Speech

Productivity Puzzles UniversityofTennessee Martin,Tennessee October26,1999 Almost everyone is aware by now of statisticisoutputperhouroflaborinput,usually the fact—and I believe it is a fact— called“laborproductivity.”Duringthe1950s thatU.S.productivitygrowthhasrisen and1960s,laborproductivityinthenonfarm substantially. The painful period of businesssectorgrewbyabout2.8percentper unusuallyslowproductivitygrowthinthe1970s year.Atthatrate,productivitydoublesinabout and 1980s is behind us. The increase in output 25years.From1973to1990,laborproductivity per hour of labor input has been high enough grewatarateofonly1.04percentperyear.At over the last few years that it is increasingly thatrate,ittakes67yearsforoutputperhourto reasonable to believe that the United States has double.Currently,itappearsthatoutputperhour indeedturnedthecorneronproductivitygrowth. isgrowingatarateofabout2percentperyear, This picture is reinforced by extensive anecdotal whichdoublesin35years.Evenasmallamount reports from across the country. ofextragrowthyieldsastonishinggainsforthe Ihavebeenstruck,however,bytheextraordi- UnitedStates.Withanextraquarterpercentage narydegreeofuncertaintywefaceonthissubject. pointoflaborproductivitygrowth,GDPwould Mypurposetodayistooutlinethepuzzlesofthe beabout$300billionhigherafteralittlemore 1970sslowdowninproductivitygrowthslow- than10years.Theimpactonthefederalbudget downandtodiscussmoreextensivelythepuzzles alonewouldbeashifttowardsurplusonthe surroundingtherisingproductivitygrowthofthe orderof$60billion. 1990s.Inthinkingaboutthesepuzzles,common Becausethegrowthinrealwagesand,there- senseandbasiceconomicreasoningareagood fore,thestandardoflivingdependsonproduc- dealmorehelpfulthanhypeaboutthe“new tivitygrowth,thetimeittakesforproductivity economy.” todoubleatvariousgrowthratestranslatesquite BeforeIgetintotheseissues,itisimportant easilyintopercapitaincome.Itmakesanenor- thatIissueadisclaimer.TheviewsIexpresshere mousdifferencetooursocietywhetherincome aremyownanddonotnecessarilyreflectofficial isdoublingevery25yearsorevery67years.Indi- positionsoftheFederalReserveSystem.I’vehad viduals,andsocietyasawhole,aremuchbetter alotofhelpwiththeseremarksfromcolleagues— offwhenthemedian-incomefamilycanenjoya especiallyJoeRitter—intheresearchdepartment standardoflivingthattheupper-incomefamily oftheSt.LouisFed;theydeservecreditforthe enjoyedagenerationortwoearlier. strengthsofmyargument.Iretainresponsibility fortheerrors. MONETARY POLICY AND ECONOMIC GROWTH WHY PRODUCTIVITY MATTERS Letmeexplainwhythistopicconcernsme Theimportanceofproductivitygrowthis asamonetarypolicymaker.Apropositionuni- easytodemonstrate.Thesimplestproductivity versallyacceptedbymonetaryeconomistsisthat 1 ECONOMICGROWTH monetarypolicyhasrelativelylittletodowith theeconomy’sunderlyinggrowthpotential.Three long-termeconomicgrowth,aslongastheinfla- percentGDPgrowthcouldbesustainable,orit tionrateremainsmodest.Ibelievethatlowinfla- couldbeaharbingerofacceleratinginflation,and tionisbetterthannot-so-lowinflation,butIam theanswermightchangefromquartertoquarter. notonewhomakestheextravagantclaimthatzero Theanswerdepends,toalargedegree,onthe inflationyieldsenormousbenefitsoversome underlyinggrowthofproductivity. modestrateofinflation.Monetarypolicycan High-frequencyeconomicdataarenotaccu- contributetogeneraleconomicstability;anda rateenough,comprehensiveenough,ortimely