Speech

The Labor Market and Economic Growth 2003WashingtonLegislativeUpdateConference InternationalFoundationofEmployeeBenefitPlans Washington,D.C. May19,2003 My purpose today is to look ahead RobertRasche,seniorvicepresidentanddirector to the longer-run performance of ofResearch—fortheirassistanceandcomments, the U.S. economy. Many aspects butIretainfullresponsibilityforerrors. of our future society depend on how high the growth rate of GDP turns out to be. The profitability of firms, and therefore trends THE NEW ECONOMY in stock prices, will depend in good part on the Manyobservers,economistsandothers,have U.S.growthrate.Soalsowilltheaveragetaxrate devotedconsiderableattentionandanalysisto we pay; given that many government outlays are thephenomenonknownasthe“NewEconomy” somewhat fixed in real terms, such as require- anditsimplicationsforgrowth.Thisdiscussion mentsfornationaldefense,thehigherthegrowth hasfocusedontheresurgenceofgrowthinlabor rateofGDP,thelowerwillbetheaveragetaxrate productivityinthesecondhalfofthe1990sand on the economy to yield the revenue required to thesustainedhighproductivitygrowththrough servicethesecommitments.Lowertaxratescreate theeconomicslowdownof2001.Oneimportant more incentives for work and investment, both question,notsatisfactorilyresolvedasyet,iswhy ofwhichpromotehighereconomicgrowth.Econ- productivitygrowthsurgedafterthelowproduc- omic growth itself generates growth in govern- tivityyearsfromroughly1974to1994.Whatever ment revenues that can yield budget surpluses thereason,theproductivitysurgedoesseemreal; that raise national saving and therefore national lookingforward,thequestioniswhethertheU.S. investment in productivity enhancing capital. Economyhastrulyenteredaperiodofsustained So,mysubjecttodayislong-rungrowth.My rapidproductivitygrowth. frameworkisasimpleone.GDPgrowthdepends Thisquestionisfundamentaltothefuture onthegrowthofhoursoflaborinputandthe economicwellbeingofoursociety,sincetherateof growthofoutputperhour—whatwecall“labor growthoflaborproductivityiswhatsustainsthe productivity.”I’lltalkbrieflyaboutproductivity growthofourstandardofliving,whichwemeas- growth,butwillconcentratemostofmyremarks ureastheaverageconsumptionperperson.The ontrendswithregardtolaborhours.Thischoice questionisalsocentraltoassessmentsofthelikely oftopicreflectsmyviewthatthesubjectoflabor futureperformanceofthefinancialmarkets.Unfor- hourshasreceivedinsufficientattention;thetopic tunately,becausenooneappearstohaveproduced ofproductivitygrowthhasgarneredmostofthe aconvincinganalysisofthecausesofthe20-year attention. slowdowninproductivity,itisnecessaryto Beforeproceeding,Iwanttoemphasizethat exercisecautionwithrespecttoprognostications theviewsIexpressherearemineanddonot aboutsustainedhighproductivitygrowthinthe necessarilyreflectofficialpositionsoftheFederal future.Suchforecastsarenecessarilyimprecise. ReserveSystem.Ithankmycolleaguesatthe Discussionsofproductivityandeconomic FederalReserveBankofSt.Louis—especially growthmostoftenconcentrateoninventionand 1 ECONOMICGROWTH innovation,andthoseareobviouslyimportant.The viewedfromtheflipsideofthecoin,thefraction functioningofthelabormarketisalsoextremely oftotaloutputthatwechoosetosave);second, important;foronething,alabormarketthatfails theaveragenumberofhourseachemployedper- toplacetherightworkersintherightjobswill sonintheeconomyworksperyear;and,third,the failtoobtainthefullbenefitsofworkers’skillsand