speeches · February 14, 2007

Speech

William Poole · President

U.S. Saving CFASocietyofNebraska Omaha,Nebraska February15,2007 Iam delighted to be here today—my first more,andthatisequallytruetoday.Therealityhas triptoOmahasincecomingtotheSt.Louis changedfar,farlessthantheheadlinessuggest. Fed in 1998. My subject is an important, Beforeproceeding,Iwanttoemphasizethat and puzzling, one. The puzzle is nicely theviewsIexpressherearemineanddonot illustrated by recent newspaper stories reporting necessarilyreflectofficialpositionsoftheFederal that the U.S. saving rate is at the lowest level in ReserveSystem.Ithankmycolleaguesatthe 73 years—that is, since 1933, the bleakest year FederalReserveBankofSt.Louisfortheircom- oftheGreatDepression.Butletmeaskfiveques- ments.RobertH.Rasche,seniorvicepresident tions: Are there signs of distress all around, as anddirectorofResearch,andMassimoGuidolin, therewere73yearsago?Hastherebeenatremen- assistantvicepresident,providedspecialassis- dous surge of bankruptcies? Has the United tance.Iretainfullresponsibilityforerrors. States become a nation of profligate spenders? Are the data wrong? Are the data screwy? Myanswerstothesefivequestionsareno, HEADLINE GRABBING DATA no,no,noandno.Buttherearesomepuzzlesto Withinthepastyearthepersonalsavingratio, explain,andthatiswhatmyremarksareabout. whichisreportedeverymonth,hasturnednega- Asyoumaybeabletotellfromtheseintro- tive(Figure1A),andthesemonthlydatareleases ductoryremarks,Iamnotgoingtoexpressdeep haveyieldedheadlinesallalongtheway.Monthly dismayinlinewiththeheadlinenews.Neverthe- dataonhouseholddebtservicepaymentsasa less,toavoidbeingmisinterpreted,Iwantto percentofpersonalincomehavereachedalltime emphasizethatmyrelaxedperspectiveonnational highs(Figure1B).Thefederalgovernmentis savingtodaydoesnotimplythatindividualhouse- runningalargebudgetdeficit(Figure1C),and holdshavenothingtoworryabout.Manyhouse- theU.S.netinternationalinvestmentposition holdswouldbemuchbetteroffiftheyhadlarger (Figure1D)isnowreportedasanegativenet assetsandlesscredit-carddebtcarryinghigh positioninexcessof20percentofGDP.Reports interestrates.Manywouldhaveahappierand inthefinancialpressdiscusstherapidaccumu- moresecureretirementiftheyconsumedlessand lationofforeignexchangereservesbyChina,held accumulatedmorewealthduringtheirworking mostlyinU.S.dollars,andspeculateontheimpact years.Butthesepointswereasvalid25yearsago onU.S.interestratesandthedollarexchange whenthepersonalsavingratewasfairlyconsis- rateshouldtheChinesechoosetodiversifyasig- tentlyabove6percentastheyaretodaywhenthe nificantfractionofsuchholdingsoutofdollars. savingrateisnegative.Moreover,Ithinktherewas Personalfinancialadvisersandothersfrequently acase25yearsagothattheUnitedStatesasa arequotedasforecastingthatthe“boomers”are wholewouldhavebeenbetteroffifithadsaved illpreparedtofinancetheirretirementyears.1 1 See,forexample,JackVanDerhei,CraigCopelandandDallasSalisbury,RetirementSecurityintheUnitedStates,WashingtonD.C.: EmployeeBenefitResearchInstitute,2006. 