Download Inventive Activity: Problems of Definition and Measurement and more Exams Statistics in PDF only on Docsity!
This PDF is a selection from an out-of-print volume from the National
Bureau of Economic Research
Volume Title: The Rate and Direction of Inventive Activity: Economic
and Social Factors
Volume Author/Editor: Universities-National Bureau Committee for
Economic Research, Committee on Economic Growth of the Social
Science Research Council
Volume Publisher: Princeton University Press
Volume ISBN: 0-87014-304-
Volume URL: http://www.nber.org/books/univ62-
Publication Date: 1962
Chapter Title: Inventive Activity: Problems of Definition and Measurement
Chapter Author: Simon Kuznets
Chapter URL: http://www.nber.org/chapters/c
Chapter pages in book: (p. 19 - 52)
InventiveActivity:ProblemsofDefinition
andMeasurement
SIMONKUZNETS
HARVARDUNIVERSITY
ProblemsofDefinition
INVENTIVEactivityislimitedinthefollowingdiscussiontoactioncon- cernedwithtechnicalinventions,yieldingnewproductstobeturned
outandnewdevicestobeusedineconomicproduction.Wethus
excludesocialinventions,newmethodsofinducinghumanbeingsto competeandcooperateinthesocialprogress—rangingfromsystems
ofpayandincentivesorarrangementsofworksequencewithina
firmtosuchfar-reachinglegalinventionsasthemoderncorporation
ortheplanningsystemofanauthoritarianeconomy.Theeffectsof
socialinventionsoneconomicproductivityareobviouslymajorand
profound,buttheoccupationalgroupsconnectedwithsocialinven-
tionsandtheinstitutionalarrangementsfortheirproduction,selec-
tion,andapplicationaresodifferentfromthoseinvolvedwithtechni-
calinventionsthatthetwocanhardlybetreatedtogether.Andit
needscarcelybementionedthatwealsoexcludecreativeworkofan estheticcharacter,inwhicheconomicuseisnotthemajoraimortest. Oftheseveralcharacteristicsusuallyformulatedindefiningtech- nicalinventions,thefirstdiscussedhereisthataninventionisanew
combinationofavailableknowledgeconcerningpropertiesofthe
materialuniverse.Twoquestionsimmediatelyarise.First,mustthe
combinationbenew?Second,mustitbeofalreadyknownproperties andprocesseswithoutcontainingnewdiscoveries? Therequirementofnewnessserves,presumably,toeliminatedupli-
cationinmeasuringtheoutputofinventiveactivity,forclearlytwo
inventionsidenticalincharacterdonotrepresenttwodistinctaddi-
tionstothestockoftechnologicalknowledge.Butifweareconcerned
withtheinputofinventiveactivity,iftwoidenticalinventionshave
beenmadeindependentlyofeachother,andifwemeasurethepro-
ductivefactorinvolvedbythecapacitytoproduceinventions,then
thesetwoinventionssignifyagreaterinputofinventiveactivityand hence,allotherconditionsbeingequal,agreatersupplyofinventive
NOTE:IamindebtedtoJacobSchmooklerforhelpfulcommentsonanearlierversion ofthispaper.
—whetheritisspecificinformationabouttheatomicweightofsome organiccompoundorawide-flungtheoryofnuclearstructure,whether itisanaccuratedescriptionofthecoastlineofthePersianGulforthe mostspeculativetheoryofgeologicalformation—ispotentiallyuseful, sinceanyadditiontoourtestedknowledgeofthematerialuniverseis exploitableineconomicproduction(thelatterbeingareshufflingofthe
materialuniversetomakeitbetterservehumanpurposes).Second,
mostscientificdiscoveriesareimmediatelyuseless—certainlysofar asthescientistisconcerned,forheisnotattemptingtheapplicationof theresultstousefulendsevenifheisawareofthem,whichoftenheis
not.Third,theoccupationalgroups,theinstitutionalarrangements,
andtheindividualmotivationsofinventorsarequitedifferentfrom
thoseofscientists.Finally,sincescientificdiscoveryisusuallygeneral
incharacter,eachdiscoveryprovidesabaseforawidevarietyof
potentialpracticaluses—technicalinventionsamongthem(themain
reasonfornonpatentability).Atleastsincethesecondhalfofthe
nineteenthcentury,therehavebeenmanymorescientificdiscoveries
thaninventionsthatcouldhavebeenbaseduponthem,andthelags
betweenscientificdiscoveriesandconsequentinventionshavebeen
numerous.Conversely,therehavebeenmanyinventionsbasedon
propertiesofthematerialuniverse,theknowledgeofwhichhasnot
beenfullyincorporatedintothecorpusofscientifictheory.Theselags andleadsareindicationsthatthetwotypesofadditiontoknowledge
havedistinctlydifferentproperties.Formoreeffectiveorganization
ofthefieldofinquiryandsharperfocusingofresearch,itwouldseem tomebettertokeepthelineofdistinctionandrestrictthedefinition
oftechnicalinventionstonewcombinationsofexistingknowledge
designedforpracticaluseinproduction—evenatthedangerofover-
lookingthecontributiontoknowledgethatmuchinventiveactivity
maysupply.
