THEORYOFEMPLOYMENT
　　thirddayasonthesecond，andsoon，eventhoughtherebenofurtherchangeinexpectations。Inthecaseofshort-termexpectationsthisisbecausechangesinexpectationarenot，asarule，sufficientlyviolentorrapid，whentheyarefortheworse，tocausetheabandonmentofworkonalltheproductiveprocesseswhich，inthelightoftherevisedexpectation，itwasamistaketohavebegun；whilst，whentheyareforthebetter，sometimeforpreparationmustneedselapsebeforeemploymentcanreachthelevelatwhichitwouldhavestoodifthestateofexpectationhadbeenrevisedsooner。Inthecaseoflong-termexpectations，equipmentwhichwillnotbereplacedwillcontinuetogiveemploymentuntilitiswornout；whilstwhenthechangeinlong-termexpectationsisforthebetter，employmentmaybeatahigherlevelatfirst，thanitwillbeaftertherehasbeentimetoadjusttheequipmenttothenewsituation。
　　Ifwesupposeastateofexpectationtocontinueforasufficientlengthoftimefortheeffectonemploymenttohaveworkeditselfoutsocompletelythatthereis，broadlyspeaking，nopieceofemploymentgoingonwhichwouldnothavetakenplaceifthenewstateofexpectationhadalwaysexisted，thesteadylevelofemploymentthusattainedmaybecalledthelong-periodemployment[46]correspondingtothatstateofexpectation。Itfollowsthat，althoughexpectationmaychangesofrequentlythattheactuallevelofemploymenthasneverhadtimetoreachthelong-periodemploymentcorrespondingtotheexistingstateofexpectation，neverthelesseverystateofexpectationhasitsdefinitecorrespondingleveloflong-periodemployment。
　　Letusconsider，firstofall，theprocessoftransition[Page49]EXPECTATION
　　ANDEMPLOYMENT
　　toalong-periodpositionduetoachangeinexpectation，whichisnotconfusedorinterruptedbyanyfurtherchangeinexpectation。Wewillfirstsupposethatthechangeisofsuchacharacterthatthenewlong-periodemploymentwillbegreaterthantheold。Now，asarule，itwillonlybetherateofinputwhichwillbemuchaffectedatthebeginning，thatistosay，thevolumeofworkontheearlierstagesofnewprocessesofproduction，whilsttheoutputofconsumption-goodsandtheamountofemploymentonthelaterstagesofprocesseswhichwerestartedbeforethechangewillremainmuchthesameasbefore。Insofarastherewerestocksofpartlyfinishedgoods，thisconclusionmaybemodified；thoughitislikelytoremaintruethattheinitialincreaseinemploymentwillbemodest。As，however，thedayspassby，employmentwillgraduallyincrease。Moreover，itiseasytoconceiveofconditionswhichwillcauseittoincreaseatsomestagetoahigherlevelthanthenewlong-periodemployment。
　　Fortheprocessofbuildingupcapitaltosatisfythenewstateofexpectationmayleadtomoreemploymentandalsotomorecurrentconsumptionthanwilloccurwhenthelong-periodpositionhasbeenreached。Thusthechangeinexpectationmayleadtoagradualcrescendointhelevelofemployment，risingtoapeakandthendecliningtothenewlong-periodlevel。Thesamethingmayoccurevenifthenewlong-periodlevelisthesameastheold，ifthechangerepresentsachangeinthedirectionofconsumptionwhichrenderscertainexistingprocessesandtheirequipmentobsolete。
　　Oragain，ifthenewlong-periodemploymentislessthantheold，thelevelofemploymentduringthetransitionmayfallforatimebelowwhatthenewlong-periodlevelisgoingtobe。Thusamerechangeinexpectationiscapableofproducinganoscillationofthesamekindofshapeasacyclicalmovement，inthecourseofworkingitselfout。ItwasmovementsofthiskindwhichIdiscussedinmyTreatiseonMoneyinconnection[Page50]THEGENERAL
　　THEORYOFEMPLOYMENT
　　withthebuildinguporthedepletionofstocksofworkingandliquidcapitalconsequentonchange。
　　Anuninterruptedprocessoftransition，suchastheabove，toanewlong-periodpositioncanbecomplicatedindetail。Buttheactualcourseofeventsismorecomplicatedstill。Forthestateofexpectationisliabletoconstantchange，anewexpectationbeingsuperimposedlongbeforethepreviouschangehasfullyworkeditselfout；sothattheeconomicmachineisoccupiedatanygiventimewithanumberofoverlappingactivities，theexistenceofwhichisduetovariouspaststatesofexpectation。
　　II
　　Thisleadsustotherelevanceofthisdiscussionforourpresentpurpose。
　　Itisevidentfromtheabovethatthelevelofemploymentatanytimedepends，inasense，notmerelyontheexistingstateofexpectationbutonthestatesofexpectationwhichhaveexistedoveracertainpastperiod。Neverthelesspastexpectations，whichhavenotyetworkedthemselvesout，areembodiedintheto-day'scapitalequipmentwithreferencetowhichtheentrepreneurhastomaketo-day'sdecisions，andonlyinfluencehisdecisionsinsofarastheyaresoembodied。Itfollows，therefore，that，inspiteoftheabove，to-day'semploymentcanbecorrectlydescribedasbeinggovernedbyto-day'sexpectationstakeninconjunctionwithto-day'scapitalequipment。
　　Expressreferencetocurrentlong-termexpectationscanseldombeavoided。
　　Butitwilloftenbesafetoomitexpressreferencetoshort-termexpectation，inviewofthefactthatinpracticetheprocessofrevisionofshort-termexpectationisagradualandcontinuousone，carriedonlargelyinthelightofrealisedresults；sothatexpectedandrealisedresultsrunintoandoverlaponeanotherintheirinfluence。For，althoughoutputandemploymentaredeterminedby[Page51]EXPECTATION
　　ANDEMPLOYMENT
　　theproducer'sshort-termexpectationsandnotbypastresults，themostrecentresultsusuallyplayapredominantpartindeterminingwhattheseexpectationsare。Itwouldbetoocomplicatedtoworkouttheexpectationsdenovowheneveraproductiveprocesswasbeingstarted；anditwould，moreover，beawasteoftimesincealargepartofthecircumstancesusuallycontinuesubstantiallyunchangedfromonedaytothenext。Accordinglyitissensibleforproducerstobasetheirexpectationsontheassumptionthatthemostrecentlyrealisedresultswillcontinue，exceptinsofarastherearedefinitereasonsforexpectingachange。Thusinpracticethereisalargeoverlapbetweentheeffectsonemploymentoftherealisedsale-proceedsofrecentoutputandthoseofthesale-proceedsexpectedfromcurrentinput；andproducers'forecastsaremoreoftengraduallymodifiedinthelightofresultsthaninanticipationofprospectivechanges。[
　　47]
　　Nevertheless，wemustnotforgetthat，inthecaseofdurablegoods，theproducer'sshort-termexpectationsarebasedonthecurrentlong-termexpectationsoftheinvestor；anditisofthenatureoflong-termexpectationsthattheycannotbecheckedatshortintervalsinthelightofrealisedresults。Moreover，asweshallseeinchapter12，whereweshallconsiderlong-termexpectationsinmoredetail，theyareliabletosuddenrevision。
　　Thusthefactorofcurrentlong-termexpectationscannotbeevenapproximatelyeliminatedorreplacedbyrealisedresults。
　　TheDefinitionofIncome[Page52]
　　Chapter6
　　THEDEFINITIONOFINCOME，SAVINGANDINVESTMENT
　　I。IncomeDuringanyperiodoftimeanentrepreneurwillhavesoldfinishedoutputtoconsumersortootherentrepreneursforacertainsumwhichwewilldesignateasA。Hewillalsohavespentacertainsum，designatedbyA1，onpurchasingfinishedoutputfromotherentrepreneurs。
　　Andhewillendupwithacapitalequipment，whichtermincludesbothhisstocksofunfinishedgoodsorworkingcapitalandhisstocksoffinishedgoods，havingavalueG。
　　Somepart，however，ofAG-A1willbeattributable，nottotheactivitiesoftheperiodinquestion，buttothecapitalequipmentwhichhehadatthebeginningoftheperiod。Wemust，therefore，inordertoarriveatwhatwemeanbytheincomeofthecurrentperiod，deductfromAG-A1acertainsum，torepresentthatpartofitsvaluewhichhasbeeninsomesensecontributedbytheequipmentinheritedfromthepreviousperiod。Theproblemofdefiningincomeissolvedassoonaswehavefoundasatisfactorymethodforcalculatingthisdeduction。
