Mr. Keynes on the Causes of Unemployment

7/28/2019 Mr. Keynes on the Causes of Unemployment

1/22

Mr. Keynes on the Causes of UnemploymentThe General Theory of Employment Interest and Money by John Maynard KeynesReview by: Jacob VinerThe Quarterly Journal of Economics, Vol. 51, No. 1 (Nov., 1936), pp. 147-167Published by: Oxford University Press

Stable URL: http://www.jstor.org/stable/1882505 .Accessed: 22/03/2013 10:02

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of

content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms

of scholarship. For more information about JSTOR, please contact [email protected].

.

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to The Quarterly

Journal of Economics.

http://www.jstor.org

This content downloaded from 146.141.15.223 on Fri, 22 Mar 2013 10:02:09 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/action/showPublisher?publisherCode=ouphttp://www.jstor.org/stable/1882505?origin=JSTOR-pdfhttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/stable/1882505?origin=JSTOR-pdfhttp://www.jstor.org/action/showPublisher?publisherCode=oup


2/22

MR. KEYNES ON THECAUSES OF UNEMPLOYMENT'The indebtedness f economists o Mr. Keynes has beengreatlyncreasedby this atest additionto hisseriesofbril-liant,original, nd provocativebooks,whose contributionoourenlightenment illprove, am sure,to have been evengreater n the longthan in the short run. This book deals

with almost everything, ut the causes of and the futureprospects of unemployment,yclical and secular, are itscentraltheme. It bringsmuchnewlight,but its displayofdialectical kill s so overwhelminghat t willhave probablymorepersuasivepowerthan it deserves, nd a concentrationon thepointswhere think can detectdefectsn theargu-ment,tho it would be unfair fpresented s an appraisal ofthe merits fthe book as a whole,may be moreuseful hanwould a catalogue which would have to be long to becomplete of ts pointsofoutstandingntellectual chieve-ment.Written hoit is by a stylist fthefirst rder, he book isnot easy to read, to master,or to appraise. An extremelywiderange ofproblems,noneof themsimpleones, are dealtwithin an unnecessarily mall number of pages. Had thebookbeen made longer, he timerequired orreading t witha fairdegreeofunderstanding ouldhave been shorter, ortheargument ften roceeds t breakneck peedandrepeatedrereadings renecessary eforet can be grasped. The book,moreover, reaks withtraditionalmodes of approachto itsproblems t a numberof points at the greatestpossiblenumber fpoints, ne suspects and no old termfor n oldconceptis used when a new one can be coined, and if oldterms re usednewmeanings re generally ssignedto them.The definitionsrovided,moreover, re sometimes f unbe-lievable complexity. The old-fashioned conomist must,therefore,truggle otonlywithnew deas and newmethods1. JohnMaynardKeynes, The GeneralTheoryof EmploymentInterest nd Money,Macmillan nd Co., London,1936.147

This content downloaded from 146.141.15.223 on Fri, 22 Mar 2013 10:02:09 AMAll use subject toJSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


3/22

148 QUARTERLY JOURNAL OF ECONOMICSofmanipulating hem,but also with a newlanguage. Thereis ample reward,however,forthe expenditure f timeandattention ecessary or ven partialmastery ftheargument.1. " INVOLUNTARY" UNEMPLOYMENTMr. Keynes claimsthatthe "classical"2economists ecog-nizedthe possibility nlyof "frictional" nd of "voluntary"unemployment, nd that a vitally importantchapter ofeconomic heory emains o be written bout a third lass ofunemployment,orwhich herewasno placeinthe"classical"scheme of things, namely, "involuntary" unemployment.The concept of "frictional" unemployment elates to theinevitable oss of time between obs, and presentsno diffi-culties. "Voluntary" unemployments defined s theunem-ployment due to the refusalor inability f a unit of labor... to accept a rewardcorresponding o the value of theproduct attributableto its marginal productivity," ut isused in such manner s to require he additionto thisdefini-tion ofthe proviso hatthe moneywage offeredmustnot bebelow what the laborerregards s a properminimum ate ofmoneywages. If laborersrefuse vailable employment t amoney rate below this minimum, r if employedlaborersrefuse o permit prevailingmoney rate to be lowered ndunemployment esultsforthemselves r forothersfrom hisrefusal,Keynes would apparently egard t as "involuntary"unemployment, ut denyits possibility r probability. Hedefines involuntary"unemployments follows:"Men areinvoluntarily nemployed f, n the event of a small rise inthe priceofwage-goodsrelatively o themoneywage, boththeaggregate upplyof abor willing o workfor hecurrentmoney-wage nd the aggregatedemand for t at that wagewould be greater han theexisting olumeofemployment."(p. 15). What he seemsto meanby this s that any unem-ploymentwhichwould disappear if real wages were to bereducedby a rise in the prices ofwage-goods,moneywagesremaining he sameorrising n lessproportion, utnot alling,2. Used by him to mean the later economists, such as J. S. Mill,Marshall, Edgeworth, Pigou, who in the main were adherents of theRicardian tradition; a usage whichI shall followhere.



