THEORYOFEMPLOYMENT
changesinmoney-wagesandchangesinrealwages。Inthecaseofachangepeculiartoaparticularindustryonewouldexpectthechangeinrealwagestobeinthesamedirectionasthechangeinmoney-wages。Butinthecaseofchangesinthegenerallevelofwages，itwillbefound，Ithink，thatthechangeinrealwagesassociatedwithachangeinmoney-wages，sofarfrombeingusuallyinthesamedirection，isalmostalwaysintheoppositedirection。Whenmoney-wagesarerising，thatistosay，itwillbefoundthatrealwagesarefalling；andwhenmoney-wagesarefalling，realwagesarerising。Thisisbecause，intheshortperiod，fallingmoney-wagesandrisingrealwagesareeach，forindependentreasons，likelytoaccompanydecreasingemployment；labourbeingreadiertoacceptwage-cutswhenemploymentisfallingoff，yetrealwagesinevitablyrisinginthesamecircumstancesonaccountoftheincreasingmarginalreturntoagivencapitalequipmentwhenoutputisdiminished。
If，indeed，itweretruethattheexistingrealwageisaminimumbelowwhichmorelabourthanisnowemployedwillnotbeforthcominginanycircumstances，involuntaryunemployment，apartfromfrictionalunemployment，wouldbenon-existent。Buttosupposethatthisisinvariablythecasewouldbeabsurd。Formorelabourthanisatpresentemployedisusuallyavailableattheexistingmoney-wage，eventhoughthepriceofwage-goodsisrisingand，consequently，therealwagefalling。Ifthisistrue，thewage-goodsequivalentoftheexistingmoney-wageisnotanaccurateindicationofthemarginaldisutilityoflabour，andthesecondpostulatedoesnotholdgood。
Butthereisamorefundamentalobjection。Thesecondpostulateflowsfromtheideathattherealwagesoflabourdependonthewagebargainswhichlabourmakeswiththeentrepreneurs。Itisadmitted，ofcourse，thatthebargainsareactuallymadeintermsofmoney，andeventhattherealwagesacceptabletolabourare[Page11]POSTULATES
OFTHECLASSICALECONOMICS
notaltogetherindependentofwhatthecorrespondingmoney-wagehappenstobe。Neverthelessitisthemoney-wagethusarrivedatwhichisheldtodeterminetherealwage。Thustheclassicaltheoryassumesthatitisalwaysopentolabourtoreduceitsrealwagebyacceptingareductioninitsmoney-wage。Thepostulatethatthereisatendencyfortherealwagetocometoequalitywiththemarginaldisutilityoflabourclearlypresumesthatlabouritselfisinapositiontodecidetherealwageforwhichitworks，thoughnotthequantityofemploymentforthcomingatthiswage。
Thetraditionaltheorymaintains，inshort，thatthewagebargainsbetweentheentrepreneursandtheworkersdeterminetherealwage；
sothat，assumingfreecompetitionamongstemployersandnorestrictivecombinationamongstworkers，thelattercan，iftheywish，bringtheirrealwagesintoconformitywiththemarginaldisutilityoftheamountofemploymentofferedbytheemployersatthatwage。Ifthisisnottrue，thenthereisnolongeranyreasontoexpectatendencytowardsequalitybetweentherealwageandthemarginaldisutilityoflabour。
Theclassicalconclusionsareintended，itmustberemembered，toapplytothewholebodyoflabouranddonotmeanmerelythatasingleindividualcangetemploymentbyacceptingacutinmoney-wageswhichhisfellowsrefuse。Theyaresupposedtobeequallyapplicabletoaclosedsystemastoanopensystem，andarenotdependentonthecharacteristicsofanopensystemorontheeffectsofareductionofmoney-wagesinasinglecountryonitsforeigntrade，whichlie，ofcourse，entirelyoutsidethefieldofthisdiscussion。Noraretheybasedonindirecteffectsduetoalowerwages-billintermsofmoneyhavingcertainreactionsonthebankingsystemandthestateofcredit，effectswhichweshallexamineindetailinchapter19。Theyarebasedonthebeliefthatinaclosedsystemareduction[Page12]THEGENERAL
THEORYOFEMPLOYMENT
inthegenerallevelofmoney-wageswillbeaccompanied，atanyrateintheshortperiodandsubjectonlytominorqualifications，bysome，thoughnotalwaysaproportionate，reductioninrealwages。
Nowtheassumptionthatthegenerallevelofrealwagesdependsonthemoney-wagebargainsbetweentheemployersandtheworkersisnotobviouslytrue。Indeeditisstrangethatsolittleattemptshouldhavebeenmadetoproveortorefuteit。Foritisfarfrombeingconsistentwiththegeneraltenoroftheclassicaltheory，whichhastaughtustobelievethatpricesaregovernedbymarginalprimecostintermsofmoneyandthatmoney-wageslargelygovernmarginalprimecost。Thusifmoney-wageschange，onewouldhaveexpectedtheclassicalschooltoarguethatpriceswouldchangeinalmostthesameproportion，leavingtherealwageandthelevelofunemploymentpracticallythesameasbefore，anysmallgainorlosstolabourbeingattheexpenseorprofitofotherelementsofmarginalcostwhichhavebeenleftunaltered。[23]Theyseem，however，tohavebeendivertedfromthislineofthought，partlybythesettledconvictionthatlabourisinapositiontodetermineitsownrealwageandpartly，perhaps，bypreoccupationwiththeideathatpricesdependonthequantityofmoney。Andthebeliefinthepropositionthatlabourisalwaysinapositiontodetermineitsownrealwage，onceadopted，hasbeenina~ntainedbyitsbeingconfusedwiththepropositionthatlabourisalwaysinapositiontodeterminewhatrealwageshallcorrespondtofullemployment，i。e。themaximumquantityofemploymentwhichiscompatiblewithagivenrealwage。
Tosumup:therearetwoobjectionstothesecondpostulateoftheclassicaltheory。Thefirstrelatestotheactualbehaviouroflabour。Afallinrealwagesdue[Page13]POSTULATES
OFTHECLASSICALECONOMICS
toariseinprices，withmoney-wagesunaltered，doesnot，asarule，causethesupplyofavailablelabouronofferatthecurrentwagetofallbelowtheamountactuallyemployedpriortotheriseofprices。Tosthatitdoesistosupposethatallthosewhoarenowunemployedthoughwillingtoworkatthecurrentwagewillwithdrawtheofferoftheirlabourintheeventofevenasmallriseinthecostofliving。YetthisstrangesuppositionapparentlyunderliesProfessorPigou'sTheoryofUnemployment，[24]anditiswhatallmembersoftheorthodoxschoolaretacitlyassuming。
Buttheother，morefundamental，objection，whichweshalldevelopintheensuingchapters，flowsfromourdisputingtheassumptionthatthegenerallevelofrealwagesisdirectlydeterminedbythecharacterofthewagebargain。Inassumingthatthewagebargaindeterminestherealwagetheclassicalschoolhavesliptinanillicitassumption。Fortheremaybenomethodavailabletolabourasawholewherebyitcanbringthewage-goodsequivalentofthegenerallevelofmoneywagesintoconformitywiththemarginaldisutilityofthecurrentvolumeofemployment。Theremayexistnoexpedientbywhichlabourasawholecanreduceitsrealwagetoagivenfigurebymakingrevisedmoneybargainswiththeentrepreneurs。
Thiswillbeourcontention。Weshallendeavourtoshowthatprimarilyitiscertainotherforceswhichdeterminethegenerallevelofrealwages。