　　Therearetwopossibleprinciplesforcalculatingit，eachofwhichhasacertainsignificance；¾；oneoftheminconnectionwithproduction，andtheotherinconnectionwithconsumption。
　　Letusconsidertheminturn。
　　iTheactualvalueGofthecapitalequipmentat[Page53]THEDEFINITION
　　OFINCOME
　　theendoftheperiodisthenetresultoftheentrepreneur，ontheonehand，havingmaintainedandimproveditduringtheperiod，bothbypurchasesfromotherentrepreneursandbyworkdoneuponitbyhimself，and，ontheotherhand，havingexhaustedordepreciateditthroughusingittoproduceoutput。Ifhehaddecidednottouseittoproduceoutput，thereis，nevertheless，acertainoptimumsumwhichitwouldhavepaidhimtospendonmaintainingandimprovingit。Letussupposethat，inthisevent，hewouldhavespentB'onitsmaintenanceandimprovement，andthat，havinghadthisspentonit，itwouldhavebeenworthG'attheendoftheperiod。Thatistosay，G'-B'isthemaximumnetvaluewhichmighthavebeenconservedfromthepreviousperiod，ifithadnotbeenusedtoproduceA。TheexcessofthispotentialvalueoftheequipmentoverG-A1isthemeasureofwhathasbeensacrificedonewayoranothertoproduceA。Letuscallthisquantity，namelyG'-B'-G-A1，whichmeasuresthesacrificeofvalueinvolvedintheproductionofA，theusercostofA。UsercostwillbewrittenU。
　　[48]Theamountpaidoutbytheentrepreneurtotheotherfactorsofproductioninreturnfortheirservices，whichfromtheirpointofviewistheirincome，wewillcallthefactorcostofA。ThesumofthefactorcostFandtheusercostUweshallcalltheprimecostoftheoutputA。
　　Wecanthendefinetheincome[49]oftheentrepreneurasbeingtheexcessofthevalueofhisfinishedoutputsoldduringtheperiodoverhisprimecost。Theentrepreneur'sincome，thatistosay，istakenasbeingequaltothequantity，dependingonhisscaleofproduction，whichheendeavourstomaximise，i。e。tohisgrossprofit[Page54]THEGENERAL
　　THEORYOFEMPLOYMENT
　　intheordinarysenseofthisterm；¾；whichagreeswithcommonsense。Hence，sincetheincomeoftherestofthecommunityisequaltotheentrepreneur'sfactorcost，aggregateincomeisequaltoA-U。
　　Income，thusdefined，isacompletelyunambiguousquantity。Moreover，sinceitistheentrepreneur'sexpectationoftheexcessofthisquantityoverhisoutgoingstotheotherfactorsofproductionwhichheendeavourstomaximisewhenhedecideshowmuchemploymenttogivetotheotherfactorsofproduction，itisthequantitywhichiscausallysignificantforemployment。
　　Itisconceivable，ofcourse，thatG-A1mayexceedG'-B'，sothatusercostwillbenegative。Forexample，thismaywellbethecaseifwehappentochooseourperiodinsuchawaythatinputhasbeenincreasingduringtheperiodbutwithouttherehavingbeentimefortheincreasedoutputtoreachthestageofbeingfinishedandsold。
　　Itwillalsobethecase，wheneverthereispositiveinvestment，ifweimagineindustrytobesomuchintegratedthatentrepreneursmakemostoftheirequipmentforthemselves。Since，however，usercostisonlynegativewhentheentrepreneurhasbeenincreasinghiscapitalequipmentbyhisownlabour，wecan，inaneconomywherecapitalequipmentislargelymanufacturedbydifferentfirmsfromthosewhichuseit，normallythinkofusercostasbeingpositive。Moreover，itisdifficulttoconceiveofacasewheremarginalusercostassociatedwithanincreaseinA，i。e。dU/dA，willbeotherthanpositive。
　　Itmaybeconvenienttomentionhere，inanticipationofthelatterpartofthischapter，that，forthecommunityasawhole，theaggregateconsumptionCoftheperiodisequaltoSA-A1，andtheaggregateinvestmentIisequaltoSA1-U。Moreover，Uistheindividualentrepreneur'sdisinvestmentand-Uhisinvestmentinrespectofhisownequipmentexclusive[Page55]THEDEFINITION
　　OFINCOME
　　ofwhathebuysfromotherentrepreneurs。ThusinacompletelyintegratedsystemwhereA1=0consumptionisequaltoAandinvestmentto-U，i。e。
　　toG-G'-B'。Theslightcomplicationoftheabove，throughtheintroductionofA1，issimplyduetothedesirabilityofprovidinginageneralisedwayforthecaseofanon-integratedsystemofproduction。
　　Furthermore，theeffectivedemandissimplytheaggregateincomeorproceedswhichtheentrepreneursexpecttoreceive，inclusiveoftheincomeswhichtheywillhandontotheotherfactorsofproduction，fromtheamountofcurrentemploymentwhichtheydecidetogive。Theaggregatedemandfunctionrelatesvarioushypotheticalquantitiesofemploymenttotheproceedswhichtheiroutputsareexpectedtoyield；andtheeffectivedemandisthepointontheaggregatedemandfunctionwhichbecomeseffectivebecause，takeninconjunctionwiththeconditionsofsupply，itcorrespondstothelevelofemploymentwhichmaximisestheentrepreneur'sexpectationofprofit。
　　Thissetofdefinitionsalsohastheadvantagethatwecanequatethemarginalproceedsorincometothemarginalfactorcost；andthusarriveatthesamesortofpropositionsrelatingmarginalproceedsthusdefinedtomarginalfactorcostsashavebeenstatedbythoseeconomistswho，byignoringusercostorassumingittobezero，haveequatedsupplyprice[50]tomarginalfactorcost。[51]
　　[Page56]THEGENERAL
　　THEORYOFEMPLOYMENT
　　iiWeturn，next，tothesecondoftheprinciplesreferredtoabove。
　　Wehavedealtsofarwiththatpartofthechangeinthevalueofthecapitalequipmentattheendoftheperiodascomparedwithitsvalueatthebeginningwhichisduetothevoluntarydecisionsoftheentrepreneurinseekingtomaximisehisprofit。Buttheremay，inaddition，beaninvoluntarylossorgaininthevalueofhiscapitalequipment，occurringforreasonsbeyondhiscontrolandirrespectiveofhiscurrentdecisions，onaccountofe。g。achangeinmarketvalues，wastagebyobsolescenceorthemerepassageoftime，ordestructionbycatastrophesuchaswarorearthquake。Nowsomepartoftheseinvoluntarylosses，whilsttheyareunavoidable，are¾；broadlyspeaking¾；notunexpected；suchaslossesthroughthelapseoftimeirrespectiveofuse，andalso'normal'obsolescencewhich，asProfessorPigouexpressesit，'issufficientlyregulartobeforeseen，ifnotindetail，atleastinthelarge'，including，wemayadd，thoselossestothecommunityasawholewhicharesufficientlyregulartobecommonlyregardedas'insurablerisks'。
　　Letusignoreforthemomentthefactthattheamountoftheexpectedlossdependsonwhentheexpectationisassumedtobeframed，andletuscallthedepreciationoftheequipment，whichisinvoluntarybutnotunexpected，i。e。theexcessoftheexpecteddepreciationovertheusercost，thesupplementarycost，whichwillbewrittenV。Itis，perhaps，hardlynecessarytopointoutthatthisdefinitionisnotthesameasMarshall'sdefinitionofsupplementarycost，thoughtheunderlyingidea，namely，ofdealingwiththatpartoftheexpecteddepreciationwhichdoesnotenterintoprimecost，issimilar。
　　[Page57]THEDEFINITION
　　OFINCOME
　　Inreckoning，therefore，thenetincomeandthenetprofitoftheentrepreneuritisusualtodeducttheestimatedamountofthesupplementarycostfromhisincomeandgrossprofitasdefinedabove。Forthepsychologicaleffectontheentrepreneur，whenheisconsideringwhatheisfreetospendandtosave，ofthesupplementarycostisvirtuallythesameasthoughitcameoffhisgrossprofit。Inhiscapacityasaproducerdecidingwhetherornottousetheequipment，primecostandgrossprofit，asdefinedabove，arethesignificantconcepts。Butinhiscapacityasaconsumertheamountofthesupplementarycostworksonhismindinthesamewayasifitwereapartoftheprimecost。Henceweshallnotonlycomenearesttocommonusagebutwillalsoarriveataconceptwhichisrelevanttotheamountofconsumption，if，indefiningaggregatenetincome，wedeductthesupplementarycostaswellastheusercost，sothataggregatenetincomeisequaltoA-U-V。
　　Thereremainsthechangeinthevalueoftheequipment，duetounforeseenchangesinmarketvalues，exceptionalobsolescenceordestructionbycatastrophe，whichisbothinvoluntaryand¾；inabroadsense¾；unforeseen。Theactuallossunderthishead，whichwedisregardeveninreckoningnetincomeandchargetocapitalaccount，maybecalledthewindfallloss。
