reason why an unassailable solution hasnot yet been put forward seems to reside in the

45


fact that all the adherents of the monetary theoryof the Trade Cycle have sought an explanationeither exclusively or predominantly in the super-ficial phenomena of changes in the value of money,while failing to pursue the far more profound andfundamental effects of the process by which moneyis introduced into the economic system, as distinctfrom its effect on prices in general. Nor did theyfollow up the consequences of the fundamentaldiversity between a money economy and thepure barter economy which is assumed in statictheory.*

Naturally it cannot be the business of thisessay to remove all defects and deficiencies fromthe monetary theories of the Trade Cycle, or todevelop a complete and unassailable theory.In these pages I shall only attempt to show thegeneral significance for this theory of the monetarystarting-point, and to refute the most important

• Similar views are expressed by W. Ropke, Kredit und Konjunktur.Jahrbiicher fiir Nationaldkonomie und Statistik, 3rd series, vol. 69, pp,264 et seq.

46


objections raised against the monetary explanationby proving that certain rightly exposed deficienciesof some monetary theories do not necessarilyfollow from the monetary approach. All that iswanted, therefore, is, first, a proof, using as ourexamples some of the best-known non-monetarytheories, that the 'real' explanations adduced bythem do not, in themselves, suffice to build up acomplete and consistent theory; secondly, ademonstration that the existing monetary theoriescontain the germ of a true explanation, although allsuffer, more or less, from that over-simplificationof the problem which results from reducing allcyclical fluctuations to fluctuations in the valueof money; finally that the monetary starting-pointmakes it possible, in fact, to show deductivelythe inevitability of fluctuation under the existingmonetary system and, indeed, under almost anyother which can be imagined. It will be shown, inparticular, that the Wicksell-Mises theory of theeffects of a divergence between the 'natural' andthe money rate of interest already contains themost important elements of an explanation, andhas only to be freed from any direct reference to a

47


purely imaginary 'general money value* (as hasalready been partly done by Prof. Mises) in orderto form the basis of a Trade Cycle theorysufficing for a deductive explanation of all theelements in the Trade Cycle.

NON-MONETARY THEORIESOF THE TRADE CYCLE

C H A P T E R I I

NON-MONETARY THEORIESOF THE TRADE CYCLE

i

ANY attempt at a general proof, within the compassof a short essay, of the assertion that non-monetarytheories of the Trade Cycle inevitably suffer froma fundamental deficiency, appears to be confrontedwith an insuperable obstacle by reason of the verymultiplicity of such theories. If it were necessaryfor our purpose to show that every one of thenumerous disequilibrating forces which have beenmade starting-points for Trade Cycle theorieswas, in fact, non-existent, then the conditions ofour success would, indeed, be impossible offulfilment; for not only would it be almost impos-sible to deal with all extant theories but no conclu-sive answer could result, seeing that we should stillhave to reckon with a new and hitherto unrefutedcrop of such theories in the future. Moreover^

5*


the existence of most of the interconnectionselaborated by the various Trade Cycle theories canhardly be denied, and our task is rather theirco-ordination in a unified logical structure thanthe development of entirely new and differenttrains of thought. In fact, it is by no meansnecessary to question the material correctness ofthe individual interconnections emphasized in thevarious non-monetary theories in order to showthat they do not afford a sufficient explanation.As has already been indicated in the first chapter,none of them is able to overcome the contradictionbetween the course of economic events as describedby them and the fundamental ideas of the theo-retical system which they have to utilize in order toexplain that course. It will, therefore, be sufficientto show, by examination of some of the best-knowntheories, that they do not answer this fundamentalquestion; nor can they ever do so by their presentmethods and by reference to the circumstanceswhich they now regard as relevant to Trade-Cycle theory. When, however, the question isanswered on different lines, viz., by reference tomonetary circumstances, it can be shown that

52

NON-MONETARY THEORIES OF TRADE CYCLE

the elements of explanation adduced by differenttheories lose their independent importance andfall into a subordinate position as necessaryconsequences of the monetary cause.

It is rather difficult to select the main types ofTrade-Cycle theory for this purpose, since wehave no theoretically satisfactory classification. Thelatest attempts at such classifications, by Mr. W. M.Persons,* Professor W. C. Mitchell,f and Mr. A.H. Hansen,J show that the usual division, whichrelies on external features and hardly touches thesolution of fundamental problems, gives far toowide a scope for arbitrary decisions. As ProfessorLowe§ has correctly emphasized (and as should beobvious from what has been said above) the onlyclassification which could be really unobjectionablewould be one which proceded according to themanner in which such theories explain the absenceof the 'normal course' of economic events, as pre-

•'Theories of Business Fluctuations* {Quarterly Journal of Economics,vol. xli, p. 923).

t Business Cycles: the Problem and its Setting (New York 1927).% Business Cycle Theory: its Development and present Status

(Boston, 1927).§ See Der GegenwartigeStandder Konjunkturforschung, p. 359 et seq.

53


sented by static theory. In fact, the various theories—as we shall hope to show later — make no attemptwhatever to do this. As there is, therefore, noclassification which would serve our purpose,our choice must be more or less arbitrary; but bychoosing some of the best known theories andexemplifying the train of thought to which ourobjection particularly applies, we should beable to make the general validity of the lattersufficiently clear. The task is made rather easierby the fact that there does exist to-day, on at leastone point, a far-reaching agreement among thedifferent theories. They all regard the emergenceof a disproportionality among the various produc-tive groups, and in particular the excessive pro-duction of capital goods, as the first and main thingto be explained. The development of theory owesa real debt to statistical research in that, to-day,there is at least no substantial disagreement as tothe thing to be explained.

There is, however, a point to be emphasizedhere. The modern habit of going beyond theactual crisis and seeking to explain the entire cycle,suffers inherently from the danger of paying less

54

NON-MONETARY THEORIES OF TRADE CYCEL

and less attention to the crucial problem. Inparticular, the attempt to give the object of thetheory as neutral a name as possible (such as'Industrial fluctuations' or * Cyclical movements ofIndustry') threatens to drive the real theoreticalproblem more into the background than was thecase in the old theory of crisis. The simple factthat economic development does not go on quiteuniformly, but that periods of relatively rapidchange alternate with periods of relative stagnation,does not in itself constitute a problem. It is suf-ficiently explained by the adjustment of the eco-nomic system to irregular changes in the data —changes whose occurrence we always have toassume and which cannot be further explained byeconomic science. The real problem presented toeconomic theory is: Why does not this adjustmentcome about smoothly and continuously, just as anew equilibrium is formed after every change inthe data? Why is there this temporary possibilityof developments leading away from equilibriumand finally, without any changes in data, necessita-ting a change in the economic trend? The phe-nomena of the upward trend of the cycle and of the

55


culminating boom constitute a problem only be-cause they inevitably bring about a slump in sales— i.e. a falling-off of economic activity — whichis not occasioned by any corresponding change inthe original economic data.

11

The prevailing disproportionality theories arein agreement in one respect. They all see thecause of the slump in the fact that, during theboom, for various reasons, the productive appara-tus is expanded more than is warranted by thecorresponding flow of consumption; there finallyappears a scarcity of finished consumption goods,thus causing a rise in the price of such goodsrelatively to the price of production goods (whichamounts to the same thing as a rise in the rate ofinterest) so that it becomes unprofitable to employthe enlarged productive apparatus or, in manycases, even to complete it. At present there ishardly a recognized theory which does not givethis idea, which we only sketch for the moment,*

•Cf. below, p. 212 et seq. esp. p. 217.

56


a decisive place in its argument, and we shouldtherefore be well advised to begin by seeing howthe various theories try to deal with the phenome-non in question. Apart from the monetary theories,which, as will be shown later, can only be consideredsatisfactory if they explain that phenomenon, thereare two groups of explanations which can be entirelydisregarded. In the first place there is nothing tobe gained from an examination of those theorieswhich seek to explain cyclical fluctuations by cor-responding cyclical changes in certain externalcircumstances, while merely using the unquestion-able methods of equilibrium theory to explain theeconomic phenomena which follow from thesechanges. To decide on the correctness of thesetheories is beyond the competence of Economics.In the second place, it is best, for the moment, toexclude from consideration those theories whoseargument depends so entirely on the assumptionof monetary changes that when the latter areexcluded no systematic explanation is left. Thiscategory includes Professor J. Schumpeter,* Pro-

• Theorie der toirUchaftlichen Enturicklung, 2nd edit., Muchenand Leipzig, 1926.

57


fessor E. Lederer,* and Professor G. Cassel,f andto a certain extent Professor W. C. Mitchell andProfessor J. Lescure.J We shall have to considerlater, with regard to this category, how far it istheoretically permissible to treat these monetaryinterconnections as determining conditions on thesame footing as the other phenomena used inexplanation.

It is, of course, impossible at this point to gointo the peculiarities of all types of theory, asworked out by their respective authors. We mustleave out of account the forms in which thevarious explanations are presented, and confineourselves to certain underlying types of theorywhich recur in a number of different guises.Inevitably, this treatment of contemporary theoriesmust fail to do full justice to the intellectualmerits exemplified in each; but for the purposes

• Konjunktur und Krisen: Grundriss der Sozialokonomik, Abt. IV,Teil I. Tubingen 1926; also Zur Morphologie der Krisen in DieWirtschaftstheorie der Gegenwart, cited above, edited by H. Mayer,vol iv, Vienna, 1928.

f Theory of Social Economy.t Des Crises g€n€rales et piriodiques de surproduction. Paris, 1913;

and Krisenlehre in Die Wirtschaftstheorie der Gegenzvart, ed. by H.Mayer, vol. iv, Vienna 1928.

58

NON-MONETARY THEORIES OF TRADE CYCLEof this chapter — that is, to show the fundamentalobjections to which all non-monetary theoriesof the Trade Cycle are open — this somewhatcursory and imperfect treatment may be enough.

We may begin our demonstration by pointingout that all those forms of disproportionalitytheory with which we have to deal here rest onthe existence of quite irregular fluctuations of'economic data' (that is, the external determiningcircumstances of the economic system, includinghuman needs and abilities). From this assumption,they try to explain in one way or another that thefluctuations in consumption or some other elementin the economic system occasioned by thesechanges are followed by relatively greater changesin the production of production goods.* Thesewide fluctuations in the industries making pro-duction goods bring about a disproportionalitybetween them and the consumption industries

• It should be noted here that the assumption of initial changesin the economic data, which no theory of the Trade Cycle can dispensewith, in itself throws no light on the proper way of explaining cyclicalfluctuations. It is not the occurrence of disturbances of equilibrium,necessitating readjustment, which presents a problem to Trade Cycletheory; it is the fact that this adjustment is brought about only after aseries of movements have taken place which cannot be considered

59


to such an extent that a reversal of the movementbecomes necessary. It is not, therefore, the simplefact of fluctuation in the production of capital goods(which is certainly inevitable in the course ofeconomic growth) which has to be explained.The real problem is the growth of excessivefluctuations in the capital goods industries out ofthe inevitable and irregular fluctuations of therest of the economic system, and the dispro-portional development, arising from these, of thetwo main branches of production. We can dis-tinguish three main types of non-monetarytheories explaining the exaggerated effect ofgiven fluctuations on capital goods industries.The most common, at the moment, are thoseexplanations which try to show that, on accountof the technique of production, an increase in thedemand for consumption goods, whether expected

'adjustments' in the sense used by the theory of economic equilibrium.'The phenomenon is never made clear until it is explained why itscause, whatever it may be, does not call forth a continuous equilibrat-ing process' (Prof. J. Schumpeter, Theorie der wirtschaftlichenEntwicklung, 2nd ed., Miinchen and Leipzig, 1926). These changesof data could serve as a complete explanation only if it could be shownthat the successive phases of the Trade Cycle are conditioned by aseries of such changes, following each other in a certain order.

60


or actual, tends to bring about a relatively largerincrease in the production of goods of a higherorder, either generally or in a certain group ofthese goods. Hardly less common, and differingonly in appearance, are explanations which seekto derive these augmented fluctuations fromspecial circumstances (non-monetary in character)arising in the field of savings and investment.Finally, as a third group, we must mention certainpsychological theories, which, for the most part,have however no pretension to rank as independentexplanations and which merely reinforce otherarguments, and are open to the same objectionsas the two other main types.

in

We shall mention only the most importantof our objections to the first type, which is theeasiest to discuss from this point of view. It iscommon to so many economists that it is hardlynecessary to mention particular representatives.The simplest way of deductively explainingexcessive fluctuations in the production of capital

61


goods is by reference to the long period of timewhich is necessary, under modern conditions, forpreparing the fixed capital goods which enablethe expansion of the productive process to takeplace.* According to a widely held view, thiscircumstance alone is enough to make everyincrease in the sales of consumption goods, whetherbrought about by an intensification of demand orby a fall in the costs of production, capable ofbringing about a more than proportional increasein the production of intermediary goods. This isexplained either by the individual producer'signorance of what his competitors are doing, or —as is common in American writings — by the'cumulative effect' of each change in the sale ofconsumption goods on the higher stages ofproduction. Owing to circumstances which willbe explained later, the leading idea in all thesetypes of explanation is that the long period which,with the present technique of production, elapsesbetween the beginning of a productive process and

• Cf. A. Aftalion: Les crises piriodiques de surproduction, Paris, 1913,Bks. ii-vi, Chaps, in to vm; and D. H. Robertson, Industrial Fluctu-ations, London, 1915, p. 14.

6*


the arrival of its final product at the market,prevents the gradual adjustment of production tochanges in demand through the agency of pricesand makes it possible, from time to time, for anexcessively large supply to be thrown on themarket. This idea is supported by another, whichhowever, can be independently and more widelyapplied; that is, that every change in demand\from the moment of its appearance, propagatesitself cumulatively through all the grades ofproduction, from the lowest to the highest. Thiscumulative effect arises because at each stage,besides the change which would be appropriateto the actual shift in demand, another changearises from the adjustment of stocks and ofproductive apparatus to the alteration in marketconditions.* An increase in the demand forconsumption goods will not merely call fortha proportional increase in the demand for goods ofa higher order: the latter will also be increased bythe amount needed to raise current stocks to a

t Cf. T. N. Carver, Quarterly Journal of Economics, 1903-4, p.492; A. Aftalion, Journal d'Economie Politique, 1909, pp. 215 ff.\W. C, Mitchell, op. cit.; and D. H. Robertson, op. cit., p. iZ2ff.

63


proportional level, and, finally, by the furtheramount by which the requirements for producingnew means of production exceed those for keepingthe existing means of production intact. (Forinstance, an extension of 10 per cent, in oneparticular year, in the machinery of a factory whichnormally renews 10 per cent of its machineryannually, causes an increase of ioo per cent in theproduction of machinery — i.e. a given increasein the demand for consumption goods occasionsa tenfold increase in the production of production-goods.) This idea is offered as an adequatereason not only for the relatively greater fluctua-tions in production-goods industries but also fortheir excessive expansion in periods of boom.Similarly, the extensive use of durable capitalequipment in the modern economy is often singledout for responsibility.* Industries using heavyequipment are prone to excessive expansion inboom periods because small increments in thisequipment are impossible; expansion must neces-sarily take place by sudden jerks. Once the newequipment is available, on the other hand, the

# Cf. D. H. Robertson: op. cit., pp. 31 et seq.

64


volume of production has little influence on totalcosts, which go on even if no production takesplace at all. New inventions and new needs, how-ever, although they are often adduced as explainingthe accelerated and excessive growth of capitalgoods industries, cannot be dealt with on the samefooting. They only represent a special group ofthe many possible causes from which the cumula-tive processes described above may originate.

IV

There is virtually no doubt that all theseinterconnections, and many others which are givenprominence in various Trade Cycle theories andwhich similarly tend to disturb economic equili-brium, do actually exist; and any Trade Cycletheory which claims to be comprehensivelyworked out must take them into consideration.But none of them get over the real difficulty —namely: Why do the forces tending to restoreequilibrium become temporarily ineffective andwhy do they only come into action again when it istoo late? They all try to explain this phenomenon

E 65


by a further, usually tacit, assumption, which oneof the advocates of these theories, Mr. C.O. Hardy,* has himself put forward as theircommon idea, by which, in my opinion, he bringsout, with the utmost clarity, their fundamentalweakness. He states that all those theories whichare based on the length of the production-periodunder modern technical conditions agree inregarding these conditions as a source of difficultyto producers in adjusting production to the stateof the market; producing, as they must, for afuture period, the market possibilities of which arenecessarily unknown to them. He then emphasizesthat in general it is the task of the price-mechanismto adjust supply to demand; he thinks, however,that this mechanism is imperfect, if a longperiod has to lapse between production and thearrival of the product at the market, because'prices and orders give information concerning theprospective state of demand compared with theknown facts of the present and future supply,

• Risk and Risk Bearing, University of Chicago Press, 1923, p. 72.See also Mr. Hardy's reply to the above criticism in the revised 2ndEdition (1931), of the same book (p. 94), which, however, does notseem to solve the fundamental difficulty.

66


but they give no clue to the changes in supplywhich they themselves are likely to cause.'* Hetries to show how periodic over- and under-production may result from an increase in demandacting as an incentive to increased production.He here states explicitly what others assumetacitly, and thus his exposition completely givesaway the question-begging nature of all sucharguments. For he holds that under free com-petition, in the case considered, more and morepeople try to profit by the favourable situation,all ignoring one another's preparations, and 'noforce intervenes to check the continual increase in

production until it reflects itself in declining orders

and falling prices.^ In this statement (accordingto which the price-mechanism comes into actiononly when the products come on to the market,while, until then, producers can regulate theextent of their production solely according to theestimated total volume of demand) the fundamentalerror which can be shown to recur in all thesetheories is plainly revealed. It arises from a mis-conception of the deliberations which regulate

• Op. cit., p. 73, t Ibid. (My italics.)

67


the entrepreneur's actions and of the significanceof the price-mechanism.

If the entrepreneur really had to guide hisdecisions exclusively by his knowledge of thequantitative increase in the total demand for hisproduct, and if the success of economic activitywere really always dependent on that knowledge,no very complicated circumstances would beneeded to produce constant disturbances inthe relation between supply and demand. But theentrepreneur in a capitalist economy is not — asmany economists seem to assume —in the samesituation as the dictator of a Socialist economy.The protagonists of this view seem to overlook thefact that production is generally guided not byany knowledge of the actual size of the totaldemand, but by the price to be obtained in themarket. In the modern exchange economy, theentrepreneur does not produce with a view tosatisfying a certain demand — even if that phraseis sometimes used — but on the basis of a calcula-tion of profitability; and it is just that calculationwhich will equilibrate supply and demand. Heis not in the least concerned with the amount by

68


which, in a given case, the total amount demandedwill alter; he only looks at the price which he canexpect to get after the change in question has takenplace. None of the theories under discussionexplains why these expectations should generallyprove incorrect. (To deduce their incorrectnessfrom the fact that over-production, arising fromfalse expectations, causes prices to fall, would bemere argument in a circle.) Nor can this general-ization be theoretically established by any othermethod. For so long, at least, as disturbingmonetary influences are not operating, we haveto assume that the price which entrepreneursexpect to result from a change in demand or froma change in the conditions of production willmore or less coincide with the equilibrium price.For the entrepreneur, from his knowledge of theconditions of production and the market, willgenerally be in a position to estimate the pricethat will rule after the changes have taken place,as distinct from the quantitative changes in thetotal volume of demand. One can only say, asto this prospective price of the product concerned,that it is just as likely to be lower than the equili-

69


brium price as to be higher and that, on the average,it should more or less coincide, since there is noreason to assume that deviations will take placeonly in one direction. But this prospective priceonly represents one factor determining the extentof production. The other factor, no less importantbut all too often overlooked, is the price theproducer has to pay for raw materials, labour-power, tools and borrowed capital — i.e. his costs.These prices, taken together, determine theextent of production for all producers operatingunder conditions of competition; and the pro-ducer's decisions as to his production must beguided not only by changes in expectations as tothe price of his product, but also by changes inhis costs. To show how the interplay of theseprices keeps supply and demand, production andconsumption, in equilibrium, is the main object ofpure economics, and the analysis cannot berepeated here in detail. It is, however, the task ofTrade Cycle theory to show under what conditionsa break may occur in that tendency towardsequilibrium which is described in pure analysis— i.e. why prices, in contradiction to the con-

70


elusions of static theory, do not bring aboutsuch changes in the quantities produced as wouldcorrespond to an equilibrium situation. In orderto show that the theories under discussion do notsolve this problem, and only as far as is necessaryfor this purpose, we shall now study the mostimportant of the interconnections which bringabout equilibrium under the assumptions ofstatic theory.

We may attempt this task by asking whatkind of reactions will be brought about by theoriginal change in the economic data, which issupposed to cause the excessive extension of theproduction of capital goods, and how, in suchcases, a new equilibrium can result. Whether theoriginal impetus comes from the demand side orthe supply side, the assumption from which wehave to start is always a price —or rather anexpected price — which renders it profitable, underthe new conditions, to extend production. Asstated above, we can assume — since none of the

71


theories in question give any reason to thecontrary — that this expected price will approxi-mate to the new equilibrium price. We can assume,that is, that if the impetus is a fall in unit costs, theproducer will consider the effects of an increasedsupply; if the impetus is an increase in demand, hewill consider the increase in the cost per unitfollowing the increase in the quantity produced.The existence of a general misconception in thisrespect would require a special explanation, andunless this is to rest on a circular argument, itcan only be accounted for by a monetary explana-tion, which we cannot consider at this point.

Now the length of time required to producemodern means of production cannot induce atendency to an excessive extension of the pro-ductive apparatus; or, more accurately, any suchtendency is bound to be effectively eliminatedby the increase in price of the factors of production.Thus we cannot give a sufficient explanation forthe occurrence of the disproportionality in theseterms. This becomes obvious as soon as we dropthe assumption that the price-mechanism beginsto function only from the moment at which the

72

NON-MONETARY THEORIES OF TRADE CYCLEincreased supply comes on the market, andconsider that whenever the price obtainable for thefinished product is correctly estimated, the adjust-ment of the prices of factors of production mustensure that the amount produced is limited towhat can be sold at remunerative prices. Themere existence of a lengthy production periodcannot be held to impair the working of the pricemechanism, so long, at any rate, as no additionalreason can be given for the occurrence of a generalmiscalculation in the same direction concerningthe effect of the original change in data on theprices of the products.