Asecondcharacteristicclaimedforatechnicalinventionisthat
itmustbetheproductofamentaleffortabovetheaverage—although
wedonotinsistonaflashofgeniusoranyotherpresumablyrare
manifestationofmanthecontriver.Theaimhereistodistinguish
betweeninventionsandthehostofimprovementsintechniquethat
aremadeinthedailyprocessofproductionandaretheresultoflow-
levelandratherobviousattentivenessorknow-how.Notthatthe
cumulativeeffectofsuchobviousimprovementsoneconomicpro-
ductivityisslight.Itismerelythataninventionissupposedtobeof
someminimummagnitude—magnitudenotofeconomicimpactbut ofinputofsomeuncommonmentalcapacityofhumanbeings.
Disregardingforthepresenttheimportanceofthischaracteristic
inconnectionwithpatentlegislationoranyothersystemofrewards, wemayaskwhythedistinctionbetweeninventionsandimprovements
isattempted.Ifwecouldmeasuretheeconomicmagnitudeofeach
itemfromthemostobviousimprovementtothemostmajo.rinvention,
wouldwestillbeinterestedinthedistinction?Theansweris"yes,"
ifweassumethattheuppersegmentofthisrange,i.e.,theinventions
whichinfactprovidethebasisfortheimprovements,canonlybe
producedbypeoplewithunusualequipment,andhencedistinguishing
themwouldleadtoamorefruitfulexaminationofthetypeofeffort
andcorrespondingtypeofpersonalityinvolved.Inparticular,the
distinctionsuggeststhatwhileinventionsandimprovementsmaybe gradedbytheireconomicmagnitude,nosuchcontinuityexistsinthe rangeofhumanabilityinvolved;thattherearequalitativedifferences betweentheeffortandcapacityassociatedwithinventionsandthose associatedwithobviousimprovements,sothatnoshiftfromtheone totheothercanbeattained,nomatterhowmuchtraining,education,
orotheruseofreproducibleresourcesisthrownintothescale.The
characteristicthusemphasizestheinputsideofinventiveactivity
andimpliesthatthedistinctivefactorcanbeidentifiedandstudied. Whetherthisisavalidassumption,whetheraneffectivelinecanbe
drawnbetweenthecapacitytoproduceinventionsandthattomake
obviousimprovements,andwhethertheparticulartypeofability
connectedwiththeformercanbeidentifiedandstudied,arequestions
bestlefttootherpapersintheconference.Butitclearlyaffectsthe
problemofmeasurement,andweshallreturntoitinthenextsection. Athirdcharacteristicofatechnicalinventionthatmaybesuggested
isthatitbeuseful,i.e.,thatitbeapracticabledevicewhich,when
employed,wouldeitherreducethecostofproducingalreadyestab-
lishedgoodsormakepossibletheproductionofnewgoodsforwhich thedemandissufficienttocovercosts.Thereasonforsuggestingthis
characteristicisobviousenough,particularlytoeconomists:weare
interestedininventionsbecauseandinsofarastheycontributetothe growthofeconomicproduction.Yetonecaneasilyenvisageinventions thatdonot.Forexample,anewdeviceforproducinghairpinswhich ismorecostly—nowandintheforeseeablefuture—thanthepresently useddevices,isnotusefulinthesenseinwhichweemploythisterm. Thedifficultywiththiscriterion—andithasnotbeenusedinpatent
butforadevicethatismoreeconomicalorattractivelynovelthan
existingones.Thisisthemajordifficultyinvolvedininventing,the
veryaspectthatmayrequirethatminimumeffortdiscussedasthe
secondcharacteristic.Andwhilesomeinventiveactivitymaybemere
tinkering,andthepotentialusefulnessofmanyinventionsmaybe
exaggerated,Iwouldstronglyurgethatthecriterionofusefulnessbe retainedasaconstituentcharacteristicinthedefinitionofatechnical
invention;thatadistinctionbemadeatleastimplicitlybetween
economicallyorientedinventiveactivityandnoneconomicallyori-
entedtinkering;andthatsomeattentionbedevotedtothedifferent
lewisofjudgmentofpotentialusefulnessintheanalysisofboth
inventiveactivityandinventions.Forthiswholematterofusefulness
has,tomyview,obviousbearinguponavarietyofquestionscon-
cernedwithmeasurement.