　　ThecausalsignificanceofnetincomeliesinthepsychologicalinfluenceofthemagnitudeofVontheamountofcurrentconsumption，sincenetincomeiswhatwesupposetheordinarymantoreckonhisavailableincometobewhenheisdecidinghowmuchtospendoncurrentconsumption。Thisisnot，ofcourse，theonlyfactorofwhichhetakesaccountwhenheisdecidinghowmuchtospend。Itmakesaconsiderabledifference，forexample，howmuchwindfallgainorlossheismakingoncapitalaccount。Butthereisadifferencebetweenthesupplementarycostandawindfalllossinthatchangesintheformerareapttoaffect[Page58]THEGENERAL
　　THEORYOFEMPLOYMENT
　　himinjustthesamewayaschangesinhisgrossprofit。Itistheexcessoftheproceedsofthecurrentoutputoverthesumoftheprimecostandthesupplementarycostwhichisrelevanttotheentrepreneur'sconsumption；
　　whereas，althoughthewindfalllossorgainentersintohisdecisions，itdoesnotenterintothemonthesamescale¾；agivenwindfalllossdoesnothavethesameeffectasanequalsupplementarycost。
　　Wemustnowrecur，however，tothepointthatthelinebetweensupplementarycostsandwindfalllosses，i。e。betweenthoseunavoidablelosseswhichwethinkitpropertodebittoincomeaccountandthosewhichitisreasonabletoreckonasawindfalllossorgainoncapitalaccount，ispartlyaconventionalorpsychologicalone，dependingonwhatarethecommonlyacceptedcriteriaforestimatingtheformer。Fornouniqueprinciplecanbeestablishedfortheestimationofsupplementarycost，anditsamountwilldependonourchoiceofanaccountingmethod。Theexpectedvalueofthesupplementarycost，whentheequipmentwasoriginallyproduced，isadefinitequantity。
　　Butifitisre-estimatedsubsequently，itsamountovertheremainderofthelifeoftheequipmentmayhavechangedasaresultofachangeinthemeantimeinourexpectations；thewindfallcapitallossbeingthediscountedvalueofthedifferencebetweentheformerandtherevisedexpectationoftheprospectiveseriesofUV。Itisawidelyapprovedprincipleofbusinessaccounting，endorsedbytheInlandRevenueauthorities，toestablishafigureforthesumofthesupplementarycostandtheusercostwhentheequipmentisacquiredandtomaintainthisunalteredduringthelifeoftheequipment，irrespectiveofsubsequentchangesinexpectation。
　　Inthiscasethesupplementarycostoveranyperiodmustbetakenastheexcessofthispredeterminedfigureovertheactualusercost。Thishastheadvantageofensuringthatthewindfallgainorlossshallbezerooverthelifeofthe[Page59]THEDEFINITION
　　OFINCOME
　　equipmenttakenasawhole。Butitisalsoreasonableincertaincircumstancestorecalculatetheallowanceforsupplementarycostonthebasisofcurrentvaluesandexpectationsatanarbitraryaccountinginterval，e。g。annually。
　　Businessmeninfactdifferastowhichcoursetheyadopt。Itmaybeconvenienttocalltheinitialexpectationofsupplementarycostwhentheequipmentisfirstacquiredthebasicsupplementarycost，andthesamequantityrecalculateduptodateonthebasisofcurrentvaluesandexpectationsthecurrentsupplementarycost。
　　Thuswecannotgetclosertoaquantitativedefinitionofsupplementarycostthanthatitcomprisesthosedeductionsfromhisincomewhichatypicalentrepreneurmakesbeforereckoningwhatheconsidershisnetincomeforthepurposeofdeclaringadividendinthecaseofacorporationorofdecidingthescaleofhiscurrentconsumptioninthecaseofanindividual。
　　Sincewindfallchargesoncapitalaccountarenotgoingtoberuledoutofthepicture，itisclearlybetter，incaseofdoubt，toassignanitemtocapitalaccount，andtoincludeinsupplementarycostonlywhatratherobviouslybelongsthere。Foranyoverloadingoftheformercanbecorrectedbyallowingitmoreinfluenceontherateofcurrentconsumptionthanitwouldotherwisehavehad。
　　ItwillbeseenthatourdefinitionofnetincomecomesveryclosetoMarshall'sdefinitionofincome，whenhedecidedtotakerefugeinthepracticesoftheIncomeTaxCommissionersand¾；broadlyspeakingtoregardasincomewhateverthey，withtheirexperience，choosetotreatassuch。Forthefabricoftheirdecisionscanberegardedastheresultofthemostcarefulandextensiveinvestigationwhichisavailable，tointerpretwhat，inpractice，itisusualtotreatasnetincome。ItalsocorrespondstothemoneyvalueofProfessorPigou'smostrecentdefinitionofthenationaldividend。[52]
　　[Page60]THEGENERAL
　　THEORYOFEMPLOYMENT
　　Itremainstrue，however，thatnetincome，beingbasedonanequivocalcriterionwhichdifferentauthoritiesmightinterpretdifferently，isnotperfectlyclear-cut。ProfessorHayek，forexample，hassuggestedthatanindividualownerofcapitalgoodsmightaimatkeepingtheincomehederivesfromhispossessionsconstant，sothathewouldnotfeelhimselffreetospendhisincomeonconsumptionuntilhehadsetasidesufficienttooffsetanytendencyofhisinvestment-incometodeclineforwhateverreason。[
　　53]Idoubtifsuchanindividualexists；but，obviously，notheoreticalobjectioncanberaisedagainstthisdeductionasprovidingapossiblepsychologicalcriterionofnetincome。ButwhenProfessorHayekinfersthattheconceptsofsavingandinvestmentsufferfromacorrespondingvagueness，heisonlyrightifhemeansnetsavingandnetinvestment。
　　Thesavingandtheinvestment，whicharerelevanttothetheoryofemployment，areclearofthisdefect，andarecapableofobjectivedefinition，aswehaveshownabove。
　　Thusitisamistaketoputalltheemphasisonnetincome，whichisonlyrelevanttodecisionsconcerningconsumption，andis，moreover，onlyseparatedfromvariousotherfactorsaffectingconsumptionbyanarrowline；andtooverlookashasbeenusualtheconceptofincomeproper，whichistheconceptrelevanttodecisionsconcerningcurrentproductionandisquiteunambiguous。
　　Theabovedefinitionsofincomeandofnetincomeareintendedtoconformascloselyaspossibletocommonusage。Itisnecessary，therefore，thatIshouldatonceremindthereaderthatinmyTreatiseonMoneyI
　　definedincomeinaspecialsense。Thepeculiarityinmyformerdefinitionrelatedtothatpartofaggregateincomewhichaccruestotheentrepreneurs，sinceItookneithertheprofitwhethergrossornetactuallyrealisedfromtheircurrentoperationsnortheprofitwhichtheyexpectedwhentheydecidedtounder-
　　[Page61]THEDEFINITION
　　OFINCOME
　　taketheircurrentoperations，butinsomesensenot，asInowthink，sufficientlydefinedifweallowforthepossibilityofchangesinthescaleofoutputanormalorequilibriumprofit；withtheresultthatonthisdefinitionsavingexceededinvestmentbytheamountoftheexcessofnormalprofitovertheactualprofit。Iamafraidthatthisuseoftermshascausedconsiderableconfusion，especiallyinthecaseofthecorrelativeuseofsaving；sinceconclusionsrelating，inparticular，totheexcessofsavingoverinvestment，whichwereonlyvalidifthetermsemployedwereinterpretedinmyspecialsense，havebeenfrequentlyadoptedinpopulardiscussionasthoughthetermswerebeingemployedintheirmorefamiliarsense。Forthisreason，andalsobecauseInolongerrequiremyformertermstoexpressmyideasaccurately，Ihavedecidedtodiscardthem¾；withmuchregretfortheconfusionwhichtheyhavecaused。
　　II。SavingandInvestmentAmidstthewelterofdivergentusagesofterms，itisagreeabletodiscoveronefixedpoint。SofarasIknow，everyoneisagreedthatsavingmeanstheexcessofincomeoverexpenditureonconsumption。Thusanydoubtsaboutthemeaningofsavingmustarisefromdoubtsaboutthemeaningeitherofincomeorofconsumption。Incomewehavedefinedabove。
　　Expenditureonconsumptionduringanyperiodmustmeanthevalueofgoodssoldtoconsumersduringthatperiod，whichthrowsusbacktothequestionofwhatismeantbyaconsumer-purchaser。Anyreasonabledefinitionofthelinebetweenconsumer-purchasersandinvestor-purchaserswillserveusequallywell，providedthatitisconsistentlyapplied。Suchproblemasthereis，e。g。whetheritisrighttoregardthepurchaseofamotor-carasaconsumer-purchaseandthepurchaseofahouseasaninvestor-purchase，hasbeenfrequentlydiscussedandIhavenothingmaterialtoaddtothediscussion。