We must next inquire what truth there is inthe alleged tendency towards a cumulative pro-pagation of the effect of every increase (or decrease)of demand from the lower to the higher stages ofproduction. The arguments given below againstthis frequently-adduced theory must serve at thesame time to refute all other theories based onsimilar technical considerations; for space will notpermit us to go into every one of these, and thereader can be trusted to apply the same reasoningas is employed in this demonstration to all similar

73

MONETARY THEORY AND THE TRADE CYCLEexplanations — such as those based on thenecessary discontinuity of the extension of pro-ductive apparatus. Does the cumulative effectof every increase in demand represent a newprice-determining factor, as a result of whichprices, and therefore quantities produced, willbe different from those needed to achieve equili-brium? Is the regulating effect of prices on theextent of production really suspended by thefact that when turnover increases merchants tryto increase stocks, and manufacturers to extendproduction? If the increase in the prices of pro-duction goods were the only counterbalancingfactor to set against the increase in the demandfor these goods, it would still be possible for moreinvestments to be undertaken than would provepermanently profitable. According to the view weare considering, there will be an increase in thequantity of factors of production demanded at anyprice, as compared with the equilibrium situation,and therefore it would appear possible that, atevery price at which producers still think they canprofitably make use of this quantity, investmentswill be undertaken to an unwarrantable extent.

74


This way of stating the position, however,entirely overlooks the fact that every attempt toextend the productive apparatus must necessarilybring about, besides the rise in factor prices, afurther checking force — viz. a rise in the rateof interest. This greatly strengthens the effectof the rise in factor prices. It makes a greatermargin between factor-prices and product-pricesnecessary just when this margin threatens todiminish. The maintenance of equilibrium isthus further secured.

For we must not forget that not only the volumeof current production, but also the size, at anygiven moment, of the productive apparatus,(including stocks, which cannot be omitted) isregulated through prices, and especially — apartfrom the above-mentioned prices for goods andservices —by the price paid for the use of capital,that is, interest. Whatever particular explanationof interest we may accept, all contemporarytheories agree in regarding the function of interestas one of equalizing the supply of capital and thedemand arising in various branches of production.Until some special reason can be adduced why

75


it should not fulfil this function in any givencase, we have to assume, in accordance with thefundamental thesis of static theory, that it alwayskeeps the supply of capital goods in equilibriumwith that of consumption goods. This assumptionis just as indispensable, and just as inevitable,as a starting-point, as the main assumption thatthe supply of and demand for any kind of goodswill be equilibrated by movements in the pricesof those goods. In our case, when we are consider-ing a tendency to enlarge the productive apparatusand the size of stocks, this function must be per-formed in such a way as to increase the rate ofinterest, and hence the necessary margin of profitbetween the price of the products and that of themeans of production. This, however, automaticallyexcludes that part of the increase in the demandfor productive goods, which would have beensatisfied despite the increase in their prices ifthe rise in the rate of interest had not taken place.None of the various Trade Cycle theories based onsome alleged peculiarity of the technique of pro-duction can even begin to explain why theequilibrium position, determined by the various

76


above-mentioned processes of price formation,

should be reached at a different point from

where it would be without these peculiarities.Now as regards the prices of goods and services

used for productive purposes, there seems to beno reason why they should not fulfil their functionof equilibrating supply and demand. For supplyand demand are here in direct relation with oneanother, so that any discrepancy which may arisebetween them, at a given price must, directly andimmediately, lead to a change in that price. Onlywhen we come to consider the second group ofprices (those paid for borrowed capital or, inother words, interest) is it conceivable thatdisturbances might creep in, since, in this case,price formation does not act directly, by equalizingthe marginal demand for and supply of capitalgoods, but indirectly, through its effect on moneycapital, whose supply need not correspond tothat of real capital. But the process by whichdivergences can arise between these two is leftunexplained by all the theories with which wehave hitherto dealt. Yet before going on to seehow far interest may present such a breach in the

77


strict system of equilibrium as may serve to explaincyclical disturbances, we must briefly examinethe explanation offered by the second importantgroup of non-monetary theories, which attemptto explain the origin of periodical disturbancesof equilibrium purely through the phenomenaarising out of the accumulation and investmentof saving.*

VI

The earlier versions of these theories startfrom the groundless and inadmissible assump-tion that unused savings are accumulated for atime and then suddenly invested, thus causingthe productive apparatus to be extended in jerks.Such versions can be passed over without furtheranalysis. For one thing, it is impossible to giveany plausible explanation why unused savingsshould accumulate for a time;f for another thing,

* In revising the above paragraphs my notice has been called to thefact that they are in many respects in accordance with the reasoningof S. Budge in his Grundzuge der theoretischen Nationaldkonomie(Jena, 1925, pp. 201 et seq.) to which I should therefore like to callattention.

j " Cf. the very effective remarks of W. Eucken in his interestingviva-voce report to the Zurich Assembly of the 'Verein fur Sozial-politik'. (Schriften, 175, p. 295 et seq.)

78


even if such an explanation were forthcoming, itwould provide no clue to the disproportionaldevelopment in the production of capital goods.The fact that the mere existence of fluctuationsin saving activity does not in itself explain thisproblem is realized (in contrast to many othereconomists) by the most distinguished exponentof these theories, Prof. A. Spiethoff. This is plainfrom his negative answer to the analogous question,whether in a barter economy an increase in savingcan create the necessary conditions for depression.*Indeed, it is difficult to see how spontaneous vari-ations in the volume of saving (which are notthemselves open to further economic explanation,and must therefore be regarded as changes ofdata) within the limits in which they are actuallyobserved can possibly create the typical disturb-ances with which Trade Cycle theory isconcerned.!

• 'Krisen' in Handwdrterbuch der Staatswissenschaften, 4th cd., vol.vi, 1925, p. 81.

t It is, however, not inconceivable, theoretically, that sudden andviolent fluctuations in the volume of saving might give rise to thephenomena of a crisis during their downward swings. On this point,see below Chap, v, p. 205.

79

MONETARY THEORY AND THE TRADE CYCLEWhere, then, according to these theories, may

we find the reason for this genesis of disequilibrat-ing disturbances in the processes of saving andinvestment? We will keep to the basis of SpiethofFstheory, which is certainly the most completeof its kind. We may disregard his simple referenceto the Complexity of capital relations', for it doesnot in itself provide an explanation. The mainbasis of his explanation is to be found in the follow-ing sentence: * If capitalists and producers of imme-diate consumption goods want to keep theirproduction in step with the supply of acquisitiveloan capital, these processes should be consciouslyadjusted to one another* * But the creation ofacquisitive loan capital ensues independently ofthe production of intermediate goods anddurable capital goods; and conversely, thelatter can be produced without the entrepreneurknowing the extent to which acquisitive capital(i.e. savings) exists and is available for investment;

* Op. cit., p. 76. My italics. The same general view, though in asomewhat different connection, has since been expressed by Mr. J. M.Keynes in several passages of his Treatise on Money, 1930. Cf., foiexample, vol. i, p. 175 and especially p. 279: 'There is, indeed, nopossibility of intelligent foresight designed to equate savings andinvestment unless it is exercised by the banking system.'

80


and thus there is always a danger that one of theseprocesses may lag behind while the other hastensforward. This reference to the entrepreneur'signorance of the situation belongs, however, tothat category of explanation which we had toreject earlier. Instead of showing why prices —and in this case, particularly, the price ofcapital, which is interest — do not fulfil, orfail to fulfil adequately, their normal functionof regulating the volume of production, it un-expectedly overlooks the fact that the extent ofproduction is regulated on the basis not of aknowledge of demand but through price deter-mination. Assuming that the rate of interestalways determines the point to which the availablevolume of savings enables productive plant to beextended — and is it only by this assumption thatwe can explain what determines the rate of interestat all — any allegations of a discrepancy betweensavings and investments must be backed up by ademonstration why, in the given case, interestdoes not fulfil this function.* Professor Spiethoff,

* Elements of the same reasoning can also be found in G. CasselTheory of Social Economy, 4th German edition, 1927, p. 575; when

F 81


like most of the theorists of this group, evadesthis necessary issue — as we shall see later — byintroducing another assumption of crucial impor-tance. It is only by means of this assumptionthat the causes which he particularly enumeratesin his analysis gain significance as an explanation;and therefore it should not have been treated asa self-evident condition, to be casually mentioned,but as the starting-point of the whole theoreticalanalysis.

V I I

Before going into this question, however,we must turn our attention to the importance inTrade Cycle theory of errors of forecast, and, inconnection with these, to a third group of theorieswhich have not been considered up to now —the psychological theories. Here, as elsewhere inour investigations, we shall only be concernedwith those theories which are endogenous — i.e.

he derives high-conjuncture from an overestimate of the supply ofcapital (i.e. savings) which is available to take over the supply ofreal capital produced.


which explain the origin of general under- andover-estimation from the economic situation itself,and not from some external circumstance suchas weather changes, etc. As we said earlier,fluctuations of economic activity which merelyrepresent an adjustment to corresponding changesin external circumstances present no problemto economic theory. The various psychologicalfactors cited are only relevant to our analysis inso far as they can cast light on its central problem:that is, how an over-estimate of future demand canoccasion a development of the productive apparatusso excessive as automatically to lead to a reaction,unprecipitated by other psychological changes.Those who are familiar with the most distinguishedof these theories, that of Professor A. C. Pigou(which, owing to lack of space, cannot be repro-duced here)* will see at once that the endogenouspsychological theories are open to the sameobjections as the two groups of theories whichwe have already examined. Professor Pigou does

• Cf. A. C. Pigou, Industrial Fluctuations, London, 1927 (2nded. 1929); and also O. Morgenstern, Qualitative und QuantitativeKonjunkturforschungy Zeitschrift fiir die gesammte Staatszvissenschaft,vol. 84, Tubingen, 1928.

83


not explain why errors should arise in estimatingthe effect, on the price of the final product, of anincrease in demand or a fall in cost; or, if theestimate is correct, why the readjustment of theprices of means of production should not checkthe expansion of production at the right point. Noone would deny, of course, that errors can ariseas regards the future movements of particularprices. But it is not permissible to assumewithout further proof that the equilibratingmechanism of the economic system will begin itswork only when the excessively increased productdue to these mistaken forecasts actually comes onthe market, the disproportional development con-tinuing undisturbed up to that time. At onepoint or another, all theories which start to explaincyclical fluctuations by miscalculations or ignor-ance as regards the economic situation fall intothe same error as those naive explanations whichbase themselves on the 'planlessness' of theeconomic system. They overlook the fact that,in the exchange economy, production is governedby prices, independently of any knowledge of thewhole process on the part of individual producers,

84

NON-MONETARY THEORIES OF TRADE CYCLEso that it is only when the pricing process is itselfdisturbed that a misdirection of production canoccur. The 'wrong' prices, on the other hand,which lead to 'wrong' dispositions, cannot in turnbe explained by a mistake. Within the frameworkof a system of explanations in which, as in allmodern economic theory, prices are merely ex-pressions of a necessary tendency towards a stateof equilibrium, it is not permissible to reintroducethe old Sismondian idea of the misleading effectof prices on production without first bringing itinto line with the fundamental system of explan-ation.

V I I I

It is, perhaps, scarcely necessary to point outthat all the objections raised against the non-monetary theories, already cited in our investi-gations, are justified by one particular assumptionwhich we had to make in order to examine theindependent validity of the so-called 'real' explan-ations. In order to see whether the 'real' causes(whose effect is always emphasized as a proof thatmonetary changes are not the cause of cyclical

85


fluctuations) can provide a sufficient explanationof the cycle, it has been necessary to study theiroperation under conditions of pure barter. Andeven if it were impossible to prove fully that,under these conditions, no non-monetary explan-ation is sufficient, enough has been said, I think,to indicate the general trend of thought whichwould refute all theories based exclusively onproductive, market, financial or psychologicalphenomena. None of these phenomena can helpus to dissolve the fundamental equilibrium-relationships which form the basis of all economicexplanation. And this dissolution is indispensableif we are to protect ourselves against objectionssuch as those outlined above.

If the various theories comprised in thesegroups are still able to offer a plausible explanationof cyclical fluctuations, and if their authors donot realize the contradictions involved, this isdue to the unconscious importation of an assump-tion incompatible with a purely 'real' explanation.This assumption is adequate to dissolve the rigidreaction-mechanisms of barter-economy, and thusmakes possible the processes described; but for

86

NON-MONETARY THEORIES OF TRADE CYCLEthis very reason it should not be treated as a self-evident condition, but as the basis of the explan-ation itself. The condition thus tacitly assumed —and one can easily prove that it is in fact assumedin all the theories examined above — is the existenceof credit which, within reasonable limits, is alwaysat the entrepreneur's disposal at an unchangedprice. This, however, assumes the absence of themost important controls which, in the bartereconomy, keep the extension of the productiveapparatus within economically permissible limits.Once we assume that, even at a single point, thepricing process fails to equilibrate supply and demand',so that over a more or less long period demand maybe satisfied at prices at which the available supplyis inadequate to meet total demand, then the marchof economic events loses its determinateness and arange of indeterminateness appears, within whichmovements can originate leading away from equili-brium. And it is rightly assumed, as we shall seelater on, that it is precisely the behaviour of interest,the price of credit, which makes possible thesedisturbances in price formation. We must not,however, overlook the fact that the range of

87


indeterminateness thus created is 'indeterminate*only in relation to the absolute determinatenessof barter economy. The new price formation,together with the new structure of productiondetermined by it, must in turn conform to certainlaws, and the apparent indeterminateness doesnot imply unfettered mobility of prices andproduction. On the contrary, every departure fromthe original equilibrium position is definitely deter-mined by the new conditioning factor. But if it isthe existence of credit which makes these variousdisturbances possible, and if the volume anddirection of new credits determines the extentof deviations from the equilibrium position, itis clearly not permissible to regard credit as akind of passive element, and its presence as aself-evident condition. One must regard itrather as the new determining factor whoseappearance causes these deviations and whoseeffects must form our starting-point when de-ducing all those phenomena which can be observedin cyclical fluctuations. Only when we havesucceeded in doing this can we claim to haveexplained the phenomena described.

88


The neglect to derive the appearance of dis-proportionality from this condition, which mustbe assumed in order to keep the argument withinthe framework of equilibrium theory, leads tocertain consequences which are best exemplifiedin the work of Professor Spiethoff. For, in histheory, all important interconnections are workedout in the fullest detail and none of the observedphenomena remains unaccounted for. But he isnot able to deduce the various phenomenadescribed from the single factor which, by virtueof its role in disturbing the inter-relationshipsof general equilibrium, should form the basis ofhis explanation. At each stage of his expositionhe calls in experience to back him up and to showwhat deviations from the equilibrium positionactually occur within the given range of indeter-minateness. Consequently it never becomes clearwhy these phenomena must always occur as theyare described; and there always remains a possi-bility that, on some other occasion, they mayoccur in a different way, or in a different order,without his being able to account for this differ-ence on the basis of his exposition. In other words

89


the latter, however accurately and pertinently itdescribes the observed phenomena, does notqualify as a theory in the rigid sense of the word,for it does not set out those conditions in whosepresence events must follow a scientifically deter-mined course.

IX

Although there is no doubt that all non-monetary Trade Cycle theories tacitly assumethat the production of capital goods has been madepossible by the creation of new credit, and al-though this condition is often emphasized in thecourse of the exposition,* no one has yet provedthat this circumstance should form the exclusivebasis of the explanation. As far as strict logicis concerned, it would not be impossible for suchtheories to make use of some other assumptionwhich is capable of dissolving the rigid inter-relationships of equilibrium and, therefore, offorming the basis of an exact theoretical analysis.But once we assume the existence of credit in our

• Cf. Spiethoff, op. cit., pp. 77-78 and 81.

90


explanation, we can attack the problem by seeinghow far the objections which were raised earlieragainst the validity of the various theories undera barter economy are invalidated when the newassumption is made. Then we shall also be ableto determine whether this assumption has neces-sarily to be made in the usual form, or whether itonly represents a special instance of a far morewidely significant extension of the assumptionsof elementary theory.

The question we have to ask ourselves is:What new price-determining factor is introducedby the assumption of a credit supply which canbe enlarged while other conditions remain un-changed — a factor capable of deflecting the ten-dency towards the establishment of equilibriumbetween supply and demand? Whether we ne-cessarily accept that answer which, to my mind,is the only possible one depends on whether weagree with a certain basic proposition, whichcould only be briefly outlined here and whose fullproof could only be given within the frameworkof a complete system of pure economics; namely,the proposition that, in a barter economy, interest

91


forms a sufficient regulator for the proportionaldevelopment of the production of capital goodsand consumption goods, respectively. If it isadmitted that, in the absence of money, interestwould effectively prevent any excessive extensionof the production of production goods, by keepingit within the limits of the available supply ofsavings, and that an extension of the stock ofcapital goods which is based on a voluntary post-ponement of consumers' demand into the futurecan never lead to disproportionate extensions,then it must also necessarily be admitted thatdisproportional developments in the productionof capital goods can arise only through the inde-pendence of the supply of free money capitalfrom the accumulation of savings; which in turnarises from the elasticity of the volume of money.*Every change in the volume of means of circu-lation is, in fact, an event to be distinguished

* 'Volume of money', in this connection, does not mean merelythe quantity of money in circulation but the volume of the moneystream or the effective circulation (in the usual terminology —quantity times velocity of circulation). Even so, certain changes inthe effective circulation may have no disturbing effect because ofcertain compensating changes in business organization. On thispoint see my Prices and Production, Lecture iv.

92


from all other real causes, for the purpose oftheoretical reasoning; for, unlike all others, itimplies a loosening of the inter-relationships ofequilibrium. No change in 'real' factors, whetherin the amount of available means of production,in consumers' preferences, or elsewhere, can doaway with that final identity of total demand andtotal supply on which every conception of econo-mic equilibrium is based. A change in the volumeof money, on the other hand, represents as it werea one-sided change in demand, which is notcounterbalanced by an equivalent change in supply.Money, being a pure means of exchange, not beingwanted by anyone for purposes of consumption,must by its nature always be re-exchanged with-out ever having entirely fulfilled its purpose; thuswhen it is present it loosens that finality and'closedness' of the system which is the funda-mental assumption of static theory, and leads tophenomena which the closed system of staticequilibrium renders inconceivable.*

* This dissolution of the 'closedness' of the system, arising becausea change in the volume of money is a one-sided change in demandunaccompanied by an equivalent change in supply, does not mean ofcourse that Lowe's plea for an 'open' system ('Wie ist Konjunktur-

93


Together with the 'closedness' of the systemthere necessarily disappears the interdependenceof all its parts, and thus prices become possiblewhich do not operate according to the self-regu-lating principles of the economic system describedby static theory. On the contrary, these prices mayelicit movements which not only do not lead toa new equilibrium position but which actuallycreate new disturbances of equilibrium. In thisway, through the inclusion of money among thebasic assumptions of exposition, it becomespossible to deduce a priori phenomena such asthose observed in cyclical fluctuations. Oneinstance of these disturbances in the price mechan-ism, brought about by monetary influences — andthe one which is most important from the pointof view of Trade Cycle theory — is that puttingout of action of the'interest brake'which is takenfor granted by the Trade Cycle theories examinedabove. How far this circumstance forms a suffi-theorie iiberhaupt moglich') has been granted. (Lowe thinks of asystem when one or several 'independent variables' are drawn in forexplanation.) This plea, which one is tempted to believe has beendictated by a desire to free theory from the trammels of exact deduc-tion, has been justly and strongly criticized by E. Carell (op. cit. pp.Z et seq. and 115).

94


cient basis for a theory of the Trade Cycle is aproblem of the concrete elaboration of monetaryexplanation, and will, therefore, be dealt with inthe next chapter, where we shall examine howfar existing monetary theories have already tackledthose problems which are relevant to a theoryof the Trade Cycle.

The purpose of the foregoing chapter was toshow that only the assumption of primary monetarychanges can fulfil the fundamentally necessarycondition of any theoretical explanation of cyclicalfluctuations — a condition which is not fulfilledby any theory based exclusively on 'real' processes.If this is true then, at the outset of theoreticalexposition, those monetary processes must berecognized as decisive causes. For we can gain atheoretically unexceptionable explanation of com-plex phenomena only by first assuming the fullactivity of the elementary economic interconnectionsas shown by the equilibrium theory, and then intro-ducing\ consciously and successively y just those

95


elements which are capable of relaxing these rigid

inter-relationships. All the phenomena whichbecome possible only as a result of this relaxationmust then be explained — as consequences of theparticular elements, through whose inclusionamong the elementary assumptions they becomeexplicable within the framework of general theory.In place of such a theoretical deduction, we oftenfind an assertion, unfounded on any system, ofa far-reaching indeterminacy in the economy.Paradoxically stated as it is, this thesis is boundto have a devastating effect on theory; for itinvolves the sacrifice of any exact theoretical de-duction, and the very possibility of a theoreticalexplanation of economic phenomena is renderedproblematic.

Similar objections of a general nature must belevelled against another large group of theorieswhich we have not yet mentioned. This grouppays close attention to the monetary inter-con-nections and expressly emphasizes them as anecessary condition for the occurrence of theprocesses described. But they fail to pass fromthis realization to the necessary conclusion; to

96


make it a starting-point for their theoreticalelaboration, from which all other particular phe-nomena have to be deduced. To this groupbelongs the theory of Professor J. Schumpeter,and certain 'underconsumption theories'* notablythat of Professor E. Lederer; and* similarly, thevarious 'realistic' theories — that is, those whichrenounce any unified theoretical deduction, suchas those of Professors G. Cassel, J. Lescure andWesley Mitchell. With regard to all these semi-monetary explanations, we must ask whether —once we have been compelled to introduce newassumptions foreign to the static system— it isnot the first task of a theoretical investigation toexamine all the consequences which must neces-sarily ensue from this new assumption, and, in sofar as any phenomena are thus proved to be logic-ally derivable from the latter, to regard them inthe course of the exposition as effects of the newcondition introduced. Only in this way is itpossible to incorporate Trade Cycle theory into

# For a detailed criticism of a representative specimen of modernunder-consumption theories, that of Messrs. W. T. Foster and W.Catchings, see my article on The 'Paradox' of Saving, 'Economica'No. 32, May 1931.

G 97


the static system which is the basis of alltheoretical economics; and, for this very reason,the monetary elements must be regarded as de-cisive factors in the explanation of cyclicalfluctuations. The contrast therefore can be re-duced to a question of theoretical presentation,and it may even seem, when comparing thesetheories, that the matter of the express recognitionof the monetary starting-point is one of purelymethodological or even terminological importance,having no bearing on the essential solution of theproblem. But the same procedure which in onecase may only lead to a lapse from theoreticalelegance, breaking the unity of the theoretical struc-ture, may in another case lead to the introductionof thoroughly faulty reasoning, against whichonly a rigid systematical procedure provides aneffective security.

