TOCHAPTER19
preservefullemployment。WithoutthisassumptionProfessorPigou'sanalysisbreaksdownandprovidesnomeansofdeterminingwhatthevolumeofemploymentwillbe。Itis，indeed，strangethatProfessorPigoushouldhavesupposedthathecouldfurnishatheoryofunemploymentwhichinvolvesnoreferenceatalltochangesintherateofinvestmenti。e。tochangesinemploymentinthenon-wage-goodsindustriesdue，nottoachangeinthesupplyfunctionoflabour，buttochangesine。g。eithertherateofinterestorthestateofconfidence。
Histitlethe'TheoryofUnemployment'is，therefore，somethingofamisnomer。Hisbookisnotreallyconcernedwiththissubject。Itisadiscussionofhowmuchemploymenttherewillbe，giventhesupplyfunctionoflabour，whentheconditionsforfullemploymentaresatisfied。Thepurposeoftheconceptoftheelasticityoftherealdemandforlabourintheaggregateistoshowbyhowmuchfullemploymentwillriseorfallcorrespondingtoagivenshiftinthesupplyfunctionoflabour。Or¾；alternativelyandperhapsbetter¾；wemayregardhisbookasanon-causativeinvestigationintothefunctionalrelationshipwhichdetermineswhatlevelofrealwageswillcorrespondtoanygivenlevelofemployment。Butitisnotcapableoftellinguswhatdeterminestheactuallevelofemployment；andontheproblemofinvoluntaryunemploymentithasnodirectbearing。
IfProfessorPigouweretodenythepossibilityofinvoluntaryunemploymentinthesenseinwhichIhavedefineditabove，as，perhaps，hewould，itisstilldifficulttoseehowhisanalysiscouldbeapplied。Forhisomissiontodiscusswhatdeterminestheconnectionbetweenxandy，i。e。betweenemploymentinthewage-goodsandnon-wage-goodsindustriesrespectively，stillremainsfatal。
Moreover，heagreesthatwithincertainlimitslabourinfactoftenstipulates，notforagivenrealwage，butforagivenmoney-wage。ButinthiscasethesupplyfunctionoflabourisnotafunctionofF'x
alonebutalsoofthemoney-priceofwage-goods；¾；withtheresultthatthepreviousanalysisbreaksdownandanadditionalfactorhastobeintroduced，withouttherebeinganadditionalequationtoprovideforthisadditionalunknown。Thepitfallsofapseudo-mathematicalmethod，whichcanmakenoprogressexceptbymakingeverythingafunctionofasinglevariableandassumingthatallthepartialdifferentialsvanish，couldnotbebetterillustrated。Foritisnogoodtoadmitlateronthatthereareinfactothervariables，andyettoproceedwithoutre-writingeverythingthathasbeenwrittenuptothat[Page276]THE
GENERALTHEORYOFEMPLOYMENT
point。Thusifwithinlimitsitisamoney-wageforwhichlabourstipulates，westillhaveinsufficientdata，evenifweassumethatn=xy，unlessweknowwhatdeterminesthemoney-priceofwage-goods。For，themoney-priceofwage-goodswilldependontheaggregateamountofemployment。Thereforewecannotsaywhataggregateemploymentwillbe，untilweknowthemoney-priceofwage-goods；andwecannotknowthemoney-priceofwage-goodsuntilweknowtheaggregateamountofemployment。
Weare，asIhavesaid，oneequationshort。Yetitmightbeaprovisionalassumptionofarigidityofmoney-wages，ratherthanofrealwages，whichwouldbringourtheorynearesttothefacts。Forexample，money-wagesinGreatBritainduringtheturmoilanduncertaintyandwidepricefluctuationsofthedecade1924-1934werestablewithinarangeof6percent，whereasrealwagesfluctuatedbymorethan20percent。Atheorycannotclaimtobeageneraltheory，unlessitisapplicabletothecasewhereortherangewithinwhichmoney-wagesarefixed，justasmuchastoanyothercase。Politiciansareentitledtocomplainthatmoney-wagesoughttobehighlyflexible；butatheoristmustbepreparedtodealindifferentlywitheitherstateofaffairs。Ascientifictheorycannotrequirethefactstoconformtoitsownassumptions。
WhenProfessorPigoucomestodealexpresslywiththeeffectofareductionofmoney-wages，heagain，palpablytomymind，introducestoofewdatatopermitofanydefiniteanswerbeingobtainable。Hebeginsbyrejectingtheargumentop。cit。p。101that，ifmarginalprimecostisequaltomarginalwage-cost，non-wage-earners'incomeswillbealtered，whenmoney-wagesarereduced，inthesameproportionaswage-earners'，onthegroundthatthisisonlyvalid，ifthequantityofemploymentremainsunaltered¾；whichistheverypointunderdiscussion。Butheproceedsonthenextpageop。cit。p。102tomakethesamemistakehimselfbytakingashisassumptionthat'attheoutsetnothinghashappenedtonon-wage-earnersmoney-income'，which，ashehasjustshown，isonlyvalidifthequantityofemploymentdoesnotremainunaltered-whichistheverypointunderdiscussionInfact，noanswerispossible，unlessotherfactorsareincludedinourdata。
Themannerinwhichtheadmission，thatlabourinfactstipulatesforagivenmoney-wageandnotforagivenrealwageprovidedthattherealwagedoesnotfallbelowacertainminimum，affectstheanalysis，canalsobeshownbypointingoutthatinthiscasetheassumptionthatmorelabourisnotavailableexceptatagreaterrealwage，whichisfundamentaltomostof[Page277]APPENDIX
TOCHAPTER19
theargument，breaksdown。Forexample，ProfessorPigourejectsop。
cit。p。75thetheoryofthemultiplierbyassumingthattherateofrealwagesisgiven，i。e。that，therebeingalreadyfullemployment，noadditionallabourisforthcomingatalowerrealwage。Subjecttothisassumption，theargumentis，ofcourse，correct。ButinthispassageProfessorPigouiscriticisingaproposalrelatingtopracticalpolicy；anditisfantasticallyfarremovedfromthefactstoassume，atatimewhenstatisticalunemploymentinGreatBritainexceeded2，000，000i。e。whentherewere2，000，000menwillingtoworkattheexistingmoney-wage，thatanyriseinthecostofliving，howevermoderate，relativelytothemoney-wagewouldcausethewithdrawalfromthelabourmarketofmorethantheequivalentofallthese2，000，000men。
ItisimportanttoemphasisethatthewholeofProfessorPigou'sbookiswrittenontheassumptionthatanyriseinthecostofliving，howevermoderate，relativelytothemoney-wagewillcausethewithdrawalfromthelabourmarketofanumberofworkersgreaterthanthatofalltheexistingunemployed。
Moreover，ProfessorPigoudoesnotnoticeinthispassageop。cit。
p。75thattheargument，whichheadvancesagainst'secondary'employmentasaresultofpublicworks，is，onthesameassumptions，equallyfataltoincreased'primary'employmentfromthesamepolicy。Foriftherealrateofwagesrulinginthewage-goodsindustriesisgiven，noincreasedemploymentwhateverispossible¾；except，indeed，asaresultofnon-wage-earnersreducingtheirconsumptionofwage-goods。
Forthosenewlyengagedintheprimaryemploymentwillpresumablyincreasetheirconsumptionofwage-goodswhichwillreducetherealwageandhenceonhisassumptionsleadtoawithdrawaloflabourpreviouslyemployedelsewhere。YetProfessorPigouaccepts，apparently，thepossibilityofincreasedprimaryemployment。Thelinebetweenprimaryandsecondaryemploymentseemstobethecriticalpsychologicalpointatwhichhisgoodcommonsenseceasestooverbearhisbadtheory。
ThedifferenceintheconclusionstowhichtheabovedifferencesinassumptionsandinanalysisleadcanbeshownbythefollowingimportantpassageinwhichProfessorPigousumsuphispointofview:'Withperfectlyfreecompetitionamongworkpeopleandlabourperfectlymobile，thenatureoftherelationi。e。betweentherealwage-ratesforwhichpeoplestipulateandthedemandfunctionforlabourwillbeverysimple。Therewillalwaysbeatworkastrongtendencyforwage-ratestobesorelatedtodemandthateverybodyisemployed。Hence，instable[Page278]THE
GENERALTHEORYOFEMPLOYMENT
conditionseveryonewillactuallybeemployed。Theimplicationisthatsuchunemploymentasexistsatanytimeisduewhollytothefactthatchangesindemandconditionsarecontinuallytakingplaceandthatfrictionalresistancespreventtheappropriatewageadjustmentsfrombeingmadeinstantaneously。'
[135]
Heconcludesop。cit。p。253thatunemploymentisprimarilyduetoawagepolicywhichfailstoadjustitselfsufficientlytochangesintherealdemandfunctionforlabour。ThusProfessorPigoubelievesthatinthelongrununemploymentcanbecuredbywageadjustments；[136]
whereasImaintainthattherealwagesubjectonlytoaminimumsetbythemarginaldisutilityofemploymentisnotprimarilydeterminedby'wageadjustments'thoughthesemayhaverepercussionsbutbytheotherforcesofthesystem，someofwhichinparticulartherelationbetweenthescheduleofthemarginalefficiencyofcapitalandtherateofinterestProfessorPigouhasfailed，ifIamright，toincludeinhisformalscheme。
Finally，whenProfessorPigoucomestothe'CausationofUnemployment'
hespeaks，itistrue，offluctuationsinthestateofdemand，muchasIdo。ButheidentifiesthestateofdemandwiththeRealDemandFunctionforLabour，forgetfulofhownarrowathingthelatterisonhisdefinition。
FortheRealDemandFunctionforLabourdependsbydefinitionaswehaveseenaboveonnothingbuttwofactors，namely1therelationshipinanygivenenvironmentbetweenthetotalnumberofmenemployedandthenumberwhohavetobeemployedinthewage-goodsindustriestoprovidethemwithwhattheyconsume，and2thestateofmarginalproductivityinthewage-goodsindustries。YetinPartVofhisTheoryofUnemploymentfluctuationsinthestateof'therealdemandforlabour'aregivenapositionofimportance。The'realdemandforlabour'isregardedasafactorwhichissusceptibleofwideshort-periodfluctuationsop。cit。
PartV，chaps。vi-xii。，andthesuggestionseemstobethatswingsin'therealdemandforlabour'are，incombinationwiththefailureofwagepolicytorespondsensitivelytosuchchanges，largelyresponsibleforthetradecycle。Tothereaderallthisseems，atfirst，reasonableandfamiliar。For，unlesshegoesbacktothedefinition，'fluctuationsintherealdemandforlabour'willconveytohismindthesamesortofsuggestionasImeantoconveyby'fluctuationsinthestateofaggregatedemand'。Butifwegobacktothedefinitionofthe'realdemandforlabour'，allthis[Page279]APPENDIX
TOCHAPTER19
losesitsplausibility。Forweshallfindthatthereisnothingintheworldlesslikelytobesubjecttosharpshort-periodswingsthanthisfactor。
ProfessorPigou's'realdemandforlabour'depends，bydefinition，onnothingbutFx，whichrepresentsthephysicalconditionsofproductioninthewage-goodsindustries，andx，whichrepresentsthefunctionalrelationshipbetweenemploymentinthewage-goodsindustriesandtotalemploymentcorrespondingtoanygivenlevelofthelatter。Itisdifficulttoseeareasonwhyeitherofthesefunctionsshouldchange，exceptgraduallyoveralongperiod。Certainlythereseemsnoreasontosupposethattheyarelikelytofluctuateduringatradecycle。ForFx
canonlychangeslowly，and，inatechnicallyprogressivecommunity，onlyintheforwarddirection；whilstxwillremainstable，unlesswesupposeasuddenoutbreakofthriftintheworkingclasses，or，moregenerally，asuddenshiftinthepropensitytoconsume。Ishouldexpect，therefore，thattherealdemandforlabourwouldremainvirtuallyconstantthroughoutatradecycle。IrepeatthatProfessorPigouhasaltogetheromittedfromhisanalysistheunstablefactor，namelyfluctuationsinthescaleofinvestment，whichismostoftenatthebottomofthephenomenonoffluctuationsinemployment。
IhavecriticisedatlengthProfessorPigou'stheoryofunemploymentnotbecauseheseemstometobemoreopentocriticismthanothereconomistsoftheclassicalschool；butbecausehisistheonlyattemptwithwhichIamacquaintedtowritedowntheclassicaltheoryofunemploymentprecisely。
Thusithasbeenincumbentonmetoraisemyobjectionstothistheoryinthemostformidablepresentmentinwhichithasbeenadvanced。TheEmploymentFunction[Page280]
Chapter20
THEEMPLOYMENTFUNCTION[137]
I
Inchapter3p。23wehavedefinedtheaggregatesupplyfunctionZ
=fN，whichrelatestheemploymentNwiththeaggregatesupplypriceofthecorrespondingoutput。Theemploymentfunctiononlydiffersfromtheaggregatesupplyfunctioninthatitis，ineffect，itsinversefunctionandisdefinedintermsofthewage-unit；theobjectoftheemploymentfunctionbeingtorelatetheamountoftheeffectivedemand，measuredintermsofthewage-unit，directedtoagivenfirmorindustryortoindustryasawholewiththeamountofemployment，thesupplypriceoftheoutputofwhichwillcomparetothatamountofeffectivedemand。ThusifanamountofeffectivedemandDwr，measuredinwage-units，directedtoafirmorindustrycallsforthanamountofemploymentNrinthatfirmorindustry，theemploymentfunctionisgivenbyNr=FrDwr。Or，moregenerally，ifweareentitledtoassumethatDwrisauniquefunctionofthetotaleffectivedemandDw，theemploymentfunctionisgivenbyNr=FrDw
Thatistosay，NrmenwillbeemployedinindustryrwheneffectivedemandisDw。
Weshalldevelopinthischaptercertainpropertiesoftheemploymentfunction。Butapartfromanyinterestwhichthesemayhave，therearetworeasonswhythesubstitutionoftheemploymentfunctionfor[Page281]THE
EMPLOYMENTFUNCTION
theordinarysupplycurveisconsonantwiththemethodsandobjectsofthisbook。Inthefirstplace，itexpressestherelevantfactsintermsoftheunitstowhichwehavedecidedtorestrictourselves，withoutintroducinganyoftheunitswhichhaveadubiousquantitativecharacter。Inthesecondplace，itlendsitselftotheproblemsofindustryandoutputasawhole，asdistinctfromtheproblemsofasingleindustryorfirminagivenenvironment，moreeasilythandoestheordinarysupplycurve¾；forthefollowingreasons。
Theordinarydemandcurveforaparticularcommodityisdrawnonsomeassumptionastotheincomesofmembersofthepublic，andhastobere-drawniftheincomeschange。Inthesamewaytheordinarysupplycurveforaparticularcommodityisdrawnonsomeassumptionastotheoutputofindustryasawholeandisliabletochangeiftheaggregateoutputofindustryischanged。When，therefore，weareexaminingtheresponseofindividualindustriestochangesinaggregateemployment，wearenecessarilyconcerned，notwithasingledemandcurveforeachindustry，inconjunctionwithasinglesupplycurve，butwithtwofamiliesofsuchcurvescorrespondingtodifferentassumptionsastotheaggregateemployment。Inthecaseoftheemploymentfunction，however，thetaskofarrivingatafunctionforindustryasawholewhichwillreflectchangesinemploymentasawholeismorepracticable。