1 ECONOMICGROWTH B.HouseholdDebtServicePaymentsasPercent ofPersonalDisposableIncome 15 14 13 12 11 10 9 8 1980:1 1982:1 1984:1 1986:1 1988:1 1990:1 1992:1 1994:1 1996:1 1998:1 2000:1 2002:1 2004:1 2006:1 Suchheadlinenewsalarmssomereadersand havecometorelyupontoexplainthesavings indeedcanpromoteageneralfeelingofunease behaviorofconsumers,andthenpresentsome aboutthefuturestandardoflivingofU.S.citizens. additionaldatarelevanttothosetheoriesthat Thenewsremindsmeofthechorusofanold suggestavastlydifferentperspective. songfrommyyouth: Youloadsixteentons,andwhatdoyouget? Anotherdayolderanddeeperindebt. THE DEFINITIONS BEHIND THE SaintPeter,don’tyoucallme’causeIcan’tgo; HEADLINE NEWS Iowemysoultothecompanystore.2 Ibeginwiththecriticaldistinctionbetween However,thereisa“restofthestory.”My savingandsavings.Economistsareoftenabit purposetodayistodiscusssomeofthedefini- tionsthatareusedtogeneratethesavingdata sloppy,inmyview,inhowtheyusethesetwo featuredintheheadlines,tooutlinesomeofthe wordsandheadlinewritersmaynotappreciate basicinsightsofthetheoriesthateconomists theimportanceofthedistinction.Savingisthe 2 tnecreP A.PersonalandPrivateBusinessSavingRatiosto GDP 10.0% 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% Jan-47 Jan-50 Jan-53 Jan-56 Jan-59 Jan-62 Jan-65 Jan-68 Jan-71 Jan-74 Jan-77 Jan-80 Jan-83 Jan-86 Jan-89 Jan-92 Jan-95 Jan-98 Jan-01 Jan-04 PDGfotnecreP PersonalSavingRatio NetBusinessSaving C.NetGov'tSavingRatiostoGDP 8.0% 6.0% 4.0% 2.0% 0.0% -2.0% -4.0% -6.0% -8.0% Jan-47 Jan-50 Jan-53 Jan-56 Jan-59 Jan-62 Jan-65 Jan-68 Jan-71 Jan-74 Jan-77 Jan-80 Jan-83 Jan-86 Jan-89 Jan-92 Jan-95 Jan-98 Jan-01 Jan-04 PDGfotnecreP D.U.S.NetInternationalInvestmentPosition 15 10 5 0 -5 -10 -15 -20 -25 NetGov'tSavingRatio NetS&LGov'tSavingRatio 197619781980198219841986198819901992199419961998200020022004 PDGfotnecreP Figure 1 2 WrittenbyMerleTravis,1947. U.S.Saving flowofafter-taxincomenotconsumed.Roughly productionofgoodsandservicesinsomepartic- speaking,savingisthepartofyourmonthlypay- ularperiod,suchasayear.TheBEAdefinesper- checkthatisleftoverafterpayingallyourbills. sonalsavingasthedifferencebetweencurrent Youmayuseyoursavingtopaydowndebtoradd personaloutlaysandcurrentdisposablepersonal toyourassets.Savings,withthe“s”ontheend, income.Thesavingrateissavingasapercentage isthestockofwealthyouhaveaccumulated— ofdisposableincome. yournetworthcalculatedbytakingthevalueof TheNIPAframeworkisthatofadouble-entry alltheassetsyouownandsubtractingyourlia- accountingsystem.Byconstruction,grossnational bilities.Savingisaflow,savingsisanaccumu- productisequaltogrossnationalincome.Con- latedstock.Thedistinctionbetweensavingand siderthevalueofacompany’soutput.Itsrevenue savingsisaselementaryandasimportantasthe fromsalesisthevalueoftheproduction.The distinctionbetweenacompany’sincomestate- revenueisthendistributedasincometoemploy- mentanditsbalancesheet. ees,dividendstoshareholdersandretainedearn- Fortheeconomist,thebasicdefinitionof ings,orundistributedcorporateincome.Ofcourse, householdincomeinaparticularyeariscon- thetaxcollectorsgetsometoo,anditshowsupin sumptionplusthechangeinnetworth.Thesav- theNIPAasgovernmentrevenue.Bydefinition, ingrateisthenincomeminusconsumptionasa thetotalvalueoftheproductequalsthetotal percentageofincome.Theissueissimplewhen valueoftheincometovariousincomerecipients. thesettingissimple.Supposeyouhaveanannual