98

MONETARY THEORIES OFTHE TRADE CYCLE

C H A P T E R I I I

M O N E T A R Y T H E O R I E S O FT H E T R A D E C Y C L E

THE argument of the foregoing chapters hasdemonstrated the main reason for the necessity ofthe monetary approach to Trade Cycle theory.It arises from the circumstance that the automaticadjustment of supply and demand can only bedisturbed when money is introduced into theeconomic system. This adjustment must beconsidered, according to the reasoning whichit most clearly expressed in Say's TMorie desDe'bouchds, as being always present in a stateof natural economy. Every explanation of theTrade Cycle which uses the methods of economictheory — which of course is only possible throughsystematic co-ordination of the former with thefundamental propositions of the latter — must,therefore, start by considering the influences

IOI


which emanate from the use of money. Byfollowing up their results it should be possibleto demonstrate the total effect on the economicsystem, and formulate the result into a co-ordin-ated whole. This must be the aim of all theorieswhich set out to explain disturbances in equili-brium which, by their very nature, cannot beregarded as immediate consequences of changesin data, but only as arising out of the developmentof the economic system itself. For that typicalform of disturbance, which experience shows tobe regularly recurrent and which can properlybe called the Trade Cycle, the influence of moneyshould be sought in the fact that when the volumeof money is elastic, there may exist a lack ofrigidity in the relationship between saving andthe creation of real capital. This is a factwhich nearly all the theories of the dispro-portional production of capital goods are agreedin emphasizing. It is, therefore, the first task ofmonetary Trade Cycle theory to show why, andhow, monetary influences directly bring aboutregular disturbances in just this part of the eco-nomic system.

102

MONETARY THEORIES OF TRADE CYCLE

11

Naturally no attempt will be made at thisstage to present such a theory systematically.This chapter is concerned with one particulartask: it attempts to show how far existing monetarytheories have already gone towards a satisfactorysolution of the problem of the Trade Cycle, andwhat corrections are needed in order to invalidatecertain objections which, up to the present, haveappeared well founded.

It should already be clear that what we expectfrom a monetary Trade Cycle theory differs con-siderably from what most of the monetary TradeCycle theories regard as the essential aim of theirexplanation. We are in no way concerned toexplain the effect of the monetary factor on tradefluctuations through changes in the value of moneyand variations in the price level — subjects whichform the main basis of current monetary theories.We expect such an explanation to emerge ratherfrom a study of all the changes originating in themonetary field — more especially variations in itsquantity — changes which are bound to disturb

103


the equilibrium inter-relationships existing inthe natural economy, whether the disturbanceshows itself in a change in the so-called 'generalvalue of money* or not. Our plea* for a monetaryapproach to all Trade Cycle theory does not,therefore, imply that henceforward such theoriesshould be exclusively, or even principally, basedon those arguments which usually predomi-nate in writings on money, and which set outto explain the general level of prices and alter-ations in the Value of money'. On the contrary,monetary theory should not merely be concernedwith money for its own sake, but should also studythose phenomena which distinguish the moneyeconomy from the equilibrium inter-relation-ships of barter economy which must always beassumed by 'pure economics\

It must of course be admitted that many TradeCycle theorists regard the importance of monetarytheory as residing precisely in its ability to explainthe cause of fluctuations by reference to changesin the general price level. Hence, it is not difficultto understand why certain economists believe that,once they have rejected this view, they have

104


settled once and for all with the monetary explan-ations of the Trade Cycle. It is not surprisingthat monetary theories of the Trade Cycle shouldbe rejected by those who, like Professor A.Spiethoff in his well-known work on the QuantityTheory as 'Haussetheorie* * identify them withthe naive quantity-theory explanations whichderive fluctuations from changes in the price-level, f Against such a conception it can rightlybe urged that there are a number of phenomenatending to bring about fluctuations, which cer-tainly do not depend on changes in the value ofmoney, and which can, in fact, exert a disturbingeffect on the economic equilibrium without thesechanges occurring at all. Again, in spite of manyassertions to the contrary, fluctuations in the

* 'Die Quantitatstheorie, insbesondere in ihrer Verwertbarkeitals Haussetheorie'. Festgaben fur A. Wagner zur 70 Wiederkehrseines Geburtstages, Leipzig, 1905, pp. 299 et seq.

f F. Burchardt, A. Lowe, and other more recent critics of monetarytrade cycle theory, also fall within this category. They recognize noother kind of monetary influence than that which manifests itselfthrough changes in the price-level; and as a result of this undoubtedlyfalse conception they quite definitely conclude that there can be nosuch thing as pure monetary trade cycle theory. In their view thetheories which are usually so called nearly always depend, in fact,on what they regard as non-monetary factors.

105


general price-level need not always be ascribedto monetary causes.*

1 1 1

But theories which explain the TradeCycle in terms of fluctuations in the generalprice-level must be rejected not only becausethey fail to show why the monetary factor disturbsthe general equilibrium, but also because theirfundamental hypothesis is, from a theoreticalstandpoint, every bit as naive as that of thosetheories which entirely neglect the influence ofmoney. They start off with a 'normal position*which, however, has nothing to do with the normalposition obtaining in the static state; and theyare based on a postulate, the postulate of a constant

* The assertion that changes in the general level of prices mustalways originate on the monetary side, as is argued for example byProfessors G. Cassel and Irving Fisher, obviously depends on circularreasoning. It starts from the postulate that the amount of moneymust be adjusted to changes in the volume of trade in such a waythat the price-level shall remain unchanged. If it is not, and the volumeof money remains unaltered, then, according to this remarkableargument, the latter becomes the cause (!) of changes in the price level.This statement is made quite baldly by Professor G. Cassel in hisbook Money and Foreign Exchange After 1914. London, 1922.

106

MONETARY THEORIES OF TRADE OYCLEprice-level, which, if fulfilled, suffices in itself tobreak down the inter-relationships of equilibrium.All these theories, indeed, are based on the idea —quite groundless but hitherto virtually unchallenged— that if only the value of money does not changeit ceases to exert a direct and independent influ-ence on the economic system. But this assumption(which is present, more or less, in the work of allmonetary theorists), so far from being the necessarystarting-point for all trade cycle theory, is perhapsthe greatest existing hindrance to a successfulexamination of the course of cyclical fluctuations.It forces us to assume variations in the effectivequantity of money as given. Such variations,however, always dissolve the equilibrium inter-relationships described by static theory; but theymust necessarily be assumed if the value of moneyis to remain constant despite changes in data;and therefore they cannot be used to explaindeviations from the course of events which statictheory lays down. The only proper starting-pointfor any explanation based on equilibrium theorymust be the effect of a change in the volume ofmoney; for this, in itself, constitutes a new state

107

MONETARY THEORY AND THE TRADE CYCLEof affairs, entirely different from that generallytreated within the framework of static theory.

In complete contrast to those economic changesconditioned by 'real* forces, influencing simul-taneously total supply and total demand, changesin the volume of money have, so to speak, a one-sided influence which elicits no reciprocal ad-justment in the economic activity of differentindividuals. By deflecting a single factor, withoutsimultaneously eliciting corresponding changesin other parts of the system, it dissolves its 'closed-ness', makes a breach in the rigid reaction mechan-ism of the system (which rests on the ultimateidentity of supply and demand) and opens a wayfor tendencies leading away from the equilibriumposition. As a theory of these one-sided influences,the theory of monetary economy should, there-fore, be able to explain the occurrence of pheno-mena which would be inconceivable in the bartereconomy, and notably the disproportional develop-ments which give rise to crises.* A starting-

• F. Wieser in Der Geldwert und seine geschichtlichen Verdnderungen(Zeitschrift fur Volkswirtschaft Sozialpolitik und Verwaltung,vol. xiii, 1907, p. 57, reprinted in F. Wieser, Abhandlungen;Tubingen, 1929, p. 178) has dealt with the special effects of a 'one-sided money supply*.

108


point for such explanations should be found inthe possibility of alterations in the quantity ofmoney occurring automatically and in the normalcourse of events, under the present organizationof money and credit, without the need for violentor artificial action by any external agency.

IV

Even if a systematic treatment of the TradeCycle problem has not yet been forthcoming, itshould be noted that, throughout the differentattempts at monetary explanation, there runs asecondary idea which is closely allied to that ofthe direct dependence of fluctuations on changesin the value of money. It is true that thisidea is used merely as a subordinate device oftechnique to assist in the explanation of fluctua-tions in the value of money. But its develop-ment included the analysis of the most impor-tant elements in the monetary factors chieflyconnected with the Trade Cycle. This was donein the teaching which began with H. Thornton*

• An Enquiry into the Nature and Effects of the Paper Credit ofGreat Britain, London, 1802; especially pp. 287 et seq. This is

109

MONETARY THEORY AND THE TRADE CYCLEand D. Ricardo* and was taken up again byH. D. Macleod,f H. Sidgwick, R. Giffen andJ. S. Nicholson,^ and finally developed by A.Marshall^ K. Wicksell,|| and L. v. Mises,fone of the most remarkable accomplishments in monetary theory,and still commands great attention; cf. the references to it by K.Wicksell in the preface to the second volume of his Vorlesungenfiber Nationalokonomie auf Grundlage des Marginalprinzipes, Jena,1922, p. xii; and Burchardt, op. cit. For a fuller discussion of theseearlier theories, see my 'Prices and Production/ London, 1931, p. iiff.

* Cf. 'The high price of bullion' (Essays, Gonner edit., p. 35),where Ricardo says that 'interest would, during that interval beunder its natural level,' and also Chapter XXVII of his Principles(McCulloch ed., p. 220), which for a long time have passed almostunnoticed but which already contained much of what is set out inlater theories.

t Theory and Practice of Banking, 1855 and later editions. Seeparticularly vol. ii, pp. 278 et seq.

$ For H. Sidgwick, R. Giffen and J. S. Nicholson, cf. J. W. Angell'sTheory of International Prices-History, Criticism and Restatement,Cambridge, 1926, pp. 117-122.

§ Cf. his evidence before the various Parliamentary Commissionswhich is collected in the volume Official Papers by Alfred Marshall,London, 1926, especially pp. 38-41, 45, 46 et seq. 273 et seq., aswell as the later account in 'Money, Credit, and Commerce,' pp.255-56.

|| Especially in Geldzins und Giiterpreise, Jena, 1898, as wellas in the second volume of his later Vorlesungen, already quoted,which has not had the influence which it deserved, mainly on accountof the exceedingly bad German translation in which it appeared.I had unfortunately no means of access to the other Swedish worksconnected with that of Wicksell, which should certainly not beoverlooked if one is to achieve a complete survey of the developmentof this theory.

If Theorie des Geldes und der Umlaufsmittel, ist edit. 1912, 2nd

IIO

MONETARY THEORIES OF TRADE CYCLEwhose works trace the development of the effectson the structure of production of a rate of interestwhich alters relatively to the equilibrium rate,as a result of monetary influences. For thepurpose of this review it is unnecessary to goback to the earlier representatives of this group;it is enough to consider the conceptions of Wickselland Mises, since both the recent improvementswhich have been effected and the errors whichstill subsist can be best examined on the basisof these studies.*

It must be taken for granted that the reader isacquainted with the works of both Wicksell andMises. Wicksell, from the outset,f regards theproblem as concerning explicitly the averagechange in the price of goods, which from the

edit. 1924; also the more recent Geldwertstabilisierung und Konjunk-turpolitik, Jena, 1928. [A translation of the former will shortly appearin the present series. Ed.]

• Professor A. Hahn, whose views regarding trade cycle theory(put forward in Volkswirtschaftlichen Theorie des Bankkredits,Tubingen, 1920) are in some respects similar to those of ProfessorMises, cannot be considered here, since we are unable to followhim in all those points in which he differs from the latter. Similartheories have also been put forward quite recently by Professor W.ROpke and S. Budge.

T See e.g. Geldzins und Giiterpreise, p. 125.

I l l


theoretical standpoint is quite irrelevant. Hestarts from the hypothesis that, in the absence ofdisturbing monetary influences, the average price-level must remain unchanged. This assumptionis based on another, only incidentally expressed,*which is not worked out and which, from thepoint of view of most of the problems dealt with,is not even permissible; i.e. the assumption ofa stationary state of the economy. His funda-mental thesis is that when the money rate ofinterest coincides with the natural rate (i.e. thatrate which exactly balances the demand for loancapital and the supply of savings)? then moneybears a completely neutral relationship to the priceof goods, and tends neither to raise nor to lower it.But, owing to the nature of his basic assumptions,this thesis enables him to show deductively onlythat every lag of the money-rate behind thenatural rate must lead to a rise in the generalprice-level, and every increase of the money-rateabove the natural rate to a fall in general price-level. It is only incidentally, in the course of his

# Ibid., p. 126.f Ibid., p. 93 and also Vorlesungen, vol. ii, p. 220.

112


analysis of the effects on the price-level of a moneyrate of interest differing from the natural ratethat Wicksell touches on the consequences ofsuch a distortion of the natural price formation(made possible by elasticity in the volume ofcurrency) on the development of particular branchesof production; and it is this question which is ofthe most decisive importance to Trade Cycletheory. If one were to make a systematic attemptto co-ordinate these ideas into an explanation ofthe Trade Cycle (dropping, as is essential, theassumption of the stationary state) a curiouscontradiction would arise. On the one hand, weare told that the price level remains unaltered whenthe money-rate of interest is the same as the naturalrate; and, on the other, that the production ofcapital goods is, at the same time, kept within thelimits imposed by the supply of real savings. Oneneed say no more in order to show that there arecases — certainly all cases of an expandingeconomy, which are those most relevant toTrade Cycle theory — in which the rate of interestwhich equilibrates the supply of real savings andthe demand for capital cannot be the rate of interest

H 113


which also prevents changes in the price-level.*In this case, stability of the price-level pre-supposes changes in the volume of money: butthese changes must always lead to a discrepancybetween the amount of real savings and thevolume of investment. The rate of interest atwhich, in an expanding economy, the amount ofnew money entering circulation is just sufficient tokeep the price-level stable, is always lower than therate which would keep the amount of availableloan-capital equal to the amount simultaneouslysaved by the public, and thus, despite the stabilityof the price-level, it makes possible a developmentleading away from the equilibrium position. ButWicksell does not recognize here a monetaryinfluence tending, independently of changes in theprice-level, to break down the equilibrium systemof barter economics: so long as the stability ofthe price level is undisturbed, everything appearsto him to be in order.f Obsessed by the notionthat the only aim of monetary theory is to explain

* Similarly also W. Eucken, op. cit. pp. 300 et seq.f Wicksell's justification of this view in Geldzins und Giiterpreise

(see p. 97) is incomprehensible to me.

114


those phenomena which cause the value of moneyto alter, he thinks himself justified in neglectingall deviations of the processes of money-economyfrom those of barter-economy, so long as theythrow no direct light on the determination of thevalue of money: and thus he shuts the door onthe possibility of a general theory covering all theconsequences of the phenomena which heindicates .* But although his thesis of a directrelationship between movements in the price-leveland deviations of the money-rate of interest from

* R. Stucken in his Theorie der Konjunkturschtcankungen (Jena,1926, p. 26) was one of the first to draw attention to the fact that therelation, indicated by Wicksell, between a money rate of interestdiverging from the natural rate, and movements in the price-levelonly exists in a stationary economy; while, if the flow of goods isincreasing, only an addition to purchasing power can secure stabilityin the price-level. He remains, however, entirely steeped in theprevalent opinion that a stable price-level is indispensable to undis-turbed economic development, and therefore holds that the additionalmoney necessary to secure that condition cannot be regarded as anelement of disturbance in the economic process. Similarly Mr. D. H.Robertson pointed out at about the same time {Banking Policy and thePrice Level, London, 1926, p. 99) that the rate of interest whichkeeps the price-level stable need not coincide with that which equatesthe supply of savings with the demand for capital. I am now informedthat, even before the war, this objection formed the basis of acriticism directed by Prof. David Davidson of Upsala againstWicksell's theory. Prof. Davidson's article and the subsequentdiscussion with Wicksell in the Swedish Economisk Tidskrift are,however, inaccessible to me.

I I


its natural level, holds good only in a stationarystate, and is therefore inadequate for an explana-tion of cyclical fluctuations, his account of theeffects of this deviation on the price structureand the development of the various branches ofproduction constitutes the most important basisfor any future monetary Trade Cycle theory.But this future theory, unlike that of Wicksell,will have to examine not movements in the generalprice-level but rather those deviations of particularprices from their equilibrium position which werecaused by the monetary factor.

The investigations of Professor Mises repre-sent a big step forward in this direction, althoughhe still regards the fluctuations in the value ofmoney as the main object of his explanation,and deals with the phenomena of disproportion-ality only in so far as they can be regarded asconsequences — in the widest sense of the term —of these fluctuations. But Professor Mises' con-ception of the intrinsic value of money extends the

116


notion of 'fluctuations in money value* far beyondthe limits of what this term is commonly under-stood to mean; and so he is in a position to describewithin the framework, or rather under the name,of a theory of fluctuations in the value of money,all monetary influences on price formation.*His exposition already contains an account ofpractically all those effects of a rate of interestaltered through monetary influences, which are

• If one follows C. Menger and now Professor Mises in disregardingordinary usage and including in the theory of the value of money allinfluences of money on prices, instead of restricting it to an explana-tion of the general purchasing power of money (by which is understoodthe absolute level of money prices as distinct from the relative pricesof particular goods) then it is correct to say that any economic theory ofmoney must be a theory of the value of money. But this use of thephrase is hardly opportune, for 'value of money' is usually taken tomean 'general purchasing power', while monetary theory has by nomeans finished its work when it has explained the absolute level of prices(or, as Wicksell would call it, the 'concrete' level); its far more importanttask is to explain those changes in the relative height of particular priceswhich are conditioned by the introduction of money. On the other hand,to avoid any possible misunderstanding we must particularly insist atthis point that in the sense of the famous contrast between suchnominalistic theories as the 'state theory' of Knapp, and the catallactictheories in general, the monetary theory which we are seeking willalso have to be exclusively a 'theory of money values'. In justice toMenger and Mises, it should be pointed out that what they mean whenthey speak of the stability of the 'inner' value of money, has nothingto do with any measurable value, in the sense of some price level; but isonly another and, as it seems to me, misleading, expression for whatI now prefer to call neutrality of money. (Cf. my Prices and Produc-

117


important for an explanation of the course of theTrade Cycle. Thus he describes the dispropor-tionate development of various branches of pro-duction and the resulting changes in the incomestructure. And yet this presentation of his theoryunder the guise of a theory of fluctuation in thevalue of money remains dangerous, partly becauseit always gives rise to misunderstandings, butmainly because it seems to bring into the fore-ground a secondary effect of cyclical fluctuations,an effect which generally accompanies the latterbut which need not necessarily do so.

This is no place to examine the extent to whichProfessor Mises escapes from this difficulty byusing the concept of the inner objective value ofmoney. For us, the only point of importance isthat the effects of an artificially lowered rate ofinterest, pointed out by Wicksell and Mises, exist

tion, pp. 27-28 and passim). This expression, first used by Wicksellin the passage quoted earlier in the text, has of late become fairlycommon in German and Dutch writings on money. Cf. L. v.Bortkiewicz, Die Frage der Reform unserer Wahrung, Brauns Analen,vol. vi, 1919, pp. S7-59J W. G. Behrens, Das Geldschopfungsproblem,Jena, 1928, pp. 228 et seq., 286 and 312; G. M. Verrijin Stuart andJ. G. Koopmans in the reports and discussions of the 1929 meetingof the 'Vereinigung voor de Staathuishoudkunde en de StatistikV

118

MONETARY THEORIES OF TRADE CYCLEwhether this same circumstance does or does noteventually react on the general value of money, inthe sense of its purchasing power. Thereforethey must be dealt with independently if they areto be properly understood.* Increases in thevolume of circulation, which in an expandingeconomy serve to prevent a drop in the price-level, present a typical instance of a change in themonetary factor calculated to cause a discrepancybetween the money and natural rate of interestwithout affecting the price-level. These changesare consequently neglected, as a rule, in dealingwith phenomena of disproportionality; but theyare bound to lead to a distribution of productiveresources between capital-goods and consumption-goods which differs from the equilibrium dis-tribution, just as those changes in the monetaryfactor which do manifest themselves in changesin the price-level. This case is particularlyimportant, because under contemporary currency

• Professor Mises recently admitted this, in principle, when heexplicitly emphasized the fact that every new issue of circulating mediabrings about a lowering of the money rate of interest in relationto the natural rate. (Geldwertstabtlisierung und Koryunkturpolitik,P- 57-)

119


systems the automatic adjustment of the value ofmoney, in the form of a flow of precious metals,will regularly make available new supplies ofpurchasing power which will depress the money-rate of interest below its natural level.*

Since a stable price-level has been regarded asnormal hitherto, far too little investigation hasbeen made into the effects of these changes in thevolume of money, which necessarily cause adevelopment different from that which would beexpected on the basis of static theory, and whichlead to the establishment of a structure of pro-duction incapable of perpetuating itself oncethe change in the monetary factor has ceased tooperate. Economists have overlooked the factthat the changes in the volume of money, which,in an expanding economy, are necessary tomaintain price stability, lead to a new state ofaffairs foreign to static analysis, so that thedevelopment which occurs under a stable pricelevel cannot be regarded as consonant with static

* Cf. also my article, 'Das intertemporale Gleichgewichtssystemder Preise und die Bewegungen des Geldwertes', WeltwirtschaftlichesArchiv, vol. 28, July 1928.

I2O


laws. Thus the disturbances described as resultingfrom changes in the value of money form only asmall part of the much wider category of deviationsfrom the static course of events brought about bychanges in the volume of money — which may oftenexist without changes in the value of money, whilethey may also fail to accompany changes in valueof money when the latter occur.

v i

As has been briefly indicated above, mostof the objections raised against monetary theoriesof cyclical fluctuations rest on the mistaken ideathat their significant contribution consists indeducing changes in the volume of productionfrom the movement of prices en bloc. In particular,the very extensive criticism recently levelled byDr. Burchardt and Professor Lowe against mone-tary Trade Cycle theory is based throughout onthe idea that this theory must start from the wave-like fluctuations of the price-level, which areconditioned mainly by monetary causes; the rise,as well as the fall, of the price-level being brought

121


about by particular new forces originating on theside of money. It is only through this specialassumption, which is also stated explicitly, thatProfessor Lowe's systematic presentation of hisobjections in his latest work* becomes compre-hensible; he is completely misleading when heasserts that, if it is to raise the monetary factor tothe rank of a conditio sine qua non of the TradeCycle, monetary theory ought to prove that theeffectiveness of all non-monetary factors dependson a previous price-boom.f We have alreadyshown that it is not even necessary, in order toascribe the cause of cyclical fluctuations to mone-tary changes, to assume that these monetary causesact through changes in the general price-level.It is therefore impossible to maintain that theimportance of monetary theories lies solely in anexplanation of price cycles.J

But even the essential point in the criticism ofLowe and Burchardt—the assertion that all mone-tary theories explain the transition from boom to

• Uber den Elnfluss monetarer Faktoren auf den Konjunkturzyklus,op. cit. pp. 361-368.

f Op. cit. p. 366.% As is maintained by Professor Lowe (op. cit., p. 364).