4/22

CAUSES OF UNEMPLOYMENT 149would be involuntary. It is with"involuntary"unemploy-ment o understood,ts causes and its remedies, hatKeynes'analysisof unemployments primarilyand almostsolelyconcerned.In Keynes' classification f unemployment y its causes,unemployment ue to downward-rigidityf money-wages,which for the "classical" economistswas the chieftype ofcyclical unemployment nd the only importanttype ofsecular or persistent nemployment,herefore indsno place.As will be seen ater, t is excluded on the ground hatresist-ance to reductions n money wage-ratesgenerallydoes notinvolve a reductionn the volumeof employment nd is, ifanything, avorable o employment ather han the reverse.The omissionchargedagainst the "classical" economistsistheir ailure onote the esserresistance f abor to reductionsin real wages ifunassociatedwithreductions n moneywagesper se, and theirfailure o recognize he existenceofa largevolumeofunemploymentorwhich he formers an availableand practicableremedy, ut not the latter. Keynes' reason-ing points obviously to the superiorityof inflationaryremedies orunemploymentvermoney-wage eductions. na world organized n accordancewith Keynes' specificationstherewould be a constant ace between he printing ress ndthe businessagentsof the trade unions,with the problemofunemploymentargely olved ftheprinting resscouldmain-tain a constant lead and if only volume of employment,irrespective f quality, s consideredmportant.The only clash here between Keynes' position and theorthodox ne is in his denial that reduction f money wagerates is a remedyfor unemployment.Keynes even followsthe classical doctrinetoo closely when he concedes that"with a given organization, quipmentand technique,realwagesand thevolume of output (and henceofemployment)are uniquely correlated, o that, in general,an increaseinemploymentanonlyoccur otheaccompanimentf declinein the rate of real wages" (p. 17). This conclusionresultsfromtoo unqualifiedan application of law-of-diminishing-returns nalysis, nd needs to be modified orcyclicalunem-



5/22

150 QUARTERLY JOURNAL OF ECONOMICSployment, s well as for the possibility hat the prices ofwage-goodsand of other goods may have divergentmove-ments. If a plant geared to workat say 80 per cent of ratedcapacity s being operated at say only 30 per cent, both theper capita andthe marginal utputof abor may well be lowerat the lowrate of operations han at the higher ate, the lawof diminishing eturns otwithstanding. here is the furtheremDiricalconsideration hat if employersoperate in theirwage policy n accordance with marginalcost analysis, t isdone only imperfectlynd unconsciously, nd the level ofwages they can be persuadedto establish s strongly nflu-enced by the profitabilityftheiroperations s a whole,andnot solely if at all - by calculationsof the marginal on-tributions f labor to output.Keynes uses the term "full employment" o signify heabsence of any involuntaryunemployment p. 16). Hedescribes t also as theconditionwhichwould prevail "whenoutputhas risen o a level at whichthe marginal eturn roma representative nit of thefactorsof productionhas fallento the minimumfigure t which a quantity of the factorssufficiento produce hisoutput s available" (p. 303). Thereare impliedhere severalquestionablepropositions.The con-cept of diminishingmarginalproductivitys generallyusedin economics n a partial differentialense to indicate thediminishingncrements f output which would result whensome particular actor r groupoffactorswas being ncreased,the remainder f the working ombination eing held con-stant. If all the factors re being increased simultaneouslyand inuniformroportions,trequires omesuchassumptionas that of the general prevalenceof externaltechnicaldis-economiesfrom ncreasedproduction fit is to be acceptedthatoutput nd return er compoundunitofthefactorsmustbe negatively correlated. There is also implied here theassumptionthat any increasein real wages (money wagesremaining onstant, rrising)willresult n an increase ntheamountof labor available. If, as widely-held pinionsincethe seventeenth enturyhas maintained,and as ProfessorPaul Douglas's recent nvestigations orurban labor in the



6/22

CAUSES OF UNEMPLOYMENT 151United States appear to confirm, he supply schedule oflabor with respectto real wages is, for part of its range atleast, negatively nclined, he volume of employment ouldconceivablybe much greaterwhen therewas "involuntary"unemploymenthan whentherewas "full" employment,ndKeynes' conditions f "full" employmentmight e metat anindefinite umber f evels of employment."Full" employment arelyoccurs, accordingto Keynes,and the mainimmediate esponsibilityorthepersistence f"'involuntary"unemployment ies with the persistenceofinterest atesat levels too high o induceemployerso bid forall the aboravailable at the prevailingmoney ates ofwages.An elaborate and strikingly ovel analysis of the causesdetermininghe level of nterest ates eads to the conclusionthat high "liquidity-preferences"f savers, an excessivedisposition o save and a lowmarginal roductivityf nvest-ment are responsible for the absence of such a relationbetween he ratesat which aversarewilling o lendand therates at which ntrepreneursre willing o borrow or nvest-ment as wouldresult n an approximation o "full" employ-ment.Mr. Keynes claims further: 1) that therecan be "full"employmentonly when entrepreneursmake investmentssufficiento absorb any excessofincome paid-out by entre-preneurs over expenditureson consumptionby income-recipients; 2) that the amountof nvestment ntrepreneursare prepared to make, or their "investmentdemand forcapital," is governedbythe relation ftheir nticipations sto the yieldof additional nvestment, rwhat Keynes callsthe "marginalefficiencyfcapital"3to the interest ates atwhich funds can be borrowed; 3) that the amountwhichincome-recipientsre willing ospendoftheir urrentncome,ortheir propensity o consume," function rimarilyftheamountoftheir ncomes,4 etermineshequantityofsaving;3. "Anticipated marginal efficiency f capital" would seem to me amore accurately descriptive abel forthe concept.4. It is, inmyopinion,probablydependent appreciablyalso on antici-pations as to the prospective trend of income, and is surely affectedsignificantly y amount ofaccumulated wealth at currentvaluations as