stable,lesscyclicaleconomyprobablyraiseslong- enoughtoanswerthequestionwithanydegree termgrowthsomewhat.Centralbanksalsomake ofreliability.Thus,thereisfarlesssciencebehind valuablecontributionstotheefficiencyandsafety ourdecisionsthanIwouldlike.IthinkIcansafely ofthepaymentssystem,whichisanessential saythateverymemberoftheFOMCwouldlike pieceofinfrastructureforamoderneconomy. tofeelmorecertainaboutwhenandhowmuch Still,asimportantasthesecentralbank toadjusttheintendedfederalfundsrate.The responsibilitiesare,itisclearthatthecentral bottomline,though,isthatinthecourseofful- government’sactivitieshavefarmoretodowith fillingourFOMCresponsibilities,wehaveto growththananythingthecentralbankdoes.The judgetheprobablestrengthorweaknessofthe soundnessandefficiencyofthelegalsystem,the economy.Wewanttheeconomytogrowasfast degreeofsafetyofcitizens,taxpolicy,government asitsresourcesandproductivitypermit;thus, spending,andregulation—alltheseaffectproduc- ongoingevaluationofunderlyingproductivity tivitygrowthtoavastlygreaterdegreethancentral growthtrendsisanimportantpartoftheartof bankpolicy. makingmonetarypolicy. Productivitygrowth,however,isterribly importanttomonetarypolicyinadifferentway. HereisthemonetarypolicyissueasIseeit.Ifwe A FRAMEWORK FOR ANALYZING knewhowtosettherateofinflationdirectly,then PRODUCTIVITY weshouldjustchooseazerorateandbedone withit.(Myguessisthatzeroinflation,properly AdamSmith,inhisWealthofNationspub- measured,translatestosomethinglikea1to11/2 lishedmorethan200yearsago,wasthefirstto percentannualincreaseintheconsumerprice arguewithclaritythatanation’swealthwasin indexastheBureauofLaborStatisticsconstructs theoutputofitspeople,notthegoldinitsvaults. thatindextoday.)ButtheFedcan’tsettherateof AndSmithcertainlyunderstoodthetremendous inflationdirectly;thatisnotpossibleinamarket- importanceofproductivitygrowth;hesoughtto basedeconomy. convincehisreadersthatcompetitivemarkets So,theFedhastoworkindirectly.Inabroad generatedwealthandthatrestrictivegovernment sense,wealtertheinflationratebycontrolling policiesmadeEnglandpoorer. growthinthestockofmoney,though,from SinceSmith’sday,we’vefilledinsomeof meetingtomeeting,theFederalOpenMarket thedetailsonhowtheeconomygrowsandhave Committeefocusesonthefederalfundsrate.But amassedahugeamountofempiricalinformation. whenisthefederalfundsratetoolow,leadingto Wehavenot,however,improveduponSmith’s toomuchmoneygrowthand,inturn,toinflation? fundamentalframeworkforunderstandingeco- IftheU.S.economywerestatic,wecouldjust nomicgrowth.Aswithsomanythings,Smithhad experimentabitandfindtherightnumber(as itright. manymacroeconomicstextbooksimply).Butin LetmeputAdamSmith’sanalysisintomod- therealworld,figuringouttherightlevelforthe ernlanguage:Thereisbroadagreementamong federalfundsrateisatoughissue.Amongother economiststhatthemainfactorsthatenablean things,weneedtotrackactualGDPgrowthagainst economytogroware 2 ProductivityPuzzles • thegrowthofthequantityoflaborinput; putintheprivatebusinesssectorwasfivetimes its1948level.Increasingquantitiesoflaborand • thegrowthofthequantityofcapitalinput; capitaleachaccountedforroughly30percentof • therateofimprovementintheprocesses thatincrease,leavingabout40percentofpostwar thatturninputsintooutputs. growth“explained”bygrowthofthismysterious Theonlyamendmentflowingfromadvances residualcategorywecalltotalfactorproductivity. ineconomicknowledgethiscentury—anditisan