TheMagnitudeofanInvention
Theprecedingsectionledtoadefinitionofinventiveactivityasbeing concernedwithtechnicalinventions,i.e.newcombinationsofexisting knowledgeindevicespotentiallyusefulineconomicproductionand resultingfromamentalperformanceabovetheaverage.Thedifficul- tiesinmeasuringanactivitysodefinedareperhapsobviousenough; butbeforewereviewtheavailableindexes,itmaybewelltodiscuss explicitlytheproblemsofmeasurementbydealingwiththemagnitude ofasingleinvention. Itfollowsfromtheearliercommentsthataninventionhasatech- nicalandaneconomicside;and,ofcourse,ithasapastandafuture. Thecombinationofthesetwosetsofaspectsgivesusfourviewsofthe
magnitudeofaninvention:(1)thetechnicalproblemovercome—
aviewofthetechnicalpast;(2)thetechnicalpotential,i.e.theeffect
oftheinventiononfurthertechnicalchangesandtheprogressof
technologyingeneral—aviewofthetechnicalfuture;(3)theecono- miccost,i.e.theresourcesconsumedorforegoneinthe"production" oftheinvention—aviewoftheeconomicpast;and(4)theeconomic
potential,i.e.thecontributionoftheinventiontocostreductionor
totheproductionofnewgoodsintheeconomy—aviewoftheecono- micfuture.Ifwecouldsetdownmeaningfulfiguresforeachinvention undereachofthefourheads,theproblemofmeasuringtheinputand outputofinventiveactivitywouldbealmostentirelyremoved,forany remainingdifficultyofidentifyinginventionscouldbeminimized.As
aneconomistIwouldbewillingtosettleformeaningfulquantities
under(3)and(4).Butknowingthatnoeasyanswersareavailable,I feelcompelledtoretain(1)and(2),forwhateverhelptheexamination ofthetechnicalproblemsandpotentialsofaninventionmaygiveus
inconsideringitseconomiccostandcontribution.
1.Considerationofthetechnicalproblemsolvedbyaninvention
bringsusrightbacktoinventionastheproductofhumanperformance ofmorethanaveragequality.Thedifficultyisobviouslyrelatedtothe existingstockofknowledge;anditisassumedthatsuchknowledge
isaccessibleandthatitspossessionisanecessarybutnot,initself,
sufficientconditionofaninvention.Giventhesearchforaneconomic- allyusefuldeviceaspartofthetask,theotherconditionistheability tousetheknowledgeinsomenewandeffectivecombination. Whileamereeconomistmusttreadwarilyhere,itdoesseemtome
thatthemagnitudesofthetechnicaldifficultiesresolvedbyvarious
inventionsdifferand,inthatsense,differentinventionsrepresent
differentmagnitudesofinventivecapacityandofitsinput.Toillus-
trate:Kay'sflyingshuttle,ausefulmodificationinthehandloom,
representedthesolutionofalessertechnicalproblemthanArk-
wright'swater-framewhichmechanizedspinning;Watt'sseparate
condensersolvedagreatertechnicalproblemthantheearlierim-
provementsinNewcomen'sengine;andtouseamodernexample,
theinventionofaneffectiveengineusingnuclearpowerforairtravel wouldrepresentthesolutionofagreatertechnicalproblemthanthe inventionofanotherandbettermechanicalpen. Theseareofcourseselectedexamples;acomparisonofthetechnical problemsresolvedbyWatt'sinventionsrelatingtothesteamengine andFaraday'sandhissuccessors'ontheelectricgeneratorwouldnot yieldunequivocalanswers.AllIamarguingisthatabovethethreshold whereobviousimprovementsceaseançlinventionsbeginthemagni- tudesofthetechnicalproblemsresolvedbyinventionsdifferwidely.