7/22

152 QUARTERLY JOURNAL OF ECONOMICSand (4) the rate of interest s determined y (a) "liquiditypreferences" nd (b) thequantityofcash available to satisfysuchpreferences.The quantityofcash is generally ssumedto be a constant. I acceptmost ofthisas valid in itsgeneraloutlines,but I am unableto accept someofKeynes' accountof how these "propensities" operate in practice or hisappraisal oftheirrelative trength.2. THE PROPENSITY TO HOARD

Keynes maintainsthat forcenturiesback the propensityto save has been so muchstronger han the inducement oinvestas to createa substantialbarrier o "full" investment.He findsfault with the " classical" economistsfor theirallegedneglect f the gulfbetween hedesire o save and thedesire to invest, .e., for theirneglect of "liquidity prefer-ences." It was a shortcoming f the Ricardianwingoftheclassicalschoolthat nthe face ofstrong riticismhey tead-fastly dhered otheir osition hathoardingwasso abnormala phenomenon s not to constitute significantontributingfactor o unemploymentvenduring periodofseveredefla-tion. In static equilibrium nalysis, n whichperfectpriceflexibilitys assumed and monetary hangesare abstractedfrom, here s no occasion forconsideration fhoarding. Inmodernmonetarytheory t is generallydealt with,withresultswhich nkind aresubstantiallydenticalwithKeynes',as a factoroperatingto reduce the "velocity" of money.There has been, I believe,common greement mongecono-miststhatwhenprice-rigiditiesre important oarding ouldpresent a serious and continuingproblem, and that it isalwaysa significantactor n thedownwardphase ofa shortbusinesscycle. Keynes,however, ttaches great mportanceto tas a barrier o "full"employmentt almost ll times, ndapparentlyrrespective f the degreeofflexibilityfprices.Thereare severalreasonswhy" liquidity references"oomso large to Keynes as a source of trouble n the economicprocess. He takes it forgranted hat they are ordinarily owellas by currentncome. See infra,?4, for urtheromments n thispoint.



8/22

CAUSES OF UNEMPLOYMENT 153strongforthe averageperson n controlof liquid resourcesthat a substantial nterest ate s required o overcome hem;and apparently hattheycannotbe overcome y any rateofinterest f a stillhigher ate ofinterest s anticipated n thenearfuture. He assignsto themtheroleofsole determinant(given the amount of cash available, whichhe treats ordi-narily s a constant)of therate of nterest. He believesthatthemarginalproductivity unction f capital and thereforethe investmentdemand for capital have little elasticity.Finally he assumes ngeneral hatnothing an satisfy iquid-itypreferencesxceptthat "cash" whose quantity s one ofthe determinants fthe interest ate.We have almostno reliable nformationboutthestrengthof iquiditypreferences ndervarying ircumstances,nd intheabsence ofstatistical nformationf a genuinely elevantcharacterdiscussionmust be based largelyon conjecture.Nevertheless venture o present series of considerationswhich, n theaggregate, eemto warrant heconclusion hatKeynes has grossly xaggerated he extent owhich iquiditypreferencesaveoperated n thepast andare ikely o operatein thefuture s a barrier o "full" employment.(a) Keynesstresses hepressurewhich s exercised y theexpectationof a rise in the interestrate on potentialpur-chasers f ecurities,eading hem opostpone heir urchasesinorder o escape a capitalloss. Thereare, however,neverycountry argenumbers f investorswho have beentaughttobuygilt-edgeecurities n thebasis of theiryieldtomaturityand to disregard hefluctuationsn theirday-to-daymarketvalues. Even investors f a speculativetypeare ordinarilyas anxiousnotto missa "low" as not to buytoo high. Therearemanyopportunitiesor nvestmentwhichare - or seemat the time to be - ofthe "now-or-never" ype. There is awidely-prevalentversionto thewaste of"dead" cash.(b) Keynes seems to exaggerate he actuarialvaluation ofpostponement f investment uringa periodof anticipatedrise in interestrates. Rising interestrates are frequentlyassociated withperiods ofgreater onfidencen the securityofthe nvestment,s far s payment fprincipal nd interest