Forletusassumetobeginwiththatthepropensitytoconsumeisgivenaswellastheotherfactorswhichwehavetakenasgiveninchapter18above，andthatweareconsideringchangesinemploymentinresponsetochangesintherateofinvestment。Subjecttothisassumption，foreverylevelofeffectivedemandintermsofwage-unitstherewillbeacorrespondingaggregateemploymentandthiseffectivedemandwillbedividedindeterminateproportionsbetweenconsumptionandinvestment。Moreover，eachlevelofeffectivedemandwillcorrespondtoagivendistribution[Page282]THE
GENERALTHEORYOFEMPLOYMENT
ofincome。Itisreasonable，therefore，furthertoassumethatcorrespondingtoagivenlevelofaggregateeffectivedemandthereisauniquedistributionofitbetweendifferentindustries。
Thisenablesustodeterminewhatamountofemploymentineachindustrywillcorrespondtoagivenlevelofaggregateemployment。Thatistosay，itgivesustheamountofemploymentineachparticularindustrycorrespondingtoeachlevelofaggregateeffectivedemandmeasuredintermsofwage-units，sothattheconditionsaresatisfiedforthesecondformoftheemploymentfunctionfortheindustry，definedabove，namelyNr=FrDwThuswehavetheadvantagethat，intheseconditions，theindividualemploymentfunctionsareadditiveinthesensethattheemploymentfunctionforindustryasawhole，correspondingtoagivenlevelofeffectivedemand，isequaltothesumoftheemploymentfunctionsforeachseparateindustry；i。e。FrDw=N
=SNr=SFrDw。Next，letusdefinetheelasticityofemployment。TheelasticityofemploymentforagivenindustryisdNrDwreer=¾；¾；¾；¾；×；¾；¾；¾；，dDwrNrsinceitmeasurestheresponseofthenumberoflabour-unitsemployedintheindustrytochangesinthenumberofwage-unitswhichareexpectedtobespentonpurchasingitsoutput。TheelasticityofemploymentforindustryasawholeweshallwritedNDwee=¾；¾；¾；¾；×；¾；¾；¾；，dDwNrProvidedthatwecanfindsomesufficientlysatisfactorymethodofmeasuringoutput，itisalsousefultodefinewhatmaybecalledtheelasticityofoutputorproduction，whichmeasurestherateatwhichoutput[Page283]THE
EMPLOYMENTFUNCTION
inanyindustryincreaseswhenmoreeffectivedemandintermsofwage-unitsisdirectedtowardsit，namelydOrDwreor=¾；¾；¾；¾；×；¾；¾；¾；，dDwrOrProvidedwecanassumethatthepriceisequaltothemarginalprimecost，wethenhave1
DDwr=¾；¾；¾；¾；DPr1-eorwherePristheexpectedprofit。[138]Itfollowsfromthisthatifeor=0，i。e。iftheoutputoftheindustryisperfectlyinelastic，thewholeoftheincreasedeffectivedemandintermsofwage-unitsisexpectedtoaccruetotheentrepreneurasprofit，i。e。DDwr=DPr；whilstifeor=1，i。e。iftheelasticityofoutputisunity，nopartoftheincreasedeffectivedemandisexpectedtoaccrueasprofit，thewholeofitbeingabsorbedbytheelementsenteringintomarginalprimecost。
Moreover，iftheoutputofanindustryisafunctionfNr
ofthelabouremployedinit，wehave[139]1-eorNrf"Nr
¾；¾；¾；¾；=-¾；¾；¾；¾；¾；¾；¾；，eerpwr{f'Nr}2wherepwristheexpectedpriceofaunitofoutputin[Page284]THE
GENERALTHEORYOFEMPLOYMENT
termsofthewage-unit。Thustheconditioneor=1meansthatf"Nr
=0，i。e。thatthereareconstantreturnsinresponsetoincreasedemployment。
Now，insofarastheclassicaltheoryassumesthatrealwagesarealwaysequaltothemarginaldisutilityoflabourandthatthelatterincreaseswhenemploymentincreases，sothatthelaboursupplywillfalloff；cet。
par。，ifrealwagesarereduced，itisassumingthatinpracticeitisimpossibletoincreaseexpenditureintermsofwage-units。Ifthisweretrue，theconceptofelasticityofemploymentwouldhavenofieldofapplication。
Moreover，itwould，inthisevent，beimpossibletoincreaseemploymentbyincreasingexpenditureintermsofmoney；formoney-wageswouldriseproportionatelytotheincreasedmoneyexpendituresothattherewouldbenoincreaseofexpenditureintermsofwage-unitsandconsequentlynoincreaseinemployment。Butiftheclassicalassumptiondoesnotholdgood，itwillbepossibletoincreaseemploymentbyincreasingexpenditureintermsofmoneyuntilrealwageshavefallentoequalitywiththemarginaldisutilityoflabour，atwhichpointtherewill，bydefinition，befullemployment。
Ordinarily，ofcourse，eorwillhaveavalueintermediatebetweenzeroandunity。Theextenttowhichpricesintermsofwage-units
willrise，i。e。theextenttowhichrealwageswillfall，whenmoneyexpenditureisincreased，depends，therefore，ontheelasticityofoutputinresponsetoexpenditureintermsofwage-units。
LettheelasticityoftheexpectedpricepwrinresponsetochangesineffectivedemandDwr，namelydpwr/dDwr
×；Dwr/pwr，bewrittene'pr。
SinceOr×；pwr=Dwr，wehavedOrDwrdpwrDwr¾；¾；¾；¾；×；¾；¾；¾；¾；¾；¾；¾；×；¾；¾；¾；=
1
dDwrOrdDwrpwrore'preor=1。
[Page285]THE
EMPLOYMENTFUNCTION
Thatistosay，thesumoftheelasticitiesofpriceandofoutputinresponsetochangesineffectivedemandmeasuredintermsofwage-units
isequaltounity。Effectivedemandspendsitsell，partlyinaffectingoutputandpartlyinaffectingprice，accordingtothislaw。
Ifwearedealingwithindustryasawholeandarepreparedtoassumethatwehaveaunitinwhichoutputasawholecanbemeasured，thesamelineofargumentapplies，sothate'peo=1，wheretheelasticitieswithoutasuffixrapplytoindustryasawhole。
Letusnowmeasurevaluesinmoneyinsteadofwage-unitsandextendtothiscaseourconclusionsinrespectofindustryasawhole。
IfWstandsforthemoney-wagesofaunitoflabourandpfortheexpectedpriceofaunitofoutputasawholeintermsofmoney，wecanwriteep=Ddp/pdD
fortheelasticityofmoney-pricesinresponsetochangesineffectivedemandmeasuredintermsofmoney，andew=DdW/WdDfortheelasticityofmoney-wagesinresponsetochangesineffectivedemandintermsofmoney。Itistheneasilyshownthatep=1=eo1-ew。[140]
Thisequationis，asweshallseeinthenextchapter，firststeptoageneralisedquantitytheoryofmoney。
[Page286]THE
GENERALTHEORYOFEMPLOYMENT
Ifeo=0orifew=1，outputwillbeunalteredandpriceswillriseinthesameproportionaseffectivedemandintermsofmoney。Otherwisetheywillriseinasmallerproportion。
II
Letusreturntotheemploymentfunction。Wehaveassumedintheforegoingthattoeveryleveloraggregateeffectivedemandtherecorrespondsauniquedistributionofeffectivedemandbetweentheproductsofeachindividualindustry。Now，asaggregateexpenditurechanges，thecorrespondingexpenditureontheproductsofanindividualindustrywillnot，ingeneral，changeinthesameproportion；¾；partlybecauseindividualswillnot，astheirincomesrise，increasetheamountoftheproductsofeachseparateindustry，whichtheypurchase，inthesameproportion，andpartlybecausethepricesofdifferentcommoditieswillrespondindifferentdegreestoincreasesinexpenditureuponthem。