Theattempttoelucidatethisproblemwillbeoneofourmainthemes。Weshallarguethattherehasbeenafundamentalmisunderstandingofhowinthisrespecttheeconomyinwhichweliveactuallyworks。
III
Thoughthestruggleovermoney-wagesbetweenindividualsandgroupsisoftenbelievedtodetermine[Page14]THEGENERAL
THEORYOFEMPLOYMENT
thegenerallevelofreal-wages，itis，infact，concernedwithadifferentobject。Sincethereisimperfectmobilityoflabour，andwagesdonottendtoanexactequalityofnetadvantageindifferentoccupations，anyindividualorgroupofindividuals，whoconsenttoareductionofmoney-wagesrelativelytoothers，willsufferarelativereductioninrealwages，whichisasufficientjustificationforthemtoresistit。Ontheotherhanditwouldbeimpracticabletoresisteveryreductionofrealwages，duetoachangeinthepurchasing-powerofmoneywhichaffectsallworkersalike；andinfactreductionsofrealwagesarisinginthiswayarenot，asarule，resistedunlesstheyproceedtoanextremedegree。Moreover，aresistancetoreductionsinmoney-wagesapplyingtoparticularindustriesdoesnotraisethesameinsuperablebartoanincreaseinaggregateemploymentwhichwouldresultfromasimilarresistancetoeveryreductioninrealwages。
Inotherwords，thestruggleaboutmoney-wagesprimarilyaffectsthedistributionoftheaggregaterealwagebetweendifferentlabour-groups，andnotitsaverageamountperunitofemployment，whichdepends，asweshallsee，onadifferentsetofforces。Theeffectofcombinationonthepartofagroupofworkersistoprotecttheirrelativerealwage。Thegenerallevelofrealwagesdependsontheotherforcesoftheeconomicsystem。
Thusitisfortunatethattheworkers，thoughunconsciously，areinstinctivelymorereasonableeconomiststhantheclassicalschool，inasmuchastheyresistreductionsofmoney-wages，whichareseldomorneverofanall-roundcharacter，eventhoughtheexistingrealequivalentofthesewagesexceedsthemarginaldisutilityoftheexistingemployment；whereastheydonotresistreductionsofrealwages，whichareassociatedwithincreasesinaggregateemploymentandleaverelativemoney-wagesunchanged，unlessthereductionproceedssofarastothreatenareductionofthereal[Page15]POSTULATES
OFTHECLASSICALECONOMICS
wagebelowthemarginaldisutilityoftheexistingvolumeofemployment。
Everytradeunionwillputupsomeresistancetoacutinmoney-wages，howeversmall。Butsincenotradeunionwoulddreamofstrikingoneveryoccasionofariseinthecostofliving，theydonotraisetheobstacletoanyincreaseinaggregateemploymentwhichisattributedtothembytheclassicalschool。
IV
Wemustnowdefinethethirdcategoryofunemployment，namely'involuntary'
unemploymentinthestrictsense，thepossibilityofwhichtheclassicaltheorydoesnotadmit。
Clearlywedonotmeanby'involuntary'unemploymentthemereexistenceofanunexhaustedcapacitytowork。Aneight-hourdaydoesnotconstituteunemploymentbecauseitisnotbeyondhumancapacitytoworktenhours。
Norshouldweregardas'involuntary'unemploymentthewithdrawaloftheirlabourbyabodyofworkersbecausetheydonotchoosetoworkforlessthanacertainrealreward。Furthermore，itwillbeconvenienttoexclude'frictional'unemploymentfromourdefinitionof'involuntary'unemployment。
Mydefinitionis，therefore，asfollows:MenareinvoluntarilyunemployedIf，intheeventofasmallriseinthepriceofwage-goodsrelativelytothemoney-wage，boththeaggregatesupplyoflabourwillingtoworkforthecurrentmoney-wageandtheaggregatedemandforitatthatwagewouldbegreaterthantheexistingvolumeofemployment。Analternativedefinition，whichamounts，however，tothesamething，willbegiveninthenextchapterp。26below。
Itfollowsfromthisdefinitionthattheequalityoftherealwagetothemarginaldisutilityofemploymentpresupposedbythesecondpostulate，realisticallyinterpreted，correspondstotheabsenceof'involuntary'
unemployment。Thisstateofaffairsweshalldescribe[Page16]THEGENERAL
THEORYOFEMPLOYMENT
as'full'employment，both'frictional'and'voluntary'unemploymentbeingconsistentwith'full'employmentthusdefined。Thisfitsin，weshallfind，withothercharacteristicsoftheclassicaltheory，whichisbestregardedasatheoryofdistributioninconditionsoffullemployment。
Solongastheclassicalpostulatesholdgood，unemployment，whichisintheabovesenseinvoluntary，cannotoccur。Apparentunemploymentmust，therefore，betheresulteitheroftemporarylossofworkofthe'betweenjobs'typeorofintermittentdemandforhighlyspecialisedresourcesoroftheeffectofatradeunion'closedshop'ontheemploymentoffreelabour。Thuswritersintheclassicaltradition，overlookingthespecialassumptionunderlyingtheirtheory，havebeendriveninevitablytotheconclusion，perfectlylogicalontheirassumption，thatapparentunemploymentapartfromtheadmittedexceptionsmustbedueatbottomtoarefusalbytheunemployedfactorstoacceptarewardwhichcorrespondstotheirmarginalproductivity。Aclassicaleconomistmaysympathisewithlabourinrefusingtoacceptacutinitsmoney-wage，andhewilladmitthatitmaynotbewisetomakeittomeetconditionswhicharetemporary；butscientificintegrityforceshimtodeclarethatthisrefusalis，nevertheless，atthebottomofthetrouble。
Obviously，however，iftheclassicaltheoryisonlyapplicabletothecaseoffullemployment，itisfallacioustoapplyittotheproblemsofinvoluntaryunemployment¾；iftherebesuchathingandwhowilldenyit？。TheclassicaltheoristsresembleEuclideangeometersinanon-Euclideanworldwho，discoveringthatinexperiencestraightlinesapparentlyparalleloftenmeet，rebukethelinesfornotkeepingstraight¾；astheonlyremedyfortheunfortunatecollisionswhichareoccurring。Yet，intruth，thereisnoremedyexcepttothrowovertheaxiomofparallelsandtoworkoutanon-Euclideangeometry。Somethingsimilarisrequiredto-dayineconomics。
Weneedtothrowover[Page17]POSTULATES
OFTHECLASSICALECONOMICS
thesecondpostulateoftheclassicaldoctrineandtoworkoutthebehaviourofasysteminwhichinvoluntaryunemploymentinthestrictsenseispossible。
V
Inemphasisingourpointofdeparturefromtheclassicalsystem，wemustnotoverlookanimportantpointofagreement。Forweshallmaintainthefirstpostulateasheretofore，subjectonlytothesamequalificationsasintheclassicaltheory；andwemustpause，foramoment，toconsiderwhatthisinvolves。
Itmeansthat，withagivenorganisation，equipmentandtechnique，realwagesandthevolumeofoutputandhenceofemploymentareuniquelycorrelated，sothat，ingeneral，anincreaseinemploymentcanonlyoccurtotheaccompanimentofadeclineintherateofrealwages。ThusIamnotdisputingthisvitalfactwhichtheclassicaleconomistshaverightlyassertedasindefeasible。