Growthofthelaborforcewashesoutofthe importantamendment—isourgreatlyincreased laborproductivitystatistics—itaffectsboththe understandingoftheimportanceofhumancapital. numeratoranddenominatorofoutputperhour It’snothardtounderstandthatthetotalvalue bythesamepercentage.Sosomethinglike70 percentofpostwargrowthhas“comefrom”the ofwhataneconomyproduceswillincreaseifthe growthoflaborproductivity.Butkeepinmind numberofpeopleworkingincreasesorifsome thatlaborproductivityisnotanindependent peopleacquirehumancapitalthrougheducation economicforce:Itmeasuresthecombinedeffects oron-the-joblearning.Similarly,providingwork- ofinvestment,learning,andinnovation. erswithmorephysicalcapitalwillincreasetheir BeforeIdigintothepuzzles,I’lldigressbriefly output;aworkercandigalongerditchinaday tomakeanobservationthatillustratesverynicely withabackhoethanwithashovel.Economists whyproductivityisacomplicatedanddifficult haveaprettygoodhandleonthesethings,both subject.Inoneofhisrecentbooks,StephenJay conceptuallyandquantitatively. Gouldtakesupthequestionofwhythe.400bat- Themysteryliesinthatthirdcategory, tingaveragehasdisappeared—nobodyhasaccom- “improvementsinprocesses.”Onemightcallit plishedthisfeatinthemajorleaguessince1941. “technologicalprogress”or“innovation,”butthat MarkMcGwire,SammySosa,andhomeruns doeslittlemorethanrenameit.It‘snotthesame aside,arebatterslessproductivethantheyused thingaslaborproductivity.Infact,theoutput- tobe?“Theproblem,”Gouldwrites,“seemsso per-hourdatacombinetheeffectsofallofthese obviousinoutline:somethingterrific,theapogee factorsexceptlaborhours. ofbattingperformance,wasoncereasonably It’shelpfultokeepthisframeworkinmind, commonandhasnowdisappeared.Therefore, becausepeopleoftentalkaboutproductivity somethingprofoundlynegativehashappenedin growthasthoughit’sjusttechnologicalprogress. baseball.”ButGoulddoesn’treallybelievethat; That’spartlybecausewehavenodirectwayto hespendsthenext50pagesexpandingonthe measurethecontributionofthatthirdcategory followingidea:The.400averageisnotapartof otherthanbysubtractingthecontributionsof thegameitself,butaverysimplestatisticpro- increasedquantitiesoflaborandcapitalfromout- ducedbyacomplexanddynamicsystem,major put.Thatexercisegivesustheresidualcategory leaguebaseball.Since1941,rules,payscales, thateconomistscall“totalfactorproductivity.” trainingregimens,schedules,andstadiumshave Whatendsupinthatresidualcategory?Well, allchanged.Imaginethedifficultyoftryingtopin it’salittlelikeart—weknowitwhenweseeit. downexactlyhow,andtowhatextent,eachof Totalfactorproductivitysoaksuptheeffectsof thesefactorscontributedtothedeclineinbatters’ everythingfromrearrangingawarehousesothat averages.Andbaseballissurelysimplecompared popularitemsareneartheloadingdocktosweep- totheentireU.S.economy! ingchangesintroducedbyinnovationslikeelec- tricityorcomputers.Itshouldn’tsurpriseusthat itisdifficulttomeasurethecontentsofthepigeon- THE 1970s PRODUCTIVITY holewherewedumptheeffectsoffuzzybutpro- PUZZLES foundconceptslikecreativityandinnovation. SowherehasU.S.growthcomefrom?First Thefactthatlaborproductivityaccounted thebigpicture—thepast50years.By1997,out- forabout70percentofgrowthoverthelasthalf 3 ECONOMICGROWTH centuryhidesoneofthemostimportantand productivityhasnotincreased.Thatmeansthe longest-runningstoriesinmacroeconomics,the highergrowthoflaborproductivityinrecent