Someoftheproblemsaredifficultandmajorbecausetheyarecon-
cernedwithharnessinganewandpreviouslyintractablesourceof
power;others,becausetheyinvolvetheconversionofavast,complex,
andtime-consumingsetofhandoperationstomachineoperation;
stillothers,becausetheyareconcernedwiththedisplacementofa
widelyusedbutexhaustiblematerialbyanewonewhosequalities,
otherthanasuitabilityforvariouspurposes,arestillnotwellknown. Bycontrast,thetechnicalproblemsinotherinventionsseemtobeof smallermagnitude,becausetheyaremodificationswithintheexisting
frameworkoftechnologyanddonotinvolvesuchnew"large"
inovercomingadifficulttechnicalproblemmaybecomeabasefor
majorsubsequentchangesinthetechniqueoftheprocess.Imaybe
wronginthinkingthattheBessemerconverter,measuredbythemag-
nitudeofthetechnicalproblemresolved,wasa"greater"invention
thantheopen-hearthfurnace;yetitwasthelatterandnottheformer thatbecamethebasisofmodernsteeltechnology,afoundationupon whichalongseriesoftechnicalimprovementsandchangeshasbeen built. 3.Theeconomiccostofaninventioncanpresumablybemeasured
bythevalueoftheresourcesusedinproducingit—thetimeofthe
inventorandhiscollaboratorsandthematerialsconsumed,valuedin termsofalternativeusesormarketprices.
Threequestionsareimmediatelysuggestedbythisformulation.
First,ifninventorsareworkingoninventionXandoneissuccessful, weshouldpresumablyincludetheresourcesusedbyallninventors, notjustbytheonesuccessful.Thiscreatesnoproblemifthecostofa
streamofinventionsoveralongperiodistobemeasuredsincewe
couldpresumablyusetheinputofallwould-beinventors,successful andotherwise.Butthespecificationofthecostofanysingleinvention, orofalimitedgroupofthem,callsforthecapacitytoidentifyallthe
inventorinputsdirectedtowarditsattainment.
Second,howfarbackinthelineofcausationshouldwegoincon-
sideringthecost?Shouldweincludeinthecostofaninventionthe
resourcesdevotedtosecuringtheknowledge,viascientificresearch
anddiscovery,thatunderliesit?Eveniftheoreticallydesirable,it
wouldbeimpossibletospecifyfullytheantecedentknowledgeem-
bodiedinagiveninvention.Butmoreimportant,ifwetreatinventive
activitylikeotherproductionactivity,thereisnomorereasonto
chargeitwiththecostsofscientificandotherknowledge(exceptas
itisalreadyreflectedinthevalueofatrainedinventor'stime)than
thereistochargetheproductionofanycommodity—whichisalso
basedonawidestockofantecedentknowledge—withsuchcosts.
Third,thereistheproblemofspecifyingthecompletionofanin-
ventionanddetermininghowfarforwardthecostcalculationshouldbe carried.Ifweadheretothedefinitionfollowedhere,aninventionis completedwhenitisshownto"work"andwhenaclaimofpossible economicusefulnessismadeforit.Substantiationofthisclaimmight requireaprolongedperiodoftrialuseandadaptation,apilotopera-
tion,andsufficientmassapplicationtopermitmostoftheimprove-
mentsthatmakeforeconomicalproduction.Theeconomiccostofa
testedinventionmay,therefore,bealargemultipleofthecostofan untestedinvention;anditisimportanttodistinguishbetweenthetwo.
Sincemanyinventionsareneverfullyappliedandtested,itisonly
theeconomiccostofuntestedinventionsthatcanbecalculatedforthe wholeuniverseofinventions;anditistheeconomiccostofuntested
inventionsthatcorrespondstoourdefinitionofinventiveactivity.
Butitisusefultokeepinmindtheadditionaldevelopmentcostsin-
volvedinthoseinventionsselectedfor andserviceas
innovationsinproductiveactivity. 4.Thegrosseconomicvalueofaninventiontotheinventororthe privateuseriseitherthecapitalsumreceivedforitorthecapitalized valueofthereturnsthatmaybeassignedtoit.Suchassignmentand
capitalizationmaycallforcomplicatedandimpreciseprocedures,
butwearemoreconcernedherewiththeeconomiccontributionto
society.'