TheNIPAfocusonbothexpenditureand salaryincomeof$100,000andyouronlyassetis incomeflowshasthepotentialtocreateaccount- abankdepositbearingnointerest.If,overthe ingdiscrepanciesbecausethedataarecollected courseoftheyear,youspend$95,000oncon- fromdifferentdatasources.Incomedataarecol- sumptiongoods,thenyoursavingis$5,000,which lectedfrompayrolldata,IRSfilings,corporate showsupasanincreaseinyourbankdeposit. taxreportsandthelike.Personaloutlaysare Wecanthinkofyourincomeasthesalaryof almostentirelypersonalconsumptionexpendi- $100,000,orasyourconsumptionof$95,000 tures.Thedollarvalueoftheseexpendituresis plustheincreaseof$5,000inyournetworth. thevalueofthegoodsandservicescompanies Thetwodifferentapproachesyieldthesame selltohouseholds.Themorereliabledataare answer.Yoursavingof$5,000yieldsasaving thosefromthedemand(expenditure)sideofthe rateof5percent. nationalaccounts.Incomedataarenotoriously Ifyouhadconsumedgoodsworth$105,000, imprecise,andtheyfailtoadduptoaggregate youwouldhavehadtodrawdownyourbank GDPbyasmuchas2to3percent.Thisdifference accountorborrow.Ineithercase,yournetworth appearsintheNIPAasanaccountcalled“statis- wouldhavedeclinedby$5,000.Yourincomewas ticaldiscrepancy.”Ifincomedataaretypically still$100,000,whethercalculateddirectlyfrom underestimated,theNIPAsavingratewillalso yoursalaryorfromyourconsumptionplusthe beunderestimated.Inanyevent,weknowthat changeinyournetworth.Yoursavingratewould thesavingrateissubjecttosubstantialmeasure- havebeenminus5percent. menterrorandtofrequentmajorrevisions. Theissuebecomesmuchmorecomplicated Theseobservationsaremeanttohelpunder- whenyourassetscanchangeinprice,yielding standthemeasurementissue,buttheydonot capitalgainsorlosses.Aswewillsee,thatis resolvetheapparentmysteryofwhythesaving muchofwhatthesavingrateissueisallabout. ratehasbecomenegative.Thestatisticaldiscrep- Theunderlyingdataoftheheadlinenews ancyintheNIPAhasnotbeengrowingovertime. aremostlyderivedfromourNationalIncome AnumberofotherissuesI’llflagdonotsolvethe andProductAccounts—theNIPA.Theseareflow mysteryeither,becausethemeasurementerrors accounts,constructedbytheBureauofEconomic donotseemtobechangingenoughovertimeto Analysis(BEA),designedtomeasurethecurrent accountforthechangeinthesavingrate.Never- 3 ECONOMICGROWTH theless,I’llgothoughsomeofthemoreimportant offcashtoinvestorseitherthroughpayingdivi- issues. dends,whichappearinpersonalincome,or Anumberofstatisticalandmeasurement throughsharerepurchases.Sharerepurchases issueshavebeendebatedintheliteratureonthe tendtoincreasestockprices,yieldingcapital evolutionoftheU.S.savingrate.Iwillfocuson gainstoshareholderswhichdonotappearin fivedistinctissuesthatmaycausethemeasured personalincome.Ifcompanieshaveincreasingly NIPAsavingratetosubstantiallydifferfroma usedsharerepurchasesinsteadofdividends— true,unobservedpersonalsavingrate. whichappearstometobethecase—theresult wouldbetocreateadownwardbiastothemeas- The NIPA Saving Rate and Realized uredsavingrate. Capital Gains NIPA Saving Rate and Pension Plans Grossdomesticincomeisdesignedtomeas- ureincomegeneratedbycurrentproduction. TheNIPAtreatcontributionstodefinedcon- Personalincomegoestohouseholdsandincludes tributionpensionplans,whethermadeby wages,salaries,rents,royalties,dividendsand