　　[Page62]THEGENERAL
　　THEORYOFEMPLOYMENT
　　Thecriterionmustobviouslycorrespondtowherewedrawthelinebetweentheconsumerandtheentrepreneur。ThuswhenwehavedefinedA1asthevalueofwhatoneentrepreneurhaspurchasedfromanother，wehaveimplicitlysettledthequestion。ItfollowsthatexpenditureonconsumptioncanbeunambiguouslydefinedasSA-A1，whereAisthetotalsalesmadeduringtheperiodandA1isthetotalsalesmadebyoneentrepreneurtoanother。Inwhatfollowsitwillbeconvenient，asarule，toomitandwriteAfortheaggregatesalesofallkinds，A1fortheaggregatesalesfromoneentrepreneurtoanotherandUfortheaggregateusercostsoftheentrepreneurs。
　　Havingnowdefinedbothincomeandconsumption，thedefinitionofsaving，whichistheexcessofincomeoverconsumption，naturallyfollows。SinceincomeisequaltoA-UandconsumptionisequaltoA-A1，itfollowsthatsavingisequaltoA1-U。Similarly，wehavenetsavingfortheexcessofnetincomeoverconsumption，equaltoA1-U-V。
　　Ourdefinitionofincomealsoleadsatoncetothedefinitionofcurrentinvestment。Forwemustmeanbythisthecurrentadditiontothevalueofthecapitalequipmentwhichhasresultedfromtheproductiveactivityoftheperiod。Thisis，clearly，equaltowhatwehavejustdefinedassaving。Foritisthatpartoftheincomeoftheperiodwhichhasnotpassedintoconsumption。WehaveseenabovethatastheresultoftheproductionofanyperiodentrepreneursendupwithhavingsoldfinishedoutputhavingavalueAandwithacapitalequipmentwhichhassufferedadeteriorationmeasuredbyUoranimprovementmeasuredby-UwhereUisnegativeasaresultofhavingproducedandpartedwithA，afterallowingforpurchasesA1fromotherentrepreneurs。DuringthesameperiodfinishedoutputhavingavalueA-A1willhavepassedintoconsumption。TheexcessofA-UoverA-A1，namelyA1-U，istheadditiontocapital[Page63]THEDEFINITION
　　OFINCOME
　　equipmentasaresultoftheproductiveactivitiesoftheperiodandis，therefore，theinvestmentoftheperiod。SimilarlyA1-U-V；
　　whichisthenetadditiontocapitalequipment，afterallowingfornormalimpairmentinthevalueofcapitalapartfromitsbeingusedandapartfromwindfallchangesinthevalueoftheequipmentchargeabletocapitalaccount，isthenetinvestmentoftheperiod。
　　Whilst，therefore，theamountofsavingisanoutcomeofthecollectivebehaviourofindividualconsumersandtheamountofinvestmentofthecollectivebehaviourofindividualentrepreneurs，thesetwoamountsarenecessarilyequal，sinceeachofthemisequaltotheexcessofincomeoverconsumption。
　　Moreover，thisconclusioninnowaydependsonanysubtletiesorpeculiaritiesinthedefinitionofincomegivenabove。Provideditisagreedthatincomeisequaltothevalueofcurrentoutput，thatcurrentinvestmentisequaltothevalueofthatpartofcurrentoutputwhichisnotconsumed，andthatsavingisequaltotheexcessofincomeoverconsumption¾；allofwhichisconformablebothtocommonsenseandtothetraditionalusageofthegreatmajorityofeconomists¾；theequalityofsavingandinvestmentnecessarilyfollows。Inshort-Income=valueofoutput=consumptioninvestment。Saving=income-consumption。Thereforesaving=investment。Thusanysetofdefinitionswhichsatisfytheaboveconditionsleadstothesameconclusion。Itisonlybydenyingthevalidityofoneorotherofthemthattheconclusioncanbeavoided。
　　Theequivalencebetweenthequantityofsavingandthequantityofinvestmentemergesfromthebilateralcharacterofthetransactionsbetweentheproducerontheonehandand，ontheotherhand，theconsumerorthepurchaserofcapitalequipment。
　　[Page64]THEGENERAL
　　THEORYOFEMPLOYMENT
　　Incomeiscreatedbythevalueinexcessofusercostwhichtheproducerobtainsfortheoutputhehassold；butthewholeofthisoutputmustobviouslyhavebeensoldeithertoaconsumerortoanotherentrepreneur；andeachentrepreneur'scurrentinvestmentisequaltotheexcessoftheequipmentwhichhehaspurchasedfromotherentrepreneursoverhisownusercost。
　　Hence，intheaggregatetheexcessofincomeoverconsumption，whichwecallsaving，cannotdifferfromtheadditiontocapitalequipmentwhichwecallinvestment。Andsimilarlywithnetsavingandnetinvestment。Saving，infact，isamereresidual。Thedecisionstoconsumeandthedecisionstoinvestbetweenthemdetermineincomes。Assumingthatthedecisionstoinvestbecomeeffective，theymustindoingsoeithercurtailconsumptionorexpandincome。Thustheactofinvestmentinitselfcannothelpcausingtheresidualormargin，whichwecallsaving，toincreasebyacorrespondingamount。
　　Itmightbe，ofcourse，thatindividualsweresotê；temonté；eintheirdecisionsastohowmuchtheythemselveswouldsaveandinvestrespectively，thattherewouldbenopointofpriceequilibriumatwhichtransactionscouldtakeplace。Inthiscaseourtermswouldceasetobeapplicable，sinceoutputwouldnolongerhaveadefinitemarketvalue，priceswouldfindnoresting-placebetweenzeroandinfinity。Experienceshows，however，thatthis，infact，isnotso；andthattherearehabitsofpsychologicalresponsewhichallowofanequilibriumbeingreachedatwhichthereadinesstobuyisequaltothereadinesstosell。Thatthereshouldbesuchathingasamarketvalueforoutputis，atthesametime，anecessaryconditionformoney-incometopossessadefinitevalueandasufficientconditionfortheaggregateamountwhichsavingindividualsdecidetosavetobeequaltotheaggregateamountwhichinvestingindividualsdecidetoinvest。
　　Clearnessofmindonthismatterisbestreached，[Page65]THEDEFINITION
　　OFINCOME
　　perhaps，bythinkingintermsofdecisionstoconsumeortorefrainfromconsumingratherthanofdecisionstosave。Adecisiontoconsumeornottoconsumetrulylieswithinthepoweroftheindividual；sodoesadecisiontoinvestornottoinvest。Theamountsofaggregateincomeandofaggregatesavingaretheresultsofthefreechoicesofindividualswhetherornottoconsumeandwhetherornottoinvest；buttheyareneitherofthemcapableofassuminganindependentvalueresultingfromaseparatesetofdecisionstakenirrespectiveofthedecisionsconcerningconsumptionandinvestment。Inaccordancewiththisprinciple，theconceptionofthepropensitytoconsumewill，inwhatfollows，taketheplaceofthepropensityordispositiontosave。
　　AppendixonUserCost[Page66]
　　AppendixonUserCostI
　　Usercosthas，Ithink，animportancefortheclassicaltheoryofvaluewhichhasbeenoverlooked。Thereismoretobesaidaboutitthanwouldberelevantorappropriateinthisplace。But，asadigression，wewillexamineitsomewhatfurtherinthisappendix。
　　Anentrepreneur'susercostisbydefinitionequaltoA1G'-B'-G，whereA1istheamountofourentrepreneur'spurchasesfromotherentrepreneurs，Gtheactualvalueofhiscapitalequipmentattheendoftheperiod，andG'thevalueitmighthavehadattheendoftheperiodifhehadrefrainedfromusingitandhadspenttheoptimumsumB'onitsmaintenanceandimprovement。NowG-G'-B'，namelytheincrementinthevalueoftheentrepreneur'sequipmentbeyondthenetvaluewhichhehasinheritedfromthepreviousperiod，representstheentrepreneur'scurrentinvestmentinhisequipmentandcanbewrittenI。ThusU，theusercostofhissales-turnoverA，isequaltoA1-IwhereA1iswhathehasboughtfromotherentrepreneursandIiswhathehascurrentlyinvestedinhisownequipment。A
　　littlereflectionwillshowthatallthisisnomorethancommonsense。
　　Somepartofhisoutgoingstootherentrepreneursisbalancedbythevalueofhiscurrentinvestmentinhisownequipment，andtherestrepresentsthesacrificewhichtheoutputhehassoldmusthavecosthimoverandabovethetotalsumwhichhehaspaidouttothefactorsofproduction。
　　Ifthereadertriestoexpressthesubstanceofthisotherwise，hewillfindthatitsadvantageliesinitsavoidanceofinsolubleandunnecessary