Ifaninventioninvolvesanincreaseinproductivityinturningout
establishedgoods,thegrossmagnitudeofitseconomiccontribution
tosocietyisthediscountedvalueoftheadditionalyieldthatitwill
permit.Giventhedataonsuchadditionalyield,itsproperevaluation requires,however,eitherknowledgeoforassumptionsconcerningthe priceelasticitesofdemandandsupply.2Ifaninventioninvolvesanew product,aroughapproximationtoitseconomicmagnitudeseemsto metobepossibleonlyifthenewproductcanbetreatedasasubstitute foranoldsothatitagainbecomesfeasibletoestimatetheadditional yieldandseekforadefensibleeconomicbasisforevaluatingit.
Grantedthedifficultiesofthetypeofapproximationjustnoted,
twofurtherobservationscanbemade.First,theestimateinvolvesa
forecast,theweightofwhichisgreaterthelongerthefutureofanin-
ventionrelativetoitspast(thefuturewithinthelimitssetbythe
discountingrate).Withalargeproportionofinventionsmadenever reachingthedevelopmentstage,theestimateoftheirgrosseconomic contributionispurelyaforecast;andthatoftheirnetcontributionis
(^1) Itmaybeofinteresttonotethatinestimatingthereturnsfromresearchanddevelop- mentexpenditurestheprevailingpracticesoffirmsfollowratherroughandready formulas,utilizingarbitrarypercentagesofsalesorreturns,arbitrarytimeperiodsover whichtocredittheinventionorimprovementwiththeeffects,andperhapsnoless arbitraryestimatesoftheprobabilityofsuccess(see,e.g.,ScienceandEngineeringin AmericanIndustry:FinalReportona1953—54Survey,NationalScienceFoundation, 56—16,Washington,1956,pp.49—52). 2Seetheingeniousanalysisforthecasesofhybridcornandrelatedinnovationsinthe paperbyZviGriliches,"ResearchCostsandSocialReturns:HybridCornandRelated Innovations,"JournalofPoliticalEconomy,October1958,pp.419—431.
contributionoftheacceleratedrateofminorimprovementsshould
obviouslybetakenintoaccount—minorimprovementsintheprocess orproductitself,fromthestandpointofthefirm;andboththeseand anacceleratedrateofminorimprovementselsewherefromthestand- pointofsocietyasawhole.
Severalconclusionswhichemergefromthisalltoosketchydiscus-
sionofthemagnitudeofaninventionseemparticularlyrelevantto
theproblemofmeasurement.
First,inventionsdifferwidelywithrespecttothemagnitudeof
technicalproblemovercome,technicalpotential,andeconomiccon- tribution.Untestedinventionsmaydifferwidelywithrespecttoecono- miccostandtestedinventionssurelydifferwidelywithrespecttotheir developmentcosts. Second,themagnitudeoftechnicalproblemsovercomemaybeposi- tivelycorrelatedwiththetechnicalpotentialofaninventionandthe
latterwithitseconomiccontribution.Wemay,therefore,assume,
althoughnonetoosecurely,apositiveassociationbetweenthemagni- tudeoftechnicalproblemsovercomeandtheeconomiccontribution ofaninvention. Third,thereisnoreasontoassumesignificantassociationbetween theeconomiccosttotheinventorofasingleuntestedinventionandits
potentialeconomiccontribution.However,somerelationcanbe
assumedifweassigntoagiveninvention,orarelatedgroupofthem, thetotalcost,includingthatofunsuccessfulattempts.Forifagiven areaseemstoawidegroupofinventorstobepotentiallyprofitable,
thecostsdevotedtothatareawouldbelargerthanthoseexpended
elsewhereandifthejudgmentiscorrect,economiccostsofinventions wouldbelargeinareaswheretheirpotentialeconomiccontribution islarge.This,ofcourse,assumesfreemarketoperationinthematch-
ingofcostsandpotentialreturnsanddisregardingtheeffectsof
governmentsubsidies,etc.Andifweincludedevelopmentcosts,there
ismorereasontoexpectthatthecostoftestedinventionswould
bepositivelyassociatedwiththepotentialeconomiccontribution,for
presumably—againontheassumptionoffreeoperationofmarket
forces—thegreaterresourcesneededfordevelopmentwouldflowin directionsthatseemmostpromisingfromthestandpointofpotential
economiccontribution—asjudgedbythecriteriaofthegivenfirm,
althoughnotnecessarilybythecriteriaofsocialusefulness.