9/22

154 QUARTERLY JOURNAL OF ECONOMICSaccording o scheduleare concerned; r in thecase ofequitysecurities,withperiods of more favorableanticipationsoflong-runyields. Hence periodsof rising nterestrates areoftenassociated with periods of risingrather than fallingprices of securities, speciallyforequity securities.Keynesseems to be in error lso whenhe assertsthat, abstractingfrom he risk of defaulton principalor interest, t will beequally profitable o hoard as to investat par in a long-termsecurity aying4 per cent fthemarket nterest ate s risingby0.16 per cent perannum. In thefirst lace,hoarding ndinvestment n a long-termecurity re not theonlyalterna-tives. Let it be provisionally ranted hathoarding nd thepurchaseat par ofa 4 percentlong-termondwouldproveequally profitable t theend of thefirstyear ifthe interestrate during hat yearhad risenby 0.16 per cent. The pur-chase at the beginning f the year of a one-yearmaturitysecurity aying nything ver 0.16 per cent would thenhavebeenmoreprofitable ven if t had to be exchangedforcashwithin ix months, nd even ifthe short-termnterest atewere also graduallyrisingby as much as 0.16 per cent perannum. Secondly,even a purchaser f the long-term percentsecuritywouldhave been richer t the end ofthe firstyearthan fhe had hoardedhiscash,unlessthe securitywerea perpetual ond.(c) Even if t be granted hat liquidity-preferencesre asstrongordinarily s Keynes indicates, theiroperationas abarrier o investmentwouldnecessarily e important nly fit be assumed (1) that liquidity-preferencesan be satisfiedsolely by the holdingofnon-investmentssets, and (2) thatthequantity f suchassetsdoesnotautomatically espond othe demand for them. Keynes takes care of this secondqualification y his assumption hat the quantityofmoney- intheassumedabsence of a positive entralmonetaryon-trol is constant. Here, indeed, he concedesmorethan isnecessary,for if liquidity preferences re assumed to bestronger uringdepressions han duringperiods ofbusinessexpansion, henthequantity fmoney, nder uchmonetarysystems s have existed n the past,varies nverselywiththe



10/22

CAUSES OF UNEMPLOYMENT 155strengthof liquidity preferences. But he does not giveadequate considerationo the first ualification.The satisfaction f liquidity preferencen the one handand of nvestmentn the other, re oppositephenomena nlyif the range of assetswhich can satisfy nvestment emandcorrespondswith the range of assets which can satisfyliquidity-preferences,o that it shall be impossible o satisfyboth by the same transaction. If liquidity-preferencesanbe satisfied ytheholding fresourceswhich renot denticalwith the "money" whose surrendersatisfiesinvestmentdemand, the satisfaction f the former oes not necessarilyentail failure o satisfy he atter. Keynes explains iquidity-preference s a wishto retain one's resourcesn the form fmoney. Thereis no systematic xamination fwhatis to beincluded as "money " for this purpose,but incidentally ohis analysis ofone particular orm f surrenderf liquidity,namely, xchangeof moneyfor debt,he states:... wecandraw he inebetween money" nd "debts" atwhateverpoint s most onvenientor andling particularroblem. orexample,we can treatas money ny command vergeneralpurchasing owerwhich he ownerhas notpartedwithfor period n excessof threemonths, nd as debtwhatcannotbe recoveredor longer eriod hanthis;orwecan substituteor threemonths" nemonth rthree aysorthree ours ranyother eriod; r wecanexclude rommoney hat-ever s not egaltender nthe pot. It is often onvenientnpractice oincludenmoneyime-depositsithbanks nd, occasionally,vensuchinstruments s (e.g.) treasurybills. As a rule, I shall . . . assume thatmoneys co-extensive ithbankdeposits p. 167,note).

If everythingwhich satisfies iquidity-preferences to beincluded as money, hen moneymust be broadlydefined oas to includenotonlydemanddeposits nd timedeposits, utalso short-termecurities, ny other ssetswhich re readilymarketablewithout eriousrisk of oss through epreciationofvalue, and even the commandover creditfrombanks orothers. But the conversion fnewly-acquiredash intoanyotherform of asset either nvolves investmentdirectlyortransfershedecision s betweenhoarding nd investmentoa bankeror other intermediaryetween the originalsaverand the ultimate borrower orinvestment. If the banker



11/22

156 QUARTERLY JOURNAL OF ECONOMICSpermitshis investments o remainconstantwhilehis cashreserves are increasing, r if he maintains the same cashreservesfor dle as for active demand deposits, or for timedeposits as fordemand deposits, r for deposits s forbank-notes in circulation, hen the propensity o hoard whichmanifeststself n the maintenance f dle bank depositsdoesoperate to check investment, ut only with the connivanceand support of the bankingmechanism.It may be objected that even if liquidity-preferencesoperate only,orin the main,to checkpurchases f ong-termsecurities, hey tilloperate s a checkto investment; ecausethe atter s and mustbe largelyndurablegoods,orinassetsfarremoved rom he stage ofthe consumers' oods. But therelation between the period of investment ntendedby thesaver andthat ntended, r nfactresulting, ytheborrowingentrepreneurs not a simpleone ofnecessary quality. It ishighlyflexible nd approachesto freevariability t thedis-cretion of the borrower. Every money market has anelaboratemachinery ortransmutinghort-termoans intolong-term nvestmentsnd long-termoans into short-terminvestments,o suit the convenience foriginal enders ndultimateborrowers.The typical ntrepreneur illshift romlong-term o short-term orrowing, r vice versa, even thothe time period involved in the particular operation isunchanged,or (as often)unknowable n advance. He mayalso be able to shift romong-termo short-termnvestmentifthe nterest ateatwhich he atter anbe financed smuchlower hanthatat whichhecan conduct dmittedlyong-termborrowing. f savers have a 5 percentperannumpreferenceforcash overinvestmentn 10-yearbondsbut onlya 1/;percent preferencefor cash over time-deposits r short-termsecurities, nd ifentrepreneurs antfundsfor 10 years andare unwilling o incurthe sacrificeof theirown liquiditywhich wouldbe involved n the attemptto finance10-yearoperationswith say 3-monthborrowings,middlemenwillstep in who are preparedto lend on long-termundswhichthey have borrowed on short-term. The modernmoneymarketis fortunately quipped to some extentwith pro-