employeesorbyemployersontheirbehalf,as interest.Theseareallincomeflowsderivedfrom disposablepersonalincomeatthetimesuch currentproductionofgoodsandservices.Dispos- contributionsaremade.Investmentincomeon ableincomeispersonalincomelessdirecttaxes, theseaccountsisalsoaccruedasdisposableper- whichincludeincometaxes,SocialSecurity sonalincome.Investmentincomeconsistsof taxesandthelike. interestanddividendearnings.3Paymentsfrom Alltransactionsinvolvingexchangeofexist- suchfundsarenotcountedascurrentincome, ingassets,however,areexcludedfrommeasured buttreatedasanexchangeofonefinancialasset— incomebecausesuchtransactionsdonothave pensionaccumulations—foranother—cash.While anassociatedproductionofgoodsorservices.If notcompletelyclear,consistencysuggeststhat ahouseholdconsumptionunitconsidersany theunrealizedcapitalgainsaccruedbydefined suchcapitalgainsorlossesasincome,thenthe contributionpensionplansarenotcountedas NIPAframeworkunderestimatesthesavingrate income,noristherealizationofsuchgainsin asperceivedbyhouseholds.And,Imightadd,it benefitpaymentscountedasretirementincome. isperfectlysensibleforahouseholdtoconsider Definedbenefitplansaretreatedsimilarly capitalgainstobeincomeavailabletobespent todefinedcontributionplans.Employercontri- onconsumptiongoods. butionstosuchplansandtheinvestmentincome Hereisoneproblem:Capitalgainstaxesare accruingarecountedaspersonaldisposable considereddirecttaxesandsubtractedfromper- income.Theadministrativeexpensesofsuch sonalincomeincalculatingdisposablepersonal plansareincludedinpersonalconsumption income.Thus,capitalgainsincomeisnotcounted, expenditures.Again,benefitpaymentsare buttaxpaidasaresultofrealizedcapitalgains excludedfrompersonaldisposableincome.4At iscounted.Inanenvironmentwherethereare onepoint,governmentsectordefinedbenefitpen- aggregatetaxablecapitalgains,theNIPAsaving sionplansweretreateddifferentlyfromprivate rateunderestimatesthetruesavingrateasseen sectordefinedbenefitplans.Thatasymmetry byhouseholds. hasnowbeenchangedandtheabovetreatment Thecapitalgainsissuehasgrown,andnot isappliedtoalldefinedbenefitplans. justbecausestockmarketgainshaveoftenbeen Notethatthistreatmentofdefinedbenefit substantialinrecentyears.Acompanycanthrow planscangenerateissuesinthetimingofincome 3 WilliamG.GaleandJohnSabelhous,“PerspectivesontheHouseholdSavingRate,”BrookingsPapersonEconomicActivity,I:1999,p.182. 4 MarshallB.Reinsdorf,“AlternativeMeasuresofPersonalSaving,”SurveyofCurrentBusiness,September2004,p.20. 4 U.S.Saving andsavingsinceatanypointintimesuchplans fortheemployee,nordotheyproduceacharge canbeunder-oroverfunded.Underfunding,and againstprofitandlossfortheemployer.Atthe subsequent“catch-up”payments,defersthe timethatnonqualifiedstockoptionsareexer- recordingofincome.Overfundingresultingfrom cised,thedifferencebetweenthemarketprice capitalgainsontheinvestmentportfoliocanpro- andtheexercisepriceoftheoptionisreported ducetemporaryorpermanentunderstatementof ascapitalgainontheemployee’sincometax incomefromsuchplans. returnandtheemployerreceivesataxdeduction Whilepaymentsoutofdefinedcontribution forthissamedifference.6Inprincipletheexer- anddefinedbenefitpensionplansarenotcounted