MeasuresofInput
Thedifficultiesinthemeasurementofinputarethoseinvolvedin
identifyingtheinputandinevaluatingit. Themajorquestionsofidentificationmaybestatedasfollows.Does theeffectiveprosecutionofinventiveactivitydependuponacertain
relativelyrarecapacityofthehumanmindforwhichthereisno
possiblesubstituteinthecombinedeffortsofseverallessgiftedindi- viduals?Orisinventiveabilitywidelydistributed,evenifatdifferent
levels,sothatthebodyofpotentialinventorsislimitedonlybythe
sizeoftheadultpopulationandtheresourcesavailableforitseduca- tionandbythetimeandfacilitiesneededforeffectiveinventing?Is
therequantitativecomparabilityininventivecapacity,i.e.,canone
man-houroflaborbyapersonatthexlevelofabilitybeequatedto
nhoursbyapersonattheylevelofability,andsoon? Ifweassumewidevariationsininventiveabilityanddisregardfor
themomentthemeasurableresourcesinvolvedintheeducational
andmaterialfacilitiesrequiredforproducinguntestedinventions,in
ordertomeasuretheinputininventiveactivity—forthepresentin
nonmonetaryterms—wewouldhaveto:(1)spotindividualswho,
overagivenperiod,engageininventiveactivity; (2)ascertainthe
hoursspentonsuchactivity;and(3)weightthesehoursbysomescale ofinventiveability.Multiplyingthehoursbytheappropriateweights
wecouldthensaythatsomanyequivalentman-hoursofinventive
capacityhavebeen"put-in"duringtheyear.
Thedisparitybetweenthedatausuallyavailableandthosesug-
gestedabovecaneasilybegleanedfromtheliterature.Inhisilluminat-
ingarticle,"Inventors,PastandPresent,"3JacobSchmooklerpre-
sentsfindingsforasmallsampleofpersonstowhompatentswere
grantedduringfourweeksinOctoberandNovember1953.These
findingsareassembledwithcross-sectioncorrelationsofstatedata
onpatentsandnumberof"technologists"(treatedaswould-bein-
ventors),anddefinedintwoways:thenarrowerdefinitionincludes
electrical,mechanical,chemical,industrial,mining,andmetallurgical engineers,pluschemists,assayers,andmetallurgists;thewidergroup includesalsocivilengineers,architects,designers,draftsmen,thevery minorgroupofinventorsproper,andsurveyors.Fromourstandpoint twooftheconclusionsareofparticularinterest.First,eveninrecent
years,"inventionremainsprimarilyapart-timeactivity.Somewhat
ReviewofEconomicsandStatistics,August1957,pp.321—333.
assumptionsimposeontheuseofthenumberoftechnologistsasan indexofphysicalvolumeofinputofinventiveactivity. Evenifwecouldestablishascaleandmeasureman-hoursofinven-
tiveactivityofequivalentcapacitywewouldfacetheproblemof
valuation.Howdowevaluetheseman-hoursforcomparisonwith
other(material)resourcesemployedininventiveactivityandforcom- parisonofthetotalwithotherinputsintheeconomy?Canweassume thatcompensationpaidbythemarkettohiredinventors(full-orpart-
time)ortherevenuewhichthemarketsecurestoanindependent
inventorreflectsayardsticksimilartothatusedforotherresourcesin theeconomy?Tobemorespecific,canweassumethatthecompensa- tionofahiredinventorisdeterminedbymarginalproductivity,with theflowofsupplysuchthatthemarginalcostofanadditionalunitwill
notexceedorfallfarshortofthemarginalreturn?Consideringthe
difficultyofestimatingtheeconomiccontributionofinventiveacti-
vitysuggestedintheprecedingdiscussionofthemagnitudeofan
invention,theassumptionofmarginalproductivitywouldsurely
strainone'scredulity.Ontheonehand,onemayarguethat,inthe
aggregate,compensationtoinventors,hiredorindependent,full-time orpart-time,iswellbelowanyreasonableestimateoftheassignable socialproductoftheiractivity—partlybecauseoftheexistenceofnon- pecuniaryrewards,andlargelybecauseofthenarrownessoftheprivate marketdemandformanytypesofsociallyusefulinventions.Onthe otherhand,onemayarguethat,insomeindustriesinwhichinventive
activityislavishlyfinancedbecauseofageneralbeliefinawide
technicalchangepotential,aggregatewagespaidtoemployedinvent- orsaretoohigh(incomparisonwithothercommoditiesandservices)
becauseoftheinabilityoffirmstodistinguishbetweenthefertile
inventivemindandtheeducatedhack.Sincethemarketmechanism
cansoeasilyfail,onemaywonderwhetheranincreaseinthereal
salariesofhiredinventorsactuallymeansanincreaseintheinputof
inventiveactivity.Itdoesrepresentanadditiontothevolumeofre-
sourcesdevotedtosustainaparticulargroupofpeoplebutitisno
assuranceofanyriseintheinputofinventiveactivitymeasuredon
somescaleofinventivecapacity.