　　accountingproblems。Thereis，Ithink，nootherwayofanalysingthecurrentproceedsofproductionunambiguously。Ifindustryiscompletelyintegratedoriftheentrepreneurhasboughtnothingfromoutside，sothatA1
　　=0，the[Page67]APPENDIX
　　ONUSERCOST
　　usercostissimplytheequivalentofthecurrentdisinvestmentinvolvedinusingtheequipment；butwearestillleftwiththeadvantagethatwedonotrequireatanystageoftheanalysistoallocatethefactorcostbetweenthegoodswhicharesoldandtheequipmentwhichisretained。Thuswecanregardtheemploymentgivenbyafirm，whetherintegratedorindividual，asdependingonasingleconsolidateddecision¾；aprocedurewhichcorrespondstotheactualinterlockingcharacteroftheproductionofwhatiscurrentlysoldwithtotalproduction。
　　Theconceptofusercostenablesus，moreover，togiveaclearerdefinitionthanthatusuallyadoptedoftheshort-periodsupplypriceofaunitofafirm'ssaleableoutput。Fortheshort-periodsupplypriceisthesumofthemarginalfactorcostandthemarginalusercost。
　　Nowinthemoderntheoryofvalueithasbeenausualpracticetoequatetheshort-periodsupplypricetothemarginalfactorcostalone。Itisobvious，however，thatthisisonlylegitimateifmarginalusercostiszeroorifsupplypriceisspeciallydefinedsoastobenetofmarginalusercost，justasIhavedefinedp。24above'proceeds'and'aggregatesupplyprice'asbeingnetofaggregateusercost。But，whereasitmaybeoccasionallyconvenientindealingwithoutputasawholetodeductusercost，thisproceduredeprivesouranalysisofallrealityifitishabituallyandtacitlyappliedtotheoutputofasingleindustryorfirm，sinceitdivorcesthe'supplyprice'ofanarticlefromanyordinarysenseofits'price'；andsomeconfusionmayhaveresultedfromthepracticeofdoingso。Itseemstohavebeenassumedthat'supplyprice'hasanobviousmeaningasappliedtoaunitofthesaleableoutputofanindividualfirm，andthematterhasnotbeendeemedtorequirediscussion。
　　Yetthetreatmentbothofwhatispurchasedfromotherfirmsandofthewastageofthefirm'sownequipmentasaconsequenceofproducingthemarginaloutputinvolvesthewholepackofperplexitieswhichattendthedefinitionofincome。For，evenifweassumethatthemarginalcostofpurchasesfromotherfirmsinvolvedinsellinganadditionalunitofoutputhastobedeductedfromthesale-proceedsperunitinordertogiveuswhatwemeanbyourfirm'ssupplyprice，westillhavetoallowforthemarginaldisinvestmentinthefirm'sownequipmentinvolvedinproducingthemarginaloutput。
　　Evenifallproductioniscarriedonbyacompletelyintegratedfirm，itisstillillegitimatetosupposethatthemarginalusercostiszero，i。e。
　　thatthemarginaldisinvestmentinequipmentduetotheproductionofthemarginaloutputcangenerallybeneglected。
　　[Page68]THEGENERAL
　　THEORYOFEMPLOYMENT
　　Theconceptsofusercostandofsupplementarycostalsoenableustoestablishaclearerrelationshipbetweenlong-periodsupplypriceandshort-periodsupplyprice。Long-periodcostmustobviouslyincludeanamounttocoverthebasicsupplementarycostaswellastheexpectedprimecostappropriatelyaveragedoverthelifeoftheequipment。Thatistosay，thelong-periodcostoftheoutputisequaltotheexpectedsumoftheprimecostandthesupplementarycost；and，furthermore，inordertoyieldanormalprofit，thelong-periodsupplypricemustexceedthelong-periodcostthuscalculatedbyanamountdeterminedbythecurrentrateofinterestonloansofcomparabletermandrisk，reckonedasapercentageofthecostoftheequipment。Orifweprefertotakeastandard'pure'rateofinterest，wemustincludeinthelong-periodcostathirdtermwhichwemightcalltherisk-costtocovertheunknownpossibilitiesoftheactualyielddifferingfromtheexpectedyield。Thusthelong-periodsupplypriceisequaltothesumoftheprimecost，thesupplementarycost，theriskcostandtheinterestcost，intowhichseveralcomponentsitcanbeanalysed。Theshort-periodsupplyprice，ontheotherhand，isequaltothemarginalprimecost。Theentrepreneurmust，therefore，expect，whenhebuysorconstructshisequipment，tocoverhissupplementarycost，hisriskcostandhisinterestcostoutoftheexcessofthemarginalvalueoftheprimecostoveritsaveragevalue；sothatinlong-periodequilibriumtheexcessofthemarginalprimecostovertheaverageprimecostisequaltothesumofthesupplementary，riskandinterestcosts。[54]
　　Thelevelofoutput，atwhichmarginalprimecostisexactlyequaltothesumoftheaverageprimeandsupplementarycosts，hasaspecialimportance，becauseitisthepointatwhichtheentrepreneur'stradingaccountbreakseven。Itcorresponds，[Page69]APPENDIX
　　ONUSERCOST
　　thatistosay，tothepointofzeronetprofit；whilstwithasmalleroutputthanthisheistradingatanetloss。
　　Theextenttowhichthesupplementarycosthastobeprovidedforapartfromtheprimecostvariesverymuchfromonetypeofequipmenttoanother。
　　Twoextremecasesarethefollowing:
　　iSomepartofthemaintenanceoftheequipmentmustnecessarilytakeplaceparipassuwiththeactofusingite。g。oilingthemachine。Theexpenseofthisapartfromoutsidepurchasesisincludedinthefactorcost。If，forphysicalreasons，theexactamountofthewholeofthecurrentdepreciationhasnecessarilytobemadegoodinthisway，theamountoftheusercostapartfromoutsidepurchaseswouldbeequalandoppositetothatofthesupplementarycost；andinlong-periodequilibriumthemarginalfactorcostwouldexceedtheaveragefactorcostbyanamountequaltotheriskandinterestcost。
　　iiSomepartofthedeteriorationinthevalueoftheequipmentonlyoccursifitisused。Thecostofthisischargedinusercost，insofarasitisnotmadegoodparipassuwiththeactofusingit。
　　Iflossinthevalueoftheequipmentcouldonlyoccurinthisway，supplementarycostwouldbezero。
　　Itmaybeworthpointingoutthatanentrepreneurdoesnotusehisoldestandworstequipmentfirst，merelybecauseitsusercostislow；sinceitslowusercostmaybeoutweighedbyitsrelativeinefficiency，i。e。byitshighfactorcost。Thusanentrepreneurusesbypreferencethatpartofhisequipmentforwhichtheusercostplusfactorcostisleastperunitofoutput。[55]Itfollowsthatforanygivenvolumeofoutputoftheproductinquestionthereisacorrespondingusercost，[56]butthatthistotalusercostdoesnotbearauniformrelationtothemarginalusercost，i。e。totheincrementofusercostduetoanincrementintherateofoutput。
　　II
　　Usercostconstitutesoneofthelinksbetweenthepresentandthefuture。
　　Forindecidinghisscaleofproductionan[Page70]THEGENERAL
　　THEORYOFEMPLOYMENT
　　entrepreneurhastoexerciseachoicebetweenusinguphisequipmentnowandpreservingittobeusedlateron。Itistheexpectedsacrificeoffuturebenefitinvolvedinpresentusewhichdeterminestheamountoftheusercost，anditisthemarginalamountofthissacrificewhich，togetherwiththemarginalfactorcostandtheexpectationofthemarginalproceeds，determineshisscaleofproduction。How，then，istheusercostofanactofproductioncalculatedbytheentrepreneur？
　　Wehavedefinedtheusercostasthereductioninthevalueoftheequipmentduetousingitascomparedwithnotusingit，afterallowingforthecostofthemaintenanceandimprovementswhichitwouldbeworthwhiletoundertakeandforpurchasesfromotherentrepreneurs。Itmustbearrivedat，therefore，bycalculatingthediscountedvalueoftheadditionalprospectiveyieldwhichwouldbeobtainedatsomelaterdateifitwerenotusednow。Nowthismustbeatleastequaltothepresentvalueoftheopportunitytopostponereplacementwhichwillresultfromlayinguptheequipment；anditmaybemore。[57]
　　Ifthereisnosurplusorredundantstock，sothatmoreunitsofsimilarequipmentarebeingnewlyproducedeveryyeareitherasanadditionorinreplacement，itisevidentthatmarginalusercostwillbecalculablebyreferencetotheamountbywhichthelifeorefficiencyoftheequipmentwillbeshortenedifitisused，andthecurrentreplacementcost。If，however，thereisredundantequipment，thentheusercostwillalsodependontherateofinterestandthecurrenti。e。re-estimatedsupplementarycostovertheperiodoftimebeforetheredundancyisexpectedtobeabsorbedthroughwastage，etc。Inthiswayinterestcostandcurrentsupplementarycostenterindirectlyintothecalculationofusercost。