ciseofanonqualifiedstockoptiongenerates aspersonalincome,suchpaymentsaresubject incomethatisreportedforpurposesofassessing toincometax.Thesetaxpaymentsreducemeas- unemploymentinsurancetaxes.Sincereported uredpersonaldisposableincomeandthesaving unemploymentinsurancewageandsalary rateatthetimethattheretirementbenefitsare incomearethebasicdatausedtoconstructcom- paid.ThelogicofthistreatmentintheNIPAis pensationintheNIPA,taxableoptionincome thatpersonalincomeisrecognizedwhenthe likelyisincludedinpersonalincome.However, retirementplancontributionsareearnedrather incomefromnontaxableincentivestockoptions thanwhenthebenefitsarepaid,oftenmanyyears doesnotgetincludedinpersonalincome. afterbeingearned. Employersmakeinvestmentdecisionsfor NIPA Saving Rate and Deferred definedbenefitplansandretaininvestmentrisk. Compensation Giventhisfact,analternativeapproach,notused Increasinguseofdeferredcompensationcan, intheNIPAaccounts,wouldbe,first,toremove inprinciple,biastheNIPAsavingratedown- employercontributionsaswellasrentalincome, ward.Compensationisaccruedinnational dividendsandinterestaccruingtosuchplans incomeatthetimeitisearned.However,itis frompersonalincomeand,second,remove onlyrecordedinPersonalIncomeatthetime administrativeexpensesfrompersonaloutlays. thatitisreceived.Growingdeferredcompensa- Then,third,underthisalternativetreatment,the tionwouldleadtoanincreasingdiscrepancy benefitspaidoutbydefinedbenefitplanswould betweenaccruedcompensationandreceived beaddedtopersonalincome.Experimentswith compensation.Ifhouseholdsdeterminetheir theseadjustmentsshowthattheimpactonthe consumptionpatternsonthebasisofearned measuredsavingrateisminor,inpartbecauseof compensation(whetherornotitisreceived) thedecreasingimportanceofdefinedbenefit thenthemeasuredsavingrateisbiaseddown- pensionsintheprivatesectorinrecentdecades.5 ward.However,since1959wageaccrualsless disbursementsasrecordedintheNIPAaccounts NIPA Saving Rate and Stock Options haveneverbeenmorethan0.3percentofper- Stockoptionsareaformofdeferredcompen- sonalincome,sodeferredcompensationcannot sationtoemployees.Atthetimetheyare havemadeasignificantcontributiontothenega- granted,theyarenottreatedasgeneratingincome tivetrendinthesavingrate.7 5 JackVanDerhei,CraigCopeland,andDallasSalisbury,RetirementSecurityintheUnitedStates,Washington,D.C.:EmployeeBenefitResearch Institute,2006,pp.44-50. 6 Somestockoptions(incentivestockoptions)donotresultinataxliabilitytotheemployeeorataxdeductionfortheemployer.Itisbelieved thatnonqualifiedstockoptionsarethemostprevalentform.ForadetaileddiscussionofthetreatmentofstockoptionsintheNIPAseeCarol Moylan,“TreatmentofEmployeeStockOptionsintheU.S.NationalIncomeAccounts,”BureauofEconomicAnalysis,U.S.Departmentof Commerce(undated). 7 SeetheNIPAtable:“RelationofGrossDomesticProduct,GrossNationalProduct,NetNationalProduct,NationalIncome,andPersonal Income.” 