Thisdefectisinherentinthemoneymeasureoftheinputofany
factorthatcannotbeeffectivelyappraisedbythemarketintermsof
itsqualityorproductivity.Obviously,itisamajordeficiencyofthe availableseriesonresearchanddevelopmentexpendituresbycorpora-
tions,governments,anduniversities.Postponingforthemoment
considerationofotherqualificationsoftheseries,Iwouldliketostress oneimplicationoftheargumentabove:thatindealingwiththeinput ofinventiveactivityanattemptmustbemadetogobehindthemoney
veilandtomeasurephysicalvolumewithallowanceforquality
differentials.True,suchanapproachwouldleadintothedifficult
fieldofinventiveintelligencetestsandpersonalitystudiesandwould,
atbest,yieldaquantitativemeasureofinventivecapacitynotcom-
binablewiththedollarvaluesofreproducibleresourcessomuch
betterevaluatedbymarketprices.Suchameasurewouldstillbemore usefulthandollarvaluetotalswhichreveallittleabouttheproductive
factorinvolved.If,asismostlikely,thesupplyofhumantalentis
thebottleneckininventiveactivity,andifwecanderiveacceptable
measuresoftheinputofthisfactoralone,weshouldeventuallybe
abletostudytheeffectofdifferencesinitsinput,incombinationwith
variousamountsofmaterialresources,ontheoutputofinventions.
Buttoreturntotheseriesonresearchanddevelopmentexpendi-
tures(RandD),theyobviouslysufferfromlimitationsotherthanthe inadequacyofthemoneyyardstickinmeasuringtheinputofinventive (thevolumeofinventiveactivity,asdistinctfromthevolume ofitsoutput).Tobeginwith,theyexcludeindividualandindependent inventors,whetherfull-orpart-time;andinthelightofSchmookler's
study,theomissionissubstantial.Then,theyincludeexpenditures
notonlyonappliedresearch(whichforpracticalpurposescouldbe
consideredafairlycloseapproximationtoinventiveactivity),butalso onbasicresearch(asmallfractionofthereportedtotal)andfarmore
important,ondevelopmentwork.Developmentexpendituresare
probablyfarlargerthanexpendituresonappliedresearchproper,
particularlyamongbusinesscorporations(whoseoutlaysdominate
thelevelandtrendofRandDexpenditures,includingandexcluding governmentfinancing).Thelineofdistinction,althoughdifficultto draw,isimportant:developmentexpendituresbeginwhenaninven- tion,havingbeenshownto"work,"hasbeenchosenforapplication
(providedthetechnicalproblemsofadjustment,prototypes,pilot
operation,etc.permitextendedoperation,i.e.,conversionofanin-
ventionintoaninnovation).Suchexpendituresarelikelytobelarge
becauseworkingoutthepatternsofoperation,eliminating"bugs",
andmakingsuccessivetrialsandremovingerrorsarealltimeconsum- •ingandrequireagreatdealofmaterialcapitalandskilledresources. IdonotmeantodisparagedevelopmentworkwhenIexcludeitfrom inventiveactivity;itcertainlymakesdemandsuponingenuity,tech-
characteristics,andoriginsofinventorsandotherinformationon
input.Moreover,wehavelongstatisticalseriesonpatentsforanum- berofcountriesaswellasvoluminousunpublisheddatainthefilesof thenationalpatentoffices.Hence,insteadofbeginningwithafurther
specificationofthedifficultiesofpropermeasurementofoutput—
whichisperhapsunnecessaryafterourearlierdiscussion—wecan
immediatelyconsiderthedataavailable,andbeginwiththestatistics onpatents. Beforenotingsomeofthelimitationsofthepatentseries,itwould bewelltopointoutitsadvantages.Apatentispresumablyissuedafter
atesthasbeenmadeofaninvention'stechnicalsoundness,i.e.,
whetheritmatchestheinventor'sclaim,and,inmostcountries,after
asearchhasrevealedthatasimilarinventionhasnotalreadybeen
patented.Wemaythereforeassume—allowingforthefallibilityof
anyhumanprocess—thatapatentrepresentsanewandtechnically
feasibledevice.Furthermore,sinceittakestimeandmoneytosecure
apatent,itsissuanceisevidencethatsomeone—eithertheinventor
orhisbackers—believesthatthepotentialeconomicvalueofthe