Inagivenstateoforganisation，equipmentandtechnique，therealwageearnedbyaunitoflabourhasauniqueinversecorrelationwiththevolumeofemployment。Thusifemploymentincreases，then，intheshortperiod，therewardperunitoflabourintermsofwage-goodsmust，ingeneral，declineandprofitsincrease。[25]Thisissimplytheobverseofthefamiliarpropositionthatindustryisnormallyworkingsubjecttodecreasingreturnsintheshortperiodduringwhichequipmentetc。isassumedtobeconstant；sothatthemarginalproductinthewage-goodindustrieswhichgovernsrealwagesneces-
[Page18]THEGENERAL
THEORYOFEMPLOYMENT
sarilydiminishesasemploymentisincreased。Solong，indeed，asthispropositionholds，anymeansofincreasingemploymentmustleadatthesametimetoadiminutionofthemarginalproductandhenceoftherateofwagesmeasuredintermsofthisproduct。
Butwhenwehavethrownoverthesecondpostulate，adeclineinemployment，althoughnecessarilyassociatedwithlabour'sreceivingawageequalinvaluetoalargerquantityofwage-goods，isnotnecessarilyduetolabour'sdemandingalargerquantityofwage-goods；andawillingnessonthepartoflabourtoacceptlowermoney-wagesisnotnecessarilyaremedyforunemployment。
Thetheoryofwagesinrelationtoemployment，towhichwearehereleadingup，cannotbefullyelucidated，however，untilchapter19anditsAppendixhavebeenreached。
VI
FromthetimeofSayandRicardotheclassicaleconomistshavetaughtthatsupplycreatesitsowndemand；¾；meaningbythisinsomesignificant，butnotclearlydefined，sensethatthewholeofthecostsofproductionmustnecessarilybespentintheaggregate，directlyorindirectly，onpurchasingtheproduct。
InJ。S。Mill'sPrinciplesofPoliticalEconomythedoctrineisexpresslysetforth:Whatconstitutesthemeansofpaymentforcommoditiesissimplycommodities。Eachperson'smeansofpayingfortheproductionsofotherpeopleconsistofthosewhichhehimselfpossesses。Allsellersareinevitably，andbythemeaningoftheword，buyers。Couldwesuddenlydoubletheproductivepowersofthecountry，weshoulddoublethesupplyofcommoditiesineverymarket；butweshould，bythesamestroke，doublethepurchasingpower。
Everybodywouldbringadoubledemandaswellassupply；everybodywouldbeabletobuytwiceasmuch，becauseeveryonewouldhavetwiceasmuchtoofferinexchange。[26]
[Page19]POSTULATES
OFTHECLASSICALECONOMICS
Asacorollaryofthesamedoctrine，ithasbeensupposedthatanyindividualactofabstainingfromconsumptionnecessarilyleadsto，andamountstothesamethingas，causingthelabourandcommoditiesthusreleasedfromsupplyingconsumptiontobeinvestedintheproductionofcapitalwealth。
ThefollowingpassagefromMarshall'sPureTheoryofDomesticValues[27]illustratesthetraditionalapproach:Thewholeofaman'sincomeisexpendedinthepurchaseofservicesandofcommodities。Itisindeedcommonlysaidthatamanspendssomeportionofhisincomeandsavesanother。Butitisafamiliareconomicaxiomthatamanpurchaseslabourandcommoditieswiththatportionofhisincomewhichhesavesjustasmuchashedoeswiththatheissaidtospend。Heissaidtospendwhenheseekstoobtainpresentenjoymentfromtheservicesandcommoditieswhichhepurchases。Heissaidtosavewhenhecausesthelabourandthecommoditieswhichhepurchasestobedevotedtotheproductionofwealthfromwhichheexpectstoderivethemeansofenjoymentinthefuture。ItistruethatitwouldnotbeeasytoquotecomparablepassagesfromMarshall'slaterwork[28]orfromEdgeworthorProfessorPigou。Thedoctrineisneverstatedto-dayinthiscrudeform。Neverthelessitstillunderliesthewholeclassicaltheory，whichwouldcollapsewithoutit。
Contemporaryeconomists，whomighthesitatetoagreewithMill，donothesitatetoacceptconclusionswhichrequireMill'sdoctrineastheirpremiss。
Theconviction，whichruns，forexample，throughalmostallProfessorPigou'swork，thatmoneymakesnorealdifferenceexceptfrictionallyandthatthetheoryofproductionandemploymentcanbe[Page20]THEGENERAL
THEORYOFEMPLOYMENT
workedoutlikeMill'sasbeingbasedon'real'exchangeswithmoneyintroducedperfunctorilyinalaterchapter，isthemodernversionoftheclassicaltradition。Contemporarythoughtisstilldeeplysteepedinthenotionthatifpeopledonotspendtheirmoneyinonewaytheywillspenditinanother。[29]Post-wareconomistsseldom，indeed，succeedinmaintainingthisstandpointconsistently；fortheirthoughtto-dayistoomuchpermeatedwiththecontrarytendencyandwithfactsofexperiencetooobviouslyinconsistentwiththeirformerview。[30]Buttheyhavenotdrawnsufficientlyfar-reachingconsequences；andhavenotrevisedtheirfundamentaltheory。
Inthefirstinstance，theseconclusionsmayhavebeenappliedtothekindofeconomyinwhichweactuallylivebyfalseanalogyfromsomekindofnon-exchangeRobinsonCrusoeeconomy，inwhichtheincomewhichindividualsconsumeorretainasaresultoftheirproductiveactivityis，actuallyandexclusively，theoutputinspecieofthatactivity。But，apartfromthis，theconclusionthatthecostsofoutputarealwayscoveredintheaggregatebythesale-proceedsresultingfromdemand，hasgreatplausibility，becauseitisdifficulttodistinguishitfromanother，similar-lookingpropositionwhichisindubitable，namelythattheincomederivedintheaggregatebyalltheelementsinthecommunityconcernedinaproductiveactivitynecessarilyhasavalueexactlyequaltothevalueoftheoutput。
Similarlyitisnaturaltosupposethattheactof[Page21]POSTULATES
OFTHECLASSICALECONOMICS
anindividual，bywhichheenricheshimselfwithoutapparentlytakinganythingfromanyoneelse，mustalsoenrichthecommunityasawhole；sothatasinthepassagejustquotedfromMarshallanactofindividualsavinginevitablyleadstoaparallelactofinvestment。For，oncemore，itisindubitablethatthesumofthenetincrementsofthewealthofindividualsmustbeexactlyequaltotheaggregatenetincrementofthewealthofthecommunity。
Thosewhothinkinthiswayaredeceived，nevertheless，byanopticalillusion，whichmakestwoessentiallydifferentactivitiesappeartobethesame。Theyarefallaciouslysupposingthatthereisanexuswhichunitesdecisionstoabstainfrompresentconsumptionwithdecisionstoprovideforfutureconsumption；whereasthemotiveswhichdeterminethelatterarenotlinkedinanysimplewaywiththemotiveswhichdeterminetheformer。
Itis，then，theassumptionofequalitybetweenthedemandpriceofoutputasawholeanditssupplypricewhichistoberegardedastheclassicaltheory's'axiomofparallels'。Grantedthis，alltherestfollows¾；thesocialadvantagesofprivateandnationalthrift，thetraditionalattitudetowardstherateofinterest，theclassicaltheoryofunemployment，thequantitytheoryofmoney，theunqualifiedadvantagesoflaissez-faireinrespectofforeigntradeandmuchelsewhichweshallhavetoquestion。
VII
Atdifferentpointsinthischapterwehavemadetheclassicaltheorytodependinsuccessionontheassumptions:1thattherealwageisequaltothemarginaldisutilityoftheexistingemployment；
2thatthereisnosuchthingasinvoluntaryunemploymentinthestrictsense；