fractionofthetotalpopulationactivelyemployed. aptitudesthatcreatehighproductivitygrowth. Intheshortrun,thepercapitaconsumption Moreover,tomakeclearanobviouspoint,if ofasocietycanbeincreasedbyincreasingthe overtimepeopleworkfewerhoursperweekand shareoftotaloutputthatisconsumed.Societies retireatincreasinglyyoungerages,laborinput sometimespursuepoliciesdeliberatelydesigned totheeconomywillgrowmoreslowly,oreven toshiftoutputfrominvestmenttoconsumption, shrink.Declininglaborinputcaneasilycancel preciselytoobtaintheshort-runadvantageof outimprovementsinproductivitygrowth,leav- largerconsumption.Fromalonger-runperspec- ingGDPgrowthunchanged,orevenlowerthan tive,suchpoliciesarecounterproductive,because before.I’mgoingtoconcentratetherestofmy theshareofoutputthatisavailableforinvestment remarksonwhatishappeningonthisfrontbut inphysicalcapitaliscorrespondinglyreduced. willnotdiscussinanydetailpolicyissuesrelat- Henceovertimethecapital-laborratiointhe ingtosuchissuesashoursofworkortheretire- economyfalls(thereverseof“capitaldeepening”) mentage. andpercapitaoutputisreduced.Economies,and individuals,thatdonotsaveandinvestreduce ANALYTICAL FRAMEWORK theirfutureoutputandthereforedeprivethem- selvesoffutureconsumption. Growthofconsumptionpercapitadepends Theaveragenumberofhoursthatemployees onanumberoffactors,eachofwhichcanbe workperyearisalsoachoicevariableforsociety. analyzedseparatelytoadegree.Theamountof Overthepastcentury,asthestandardoflivingin totalconsumptiondependsontheleveloftotal oureconomyhasincreased,workershavechosen output,orrealGDP,lesstheamountofGDP tosubstitutemoreleisureforotherconsumption devotedtoinvestmentandgovernmentspending opportunities.Theyhavedonesothroughshorter ongoodsandservices.Althoughtherearemany workweeksandlongervacationperiods.Six-day interestingandimportantissuesconcerningpri- workweeksarelonggone.Interestingly,though, vateinvestmentandthechoicesgovernments intheUnitedStates,averagehoursperworkweek make,whatIwanttofocusonisthegrowthof stoppedfallinginabout1940. totalGDPandtheroleofthelabormarketin Inothersocieties,thedeclineinaverage advancing,orretarding,thatgrowth. hourshasgonemuchfurtherthanintheUnited Overthelongrun,hoursoflaborinputare States.Workersinthesesocietiesenjoymorepaid determinedprimarilybythegrowthofthetotal population.Givenpopulationgrowth,productiv- holidays,longervacationallotmentsand,insome itygrowthisthesourceofsustainedincreasesin cases,shorterworkweeksthanhere.Insome thestandardoflivinginasociety.Nevertheless, places,notablycontinentalEuropeansocieties, overperiodsthatcanbedecadeslong,otherfactors thetrendtowardgreaterconsumptionofleisure cananddoaffectthelevelofpercapitaconsump- hasbeenlegislated.Thistrend,apparentinmany tion—thatis,theseotherfactorscangenerate countriesoverthepasthalfcentury,hasreduced increasesordecreasesinthegrowthrateofour thestandardoflivingasmeasuredbytheaverage standardoflivingforaperiodofsomeyears.Such levelofconsumptionperpersonrelativetowhat factorsinclude,first,thefractionoftheoutput itwouldhavebeenwithoutthereductioninwork ofoureconomythatwechoosetoconsume(or hours. 