122

MONETARY THEORIES OF TRADE CYCLEdepression not in terms of monetary causes butin terms of other causes super-added to themonetary explanation—rests exclusively on the ideathat only general price changes can be recognizedas monetary effects. But general price changesare no essential feature of a monetary theory ofthe Trade Cycle; they are not only unessential, butthey would be completely irrelevant if only they werecompletelylgeneral— that is> if they affected allpricesat the same time and in the same proportion. Thepoint of real interest to Trade Cycle theory is theexistence of certain deviations in individual price-relations occurring because changes in the volumeof money appear at certain individual points; devia-tions, that is, away from the position which isnecessary to maintain the whole system in equili-brium. Every disturbance of the equilibrium ofprices leads necessarily to shifts in the structure ofproduction, which must therefore be regarded aaconsequences of monetary change, never asadditional separate assumptions. The nature ofthe changes in the composition of the existingstock of goods, which are effected through suchmonetary changes, depends of course on the

123


point at which the money is injected into theeconomic system.

There is no doubt that the emphasis placed onthis phenomenon marks the most importantadvance made by monetary science beyond theelementary truths of the quantity theory. Mone-tary theory no longer rests content with deter-mining the final reaction of a given monetarycause on the purchasing power of money, butattempts instead to trace the successive alterationsin particular prices, which eventually bring abouta change in the whole price system.* The assump-tion of a 'time lag1 between the successive changesin various prices has not been spun out of thin airsolely for the purposes of Trade Cycle theory; itis a correction, based on systematic reasoning, ofthe mistaken conceptions of older monetarytheories.*}- Of course, the expression 'time lag,'

* On the development of this point of view, see Chap. I of myPrices and Production.

f We cannot, therefore, regard Mises' pronouncement at the Zurichdebate of the Verein fiir Sozialpolitik as a surrender of the monetarystandpoint. On this occasion, he not only admitted but indeedemphasized the fact that monetary causes can only act by producing a'lag' between various prices, wages and interest rates. (Cf. Verhand-lungendes Vereinsfur Sozialpolitik in Zuricht Sch»iften des Vereins, vol.175, Munich and Leipzig, 1929.)

124


borrowed from Anglo-American writers and deno-ting a temporary lagging behind of the changesin the price of some goods relatively to thechanges in the price of other goods, is a veryunsuitable expression when the shifts in relativeprices are due to changes in demand which arethemselves conditioned by monetary changes. Forsuch shifts are bound to continue so long as thechange in demand persists. They disappear onlywith the disappearance of the disturbing monetaryfactor. They cease when money ceases toincrease or diminish further; not, however,when the increase or diminution has itself beenwiped out. But, whatever expression we may useto denote these changes in relative prices and thechanges in the structure of production conditionedby them, there can be no doubt that they are, inturn, conditioned by monetary causes, whichalone make them possible.

The only plausible objection to this argumentwould be that the shifts in price-relationshipsoccurring at any point in the economic systemcould not possibly cause those typical, regularlyrecurring, shifts in the structure of production

125


which we observe in cyclical fluctuations. Inopposition to this view, as we shall show in moredetail later, it can be urged that those changeswhich are constantly taking place in our moneyand credit organization cause a certain price, therate of interest, to deviate from the equilibriumposition, and that deviations of this kind necessarilylead to such changes in the relative position ofthe various branches of production as are boundlater to precipitate the crisis.* There is oneimportant point, however, which must beemphasized against the above-named critics;namely, that it is not only when the crisisis directly occasioned by a new monetary factor,separate from that which originally brought aboutthe boom, that it is to be regarded as conditionedby monetary causes. Once the monetary causeshave brought about that development in the whole

* It is not essential, as Burchardt maintains (op. cit., p. 124) tobase this analysis on any particular theory of interest, such as thatof Bohm-Bawerk; it is equally consonant with all modern interesttheories. The reason why, under the circumstances assumed, interestfails to equilibrate production for the future and production for thepresent, is bound up not with the special form in which interest ingeneral is explained, but with the deviations, due to monetary causes,of the current rate of interest from the equilibrium rate,


economic system which is known as a boom,sufficient forces have already been set in motionto ensure that, sooner or later, when the monetaryinfluence has ceased to operate, a crisis must occur.The 'cause' of the crisis is, then, the disequilibriumof the whole economy occasioned by monetarychanges and maintained through a longer period,possibly, by a succession of further monetarychanges — a disequilibrium the origin of which canonly be explained by monetary disturbances.

Professor Lowe's most important argumentagainst the monetary theory of the Trade Cycle —an argument which so far as most existingmonetary theories are concerned is unquestionablyvalid — will be discussed in more detail later. Thesole purpose of the next chapter of this book isto show that the cycle is not only due to 'mistakenmeasures by monopolistic bodies' (as ProfessorLowe assumes),* but that the reason for itscontinuous recurrence lies in an 'immanentnecessity of the monetary and credit mechanism.'

• Op. cit., p. 365 et seq.

127


VII

Among the phenomena which are funda-mentally independent of changes in the value ofmoney, we must include, first of all, the effects ofa rate of interest lowered by monetary influences,which must necessarily lead to the excessiveproduction of capital goods. Wicksell and Misesboth rightly emphasize the decisive importanceof this factor in the explanation of cyclical pheno-mena, as its effect will occur even when theincrease in circulation is only just sufficient toprevent a fall in the price level. Besides this,there exist a number of other phenomena, byvirtue of which a money economy (in the senseof an economy with a variable money supply)differs from a static economy, which for thisreason are important for a true understandingof the course of the Trade Cycle. They have beenpartly described already by Mises, but they canonly be clearly observed by taking as the centralsubject of investigation not changes in generalprices but the divergences of the relation ofparticular prices as compared with the price

128

MONETARY THEORIES OF TRADE CYCLEsystem of static equilibrium. Phenomena of thissort include the changes in the relation of costs andselling prices and the consequent fluctuations inprofits, which Professors Mitchell and Lescurein particular have made the starting point of theirexposition; and the shifts in the distribution ofincomes which Professor Lederer investigates —both of these phenomena depending for theirexplanation on monetary factors,* while neitherof them can be immediately connected withchanges in the general value of money. It is,perhaps, for this very reason that their authors,although perfectly realizing the monetary originof the phenomena which they described, did notpresent their views as monetary theories. Whilewe cannot attempt here to show the positionwhich these phenomena would occupy in a system-atically developed Trade Cycle theory (a taskwhich really involves the development of a newtheory, and which is unnecessary for the purposesof our present argument) it is not difficult to see

# That Professor Lederer himself sees this clearly is evident fromhis analysis, mentioned earlier, contained in the Grundriss der Sozial-okonomik, vol. iv, part i, pp. 390-91.

I 129


that all of them can be logically deduced from aninitiating monetary disturbance,* which, in anycase, we are compelled to assume in studyingthem. The special advantages of the monetaryapproach consist precisely in the fact that, bystarting from a monetary disturbance, we are ableto explain deductively all the different peculiaritiesobserved in the course of the Trade Cycle, and soto protect ourselves against objections such as wereraised in an earlier chapter against non-monetarytheories. It makes it possible to look uponempirically recognized interconnections, whichwould otherwise rival one another as independentclues to an explanation, as necessary consequencesof one common cause.

Much theoretical work will have to be donebefore such a theoretical system can be worked outin such detail that all the empirically observedcharacteristics of the Trade Cycle can find theirexplanation within its framework. Up to now, themonetary theories have unduly narrowed the fieldof phenomena to be explained, by limiting research

* Cf. Mise&' presentation of the social effects of changes in thevalue of money: Theorie des Geldes undder Umlaufsnrittel, pp. 178-200.

130


to those monetary changes which find theirexpression in changes in the general value ofmoney. Thus they are prevented from showingthe deviations of a money economy from a staticeconomy in all their multiplicity. The problem ofcyclical fluctuations can only be solved satisfactorilywhen a theory of the money economy itself — stillalmost entirely lacking at present —has beenevolved, comprising a detailed discussion of allthose points in which it differs from the equilibriumanalysis worked out on the assumption of a purebarter economy. The full elaboration of this inter-mediate step of theoretical exposition is indispens-able before we can achieve a Trade Cycle theory,which — as Bohm-Bawerk has expressed it in aphrase, often quoted but hardly ever taken toheart — must constitute the last chapter of thecomplete theory of social economy.* In myopinion, the most important step towards such atheory, which would embrace all new phenomenaarising from the addition of money to the

• This expression occurs in connection with a review of E. v.Bergmann's Geschichte der Nationaldkonomisthen Krisentheorien.(Zeitschrift fur Volkswirtschaft, Sozialpolitik und Vcrwaltung, vol.vii, 1898, p. 11 a.)


conditions assumed in elementary equilibriumtheory, would be the emancipation of the theory ofmoney from the restrictions which limit its scopeto a discussion of the value of money.

V I I I

Once, however, we have accomplished thisurgently necessary displacement of the problem ofmonetary value from its present central positionin monetary theory, we find ourselves in a positionto come to an understanding with the mostimportant non-monetary theorists of the TradeCycle; for the effect of money on the 'real' econo-mic processes will automatically be brought moreto the surface, while monetary theory will no longerappear to be insisting on the immediate depend-ence of Trade Cycle phenomena on changes in thevalue of money--a claim which is certainlyunjustified. On the other hand, a number of non-monetary theories do not question in the least thedependence of the processes which they describeon certain monetary assumptions; and in their

132

MONETARY THEORIES OF TRADE CYCLEcase the only conflict now arising concerns thesystematic presentation of these. It should be thetask of our analysis to show that the placing of themonetary factor in the centre of the exposition isnecessary in the interest of the unity of the system,and that the various 'real' interconnections, which,in certain theories, form the main basis of theexplanation, can only find place in a closed systemas consequences of the original monetary influ-ences. There can hardly be any question, in thepresent state of research, as to what should be thebasic idea of a completely developed theory ofmoney. One can abandon those parts of theWicksell-Mises theory which aim at explainingthe movements in the general value of money, anddevelop to the full the effects of all discrepanciesbetween the natural and money rates of intereston the relative development of the production ofcapital goods and consumption goods — a theorywhich has already been largely elaborated byProfessor Mises. In this way, one can achieve, bypurely deductive methods, the same picture of theprocess of cyclical fluctuations which the morerealistic theories of Spiethoff and Cassel have al-


ready deduced from experience. Wicksell himself*drew attention to the way in which the processesdeduced from his own theory harmonize with theexposition of Spiethoff; and conversely, Spiethoff,in a statement already quoted, has emphasized thefact that the phenomena which he describes are allconditioned by a change in monetary factors. Butit is only by placing monetary factors first that suchexpositions as those of Spiethoff and Cassel can beincorporated into the general system of theoreticaleconomics. A final point of decisive importanceis that the choice of the monetary starting pointenables us to deduce simultaneously all the otherphenomena, such as shifts in relative prices andincomes, which are more empirically determinedand utilized as independent factors; and thus therelations existing between them can be classifiedand their relative position and importance deter-mined within the framework of the theory.

* Vorlesungen, vol. ii, p. 238; Wicksell's review of CasseTstextbook, which has since appeared in a German translation (Schmol-lers Jahrbuch fur Volkswirtschaft, Gesetzgebung und Verwaltung,vol. 52, Munich, 1928), shows that, although he rightly opposesCassel's general system, he agrees to a large extent with his theory ofthe Trade Cycle.


Even when these phenomena are, as yet, muchfurther from a satisfactory explanation than arethe disproportionalities in the development ofproduction, which are cleared up in a greaterdegree, there can be no doubt that it will becomepossible to incorporate them also into a self-sufficient theory of the effects of monetary dis-turbances. These effects, however, althoughultimately caused by monetary factors, do not fallwithin the narrower field of monetary theory.A well-developed theory of the Trade Cycle oughtto deal thoroughly with them; but as this book isexclusively concerned with the monetary theoriesthemselves, we shall, in the following chapters,only study the reasons why these monetary causesof the Trade Cycle inevitably recur under theexisting system of money and credit organization,and what are the main problems with which futureresearch is faced by reason of the realization of thedetermining role played by money.

THE FUNDAMENTAL CAUSEOF CYCLICAL

FLUCTUATIONS

C H A P T E R I V

THE FUNDAMENTAL CAUSEOF CYCLICAL

FLUCTUATIONS

So far we have not answered, or have onlyhinted at an answer to the question why, under theexisting organization of the economic system, weconstantly find those deviations of the money rateof interest from the equilibrium rate* which, aswe have seen, must be regarded as the causeof the periodically recurring disproportionalitiesin the structure of production. The problem is,then, to discover the gap in the reaction mechanismof the modern economic system which is respon-

• The term 'equilibrium rate of interest' which, I believe, wasintroduced into Germany in this connection, by K. Schlesinger in hisTheorie der Geld-und Kreditzvirtschaft (Miinchen and Leipzig,1914, p. 128) seems to me preferable in this case to the usual expres-sion of 'natural rate* or 'real rate.' Alfred Marshall used the term'equilibrium level' as early as 1887 (cf. Official Papers of AlfredMarshall, p. 130). Cf. also chap. v. of the present work.

MONETARY THEORY AND THE TRADE CYCLEsible for the fact that certain changes of data, sofar from being followed by a prompt readjustment(i.e. the formation of a new equilibrium) are,actually, the cause of recurrent shifts in economicactivity which subsequently have to be reversedbefore a new equilibrium can be established.

The analysis of the foregoing chapters hasshown that when it is possible to detect, in theorganization of our economy, a dislocation in thereaction mechanism described by equilibriumtheory, it should be possible (and should, indeed,be the object of a fully developed Trade Cycletheory) to describe deductively, as a necessaryeffect of the disturbance — quite apart from theirobserved occurrence — all the deviations in thecourse of economic events conditioned by thisdislocation. It has been shown, in addition, thatthe primary cause of cyclical fluctuations must besought in changes in the volume of money, whichare undoubtedly always recurring and which, bytheir occurrence, always bring about a falsificationof the pricing process, and thus a misdirection ofproduction. The new element which we areseeking is, therefore, to be found in the 'elasticity'

140

CAUSE OF CYCLICAL FLUCTUATIONS

of the volume of money at the disposal of theeconomic system. It is this element whosepresence forms the 'necessary and sufficient*condition for the emergence of the Trade Cycle.*

The question which we now have to examine iswhether this elasticity in the volume of money is animmanent characteristic of our present moneyand credit system; whether, given certain con-ditions, changes in the volume of money and theresulting differences between the natural and themonetary rate of interest must necessarily occur,or whether they represent, so to speak, casualphenomena arising from arbitrary interferencesby the authorities responsible for the regulation

* Mr. R. G. Hawtrey regards the following theses as importantfor monetary Trade Cycle theories: (i) That certain monetary andcredit movements are necessary and sufficient conditions of theobserved phenomena of the Trade Cycle; and (2) that the periodicityof those phenomena can be explained by purely monetary tendencieswhich cause the movements to take place successively and to be spreadover a considerable period, of years. ('The Monetary Theory of theTrade Cycle and its Statistical Test': Quarterly Journal of Economics,vol. 41, p. 472). This entirely correct definition of Mr. Hawtrey'sshould have prevented Dr. Burchardt and Prof. Lowe, who expresslyfasten on this point in their criticism of monetary Trade Cycletheories, from looking from monetary influences to changes in thegeneral value of money, while disregarding the changes in the dis-tributive process which are conditioned by monetary causes,

MONETARY THEORY AND THE TRADE CYCLEof the volume of currency media. Is it an inherentnecessity of the existing monetary and creditsystem that its reaction to certain changes in datais different from what we should expect on thebasis of economic equilibrium theory; or arethese discrepancies to be explained by specialassumptions regarding the nature of the monetaryadministration, i.e. by a series of what might becalled 'political' assumptions? The questionwhether the recurrence of credit cycles is, or isnot, due to an unavoidable characteristic of theexisting economic organization, depends onwhether the existing monetary and credit organiza-tion in itself necessitates changes in the currencymedia, or whether these are brought about onlyby the special interference of external agencies.The answer to this question will also decideinto which of the most commonly acceptedcategories a given Trade Cycle theory is to beplaced. We must deal briefly with this pointbecause a false classification, which is largely thefault of the exponents of the monetarytheories, has contributed much to make themmisunderstood.

142


II

If we are to understand the present statusof monetary theories of the Trade Cycle, we mustpay special attention to the assumptions uponwhich they are based. At the present day, mone-tary theories are generally regarded as fallingwithin the class of so-called 'exogenous' theories,i.e. theories which look for the cause of the cyclenot in the interconnections of economic pheno-mena themselves but in external interferences.Now it is, no doubt, often a waste of time to discussthe merits of classifying a theory in a given category.But the question of classification becomes im-portant when the inclusion of a theory in one classor another implies, at the same time, a judgmentas to the sphere of validity of the theory inquestion. This is undoubtedly the case with thedistinction, very general to-day, between endogen-ous and exogenous theories — a distinction intro-duced into economic literature some twentyyears ago by Bouniatian.* Endogenous theories,

• Studien zur Theorie und Geschichte der Wirtschaftskrisen, Munich,1908, p- 3-

MONETARY THEORY AND THE TRADE CYCEL

in the course of their proof, avoid making use ofassumptions which cannot either be decided bypurely economic considerations, or regarded asgeneral characteristics of our economic system —and hence capable of general proof. Exogenoustheories, on the other hand, are based on concreteassertions whose correctness has to be provedseparately in each individual case. As comparedwith an endogenous theory, which, if logicallysound, can in a sense lay claim to general validity,an exogenous theory is at some disadvantage,inasmuch as it has, in each case, to justify theassumptions on which its conclusions arebased.

Now as far as most contemporary monetarytheories of the cycle are concerned, their opponentsare undoubtedly right in classifying them, as doesProfessor Lowe* in his discussion of the theoriesof Professors Mises and Hahn, among the exogen-ous theories; for they begin with arbitrary inter-ferences on the part of the banks. This is, perhaps,one of the main reasons for the prevailing scepti-

* Der gegemo&rtige Stand der Konjunkturforschung in Deutschland,op. tit. p. 349.

144


cism concerning the value of such theories. Atheory which has to call upon the deus ex machina*of a false step by bankers, in order to reach itsconclusions is, perhaps, inevitably suspect. YetProfessor Mises himself—who is certainly to beregarded as the most respected and consistentexponent of the monetary theory of the TradeCycle in Germany — has, in his latest work,afforded ample justification for this view of histheory by attributing the periodic recurrenceof the Trade Cycle to the general tendency ofCentral Banks to depress the money rate of interestbelow the natural rate.f Both the protagonistsand the opponents of the Monetary Theory ofthe Trade Cycle thus agree in regarding theseexplanations as falling ultimately within theexogenous and not the endogenous group. Thefact that this is not an inherent necessity of themonetary starting-point is however shown by the

* Cf. Neisser, Der Tauschwert des Geldes, Jena, 1928, p. 16 x.f While it seems to me that in the analysis of the effects of a money

rate of interest diverging from the natural rate Professor Mises hasmade considerable progress as compared with the position adoptedby Wicksell, the latter succeeded better than Mises did in explainingthe origin of this divergence. We shall go into Wicksell's explanationin somewhat more detail below.

K I 4 S


undoubtedly endogenous nature of the variousolder Trade Cycle theories, such as that ofWicksell. But since this suffers from otherdeficiencies, which have already been indicated,the question whether the exogenous character ofmodern theories is, or is not, an inherent necessityof their nature remains an open one.* It seemsto me that this classification of monetary TradeCycle theory depends exclusively on the fact that asingle, specially striking, case is treated as thenormal; while, in fact, it is quite unnecessary toadduce interference on the part of the banks inorder to bring about a situation of alternatingboom and crisis. By disregarding those diver-gencies between the natural and money rate ofinterest which arise automatically in the course ofeconomic development, and by emphasizing thosecaused by an artificial lowering of the money rate,the Monetary Theory of the Trade Cycle deprivesitself of one of its strongest arguments; namely,the fact that the process which it describes must

• Part of the two following paragraphs repeats word for word mycontribution to the discussion on 'Credit and the Trade Cycle' at theZurich Assembly of the 'Verein fur Sozialpolitik,' (cf. Schriften4e$ Vereins fur Sozialpolitik, vol. 175, p. 370-71).

146


always recur under the existing credit organization,and that it thus represents a tendency inherentin the economic system, and is in the fullestsense of the word an endogenous theory.

It is an apparently unimportant difference inexposition which leads one to this view that theMonetary Theory can lay claim to an endogenousposition. The situation in which the money rateof interest is below the natural rate need not,by any means, originate in a deliberate loweringof the rate of interest by the banks. The sameeffect can be obviously produced by an improve-ment in the expectations of profit or by a diminu-tion in the rate of saving, which may drive the* natural rate* (at which the demand for and thesupply of savings are equal) above its previouslevel; while the banks refrain from raising theirrate of interest to a proportionate extent, butcontinue to lend at the previous rate, and thusenable a greater demand for loans to be satisfiedthan would be possible by the exclusive use of theavailable supply of savings. The decisive sig-nificance of the case quoted is not, in my view,due to the fact that it is probably the commonest


in practice, but to the fact that it must inevitablyrecur under the existing credit organization.

I I I

The notion that the increase in circulation is dueto arbitrary interference by the banks owes itsorigin to the widespread view that Banks of Issueare the exclusive or predominant agencies whichcan change the volume of the circulation; andthat they do so of their own free will. But theCentral Banks are by no means the only factorcapable of bringing about a change in the volume ofcirculating media5*; they are, in their turn, largelydependent upon other factors, although they caninfluence or compensate for these to a great extent.Altogether, there are three elements which regulatethe volume of circulating media within a country— changes in the volume of cash, caused by

* This fact has already been pointed out by the representatives ofthe Banking School, and later by C. Juglar (DM change et de la liberttd'dmission, Paris, 1868. Chap. I l l , passim; and Des crises commercialeset leur retour periodique, 2nd edit. Paris, 1889, p. 57). Wicksell (Geldzinsund Guterpreise, p. 101) also points, first of all, to the deposit businessof the banks as the cause of the 'elasticity' of the volume of currencymedia.

148


inflows and outflows of gold; changes in the notecirculation of the Central Banks: and last, and inmany ways most important, the often-disputed'creation' of deposits by other banks. The inter-relations of these are, naturally, complicated.