3thatsupplycreatesitsowndemandinthesense[Page22]THEGENERAL
THEORYOFEMPLOYMENTthattheaggregatedemandpriceisequaltotheaggregatesupplypriceforalllevelsofoutputandemployment。Thesethreeassumptions，however，allamounttothesamethinginthesensethattheyallstandandfalltogether，anyoneofthemlogicallyinvolvingtheothertwo。
ThePrincipleofEffectiveDemand[Page23]
Chapter3
THEPRINCIPLEOFEFFECTIVEDEMAND
I
Weneed，tostartwith，afewtermswhichwillbedefinedpreciselylater。Inagivenstateoftechnique，resourcesandcosts，theemploymentofagivenvolumeoflabourbyanentrepreneurinvolveshimintwokindsofexpense:firstofall，theamountswhichhepaysouttothefactorsofproductionexclusiveofotherentrepreneursfortheircurrentservices，whichweshallcallthefactorcostoftheemploymentinquestion；
andsecondly，theamountswhichhepaysouttootherentrepreneursforwhathehastopurchasefromthemtogetherwiththesacrificewhichheincursbyemployingtheequipmentinsteadofleavingitidle，whichweshallcalltheusercostoftheemploymentinquestion。[31]Theexcessofthevalueoftheresultingoutputoverthesumofitsfactorcostanditsusercostistheprofitor，asweshallcallit，theincomeoftheentrepreneur。Thefactorcostis，ofcourse，thesamething，lookedatfromthepointofviewoftheentrepreneur，aswhatthefactorsofproductionregardastheirincome。Thusthefactorcostandtheentrepreneur'sprofitmakeup，betweenthem，whatweshalldefineasthetotalincomeresultingfromtheemploymentgivenbytheentrepreneur。Theentrepreneur'sprofitthusdefinedis，asitshouldbe，thequantitywhichheendeavourstomaximisewhenheisdecidingwhatamountofemploy-
[Page24]THEGENERAL
THEORYOFEMPLOYMENT
menttooffer。Itissometimesconvenient，whenwearelookingatitfromtheentrepreneur'sstandpoint，tocalltheaggregateincomei。e。
factorcostplusprofitresultingfromagivenamountofemploymenttheproceedsofthatemployment。Ontheotherhand，theaggregatesupplyprice[32]oftheoutputofagivenamountofemploymentistheexpectationofproceedswhichwilljustmakeitworththewhileoftheentrepreneurstogivethatemployment。[33]
Itfollowsthatinagivensituationoftechnique，resourcesandfactorcostperunitofemployment，theamountofemployment，bothineachindividualfirmandindustryandintheaggregate，dependsontheamountoftheproceedswhichtheentrepreneursexpecttoreceivefromthecorrespondingoutput。
[34]Forentrepreneurswillendeavourtofixtheamountofemploy-
[Page25]THEPRINCIPLE
OFEFFECTIVEDEMAND
mentatthelevelwhichtheyexpecttomaximisetheexcessoftheproceedsoverthefactorcost。
LetZbetheaggregatesupplypriceoftheoutputfromemployingNmen，therelationshipbetweenZandNbeingwrittenZ=fN，whichcanbecalledtheaggregatesupplyfunction。[35]Similarly，letDbetheproceedswhichentrepreneursexpecttoreceivefromtheemploymentofNmen，therelationshipbetweenDandNbeingwrittenD=fN，whichcanbecalledtheaggregatedemandfunction。
NowifforagivenvalueofNtheexpectedproceedsaregreaterthantheaggregatesupplyprice，i。e。ifDisgreaterthanZ，therewillbeanincentivetoentrepreneurstoincreaseemploymentbeyondNand，ifnecessary，toraisecostsbycompetingwithoneanotherforthefactorsofproduction，uptothevalueofNforwhichZhasbecomeequaltoD。Thusthevolumeofemploymentisgivenbythepointofintersectionbetweentheaggregatedemandfunctionandtheaggregatesupplyfunction；foritisatthispointthattheentrepreneurs'
expectationofprofitswillbemaximised。ThevalueofDatthepointoftheaggregatedemandfunction，whereitisintersectedbytheaggregatesupplyfunction，willbecalledtheeffectivedemand。
SincethisisthesubstanceoftheGeneralTheoryofEmployment，whichitwillbeourobjecttoexpound，thesucceedingchapterswillbelargelyoccupiedwithexaminingthevariousfactorsuponwhichthesetwofunctionsdepend。
Theclassicaldoctrine，ontheotherhand，whichusedtobeexpressedcategoricallyinthestatementthat'SupplycreatesitsownDemand'andcontinuestounderlieallorthodoxeconomictheory，involvesaspecialassumptionastotherelationshipbetweenthesetwofunctions。For'SupplycreatesitsownDemand'mustmeanthatfNandfN
areequalforallvalues[Page26]THEGENERAL
THEORYOFEMPLOYMENT
ofN，i。e。foralllevelsofoutputandemployment；andthatwhenthereisanincreaseinZ=fN
correspondingtoanincreaseinN，D=fN
necessarilyincreasesbythesameamountasZ。Theclassicaltheoryassumes，inotherwords，thattheaggregatedemandpriceorproceeds
alwaysaccommodatesitselftotheaggregatesupplyprice；sothat，whateverthevalueofNmaybe，theproceedsDassumeavalueequaltotheaggregatesupplypriceZwhichcorrespondstoN。Thatistosay，effectivedemand，insteadofhavingauniqueequilibriumvalue，isaninfiniterangeofvaluesallequallyadmissible；andtheamountofemploymentisindeterminateexceptinsofarasthemarginaldisutilityoflaboursetsanupperlimit。
Ifthisweretrue，competitionbetweenentrepreneurswouldalwaysleadtoanexpansionofemploymentuptothepointatwhichthesupplyofoutputasawholeceasestobeelastic，i。e。whereafurtherincreaseinthevalueoftheeffectivedemandwillnolongerbeaccompaniedbyanyincreaseinoutput。Evidentlythisamountstothesamethingasfullemployment。Inthepreviouschapterwehavegivenadefinitionoffullemploymentintermsofthebehaviouroflabour。Analternative，thoughequivalent，criterionisthatatwhichwehavenowarrived，namelyasituationinwhichaggregateemploymentisinelasticinresponsetoanincreaseintheeffectivedemandforitsoutput。ThusSay'slaw，thattheaggregatedemandpriceofoutputasawholeisequaltoitsaggregatesupplypriceforallvolumesofoutput，isequivalenttothepropositionthatthereisnoobstacletofullemployment。
If，however，thisisnotthetruelawrelatingtheaggregatedemandandsupplyfunctions，thereisavitallyimportantchapterofeconomictheorywhichremainstobewrittenandwithoutwhichalldiscussionsconcerningthevolumeofaggregateemploymentarefutile。
[Page27]THEPRINCIPLE
OFEFFECTIVEDEMAND
II
Abriefsummaryofthetheoryofemploymenttobeworkedoutinthecourseofthefollowingchaptersmay，perhaps，helpthereaderatthisstage，eventhoughitmaynotbefullyintelligible。Thetermsinvolvedwillbemorecarefullydefinedinduecourse。Inthissummaryweshallassumethatthemoney-wageandotherfactorcostsareconstantperunitoflabouremployed。Butthissimplification，withwhichweshalldispenselater，isintroducedsolelytofacilitatetheexposition。Theessentialcharacteroftheargumentispreciselythesamewhetherornotmoney-wages，etc。，areliabletochange。