2 TheLaborMarketandEconomicGrowth LONG-TERM TRENDS IN THE increasedtoabout76percentin2001.Thistrend persisted—indeedwasmostprominent—during EMPLOYMENT/POPULATION theslowproductivitygrowthperiodfromthe RATIO IN THE UNITED STATES mid-70stothemid-90s.From1974to1994,the Let’sconsidertrendsintheemployedfraction employmentratiorosefrom68to75percent, ofthepopulationandsomeimplicationsofthose therebysupportingthegrowthoftheaverage standardoflivingintheUnitedStatesduring trendsforthegrowthpotentialofoureconomy. thisperiod. Toofewpeopleseemtorealizejusthowdifferent Aprincipalsourceoftherisingemployment thesituationintheUnitedStatesisinthisregard ratioduringthe1970sand1980swastheremark- frommanyothercountries. ablerateofintegrationoffemaleworkersinto Iwillfocusprimarilyonthe20-to-64year employment.Startingfromonly40percentin agegroup.Thisisnottheconventionalmeasure 1960,theemployedfractionoffemalesaged20 thatisusuallyemployedinsuchdiscussions. to64increasedto69percentby1997,afterwhich Thedatathatarecommonlycitedrefereitherto itleveledoff.Incontrast,theemploy- the“workingage”population—currentlydefined ment/populationratioformalesaged20to64, as16yearsofageandolderintheUnitedStates, whichwas89percentin1960,declinedfromthe orthe“primeworkingage”population—16 through64yearsofage. mid-1970sthroughthemid-1980stothelow80 Myexclusionoftheteenagepopulationfrom percentlevel,aroundwhichithasfluctuated thisdiscussionisdeliberate.First,thechoiceof withouttrendeversince.Muchofthedeclinein age16inouremploymentstatisticsisanarbitrary themaleemploymentratiowasaconsequence convention.Before1967theemploymentstatis- ofearlierretirement. ticscompiledbytheBureauofLaborStatistics Incidentally,whenIwasteachingIusedto werebasedonages14andolder.Thedecisionto enjoytellingmybeginningeconomicsstudents redefinethestatisticspresumablyreflectedthe thatthesocialrevolutionthatsodramatically ideathatasoursocietybecamelessagrarianand increasedthefractionofwomenatworkmade ahigherpercentageofjobsrequiredmoreskilled possibletheearlyretirementofmen.Thatline labor,theprincipalactivityof14to16yearolds rarelygotevenanembarrassedgiggle.Butit hadbecomefull-timeschoolratherthaninfull- reallyistruethatwithouttheproductiongains timeemployment. fromtherisingfractionofwomenatworkour Theskillrequirementsofourlabormarket societycouldnothaveaffordedincreasingly havecontinuedtoincrease,perhapsatanaccel- earlyretirementformen. eratingrate,overtheyearssincethelaborforce Thenetresultofthesetrendsisthatthe wasredefined.Today,employmentprospects employment/populationratiofortheentirepop- andstandardsoflivingarebleakfortheperson ulationaged20to64increasedfrom64percent lackingahighschooleducation.Incomedifferen- in1960to76percentin2001.Thisexperience tialsbetweenhighschoolandcollegegraduates contradictsapopularhypothesisthatthereisonly haveincreasedsubstantiallyoverthepastdecade. afixedamountofworkopportunitiesavailable Consequently,anincreasinglylargefractionof intheeconomy.Fromthatperspective,public thoseinthe16to19agegroupareinschool,and policiesshoulddiscourageorexcludecertain itmakessensetoconcentratelaborforceanalysis groupsofindividualsfromemploymentinorder onthoseage20andabove. to“makeroom”forothergroupsofworkers, Overthepast40yearsintheUnitedStates particularlyyoungerworkers.Analternative therehasbeenasteadyupwardtrendinthe hypothesisisthatwelldesignedpublicpolicies employment/populationratioofthose20to64 thatpromotemaximumsustainableeconomic yearsold.Startingfromabout64percentin1960, growthwillprovideanamplesupplyofemploy- theemployedshareofthispopulationgroup mentopportunitiesfortheentireavailablepopu- 3 ECONOMICGROWTH lationofskilledworkers.Certainly,myviewis genarianswillbeenticedtoworkregularlyas thatgovernmentshouldpursuepoliciesthatcre- greetersatWal-Mart! ateemploymentopportunitiesforeveryonewill- Inspiteofage-discriminationlaws,theabo- ingandabletowork. litionofmandatoryretirementprovisions,and I’vehintedthatearlyretirementisadisturb- increasinglongevity,thefractionofindividuals