As regards original changes in the first twofactors —that is, changes which are not set inmotion by changes in one of the other factors —there is comparatively little to say. It has alreadybeen pointed out that, in principle, an increasein the volume of cash, occasioned by an increasein the volume of trade, also implies a lowering ofthe money rate of interest — which gives rise toshifts in the structure of production which seem,though only temporarily, to be advantageous.It must certainly appear very problematicalwhether the deviations in the money rate of interestthus occasioned would, as a rule, be large enoughto cause fluctuations of an empirically ascertainablemagnitude. Central Banks, on the other hand, areby law or custom bound to preserve such a closeconnection between note issues and cash holdingsthat we have no reason to assume that they, andthey alone, provide the original impetus. Of

149


course, it is possible to assume, with ProfessorMises, that the Central Banks, under the pressureof an inflationist ideology, are always trying toexpand credit and thus provide the impetus fora new upward swing of the Trade Cycle; and thisassumption may be correct in many cases. Thecredit expansion is then conditioned by specialcircumstances, which need not always be present;and the cyclical fluctuations caused by it are,therefore, not the necessary consequence of aninherent tendency of our credit system, for theremoval of the special circumstances would elimin-ate them. But before deciding in favour of thisspecial assumption — which requires a proof ofits own, to be given separately in the case of eachcycle — we have to ask whether, in some other partof our credit system, such extensions may nottake place automatically under certain conditions— without the necessity for any special assumptionof the inadequate functioning of any part of thesystem. To me this certainly appears to be trueas regards the third factor of money expansion —the 'credit creation* of the commercial banks.

There are few questions upon which scientific

150

CAUSE OF CYCLICAL FLUCTUATIONSliterature, especially in Germany, is so lacking inclarity as on the possibility and importance of anincrease in circulating media due to the grantingof additional credits by the banks of deposit.To give an answer to the question whether credit-creation is a regular consequence of the existingorganization of banking, we shall have to attemptto clear up our conception of the methods andextent of such credit creation by deposit banks.Besides dealing with the fundamental question ofthe possibility of credit creation and the limits towhich it can extend, we shall have to discuss twospecial questions which are important for ourfurther investigations: namely, whether the practi-cal importance of credit creation depends uponcertain practices of banking technique, as is oftenassumed; and secondly, whether it is, in fact,possible to determine whether a given issue ofcredit represents credit freshly created or not.

If in the course of our investigation, it is possibleto prove that the rate of interest charged by thebanks to their borrowers is not promptly adjustedto all changes in the economic data (as it would beif the volume of money in circulation were constant)


•—either because the supply of bank credits is,within certain limits, fundamentally independentof changes in the supply of savings, or becausethe banks have no particular interest in keepingthe supply of bank credit in equilibrium with thesupply of savings and because it is, in any case,impossible for them to do so — then we shallhave proved that, under the existing creditorganization, monetary fluctuations must inevit-ably occur and must represent an immanentfeature of our economic system — a feature deserv-ing of the closest examination.

IV

The main reason for the existing confusionwith regard to the creation of deposits is to befound in the lack of any distinction between thepossibilities open to a single bank and those opento the banking system as a whole.* This is

* As it is impossible to deal exhaustively with this problem, it mustbe sufficient to draw attention to the main literature of the subject.The first author known to me who definitely stated that 'the balancesin the bank are to be considered in very much the same light with thepaper circulation,' was Henry Thornton (see his evidence before theCommittee on the Bank Restriction, 1797). The development of amore definite theory of credit creation by the banks began, however,


connected with the fact that, in Germany, thewhole theory has been taken over bodily fromEngland, where, owing to differences in banking

with the criticisms levelled by the Banking School against the CurrencySchool, and represent the former's only correct contribution to thescience of economics. As Professor T. E. Gregory has recently shown{Introduction to Tooke and Newmarch's History of Prices, London, 1928,pp. 11 et seq.) it was James Pennington who originally developed thisthesis, first in an appendix to T. Tooke's Letter to Lord Grenvilleon the Effects ascribed to the Resumption of Cash Payments, then infurther contributions to R. Torrens' Letter to the Rt. Hon. ViscountMelbourne (London, 1837) and finally in an appendix to the thirdvolume of Tooke's History of Prices (1838). If one wanted to tracethe further progress of this theory during the nineteenth century,one would have to draw particular attention to the writings of H. D.Macleod (cf., in particular, his Dictionary of Political Economy,London 1863, article on Credit), C. F. Dunbar and F. Ferrara.

Modern developments follow the exposition of H. J. Davenport{The Economics of Enterprise, New York, 1915, pp. 250 et seq.); andmention should, in particular, be made of C. O. Pnillips'sBan^ Credit,New York, 1920 (especially Chap. 111, 'The Philosophy of BankCredit'), of W. F. Crick {The Genesis of Bank Deposits, 'Economica',vol. vii, No. 20-June 1927) and R. G. Rodkey {The Banking Process,New York, 1928). Apart from these, we must include in our list thewell-known works of Hartley Withers, Irving Fisher and R. G. Haw-trey and, in German literature, K. Wicksell {Geldzins und Giiterpreise,p. 101), A. Weber {Depositenbanken und Spekulationsbanken, 2nd edit.,1922), the works which we have already mentioned of Mises and Hahn,G. Haberler's essay on the latter {Hahns Volkawirtschaftliche Theoriedes Bankkredits, Archiv fur Sozialwissenschaften, vol. 57, 1927) and,finally, H. Neisser {Der Tauschwert des Geldes, Jena, 1928).

The theory has been severely criticized especially by ProfessorCarman, W. Leaf, and more recently by R. Reisch {Die 'Deposit* -legende in der Banktheoriet Zeitschrift fur Nationalokonomie, vol. i,1930.)

153


technique, the limits imposed on any individualbank are, perhaps, somewhat less narrow, so thatthe general possibilities open to the bankingsystem as a whole have not been indicated withthe degree of emphasis which their importancedeserves. In Germany, following the popular ex-position of Mr. Hartley Withers, the most gener-ally accepted view starts from English banking prac-tice which (except in the case ofloverdrafts') creditsthe account of the customer with the amountborrowed before the latter is actually utilized.Granted this assumption, the process leading toan increase of circulating media is comparativelyeasy to survey and therefore hardly ever disputed.So long and in so far as the credits which a bankis able to grant, considering its cash position,remain on current account —and in the UnitedStates, for example, it is a regular condition for thegranting of a loan that the current account of theborrower shall never fall below a certain relativelyhigh percentage of the sum borrowed* — everynew grant of credit must, of course, bring aboutan equivalent increase of deposits and a propor-

• Cf. C. O. Phillips, op. cit., p. 50.

CAUSE OF CYCLICAL FLUCTUATIONStionately smaller diminution of cash reserves.Against these 'deduced deposits' (Phillips) whichregularly occur in the normal course of business,the banks naturally have to keep only a certainpercentage of cash reserve; and thus it is clearthat every bank can, on the basis of a given increaseof deposits resulting from public payments, grantnew credits to an amount exceeding this increasein deposits.

Against this method of proof it can rightly beobjected that, while banking practices of this kindmay well lead to the possibility of credit creation,the conditions which this argument assumes are notpresent on the Continent. It has been justifiablyand repeatedly emphasized that there is no reasonwhy the borrower, so long as he is not forced to doso, should borrow money at a higher rate ofinterest merely to leave that money on deposit ata lower rate.f

If the possibility of creating credit dependedt R. Reisch in 'Die Wirtschaftliche Bedeutung des Kredites im

Lichte von Theorie und Praxis' (Mitteilungen des Verbandes Sster-reickischen Bankenund Bankiers, ioth year, Nos. 2-3, Vienna, 1928,p. 38) and A. Johr in his verbal report on Credit and the Cycle, in theZurich Assembly of the Verein fiir Sozialpolitik (Schriften, vol. 175,P- 3 " ) .


only on the fact that borrowers leave part oftheir loans on current account for a time, thencredit creation would be practically impossibleon the Continent; * while even in England and theUnited States it would have only a very secondaryimportance. It should be noted that this appliesto the case in which the borrower pays the sumborrowed into another account in the same bank,so that it is transferred from one to the otherwithout diminishing the total volume of depositsin the bank concerned. We need not, therefore,go separately into this case.

But, in adopting this line of argument, by farthe most important process by which deposits arecreated in the course of current banking businesseven in Anglo-Saxon countries is neglected, andthe sole way in which they are created on theContinent is left entirely out of consideration.The latter could easily be overlooked, since theability of individual banks to make an increasein their deposits the basis of a far greater amount

* As Bouniatian, evidently for this reason, actually assumes, (cf.his essay, 'Industrielle Schwankungen, Bankkredit und Warenpreise',Archiv fur Sozialwissenschaften und Sozialpolitik, vol. 58, Tubingen,1927, p. 463.)

156

CAUSE OF CYCLICAL FLUCTUATIONSof new credit can only be accounted for by meansof the assumptions used above, while in the bank-ing system as a whole the same process occursindependently. In the following pages, therefore,we shall examine how an increase in deposits,paid in in cash, influences the lending capacityof the whole banking system; starting from theassumption, more appropriate to Continentalconditions, that the sums granted will be creditedto the account of the borrower only at the timewhen, and to the extent that, he makes use of them.

We may start as before by examining the pro-cedure of a single bank. At this bank a certainamount of cash is newly deposited; a sum, let ussay, equal to 5 per cent of its previous totaldeposits. If the policy of the bank was to keep areserve of 10 per cent against deposits, that ratiohas now been increased, by the new deposit,to 14.3 per cent, and the bank is therefore in aposition, in accordance with its policy, to grant newcredits. If we assume further that it re-lends90 per cent of the newly deposited money and that

MONETARY THEORY AND THE TRADE CYCLEthe whole of this is immediately utilized by theborrower (in order, let us say, to increase hispurchases of raw materials) then the ratio of cashto deposits has again sunk to 10 per cent. In sofar as the bank does not change its policy itsindividual lending capacity is exhausted, in thesecircumstances, before it has even re-lent the wholeof the amount newly deposited.

The effect of the sums newly deposited at onebank on the lending capacity of the whole bankingsystem is, however, not exhausted by this trans-action. If the borrower does not use the creditin a way which leads quickly to the market forconsumers' goods, such as wage payments, butdevotes it instead to the purchase of raw materialsor half-finished products, then it is to be assumedthat payment will be made by cheque and that theseller will hand over the sum received to his ownbank for encashment, the amount being credited tohis own account. The next consequence must bethat the clearing-house position of this bankimproves by exactly the amount transferred, andit therefore obtains an equivalent amount of cashfrom the bank which originally granted the credit.

158


For the second bank, therefore, the sum originat-ing in the granting of credit and paid into itsaccounts (representing, as we remember, 90 percent of the original deposit) is just as much anoriginal deposit, based on cash payments, as itwas to the bank which we originally considered.It will, therefore, be regarded as a basis foradditional lending and used in just the same wayas any other new deposit. If the second bank alsokeeps 10 per cent of its deposits as cash reserves, ittoo will be in a position to lend 90 per cent of the newdeposit, and the same process will be continuedas long as the amounts are merely transferredfrom bank to bank and are not taken out in cash.As every bank re-lends 90 per cent of the amountpaid into it and thus causes an equivalent inoreasein deposits for some other bank, the originaldeposit will give rise to credits representingo.9+o.92+o.93-fo.9+ times the originalamount. As the sum of this converging infiniteseries is 9, the banks will be enabled, in anextreme case, to create, against an amount ofcash flowing in from an outside source, creditsequal to nine times that amount. This becomes


clear when we consider that the process can onlystop when the last part of this cash is required forthe 10 per cent reserve of the deposits.

For simplicity's sake we have made use of anassumption which is undoubtedly incorrect, butwhich affects our conclusion only in so far as itreduces the actual amount of new credit which thebanks can create with a reserve ratio of 10 percent. Its omission leaves our fundamental con-clusion intact; i.e. that they can grant credit toan amount several times greater than the sumoriginally deposited. In fact some part of thecredit at least, if not on the first then on subsequentoccasions, will always be withdrawn in cash andnot deposited with other banks. For example,if 70 per cent is always redeposited instead of ̂ befull 90 per cent this amount being re-lent byevery bank and the remainder being used in cashtransactions, then the increase in deposits will giverise to additional credits equal to only 0.7+0.7*+0.73 times (i.e. two and one-third times) theoriginal. So long as any part of the credits grantedare not withdrawn in cash but redeposited withthe banks, the latter will be able to create addi-

160


tional credits, of a larger or smaller amount, as aconsequence of every increase in their cash hold-ings.* The lifetime of this pyramid of credit islimited to that of the first credit granted, save inthe case (which can be assumed as long as thereare no withdrawals from deposits) where it isimmediately replaced by a fresh credit. If,however, deposits unexpectedly diminish at anypart of the banking system, the process will bereversed, and the original diminution of depositswill occasion a contraction of credit correspond-ingly exceeding the amount withdrawn.f

# The maximum amount of credit, to the creation of which theincrease in the cash holdings of the banks may give rise under such anassumption, is easily found by inserting the factor representing theproportion of the original deposit which is re-lent and redepositedwith another bank into the mathematical formula expressing thelimit which a convergent geometrical series approaches, viz., x=rt

The result gives the total of credits which originate in the series oftransactions, including the original deposit; and, in order to arrive atthe amount of additional credits, i has to be subtracted from theresult. It is thus easily seen that even if, for example, only i-oth ofthe 90 per cent re-lent by the first bank, or 10 per cent of the originaldeposit, is redeposited with another bank-and this process isrepeated, additional credits amounting to o . m times the originaldeposit will be created.

f On this question, and on the interesting effects of a transferenceof deposits from one bank to another, cf. the more elaborate treatmentof C. O. Phillips, op. cit, p. 64 seq.; also the remarks of W. F. Crick,loc. citt p. 196.


In this connection we must note for furtheremphasis later the fact that the proportion inwhich the credits granted are transferred to otheraccounts —and not paid out in cash —must beregarded as subject to very wide fluctuations asbetween different individuals at a given moment,as well as between various periods of time for theeconomic system as a whole. We return later tothe significance of this fact.

What has been said above should be sufficientto show that the possibility of creating creditsover and above the sums deposited — which,under Continental banking conditions, is notopen to any individual bank — is, however, opento the whole banking system of the country to aconsiderable extent. The fact that a single bankcannot do what is automatically done by the bank-ing system as a whole also explains another circum-stance, which might otherwise easily be cited as aproof of the impossibility of additional creditcreation. If every bank could re-lend severaltimes the amount deposited, there would be noreason against its offering a much higher rate ofinterest on deposits than it actually does, or, in

162

CAUSE OF CYCLICAL FLUCTUATIONSparticular, under the existing discount rates of theCentral Banks, against its procuring cash inunlimited quantities by way of re-discount; for itwould only have to charge its customers a smallpart of the rate of interest charged by the banksin order to make the business pay. This apparentcontradiction between theory and practice iscleared up as soon as one realizes that an increaseof deposits by a single bank only offers possibilitiesfor credit creation to the banking system as a whole.But the importance of this circumstance transcendsthe mere clearing up of this difficulty.

VI

As credits created on the basis of additionaldeposits do not normally appear in the accountsof the same bank which granted the credit, it isfundamentally impossible to distinguish, in in-dividual cases, between * those deposits whicharose through cash payment and those whichfind their origin in credit.'* But this consideration

• Neisser (op. cit.,p. 53) deserves credit for clearing up an untenableconception, which was quite recently held by no less an authoritythan Professor J. Schumpeter (Theorie der toirtschaftlichen Entwick-lung, Munich and Leipzig, 2nd edit., 1926, p. 144).

•63


rules out, a priori, the possibility of bankerslimiting the amount of credit granted by them tothe amount of 'real' accumulated deposits — thatis, those arising from the accommodation oftemporarily unused money. The same fact enablesus to understand why it is generally just thoseeconomic writers who are also practical bankerswho are most unwilling to admit in any circum-stances that they are in a position to create credits.*'The banker simply does not notice that throughthis process there is an increase in the amount ofmoney in circulation.'f Once the impetus hasbeen given to any part of the banking system, mereadherence to the routine of banking techniquewill lead to the creation of additional deposits

• Cf., for example, Walter Leaf, the late chairman of the West-minster Bank, in his book Banking (Home University Library,London, 1926), or the contributions of A. Johr and B. Dernburg tothe Zurich Debate on the Trade Cycle. (Schrtften des Vereines fUrSozialpolitik, vol. 175, 1929, pp. 311 and 329). These argumentswere perfectly correctly answered by another 'practical* banker, K.Schlesinger (ibid., p. 355). Professor A. Hahn, on the other hand, fallsinto the opposite error. The standpoint of Professor R. Reisch will bediscussed later.

f Neisser, op. cit., p. 54. He goes on to say, quite correctly, that'the mere fact that cheque-deposits represent money, without beingcovered by cash up to 100 per cent, already explains the money-creat-ing nature of bank credit.'

164


without the possibility arising, at any point, ofdetermining whether any particular credit shouldproperly be regarded as 'additional.' Every timemoney which has been deposited is re-lent —provided that the depositor is not preventedfrom using his deposits for making payments —this process is to be regarded as the creation ofadditional purchasing power; and it is merelythis comparatively simple operation which is atthe root of the banks' ability to create purchasingpower —- although the process appears so mysteri-ous to many people. It is thus by no meansnecessary that the banks should grant thesecredits, as Dr. Dernburg seems to assume, in an'improper or wanton* way.

It is of course quite another question whetherbankers can, or do, create additional credits oftheir own free will. The objections to this theoryof additional credits, which are levelled againstthe statement that the banks create credit 'asthey please', although holding good at a givenrate of interest, do not in the least affect thatpart of the theory which we need for our analysis.If Professor Reisch, for example, emphasizes that

165


bank deposits generally increase only * accordingto the needs of business/* or if Prof. Bouniatianobjects that 'it does not depend on the banks,but on the demands made by commerce andindustry, how far banks expand credit/f then theseassertions, coming as they do from opponentsof the theory of bank credits, already contain allthat is needed for a deductive proof of the necessityfor the recurrence of credit cycles. What interestsus is precisely the question whether the banksare able to satisfy the increased demands ofbusiness men for credits without being obligedimmediately to raise their interest charges —aswould be the case if the supply of savings and thedemand for credits were to be in direct contact,without the agency of the banks (as for example inthe hypothetical Savings market' of theory); orwhether it is even possible for the banks to raisetheir interest charges immediately the demand forcredits increases. Even the bitterest opponentsof this theory of bank credit are forced to admitthat * there can be no doubt that, with the upward

• Op. cit., p. 39. t Op. cit., p. 465.

166


swing of the Trade Cycle, a certain expansion of

bank credits takes place.'*We must not, however, be satisfied with register-

ing the general agreement of opinion on thispoint. Before passing on to analyse the conse-quences of this phenomenon we must ask whetherthe causes which bring it about that banks increasetheir deposits through additional credits in periodsof boom and thus postpone, at any rate temporarily,the rise in the rate of interest which would other-wise necessarily take place, are inherent in thenature of the system, or not.

v 11

So far, the starting point of our argumentconcerning the origin of additional credits hasbeen the assumption that the banks receive anincreased in-flow of cash which they then use as abasis for new credits on a much larger scale. Wemust now inquire how banks behave when anincreased demand for credit makes itself felt.

* Dernburg, op. cit.t p. 329. He merely adds to this statement theremark that the banks and the Central Bank should* see to it that thisexpansion is 'kept in order'!

167

MONETARY THEORY AND THE TRADE CYCLEAssuming, as is preferable, that this increaseddemand was not caused by a lowering of their owninterest rates, this additional demand is always asign that the natural rate of interest has risen —that is, that a given amount of money can nowfind more profitable employment than hitherto.The reasons for this can be of very different kinds.*New inventions or discoveries, the opening up ofnew markets, or even bad harvests,f the appearanceof entrepreneurs of genius who originate 'newcombinations' (Schumpeter), a fall in wage ratesdue to heavy immigration; and the destruction ofgreat blocks of capital by a natural catastrophe,or many others. We have already seen that none ofthese reasons is in itself sufficient to accountfor an excessive increase of investing activity,which necessarily engenders a subsequent crisis;but that they can lead to this result only throughthe increase in the means of credit which theyinaugurate.

* 'A great variety of causes,' observes R. G. Hawtrey, verycorrectly (Trade and Credit, London, 1928, p. 175).

t Regarding the influence of harvests on the Trade Cycle, cf. theuseful compilation of various contradictory theories by V. P. Timos-henko, The Role of Agricultural Fluctuations in The Business Cycle(Michigan Business Studies, vol. ii, No. 9, 1930).

168


But how is it possible for the banks to extendcredit, as they undoubtedly do, following anincrease in demand, when no additional cash isflowing into their vaults? There is no reason toassume that the same cause which has led to anincreased demand for credit will also influenceanother factor, the cash position of the banks —which as we know is the only factor determiningthe extent to which credit can be granted.* Solong as the banks maintain a constant proportionbetween their cash reserves and their deposits itwould be impossible to satisfy the new demand forcredit. The fact that in reality deposits always doexpand relatively to cash reserves, in the course ofthe boom, so that the liquidity of the banks isalways impaired in such periods, does not ofcourse constitute a sufficient starting point for anargument in which the increase in credits is

* It is of course possible that an improvement in the conditions ofproduction and profit-making will also indirectly cause an increasedflow of cash to the banks, for a flow of funds for investment, as wellas an increased flow of payments for goods, can be expected fromabroad. But, in the first place, this increased flow of cash can only beexpected in a comparatively late stage of the boom, so that it canhardly explain the latter's origin; and in the second place, such anexplanation could only be adduced in the case of a single country,and not for the world economy as a whole, or in a closed system.

MONETARY THEORY AND THE TRADE CYCLEregarded as the decisive factor determining thecourse and extent of the cyclical movement. Wemust attempt to understand fully the causes andnature of this credit expansion and in particular,its limits.

The key to this problem can only be found inthe fact that the ratio of reserves to deposits doesnot represent a constant magnitude, but, asexperience shows, is itself variable. But we shallachieve a satisfactory solution only by showingthat the reason for this variability in the reserveis not based on the arbitrary decisions of thebankers, but is itself conditioned by the generaleconomic situation. Such an examination of thecauses determining the size of the reserve ratiodesired by the banks is all the more importantsince we had no theoretical warrant for ourprevious assumption that it always tends to beconstant.

It is best to begin our investigation by consider-ing once again the situation of a single bank, andasking how the manager will react when the creditrequirements of the customers increase in conse-quence of an all-round improvement in the

170

CAUSE OF CYCLICAL FLUCTUATIONSbusiness situation.* For reasons which willshortly become clear, we must assume that thebank under consideration is the first to feel thenew credit-requirements of industry, because,let us say, its customers are drawn from just thoseindustries which first feel the effects of the newrecovery. Among the factors which determinethe volume of loans granted by the bank, only onehas changed; whereas previously, at the same rateof interest and with the same security, no newborrowers came forward, now, under the sameconditions of borrowing, more loans can be placed.On the other hand, the cash holdings of the bankremain unchanged. This does not mean, however,that the considerations of liquidity which dictatethe amount of loans to be granted will lead to thesame result now as when fresh loans could onlyhave been placed at a lower rate of interest or withinferior security than was the case with loansalready granted. In this connection, finally, wemust mention that the sums which we have, for

* The problems with which the manager of a single bank is con-fronted in deciding the bank's credit policy are very neatly analysedby Mr. W. F. Crick, op. cit., p. 197, et seq.