Theoutlineofourtheorycanbeexpressedasfollows。Whenemploymentincreases，aggregaterealincomeisincreased。Thepsychologyofthecommunityissuchthatwhenaggregaterealincomeisincreasedaggregateconsumptionisincreased，butnotbysomuchasincome。Henceemployerswouldmakealossifthewholeoftheincreasedemploymentweretobedevotedtosatisfyingtheincreaseddemandforimmediateconsumption。Thus，tojustifyanygivenamountofemploymenttheremustbeanamountofcurrentinvestmentsufficienttoabsorbtheexcessoftotaloutputoverwhatthecommunitychoosestoconsumewhenemploymentisatthegivenlevel。Forunlessthereisthisamountofinvestment，thereceiptsoftheentrepreneurswillbelessthanisrequiredtoinducethemtoofferthegivenamountofemployment。Itfollows，therefore，that，givenwhatweshallcallthecommunity'spropensitytoconsume，theequilibriumlevelofemployment，i。e。thelevelatwhichthereisnoinducementtoemployersasawholeeithertoexpandortocontractemployment，willdependontheamountofcurrentinvestment。Theamountofcurrentinvestmentwilldepend，inturn，onwhatweshallcalltheinducementtoinvest；andtheinducementtoinvestwill[Page28]THEGENERAL
THEORYOFEMPLOYMENT
befoundtodependontherelationbetweenthescheduleofthemarginalefficiencyofcapitalandthecomplexofratesofinterestonloansofvariousmaturitiesandrisks。
Thus，giventhepropensitytoconsumeandtherateofnewinvestment，therewillbeonlyonelevelofemploymentconsistentwithequilibrium；
sinceanyotherlevelwillleadtoinequalitybetweentheaggregatesupplypriceofoutputasawholeanditsaggregatedemandprice。Thislevelcannotbegreaterthanfullemployment，i。e。therealwagecannotbelessthanthemarginaldisutilityoflabour。Butthereisnoreasoningeneralforexpectingittobeequaltofullemployment。Theeffectivedemandassociatedwithfullemploymentisaspecialcase，onlyrealisedwhenthepropensitytoconsumeandtheinducementtoinveststandinaparticularrelationshiptooneanother。Thisparticularrelationship，whichcorrespondstotheassumptionsoftheclassicaltheory，isinasenseanoptimumrelationship。
Butitcanonlyexistwhen，byaccidentordesign，currentinvestmentprovidesanamountofdemandjustequaltotheexcessoftheaggregatesupplypriceoftheoutputresultingfromfullemploymentoverwhatthecommunitywillchoosetospendonconsurnptionwhenitisfullyemployed。
Thistheorycanbesummedupinthefollowingpropositions:
1Inagivensituationoftechnique，resourcesandcosts，incomebothmoney-incomeandrealincomedependsonthevolumeofemploymentN。
2Therelationshipbetweenthecommunity'sincomeandwhatitcanbeexpectedtospendonconsumption，designatedbyD1，willdependonthepsychologicalcharacteristicofthecommunity，whichweshallcallitspropensitytoconsume。Thatistosay，consumptionwilldependonthelevelofaggregateincomeand，therefore，onthelevelofemploymentN，exceptwhenthereissomechangeinthepropensitytoconsume。
[Page29]THEPRINCIPLE
OFEFFECTIVEDEMAND
3TheamountoflabourNwhichtheentrepreneursdecidetoemploydependsonthesumDoftwoquantities，namelyD1，theamountwhichthecommunityisexpectedtospendonconsumption，andD2，theamountwhichitisexpectedtodevotetonewinvestment。Diswhatwehavecalledabovetheeffectivedemand。
4SinceD1D2=D=fN，whereistheaggregatesupplyfunction，andsince，aswehaveseenin2above，D1isafunctionofN，whichwemaywritecN，dependingonthepropensitytoconsume，itfollowsthatfN-cN
=D2。
5Hencethevolumeofemploymentinequilibriumdependsoni
theaggregatesupplyfunction，，iithepropensitytoconsume，，andiiithevolumeofinvestment，D2。ThisistheessenceoftheGeneralTheoryofEmployment。
6ForeveryvalueofNthereisacorrespondingmarginalproductivityoflabourinthewage-goodsindustries；anditisthiswhichdeterminestherealwage。5is，therefore，subjecttotheconditionthatNcannotexceedthevaluewhichreducestherealwagetoequalitywiththemarginaldisutilityoflabour。ThismeansthatnotallchangesinDarecompatiblewithourtemporaryassumptionthatmoney-wagesareconstant。Thusitwillbeessentialtoafullstatementofourtheorytodispensewiththisassumption。
7Ontheclassicaltheory，accordingtowhichD
=fNforallvaluesofN，thevolumeofemploymentisinneutralequilibriumforallvaluesofNlessthanitsmaximumvalue；sothattheforcesofcompetitionbetweenentrepreneursmaybeexpectedtopushittothismaximumvalue。Onlyatthispoint，ontheclassicaltheory，cantherebestableequilibrium。
8Whenemploymentincreases，D1willincrease，butnotbysomuchasD；sincewhenourincomeincreasesourconsumptionincreasesalso，butnotbysomuch。Thekeytoourpracticalproblemistobefoundinthis[Page30]THEGENERAL
THEORYOFEMPLOYMENT
psychologicallaw。ForitfollowsfromthisthatthegreaterthevolumeofemploymentthegreaterwillbethegapbetweentheaggregatesupplypriceZofthecorrespondingoutputandthesumD1
whichtheentrepreneurscanexpecttogetbackoutoftheexpenditureofconsumers。Hence，ifthereisnochangeinthepropensitytoconsume，employmentcannotincrease，unlessatthesametimeD2isincreasingsoastofilltheincreasinggapbetweenZandD1。
Thus¾；exceptonthespecialassumptionsoftheclassicaltheoryaccordingtowhichthereissomeforceinoperationwhich，whenemploymentincreases，alwayscausesD2toincreasesufficientlytofillthewideninggapbetweenZandD1¾；theeconomicsystemmayfinditselfinstableequilibriumwithNatalevelbelowfullemployment，namelyatthelevelgivenbytheintersectionoftheaggregatedemandfunctionwiththeaggregatesupplyfunction。
Thusthevolumeofemploymentisnotdeterminedbythemarginaldisutilityoflabourmeasuredintermsofrealwages，exceptinsofarasthesupplyoflabouravailableatagivenrealwagesetsamaximumleveltoemployment。
Thepropensitytoconsumeandtherateofnewinvestmentdeterminebetweenthemthevolumeofemployment，andthevolumeofemploymentisuniquelyrelatedtoagivenlevelofrealwages¾；nottheotherwayround。Ifthepropensitytoconsumeandtherateofnewinvestmentresultinadeficienteffectivedemand，theactuallevelofemploymentwillfallshortofthesupplyoflabourpotentiallyavailableattheexistingrealwage，andtheequilibriumrealwagewillbegreaterthanthemarginaldisutilityoftheequilibriumlevelofemployment。
Thisanalysissuppliesuswithanexplanationoftheparadoxofpovertyinthemidstofplenty。Forthemereexistenceofaninsufficiencyofeffectivedemandmay，andoftenwill，bringtheincreaseofemploymenttoastandstillbeforealeveloffullemploy[Page31]THEPRINCIPLE
OFEFFECTIVEDEMAND
menthasbeenreached。Theinsufficiencyofeffectivedemandwillinhibittheprocessofproductioninspiteofthefactthatthemarginalproductoflabourstillexceedsinvaluethemarginaldisutilityofemployment。
Moreoverthericherthecommunity，thewiderwilltendtobethegapbetweenitsactualanditspotentialproduction；andthereforethemoreobviousandoutrageousthedefectsoftheeconomicsystem。Forapoorcommunitywillbepronetoconsumebyfarthegreaterpartofitsoutput，sothataverymodestmeasureofinvestmentwillbesufficienttoprovidefullemployment；whereasawealthycommunitywillhavetodiscovermuchampleropportunitiesforinvestmentifthesavingpropensitiesofitswealthiermembersaretobecompatiblewiththeemploymentofitspoorermembers。