ingcharacteristicofemploymenttrendsoverthe 65andoveremployedhasdeclinedfromaround lastseveraldecades.Indeed,thedeclineinthe 17percentinthemid-1960stounder13percent employment/populationratioofmaleswasheav- in2001.Thistrendprimarilyreflectsadecline ilyconcentratedinthe55-to-64agegroup.Inthe intheemployment/populationratioofmales65 1960s,theemploymentratiooftheseindividuals yearsofageandolderfromover25percentin wasover80percent,lessthan10percentage themid-1960stothe15to17percentrangesince themid-1980s.Theemployment/populationratio pointslowerthanthatoftheentiremalepopula- offemales65yearsandolderhasbeentrendless tionages20to64.By2001,theemploymentratio atunder10percentduringthelast40years. ofthisgroupwasonly65percent,almost20per- Youmayrecallthatthelastoccupationtobe centagepointslessthantheemploymentratio freedfrommandatoryrequirementprovisions fortheentiremalepopulationages20to64. wastenuredcollegeprofessors—myprevious Ifindthesubstantialincreaseinearlyretire- occupationalexperience.Universityadministra- mentdisturbingforseveralreasons.Individualsin torsexpressedconsiderableapprehensionatthe thisagegrouphaveconsiderableworkexperience prospectofdodderingprofessorsrefusingtoretire andlikelyhaveaccumulatedsubstantialskills. tomakeroominthetenureranksfornew,younger Theimprovedhealthstatusofthepopulationand blood.Suchconcernsprovedbaseless.Relatively increasedlifeexpectancymightbeexpectedto fewacademicshavechosentopursuefull-time yieldlongerratherthanshorterworkinglives.In teachingmuchbeyondage70. theyearstocome,asthebaby-boomgeneration Smallastheemployedfractionoftheolder reachesnormalretirementage,thefractionofthe agegroupsisintheUnitedStates,wewillseelater totalpopulationatworktosupportthosewhoare thatitissubstantialcomparedwiththesituation retiredwillfall.Theburdenofthedependentpop- inmanyothereconomies.Thisfactissignificant ulation—boththeyoungandthoseretired—onthe inaperiodofgenerallyagingpopulations.With workingpopulationwillgrow.Inmyopinion,we increasinglongevity,stableemploymentratios willhavetoconsiderwhetherthegovernment forrelativelyyoungagegroupsimplyanincreas- shouldadoptpoliciestoincreaseincentivesfor ingdependencyratio:theratioofthepopulation olderworkerstoremainemployed,perhapsin notemployedtothatemployed.Withlowor part-timeemploymentifthatispreferred.Inany decliningemploymentratiosforyoungerpersons event,U.S.GDPgrowthwilldependinparton remaininginschoollonger,thedependencyratio whetherthetrendtoearlierretirementcontinues. isevenlarger.Highdependencyratioscausesig- Attheupperagedefinitionofthelaborforce, nificantproblemsforthesolvencyofpay-as-you- IconfessthatIamnotaltogethercomfortablewith gogovernmentbenefitprogramsforthesenior excludingindividuals65yearsofageandolder. population,suchasSocialSecurityandMedicare Onapersonallevel,thisdefinitionsuggeststhat intheUnitedStates. Ishouldgracefullymoveaside,aprospectthat Asthedependencyratioinasocietyincreases, doesnotappealtomeintheimmediatefuture. ever-highertaxesontheemployedfractionofthe Someindividualschoosetoparticipateinthe populationarerequiredtomaintainthesolvency laborforcewellbeyondage65—Chairman ofsuchprograms.Highermarginaltaxrateson Greenspan,forexample.Clearly,asseniorsage, theworkingfractionofthepopulationinturncan relativelyfewerindividualswillchoosetocon- reduceparticipationinemploymentandgenerate tinueemployment.Itisunlikelythatmanynona- evenhigherdependencyratios.Suchsituations 4 TheLaborMarketandEconomicGrowth