5 ECONOMICGROWTH NIPA Saving Rate and Purchases of goodsasconsumptionhavebeenconstructed.8 Consumer Durable Goods Thesealternativeestimatesshowahigherlevel ofthesavingrateanddifferentbehaviorofthe Theaccumulationofreproduciblecapital savingrateoverbusinesscycles.However,the goodspurchasedbyhouseholdsisnottreated negativetrendobservedintheNIPAsavingrate symmetricallyintheNIPA.Purchasesofnewly overthepast10to15yearsisevidentinthis constructedhousesareconsideredintheNIPA alternativemeasure. asaninvestmentactivity,notaconsumption expenditureatthetimeofpurchase.Housesare treatedasassetsthatgeneratehousingservicesor A LITTLE BIT OF rentalincomeanddepreciateovertimeregard- CONSUMPTION THEORY lessofwhethertheunitsareowner-occupiedor areownedbyarentalbusiness.Consequently, Thefoundationsofmoderneconomictheory overtheusefullifeofthehouseowner-occupiers ofhouseholdconsumptionbehaviorwereestab- aretreatedasiftheyarebothlandlordsandten- lishedahalfcenturyago.Thepioneeringworkin antsrentingthepropertyfromthemselves.An thisfieldwasdoneinthe1950sandearly1960s imputationismadeforrentfromtheowner- byMiltonFriedmanandbyFrancoModigliani occupiedunit,basedonmarketdatafromcom- andcoauthors.9Thecommonstartingpointfor parablerentalunits.Thisimputedrentisincluded thesetheoriesisthathouseholdsderiveutility, inpersonalconsumptionexpendituresforeach orsatisfaction,fromtheirconsumptionover period.Ontheincomesideoftheaccountsthis multipleperiods.Thetheoreticalconceptof imputationisincludedinrentalincome. consumptionincludestheflowofservicesfrom Inprinciple,theaccountingappliedtohouses consumerdurables,nottheexpenditureson shouldapplytoallpurchasesofdurablegoods durablesmeasuredintheNIPA.InModigilani’s byhouseholds,includingautomobiles,household formulation,themultipleperiodsspanthelife- appliances,electronicequipmentandsoforth. timeofthehouseholdandencompassboth Inpracticeconsumerdurablespurchasesare workingyearsandaperiodofretirement.The treatedlikefoodpurchases—acurrentoutlay motivationassignedtohouseholdsistomaximize thatisquicklyconsumed.Thus,intheNIPApur- theutilityderivedfromconsumptionoverthe chasesofnewlyproducedconsumerdurables multi-periodhorizon.Underthisassumption, areincludedinentiretyatthetimeofpurchase, householdssmooththeirconsumptionovertime. ratherthanimputingaflowofservicesfromthe Facedwithfluctuationsinincome,households assetsovertheirusefullives.Ifcars,forexample, useborrowing,saving,andwealthaccumulation weretreatedlikeahouseintheNIPA,thecon- asthetoolsbywhichtheyachievethedesired sumptionentrywouldbeimputedtransportation timeprofileofconsumption. Inthisframework, services,includingdepreciation,fromcarsrather theprincipaldeterminantofconsumptionisnot thantheamountspentonthecaritselfattimeof currentpersonaldisposableincome,butrather purchase. householdnetworth. Estimatesofthehouseholdsavingratethat AndoandModiglianishowthatthistheoret- treatpurchaseofconsumerdurablegoodsas icalframeworkcanbeusedtoderivehypotheses consumptionandimputeserviceflowsfromsuch aboutaggregatehouseholdconsumptionand 8 See,forexample,MarshallB.Reinsdorf,“AlternativeMeasuresofPersonalSaving,”SurveyofCurrentBusiness,September2004,Chart5,p.23. 9 M.Friedman,ATheoryoftheConsumptionFunction,Princeton,NJ:PrincetonUniversityPress,1957.FrancoModiglianiandRichard Brumberg,“UtilityAnalysisandtheConsumptionFunction:AnInterpretationofCross-SectionData,”inKennethK.Kurihara,ed.,Post KeynesianEconomics,NewBrunswick,NJ:RutgersUniversityPress,1954,pp.388-436.AlbertAndoandFrancoModigliani,“The‘Life- Cycle’HypothesisofSaving:AggregateImplicationsandTests,”AmericanEconomicReview,March1963,53(Part1),pp.55-84. 