　　Thecalculationisexhibitedinitssimplestandmostintelligibleformwhenthefactorcostiszero，e。g。inthecaseofaredundantstockofarawmaterialsuchascopper，onthelineswhichIhaveworkedoutinmyTreatiseonMoney，vol。ii。chap。29。Letustaketheprospectivevaluesofcopperatvariousfuturedates，aserieswhichwillbegovernedbytherateatwhichredundancyisbeingabsorbedandgraduallyapproachestheestimatednormalcost。Thepresentvalueorusercostofatonofsurpluscopperwillthenbeequaltothegreatestofthevaluesobtainablebysubtractingfromtheestimatedfuturevalueatany[Page71]APPENDIX
　　ONUSERCOST
　　givendateofatonofcoppertheinterestcostandthecurrentsupplementarycostonatonofcopperbetweenthatdateandthepresent。
　　Inthesamewaytheusercostofashiporfactoryormachine，whentheseequipmentsareinredundantsupply，isitsestimatedreplacementcostdiscountedatthepercentagerateofitsinterestandcurrentsupplementarycoststotheprospectivedateofabsorptionoftheredundancy。
　　Wehaveassumedabovethattheequipmentwillbereplacedinduecoursebyanidenticalarticle。Iftheequipmentinquestionwillnotberenewedidenticallywhenitiswornout，thenitsusercosthastobecalculatedbytakingaproportionoftheusercostofthenewequipment，whichwillbeerectedtodoitsworkwhenitisdiscarded，givenbyitscomparativeefficiency。
　　III
　　Thereadershouldnoticethat，wheretheequipmentisnotobsolescentbutmerelyredundantforthetimebeing，thedifferencebetweentheactualusercostanditsnormalvaluei。e。thevaluewhenthereisnoredundantequipmentvarieswiththeintervaloftimewhichisexpectedtoelapsebeforetheredundancyisabsorbed。Thusifthetypeofequipmentinquestionisofallagesandnot'bunched'，sothatafairproportionisreachingtheendofitslifeannually，themarginalusercostwillnotdeclinegreatlyunlesstheredundancyisexceptionallyexcessive。Inthecaseofageneralslump，marginalusercostwilldependonhowlongentrepreneursexpecttheslumptolast。Thustheriseinthesupplypricewhenaffairsbegintomendmaybepartlyduetoasharpincreaseinmarginalusercostduetoarevisionoftheirexpectations。
　　Ithassometimesbeenargued，contrarytotheopinionofbusinessmen，thatorganisedschemesforscrappingredundantplantcannothavethedesiredeffectofraisingpricesunlesstheyapplytothewholeoftheredundantplant。Buttheconceptofusercostshowshowthescrappingofsayhalftheredundantplantmayhavetheeffectofraisingpricesimmediately。
　　Forbybringingthedateoftheabsorptionoftheredundancynearer，thispolicyraisesmarginalusercostandconsequentlyincreasesthecurrentsupplyprice。Thusbusinessmenwouldseemtohavethenotionofusercostimplicitlyinmind，thoughtheydonotformulateitdistinctly。Ifthesupplementarycostisheavy，itfollowsthatthemarginalusercostwillbelowwhenthereissurplusequipment。Moreover，[Page72]THEGENERAL
　　THEORYOFEMPLOYMENT
　　whenthereissurplusequipment，themarginalfactorandusercostsareunlikelytobemuchinexcessoftheiraveragevalue。Ifboththeseconditionsarefulfilled，theexistenceofsurplusequipmentislikelytoleadtotheentrepreneur'sworkingatanetloss，andperhapsataheavynetloss。Therewillnotbeasuddentransitionfromthisstateofaffairstoanormalprofit，takingplaceatthemomentwhentheredundancyisabsorbed。
　　Astheredundancybecomesless，theusercostwillgraduallyincrease；
　　andtheexcessofmarginaloveraveragefactorandusercostmayalsograduallyincrease。
　　IV
　　InMarshall'sPrinciplesofEconomics6thed。p。360apartofusercostisincludedinprimecostundertheheadingof'extrawear-and-tearofplant'。Butnoguidanceisgivenastohowthisitemistobecalculatedorastoitsimportance。InhisTheoryofUnemploymentp。42ProfessorPigouexpresslyassumesthatthemarginaldisinvestmentinequipmentduetothemarginaloutputcan，ingeneral，beneglected:'Thedifferencesinthequantityofwear-and-tearsufferedbyequipmentandinthecostsofnon-manuallabouremployed，thatareassociatedwithdifferencesinoutput，areignored，asbeing，ingeneral，ofsecondaryimportance'。[
　　58]Indeed，thenotionthatthedisinvestmentinequipmentiszeroatthemarginofproductionrunsthroughagooddealofrecenteconomictheory。
　　Butthewholeproblemisbroughttoanobviousheadassoonasitisthoughtnecessarytoexplainexactlywhatismeantbythesupplypriceofanindividualfirm。
　　Itistruethatthecostofmaintenanceofidleplantmayoften，forthereasonsgivenabove，reducethemagnitudeofmarginalusercost，especiallyinaslumpwhichisexpectedtolastalongtime。Neverthelessaverylowusercostatthemarginisnotacharacteristicoftheshortperiodassuch，butofparticularsituationsandtypesofequipmentwherethecostofmaintainingidleplanthappenstobeheavy，andofthosedisequilibriawhicharecharacterisedbyveryrapidobsolescenceorgreatredundancy，especiallyifitiscoupledwithalargeproportionofcomparativelynewplant。
　　Inthecaseofrawmaterialsthenecessityofallowingforusercostisobvious；¾；ifatonofcopperisusedupto-dayitcannotbe[Page73]APPENDIX
　　ONUSERCOST
　　usedto-morrow，andthevaluewhichthecopperwouldhaveforthepurposesofto-morrowmustclearlyhereckonedasapartofthemarginalcost。Butthefacthasbeenoverlookedthatcopperisonlyanextremecaseofwhatoccurswhenevercapitalequipmentisusedtoproduce。Theassumptionthatthereisasharpdivisionbetweenrawmaterialswherewemustallowforthedisinvestmentduetousingthemandfixedcapitalwherewecansafelyneglectitdoesnotcorrespondtothefacts；¾；especiallyinnormalconditionswhereequipmentisfallingdueforreplacementeveryyearandtheuseofequipmentbringsnearerthedateatwhichreplacementisnecessary。
　　Itisanadvantageoftheconceptsofusercostandsupplementarycostthattheyareasapplicabletoworkingandliquidcapitalastofixedcapital。
　　Theessentialdifferencebetweenrawmaterialsandfixedcapitalliesnotintheirliabilitytouserandsupplementarycosts，butinthefactthatthereturntoliquidcapitalconsistsofasingleterm；whereasinthecaseoffixedcapital，whichisdurableandusedupgradually，thereturnconsistsofaseriesofusercostsandprofitsearnedinsuccessiveperiods。
　　TheMeaningofSavingandInvestment[Page74]
　　Chapter7
　　THEMEANINGOFSAVINGANDINVESTMENT
　　FURTHERCONSIDERED
　　I
　　Inthepreviouschaptersavingandinvestmenthavebeensodefinedthattheyarenecessarilyequalinamount，being，forthecommunityasawhole，merelydifferentaspectsofthesamething。SeveralcontemporarywritersincludingmyselfinmyTreatiseonMoneyhave，however，givenspecialdefinitionsofthesetermsonwhichtheyarenotnecessarilyequal。
　　Othershavewrittenontheassumptionthattheymaybeunequalwithoutprefacingtheirdiscussionwithanydefinitionsatall。Itwillbeuseful，therefore，withaviewtorelatingtheforegoingtootherdiscussionsoftheseterms，toclassifysomeofthevarioususesofthemwhichappeartobecurrent。
　　SofarasIknow，everyoneagreesinmeaningbysavingtheexcessofincomeoverwhatisspentonconsumption。Itwouldcertainlybeveryinconvenientandmisleadingnottomeanthis。Noristhereanyimportantdifferenceofopinionastowhatismeantbyexpenditureonconsumption。Thusthedifferencesofusageariseeitheroutofthedefinitionofinvestmentoroutofthatofincome。
　　II
　　Letustakeinvestmentfirst。Inpopularusageitiscommontomeanbythisthepurchaseofanasset，old[Page75]THEMEANING
　　OFSAVINGANDINVESTMENT