productivityslowdownthatstartedaround1970. yearsprimarilyreflectstheinvestmentboomof Startingintheearly1970s,thetrendrateofpro- recentyears—morecapital—butnotahigher ductivitygrowthfellbyalmost60percentinthe growthrateoftotalfactorproductivity.Theslow- nonfarmbusinesssector.After1975,itgradually downinthe1970sshoweduponbothdimen- becameapparentthatproductivitygrowthhad sions—capitalaccumulationandtotalfactor slowedtoacrawl,comparedwiththerapidpace productivity—butourproductivityrecovery ofthe1950sand60s.Someexpertsflaggedthe seemstobereversingonlyonedimensionofthe changedtrendquickly;othersinsistedforseveral slowdown. yearsthatlowerproductivitygrowthwaslikely Butwait.Itgetsworse.Thedeeperwedig, atemporaryphenomenon.Notuntil1978orso themorepuzzleswefind,andcomputersareat didmosteconomistsagreethatsomethingserious thecenterofthesepuzzles.Inanefforttounder- hadinfacttakenplace. standwhatisgoingon,productivityexpertsdrill Economistshavedebatedthecausesofthe downintothedata.ThedataI’vebeendiscussing sofarrefertotheentireeconomyexceptforthe slowdownforyears,andtheissueisstillunre- governmentsector,forwhichnooverallproduc- solved.Partoftheslowdownwasclearlydueto tivitymeasuresareavailable. slowercapitalaccumulation,butthatonlypushes Forthemanufacturingpartofthetotalbusiness thequestionbackastep.Inanycase,slowercapi- sector,itisapparentthatmostoftheproductivity talaccumulationwascertainlynotthewhole slowdownhasevaporated.Inmanufacturing,both story;totalfactorproductivity—thatresidual laborproductivityandtotalfactorproductivity category—sloweddramaticallyaswell. havebeengrowingrapidlyforseveralyears.But, Somepeopleareconvincedthattheexplana- andthisistrulyastonishing,RobertGordonin tionliesintheenergycrisesofthe1970s;some hisrecentresearcharguesthatwithinmanufac- believethatapolicyenvironmentunfriendlyto turing,afterallowingfornormalcyclicaleffects, businessbearsmuchoftheblame.Otherspoint almostallofthelaborproductivitygrowthhas tothehigherinflationrateofthe1970s,andstill beeninthesectorthatproducescomputers.When otherstoenvironmentalcontrols.Wehavemore youtakeoutthedurablesmanufacturingsector, theoriesthandatapoints.Tothisday,thedecline wherecomputerscomefrom,andlookatwhat’s inproductivitygrowththatoccurredafter1972 left,productivitygrowthlooksdownrighttepid. isapuzzle. Inotherwords,thereisproductivitygrowthwhere computersaremade,butnotwheretheyareused. Thefirstpartofthisisentirelycredible—wedon’t THE 1990s PRODUCTIVITY needgovernmentstatisticianstotellusthatpro- PUZZLES ductivitygainsintheelectronicsindustryhave beenastonishing. Thesecondsetofproductivitypuzzleshas Tounderstandjusthowamazingthispuzzleis, beenunfoldingforthepastseveralyears.They I’llputthepointthisway:All,ormost—depend- aresummarizedinthefamousquipbyRobert ingonyourchoiceofexpert—oftheincreasein Solow:“Youcanseethecomputerageeverywhere productivitygrowthfortheentireU.S.economy butintheproductivitystatistics.” canbeattributedtoasingleindustry—computer Didtheproductivityslowdowntrulyendin manufacturing—thatamountstoabout11/4per- the1990s?Atfirstglance,theanswerappearsto centoftheeconomy!Thatisatrulyremarkable be“partly.”Thelastfewyearshavecertainlyseen finding,anditreallydoesn’tmattermuchwhether strongerlaborproductivitygrowth.Butifwelook thetruthis“all”or“most.” atthedatainmoredetail,thingsdon’tlookas Theideathattherecoveryofproductivity reassuring.Itappearsthatgrowthoftotalfactor