Itfollowsfromthisthattheassumptionuponwhichwehaveworkedhitherto，thatchangesinemploymentdependsolelyonchangesinaggregateeffectivedemandintermsofwage-units，isnobetterthanafirstapproximation，ifweadmitthatthereismorethanonewayinwhichanincreaseofincomecanbespent。Forthewayinwhichwesupposetheincreaseinaggregatedemandtobedistributedbetweendifferentcommoditiesmayconsiderablyinfluencethevolumeofemployment。If，forexample，theincreaseddemandislargelydirectedtowardsproductswhichhaveahighelasticityofemployment，theaggregateincreaseinemploymentwillbegreaterthanifitislargelydirectedtowardsproductswhichhavealowelasticityofemployment。
Inthesamewayemploymentmayfalloffwithouttherehavingbeenanychangeinaggregatedemand，ifthedirectionofdemandischangedinfavourofproductshavingarelativelylowelasticityofemployment。
[Page287]THE
EMPLOYMENTFUNCTION
Theseconsiderationsareparticularlyimportantifweareconcernedwithshort-periodphenomenainthesenseofchangesintheamountordirectionofdemandwhicharenotforeseensometimeahead。Someproductstaketimetoproduce，sothatitispracticallyimpossibletoincreasethesupplyofthemquickly。Thus，ifadditionaldemandisdirectedtothemwithoutnotice，theywillshowalowelasticityofemployment；althoughitmaybethat，givensufficientnotice，theirelasticityofemploymentapproachesunity。
ItisinthisconnectionthatIfindtheprincipalsignificanceoftheconceptionofaperiodofproduction。Aproduct，Ishouldprefertosay，[141]hasaperiodofproductionnifntime-unitsofnoticeofchangesinthedemandforithavetobegivenifitistoofferitsmaximumelasticityofemployment。Obviouslyconsumption-goods，takenasawhole，haveinthissensethelongestperiodofproduction，sinceofeveryproductiveprocesstheyconstitutethelaststage。Thusifthefirstimpulsetowardstheincreaseineffectivedemandcomesfromanincreaseinconsumption，theinitialelasticityofemploymentwillbefurtherbelowitseventualequilibrium-levelthaniftheimpulsecomesfromanincreaseininvestment。Moreover，iftheincreaseddemandisdirectedtoproductswitharelativelylowelasticityofemployment，alargerproportionofitwillgotoswelltheincomesofentrepreneursandasmallerproportiontoswelltheincomesofwage-earnersandotherprime-costfactors；withthepossibleresultthattherepercussionsmaybesomewhatlessfavourabletoexpenditure，owingtothelikelihoodofentrepreneurssavingmoreoftheirincrementofincomethanwage-earnerswould。Neverthelessthedistinctionbetweenthetwocasesmustnotbeover-stated，sincealargepartofthereactionswillbemuchthesameinboth。[142]
[Page288]THE
GENERALTHEORYOFEMPLOYMENT
Howeverlongthenoticegiventoentrepreneursofaprospectivechangeindemand，itisnotpossiblefortheinitialelasticityofemployment，inresponsetoagivenincreaseofinvestment，tobeasgreatasitseventualequilibriumvalue，unlesstherearesurplusstocksandsurpluscapacityateverystageofproduction。Ontheotherhand，thedepletionofthesurplusstockswillhaveanoffsettingeffectontheamountbywhichinvestmentincreases。Ifwesupposethatthereareinitiallysomesurplusesateverypoint，theinitialelasticityofemploymentmayapproximatetounity；thenafterthestockshavebeenabsorbed，butbeforeanincreasedsupplyiscomingforwardatanadequateratefromtheearlierstagesofproduction，theelasticitywillfallaway；risingagaintowardsunityasthenewpositionofequilibriumisapproached。Thisissubject，however，tosomequalificationinsofarastherearerentfactorswhichabsorbmoreexpenditureasemploymentincreases，oriftherateofinterestincreases。
Forthesereasonsperfectstabilityofpricesisimpossibleinaneconomysubjecttochange¾；unless，indeed，thereissomepeculiarmechanismwhichensurestemporaryfluctuationsofjusttherightdegreeinthepropensitytoconsume。Butprice-instabilityarisinginthiswaydoesnotleadtothekindofprofitstimuluswhichisliabletobringintoexistenceexcesscapacity。Forthewindfallgainwillwhollyaccruetothoseentrepreneurswhohappentopossessproductsatarelativelyadvancedstageofproduction，andthereisnothingwhichtheentrepreneur，whodoesnotpossessspecialisedresourcesoftherightkind，candotoattractthisgaintohimself。Thustheinevitableprice-instabilityduetochangecannotaffecttheactionsofentrepreneurs，butmerelydirectsadefactowindfallofwealthintothelapsoftheluckyonesmutatismutandiswhenthesupposedchangeisintheotherdirection。Thisfacthas，Ithink，beenoverlookedinsomecontemporarydiscussionsofapracticalpolicyaimedatstabilisingprices。
[Page289]THE
EMPLOYMENTFUNCTION
Itistruethatinasocietyliabletochangesuchapolicycannotbeperfectlysuccessful。Butitdoesnotfollowthateverysmalltemporarydeparturefrompricestabilitynecessarilysetsupacumulativedisequilibrium。
III
Wehaveshownthatwheneffectivedemandisdeficientthereisunder-employmentoflabourinthesensethattherearemenunemployedwhowouldbewillingtoworkatlessthantheexistingrealwage。Consequently，aseffectivedemandincreases，employmentincreases，thoughatarealwageequaltoorlessthantheexistingone，untilapointcomesatwhichthereisnosurplusoflabouravailableatthethenexistingrealwage；i。e。nomoremenorhoursoflabouravailableunlessmoney-wagesrisefromthispointonwardsfasterthanprices。ThenextproblemistoconsiderwhatwillhappenIf，whenthispointhasbeenreached，expenditurestillcontinuestoincrease。
Uptothispointthedecreasingreturnfromapplyingmorelabourtoagivencapitalequipmenthasbeenoffsetbytheacquiescenceoflabourinadiminishingrealwage。Butafterthispointaunitoflabourwouldrequiretheinducementoftheequivalentofanincreasedquantityofproduct，whereastheyieldfromapplyingafurtherunitwouldbeadiminishedquantityofproduct。Theconditionsofstrictequilibriumrequire，therefore，thatwagesandprices，andconsequentlyprofitsalso，shouldallriseinthesameproportionasexpenditure，the'real'position，includingthevolumeofoutputandemployment，beingleftunchangedinallrespects。Wehavereached，thatistosay，asituationinwhichthecrudequantitytheoryofmoneyinterpreting'velocity'tomean'income-velocity'isfullysatisfied；
foroutputdoesnotalterandpricesriseinexactproportiontoMV。
Neverthelesstherearecertainpracticalqualifica-
[Page290]THE
GENERALTHEORYOFEMPLOYMENT
tionstothisconclusionwhichmustbeborneinmindinapplyingittoanactualcase:
1Foratimeatleast，risingpricesmaydeludeentrepreneursintoincreasingemploymentbeyondthelevelwhichmaximisestheirindividualprofitsmeasuredintermsoftheproduct。Fortheyaresoaccustomedtoregardrisingsale-proceedsintermsofmoneyasasignalforexpandingproduction，thattheymaycontinuetodosowhenthispolicyhasinfactceasedtobetotheirbestadvantage；i。e。theymayunderestimatetheirmarginalusercostinthenewpriceenvironment。
2Sincethatpartofhisprofitwhichtheentrepreneurhastohandontotherentierisfixedintermsofmoney，risingprices，eventhoughunaccompaniedbyanychangeinoutput，willredistributeincomestotheadvantageoftheentrepreneurandtothedisadvantageoftherentier，whichmayhaveareactiononthepropensitytoconsume。This，however，isnotaprocesswhichwillhaveonlybegunwhenfullemploymenthasbeenattained；¾；itwillhavebeenmakingsteadyprogressallthetimethattheexpenditurewasincreasing。Iftherentierislesspronetospendthantheentrepreneur，thegradualwithdrawalofrealincomefromtheormerwillmeanthatfullemploymentwillbereachedwithasmallerincreaseinthequantityofmoneyandasmallerreductionintherateofinterestthanwillbethecaseiftheoppositehypothesisholds。Afterfullemploymenthasbeenreached，afurtherriseofpriceswill，ifthefirsthypothesiscontinuestohold，meanthattherateofinterestwillhavetorisesomewhattopreventpricesfromrisingindefinitely，andthattheincreaseinthequantityofmoneywillbelessthaninproportiontotheincreaseinexpenditure；whilstifthesecondhypothesisholds，theoppositewillbethecase。Itmaybethat，astherealincomeoftherentierisdiminished，apointwillcomewhen，asaresultofhisgrowingrelativeimpoverishment，therewillbeachange-overfromthefirsthypothesistothesecond，which[Page291]THE
EMPLOYMENTFUNCTION
pointmaybereachedeitherbeforeorafterfullemploymenthasbeenattained。
IV
Thereis，perhaps，somethingalittleperplexingintheapparentasymmetrybetweeninflationanddeflation。Forwhilstadeflationofeffectivedemandbelowthelevelrequiredforfullemploymentwilldiminishemploymentaswellasprices，aninflationofitabovethislevelwillmerelyaffectprices。Thisasymmetryis，however，merelyareflectionofthefactthat，whilstlabourisalwaysinapositiontorefusetoworkonascaleinvolvingarealwagewhichislessthanthemarginaldisutilityofthatamountofemployment，itisnotinapositiontoinsistonbeingofferedworkonascaleinvolvingarealwagewhichisnotgreaterthanthemarginaldisutilityofthatamountofemployment。
TheTheoryofPrices[Page292]
Chapter21
THETHEORYOFPRICES
I
Solongaseconomistsareconcernedwithwhatiscalledthetheoryofvalue，theyhavebeenaccustomedtoteachthatpricesaregovernedbytheconditionsofsupplyanddemand；and，inparticular，changesinmarginalcostandtheelasticityofshort-periodsupplyhaveplayedaprominentpart。ButwhentheypassinvolumeII，ormoreofteninaseparatetreatise，tothetheoryofmoneyandprices，wehearnomoreofthesehomelybutintelligibleconceptsandmoveintoaworldwherepricesaregovernedbythequantityofmoney，byitsincome-velocity，bythevelocityofcirculationrelativelytothevolumeoftransactions，byhoarding，byforcedsaving，byinflationanddeflationethocgenusomne；andlittleornoattemptismadetorelatethesevaguerphrasestoourformernotionsoftheelasticitiesofsupplyanddemand。Ifwereflectonwhatwearebeingtaughtandtrytorationaliseit，inthesimplerdiscussionsitseemsthattheelasticityofsupplymusthavebecomezeroanddemandproportionaltothequantityofmoney；whilstinthemoresophisticatedwearelostinahazewherenothingisclearandeverythingispossible。Wehaveallofusbecomeusedtofindingourselvessometimesontheonesideofthemoonandsometimesontheother，withoutknowingwhatrouteorjourneyconnectsthem，related，apparently，afterthefashionofourwakingandourdreaminglives。
[Page293]THE
THEORYOFPRICES
Oneoftheobjectsoftheforegoingchaptershasbeentoescapefromthisdoublelifeandtobringthetheoryofpricesasawholebacktoclosecontactwiththetheoryofvalue。Thedivisionofeconomicsbetweenthetheoryofvalueanddistributionontheonehandandthetheoryofmoneyontheotherhandis，Ithink，afalsedivision。Therightdichotomyis，Isuggest，betweenthetheoryoftheindividualindustryorfirmandoftherewardsandthedistributionbetweendifferentusesofagivenquantityofresourcesontheonehand，andthetheoryofoutputandemploymentasawholeontheotherhand。Solongaswelimitourselvestothestudyoftheindividualindustryorfirmontheassumptionthattheaggregatequantityofemployedresourcesisconstant，and，provisionally，thattheconditionsofotherindustriesorfirmsareunchanged，itistruethatwearenotconcernedwiththesignificantcharacteristicsofmoney。Butassoonaswepasstotheproblemofwhatdeterminesoutputandemploymentasawhole，werequirethecompletetheoryofamonetaryeconomy。
Or，perhaps，wemightmakeourlineofdivisionbetweenthetheoryofstationaryequilibriumandthetheoryofshiftingequilibrium¾；meaningbythelatterthetheoryofasysteminwhichchangingviewsaboutthefuturearecapableofinfluencingthepresentsituation。Fortheimportanceofmoneyessentiallyflowsfromitsbeingalinkbetweenthepresentandthefuture。Wecanconsiderwhatdistributionofresourcesbetweendifferentuseswillbeconsistentwithequilibriumundertheinfluenceofnormaleconomicmotivesinaworldinwhichourviewsconcerningthefuturearefixedandreliableinallrespects；¾；withafurtherdivision，perhaps，betweenaneconomywhichisunchangingandonesubjecttochange，butwhereallthingsareforeseenfromthebeginning。
Orwecanpassfromthissimplifiedpropaedeutictotheproblemsoftherealworldinwhichourpreviousexpectationsareliabletodisappoint-
[Page294]THE
GENERALTHEORYOFEMPLOYMENT
mentandexpectationsconcerningthefutureaffectwhatwedoto-day。
Itiswhenwehavemadethistransitionthatthepeculiarpropertiesofmoneyasalinkbetweenthepresentandthefuturemustenterintoourcalculations。But，althoughthetheoryofshiftingequilibriummustnecessarilybepursuedintermsofamonetaryeconomy，itremainsatheoryofvalueanddistributionandnotaseparate'theoryofmoney'。Moneyinitssignificantattributesis，aboveall，asubtledeviceforlinkingthepresenttothefuture；andwecannotevenbegintodiscusstheeffectofchangingexpectationsoncurrentactivitiesexceptinmonetaryterms。Wecannotgetridofmoneyevenbyabolishinggoldandsilverandlegaltenderinstruments。Solongasthereexistsanydurableasset，itiscapableofpossessingmonetaryattributes[143]and，therefore，ofgivingrisetothecharacteristicproblemsofamonetaryeconomy。
II
Inasingleindustryitsparticularprice-leveldependspartlyontherateofremunerationofthefactorsofproductionwhichenterintoitsmarginalcost，andpartlyonthescaleofoutput。Thereisnoreasontomodifythisconclusionwhenwepasstoindustryasawhole。Thegeneralprice-leveldependspartlyontherateofremunerationofthefactorsofproductionwhichenterintomarginalcostandpartlyonthescaleofoutputasawhole，i。e。takingequipmentandtechniqueasgivenonthevolumeofemployment。Itistruethat，whenwepasstooutputasawhole，thecostsofproductioninanyindustrypartlydependontheoutputofotherindustries。Butthemoresignificantchange，ofwhichwehavetotakeaccount，istheeffectofchangesindemandbothoncostsandonvolume。