　　ornew，byanindividualoracorporation。Occasionally，thetermmightberestrictedtothepurchaseofanassetontheStockExchange。Butwespeakjustasreadilyofinvesting，forexample，inahouse，orinamachine，orinastockoffinishedorunfinishedgoods；and，broadlyspeaking，newinvestment，asdistinguishedfromreinvestment，meansthepurchaseofacapitalassetofanykindoutofincome。Ifwereckonthesaleofaninvestmentasbeingnegativeinvestment，i。e。disinvestment，myowndefinitionisinaccordancewithpopularusage；sinceexchangesofoldinvestmentsnecessarilycancelout。Wehave，indeed，toadjustforthecreationanddischargeofdebtsincludingchangesinthequantityofcreditormoney；butsinceforthecommunityasawholetheincreaseordecreaseoftheaggregatecreditorpositionisalwaysexactlyequaltotheincreaseordecreaseoftheaggregatedebtorposition，thiscomplicationalsocancelsoutwhenwearedealingwithaggregateinvestment。Thus，assumingthatincomeinthepopularsensecorrespondstomynetincome，aggregateinvestmentinthepopularsensecoincideswithmydefinitionofnetinvestment，namelythenetadditiontoallkindsofcapitalequipment，afterallowingforthosechangesinthevalueoftheoldcapitalequipmentwhicharetakenintoaccountinreckoningnetincome。
　　Investment，thusdefined，includes，therefore，theincrementofcapitalequipment，whetheritconsistsoffixedcapital，workingcapitalorliquidcapital；andthesignificantdifferencesofdefinitionapartfromthedistinctionbetweeninvestmentandnetinvestmentareduetotheexclusionfrominvestmentofoneormoreofthesecategories。
　　MrHawtrey，forexample，whoattachesgreatimportancetochangesinliquidcapital，i。e。toundesignedincrementsordecrementsinthestockofunsoldgoods，hassuggestedapossibledefinitionof[Page76]THEGENERAL
　　THEORYOFEMPLOYMENT
　　investmentfromwhichsuchchangesareexcluded。Inthiscaseanexcessofsavingoverinvestmentwouldbethesamethingasanundesignedincrementinthestockofunsoldgoods，i。e。asanincreaseofliquidcapital。MrHawtreyhasnotconvincedmethatthisisthefactortostress；foritlaysalltheemphasisonthecorrectionofchangeswhichwereinthefirstinstanceunforeseen，ascomparedwiththosewhichare，rightlyorwrongly，anticipated。MrHawtreyregardsthedailydecisionsofentrepreneursconcerningtheirscaleofoutputasbeingvariedfromthescaleofthepreviousdaybyreferencetothechangesintheirstockofunsoldgoods。Certainly，inthecaseofconsumptiongoods，thisplaysanimportantpartintheirdecisions。ButIseenoobjectinexcludingtheplayofotherfactorsontheirdecisions；andIprefer，therefore，toemphasisethetotalchangeofeffectivedemandandnotmerelythatpartofthechangeineffectivedemandwhichreflectstheincreaseordecreaseofunsoldstocksinthepreviousperiod。Moreover，inthecaseoffixedcapital，theincreaseordecreaseofunusedcapacitycorrespondstotheincreaseordecreaseinunsoldstocksinitseffectondecisionstoproduce；andIdonotseehowMrHawtrey'smethodcanhandlethisatleastequallyimportantfactor。
　　Itseemsprobablethatcapitalformationandcapitalconsumption，asusedbytheAustrianschoolofeconomists，arenotidenticaleitherwithinvestmentanddisinvestmentasdefinedaboveorwithnetinvestmentanddisinvestment。Inparticular，capitalconsumptionissaidtooccurincircumstanceswherethereisquiteclearlynonetdecreaseincapitalequipmentasdefinedabove。Ihave，however，beenunabletodiscoverareferencetoanypassagewherethemeaningofthesetermsisclearlyexplained。Thestatement，forexample，thatcapitalformationoccurswhenthereisalengtheningoftheperiodofproductiondoesnotmuchadvancematters。
　　[Page77]THEMEANING
　　OFSAVINGANDINVESTMENT
　　III
　　Wecomenexttothedivergencesbetweensavingandinvestmentwhichareduetoaspecialdefinitionofincomeandhenceoftheexcessofincomeoverconsumption。MyownuseoftermsinmyTreatiseonMoneyisanexampleofthis。For，asIhaveexplainedonp。60above，thedefinitionofincome，whichIthereemployed，differedfrommypresentdefinitionbyreckoningastheincomeofentrepreneursnottheiractuallyrealisedprofitsbutinsomesensetheir'normalprofit'。ThusbyanexcessofsavingoverinvestmentImeantthatthescaleofoutputwassuchthatentrepreneurswereearningalessthannormalprofitfromtheirownershipofthecapitalequipment；andbyanincreasedexcessofsavingoverinvestmentImeantthatadeclinewastakingplaceintheactualprofits，sothattheywouldbeunderamotivetocontractoutput。
　　AsInowthink，thevolumeofemploymentandconsequentlyofoutputandrealincomeisfixedbytheentrepreneurunderthemotiveofseekingtomaximisehispresentandprospectiveprofitstheallowanceforusercostbeingdeterminedbyhisviewastotheuseofequipmentwhichwillmaximisehisreturnfromitoveritswholelife；whilstthevolumeofemploymentwhichwillmaximisehisprofitdependsontheaggregatedemandfunctiongivenbyhisexpectationsofthesumoftheproceedsresultingfromconsumptionandinvestmentrespectivelyonvarioushypotheses。InmyTreatiseonMoneytheconceptofchangesintheexcessofinvestmentoversaving，astheredefined，wasawayofhandlingchangesinprofit，thoughIdidnotinthatbookdistinguishclearlybetweenexpectedandrealisedresults。[59]Itherearguedthatchangeintheexcessof[Page78]THEGENERAL
　　THEORYOFEMPLOYMENT
　　investmentoversavingwasthemotiveforcegoverningchangesinthevolumeofoutput。Thusthenewargument，thoughasInowthinkmuchmoreaccurateandinstructive，isessentiallyadevelopmentoftheold。ExpressedinthelanguageofmyTreatiseonMoney，itwouldrun:theexpectationofanincreasedexcessofinvestmentoversaving，giventheformervolumeofemploymentandoutput，willinduceentrepreneurstoincreasethevolumeofemploymentandoutput。Thesignificanceofbothmypresentandmyformerargumentsliesintheirattempttoshowthatthevolumeofemploymentisdeterminedbytheestimatesofeffectivedemandmadebytheentrepreneurs，anexpectedincreaseofinvestmentrelativelytosavingasdefinedinmyTreatiseonMoneybeingacriterionofanincreaseineffectivedemand。ButtheexpositioninmyTreatiseonMoneyis，ofcourse，veryconfusingandincompleteinthelightofthefurtherdevelopmentsheresetforth。
　　MrD。H。Robertsonhasdefinedto-day'sincomeasbeingequaltoyesterday'sconsumptionplusinvestment，sothatto-day'ssaving，inhissense，isequaltoyesterday'sinvestmentplustheexcessofyesterday'sconsumptionoverto-day'sconsumption。Onthisdefinitionsavingcanexceedinvestment，namely，bytheexcessofyesterday'sincomeinmysenseoverto-day'sincome。ThuswhenMrRobertsonsaysthatthereisanexcessofsavingoverinvestment，hemeansliterallythesamethingasImeanwhenIsaythatincomeisfalling，andtheexcessofsavinginhissenseisexactlyequaltothedeclineofincomeinmysense。Ifitweretruethatcurrentexpectationswerealwaysdeterminedbyyesterday'srealisedresults，to-day'seffectivedemandwouldbeequaltoyesterday'sincome。ThusMrRobertson'smethodmightberegardedasanalternative。attempttominebeing，perhaps，afirstapproximationtoittomakethesamedistinction，sovitalforcausalanalysis，thatIhavetried[Page79]THEMEANING
　　OFSAVINGANDINVESTMENT
　　tomakebythecontrastbetweeneffectivedemandandincome。[60]
　　IV
　　Wecomenexttothemuchvaguerideasassociatedwiththephrase'forcedsaving'。Isanyclearsignificancediscoverableinthese？InmyTreatiseonMoneyvol。1，p。171，footnote[JMK，vol。V，p。154]Igavesomereferencestoearlierusesofthisphraseandsuggestedthattheyboresomeaffinitytothedifferencebetweeninvestmentand'saving'inthesenseinwhichIthereusedthelatterterm。IamnolongerconfidentthattherewasinfactsomuchaffinityasIthensupposed。Inanycase，Ifeelsurethat'forcedsaving'andanalogousphrasesemployedmorerecentlye。g。byProfessorHayekorProfessorRobbinshavenodefiniterelationtothedifferencebetweeninvestmentand'saving'inthesenseintendedinmyTreatiseonMoney。Forwhilsttheseauthorshavenotexplainedexactlywhattheymeanbythisterm，itisclearthat'forcedsaving'，intheirsense，isaphenomenonwhichresultsdirectlyfrom，andismeasuredby，changesinthequantityofmoneyorbank-credit。