growthhasbeenlacklusteramongcomputer-using 4 ProductivityPuzzles firmsdoesn’tseemright,though,doesit?Wesee trendsfromsmallnumbersofobservations.It’s productivityimprovementsallaroundus.And notinconceivablethatitwilltakeusalongtime weseeinnovation,notjustthecapitaldeepening tobesureofaturnaroundinthe1990s. impliedbytheinvestmentboom.Thereisa Asecondangleonthesenumbersistothink moretechnicalwaytoexpressthisdiscomfort: aboutwhetherthemeasurementofproductivity Rememberthatlaborproductivitycombinesthe isdistorted.ZviGriliches,whoisoneofthelead- effectsofcapitalinvestmentandinnovation. ingresearchersinthisarea,arguesthatthepart Sincewe’vebeeninaninvestmentboom,slow oftheeconomyhecalls“reasonablymeasurable” laborproductivitygrowthimpliesthattotalfactor hasdeclinedfromabouthalftolessthan30per- productivityiscompletelystalledorgoingback- centsinceWorldWarII.Theproblemisthatmuch wards!Thisjustdoesnotmakesense.Further- oftheeconomyproducesthingsthatareextremely more,whywouldbusinessesinvestbillionsof difficulttomeasure,andtheshareofthissector— dollarsincomputersthatdon’tincreaseproduc- services,broadlyspeaking—keepsgrowing.More- tivity?Thatis,whyshouldbusinessesnowbe over,theproductivityslowdownappearstobe investingsoheavilyininformationtechnology persistinginthesedifficult-to-measureindustries. iftherateofreturntosuchinvestmentswereno Griliches’bottomlineisthatoutsideofsectors higherthaninthe1970sand1980s?Iftherateof likeagricultureandmanufacturing,whereit’s returnishighertoday,thathigherreturnshould moreorlesspossibletocountthingsinorderto showupintotalfactorproductivity. measureoutput,weshouldbeextremelysuspi- Sohereisthecruxofoursecondproductivity ciousofproductivitynumbers. puzzle:Outsidemanufacturing,productivity Beyondthisconceptualproblem,it’sno growth,asmeasuredbygovernmentstatistics, secretthatstatisticalagencieshaveahardtime looksslow.Withinmanufacturing,productivity keepingupwithinnovationsintheeconomy.In growthisconcentratedinthemanufacturingof particular,ittakesawhiletofigureouthowto computers.Althoughtheworkerswhoproduce measurenewkindsofoutput.Justthismonth,in computershavebecomeimmenselyproductive, fact,theCommerceDepartmenthassignificantly theoverallproductivitydatadon’tseemtosupport updatedhowitmeasuresthecontributionofsoft- theideathatcomputersenhanceproductivity waretotheU.S.economy.So,inabroadsense,it growth.Butthisconclusiondoesn’tseemconsis- shouldn’tsurpriseusverymuchthatoutputand tentwithbusinesseschoosingtoinvestincom- productivitystatisticsareabitslowtocapture puters.Andtheoverallpicturedoesn’tseem rapidinnovation. consistentwithathriving,healthyeconomy. Asyoucansee,wepolicymakershaveareal Therearetwowaystointerpretthisdiscrep- problem.Wehavetodecidehowmuchweightto ancy.Ahard-coredatahoundmightconclude givetooureyesandearsandhowmuchtoformal thatthenumbersarecorrect;manufacturing,in statistics.Weknowthestatisticsmaybemislead- general,anddurablesmanufacturing,inparticular, ing,butwealsoknowthatthedisciplineimposed reallyhasbeenmoreinnovative—streamlining bystatisticsisourmainbulwarkagainsteyesthat productionprocessesandsoforth.Thereisprob- lookthroughrose-coloredglassesandearsthat ablysometruthtothat,butifit’sthewholestory, turntotinwhentheyhearwhatwedonotwant therestoftheprivatesectorisdoingverybadly, tohear—wishfulthinkingofallsorts. indeed.AsIsaid,myobservationssuggestthat innovationandimprovedproductivityareall