12/22

CAUSES OF UNEMPLOYMENT 157ceduresfor atisfyingiquidity-preferencesithout rovidinggenuine iquidity.(d) The propensity o hoard exercises ts influence s arestraintn investmenthroughtstendency o raise nterestrates. But in whatseems to methe mostvulnerablepartofhis analysis,his explanationof thedetermination fthe rateof nterest, eynesassignsto thedesirefor ashforhoardingpurposes grossly xaggeratedmportance.Keynesdenies thevalidityof the "classical" doctrine hatinterest s the rewardforsavingand is directly eterminedbythesupply cheduleof savingswithrespect o the nterestrate and the nvestment emandschedulefor apital,and hisexpositioneaves the impressionhat the interest ate is notdependent o any important xtenton these twofactors.Hedeniesthat nterest s the"rewards for avingon thegroundthat, fa manhoardshis savings n cash,heearns no interest,thohe saves just as much as before p. 167), and claimsthat,on the contrary, t is the rewardforsurrender f liquidity.By analogous reasoninghe could deny that wages are therewardfor abor, or thatprofit s therewardforrisk-taking,because labor is sometimesdone withoutanticipationorrealizationof a return, nd men who assumefinancial iskshave beenknown o incur osses as a result nstead ofprofits.Without avingtherecan be no liquidity o surrender.Thesaver who has no concern about liquiditygets the samereward s the personwho saved with iquidity s his initialobjectivebut is persuadedby the interest ate to lend; andthe return s grantedfor loans irrespectivewhether t isreluctance to postpone consumptionor reluctanceto sur-render iquiditywhichkeepsthe supplyoffundsfor nvest-mentdown to thelevelat whichborrowersrewilling o paytheprevailing ate ofinterest or t. The rate of interest sthereturn or avingwithout iquidity.Keynes explainsthe rate ofinterest s determined y theschedule f iquidity-preferencesndtheavailable quantity fmoney, heprevailing ateof nterest eing imply hatpricefor he sacrificef iquidity t which hedesire o hold cashisequated with the quantityof available cash (p. 167). The



13/22

158 QUARTERLY JOURNAL OF ECONOMICSrate of nterest etermines he amount of investment, iventhe investmentdemand for capital; but a change in theinvestmentdemand for capital will not affect he interestrate "if nothing has happened to the state of liquidity-preference nd the quantity of money." (See especiallythefigure n p. 180, and the text on p. 181).There have been previous ttempts o discover basis onwhichthe interest ate could be held to be determinednde-pendently fthedemandforcapital,the level ofwages,andother mportant lements n the economy,but the growingrecognition f the basic interdependence f all the importanteconomicvariableshas led to widespread kepticismhatanysuch attempt ould succeed. In Keynes' present ttempt hefatalflaws, to repeat, he exaggerated mportance ttributedto hoarding.In his discussion f iquidity-preferenceseynesdistinguishes etween he desire or ashforuse inthe currenttransaction of personal and business exchanges,and thedesirefor cash as a security gainst loss fromunsuccessfulinvestment. As I have already argued,the latter considera-tion shouldnotoperate s a barrier o short-termnvestment,and while tmay inducea high ong-termnterest ate, t willbe compensatedfor n part by a shiftofborrowing o theshort-term arket. The pattern f behaviorofthe desirefortransaction-liquiditys probably very argely he inverseofthat of security-liquidity,r hoarding proper. As D. H.Robertsonpoints out in his contributiono thissymposium,thetransactions-desireor ash is for ash to be usedand notfor ash to be held unused. It mustthereforeary positivelywith he volumeof nvestment,f ncome, ndof xpendituresforconsumption. n so faras it consists f demandforcashfrom ntrepreneursor businessuses, it is but a reflectionftheirinvestmentdemand forcapital. In so far as it is ademandfor ash from onsumerswho are livingbeyondtheircurrentncome, t is the demandforconsumptionoans ofolder theory. Whatever its origin,demand for cash fortransaction urposes s, dollarfordollar, fequal influencentherateof nterest s demandfor ashforhoarding urposes.The demandfor apital and thepropensityo save (which s



14/22

CAUSES OF UNEMPLOYMENT 159thereciprocal fthepropensityoconsume) rethus restored-tho, I admit, in somewhat modifiedand improvedfashion to theirtraditionalr6les as determinants f therateof nterest.While (to repeatagain) relevant tatistical nformationsscarce,whatwe do knowabout theholdersofcash balancesintheUnitedStatespoints tronglyo the mportance fthetransactions-motiveor iquidity nd to therelative nsignifi-cance in ordinary imesofhoarding. It is the corporations,institutions,nd governmentshatholdat all times hebulkofthecashbalances, speciallyf avingsdeposits reexcludedas constitutingnvestments atherthan cash. MoreoverIsuspect I know of no data on the question)thatat least inprosperous imesthe savers thosewho add each year totheirestates -who are supposedby Keynesto be a sourceof so muchtroublebecause of their hoardingpropensities,typicallyholdin cash a smallerpercentage f their ncomes,let alone oftheir otalresources,hando thespenders. Theformer ave investment abits, nd abhor dlecashas natureabhors vacuum. The latterhold cashuntilthebillscome nfor settlement. It would at least be interesting o knowwhether heseare factsorfancies.The importancefthetransactions-demandor ashmakesiteasytoexplain whole eries fhistorical henomenawhichdo not fit ntoKeynes'theory.Because thedemandfor ashforbusinessusevariespositivelywith he nvestmentemandforcapital, and thedemandforcash forpersonaluse variespositivelywiththe level of incomeand ofexpenditures orconsumption, here s no need fortreating s a perplexingpuzzlethefacts, hatbusinesss activewhen nterest atesarehigh and slack when interestrates are low, and that thequantity of money and the interestrate are historicallycorrelatedpositivelyratherthan negatively. There is animportant tabilizing nfluence,moreover,n these circum-stances. During a depressionentrepreneursnd spendersrelease omeof thecashto supply he demandofhoarders orsecurity,nd during n expansion fbusiness heabsorptionof cash by businessand by spenders, erving s it does to