Itisonthesideofdemandthatwehavetointroducequitenewideaswhenwearedealingwithdemandasawholeand[Page295]THE
THEORYOFPRICES
nolongerwiththedemandforasingleproducttakeninisolation，withdemandasawholeassumedtobeunchanged。
III
Ifweallowourselvesthesimplificationofassumingthattheratesofremunerationofthedifferentfactorsofproductionwhichenterintomarginalcostallchangeinthesameproportion，i。e。inthesameproportionasthewage-unit，itfollowsthatthegeneralprice-leveltakingequipmentandtechniqueasgivendependspartlyonthewage-unitandpartlyonthevolumeofemployment。Hencetheeffectofchangesinthequantityofmoneyontheprice-levelcanbeconsideredasbeingcompoundedoftheeffectonthewage-unitandtheeffectonemployment。
Toelucidatetheideasinvolved，letussimplifyourassumptionsstillfurther，andassume1thatallunemployedresourcesarehomogeneousandinterchangeableintheirefficiencytoproducewhatiswanted，and2
thatthefactorsofproductionenteringintomarginalcostarecontentwiththesamemoney-wagesolongasthereisasurplusofthemunemployed。
Inthiscasewehaveconstantreturnsandarigidwage-unit，solongasthereisanyunemployment。Itfollowsthatanincreaseinthequantityofmoneywillhavenoeffectwhateveronprices，solongasthereisanyunemployment，andthatemploymentwillincreaseinexactproportiontoanyincreaseineffectivedemandbroughtaboutbytheincreaseinthequantityofmoney；whilstassoonasfullemploymentisreached，itwillthenceforwardbethewage-unitandpriceswhichwillincreaseinexactproportiontotheincreaseineffectivedemand。Thusifthereisperfectlyelasticsupplysolongasthereisunemployment，andperfectlyinelasticsupplysosoonasfullemploymentisreached，andifeffectivedemandchangesinthesameproportionasthequantityofmoney，thequantitytheoryofmoneycan[Page296]THE
GENERALTHEORYOFEMPLOYMENT
beenunciatedasfollows:'Solongasthereisunemployment，employmentwillchangeinthesameproportionasthequantityofmoney；andwhenthereisfullemployment，priceswillchangeinthesameproportionasthequantityofmoney'。
Having，however，satisfiedtraditionbyintroducingasufficientnumberofsimplifyingassumptionstoenableustoenunciateaquantitytheoryofmoney，letusnowconsiderthepossiblecomplicationswhichwillinfactinfluenceevents:
1Effectivedemandwillnotchangeinexactproportiontothequantityofmoney。
2Sinceresourcesarenothomogeneous，therewillbediminishing，andnotconstant，returnsasemploymentgraduallyincreases。
3Sinceresourcesarenotinterchangeable，somecommoditieswillreachaconditionofinelasticsupplywhilsttherearestillunemployedresourcesavailablefortheproductionofothercommodities。
4Thewage-unitwilltendtorise，beforefullemploymenthasbeenreached。
5Theremunerationsofthefactorsenteringintomarginalcostwillnotallchangeinthesameproportion。
Thuswemustfirstconsidertheeffectofchangesinthequantityofmoneyonthequantityofeffectivedemand；andtheincreaseineffectivedemandwill，generallyspeaking，spenditselfpartlyinincreasingthequantityofemploymentandpartlyinraisingthelevelofprices。Thusinsteadofconstantpricesinconditionsofunemployment，andofpricesrisinginproportiontothequantityofmoneyinconditionsoffullemployment，wehaveinfactaconditionofpricesrisinggraduallyasemploymentincreases。
Thetheoryofprices，thatistosay，theanalysisoftherelationbetweenchangesinthequantityofmoneyandchangesintheprice-levelwithaviewtodeterminingtheelasticityofpricesinresponsetochangesinthequantity[Page297]THE
THEORYOFPRICES
ofmoney，must，therefore，directitselftothefivecomplicatingfactorssetforthabove。
Wewillconsidereachoftheminturn。Butthisproceduremustnotbeallowedtoleadusintosupposingthattheyare，strictlyspeaking，independent。
Forexample，theproportion，inwhichanincreaseineffectivedemandisdividedinitseffectbetweenincreasingoutputandraisingprices，mayaffectthewayinwhichthequantityofmoneyisrelatedtothequantityofeffectivedemand。Or，again，thedifferencesintheproportions，inwhichtheremunerationsofdifferentfactorschange，mayinfluencetherelationbetweenthequantityofmoneyandthequantityofeffectivedemand。
Theobjectofouranalysisis，nottoprovideamachine，ormethodofblindmanipulation，whichwillfurnishaninfallibleanswer，buttoprovideourselveswithanorganisedandorderlymethodofthinkingoutparticularproblems；
and，afterwehavereachedaprovisionalconclusionbyisolatingthecomplicatingfactorsonebyone，wethenhavetogobackonourselvesandallow，aswellaswecan，fortheprobableinteractionsofthefactorsamongstthemselves。
Thisisthenatureofeconomicthinking。Anyotherwayofapplyingourformalprinciplesofthoughtwithoutwhich，however，weshallbelostinthewoodwillleadusintoerror。Itisagreatfaultofsymbolicpseudo-mathematicalmethodsofformalisingasystemofeconomicanalysis，suchasweshallsetdowninsectionviofthischapter，thattheyexpresslyassumestrictindependencebetweenthefactorsinvolvedandlosealltheircogencyandauthorityifthishypothesisisdisallowed；whereas，inordinarydiscourse，wherewearenotblindlymanipulatingbutknowallthetimewhatwearedoingandwhatthewordsmean，wecankeep'atthebackofourheads'thenecessaryreservesandqualificationsandtheadjustmentswhichweshallhavetomakelateron，inawayinwhichwecannotkeepcomplicatedpartialdiffer-
[Page298]THE
GENERALTHEORYOFEMPLOYMENT
entials'attheback'ofseveralpagesofalgebrawhichassumethattheyallvanish。Toolargeaproportionofrecent'mathematical'economicsaremerelyconcoctions，asimpreciseastheinitialassumptionstheyreston，whichallowtheauthortolosesightofthecomplexitiesandinterdependenciesoftherealworldinamazeofpretentiousandunhelpfulsymbols。
IV
1Theprimaryeffectofachangeinthequantityofmoneyonthequantityofeffectivedemandisthroughitsinfluenceontherateofinterest。Ifthisweretheonlyreaction，thequantitativeeffectcouldbederivedfromthethreeelements¾；a
thescheduleofliquidity-preferencewhichtellsusbyhowmuchtherateofinterestwillhavetofallinorderthatthenewmoneymaybeabsorbedbywillingholders，bthescheduleofmarginalefficiencieswhichtellsusbyhowmuchagivenfallintherateofinterestwillincreaseinvestment，andctheinvestmentmultiplierwhichtellsusbyhowmuchagivenincreaseininvestmentwillincreaseeffectivedemandasawhole。