arenotstableenvironments.Wearewitnessing ofseniors,startingfromsubstantiallyhigher thiskindofprobleminanumberofeconomies employment/populationratiosintheearly1970s. today,wherethereisanactivedebateonwhether Forthose64andyounger,in2001theemploy- variousgovernmentprogramsthatbenefitsen- mentratiosareremarkablysimilarfortheUnited iorscanbemaintainedatexistinglevels,orhow States,Canada,Japan,andtheUnitedKingdom. thecostofsuchprogramscanbesignificantly Forallthesecountries,theratioofemployment contained. tototalpopulationisapproximately75percent. Thedistributionofemploymentbetweenmales andfemalesin2001isroughlythesameinthe INTERNATIONAL COMPARISONS UnitedStates,Canada,andtheUnitedKingdom— intherangeof79to83percentformalesand67 Therearestrikingdifferencesacrosscountries to69percentforfemales.InJapan,maleemploy- intheutilizationofthelaborresourcesavailable. mentratiosaresubstantiallyhigher,(87percent) Iwillfocusfirstonthesenioragegroup—those65 andfemaleemploymentratiosaresubstantially andolder.TheJapaneseeconomyhasthehighest lower(at61percent)thanintheotherthree utilizationoftheseworkers.OECDdataforJapan countries. startin1968,atwhichtimetheemploymentratios Employmentratiosforthose64andyounger forJapanesemalesandfemales65andoverwere inGermanyandFranceare69percentand67 52percentand19percent,respectively,more percent,respectively,whileforItalytheoverall thantwicethecorrespondingemploymentratios ratioisonly58percent.InItalytherehasbeen intheUnitedStatesatthesametime.Theseratios notrendinthesestatisticsovertheperiodsince inJapantrendeddownwardoverthelastthirdof 1980. the20thcentury,reachinglowsof31percentand InFranceandGermanytheoverallemploy- 14percentformalesandfemales,respectively, mentratioin1980wasvirtuallyidenticaltothat in2001.WhilethenegativetrendsinJapanwere intheUnitedStates.Thisfactisparticularlyinter- strongerthanintheUnitedStates,theoverall estingbecauseinthe1960sand1970sEuropean employmentratiofortheseniorpopulationin countrieswerecitedasexamplesof“lowunem- Japanremainsabouttwo-thirdshigherthanin ploymenteconomies”thatsomeregardedas theUnitedStates. theenvyofcountries—inparticulartheUnited Canadiandataareavailableonlyfortheperiod States—saidtobemiredinchronicallyhigh since1976.Atthattime,theemploymentratios unemployment.Yetin1980,regardlessofthe formalesandfemales65andoverwere15percent reportedunemploymentnumbers,theemployed and4percent,respectively.Thesefractionsare “youngerworkers”inthethreecountriesisthe fairlycomparabletothoseintheUnitedStatesat samefractionofthepopulation.Inrecentyears, thatsametime:19percentand8percent,respec- theperspectiveontheunemploymentsituation tively.Asof2001,theoverallemploymentratio hasreversed:GermanyandFranceareconsidered forthose65andoverisonly6percentinCanada, highunemploymenteconomies,withreported lessthanhalfthatintheUnitedStates. unemploymentratesatorneardouble-digitlevels, Sincethemid-1980s,theemploymentratios whilereportedunemploymentintheUnited forseniorsinmanycountrieshavebeenextremely Stateshasonlyrecentlycreptupto6percent. low;5percentorless.Forpracticalpurposes,indi- Comparisonsofemploymentratiosprovide vidualsofthisagedonotparticipateinemploy- importantinformationonhowwellcountries mentinthesecountries.InItaly,thisenvironment utilizetheirlaborforces.Alsoimportantarecom- hasprevailedsinceatleast1970.IntheUnited parisonsacrosscountriesonaveragehoursworked Kingdom,dataforpreviousyearsarenotavailable. perworkerperyear.Availabledataneedtobe Intwocountries,FranceandGermany,therewere