MONETARY THEORY AND THE TRADE CYCLEsimplicity's sake, hitherto called cash balances,and which form the bank's liquid reserve, are byno means exclusively composed of cash —andare not even of a constant magnitude, unrelated tothe size of the profits which they make possible.The danger that, in case of need, the reservesmay have to be replenished by rediscounting billsthrough the Central Bank*; or that, in order tocorrect an unfavourable clearing-house balance,day-money may have to be borrowed at a givenrate of interest, is far less abhorrent when it ispossible to extend credits at an undiminished rateof interest than when such an extension wouldinvolve a lowering of that rate. But even disregard-ing this possibility and assuming that the bankrecognizes that it can satisfy its eventual need forcash only at correspondingly higher rates, we cansee that the greater loss of profit entailed by keep-ing the cash reserve intact will, as a rule, lead the

# On this point see J. S. Lawrence, 'Borrowed Reserves and BankExpansion* (Quarterly Journal of Economics, vol. xlii, Cambridge,Mass., 1928) where Mr. Phillips's exposition, mentioned above, isextensively criticized; also the rejoinder of Mr. F. A. Bradford,published under the same title in the next volume (xliii) of thesame journal.

172


bank to a policy which involves diminishing thesize of this non-earning asset. Besides this, wehave the consideration that, in the upward phaseof the cycle, the risks of borrowing are less; andtherefore a smaller cash reserve may suffice toprovide the same degree of security. But it isabove all for reasons of competition that the bankwhich first feels the effect of an increased demandfor credits cannot afford to reply by putting up itsinterest charges; for it would risk losing its bestcustomers to other banks which had not yetexperienced a similarly increased demand forcredits. There can be little doubt, therefore, thatthe bank or banks which are the first to feel theeffects of new credit requirements will be forcedto satisfy these even at the cost of reducing theirliquidity.

V I I I

But once one bank or group of banks hasstarted the expansion, then all the other banksreceive, as already described, a flow of cash whichat first enables them to expand credit on their ownaccount without impairing their liquidity. They

173


make use of this possibility the more readily sincethey, in turn, soon feel the increased demand forcredit. Once the process of expansion has becomegeneral, however, the banks soon realize that, forthe moment at any rate, they can safely modifytheir ideas of liquidity. While expansion by asingle bank will soon confront it with a clearing-house deficit of practically the same magnitudeas the original new credit, a general expansioncarried on at about the same rate by all bankswill give rise to clearing-house claims which,although larger, mainly compensate one anotherand so induce only a relatively unimportant cashdrain. If a bank does not at first keep pace withthe expansion it will, sooner or later, be inducedto do so, since it will continue to receive cash atthe clearing house as long as it does not adjustitself to the new standard of liquidity.

So long as this process goes on, it is practicallyimpossible for any single bank, acting alone, toapply the only control by which the demand forcredit can, in the long run, be successfully keptwithin bounds; that is, an increase in its interestcharges. Concerted action in this direction,

CAUSE OF CYCLICAL FLUCTUATIONSwhich for competitive reasons is the only actionpossible, will ensue only when the increased cashrequirements of business compel the banks toprotect their cash balances by checking furthercredit expansion, or when the Central Bank haspreceded them by raising its discount rate. This,again, will only happen, as a rule, when the bankshave been induced by the growing drain on theircash to increase their re-discount. Experienceshows, moreover, that the relation between cheque-payments and cash payments alters in favour ofthe latter as the boom proceeds, so that anincreased proportion of the cash is finally with-drawn from the banks.*

This phenomenon is easily explained in theoryby the fact that a low rate of interest first raisesthe prices of capital goods and only subsequentlythose of consumption goods, so that the first

* Cf. the statements contained in the well-known ioth yearlyReport of the Federal Reserve Board, for 1923 (Washington, 1924) p.25: 'This is the usual sequence - an increase in deposits followed by anincrease in currency. Ordinarily the first effect of an increase inbusiness activity on the banking position is a growth in loans anddeposits . . . Then comes a time when the increase in business activityand the fuller employment of labour and increased pay-roll call foran increase in actual pocket money to support the increased wagedisbursements and the increased volume of purchases in detail.'

X75


increases occur in the kind of payments which areeffected in large blocks.* It may lead to theconsequence that banks are not only preventedfrom granting new credits, but even forced todiminish credits already granted. This fact maywell aggravate the crisis; but it is by no meansnecessary in order to bring it about. For this it isquite enough that the banks should cease to extendthe volume of credit; and sooner or later this musthappen. Only so long as the volume of circulatingmedia is increasing can the money rate of interestbe kept below the equilibrium rate; once it hasceased to increase, the money rate must, despitethe increased total volume in circulation, riseagain to its natural level and thus render unprofit-able (temporarily, at least) those investmentswhich were created with the aid of additionalcredit.f

# Neisser {op. cit.t p. 162) doubts this, but his criticism resultsfrom an inadequate grasp of the effects of an unduly low money rate ofinterest. But even if he were right on this point, the arguments ofmonetary Trade Cycle theory would remain unaffected, since thelatter, as is shown in the text, does not depend on this assumption forits proof.

t We need not stay to examine the case of a continuous increase incirculating media, which can only occur under a free paper standard.

176


I X

The assertion which forms the starting point ofthe Additional Credit Theory of the Trade Cycle',and whose proof has been attempted in thepreceding pages, has never in fact been seriouslyquestioned; but hardly any attempts have beenmade to follow up all the unpleasant consequencesof the state of affairs it indicates. Yet what isimplied when the beneficial effects of bank creditsare praised but that thanks to the activities of banksan increased demand for credit is followed by agreater increase in its supply than would bewarranted by the supply of contemporary saving?Wherein lie the often praised effects of credit,if not in the fact that it provides means for enter-prises for which no provision could be found ifthe choices of the different economic subjectswere strictly followed? By creating additionalcredits in response to an increased demand, andthus opening up new possibilities of improvingand extending production, the banks ensure thatimpulses towards expansion of the productive

M 177


apparatus shall not be so immediately and insuper-ably balked by a rise of interest rates as they wouldbe if progress were limited by the slow increasein the flow of savings. But this same policystultifies the automatic mechanism of adjustmentwhich keeps the various parts of the system inequilibrium, and makes possible disproportionatedevelopments which must, sooner or later, bringabout a reaction.

Elasticity in the credit supply of an economicsystem, is not only universally demanded but also— as the result of an organization of the creditsystem which has adapted itself to this require-ment — an undeniable fact, whose necessity oradvantages are not discussed here.* But we mustbe quite clear on one point. An economic systemwith an elastic currency must, in many instances,react to external influences quite differently from aneconomy in which economic forces impinge on goodsin their full force — without any intermediary; andwe musty a priori, expect any process started by an

* Cf. K. Wicksell, Geldzins und Giiterpretse, p. 101, 'The moreelastic is the currency system the longer can a more or less constantdifference persist between the two interest rates and the greater,therefore, will be the influence of this discrepancy on prices.'

178


outside impulse to run an entirely different course insuch an economy from that described by a theorywhich only takes into account changes originatingon the side of goods. Once, owing to the disturbinginfluence of money, even a single price has beenfixed at a different level from that which it wouldhave formed in a barter economy, a shift in thewhole structure of production is inevitable; andthis shift, so long as we make use of static theoryand the methods proper to it, can only be ex-plained as an exclusive consequence of thepeculiar influence of money. The immediateconsequence of an adjustment of the volume ofmoney to the 'requirements' of industry is thefailure of the 'interest brake' to operate as promptlyas it would in an economy operating withoutcredit. This means, however, that new adjust-ments are undertaken on a larger scale than canbe completed; a boom is thus made possible, withthe inevitably recurring 'crisis.' The determiningcause of the cyclical fluctuation is, therefore, thefact that on account of the elasticity of the volumeof currency media the rate of interest demanded bythe banks is not necessarily always equal to the

179


equilibrium rate, but is, in the short run, determined

by considerations of banking liquidity*

* In a previous work (Die Wahrungspolitik der VereinigtenStaaten, op. cit.t p. 260), I have already dealt with the elasticity ofbank credit as the cause of cyclical fluctuations. This view of itsdetermining importance is now also put forward by Professor F. A.Fetter in a very interesting essay, 'Interest Theories and PriceMovements' (American Economic Review, vol. xvii, supplement,March 1927; see especially, pp. 95 et seq.). Prof. Fetter, of course, isalso under the influence of the prevailing dogma which holds that theexistence of a stable price level is sufficient proof of the absence of allmonetary influences. The crucial part of his argument, not havingreceived the attention which it deserves in recent monetary literature,is reprinted here:

'The foregoing presents the extreme case of the expansion andcontraction of bank loans in relation to prices, but in principle quitesmall changes in the loan policies of banks affecting the volume of com-mercial loans, discount rates, and percentages of reserves, are of thesame nature. They cause and constitute inflation and deflation of theexchange medium and of commercial purchasing power, not originat-ing in the amount of standard money but in the elasticity of bankingloan funds. This word "elasticity" has long been used in discussions ofbanking policy to designate a quality assumed to be wholly desirable inbank note issues and customers' credits, but with only vague suggestionsas to what is the need, standard, or means, with reference to whichbank loans should expand and contract.

'Rather, it may be more exact to say, the tacit assumption has beenthat the bank loan funds should be elastic in response to the "needs ofbusiness." But"theneeds of business" appears to be nothing but anothername for changes in customers' eagerness for loans; and this eagernessincreases when prices are beginning, or are expected, to rise and oftencontinues to gather momentum while prices rise and until, because ofvanishing reserve percentages (and other factors), the limit of thiselasticity and also the limit of price increase are in sight. In thissituation the most conservative business operations become inter-mixed with elements of investment speculation, motivated by the

CAUSE OF CYCLICAL FLUCTUATIONSThe main question set by this inquiry is thus

answered. A deductive explanation embracingall the phenomena of the Trade Cycle wouldrequire far-reaching logical investigations entirelytranscending the scope of this work, which aimsmerely at an exposition of the monetary basis ofTrade Cycle theory. For the present, we mustcontent ourselves with a reference to existingliterature on the subject.* In the present work

rise of prices and the hope of profit that will be made possible by afurther rise. Throughout this process the much-esteemed elasticity ofbank funds is the very condition causing, or making possible, the risingprices which stimulate the so-called "needs of business". Truly a viciouscircle, to be broken only by crisis and collapse when bank loans reach alimit and prices fall.' (My italics.)

Further, we should point out the connection between our theoryand a famous thesis of Mr. R. G. Hawtrey. The phrase 'so long ascredit is regulated with reference to reserve proportions, the tradecycle is bound to recur* (Monetary Reconstruction, 2nd edit.,London, 1926, p. 135) is undoubtedly correct, though perhaps in asense somewhat different from that intended by the author; for aregulation of this volume of loans exclusively from the point of viewof liquidity can never effect a prompt adjustment of the rates chargedon loans to the changes in the equilibrium rate, and thus cannothelp providing opportunities for the temporary creation of additionalcredits as soon as (at a given rate of interest) the demand for creditsurpasses the accumulation of savings; that is, when the natural rateof interest has risen. See, finally, the remarks of Professor W.Ropke, Kredit und Konjunktur, op. cit., p. 274.

* Besides Professor Mises' Theorie des Geldes und der Umlaufs-mitt el we must mention the last chapter of S. Budge's Griindzugeder Theoretischen Nationalokonomie (Jena, 1925) and Prof. Strigl's

181


we shall only draw a few conclusions which followfrom our previous arguments, some with regard topractical policy, some with regard to furtherscientific research. Before going on to this, how-ever, we shall venture a few remarks on thequestion whether the result of our investigationsunequivocally settles the controversy between theprotagonists and opponents of the monetaryTrade Cycle theory in favour of the former.

xIt must be emphasized first and foremost that

there is no necessary reason why the initiatingchange, the original disturbance eliciting a cyclicalfluctuation in a stationary economy, should be ofmonetary origin. Nor, in practice, is this evengenerally the case. The initial change need haveno specific character at all, it may be any oneamong a thousand different factors which may at

paper on "Die Produktion unter dem Einfluss einer Kredit-expansion' in vol. 173-ii of the Schriften des Vereins fur Sozial-politik, concerning Trade Cycle theory and business research(Munich and Leipzig, 1928), a volume which has been repeatedlyquoted above. Since the above was written, I have tried to carry theanalysis of these phenomena a step further in Prices and Production(London, 1931).

182

CAUSE OF CYCLICAL FLUCTUATIONSany time increase the profitability of any group ofenterprises. For it is not the occurrence of a* change of data' which is significant, but the factthat the economic system, instead of reacting tothis change with an immediate 'adjustment*(Schumpeter) — i.e. the formation of a newequilibrium — begins a particular movement of'boom' which contains, within itself, the seeds ofan inevitable reaction. This phenomenon, as wehave seen, should undoubtedly be ascribed tomonetary factors, and in particular to 'additionalcredits' which also necessarily determine theextent and duration of the cyclical fluctuation.Once this point is agreed upon, it naturallybecomes quite irrelevant whether we label thisexplanation of the Trade Cycle as a monetarytheory or not. What is important is to recognizethat it is to monetary causes that we must ascribethe divergences of the pricing process, during theTrade Cycle, from the course deduced in statictheory.

From the particular point of view from whichwe started, our theory must be regarded mostdecisively as a monetary one. As to the incorpor-

183


ation of Trade Cycle theory into the generalframework of static equilibrium theory (for theclear formulation of which we are indebted toProfessor A. Lowe, one of the strongest opponentsof monetary Trade Cycle theory), we mustmaintain, in opposition to his view, not onlythat our own theory is undoubtedly a monetaryone but that a theory other than monetary ishardly conceivable.* It must be conceded thatthe monetary theory as we have presented it —whether one prefers to call it a monetary theoryor not, and whether or not one finds it asufficient explanation of the empirically determinedfluctuations-has this definite advantage: it dealswith problems which must, in any case, be dealtwith for they are necessarily given when the centralapparatus of economic analysis is applied to theexplanation of the existing organization of exchange.Even if we had never noticed cyclical fluctuations,even if all the actual fluctuations of history wereaccepted as the consequences of natural events, a

• Cf. my report: 'Uber den Einfluss monetarer Faktoren auf denKonjunkturzyklus' {Schriften des Vereins fur Sozialpolitik, vol. 173-ii, p. 362 et seq.).

184


consequential analysis of the effects which follow fromthe peculiar workings of our existing credit organiza-tion would be bound to demonstrate that fluctuationscaused by monetary factors are unavoidable.

It is, of course, an entirely different questionwhether these monetary fluctuations would, ifnot reinforced by other factors, attain the extentand duration which we observe in the historicalcycles; or whether in the absence of these supple-mentary factors they would not be much weakerand less acute than they actually are. Perhaps theempirically observed strength of the cyclicalfluctuations is really only due to periodic changesin external circumstances, such as short-periodvariations of climate, or changes in subjective data(as e.g., the sudden appearance of entrepreneursof genius) or perhaps the interval between in-dividual cyclical waves may be due to some naturallaw.* Whatever further hypothetical causes areadduced to explain the empirically observedcourse of the fluctuations, there can be no doubt

* From now to the end of the section the exposition follows, inpart word for word, my contribution to the Zurich discussion of the'Verein fur Sozialpolitik.' {pp. cit., p. 37a et seq.)

185


(and this is the important and indispensablecontribution of monetary Trade Cycle theory)that the modern economic system cannot beconceived without fluctuations ascribable to mone-tary influences; and therefore any other factorswhich may be found necessary to explain theempirically observed phenomena will have to beregarded as causes additional to the monetarycau$e. In other words, any non-monetary TradeCycle theory must superimpose its system ofexplanation on that of the monetarily determinedfluctuations; it cannot start simply from the staticsystem as presented by pure equilibrium theory.

Once this is admitted, however, the questionwhether the monetary theory of the Trade Cycleis correct or not must, at any rate, be presented in adifferent form. For if the correctness of the inter-connections described by monetary theory isunquestioned, there still remains the problemwhether it is also sufficient to explain all thosephenomena which are observed empirically in thecourse of the Trade Cycle; it may perhaps needsupplementing in order to make it an instrumentsuitable to explain the working of the modern

186


economic system. It seems to me, however, thatbefore we can successfully tackle this problem weought to know exactly how much of the empiricallyobserved fluctuations is due to the monetary factor,which is actually always at work; and therefore weshall have to work out in the fullest detail thetheory of monetary fluctuations. It is hardlypermissible, methodologically speaking, to go insearch of other causes whose existence we mayconjecture, before ascertaining exactly how far,and to what extent, the monetary factors areoperative. It is our duty to work out in detail thenecessary consequences of those causes ofdisturbance which we know, and to make thistrain of thought a definite part our logical system,before attempting to incorporate any other factorswhich may come into play.

XI

The fact, simple and indisputable as it is, thatthe Elasticity' of the supply of currency media,resulting from the existing monetary organization,offers a sufficient reason for the genesis and

187


recurrence of fluctuations in the whole economyis of the utmost importance — for it implies that nomeasure which can be conceived in practice wouldbe able entirely to suppress these fluctuations.

It follows particularly from the point of view ofthe monetary theory of the Trade Cycle, that itis by no means justifiable to expect the total dis-appearance of cyclical fluctuations to accompany astable price-level — a belief which Professor Lowe*seems to regard as the necessary consequence ofthe Monetary Theory of the Trade Cycle.Professor Ropke is undoubtedly right when heemphasizes the fact that 'even if a stable pricelevel could be successfully imposed on thecapitalist economy the causes making for cyclicalfluctuations would not be removed.*f But torealize this, as the preceding argument shows,is by no means 'equivalent to a rejection of a iooper cent monetary Trade Cycle theory.'^ Onthe contrary, on this view, we must regard Pro-fessor Ropke's theory, which coincides in the

*'Uber den Einfluss monetarer Faktoren auf den Konjunkturzyklus.'op. dt.y p. 369.

f Op. cit., p. 265. X ̂ i d . , p. 278

188


more important points with our own,* as itselfconstituting such a ioo per cent monetary TradeCycle theory.

Once this is realized, we can also see hownonsensical it is to formulate the question of thecausation of cyclical fluctuations in terms of* guilt,'and to single out, e.g., the banks as those'guilty* of causing fluctuations in economicdevelopment.f Nobody has ever asked them topursue a policy other than that which, as we haveseen, gives rise to cyclical fluctuations; and it isnot within their power to do away with suchfluctuations, seeing that the latter originate notfrom their policy but from the very nature ofthe modern organization of credit. So long as wemake use of bank credit as a means of furtheringeconomic development we shall have to put upwith the resulting trade cycles. They are, in asense, the price we pay for a speed of developmentexceeding that which people would voluntarilymake possible through their savings, and which

* Cf. especially pp. 274 et seq. of the work mentioned,t As Prof. S. Budge seems inclined to do (op. cit.t p. 216). His

exposition in other respects largely coincides with ours.

189


therefore has to be extorted from them. And evenif it is a mistake — as the recurrence of criseswould demonstrate — to suppose that we can, inthis way, overcome all obstacles standing in theway of progress, it is at least conceivable that thenon-economic factors of progress, such as technicaland commercial knowledge, are thereby benefitedin a way which we should be reluctant to forgo.

If it were possible, as has been repeatedlyasserted in recent English literature,* to keep thetotal amount of bank deposits entirely stable,that would constitute the only means of gettingrid of cyclical fluctuations. This seems to uspurely Utopian. It would necessitate the completeabolition of all bank-money-—i.e. notes andcheques — and the reduction of the banks to therole of brokers, trading in savings. But even ifwe assume the fundamental possibility of thisstate of things, it remains very questionablewhether many would wish to put it into effectif they were clear about its consequences. Thestability of the economic system would be obtained

• Certain statements of Mr. R. G. Hawtrey seem to point to this,especially op. cit., p. 121.

190


at the price of curbing economic progress. Therate of interest would be constantly above the levelmaintained under the existing system (for, gener-ally speaking, even in times of depression someextension of credit takes place)*. The utilizationof new inventions and the 'realization of newcombinations' would be made more difficult,and thus there would disappear a psychologicalincentive towards progress, whose importancecannot be judged on purely economic grounds.It is no exaggeration to say that not only would itbe impossible to put such a scheme into practicein the present state of economic enlightenment ofthe public, but even its theoretical justificationwould be doubtful.

As regards the practical bearing of our analysison the Trade Cycle policy of the banks, all that canbe deduced from it is that bankers will have toweigh carefully the relative advantages and dis-advantages of granting credits on an increasing

• Cf. Professor A. C. Pigou Industrial Fluctuations, 2nd edit., p.145: 'Banks do not in bad times reduce the amount of new real capitalflowing to business men below what it would have been had therebeen no banks, but merely increase it to a smaller extent than theydo in good times.'


scale, and to take into account the demand, nowfairly widespread, for the early application of acheck to credit expansion. But the utmost that canbe achieved on these lines is only a mitigation,never the abolition, of the Trade Cycle. Apartfrom this, the only way of minimizing damageis through a far-reaching adjustment of theeconomic system to the recognized existence ofcyclical movements; and for this purpose themost important condition is an increased insightinto the nature of the Trade Cycle and a knowledgeof its actual phase at any particular moment.*

* In this connection, apart from empirical research, the greatestconsideration should be given to the plea made by O. Morgenstern(op. cit., p. 123 et seq.) for giving increased publicity to companydevelopments.

19a

UNSETTLED PROBLEMS OFTRADE CYCLE THEORY

CHAPTER V

U N S E T T L E D P R O B L E M S O FT R A D E C Y C L E T H E O R Y

So much has already been said (in Ch. in, 5 4 and6) about the most important of the outstandingproblems of monetary influences on economicphenomena, that only a brief supplement isneeded at this point. With regard to theproblems of monetary theory in the narrowersense I may restrict myself chiefly to whathas been said above, as I hope to publish theresults of a separate investigation concerning thisproblem elsewhere.* A few remarks may, how-ever, be ventured merely as a summing-up ofwhat has already been said, and in doing this weshall touch on a number of other important

* Cf. Das intertemporale Gleichgewichtssystem der Preise und dieBewegungen des Geldtuertes, Weltwirtschaftliches Archiv 1928; and moreespecially Prices and Production, London, 1931, where I haveattempted to develop some of the points touched upon in this chapter.


problems. The most significant result of ourinvestigation must be the grasp of the elementaryfact that we have no right to assume that an economicsystem with an 'elastic' currency will ever exhibitthose movements which can be immediately deducedfrom the propositions of static theory. On thecontrary, it is to be expected that movements willarise which would not be possible under theconditions usually assumed by that theory. It isparticularly important to realize that this pro-position is true whether the changes in the volumeof money also effect changes in the so-called* general value of money* or not. With the dis-appearance of the idea that money can onlyexert an active influence on economic movementwhen the value of money (as measured by onekind of price level) is changing, the theory that thegeneral value of money is the sole object ofexplanation for monetary theory must fall to theground. Its place must, henceforth, be taken byan analysis of all the effects of money on the courseof economic development. All changes in thevolume of effective monetary circulation, andonly such changes, will therefore rank for con-

196

UNSETTLED PROBLEMS OF THEORY

sideration as changes in economic data capableof originating 'monetary influences'.