Ifinapotentiallywealthycommunitytheinducementtoinvestisweak，then，inspiteofitspotentialwealth，theworkingoftheprincipleofeffectivedemandwillcompelittoreduceitsactualoutput，until，inspiteofitspotentialwealth，ithasbecomesopoorthatitssurplusoveritsconsumptionissufficientlydiminishedtocorrespondtotheweaknessoftheinducementtoinvest。
Butworsestill。Notonlyisthemarginalpropensitytoconsume[36
]weakerinawealthycommunity，but，owingtoitsaccumulationofcapitalbeingalreadylarger，theopportunitiesforfurtherinvestmentarelessattractiveunlesstherateofinterestfallsatasufficientlyrapidrate；
which'bringsustothetheoryoftherateofinterestandtothereasonswhyitdoesnotautomaticallyfalltotheappropriatelevel，whichwilloccupyBookIV。
Thustheanalysisofthepropensitytoconsume，thedefinitionofthemarginalefficiencyofcapitalandthetheoryoftherateofinterestarethethreemaingapsinourexistingknowledgewhichitwillbenecessarytofill。Whenthishasbeenaccomplished，[Page32]THEGENERAL
THEORYOFEMPLOYMENT
weshallfindthatthetheoryofpricesfallsintoitsproperplaceasamatterwhichissubsidiarytoourgeneraltheory。Weshalldiscover，however，thatmoneyplaysanessentialpartinourtheoryoftherateofinterest；andweshallattempttodisentanglethepeculiarcharacteristicsofmoneywhichdistinguishitfromotherthings。
III
TheideathatwecansafelyneglecttheaggregatedemandfunctionisfundamentaltotheRicardianeconomics，whichunderliewhatwehavebeentaughtformorethanacentury。Malthus，indeed，hadvehementlyopposedRicardo'sdoctrinethatitwasimpossibleforeffectivedemandtobedeficient；
butvainly。For，sinceMalthuswasunabletoexplainclearlyapartfromanappealtothefactsofcommonobservationhowandwhyeffectivedemandcouldbedeficientorexcessive，hefailedtofurnishanalternativeconstruction；
andRicardoconqueredEnglandascompletelyastheHolyInquisitionconqueredSpain。Notonlywashistheoryacceptedbythecity，bystatesmenandbytheacademicworld。Butcontroversyceased；theotherpointofviewcompletelydisappeared；itceasedtobediscussed。ThegreatpuzzleofeffectivedemandwithwhichMalthushadwrestledvanishedfromeconomicliterature。YouwillnotfinditmentionedevenonceinthewholeworksofMarshall，EdgeworthandProfessorPigou，fromwhosehandstheclassicaltheoryhasreceiveditsmostmatureembodiment。Itcouldonlyliveonfurtively，belowthesurface，intheunderworldsofKarlMarx，SilvioGesellorMajorDouglas。
ThecompletenessoftheRicardianvictoryissomethingofacuriosityandamystery。Itmusthavebeenduetoacomplexofsuitabilitiesinthedoctrinetotheenvironmentintowhichitwasprojected。Thatit[Page33]THEPRINCIPLE
OFEFFECTIVEDEMAND
reachedconclusionsquitedifferentfromwhattheordinaryuninstructedpersonwouldexpect，added，Isuppose，toitsintellectualprestige。Thatitsteaching，translatedintopractice，wasaustereandoftenunpalatable，lentitvirtue。Thatitwasadaptedtocarryavastandconsistentlogicalsuperstructure，gaveitbeauty。Thatitcouldexplainmuchsocialinjusticeandapparentcrueltyasaninevitableincidentintheschemeofprogress，andtheattempttochangesuchthingsaslikelyonthewholetodomoreharmthangood，commendedittoauthority。Thatitaffordedameasureofjustificationtothefreeactivitiesoftheindividualcapitalist，attractedtoitthesupportofthedominantsocialforcebehindauthority。
Butalthoughthedoctrineitselfhasremainedunquestionedbyorthodoxeconomistsuptoalatedate，itssignalfailureforpurposesofscientificpredictionhasgreatlyimpaired，inthecourseoftime，theprestigeofitspractitioners。Forprofessionaleconomists，afterMalthus，wereapparentlyunmovedbythelackofcorrespondencebetweentheresultsoftheirtheoryandthefactsofobservation；¾；adiscrepancywhichtheordinarymanhasnotfailedtoobserve，withtheresultofhisgrowingunwillingnesstoaccordtoeconomiststhatmeasureofrespectwhichhegivestoothergroupsofscientistswhosetheoreticalresultsareconfirmedbyobservationwhentheyareappliedtothefacts。
Thecelebratedoptimismoftraditionaleconomictheory，whichhasledtoeconomistsbeinglookeduponasCandides，who，havingleftthisworldforthecultivationoftheirgardens，teachthatallisforthebestinthebestofallpossibleworldsprovidedwewillletwellalone，isalsotobetraced，Ithink，totheirhavingneglectedtotakeaccountofthedragonprosperitywhichcanbeexercisedbyaninsufficiencyofeffectivedemand。Fortherewouldobviouslybeanaturaltendencytowardstheoptimumemploymentofresourcesinasocietywhichwasfunctioningafterthe[Page34]THEGENERAL
THEORYOFEMPLOYMENT
manneroftheclassicalpostulates。Itmaywellbethattheclassicaltheoryrepresentsthewayinwhichweshouldlikeoureconomytobehave。
Buttoassumethatitactuallydoessoistoassumeourdifficultiesaway。
TheChoiceofUnits[Page37]
Chapter4
THECHOICEOFUNITS
I
Inthisandthenextthreechaptersweshallbeoccupiedwithanattempttoclearupcertainperplexitieswhichhavenopeculiarorexclusiverelevancetotheproblemswhichitisourspecialpurposetoexamine。Thusthesechaptersareinthenatureofadigression，whichwillpreventusforatimefrompursulngourmaintheme。Theirsubject-matterisonlydiscussedherebecauseitdoesnothappentohavebeenalreadytreatedelsewhereinawaywhichIfindadequatetotheneedsofmyownparticularenquiry。
Thethreeperplexitieswhichmostimpededmyprogressinwritingthisbook，sothatIcouldnotexpressmyselfconvenientlyuntilIhadfoundsomesolutionforthem，are:firstly，thechoiceoftheunitsofquantityappropriatetotheproblemsoftheeconomicsystemasawhole；secondly，thepartplayedbyexpectationineconomicanalysis；and，thirdly，thedefinitionofincome。
II
Thattheunits，intermsofwhicheconomistscommonlywork，areunsatisfactorycanbeillustratedbytheconceptsofthenationaldividend，thestockofrealcapitalandthegeneralprice-level:
iThenationaldividend，asdefinedbyMarshall[Page38]THEGENERAL
THEORYOFEMPLOYMENT
andProfessorPigou，[37]measuresthevolumeofcurrentoutputorrealincomeandnotthevalueofoutputormoney-income。[38]Furthermore，itdepends，insomesense，onnetoutput；¾；onthenetaddition，thatistosay，totheresourcesofthecommunityavailableforconsumptionorforretentionascapitalstock，duetotheeconomicactivitiesandsacrificesofthecurrentperiod，afterallowingforthewastageofthestockofrealcapitalexistingatthecommencementoftheperiod。Onthisbasisanattemptismadetoerectaquantitativescience。