This content downloaded from 146.141.15.223 on Fri, 22 Mar 2013 10:02:09 AMAll use subject to JSTOR Terms and Conditions
http://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsphttp://www.jstor.org/page/info/about/policies/terms.jsp


15/22

160 QUARTERLY JOURNAL OF ECONOMICSraise the nterest ate,keeps theexpansion rom oingbeyondbounds; or, Keynes would say, from ven approaching ea-sonablebounds.3. MONEY-WAGE FLEXIBILITY AND VOLUME OFEMPLOYMENTKeynes expresses sweepingdissentwith the "classical"doctrine hat money-wage igiditys a major cause both ofcyclical and of secularunemployment,lthohe freely rantsthatingeneral ncreased mploymentmustmean lowerrealwages. He maintains hat laborstrongly esistsmoneywagereductionsbut takes reductions n real wages much morecalmly, and thereforehat even if money wage-reductionswere ogically remedy orunemploymentheywouldnotbea practicable ne. His view s that a loweringfmoneywage-rates,unless it proceededsimultaneously nd uniformlyllalong theline, would chiefly lter therelativeratesofwagesofdifferentaborgroups. It wouldnotbe likelyto increasethe aggregatevolumeof employment f labor, and on thebalance of probabilitieswould be more likelyto reduce it.He does not discuss the effects n employmentwhichwouldresultfrom ressure rom abor for ncreasesnmoney-wages,or from ncreasesofmoney-wagesmade voluntarily n thepart of employers,whether for humanitarianreasons orbecause of belief that high wages mean prosperity r inresponse o public opinion.Keynes presentshis own positionmainlyin termsof acriticism f a theorywhich he imputesto the "classical"economists, ccording o which a reduction fmoneywagesand a simultaneousorrespondingeductionn priceswouldincreaseemployment ecause the same volume ofmonetaryexpenditureswould purchasea greater physical output ofcommodities.He easily demolishes hisby pointing utthat,ifmoneywages paid out were to fall in amount and invest-ment by entrepreneursmeasuredin wage-units)did notincrease, he amountofmoney ncomeavailable for xpendi-tureswould fall to an equivalent extent.His discussion ftheeffectsf thewage-reductionn thevolumeof nvestments



16/22

CAUSES OF UNEMPLOYMENT 161mainly in terms of its influenceon the expectations ofentrepreneurss to the future rend ofwages,and he con-cedesthat ifentrepreneursre led to expectfurther hangesto be in an upward direction ts effectwillbe favorable. Heurges,however, hat "it would be much betterthat wagesshould be rigidlyfixed and deemed incapable of materialchangesthan that depressions hould be accompanied by agradualdownward endencyofmoney-wages" p. 265).This does not meet theargument orwage-reduction orrathermoney-cost eduction' during depressionwhichhad understood o be the prevalentone in recentyears. Inthis otherdoctrine, actor-pricesre to be reduced,but not,ornot nthe same degree, he pricesofconsumers' oods. InKeynes' analysisperfect nd active competition s assumed,and prices are supposed to fallimmediately nd in full pro-portion to the fall in marginal variable' costs. If thisoccurred, and output remained the same, prices per unitwould fall in greater bsolute amountthan would average

5. From thepointofviewofeffectn output, hereductionfanypartofvariable osts s dollarfordollar f thesame mportances thereductionf nyother artof uchcosts, nd t sonly s against educ-tion of outstandingixed osts, to the extent hat they also do notconsist f aborcosts, hatthere s anythingo be saidfor eduction flaborcosts npreferenceo other osts. But fromhepointofviewofthe ffectn theemploymentf abor, hereduction f abor ost s morefavorable han hereductionoan equivalentmount f nyother ost,because t willtendto lead to a substitutionf abor for ther actors,tho t willnotbe as favorables thereduction fbothora fortiorif allcosts imultaneouslynd in thesameproportions.6. Keynesdistinguishesetween factor osts"and "usercosts," hetwocombined omprisingprime"costs. By user costshemeanstheamounts aid out to other ntrepreneursorpurchases rom hem ndsacrificesncurred extrawear and tear presumably)n employingequipmentnsteadof eaving t idle. He claims hateconomists avegenerallyquatedsupplypricewithmarginal actor ost, gnoring sercost,whereastshould e equatedwith rime ost. I seenopointnthedistinctionetween urchases romntrepreneursnd direct urchasesof he ervices f hefactors.What sthepoint ndistinguishingetweenthe cost ofcoal to a steelmill ccording s it isbought rom n outsidemineorproduced n its owncollieries?Where s the ine to be drawnbetween ntrepreneursnd "factors"? I am sceptical s to whetheranyeconomistsave,explicitlyrbyimplication,xcluded ost ofpur-chasedmaterials r depreciation fequipmenthrough se from hecostssupposed o determineupply rice.