Butthisanalysis，thoughitisvaluableinintroducingorderandmethodintoourenquiry，presentsadeceptivesimplicity，ifweforgetthatthethreeelementsa，bandcarethemselvespartlydependentonthecomplicatingfactors2，3，4and5whichwehavenotyetconsidered。Forthescheduleofliquidity-preferenceitselfdependsonhowmuchofthenewmoneyisabsorbedintotheincomeandindustrialcirculations，whichdependsinturnonhowmucheffectivedemandincreasesandhowtheincreaseisdividedbetweentheriseofprices，theriseofwages，andthevolumeofoutputandemployment。Furthermore，thescheduleofmarginalefficiencieswillpartlydependontheeffectwhichthecircumstancesattendantontheincreaseinthequantityofmoneyhaveonexpectationsofthefuturemonetary[Page299]THE
THEORYOFPRICES
prospects。Andfinallythemultiplierwillbeinfluencedbythewayinwhichthenewincomeresultingfromtheincreasedeffectivedemandisdistributedbetweendifferentclassesofconsumers。Nor，ofcourse，isthislistofpossibleinteractionscomplete。Nevertheless，ifwehaveallthefactsbeforeus，weshallhaveenoughsimultaneousequationstogiveusadeterminateresult。Therewillbeadeterminateamountofincreaseinthequantityofeffectivedemandwhich，aftertakingeverythingintoaccount，willcorrespondto，andbeinequilibriumwith，theincreaseinthequantityofmoney。Moreover，itisonlyinhighlyexceptionalcircumstancesthatanincreaseinthequantityofmoneywillbeassociatedwithadecreaseinthequantityofeffectivedemand。
Theratiobetweenthequantityofeffectivedemandandthequantityofmoneycloselycorrespondstowhatisoftencalledthe'income-velocityofmoney'；¾；exceptthateffectivedemandcorrespondstotheincometheexpectationofwhichhassetproductionmoving，nottotheactuallyrealisedincome，andtogross，notnet，income。Butthe'income-velocityofmoney'is，initself，merelyanamewhichexplainsnothing。Thereisnoreasontoexpectthatitwillbeconstant。Foritdepends，astheforegoingdiscussionhasshown，onmanycomplexandvariablefactors。Theuseofthistermobscures，Ithink，therealcharacterofthecausation，andhasledtonothingbutconfusion。
2Aswehaveshownabovep。42，thedistinctionbetweendiminishingandconstantreturnspartlydependsonwhetherworkersareremuneratedinstrictproportiontotheirefficiency。Ifso，weshallhaveconstantlabour-costsintermsofthewage-unitwhenemploymentincreases。Butifthewageofagivengradeoflabourersisuniformirrespectiveoftheefficiencyoftheindividuals，weshallhaverisinglabour-costs，irrespectiveoftheefficiencyoftheequipment。Moreover，ifequipmentisnon-homogeneousandsomepartofitinvolvesagreaterprimecostper[Page300]THE
GENERALTHEORYOFEMPLOYMENT
unitofoutput，weshallhaveincreasingmarginalprimecostsoverandaboveanyincreaseduetoincreasinglabour-costs。
Hence，ingeneral，supplypricewillincreaseasoutputfromagivenequipmentisincreased。Thusincreasingoutputwillbeassociatedwithrisingprices，apartfromanychangeinthewage-unit。
3Under2wehavebeencontemplatingthepossibilityofsupplybeingimperfectlyelastic。Ifthereisaperfectbalanceintherespectivequantitiesofspecialisedunemployedresources，thepointoffullemploymentwillbereachedforallofthemsimultaneously。But，ingeneral，thedemandforsomeservicesandcommoditieswillreachalevelbeyondwhichtheirsupplyis，forthetimebeing，perfectlyinelastic，whilstinotherdirectionsthereisstillasubstantialsurplusofresourceswithoutemployment。Thusasoutputincreases，aseriesof'bottle-necks'willbesuccessivelyreached，wherethesupplyofparticularcommoditiesceasestobeelasticandtheirpriceshavetorisetowhateverlevelisnecessarytodivertdemandintootherdirections。
Itisprobablethatthegenerallevelofpriceswillnotriseverymuchasoutputincreases，solongasthereareavailableefficientunemployedresourcesofeverytype。Butassoonasoutputhasincreasedsufficientlytobegintoreachthe'bottle-necks'，thereislikelytobeasharpriseinthepricesofcertaincommodities。
Underthisheading，however，asalsounderheading2，theelasticityofsupplypartlydependsontheelapseoftime。Ifweassumeasufficientintervalforthequantityofequipmentitselftochange，theelasticitiesofsupplywillbedecidedlygreatereventually。Thusamoderatechangeineffectivedemand，comingonasituationwherethereiswidespreadunemployment，mayspenditselfverylittleinraisingpricesandmainlyinincreasingemployment；whilstalargerchange，which，beingunforeseen，causessometemporary[Page301]THE
THEORYOFPRICES
'bottle-necks'tobereached，willspenditselfinraisingprices，asdistinctfromemployment，toagreaterextentatfirstthansubsequently。
4Thatthewage-unitmaytendtorisebeforefullemploymenthasbeenreached，requireslittlecommentorexplanation。Sinceeachgroupofworkerswillgain，cet。par。，byariseinitsownwages，thereisnaturallyforallgroupsapressureinthisdirection，whichentrepreneurswillbemorereadytomeetwhentheyaredoingbetterbusiness。Forthisreasonaproportionofanyincreaseineffectivedemandislikelytobeabsorbedinsatisfyingtheupwardtendencyofthewage-unit。
Thus，inadditiontothefinalcriticalpointoffullemploymentatwhichmoney-wageshavetorise，inresponsetoanincreasingeffectivedemandintermsofmoney，fullyinproportiontotheriseinthepricesofwage-goods，wehaveasuccessionofearliersemi-criticalpointsatwhichanincreasingeffectivedemandtendstoraisemoney-wagesthoughnotfullyinproportiontotheriseinthepriceofwage-goods；andsimilarlyinthecaseofadecreasingeffectivedemand。Inactualexperiencethewage-unitdoesnotchangecontinuouslyintermsofmoneyinresponsetoeverysmallchangeineffectivedemand；butdiscontinuously。Thesepointsofdiscontinuityaredeterminedbythepsychologyoftheworkersandbythepoliciesofemployersandtradeunions。Inanopensystem，wheretheymeanachangerelativelytowage-costselsewhere，andinatradecycle，whereeveninaclosedsystemtheymaymeanachangerelativelytoexpectedwage-costsinthefuture，theycanbeofconsiderablepracticalsignificance。
Thesepoints，whereafurtherincreaseineffectivedemandintermsofmoneyisliabletocauseadiscontinuousriseinthewage-unit，mightbedeemed，fromacertainpointofview，tobepositionsofsemi-inflation，havingsomeanalogythoughaveryimperfectonetotheabsoluteinflationcf。p。303below
[Page302]THE
GENERALTHEORYOFEMPLOYMENT
whichensuesonanincreaseineffectivedemandincircumstancesoffullemployment。Theyhave，moreover，agooddealofhistoricalimportance。
Buttheydonotreadilylendthemselvestotheoreticalgeneralisations。
5Ourfirstsimplificationconsistedinassumingthattheremunerationsofthevariousfactorsenteringintomarginalcostallchangeinthesameproportion。Butinfacttheratesofremunerationofdifferentfactorsintermsofmoneywillshowvaryingdegreesofrigidityandtheymayalsohavedifferentelasticitiesofsupplyinresponsetochangesinthemoney-rewardsoffered。Ifitwerenotforthis，wecouldsaythattheprice-leveliscompoundedoftwofactors，thewage-unitandthequantityofemployment。