6 U.S.Saving householdnetworthaccumulation.Inparticular, theseaccountsincludecapitalgains/losseson theyconcludethat,withastableagedistribution bothhousesandequities. ofthepopulationandcontinuingproductivity RaymondGoldsmithalsoconstructedesti- growth,overtimehouseholdnetworthshould matesofnetworthonnonfarmhouseholdsand tendtogrowataconstantrate.10 nonprofitinstitutionsinhismonumentalStudy ofSaving.13Thesedataareavailableforselected yearsfrom1900through1949.AlbertAndoand ANOTHER STORY: E.CaryBrownextendedthesedatatoannualtime seriesfrom1929through1958.14Andoand EVIDENCE FROM THE FLOW Brownalsoconstructedmeasuresofconsumption OF FUNDS ACCOUNTS definedastotalpersonalconsumptionexpendi- Theflowoffundsaccountspublishedbythe tures,lesspersonalconsumptionexpenditures BoardofGovernorsareanothersourceofdataon ondurablegoodsplusdepreciationofthestock personalsavingandhouseholdwealthaccumu- ofdurablegoodsvaluedatreplacementcost.15 lation.Thepersonalsavingconceptinthese Sincethestockofconsumerdurablegoods accountsisconceptuallythesameastheNIPA recordedintheflowoffundsaccountsismeas- measure,butbecausedifferentsourcedataare uredatreplacementcost,acomparableseriesfor used,thenumbersarenotidentical.11Fromthe annualconsumptioncanbeconstructedbysub- flowoffundsaccountsitispossibletoconstruct tractingtheannualchangeintheflowoffunds anotherconceptofpersonalincomeandcon- measureofthestockofconsumerdurablesfrom sumption,andhencesaving,thattreatsexpendi- theannualNIPAtotalpersonalconsumption turesonconsumerdurablesasinvestmentand expendituremeasure.Thetimeseriesofthese measuresconsumptionasaflowofservices,as measuresofannualconsumptiontoend-of-year suggestedbytheoriesofconsumptionbehavior. householdnetworthareshowninFigure2onan Yetanotherperspectiveisavailablefromthe annualandfive-yearaveragebasis.Thebehavior balancesheetdataintheflowoffundsaccounts. ofthesetimeseriesisquiteconsistentwiththe End-of-yearbalancesheetsforthehousehold theorythatconsumptionshouldbeproportional (andnonprofitinstitution)sectorareavailable tonetworth.Therearesmallfluctuationsinthe from1946.End-of-quarterbalancesheetsare seriesfromyeartoyearthatlargelyaverageout availablebeginningin1952.Thesetablescontain overfiveyears.Thenotableexceptionsareduring estimatesofreproducibleassets,financialassets, theGreatDepressionandduringthesecondhalf liabilitiesandnetworthforthevarioussectors.12 ofthe1990swhenhouseholdnetworthchanged Marketvaluesareusedforhousingassetsand rapidlywithmajorstockmarketfluctuations. corporateequityintheseaccounts,thoughhold- Theannualpercentagechangeinboththe ingsofbondsarereportedatfacevalue.Thus, Ando-Brownandtheflowoffundsseries,less thechangesinhouseholdnetworthmeasuredin theDecember-to-DecemberrateofCPIinflation 10SeeAndoandModigliani,footnote30,equation(b),p.77. 11DanielLarkins,“NoteonthePersonalSavingRate,”SurveyofCurrentBusiness,February1999,p.8. 12Forannualdata,seeTableB-100. 13R.W.Goldsmith,D.S.Brady,andH.Menderhausen,AStudyofSavingintheUnitedStates,VolumeIII,Princeton,NJ:PrincetonUniversity Press,1956,TableW-22. 14AlbertAndo,E.CaryBrown,RobertM.SolowandJohnKareken,“LagsinFiscalandMonetaryPolicy,”StabilizationPolicies,Commission onMoneyandCredit,EnglewoodCliffs,NJ:Prentice-Hall,Inc.,1963,p.150. 