17/22

162 QUARTERLY JOURNAL OF ECONOMICSvariable costs,7 nd even more, f current abor cost were anegligible lement in the fixed costs, than would averageaggregate costs. The profit tatus of entrepreneurs ouldthenbe less favorable han before. What I understand o bethe current octrine s different. t looks to wage-reductionsduring depression o restore rofit-margins,hus to restorethe investment-moralef entrepreneurs nd to give themagain a creditstatuswhich will enable them to finance nyinvestment heymaywishto make. It reliesupon the occur-rence of a lag betweenthe reduction n wage-rates nd aresponse n reducedvolume of sales at the previousprices,duringwhich nterval ntrepreneursindprices to be higherthan marginal costs and extensions of output thereforeprofitable, rovidedbuyers can be found for the increasedoutput. Increase in expenditures o restoredepleted nven-tories and to replaceinefficientquipment s relied upon toincrease pay rolls sufficientlyo provide the incomeswithwhich the increasedoutput can be bought, and the gain inemployment and in security of employmentfor thosepreviously mployed is expected o releasefor xpenditurethe emergency eservesof the wage-earning lass. On theassumptionthat a largepart of an entrepreneur's xpendi-turesare ordinarily f the postponable lass in the sensethatthey can be deferredwithoutforcing reduction of thescheduled rate of current utput, even tho not without n-creasingthe current ost of production; nd on thefurtherassumptionthat operationsat a loss are conduciveto thepostponement fevery expenditure ot essentialforcurrentoperation,the supportersof this doctrinemaintain thatrecovery f a profitmargin an lead for time to an increasein entrepreneurs'xpendituresmany times the increase intheir net income, or, alternatively, he reduction n theirnet loss. They do not contendthat this s certainto occur,but on the groundthat the chieffactor n governingtheactionofentrepreneurs ithrespect o postponable xpendi-

7. Becausemarginal ostswouldfall n the same proportions aver-agevariable ostsbutwouldbegreaternamount erunit han veragevariable osts.



18/22

CAUSES OF UNEMPLOYMENT 163tures s the currentprofit tatus of their perations s com-pared to their mmediatelypreceding xperience, hey saythat it is a reasonable probability. Where externalpressureon pricesin the face of rigid costs has been an importantfactor n the depression, hey lso expect favorable nfluenceon the volume of employment rom the effect f a wage-reductionon profits nd therefore n the volume of post-ponable expenditures, atherthan from ts effect n prices.WhileKeynes' analysis providesmaterialsfor trengtheningthis doctrine t a numberof points, cannot find n it anyrefutation f its general validity.4. PROPENSITY TO CONSUME

Mr. Keynes himself ells us that the functional elation-ships of thevariouseconomicvariables are more complex nfact han sformally ecognizednhis analysis. Simplificationofthis ort s inevitable,fanalysis s toproceed t all. In thecase, however, fKeynes' "propensity o consume"function,it seemsto me that the simplificationas been carried urtherthan is necessaryto preventthe analysis frombecomingentangled n its own complexities,nd furtherhan is per-missible ftheconcept s to be used fruitfullyn theanalysisof the shortcycle.Keynes explainsthepropensityo consume s a functionalrelationship etweenthe amount of consumptionmeasuredin money-wage nits and the amount of income similarlymeasured. On the assumption hat ncome nterms fmoneywage-units orresponds ubstantially n its variationswiththe variations n level of employment,t is concludedthatincome, onsumption,nd level of employment re relatedtoeach other n a simple pattern. WritingCw,foramount ofconsumptionnwage-units nd Yw,or ncome n wage-units,and accepting s a close approximationhat Y, is a uniquefunctionf the evel ofemployment, e statesthepropensityto consumefunction s: Cw,= X (Yw) (p. 90).Keynes lists a number ffactors, p. 96) "subjective" and"objective,"whichmight ffecthe value ofX, Yw remaining



19/22

164 QUARTERLY JOURNAL OF ECONOMICSconstant,but he assumes in general that the "subjective"factors emain onstant, t least over shortperiods, nd that,given Y,,, X depends only on changes in the "objective"factors, which in the aggregate he takes to be of minorimportance s comparedto changes n Y, Several "objec-tive" factorswhich he does not appear to have taken intoaccount seem importantenough in the short cycle to bedeserving t least ofmention.Keynes believes that, apart from heeffect fa change nthe wage-uniton the distribution f income between entre-preneurs nd rentiers, ho mighthave differentropensitiesto consume,he has made adequate allowance n hisformulafor changes in expenditureresultingfromchanges in thewage-unitby measuringboth consumption nd income inwage-units. This disregards he possibility hat, forshortperiodsat least, the distinctionwhichKeynes' makes in hissupplyfunction f aborbetween heresponse f abor (1) tochanges n realwages accompaniedbycorrespondinghangesin moneywages, and (2) to changes n real wages resultingfrom he changes n the pricesofwage-goods,moneywagesremaining he same, may have a parallel in the propensityto consume function. The response of consumptionto areduction n real income may be, for a time, substantiallydifferentfthe reductiontakes the form of a decrease inmoney-income, rices remaining he same, fromwhat itwould be if money-income emainedthe same but pricesincreased.\ier.eynesclaimsthat ngeneralrichcountries reworseoffhanpoorcountrieswithrespect o avoidanceof" involun-tary" employment ecause ofthe lesserpropensitieso con-sume in theformer han in the latter, nd thusthegreaterpotential mportance fhoarding.Since wouldcontend hatover long periods, given a flexibleprice system,the pro-pensity o consumewill affect he rate ofcapital accumula-tion ratherthan thevolumeof employment, willconfinemyself o a considerationf thecomparative ituation f therich and poor countrieswithrespect o theshortcycle. Thepossessionof largeaccumulatedresources houldoperateto