aroundus—inmanufacturingandelsewhere.It THE FUTURE OF PRODUCTIVITY isusefultorememberthatittookeconomists almostadecadetorecognizetheproductivity I’vetalkedabitaboutthestatistics,andso slowdownofthe1970s;thedataareverynoisy letmenowturntooureyesandearsandhowwe andverycyclical,makingitdifficulttoextract mightbestthinkaboutwhatweseeandhear.We 5 ECONOMICGROWTH seenewelectronictechnologyallaroundus.News Thehistoryof“general-purposetechnologies,” storiesabouttheInternetareincessant.Onewould aseconomistscallthem,tellsuswhyevolutionis surelybejustifiedinsuspectingthatallofthis abetterwordthanrevolution.Onesuchgeneral- representsaproductivityrevolutionofsorts.And purposetechnology,electricity,hasbeenstudied Iampartlysympathetictothisview.DoIthink extensivelybyeconomichistorianPaulDavid. then,thattheproductivitydatahavenothingto AccordingtoDavid,lessthan5percentof tellus?DoIthinkthat,despitetheevidence,we mechanicalpowerinthenation’sfactorieswas areseeingthebirthofsomesortof“neweconomy,” providedbyelectricmotorsin1899.Ittookabout beyondtheboundsofhistoricalexperienceand 20yearsforthatnumbertoriseto50percent. thelawsofeconomics?Hardly.Infact,Ithink Davidaddressedtwointerrelatedquestionsabout thathistoryandsensibleeconomicscantellusa thespreadofelectricityuseingeneralandthe lotaboutthefutureofproductivity. useofelectricmotorsinparticular.First,why First,it’smybeliefthatthepolicyenviron- wastheadoptionratesolow?Second,whatis mentisconsiderablymorehospitabletotheeco- specialaboutthespreadofageneral-purpose nomicactivitiesthatgenerateproductivitygrowth technologylikeelectricity? thanitwasinthe1970s.Thereiswiderunder- Thinkaboutwhatafactorywaslikebefore theelectricmotor.Therewastypicallyasingle standingthaninthepastthatbadpolicy—tax sourceofmechanicalenergy:awaterwheelor, policy,financialpolicy,environmentalpolicy, later,asteamengine.Thisenergyhadtobedis- tradepolicy—canprofoundlyaffectafirm’sincen- tributedaroundtheplantbymechanicalmeans— tivetoinvestinproductiveassets.Tooofteninthe gears,driveshafts,belts,andpulleys.Becauseof past,conflictingpolicygoalshavebeenresolved thenumberofinterconnectedmovingparts,this withoutregardforeconomicefficiency.Although systemwasexpensivetobuild,inflexible,and Ithinkwestillhavealongwaytogointhisregard, dangerous.Buttheinitialexpensewasasunkcost, todaywearemorelikelytoseeinnovativepoli- andonceinplacethesystemdidn’tcostmuchto ciesliketradablepollutionpermits.Thiskindof run.Soinmostcasesitdidn’tmakesensetoscrap market-basedapproachisfarlessdamagingto anoldplantuntilitphysicallyworeout,even productivitythanthestyleofregulationthatsays thoughthenewtechnologywasmarkedlysupe- simply,“Thoushaltnotpollutemorethan3parts rior.Electricmotorsspreadrapidlyinindustries perbillion.”Inshort,wehavelearnedalotabout thatwereexpanding,butelsewheretheoldtech- howtoprotecttheenvironmentatlesscostthan nologycontinuedtoprevail.Thereisatremendous inthepast. amountofinertiainthissortofthing.Thatisthe Turningtotheinnovationprocessitself,we firstlessonaboutthespreadoftechnology:It’s shouldstartbyrecognizingthat,inthemacro- simplytooexpensiveforanindustrialeconomy economicsphere,revolutionstakedecades.Most torearrangeitsproductionandscrapalargepart peoplecallthatevolution.Ibelievethatinforma- ofitscapitalstockovernight,nomatterhow tiontechnologywillbeagenuineengineofgrowth excitingthenewtechnologyis. fordecades,buttherehasn’tbeenandwon’tbea