Butitisagraveobjectiontothisdefinitionforsuchapurposethatthecommunity'soutputofgoodsandservicesisanon-homogeneouscomplexwhichcannotbemeasured，strictlyspeaking，exceptincertainspecialcases，asforexamplewhenalltheitemsofoneoutputareincludedinthesameproportionsinanotheroutput。
iiThedifficultyisevengreaterwhen，inordertocalculatenetoutput，wetrytomeasurethenetadditiontocapitalequipment；forwehavetofindsomebasisforaquantitativecomparisonbetweenthenewitemsofequipmentproducedduringtheperiodandtheolditemswhichhaveperishedbywastage。Inordertoarriveatthenetnationaldividend，ProfessorPigou[39]deductssuchobsolescence，etc。，'asmayfairlybecalled"normal"；andthepracticaltestofnormalityisthatthedepletionissufficientlyregulartobeforeseen，ifnotindetail，atleastinthelarge'。But，sincethisdeductionisnotadeductionintermsofmoney，heisinvolvedinassumingthattherecanbeachangeinphysicalquantity，althoughtherehasbeennophysicalchange；i。e。heiscovertlyintroducingchangesinvalue。
[Page39]THECHOICE
OFUNITS
Moreover，heisunabletodeviseanysatisfactoryformula[40]toevaluatenewequipmentagainstoldwhen，owingtochangesintechnique，thetwoarenotidentical。IbelievethattheconceptatwhichProfessorPigouisaimingistherightandappropriateconceptforeconomicanalysis。
But，untilasatisfactorysystemofunitshasbeenadopted，itsprecisedefinitionisanimpossibletask。Theproblemofcomparingonerealoutputwithanotherandofthencalculatingnetoutputbysettingoffnewitemsofequipmentagainstthewastageofolditemspresentsconundrumswhichpermit，onecanconfidentlysay，ofnosolution。
iiiThirdly，thewell-known，butunavoidable，elementofvaguenesswhichadmittedlyattendstheconceptofthegeneralprice-levelmakesthistermveryunsatisfactoryforthepurposesofacausalanalysis，whichoughttobeexact。
Neverthelessthesedifficultiesarerightlyregardedas'conundrums'。
Theyare'purelytheoretical'inthesensethattheyneverperplex，orindeedenterinanywayinto，businessdecisionsandhavenorelevancetothecausalsequenceofeconomicevents，whichareclear-cutanddeterminateinspiteofthequantitativeindeterminacyoftheseconcepts。Itisnatural，therefore，toconcludethattheynotonlylackprecisionbutareunnecessary。
Obviouslyourquantitativeanalysismustbeexpressedwithoutusinganyquantitativelyvagueexpressions。And，indeed，assoonasonemakestheattempt，itbecomesclear，asIhopetoshow，thatonecangetonmuchbetterwithoutthem。
Thefactthattwoincommensurablecollectionsofmiscellaneousobjectscannotinthemselvesprovidethematerialforaquantitativeanalysisneednot，ofcourse，preventusfrommakingapproximatestatisticalcomparisons，dependingonsomebroadelementofjudgmentratherthanofstrictcalculation，whichmaypossesssignificanceandvaliditywithincertainlimits。
[Page40]THEGENERAL
THEORYOFEMPLOYMENT
Buttheproperplaceforsuchthingsasnetrealoutputandthegenerallevelofpriceslieswithinthefieldofhistoricalandstatisticaldescription，andtheirpurposeshouldbetosatisfyhistoricalorsocialcuriosity，apurposeforwhichperfectprecision¾；suchasourcausalanalysisrequires，whetherornotourknowledgeoftheactualvaluesoftherelevantquantitiesiscompleteorexact¾；isneitherusualnornecessary。Tosaythatnetoutputto-dayisgreater，buttheprice-levellower，thantenyearsagooroneyearago，isapropositionofasimilarcharactertothestatementthatQueenVictoriawasabetterqueenbutnotahappierwomanthanQueenElizabeth¾；apropositionnotwithoutmeaningandnotwithoutinterest，butunsuitableasmaterialforthedifferentialcalculus。Ourprecisionwillbeamockprecisionifwetrytousesuchpartlyvagueandnon-quantitativeconceptsasthebasisofaquantitativeanalysis。
III
Oneveryparticularoccasion，letitberemembered，anentrepreneurisconcernedwithdecisionsastothescaleonwhichtoworkagivencapitalequipment；andwhenwesaythattheexpectationofanincreaseddemand，i。e。araisingoftheaggregatedemandfunction，willleadtoanincreaseinaggregateoutput，wereallymeanthatthefirms，whichownthecapitalequipment，willbeinducedtoassociatewithitagreateraggregateemploymentoflabour。Inthecaseofanindividualfirmorindustryproducingahomogeneousproductwecanspeaklegitimately，ifwewish，ofincreasesordecreasesofoutput。Butwhenweareaggregatingtheactivitiesofallfirms，wecannotspeakaccuratelyexceptintermsofquantitiesofemploymentappliedtoagivenequipment。Theconceptsofoutputasawholeanditsprice-levelarenotrequiredinthiscontext，sincewehavenoneedofanabsolute[Page41]THECHOICE
OFUNITS
measureofcurrentaggregateoutput，suchaswouldenableustocompareitsamountwiththeamountwhichwouldresultfromtheassociationofadifferentcapitalequipmentwithadifferentquantityofemployment。When，forpurposesofdescriptionorroughcomparison，wewishtospeakofanincreaseofoutput，wemustrelyonthegeneralpresumptionthattheamountofemploymentassociatedwithagivencapitalequipmentwillbeasatisfactoryindexoftheamountofresultantoutput；¾；thetwobeingpresumedtoincreaseanddecreasetogether，thoughnotinadefinitenumericalproportion。
IndealingwiththetheoryofemploymentIpropose，therefore，tomakeuseofonlytwofundamentalunitsofquantity，namely，quantitiesofmoney-valueandquantitiesofemployment。Thefirstoftheseisstrictlyhomogeneous，andthesecondcanbemadeso。For，insofarasdifferentgradesandkindsoflabourandsalariedassistanceenjoyamoreorlessfixedrelativeremuneration，thequantityofemploymentcanbesufficientlydefinedforourpurposebytakinganhour'semploymentofordinarylabourasourunitandweightinganhour'semploymentofspeciallabourinproportiontoitsremuneration；
i。e。anhourofspeciallabourremuneratedatdoubleordinaryrateswillcountastwounits。Weshallcalltheunitinwhichthequantityofemploymentismeasuredthelabour-unit；andthemoney-wageofalabour-unitweshallcallthewage-unit。[41]Thus，ifEisthewagesandsalaries
bill，Wthewage-unit，andNthequantityofemployment，E=N×；W。
Thisassumptionofhomogeneityinthesupplyoflabourisnotupsetbytheobviousfactofgreatdifferencesinthespecialisedskillofindividualworkersandintheirsuitabilityfordifferentoccupations。For，[Page42]THEGENERAL
THEORYOFEMPLOYMENT