The next task of monetary theory is, therefore, asystematic investigation of the effects of changesin the volume of money. In the course of thisapproach, relationships will inevitably be con-templated which do not have the permanence ofthe equilibrium relationships. All these results,however (and this must be emphasized to preventmisunderstanding) will be reached by the aid ofthe methods of static analysis, for these are theonly instruments available to economic theory.The only difference is that these methods will beapplied to an entirely new set of circumstanceswhich have never, up till now, received the atten-tion they deserve. It is vitally necessary that suchan investigation should keep clear of the notionthat the adjustment of the supply of money tochanges in 'money requirements' is an essentialcondition for the smooth working of the equilibrat-ing process of the system, as presented in equili-brium theory.* It must always start from the

• This notion rests on a confusion between the demand for moneyand the demand for cash, i.e. that portion of the total amount of money

197


assumption that the natural determining factorswill exert their full effect only when the effectivevolume of money remains unchanged, whatevermay be the actual changes in the extent of eco-nomic activity.

Precise propositions as to the effects of changesin the volume of money can be laid down onlywhen accurate information is available both asto the genesis of the change and the part of theeconomic system where it took place. For thisreason little can be said about changes resultingfrom the decumulation of hoarded treasure or thediscovery of new gold deposits. The way inwhich an individual will elect to spend moneycoming to him as a gift or as a result of other non-economic motives cannot be determined fromdeductive considerations. Similarly, little can besaid a priori about bank credits granted to theState, so long as we have no information as to howthey are to be used. The situation is different,

which at any given moment is utilized in cash, and which undergoessharp seasonal fluctuations. This phenomenon, however, is itself aconsequence of the use of bank credit. For a somewhat more detaileddiscussion of these problems, cf. Lecture IV, of Prices and Production.

198

UNSETTLED PROBLEMS OF THEORYhowever, when we are dealing with productivecredits granted by the banks to industry — whichconstitute the most frequent form of increase in thevolume of circulating media. These credits are onlygiven when and where their utilization is profitable,or at least appears to be so. Profitability is deter-mined, however, by the ratio of the interest paid onthese credits to the profits earned by their use. Solong as the amount of credit obtainable at any givenrate of interest is limited, competition will ensurethat only the most profitable employments arefinanced out of a given amount of credit. The usesto which the additional money can be put are thusdetermined by the rate of interest, and the amountwhich can be said about those uses will thereforedepend, in turn, on how much is known about theimportance and the effects of interest. Whatevermay have been written or thought on this oldproblem of theoretical economics, it is undeniablethat those particular aspects of interest theorywhich are important for our analysis have sofar received less attention and even less recogn-i t ion than is their due. It is not practicableto work out, within the limits of this essay,

199


the supplementary analysis which seems to meto be necessary in this field; but I should like, atleast, to indicate, before I conclude, some anglesof approach which appear to have been undulyneglected hitherto. Needless to say, the sectionswhich follow have even less claim than theirpredecessors to be regarded as comprehensive.

11

In the economic system of to-day, interestdoes not exist in the form in which it is presented bypure economic theory. Not only do we find,instead of one uniform rate, a great number ofdiffering rates, but, beyond this, none of thevarious rates of interest existing is entitled torank as the rate of interest described by statictheory, on which all other rates depend, differingonly to the extent to which they are affected byspecial circumstances. The process of interestfixation, which is at the basis of pure theory,never in fact follows the same course in a moderncredit economy; for in such an economy the supplyof, and the demand for, savings never directlyconfront each other.

200

UNSETTLED PROBLEMS OF THEORYAll existing theories of interest, with a few not

very successful exceptions, restrict themselvesto the explanation of that imaginary rate of in-terest which would result from such an immediateconfronting of supply and demand. The factthat the rate of interest which these theoriesexplain is one never found in practice does notmean that they are of no importance, or eventhat any explanation of the actual rates can affordto ignore them. On the contrary, an adequateexplanation of that 'natural rate' is the indis-pensable starting point for any realization of theconditions necessary to the achievement of equili-brium, and for an understanding of the effectswhich every rate of interest actually in forceexerts on the economic system. It is true thatit does not suffice to explain all empirically observedrates since it takes into consideration only oneof the factors determining those rates (thoughthat factor is of course, the one which is alwaysoperative); but any consideration of ruling interestrates which did not relate its analysis to that ofthe imaginary interest rate of static theory wouldhang entirely in the air. For the most part,

201


however, no solution has been found to the widerproblem of building up on the basis of the theoryof an equilibrium rate of interest, which can bededuced from the credit-less economy, the struc-ture of different rates which can be simultane-ously observed in a modern economy. The solu-tion of this particular problem should providea most valuable contribution to a deeper insightinto cyclical fluctuations.

But before we set out to explore on the onehand the difference between the natural rate ofinterest and the actual rate, and on the otherhand that between the various kinds of the latter,we must say something about the importance ofchanges in the equilibrium rate itself, since somevery confused ideas prevail as to the functionof the equilibrium rate of interest in a dynamiceconomy. This is not very surprising since, aswe have seen above, an insufficient appreciationof the role of interest is the cause of most misunder-standing in Trade Cycle theory. Perhaps it isnot too much to say that the importance whichan economist attaches to interest as a regulatorof economic development is the best criterion of

202

UNSETTLED PROBLEMS OF THEORYhis theoretical insight. It is therefore all the moreregrettable that recent economic literature hasbeen quite fruitless so far as the theory of interestis concerned.* This too is, perhaps, due in partto the fact that the earlier economists, to whomwe owe our present knowledge of interest theory,stopped short in their investigations and nevercame to the point of explaining the actual rates.

1 1 1

Under the rubric of pure interest theory (bywhich we understand the explanation of that rateof interest which is not modified by monetaryinfluences, although paid, of course, on capitalreckoned in money terms) we shall have to dealbriefly with the question of the effect of transitoryfluctuations in the natural rate of interest, condi-tioned by 'real' factors. This question is of great

* The best confirmation of this view is given by Mr. G. Heinze,who, in his recent study, Static or Dynamic Interest Theory (Leipzig,1928), conies to the correct conclusion that 'In spite of all the partlyjustified criticism which was levelled against the interest theory ofBohm-Bawerk, the latter still represents the most logically perfecteconomic explanation of the phenomenon of interest, and is, moreoverthe one which comes nearest to the observed facts' (p. 165).

203


importance, taking as it does a decisive place insome of the best known Trade Cycle theories of ourday. In particular Professor Cassel's view (men-tioned above, p. 80) that the real cause of cyclicalfluctuations lies in an over-estimate of the supplyof new capital, is based on the assumption that atemporary fall in the rate of interest conditionedby real causes can bring about over-investmentin the same way as a rate of interest artificiallylowered by monetary factors. This view, whichseems to be supported by a considerable body ofexperts, has to be judged quite differently accord-ing as the changes which elicit fluctuations in therate of interest originate on the demand or thesupply side. Fluctuations caused by changes onthe demand side, which Professor Cassel uses asan explanation in his trade cycle theory, certainlycannot be regarded as an adequate explanationof the cycle; for, as Professor Amonn has alreadypointed out,* this is no reason why enterpreneurs

* CasseVs System der Theorettschen Nationalokonomie (Archivfur Sozialwissenschaften und Sozialpolitik, vol. 51, Tubingen1924, p. 348 et seq.) In order to remain within the scope of our workwe have to forgo the very alluring task of criticizing Professor Cassel'sargument. Such a criticism would also have to deal with the very

204


should (assuming an unchanged rate of interest)expect to obtain more credits in the future thanthey can now. However there can be no doubtthat violent fluctuations in savings and theconsequent temporary changes in the equilibriumrate of interest act similarly to an artificial loweringof the money rate of interest in causing an exten-sion of capital investments which cannot bemaintained later owing to the diminished supplyof savings.* In this case, therefore, it is permissibleto speak of non-monetary cyclical fluctuations.This differs, however, from the conception ofcyclical fluctuations employed hitherto, in thatthe passage from boom into depression is not anecessary consequence of the boom itself, but isconditioned by 'external circumstances'. A down-ward turn of this sort can occur just as well in a

ingenious theoretical interpretation of this argument by Dr. G. Halm(Das Zinsproblem am Geld- und Kapitalmarkt, Jakrbucher fttrNationalokonomie und Statistik, 3rd series, vol. 70, Jena, 1926, p.16 et seq.) Here we may only point out that in this study Halm isdriven to make use of the old hypothesis that savings accumulatefor a time and are then suddenly utilized 'at the moment when thereal boom begins' (p. 21).

*Cf. Dr. A. Lampe, Zur Theorie des Sparprozesses und derKrediUchfipfung (Jena, 1926, p. 67 seq.)

205


hitherto stationary economy or during a depressionas at the end of a boom; and it should thereforebe regarded less as an example of a cyclicalmovement than as a particularly complicatedcase of the direct process of adjustment to changesin data. In any case, for reasons given above,such an explanation, as compared with an endo-genous theory, would only come into play whenthe latter had proved insufficient to explain a givenconcrete phenomenon.

But there can be no doubt that such fluctuationsin the natural rate, conditioned by changes inthe rate of saving activity, present some very im-portant problems in interest theory, the solutionof which would be an important aid in estimatingthe effect of fluctuations conditioned by monetarychanges. We have entirely disregarded the cir-cumstances determining the supply of savingsand the fluctuations in this supply; and theexamination of these is a promising field for futureresearch. It might even be possible to showthat fluctuations in saving activity are a neces-sary concomitant of economic progress, and thusto give a firm basis to the theories which we

206


have mentioned. This is, perhaps, not very

probable.In direct relation to the above problem stands

the question of the effects of alterations in therate of interest on the price system as a whole.An examination of this subject should throw lighton the point of view, emphasized by ProfessorFetter,*5 that the height of interest rate, at anygiven moment expresses itself in the wholestructure of price relationships, while everychange in that rate must part passu bring aboutchanges in the relation between particular pricesand thus in the quantitative relationships of thewhole economy.

But here we must content ourselves withdrawing attention to the problems arising outof the changes in the natural rate of interest,without contributing further to their solution.We shall only venture a further remark on a ques-tion concerning not the consequences but thecauses of these changes, since this is importantin what follows. This is the question whether

* Op. cit., especially, p. 78. Cf. also, Lecture III, of Prices andProduction.

207


the rate of interest at any moment depends onthe total amount of capital existing at that momentor only on the amount of free capital available fornew investment.* We mention this here only inorder to emphasize the untenability of the wide-spread view that the determining factor, on thesupply side, is the whole existing stock of capital.If that were so it would hardly be possible toexplain any large fluctuations of the rate ofinterest, since the relative changes which theexisting capital stock undergoes within briefperiods and under normal circumstances is in-significant. A thorough investigation of the inter-connections in this field must show that the actualrate of interest depends (apart from the demandfor loan capital) only on the supply of newlyproduced or reproduced capital. The existingstock of fixed capital affects only the demandside, by determining the yields to be expectedfrom new investments. This explains how, in acountry which is well equipped with fixed capital,

•'Capital disposable for investment' was the phrase usuallyemployed by the classical writers to distinguish this free capitalfrom the stock of real capital. Cf., e.g., J. S. Mill, Essays on someunsettled questions of Political Economy, London, 1844, pp. 113 ff.

2O8

UNSETTLED PROBLEMS OF THEORYthe rate of interest can temporarily rise higherthan that obtainable in a country which is poorlyequipped, provided that there is relatively morefree capital available for new investment in thelatter than in the former. This fact has somesignificance in connection with the phenomenonof enforced saving with which we shall deal later.

i v

As regards the relationship of the naturalor equilibrium rate of interest to the actual rate,it should be noted, in the first place, that eventhe existence of this distinction is questioned.The objections, however, mainly arise from amisunderstanding which occurred because K.Wicksell, who originated the distinction, made usein his later works of the term 'real rate' (whichto my mind is less suitable than 'natural rate') andthis expression became more widespread thanthat which we have used.* The expression 'realrate of interest' is also unsuitable, since it coincides

• Cf. the works of Professor Ropke and Dr. Burchardt, mentionedabove; also £. Egner, Zur Lehre vom Zwangsparen, 1928, p. 537.Occasionally {Geldzins und Guterpreise, p. i n ) Wicksell also uses theexpression 'normal rate*.

o 209

MONETARY THEORY AND THE TRADE CYCLEwith Professor Fisher's 'real interest',* which, asis well known, denotes the actual rate plus therate of appreciation or minus the rate of depre-ciation of money, and is thus in accordance withcommon usage, which employs the term 'realwages' or 'real income' in the same sense. Un-fortunately Wicksell's change in terminology isalso linked up with a certain ambiguity inhis definition of the 'natural rate'. Havingcorrectly defined it once as 'that rate at which thedemand for loan capital just equals the supplyof savings'f he redefines it, on another occasion,as that rate which would rule 'if there were nomoney transactions and real capital were lent innatura\% If this last definition were correct, Dr.G. Halm§ would be right in raising, against theconception of a 'natural rate', the objection thata uniform rate of interest could develop only ina money economy, so that the whole analysis is

• Cf. especially Appreciation and Interest (Publications of theAmerican Economic Association, 3rd series, vol. xi, No. 4, NewYork, 1896).

f Vorlesungen, vol. ii, p. 220.X Geldzins und Gtiterpreise, p. 93.§ Op. cit., p. 7, footnote.

210


irrelevant. If Dr. Halm, instead of clinging tothis unfortunate formula, had based his reasoningon the correct definition which is also to be foundin Wicksell, he would have reached the same con-clusion as Professor Adolf Weber—the distinguishedhead of the school of which he is a member; thatis, that the natural rate is a conception 'which isevolved automatically from any clear study ofeconomic interconnections'.* In accordance withthis view, Wicksell's conception must be creditedwith fundamental significance in the study ofmonetary influences on the economic system;especially if one realizes the practical importanceof a money rate of interest depressed below thenatural rate by a constantly increasing volume ofcirculating media. Unfortunately, although Wick-sell's solution cannot be regarded as adequate atall points, the attention which it has receivedsince he propounded it has borne no relation toits importance. Apart from the works of ProfessorMises, mentioned above, the theory has made noprogress at all, although many questions concern-

* Depositenbanken und Spekulationsbanken, 3rd edit., 1922, p. 171.

211


ing it still await solution.* This may be due to thefact (on which we have touched already) that theproblem had become entangled with that offluctuations in the general price level. We havealready stated our views on this point, (p. 196)and indicated what is necessary for the furtherdevelopment of the theory. Here, we shall tryto restate the problem in its correct form, freedfrom any reference to movements in the pricelevel.

Every given structure of production, i.e.every given allocation of goods as betweendifferent branches and stages of production,requires a certain definite relationship betweenthe prices of the finished products and those of themeans of production. In a state of equilibrium,

• Another attempt to develop Wicksell's theory — of which I havelearnt only since the above was written — was made, at roughly thetime when Mises' work was published, by Prof. M. Fanno of Paduain a work entitled Le banche e il mercato monetario, Rome 1912. Anabridged restatement of Prof. Fanno's theory will shortly appear, inGerman, in a volume of essays on monetary theory by a number ofDutch, Italian and Swedish authors, edited by the author of thepresent essay.

212

UNSETTLED PROBLEMS OF THEORYthe difference necessarily existing between thesetwo sets of prices must correspond to the rate ofinterest, and at this rate, just as much must besaved from current consumption and made avail-able for investment as is necessary for the main-tenance of that structure of production. Thelatter condition necessarily follows from the ful-filment of the former, since the prices paid for themeans of production, plus interest, can onlycorrespond to the prices of the resulting productswhen, at the given prices and rate of interest,the supply of producers' goods is exactly adequateto maintain production on the existing scale.The price margins between means of productionand products, therefore, can only remain constantand in correspondence with the rate of interestso long as the proportion of current income, whichat the given rate of interest is not consumed butreinvested in production, remains exactly equalto the necessary capital required to carry onproduction. Every change in this proportionmust begin by impairing the correspondence ofprice margins and the interest rate; for it influencesboth in opposite directions, and so leads to further

213


shifts in the whole structure of production,representing an adjustment to altered price-relationships. These resulting changes in thestructure of production will not always be thesame; they will vary according to whether thechange in the proportions of the social incomegoing respectively to consumption goods andinvestment goods corresponds to real changes inthe decisions of individuals as to spending andsaving, or whether it was brought about artificially,without any corresponding changes in individualsaving activity.

Apart from individual saving activity (whichincludes, of course the savings of corporations,of the State and of other bodies entitled to raisecompulsory contributions) the proportions be-tween consumption and capital creation can onlychange as a result of alterations in the effectivequantity of money.* When changes in the divisionof the total social dividend, in favour of capitalcreation, result from changes in the saving activityof individuals, they are self-perpetuating. This

• Very instructive investigations of the problems considered herewere carried out by A. Lampe: Zur Theorie des Sparprozesses undder Kreditschdpfung, Jena, 1926.

214

UNSETTLED PROBLEMS OF THEORYis not true of such variations between 'consumptionand accumulation* (if we may use for once theterminology of Marxian literature) as are due toadditional credits granted to the entrepreneurs;these can be assumed to persist only so long asthe proportion is kept artificially high by a pro-gressively increasing rate of credit creation. Suchan injection of money into circulation acts onlytemporarily — until the additional money becomesincome. At that moment, the proportion of capitalcreation must relapse to the level of voluntarysaving activity, unless new credits are grantedbearing the same relation to the new total ofmoney incomes as the first injection bore to theformer total.*

* The argument presented in the text (and put in this form forbrevity's sake) is imperfect in two respects. Firstly, the flow ofvoluntary saving can itself vary as a result of a single change in theproportion of capital formation. This factor, however, is unlikely tobecome important enough for its omission to affect the expositiongiven in the text. Secondly, the way in which the additional money,which was given in the first instance to entrepreneurs and used bythem to lengthen the period of production, will always swell incomesin the long run, needs further elaboration. As a general proposition,however, it is obvious that whoever uses the additional credits tomake additional investment goods can do so only by employingadditional factors of production; and therefore, since there is in ourcase no compensating decrease in the demand for factors elsewhere,the total incomes of the factors must increase.

215


It is clear that such a process of progressiveincrease in the supply of money cannot be main-tained under our existing credit system, especiallysince, as it proceeds, more extensive use will bemade of cash. On the other hand, a mere cessationof further increases — not, therefore, a reversalof credit policy, towards deflation —is sufficientto bring back the proportion of total incomeavailable for capital formation to the extent ofvoluntary savings.

The differences in the effects of these two kindsof variation between consumption and capitalformation manifest themselves first of all on theprice system, and thus on the natural or equilibriumrate of interest. The first effect of a diminutionof the rate on loans arising from increased savingactivity — so long as the structure of productionremains unchanged — is to bring that rate belowthe margin between the prices of means of pro-duction and of products. The increased savingactivity, however, must soon cause on the onehand a falling off in the demand for consumptiongoods, and hence a tendency for their prices tofall (a tendency which may merely find expression

216

UNSETTLED PROBLEMS OF THEORYin decreasing sales at existing prices) and,on the other hand, an increase in the demand forinvestment goods and thus a rise in their prices.The extension of production will have a furtherdepressing effect on the prices of consumptiongoods, as the new products come on the market,until, finally, the difference between the respectiveprices has shrunk to a magnitude correspondingto the new, lower, interest rate. If, however, thefall in the rate of interest is due to an increase inthe circulating media, it can never lead to a corre-sponding diminution in the price margin, or toa readjustment of the two sets of prices to thelevel of an equilibrium rate of interest whichwill endure. In this case, moreover, the increaseddemand for investment goods will bring about anet increase in the demand for consumptiongoods; and therefore the price margin cannot benarrowed more than is permitted by the time-lagin the rise of consumption goods prices — a lagexisting only as long as the process of inflationcontinues. As soon as the cessation of creditinflation puts a stop to the rise in the prices ofinvestment goods, the difference between these

217


and the prices of consumption goods will increaseagain, not only to its previous level but beyond,since, in the course of inflation, the structure ofproduction has been so shifted that in comparisonwith the division of the social income betweenexpenditure and saving the supply of consumptiongoods will be relatively less, and that of productiongoods relatively greater, than before the inflationbegan.*

VI

There have recently been increasingly fre-quent objections to this account of the effects ofan increased volume of currency, and the artificiallowering of interest rates conditioned by it, on thegrounds that it disregards certain supposedlybeneficial effects which are closely connected withthis phenomenon. What the objectors have inmind is the phenomenon of so-called 'forcedsaving' which has received great attention inrecent literature.f This phenomenon, we are

• Cf. my article on The *Paradox' of Saving, 'Economical No. 32,p. 160.

f Besides Leon Walras - the originator of this theory (Cf. hisEtudes d'economic politique appliquee 1898, pp. 348-356),

218


to understand, consists in an increase in capitalcreation at the cost of consumption, through thegranting of additional credit, without voluntaryaction on the part of the individuals who forgoconsumption, and without their deriving anyimmediate benefit. According to the usual pre-sentation of the theory of forced saving, thisoccurs through a fall in the general value of money ,which diminishes the consumers' purchasingpower; the volume of goods thus freed can be usedby the producers who obtained additional credits.We must, however, raise the same objection tathis theory which we raised against the usualaccount of the effects of an artificial lowering ofthe money rate of interest, i.e. that, in principle,,forced saving takes place whenever the volumeof money is increased, and does not need to

Wicksell and the well-known works of Professors Misesand Schumpeter, one must mention the recent works of Pro-fessor Ropke, Dr. Egner, and Dr. Neisser; and in Anglo-Saxon litera-ture, Mr. D. H. Robertson's Banking Policy and the Price Level(London, 1926). As I have pointed out, however, in Lecture I of Pricesand Production and - at somewhat greater length - in A Note on theDevelopment of the Theory of Forced Saving, in the Quarterly Journalof Economics,.November, 1932, the concept of'forced saving' was alreadyknown to J. Bentham, H. Thornton, T. R. Malthus and a number ofother writers in the early 19th century, down to J. S. Mill.

219


manifest itself in changes in the value ofmoney.*

The 'depreciation' of money in the hands ofthe consumer can be, and frequently will be, onlyrelative, in the sense that those diminutions inprice which would otherwise have occurred areprevented from occurring. Even this causes apart of the social dividend to be distributed toindividuals who have not acquired legitimateclaim to it through previous services, nor takenthem over from others legitimately entitled tothem. It is thus taken away from this part of thecommunity against its will. After what has beensaid above, this process needs no further illumin-ation.