Buttheramificationsoftheelectricmotorfor suddenswervetowardsomesortof“newecon- manufacturing,whenitwasfinallyadopted,were omy.”Itisimportanttounderstandthatsomeof immense.Mechanicalenergydidn’thavetobe thenewtechnologiesreducetheproductivityof distributedfromacentralsource;youjustputa oldertechnologies.Forexample,fortheeconomy motorwhereyouneededit,soyoucouldeasily asawhole,productivitygrowthwillreflectthe reconfiguretheproductionprocess.Theproduc- highproductivityofInternetcommerceandthe tionprocesscouldbephysicallystretchedout, decliningproductivityofthehugeexistinginvest- allowingthedevelopmentofatrueassemblyline. mentinbricksandmortarandpeopleengaged Newfactorystructuresneededonlytokeepthe intraditionalretailing. rainoff;theynolongerneededbracingforheavy, 6 ProductivityPuzzles rapidlymovingpower-distributionmachinery. moneyisapeculiaranimal,buttherealreason Maintenancecouldbeperformedonasingle Ifoundthisstoryinterestingisthatverylittleof machine,withoutshuttingdowntheentirefactory. ithadanythingtodowiththeInternet,perse. Alloftheseadvantageswereclearinprinciple Instead,thethingsthatREIhasfounddifficult attheturnofthecentury,buteachbusinesshad areproblemsanyretailerwouldunderstand. tofigureouthowtoadaptthetechnologytoits Already,theproblemsoffastserversandbroad- needs(aswellasneedingtoamortizeolderinvest- bandconnectionsarelookingsimilartotheprob- ments).Thus,thoughtheimpactofelectricityon lemofrentingastorethathaselectricalservice. manufacturinganddailylifewasprofound,it Therealproblemisfiguringoutwhattodowith wasspreadovermanyyears. them.REI,forexample,foundthatpicturestaken Amorerecentexampleillustratesaslightly foritsprintcatalogdidn’tworkwellonthe differenttheme.Whenthelaserwasinventedin Internet.Tomethissoundsmorelikeolderprob- 1957,noonerecognizeditasageneral-purpose lemssuchaslearninghowtosellthingsinmalls technology.Indeed,BellLabsdidn’tevenpatent thanlikea“neweconomy.” itsinvention.Forsomeyears,thelaserwas Idon’tfeelascurmudgeonlyaboutnewtech- regardedasanextremelyspecializedtool.Infact, nologyasthatsounds.Todayweareinthemiddle itwasbidingitstime,waitingforcomplementary oftheadoptioncycleforaremarkablesetoftech- developmentsinthesemiconductorindustry. nologicalinnovationsinmicroprocessorsand Wheninexpensivesemiconductorlasersbecame communications.Itisdifficulttobelievethat available,thelaserbecameubiquitous.Although thesethingswillnotspureconomicgrowth.But wetendtoconnectthelaserwithgee-whizinven- let’snotkidourselves:Wehaveyettofigureout tionsandweapons,probablyitsmostprofound whattodowithallofthiscomputerpowerand effectontheU.S.economyhasbeenviathe theInternet,andit’sgoingtotaketimetofigure humblebar-codereader,whichwasnotpractical outwhatworksandwhatdoesn’t.Ineffect,we beforecheaplasers.Thisinnovationhasaltered mustwritetheeconomicsoftwareforthistech- theeconomiclandscapeinretailstores,libraries, nology.Thatwilltakealongtime,andwewon’t thepostoffice,evenRedCrossbloodcollection. understandhowithasshapedoureconomyuntil Virtuallyanywhereweneedtokeeptrackofthe ithasalreadyhappened.Thatseemstobethe movementofphysicalobjects,you’llseebarcodes waythesethingshavealwaysbeen. ofonesortoranother. WhatisthebottomlineforU.S.growth Ofcourse,cheaperandcheapercomputing prospects?Optimistsandpessimistsamongseri- powerenableswiderspreadofbar-codescanners, ousstudentsofeconomicgrowtharenotasfar justasbar-codescannersallowbusinessestobring