iftheremunerationoftheworkersisproportionaltotheirefficiency，thedifferencesaredealtwithbyourhavingregardedindividualsascontributingtothesupplyoflabourinproportiontotheirremuneration；whilstif，asoutputincreases，agivenfirmhastobringinlabourwhichislessandlessefficientforitsspecialpurposesperwage-unitpaidtoit，thisismerelyonefactoramongothersleadingtoadiminishingreturnfromthecapitalequipmentintermsofoutputasmorelabourisemployedonit。Wesubsume，sotospeak，thenon-homogeneityofequallyremuneratedlabourunitsintheequipment，whichweregardaslessandlessadaptedtoemploytheavailablelabourunitsasoutputincreases，insteadofregardingtheavailablelabourunitsaslessandlessadaptedtouseahomogeneouscapitalequipment。Thusifthereisnosurplusofspecialisedorpractisedlabourandtheuseoflesssuitablelabourinvolvesahigherlabourcostperunitofoutput，thismeansthattherateatwhichthereturnfromtheequipmentdiminishesasemploymentincreasesismorerapidthanitwouldbeifthereweresuchasurplus。[42]Eveninthelimitingcasewheredifferentlabourunitsweresohighlyspecialisedastobealtogetherincapableofbeingsubstitutedforoneanother，thereisnoawkwardness；forthismerelymeansthattheelasticityofsupplyofoutputfromaparticulartypeofcapitalequipmentfallssuddenlytozerowhenalltheavailablelabourspecialisedtoitsuseisalreadyemployed。[43]Thusourassumptionofahomo-
[Page43]THECHOICE
OFUNITS
geneousunitoflabourinvolvesnodifficultiesunlessthereisgreatinstabilityintherelativeremunerationofdifferentlabour-units；andeventhisdifficultycanbedealtwith，ifitarises，bysupposingarapidliabilitytochangeinthesupplyoflabourandtheshapeoftheaggregatesupplyfunction。
Itismybeliefthatmuchunnecessaryperplexitycanbeavoidedifwelimitourselvesstrictlytothetwounits，moneyandlabour，whenwearedealingwiththebehaviouroftheeconomicsystemasawhole；reservingtheuseofunitsofparticularoutputsandequipmentstotheoccasionswhenweareanalysingtheoutputofindividualfirmsorindustriesinisolation；
andtheuseofvagueconcepts，suchasthequantityofoutputasawhole，thequantityofcapitalequipmentasawholeandthegenerallevelofprices，totheoccasionswhenweareattemptingsomehistoricalcomparisonwhichiswithincertainperhapsfairlywidelimitsavowedlyunpreciseandapproximate。
[Page44]THEGENERAL
THEORYOFEMPLOYMENT
Itfollowsthatweshallmeasurechangesincurrentoutputbyreferencetothenumberofhoursoflabourpaidforwhethertosatisfyconsumersortoproducefreshcapitalequipmentontheexistingcapitalequipment，hoursofskilledlabourbeingweightedinproportiontotheirremuneration。
Wehavenoneedofaquantitativecomparisonbetweenthisoutputandtheoutputwhichwouldresultfromassociatingadifferentsetofworkerswithadifferentcapitalequipment。Topredicthowentrepreneurspossessingagivenequipmentwillrespondtoashiftintheaggregatedemandfunctionitisnotnecessarytoknowhowthequantityoftheresultingoutput，thestandardoflifeandthegenerallevelofpriceswouldcomparewithwhattheywereatadifferentdateorinanothercountry。
IV
Itiseasilyshownthattheconditionsofsupply，suchasareusuallyexpressedintermsofthesupplycurve，andtheelasticityofsupplyrelatingoutputtoprice，canbehandledintermsofourtwochosenunitsbymeansoftheaggregatesupplyfunction，withoutreferencetoquantitiesofoutput，whetherweareconcernedwithaparticularfirmorindustryorwitheconomicactivityasawhole。FortheaggregatesupplyfunctionforagivenfirmandsimilarlyforagivenindustryorforindustryasawholeisgivenbyZr=frNr，whereZristheproceedsnetofusercosttheexpectationofwhichwillinducealevelofemploymentNr。If，therefore，therelationbetweenemploymentandoutputissuchthatanemploymentNrresultsinanoutputOr，whereOr=yrNr，itfollowsthatZrUrNrfrNrUrNr
p=¾；¾；¾；¾；¾；¾；¾；
=¾；¾；¾；¾；¾；¾；¾；¾；¾；
OryrNr
[Page45]THECHOICE
OFUNITS
istheordinarysupplycurve，whereUrNr
istheexpectedusercostcorrespondingtoalevelofemploymentNr。
Thusinthecaseofeachhomogeneouscommodity，forwhichOr=yrNr
hasadefinitemeaning，wecanevaluateZr=frNrintheordinaryway；butwecanthenaggregatetheNr'sinawayinwhichwecannotaggregatetheOr's，sinceSOrisnotanumericalquantity。Moreover，ifwecanassumethat，inagivenenvironment，agivenaggregateemploymentwillbedistributedinauniquewaybetweendifferentindustries，sothatNrisafunctionofN，furthersimplificationsarepossible。
ExpectationandEmployment[Page46]
Chapter5
EXPECTATIONASDETERMINING
OUTPUTANDEMPLOYMENT
I
Allproductionisforthepurposeofultimatelysatisfyingaconsumer。
Timeusuallyelapses，however¾；andsometimesmuchtime¾；betweentheincurringofcostsbytheproducerwiththeconsumerinviewandthepurchaseoftheoutputbytheultimateconsumer。Meanwhiletheentrepreneurincludingboththeproducerandtheinvestorinthisdescriptionhastoformthebestexpectations[44]hecanastowhattheconsumerswillbepreparedtopaywhenheisreadytosupplythemdirectlyorindirectlyaftertheelapseofwhatmaybealengthyperiod；andhehasnochoicebuttobeguidedbytheseexpectations，ifheistoproduceatallbyprocesseswhichoccupytime。
Theseexpectations，uponwhichbusinessdecisionsdepend，fallintotwogroups，certainindividualsorfirmsbeingspecialisedinthebusinessofframingthefirsttypeofexpectationandothersinthebusinessofframingthesecond。Thefirsttypeisconcernedwiththepricewhichamanufacturercanexpecttogetforhis'finished'outputatthetimewhenhecommitshimselftostartingtheprocesswhichwillproduceit；outputbeing'finished'fromthepointofviewofthemanufacturerwhenitisreadytobeusedortobesoldtoasecondparty。The[Page47]EXPECTATION
ANDEMPLOYMENT
secondtypeisconcernedwithwhattheentrepreneurcanhopetoearnintheshapeoffuturereturnsifhepurchasesor，perhaps，manufactures
'finished'outputasanadditiontohiscapitalequipment。Wemaycalltheformershort-termexpectationandthelatterlong-termexpectation。
Thusthebehaviourofeachindividualfirmindecidingitsdaily[45
]outputwillbedeterminedbyitsshort-termexpectations¾；expectationsastothecostofoutputonvariouspossiblescalesandexpectationsastothesale-proceedsofthisoutput；though，inthecaseofadditionstocapitalequipmentandevenofsalestodistributors，theseshort-termexpectationswilllargelydependonthelong-termormedium-termexpectationsofotherparties。Itisuponthesevariousexpectationsthattheamountofemploymentwhichthefirmsofferwilldepend。Theactuallyrealisedresultsoftheproductionandsaleofoutputwillonlyberelevanttoemploymentinsofarastheycauseamodificationofsubsequentexpectations。Nor，ontheotherhand，aretheoriginalexpectationsrelevant，whichledthefirmtoacquirethecapitalequipmentandthestockofintermediateproductsandhalf-finishedmaterialswithwhichitfindsitselfatthetimewhenithastodecidethenextday'soutput。Thus，oneachandeveryoccasionofsuchadecision，thedecisionwillbemade，withreferenceindeedtothisequipmentandstock，butinthelightofthecurrentexpectationsofprospectivecostsandsale-proceeds。
Now，ingeneral，achangeinexpectationswhethershort-termorlong-termwillonlyproduceitsfulleffectonemploymentoveraconsiderableperiod。Thechangeinemploymentduetoachangeinexpectationswillnotbethesameontheseconddayafterthechangeasonthefirst，orthesameonthe[Page48]THEGENERAL