Nor do we need to adduce further proof that everygrant of additional credit induces 'forced saving*— even if we have avoided using this ratherunfortunate expression in the course of our argu-ment. There is only one further point — the effectof this artificially induced capital accumulation —on which a few remarks should be added. It has

• Cf. D. H. Robertson, Money, 2nd edit., p. 99; and A. C. Pigou,Industrial Fluctuations, 2nd edit., 1929, pp. 251-257.

220

UNSETTLED PROBLEMS OF THEORYoften been argued that the forced saving arisingfrom an artificially lowered interest rate wouldimprove the capital supply of the economy to suchan extent that the natural rate of interest wouldhave to fall finally to the level of the money rateof interest, and thus a new state of equilibriumwould be created —that is, the crisis could beavoided altogether. This view is closely connectedwith the thesis, which we have already rejected,that the level of the natural rate of interest dependsdirectly upon the whole existing stock of realcapital. Forced saving increases only the existingstock of real capital goods, but not necessarily thecurrent supply of free capital disposable for in-vestment — that portion of total income which isnot consumed but used as a provision for theupkeep and depreciation of fixed plant. But anyaddition to the supply of free capital availablefor new investment or reinvestment must comefrom those of the investments induced by forcedsavings which already yield a return; a returnlarge enough to leave over, after providing forsupplementary costs connected with the newmeans of production, a surplus for depreciation

221

MONETARY THEORY AND THE TRADE CYCLEand for interest payments on the capital. If thecapital supply from this source is to lower thenatural rate of interest, it must not, of course,be offset by a diminution elsewhere — resultingfrom the decline of other undertakings confrontedwith the reinforced competition of those newlysupplied with capital.

The assumption that an artificial increase offixed capital (i.e. one caused by additional credits)tends to diminish the natural rate of interest inthe same way as one effected through voluntarysavings activity presupposes, therefore, that thenew capital must be incorporated into the eco-nomic system in such a way that the prices of theproducts imputed to it shall cover interest anddepreciation. Now a given stock of capital goodsis not a factor which will maintain and renewitself automatically, irrespective of whether it isin accordance with the current supply of savings ornot. The fact that investments have been under-taken which cannot be 'undone* offers no guaranteewhatever that this is the case. Whether capital canbe created beyond the limits set by voluntary savingdepends — and this is just as true for its renewal

222

UNSETTLED PROBLEMS OF THEORYas for the creation of new plant — on whether theprocess of credit creation continues in a steadilyincreasing ratio. If the new processes of produc-tion are to be completed, and if those already inexistence are to continue in employment, it isessential that additional credits should be con-tinually injected at a rate which increases fast enoughto keep ahead, by a constant proportion, of the ex-panding purchasing power of the consumer. If anew process of roundabout production can be com-pleted while these conditions still hold good, itcan contribute temporarily to a lowering of thenatural rate of interest; but this provides no finalsolution of the difficulty.

For, eventually, a moment must inevitablyarrive when the banks are unable any longer tokeep up the rate of inflation required, and atthat moment there must always be some processesof production, newly undertaken and not yetcompleted,* which were only ventured because

* The existence of new long processes, which have not yet beencompleted, is not a necessary condition in order that the relativeincrease in the demand for consumers' goods may lead to the abandon-ment of such processes and, therefore, to the destruction of parirof thecapital employed there; but it is the case which will always be given in

223


the rate of interest was kept artificially low. Itdoes not follow, of course, that these processes inparticular will be left unfinished because of thesubsequent rise in that rate; on the other hand,their existence does cause the rate of interest tobe higher than it would be in their absence, whencapital would be required only by processes madepossible by voluntary saving without any competingdemand arising from processes which were onlyenabled to start by 'forced saving.' The capitalinvested in new and not yet completed processesof production will thus merely intensify the demandfor further supplies by calling for the capitalnecessary to complete them — an effect whichwill be the more pronounced the greater the

practice and where this effect is most easily seen. In this connectionit should, however, be noted that the introduction of a longer round-about process of production will, in almost all cases, affect not onlya single enterprise but a series of enterprises representing successivestages of production. Even a completed plant may, in this sense,represent part of an incomplete process—if the capital is lackingwhich would have to be invested in the machines or other capitalgoods to be produced by this plant. Plant equipped to satisfy ademand for machinery which cannot be permanently maintained is, inthis sense, part of a roundabout process which cannot be continued.For a fuller description, see Prices and Production, Lecture III and,especially, my article, Kapitalaufzehrung in the WeltwirtschaftlichesArchiv, July 1932.

224


ratio of capital invested to capital still required.It may therefore quite easily come about that, inorder to complete these newly initiated processes,capital may be diverted from the maintenanceof complete and old-established undertakings,so that new plant is put into operation and oldplant closed down, although the latter would havebeen kept up, and the former never put in hand,if it had been a question of building up the wholecapital equipment of the economy from the start.This does not merely mean that the total returncomes to less than it otherwise would; it alsomeans, primarily, that production is forced intochannels to which it will only keep for as long asthe new and spuriously produced stock of fixedcapital can remain in use. The value of capitalinvested in processes which can be continued,and, still more, that in processes where continuanceis impracticable, will shrink rapidly in value —this shrinkage being accompanied by the phe-nomena of a crisis. Thus on purely technicalgrounds it will become uneconomic to maintainthem. It should be particularly remarked that,from the point of view of the fate of individual

p 225


enterprises, capital invested in fixed plant, butraised by borrowing, is of precisely the sameimportance as working capital, i.e. the loss ofvalue does not merely necessitate writing down,it generally makes it impossible to carry on at all.

The cause of this development is, evidently,that an unwarranted accumulation of capital hasbeen taking place; though people may regard it(under the alluring name of * forced saving') as athoroughly desirable phenomenon. After whathas been said above it is probably more proper toregard forced saving as the cause of economic crisesthan to expect it to restore a balanced structure ofproduction.

V I I

There remains one problem of interest theory,in the wider sense of the word, which we needto examine more closely than we have yet done— in order to exhibit a problem of first-class im-portance to the progress of Trade Cycle theory.This is the problem of the varying height and inde-pendent movements of rates of interest rulingat the same place and at the same time. We are

226


not thinking, of course, of differences conditionedeither by the unequal standing of borrowers orby the fact that under the name of interest paymentsare also made for the services or costs connectedwith the granting of credit. We are interestedonly in the problem of variations arising withinthe pure or net rate of interest, as they can beobserved between credits of varying duration —the problem usually known in economic literatureas the problem of interest rates, in the money, andin the capital (investment) market, respectively.

In this respect we may repeat what we havealready said at the beginning of this chapter —that the theoretical investigations of interest havebeen broken off at far too early a stage to affordmuch understanding of the rates actually rulingat any given moment. It is very remarkable thatnone of the great theorists to whom we owe ourinsight into the fundamental factors determiningthe equilibrium rate of interest made the slightestattempt to explain these differences betweeninterest rates. Systematic investigation of thisproblem came much later and then character-istically the investigation related chiefly to the

PI 227


question of the 'external order of the capital ormoney market'; and it is only recently that Dr.G. Halm* has treated the simultaneous existence ofvarying interest rates 'as a problem of interesttheory'. Although Dr. Halm deserves full creditfor the undeniable service he has rendered inputting the problem in the proper form fordiscussion, his attempt at solution can hardly beregarded as fully successful. Thus we still stand

* 'Das Zinsproblem am Geld und Kapitalmarkt' Jahrbucher fiirNationalokonomie und Statistik, 3rd series, No. 125, vol. 70, Jena1926. Of the comprehensive bibliography given by Halm, the follow-ing, together with some more recent additions, are worth mention:A. Spiethoff: (1) Die dussere Ordnung des Kapital und Geldmarktes;(2) Das Verhdltnis von Kapital, Geld und Gutertoelt; (3) Der Kapital-mangel in seinem Verhdltnis zur Giiterwelt; all in Schmloler's Jahr-buch, vol. 33 (Miinchen, 1909). Also, by the same author: DerBegriff des Kapital und Geldmarktes, Schmoller's Jahrbuch, vol. 44,1920. H. von Beckerath Kapital und Geldmarkt (Jena 1916).

Professor Schumpeter's Dr. Neisser's and Professor Fetter's worksalready mentioned; A. Hahn: Zur Theorie des Geldmarktes (Archivfiir Sozialwissenschaften und Sozialpolitik, vol. 51, Tubingen, 1924).Karin Kock: A Study of Interest Rates, Stockholm EconomicStudies, No. 1, London, 1929.

W. W. Riefler, Money Rates and Markets in the United States(New York and London, 1930).

The problems arising out of empirical research are well summar-ized by O. Dormer and A. Hanau, 'Untersunchung zur Frage derMarktzusammenhange' (Vierteljahrshefte zur Konjunkturforschung,3rd year, No. 3 A, Institut fur Konjunkturforschung, Berlin 1928), aninvestigation which is a model of its kind.

228


at the beginning of a crucially important develop-ment of a special theory of money rates ofinterest.

The clearing up of these interconnections isof primary importance to Trade Cycle theory,since the discrepancies between the expectedyields of existing means of production and theactual yield obtainable from the available liquidcapital must necessarily arise in the course of thecycle. Given a sufficient insight into the influencesdetermining the yields of both types of investment,the simultaneous changes in the height of bothkinds of interest rates should afford extraordinarilyvaluable material for the diagnosis of any actualsituation, and thus the growth of this part ofinterest theory would provide an important basisfor the development of empirical research andforecasting. A particularly promising approachmight consist in an examination of the questionfrom the point of view of an equalization of thetime-differences between the rates of interestwhich would prevail if the whole supply of capital,at any time, had to be invested for a longer period.Such an equalization would be brought about by

229


a kind of arbitrage for which, naturally, onlymoney lent at call or at short notice could beconsidered.

In this field, too, the extension of equilibriumanalysis to successively occurring phenomena,which I have attempted in another work,* mayprove fruitful. At any rate, an explanation of thisarbitrage could also explain why the rates on short-term credits can be temporarily lower, or on theother hand higher, than the long-term rate, sinceboth borrower and lender would find such anarbitrage to their advantage. This view cannotbe refuted by the objection that the rates on short-term credits not only change earlier but alsochange to a greater degree than those in thecapital market; for it may be economically entirelyjustifiable to pay higher rates or obtain lowerones, for a short term, than one expects for a longterm, since the expectation of getting better termsat a later period, under more favourable conditions,may compensate for the relative disadvantagessuffered in the short run.

• Cf. 'Das intertemporale Gleichgewichtssystem der Preise und dieBewegungen des Geldwertes'- already quoted.

230


V I I I

Finally, we should like to point out quitebriefly certain tasks in the field of statisticalresearch which according to our theoretical analysisseem likely to be particularly fruitful. In connectionwith the last question dealt with, we should drawattention to the statistics of the money market,which are still, unfortunately, in a very elementarystage, partly for technical reasons but mainlybecause of difficulties of interpretation. Theselatter arise largely from the fact that the statisticaldetermination of the absolute height of theinterest rate, or even of its movements, disclosesalmost nothing as to its bearing on the economicsystem.# The same rate of interest which at onemoment may be too low in relation to the wholeeconomic situation may be too high at the next,or vice versa. Misunderstandings on this pointmay be responsible for certain erroneous views,

* The statistical determination of nominal changes in the interestrate is also rendered very difficult by the fact that changes can takeplace in the form of changes in stipulations as to the quality of thebills discounted at a given rate, and so on. The same rate may bemerely applied to a better class of borrowers, or the same borrowersmay be required to pay a higher rate.

231


concerning the alleged insignificance of the heightof interest rates, which are often held by statis-tical economists. The innumerable attempts tominimize the significance of interest rates bymeans of statistical investigations, which aboundin the United States* (where they do not evenshrink from such absurdities as an attempt tofind an explanation (!) of interest by way of statis-tical investigation) would be impossible but forthe complete confusion persisting as to the limitsof statistical research. Here again we have torepeat what was asserted at the beginning of thisbook: statistics can never prove or disprove a theo-retical explanation, they can only present problemsor offer fields for theoretical research.

For precisely this reason —viz. that the abso-lute height of interest rates tells us nothing oftheir significance — an examination of the extentand regularity of shifts between various interestrates offers a promising field for statistical tech-nique. An interesting first attempt in this direction

• Cf. e.g. Snyder, 'The Influence of the Interest Rate on theBusiness Cycle,' American Economic Review, vol. xv, December1925; reprinted in Business Cycles and Business Measurements (NewYork, 1927).

232

UNSETTLED PROBLEMS OF THEORYis the famous * Three Market Barometer' of theHarvard Economic Service, which uses the trendof the long-term interest rate as a base-line inplotting the curve of the money market rates.Such an empirical consideration of the differencesbetween interest rates does not, of course, exhaustthe lines of approach which a complete theoreticalexplanation of these rates might indicate assuitable for empirical research. The fact thattheoretical research itself can be stimulated andawakened to new problems to an importantdegree from the application of our sketchy know-ledge to statistical investigations is amply demon-strated by the investigations of Donner and Hanau,which we have already mentioned.

It is in the statistics of private banking, however,that the heaviest task presents itself. In Europewe are still worse supplied with these than withthose of the money market proper. In the UnitedStates, on the other hand, some pioneer workhas been done in this field,* since the ample

• Cf. first of all A. A. Young, 'An Analysis of Bank Statistics for theUnited States' (reprinted from The Review of Economic Statistics,October 1924, January and April 1925, and July 1927) (Cambridge1928); H. Working/Prices and the Quantity of the Circulating Medium,

233


statistical material available there provided isitself a sufficient incentive for such investigations.In Europe the lack of any kind of material makeseven a first step in this direction impossible.

In many respects the most remarkable of theseenquiries are those of Mr. Holbrook Working.Using the data concerning the state of deposits inthe 'National Banks', which are available for manyyears past and at intervals of only a few months,he succeeds in establishing a far-reaching parallel-ism between the movements of deposits and thefluctuations of the wholesale price-level. Likemost theoretical investigations in the same field,however, his results are distorted by the super-ficial assumption that monetary influences can onlymanifest themselves in movements of the price

1890-1921' {The Quarterly Journal of Economics, vol. xxxvii,Cambridge, 1923) and 'Bank Deposits as a Forecaster of the GeneralWholesale Price Level' {The Review of Economic Statistics, Cam-bridge, Mass., 1926) of the same author, C. Snyder, 'DepositsActivity as a Measure of Business Activity' {Review of EconomicStatistics, Cambridge, Mass., 1924, reprinted in Business Cyclesand Business Measurements. New York, 1927), W. M. Persons,'Cyclical Fluctuations of the Ratio of Bank Loans to Deposits'{Review of Economic Statistics, Cambridge, Mass., 1924). L. A.Hahn, 'Zur Frage des volkswirtschaftlichen Erkenntnisinhaltes derBankbilanzziffern' {Vierteljahrshefte zur Konjunkturforschung, 1,Jahrgang, Erganzungsheft 4, Berlin, 1927).

234

UNSETTLED PROBLEMS OF THEORYlevel, while those changes in the volume of bankcredits which are just sufficient to prevent changesin the price level are supposed, on this assumption,to exercise no active influence on the TradeCycle. It should be mentioned —as havingparticular bearing upon the views developed inthis essay — that, according to Mr. Working'scalculations, before the War a yearly increase indeposits of more than 5 per cent would have beennecessary in order to keep the price-level steady;that is to say, additional credits would have hadto be created to an extent which must have causedconsiderable changes in the structure of pro-duction.

If the results of our theoretical analysis were tobe subjected to statistical investigation, it is notthe connection between changes in the volume ofbank credit and movements in the price levelwhich would have to be explored. Investigationwould have to start on the one hand from alter-ations in the rate of increase and decrease in thevolume and turnover of bank deposits and, on theother, from the extent of production in thoseindustries which as a rule expand excessively as

235


a result of credit injection.*8 Every increase in thecirculating media brings about the same effect,so long as each stands in the same proportion tothe existing volume; and only an increase in thisproportion makes possible a further increase ininvestment activity. On the other hand, everydiminution of the rate of increase in itself causessome portion of existing investment, made possiblethrough credit creation, to become unprofitable.It follows that a curve exhibiting the monetaryinfluences on the course of the cycle ought toshow, not the movements in the total volume ofcirculating media, but the alteration in the rate ofchange of this volume.*)* Every up-turn of thiscurve would show that an artificial lowering ofthe money rate of interest or, if the curve wasalready rising, a further lowering of the moneyrates, was making possible additional investments

• Cf. the instructive graphs given by Harold L. Reed in his FederalReserve Policy 1921-1930 (New York and London 1930), pp. 181et seq.

t Mathematically speaking, the question is one of the graphicalpresentation, in place of the curve showing the original movementsat any moment, of the first differential of this function. On thesubject of this method, which has been frequently used of late, cf.I. Fisher, The Business Cycle Largely a 'Dance of the Dollar* (QuarterlyPublication of the American Statistical Association. December, 1923).

236

UNSETTLED PROBLEMS OF THEORYfor which voluntary savings would not suffice;and every down-turn would show that currentcredit-creation was no longer sufficient to ensurethe continuance of all the enterprises which itoriginally called into existence. It would be ofgreat interest to correlate this presentation of theinfluence causing an excessive production of capitalgoods with actual changes in the production ofthese goods, on the basis of available data.

The possible contributions of banking statisticsto Trade Cycle research are by no means exhaustedby the chance they offer of observing the imme-diate connection between the granting of credit andthe movements of production, though these maysome day constitute the most important basis ofbusiness forecasting. No less important would bean investigation into the volume, at any givenmoment, of those factors which determine creditexpansion, under the other headings of bankbalance sheets, and, in particular, an examinationof the relation between the total amount of earningassets and the current accounts, the relationbetween these and the cash-circulation, and so on.Such an investigation, if it were not merely to

237


exhibit their movements in time but also toanalyse the deeper connections between them, andmost especially if it were to clear up the relation-ship between interest rates, profits, and theliquidity of the banks, would further our insightinto the factors determining credit expansion aswell as our knowledge of their limits, and thusmake it possible to forecast movements in thefactors determining the total development of theeconomic situation.

It is very unfortunate that such inquiries,especially on the continent of Europe, are almostimpossible owing to the lack of necessary datain the form of returns showing the state of thebanks, and published at short intervals; in so faras they are possible at all it is only in a few coun-tries and for a very short period. As soon as it isrealized that, owing to the existence of banks, theequilibrating forces of the economic systemcannot bring about that automatic adjustment ofall its branches to the actual situation, which isdescribed by classical economic theory, it isjustifiable even from a liberal point of view thatthe banks should be subjected to degrees of pub-

238

UNSETTLED PROBLEMS OF THEORYlicity as to the state of their business which is notnecessary in the case of other undertakings; andthis would by no means imply a violation of theprinciple of business secrecy, since it would bequite sufficient for this purpose if the authoritieswere to adopt the United States' plan of publishingsummary returns for all banks at frequent intervals.Our reflections thus yield the conclusion that analleviation of cyclical fluctuations should beexpected pre-eminently from a greater publicityamong business enterprises, and particularlyamong the banks. The example of the UnitedStates, which is far ahead in this respect in all thebranches of its economic system, will not onlysilence in time the objections raised against suchpublicity, but sooner or later will force us tofollow in their path.

239

INDEX OF AUTHORS CITED

INDEX OF AUTHORS CITEDAFTALION, A., 62, 63Altschul, E., 31Amonn, A., 204Angell, J. W., n o

BECKERATH, H. U., 228Behrens, W. G., 118Bergmann, E. v.t 131Bdhm-Bawerk, E. v.t 126, 131,

203Bortkiewiez, L. v.t 118Bouniatian, M., 143, 156, 166Bradford, F. A., 172Budge, S., 78, i n , 181, 189Bullock, C. J., 38

FANNO, M., 212Ferrara, F., 153Fetter, F. A., 180, 207, 228Fisher, I., 106, 153, 210, 236Foster, W. T., 97

GREGORY, T. E., 153

Giffen, R., no

HABERLER, G., 153

Hahn, L. A., 111, 144, 153, 164,228, 234

Halm, G., 205, 210, 211, 228Hanau, A., 228, 233Hansen, A. H., 53

Burchardt, F., 105, 121, 122, Hardy, C. O., 66126, 141, 209 Hawtrey, R. G., 141, 153, 168,

181, 190CANNAN, E., 153 Heinze, G., 203Carell, E., 33, 94Carver, T. N., 63 J6HR, A., 155, 164Cassel, G., 41, 58, 81, 97, 106, Juglar, C , 148

133, i34. 204Crick, W. F., 153, 161, 171

DAVENPORT, H. J., 153Davidson, D., 115Dernburg, B., 164, 165,167Dormer, O., 228, 233Dunbar, C. F., 153

EGNER, E., 209, 219Eucken, W., 78, 114

KEYNES, J. M., 80Knapp, G. F., 117Kock, K., 228Koopmans, J. G., 118

LAMPE, A., 205, 214Lawrence, J. S., 172Leaf, W., iS3, 164Lederer, E., 58, 97, 129Lescure, J., 58, 97, 129

243

INDEX

L6we, A., 27, 29, 33, 42, S3, 93,94, 105, 121, 122, 127, 141,144, 184, 188

MACLEOD, H. D., no , 153Marshall, A., no , 139Menger, C , 117Miksch, L., 42Mill, James, 42Mill, J. S., 208, 219Miller, A. C , 22Mises, L., 47, 48, no , m , 116,

117, 118, 119, 124, 128, 133,144, 145, 150, 153, x8x, 211,219

Mitchell, W. C , 53, 58, 63, 97,129

Morgenstern, O., 36, 83, 192

NEISSER, H., 145, 153, 163, 164,176, 219, 228

Nicholson, J. S., n o

PENNINGTON, J., 153Persons, W. M., 38, 53, 234Phillips, C. O., 153, 154, 161, 172Pigou, A. C , 31, 83, 191, 220

REED, H. L., 236Reisch, R., 153, 155, 164, 165Ricardo, D., n oRiefler, W. W., 228

Robertson, D. H., 62, 63, 64115, 219, 220

Rodkey, R. G., 153Rdpke, W., 46, i n , 181, 188,

209, 219

SAY, J. B., 42, 101Schlesinger, K., 139, 164Schumpeter, J., 57, 60, 97, 163,

168, 183, 219, 228.Sidgwick, H., n oSnyder, C, 232, 234Spiethoff, A., 41, 79, 80, 89, 90,

105, 133, 134, 228Strigl, R., 181Stucken, R., 115

THORNTON, H., 109, 152, 219Timoshenko, V. P., 168Tooke, T., 153Torrens, R., 153

WALRAS, L., 218Weber, A., 153, 211Wicksell, K., 34, 47, no , i n ,

ii3#, 128, 133, 134, 145, 146,148, 153, 178, 209, 210, 211,212, 219

Wieser, F., 108Withers, H., 153, 154Working, H., 233, 234, 235

YOUNG, A. A., 233

244

Monetary Theory and the Trade Cycle

Documents

Transcript of Monetary Theory and the Trade Cycle