apartasthepopularpresswouldhaveyoubelieve. computingpowertobearoninventorycontrol, Pessimistsbelievethattherecenthigherproduc- marketing,andsales.Who’dhaveimaginedthat tivitygrowthreflectsatransitorycyclicalphenom- theircombinedpowerwouldbemostvisiblein enonandthattheunderlyingortrendgrowthof thegrocerycheckoutlane?That’sthesecondbig laborproductivityremainsboggeddownatabout lessonabouttechnologicalchangethatItakefrom thelevelofthe1970sand1980s.Thatrateisin economichistory:It’shardtopredictthebiggest therangeof1to11/2percentperyearandyields effectsuntilyou’rerightinthemiddleofthem. likelytrendGDPgrowthofabout21/2percentper Letmegiveyouafinalexample,directly year.Optimistsbelievethatthegrowthrateof relatedtotheInternet.Afewmonthsagothe productivityhasrisentothe21/2to3percent NewYorkTimesranalongstoryaboutdifficulties range,whichtranslatesintoaverageGDPgrowth facedbytheonlinesalesoperationsofRecreational inthe3to4percentrangeoverthenextfew EquipmentInc.,anoutdoorequipmentretailer. years.AlthoughIamanoptimistongrowth,my REI’sonlineoperationhasbeenprofitable,though instinctsasapolicymakercompelmetoconcen- notwildlyso.AnInternetretailerthatmakes trateonthemidrangeofinformedopinion—pro- 7 ECONOMICGROWTH ductivitygrowthofabout2percentperyearand tivelylow.Nevertheless,theFederalReserve,in trendGDPgrowthofabout3percentperyear. settingtheintendedfederalfundsrate,cannot Thattomeistheappropriatebasisformonetary avoidmakingsomejudgmentsabouttheproduc- policydecisions.ButIdowanttoemphasizethe tivitygrowth. importanceoftheword“about”inthesejudg- Thecrudestateofourknowledgeaboutcur- ments.WithallI’vesaid,itshouldbeclearthat rentandfutureproductivitytrendsisuncomfort- analystsshouldnotlockintheirviewofthese ableformeasapolicymaker.Iseelittlereasonfor estimates. pessimism,however.Mybestjudgmentisthatthe productivityslowdownofthe1970sand1980s isover.However,wehavetoberealisticabout SUMMARY themagnitudeoftheimprovement.Withallof theoptimismthatsomarksU.S.culture,andwith Productivitygrowthisaterriblyimportant oursatisfactionaboutthefineperformanceof subjectfortheUnitedStates,indeedforevery theeconomyinrecentyears,wemustnotallow country.Butthisisnotaneasysubject.Thepuz- ourselvestobelulledintowishfulthinkingabout zlesofthe1970sslowdowninproductivitygrowth productivityandeconomicgrowth.Evenmodest havenotbeenresolved.Thepuzzlesofthe1990s improvementsinproductivitygrowthareimpor- increaseinproductivitygrowthseemonlyto tant,particularlyiftheycanbesustainedforthe deepenwithfurtherresearchintowhatisgoingon. longrun.Andwecertainlydonotwantoureyes Itisamazing,butstillafactatthistime,thatmost andearsandheadstobeclosedtothepossibility ofthereportedincreaseinproductivitygrowth thattrendproductivitygrowthmightnowbehigh canbeattributedtothecomputer-producing enoughthateconomichistoriansmayintime industryandlittletothecomputer-usingindus- refertothe1990sasthebeginningofanewage. tries—thatis,thewholerestoftheeconomy. Whatawonderfuloutcomethatwouldbe. Thecentralbankisreallyabitplayerinthe growthprocess,providedinflationiskeptrela- 8

William Poole (1999, October 25). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_19991026_poole

@misc{wtfs_speech_19991026_poole,
  author = {William Poole},
  title = {Speech},
  year = {1999},
  month = {Oct},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_19991026_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}