　　Itisevidentthatachangeinthevolumeofoutputandemploymentwill，indeed，causeachangeinincomemeasuredinwage-units；thatachangeinthewage-unitwillcausebotharedistributionofincomebetweenborrowersandlendersandachangeinaggregateincomemeasuredinmoney；andthatineithereventtherewillormaybeachangeintheamountsaved。Since，therefore，changesinthequantityofmoneymayresult，throughtheireffectontherateofinterest，inachangeinthevolumeanddistributionofincomeasweshallshowlater，suchchangesmayinvolve，indirectly，achangeintheamountsaved。Butsuch[Page80]THEGENERAL
　　THEORYOFEMPLOYMENT
　　changesintheamountssavedarenomore'forcedsavings'thananyotherchangesintheamountssavedduetoachangeincircumstances；andthereisnomeansofdistinguishingbetweenonecaseandanother，unlesswespecifytheamountsavedincertaingivenconditionsasournormorstandard。Moreover，asweshallsee，theamountofthechangeinaggregatesavingwhichresultsfromagivenchangeinthequantityofmoneyishighlyvariableanddependsonmanyotherfactors。
　　Thus'forcedsaving'hasnomeaninguntilwehavespecifiedsomestandardrateofsaving。Ifweselectasmightbereasonabletherateofsavingwhichcorrespondstoanestablishedstateoffullemployment，theabovedefinitionwouldbecome:'Forcedsavingistheexcessofactualsavingoverwhatwouldbesavediftherewerefullemploymentinapositionoflong-periodequilibrium'。Thisdefinitionwouldmakegoodsense，butasenseinwhichaforcedexcessofsavingwouldbeaveryrareandaveryunstablephenomenon，andaforceddeficiencyofsavingtheusualstateofaffairs。
　　ProfessorHayek'sinteresting'NoteontheDevelopmentoftheDoctrineofForcedSaving'[61]showsthatthiswasinfacttheoriginalmeaningoftheterm。'Forcedsaving'or'forcedfrugality'was，inthefirstinstance，aconceptionofBentham's；andBenthamexpresslystatedthathehadinmindtheconsequencesofanincreaseinthequantityofmoneyrelativelytothequantityofthingsvendibleformoneyincircumstancesof'allhandsbeingemployedandemployedinthemostadvantageousmanner'。
　　[62]Insuchcircumstances，Benthampointsout，realincomecannotbeincreased，and，consequently，additionalinvestment，takingplaceasaresultofthetransition，involvesforcedfrugality'attheexpenseofnationalcomfortandnationaljustice'。Allthenineteenth-century[Page81]THEMEANING
　　OFSAVINGANDINVESTMENT
　　writerswhodealtwiththismatterhadvirtuallythesameideainmind。
　　Butanattempttoextendthisperfectlyclearnotiontoconditionsoflessthanfullemploymentinvolvesdifficulties。Itistrue，ofcourseowingtothefactofdiminishingreturnstoanincreaseintheemploymentappliedtoagivencapitalequipment，thatanyincreaseinemploymentinvolvessomesacrificeofrealincometothosewhowerealreadyemployed，butanattempttorelatethislosstotheincreaseininvestmentwhichmayaccompanytheincreaseinemploymentisnotlikelytobefruitful。AtanyrateI
　　amnotawareofanyattempthavingbeenmadebythemodernwriterswhoareinterestedin'forcedsaving'toextendtheideatoconditionswhereemploymentisincreasing；andtheyseem，asarule，tooverlookthefactthattheextensionoftheBenthamiteconceptofforcedfrugalitytoconditionsoflessthanfullemploymentrequiressomeexplanationorqualification。
　　V
　　Theprevalenceoftheideathatsavingandinvestment，takenintheirstraightforwardsense，candifferfromoneanother，istobeexplained，Ithink，byanopticalillusionduetoregardinganindividualdepositor'srelationtohisbankasbeingaone-sidedtransaction，insteadofseeingitasthetwo-sidedtransactionwhichitactuallyis。Itissupposedthatadepositorandhisbankcansomehowcontrivebetweenthemtoperformanoperationbywhichsavingscandisappearintothebankingsystemsothattheyarelosttoinvestment，or，contrariwise，thatthebankingsystemcanmakeitpossibleforinvestmenttooccur，towhichnosavingcorresponds。
　　Butnoonecansavewithoutacquiringanasset，whetheritbecashoradebtorcapital-goods；andnoonecanacquireanassetwhichhedidnotpreviouslypossess，unlesseitheranassetofequalvalueisnewlyproducedorsomeoneelse[Page82]THEGENERAL
　　THEORYOFEMPLOYMENT
　　partswithanassetofthatvaluewhichhepreviouslyhad。Inthefirstalternativethereisacorrespondingnewinvestment:inthesecondalternativesomeoneelsemustbedis-savinganequalsum。Forhislossofwealthmustbeduetohisconsumptionexceedinghisincome，andnottoalossoncapitalaccountthroughachangeinthevalueofacapital-asset，sinceitisnotacaseofhissufferingalossofvaluewhichhisassetformerlyhad；heisdulyreceivingthecurrentvalueofhisassetandyetisnotretainingthisvalueinwealthofanyform，i。e。hemustbespendingitoncurrentconsumptioninexcessofcurrentincome。Moreover，ifitisthebankingsystemwhichpartswithanasset，someonemustbepartingwithcash。Itfollowsthattheaggregatesavingofthefirstindividualandofotherstakentogethermustnecessarilybeequaltotheamountofcurrentnewinvestment。
　　Thenotionthatthecreationofcreditbythebankingsystemallowsinvestmenttotakeplacetowhich'nogenuinesaving'correspondscanonlybetheresultofisolatingoneoftheconsequencesoftheincreasedbank-credittotheexclusionoftheothers。Ifthegrantofabankcredittoanentrepreneuradditionaltothecreditsalreadyexistingallowshimtomakeanadditiontocurrentinvestmentwhichwouldnothaveoccurredotherwise，incomeswillnecessarilybeincreasedandataratewhichwillnormallyexceedtherateofincreasedinvestment。Moreover，exceptinconditionsoffullemployment，therewillbeanincreaseofrealincomeaswellasofmoney-income。Thepublicwillexercise'afreechoice'astotheproportioninwhichtheydividetheirincreaseofincomebetweensavingandspending；
　　anditisimpossiblethattheintentionoftheentrepreneurwhohasborrowedinordertoincreaseinvestmentcanbecomeeffectiveexceptinsubstitutionforinvestmentbyotherentrepreneurswhichwouldhaveoccurredotherwise
　　atafasterratethanthepublicdecidetoincreasetheir[Page83]THEMEANING
　　OFSAVINGANDINVESTMENT
　　savings。Moreover，thesavingswhichresultfromthisdecisionarejustasgenuineasanyothersavings。Noonecanbecompelledtoowntheadditionalmoneycorrespondingtothenewbank-credit，unlesshedeliberatelypreferstoholdmoremoneyratherthansomeotherformofwealth。Yetemployment，incomesandpricescannothelpmovinginsuchawaythatinthenewsituationsomeonedoeschoosetoholdtheadditionalmoney。Itistruethatanunexpectedincreaseofinvestmentinaparticulardirectionmaycauseanirregularityintherateofaggregatesavingandinvestmentwhichwouldnothaveoccurredifithadbeensufficientlyforeseen。Itisalsotruethatthegrantofthebank-creditwillsetupthreetendencies¾；1
　　foroutputtoincrease，2forthemarginalproducttoriseinvalueintermsofthewage-unitwhichinconditionsofdecreasingreturnmustnecessarilyaccompanyanincreaseofoutput，and3forthewage-unittoriseintermsofmoneysincethisisafrequentconcomitantofbetteremployment；
　　andthesetendenciesmayaffectthedistributionofrealincomebetweendifferentgroups。Butthesetendenciesarecharacteristicofastateofincreasingoutputassuch，andwilloccurjustasmuchiftheincreaseinoutputhasbeeninitiatedotherwisethanbyanincreaseinbank-credit。
　　Theycanonlybeavoidedbyavoidinganycourseofactioncapableofimprovingemployment。Muchoftheabove，however，isanticipatingtheresultofdiscussionswhichhavenotyetbeenreached。
　　Thustheold-fashionedviewthatsavingalwaysinvolvesinvestment，thoughincompleteandmisleading，isformallysounderthanthenew-fangledviewthattherecanbesavingwithoutinvestmentorinvestmentwithout'genuine'saving。Theerrorliesinproceedingtotheplausibleinferencethat，whenanindividualsaves，hewillincreaseaggregateinvestmentbyanequalamount。Itistrue，that，whenanindividualsavesheincreaseshisownwealth。Buttheconclusionthathe[Page84]THEGENERAL