20/22

CAUSES OF UNEMPLOYMENT 165level out the rate of consumptionn the face of fluctuationsnincome, nd therefore o check both the downward nd theupward phases of the cycle. Corresponding o the chargesagainstthe entrepreneur'sudgetwhich re fixed n aggregatemonetary mount regardless f current utput, there are inthe ordinary onsumer'sbudget tems of monetary xpendi-ture which are fixed for a time, very much regardlessofchanges n hismoney ncomeas faras reductions herein reconcerned, nd whichtendto be increasedonlyas theresultofcareful eliberationnresponse o anticipation fa changeof someduration n the individual'seconomic tatus. Asidefrom he probability hat such fixed harges re ordinarilygreater roportion f the expenditures f the richthan of thepoor,the poor in timesof severedepressionhave a partialmeansofescape from hem, nthe form fdefaults, o whichthosewithresources ubjectto levycannotresort. What thisamountsto is that C., should be treatedas a function otonlyofYb,, ut also oftheamountofaccumulatedresourcesmeasured n wageunitsheldby theindividual. In so far asthepossession fresources perates n themanner uggestedhere,wealth becomesa stabilizingrather han a disturbingfactor. The explanationofthe apparently ndisputablefactthat thecyclicaldisturbancesremoresevere n richthan npoorcountrieswould then have to be soughtelsewhere hanin the differencesetweenrich and poor in propensities oconsume. My ownguess s that it is to be sought argely nthe differencesetweenthe cyclical behavior of rich andpoorwithrespect o thedisposition fthe ncomewhich heydo not spend. The rich hoard only duringdepressions nddishoardfor nvestment uringprosperity, hereasthepoorhoardsomeoftheir mergencyeserves uring rosperitynddishoardduringdepression.Mir.Keynes says that a fundamental sychological aw,upon which we have a rightto depend both on a priorigrounds nd on thebasis ofexperiences thatdyw spositivedYwand less than unity; i.e., that in termsof wage-units on-



21/22

166 QUARTERLY JOURNAL OF ECONOMICSsumptionvaries in the same directionas income,but insmallerabsolute amount than income (p. 96). This seemsaltogetherreasonable. It leaves unanswered,however,aquestionofsome nterest: oesC,,ever, xceptperhapsunderwar conditions, xceed YW? Since the community xcess ofY,,overC, constitutes ew nvestment,fC,,, everexceededYW herewouldbe continuous,ho fluctuating,ccumulationofcapital resources, ven through he depthsofdepression.Mr. Keynes apparentlymustbelievethatforthe world as awholetheC,,,'smustoften nd substantiallyxceed theY,,'s,forhe holds that in spite of"severalmilleniaofsteady ndi-vidual saving" the world is poor in accumulated capitalassets.8 But what evidencethere s seems to indicatethat, fany acceptable mode of measuring physical amount ofcapitalcould be found and applied,it would show that thewesternworld has been gettingwealthierfairlysteadilyduring ay the past century nd a half,not only n termsofaggregateresourcesbut per capita, in spite of a three-orfour-foldncreaseofpopulation.In connectionwiththepropensity o consumeconcept,aswithmost of Keynes' concepts,the question arises in mymind howtheseconceptswouldhave to be restated n orderto providespecifications or the construction f statisticalseriesby whichhis conclusions s to the natureand modeofbehavior throughtime of the various functions ould beinductively ested,and I regret hat no suggestions f thissortareprovidednthisbook. I am disposedto supportMr.Robertson in his claim that concepts expressedin more"monetary" terms, and expressionsfor the relationshipsbetweenvariableswhichmakespecificllowancefor ime-lagsinsteadof assigninguniform ime-units o all the variables,have forpurposesofa prior analysissome pointsofsuperi-ority verKeynes' "propensity" oncepts xpressednterms

8. "That theworld ftereveralmillennia f teadyndividualaving,is so pooras it is in accumulated apital-assets,s to be explained,nmyopinion, either y the mprovidentropensitiesfmankind, oreven by thedestructionfwar,but by thehigh iquidity-premiumsformerlyttaching o the ownership f land and now attaching omoney" p. 242).



22/22

CAUSES OF UNEMPLOYMENT 167ofa singletime-unit.For purposesof nductiveverification,assuming hatthe statisticaldata available willever be in aform elevant o the answerof mportant uestions, t seemsobvioustomethat theanalysiswouldhave to be extensivelyrestated n termsofdirections nd degreesoftime-lags.

JACOB VINER.UNIVERSITY OF CHICAGO

Documents

Mr. Keynes on the Causes of Unemployment