patentwarrantstheexpenditure.Tobesurethecostsarenotlargeand
somepatentsaretakenoutfortheirnuisanceorblackmailvaluebut
unlesstheproportionofsuchpatentsislargeorvariable,thequali-
ficationtheyimposeuponthestatisticsofpatentsasacountofnew,
technicallyfeasibledeviceswithsomepotentialeconomicvalueis
limited. Notethattheseandsubsequentcommentsreferprimarilytopatents issued—nottopatentapplications.Althoughsometimedistortionis introducedbythevaryingefficiencyofthepatentofficeintotheseries ofpatentsissuedascomparedwithapplications,itistheissuanceof apatentthatassuresatestofsoundnessandnewness.Eventhatmay beupsetbysubsequentchallengeahdlitigation,andperhapstheseries ofpatentsissuednetofupsetsisthemostunequivocalindexofnew devicesofpotentialeconomicvalue.Butthereisnoneedtodwellhere ontechnicaldetailsandonthepossibilitiesofusingoneseriesasan approximationtotheother. Theimpressiveadvantagesofpatentissuesasameasureofoutput
ofinventiveactivity,areoffsetbydisadvantagesundertheheadsof
bothidentificationandvaluation.First,notallinventionsarepatented.
(^8) ThebestrecentdiscussionofthepatentdataisthatbyJacobSchmooklerinan unpublishedpaper,"ACritiqueofPatentStatisticsandaReviewoftheLiterature."I haveprofitedgreatlyfromit,althoughIdisagreeatpointswithSchmookler'sconclusions.
Giventhecostsinvolvedinsecuringapatent,thedangerofcompeti-
tiveimitationresultingfromthecompulsoryrevelationofthenew
device,thecostofpossiblelitigationindelendingthepatent,andthe
abilityofanenterpriseoriginatingorpossessingthenewdeviceto
reapmostofthebenefitofpioneeringwithoutpatentprotection,the numberofinventionsthatarenotpatentedmaybelarge.Ithasbeen suggestedthatthefailureofthenumberofpatentsissuedinthiscoun- trytorisesignificantlysincethe1920'smaybeduetoanincreasing volumeofinventionsforwhichnopatentissought.Thishypothesis gainssomesupportfromtheincreasingproportionofpatentsassigned
tofirms,whichimpliesthatagrowingnumberofinventionsare
originatingunderconditionsinwhichapatentisnotindispensable
(asitistoanindependentinventor)toassuretheeconomicbenefits
tobederived.Acontributoryfactormaybethatmanyinventionsare originatingwithlargefirms(whichhavethedominantproportionof privateresearchdepartments)whofearthatsecuringofpatentsmay
exposethemtoaccusationsofviolationofanti-trustregulations.
Anotherfactormaybethegrowingfinancingofprivateresearchby
government.Alsorelevantmaybetheincreasingextenttowhich
scientificdiscoveryinsomefieldsfacilitatesthedevelopmentofalter- nativedevices,sothatpatentingthepioneeringdeviceisnoprotection
againsteffectivecompetition.Whateverthereasons,thereprobably
arevariationsintheextenttowhichtheseriesofpatentsissuedcovers allthenewtechnicallyfeasibleandpotentiallyprofitableinventions.
Itisalsolikelythatthepatentsysteminthiscountryhasbecome
increasinglydeficientincoverageinrecentdecades. Butthemaindifficultywithpatentstatisticsis,ofcourse,theenorm- ousrangeinthemagnitudeoftheinventionscovered.Obviously,we
cannotassumethatonepatentedinventionis,inanymeaningful
economicsense,equivalenttoanother.Itisinthiseconomicsense,the potentialeconomiccontributionofaninvention,thatwearenaturally interested.Asalreadyindicated,thepotentialeconomiccontribution ofaninventionwhenitiscompletedbutbeforeithasbeentestedand
appliedonanadequatescaleisaratherconjecturalquantity,andit
wouldbedifficulttoattachsuchanestimatetoeachpatent.Yet
patentedinventionsdodifferwidelyintheirpotentialeconomicmag- nitude:apatentforanewcorkscreworplowsulkycannotbeequated withoneforanewgascombustionengine(althoughtheformermay
eventuallybeeconomicallymoreprofitable,privatelyandsocially,
thanthelatter).Evenatpresent,witharestrictedandselectiveuseof