interpretedcautiouslybecauseofdifferentstatis- strongnegativetrendsintheemploymentratios ticalsourcesindifferentcountries.Nevertheless, 5 ECONOMICGROWTH thenumbersofhoursworkedperworkerperyear of3.8percentagepoints.Thecorresponding appeartodiffersubstantiallyindifferentcountries. employmentratioforfemalesdeclinedfrom70.6 AccordingtoOECDdata,averageannualhours percentinAprilof2000to68.6percentinApril in2001intheUnitedStateswere1,821.Corres- 2003,adeclineof2.0percentagepoints.Thetotal pondingdataforsomeothercountrieswere1,532 employmentratiofortheseagegroupsdeclined forFrance,1,467forGermanyand1,346forthe fromahighof77.4percentinAprilof2000to Netherlands.Datafor2001arenotavailablefor 74.6percentinApril2003,adeclineof2.8per- CanadaandJapan,butdatafor2000indicate centagepoints.Thedeclineintheoverallemploy- thataverageannualhoursinthosetwocountries mentratiofortheseagegroupscompareswithan areroughlythesameasintheUnitedStates. increaseof2.2percentagepointsintheoverall Overthepast20years,averageannualhours unemploymentrateintheU.S.economyoverthe havechangedlittleintheUnitedStatesand samethree-yearperiod. Canada,buthavedeclinedsignificantlyinthe Idon’tknowwhether,orhowquickly,theU.S. othercountriesI’vementionedinthiscontext. economywillreturntothesamehighemploy- Togainafeelforthequantitativeimportanceof mentratioexperiencedthreeyearsago.Butthere thedeclineinaverageannualhours,forFrance islittlequestionthatthelaborforceissignificantly thedeclinebetween1979and2001amountedto underemployedtoday.Thus,itisreasonableto 0.75percentperyear.Thatdeclineissignificant expectthattotalhourscouldincreasesubstantially foraneconomywithatrendrateofgrowthinthe overthenextseveralyears.Wehavethepotential neighborhoodof2percentperyear.Whenwe forlaborhourstoincreaseby2percentperyear considerboththedeclineinaverageannualhours forseveralyears—1percentperyearfromthe andthedeclineinthefractionofthepopulation longer-rungrowthofthepopulation,andanother employed,thetwoeffectstakentogetheraddup 1percentperyeartomakeupforthedeclines toasignificantdeclineinlaborinputovertime. duringtherecentslow-economyyears.Adding Formanycountries,thedeclineinlaborinputis totheincreaseinlaborhoursthegrowthinlabor everybitasimportant,andinmanycasesmore productivityof2to2.5percentperyeargivesus important,thananydeclineinlaborproductivity thepotentialforrealGDPgrowthof4to4.5per- growth. centperyearforseveralyears. Thesesimplecalculationsmakeclearthat theUnitedStateshasthepotentialtogrowsub- RECENT AND PROSPECTIVE stantiallyoverthenextseveralyears,andthata majorpartofthatgrowthwillcomefromgrowth CHANGES IN THE oflaborhours.Productivitygrowthisthecritical EMPLOYMENT/POPULATION elementofourlonger-runfuture,butoverthe RATIO IN THE UNITED STATES immediatefuturelaborutilizationplaysanequally Employment/populationratiosintheUnited importantrole.Weneedtomakesurethatpublic Statespeakedpriortothemostrecentbusiness policyencouragesproductivitygrowthandfull cyclepeakinMarch2001.Theemploymentratio utilizationoflabor,bothfortheimmediatefuture formalesages20to64declinedfrom84.7percent andforthelongrun.Ithinkwe’reontheright inJune2000to81percentinApril2002,adecline track,andhaveamplereasontobeoptimistic. 6

William Poole (2003, May 18). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20030519_poole

@misc{wtfs_speech_20030519_poole,
  author = {William Poole},
  title = {Speech},
  year = {2003},
  month = {May},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20030519_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}