Perhapsthemostimportantelementinmarginalcostwhichislikelytochangeinadifferentproportionfromthewage-unit，andalsotofluctuatewithinmuchwiderlimits，ismarginalusercost。Formarginalusercostmayincreasesharplywhenemploymentbeginstoimprove，ifaswillprobablybethecasetheincreasingeffectivedemandbringsarapidchangeintheprevailingexpectationastothedatewhenthereplacementofequipmentwillbenecessary。
Whilstitisformanypurposesaveryusefulfirstapproximationtoassumethattherewardsofallthefactorsenteringintomarginalprime-costchangeinthesameproportionasthewage-unit，itmightbebetter，perhaps，totakeaweightedaverageoftherewardsofthefactorsenteringintomarginalprime-cost，andcallthisthecost-unit。Thecost-unit，or，subjecttotheaboveapproximation，thewage-unit，canthusberegardedastheessentialstandardofvalue；andtheprice-level，giventhestateoftechniqueandequipment，willdependpartlyonthecost-unit，andpartlyonthescaleofoutput，increasing，whereoutputincreases，morethaninproportiontoanyincreaseinthecost-unit，inaccordancewiththeprincipleofdiminishingreturnsintheshort[Page303]THE
THEORYOFPRICES
period。Wehavefullemploymentwhenoutputhasrisentoalevelatwhichthemarginalreturnfromarepresentativeunitofthefactorsofproductionhasfallentotheminimumfigureatwhichaquantityofthefactorssufficienttoproducethisoutputisavailable。
V
Whenafurtherincreaseinthequantityofeffectivedemandproducesnofurtherincreaseinoutputandentirelyspendsitselfonanincreaseinthecost-unitfullyproportionatetotheincreaseineffectivedemand，wehavereachedaconditionwhichmightbeappropriatelydesignatedasoneoftrueinflation。Uptothispointtheeffectofmonetaryexpansionisentirelyaquestionofdegree，andthereisnopreviouspointatwhichwecandrawadefinitelineanddeclarethatconditionsofinflationhavesetin。Everypreviousincreaseinthequantityofmoneyislikely，insofarasitincreaseseffectivedemand，tospenditselfpartlyinincreasingthecost-unitandpartlyinincreasingoutput。
Itappears，therefore，thatwehaveasortofasymmetryonthetwosidesofthecriticallevelabovewhichtrueinflationsetsin。Foracontractionofeffectivedemandbelowthecriticallevelwillreduceitsamountmeasuredincost-units；whereasanexpansionofeffectivedemandbeyondthislevelwillnot，ingeneral，havetheeffectofincreasingitsamountintermsofcost-units。Thisresultfollowsfromtheassumptionthatthefactorsofproduction，andinparticulartheworkers，aredisposedtoresistareductionintheirmoney-rewards，andthatthereisnocorrespondingmotivetoresistanincrease。Thisassumptionis，however，obviouslywellfoundedinthefacts，duetothecircumstancethatachange，whichisnotanall-roundchange，isbeneficialtothespecialfactorsaffectedwhenitisupwardandharmfulwhenitisdownward。
If，onthecontrary，money-wagesweretofallwith-
[Page304]THE
GENERALTHEORYOFEMPLOYMENT
outlimitwhenevertherewasatendencyforlessthanfullemployment，theasymmetrywould，indeed，disappear。Butinthatcasetherewouldbenoresting-placebelowfullemploymentuntileithertherateofinterestwasincapableoffallingfurtherorwageswerezero。Infactwemusthavesomefactor，thevalueofwhichintermsofmoneyis，ifnotfixed，atleaststicky，togiveusanystabilityofvaluesinamonetarysystem。
Theviewthatanyincreaseinthequantityofmoneyisinflationaryunlesswemeanbyinflationarymerelythatpricesarerisingisboundupwiththeunderlyingassumptionoftheclassicaltheorythatwearealwaysinaconditionwhereareductionintherealrewardsofthefactorsofproductionwillleadtoacurtailmentintheirsupply。
VI
Withtheaidofthenotationintroducedinchapter20wecan，ifwewish，expressthesubstanceoftheaboveinsymbolicform。
LetuswriteMV=DwhereMisthequantityofmoney，Vitsincome-velocitythisdefinitiondifferingintheminorrespectsindicatedabovefromtheusualdefinitionandDtheeffectivedemand。If，then，Visconstant，priceswillchangeinthesameproportionasthequantityofmoneyprovidedthatep=Dpd/pdDisunity。Thisconditionissatisfiedseep。286aboveifeo=0orifew=1。Theconditionew=1meansthatthewage-unitintermsofmoneyrisesinthesameproportionastheeffectivedemand，sinceew=DdW/WdDandtheconditioneo=0meansthatoutputnolongershowsanyresponsetoafurtherincreaseineffectivedemand，sinceeo=DdO/OdD。Outputineithercasewillbeunaltered。Next，wecandealwiththecasewhereincome-velocityisnotconstant，byintroducingyetafurther[Page305]THE
THEORYOFPRICES
elasticity，namelytheelasticityofeffectivedemandinresponsetochangesinthequantityofmoney，MdD
ed=¾；¾；¾；¾；
DdMThisgivesusMdp¾；¾；¾；¾；=ep×；edwhereep=1-ee×；eo1-ew；
pdMsothate=ed-1-ewed×；eeeo=ed1-eeeoeeeo×；ewwhereewithoutsuffix=Mdp/pdM
standsfortheapexofthispyramidandmeasurestheresponseofmoney-pricestochangesinthequantityofmoney。
Sincethislastexpressiongivesustheproportionatechangeinpricesinresponsetoachangeinthequantityofmoney，itcanberegardedasageneralisedstatementofthequantitytheoryofmoney。Idonotmyselfattachmuchvaluetomanipulationsofthiskind；andIwouldrepeatthewarning，whichIhavegivenabove，thattheyinvolvejustasmuchtacitassumptionastowhatvariablesaretakenasindependentpartialdifferentialsbeingignoredthroughoutasdoesordinarydiscourse，whilstIdoubtiftheycarryusanyfurtherthanordinarydiscoursecan。Perhapsthebestpurposeservedbywritingthemdownistoexhibittheextremecomplexityoftherelationshipbetweenpricesandthequantityofmoney，whenweattempttoexpressitinaformalmanner。Itis，however，worthpointingoutthat，ofthefourtermsed，ew，eeandeouponwhichtheeffectonpricesofchangesinthequantityofmoneydepends，edstandsfortheliquidityfactorswhichdeterminethedemandformoneyineachsituation，ewforthelabourfactorsor，morestrictly，thefactorsenteringintoprime-cost
whichdeterminetheextenttowhichmoney-wagesareraisedasemploymentincreases，andeeandeoforthephysicalfactorswhichdeterminethe[Page306]THE
GENERALTHEORYOFEMPLOYMENT
rateofdecreasingreturnsasmoreemploymentisappliedtotheexistingequipment。
Ifthepublicholdaconstantproportionoftheirincomeinmoney，ed=1；ifmoney-wagesarefixed，ew=
0；ifthereareconstantreturnsthroughoutsothatmarginalreturnequalsaveragereturn，eeeo=
1；andifthereisfullemploymenteitheroflabourorofequipment，eeeo=0。
Nowe=1，ifed=
1，andew=1；orifed=1，ew=0andee×；eo=0；orifed=1andeo=0。Andobviouslythereisavarietyofotherspecialeasesinwhiche=1。
Butingeneraleisnotunity；anditis，perhaps，safetomakethegeneralisationthatonplausibleassumptionsrelatingtotherealworld，andexcludingthecaseofa'flightfromthecurrency'inwhichedandewbecomelarge，eis，asarule，lessthanunity。