15AlbertAndo,E.CaryBrown,RobertM.SolowandJohnKareken,“LagsinFiscalandMonetaryPolicy,”StabilizationPolicies,Commission onMoneyandCredit,EnglewoodCliffs,NJ:Prentice-Hall,Inc.,1963,TableI-A1,footnote6,p.152. 7 ECONOMICGROWTH 25.0% 20.0% 15.0% 10.0% 5.0% 0.0% 1 929 1 933 1 937 1 941 1 945 1 949 1 953 1 957 1 961 1 965 1 969 1 973 1 977 1 981 1 985 1 989 1 993 1 997 2001 2005 8 tnecreP Figure 2 Consumption–Household Net Worth Ratio CMC=Ando/BrownDatafromCommissiononMoneyandCreditStudy FoF=FlowofFundsData C/NW(CMC) C/NW(FoF) 5YearAverage(CMC) 5YearAverage(FoF) 15.0 10.0 5.0 0.0 -5.0 -10.0 -15.0 1 930 1 934 1 938 1 942 1 946 1 950 1 954 1 958 1 962 1 966 1 970 1 974 1 978 1 982 1 986 1 990 1 994 1 998 2002 tnecreP Figure 3 Growth of Household Net Worth Less CPI Inflation CMC=Ando/BrownDatafromCommissiononMoneyandCreditStudy FoF=FlowofFundsData AverageCMC(3.0%) RealNetWorthGrowthRate(CMC) AverageFoF(3.6%) RealNetWorthGrowthRate(FoF) U.S.Saving isshowninFigure3.Theseseriesmeasurethe islargelyaconsequenceofcapitalgainsoncor- real(i.e.inflationadjusted)rateofchangeof porateequities.16 householdnetworth.Bothseriesarequitevolatile. Asiswellknown,thestockmarketboom However,neitherseriesexhibitsatrend,andthe collapsedin2001,andequitypricesareonly meansoftheseriesarequiteclose.Themean nowreturningtothepreviouspeaklevels.Sub- realgrowthofhouseholdnetworthfrom1929 sequenttothestockmarketboom,however,there through1958is3.0percent.Themeanrealgrowth wasamajorhousingboomintheUnitedStates oftheflowoffundsmeasureofhouseholdnet bothintermsofconstructionandpropertyvalues. worthfrom1946through2005is3.6percent. Itismoreproblematictoarguethattherecent Again,thedataareconsistentwiththebroad growthinhouseholdnetworth,supportedby implicationsofthereceivedeconomictheoryon capitalgainsinhousing,reflectsimprovedpro- thedeterminantsofaggregateconsumption. ductivitytrends.Nevertheless,inthepastseveral Thereisnotuniformagreementamongecon- years,consumptiondemandappearstohave omiststhatincreasesinnetworthgeneratedby respondedtothecapitalgainsinhousing.There capitalgainsshouldbeconsideredsaving.Some issubstantialevidenceof“equityextraction”by arguethattherelevantcapitalgainsarethose homeownersduringtherecenthousingboom, generatedbyincreasedproductivityoftheunder- andcross-sectionevidenceoflargeresponsesof lyingasset,andcapitalgainsthatdonotcon- consumptionbysomegroupsofhouseholdsto tributetoincreasedfutureincomeshouldnotbe increasesinhousevalues.17 treatedassaving.Ofcourse,thereisnoobvious Asanalternativemeasureofthesavingrate, waytoseparateobservedcapitalgainsintothose wecanconsidertheratioofthechangeinhouse- generatedbyhigherproductivityofcapitaland holdnetworthfromtheflowoffundsaccounts totheNIPAmeasureofpersonaldisposable thosethatcouldresultfromchangesintastesor income.Theannualandfive-yearaveragedata riskpremiums.Thereisnowgeneralagreement forthisseriesareshowninFigure4.Yeartoyear, amongeconomiststhatasignificantincreasein themovementsarequitevolatile,butthenega- trendproductivityoccurredintheU.S.economy tivetrendcharacteristicoftheheadlinesaving startingaroundthemid1990s.Amajorstock rateisnotpresent.Indeed,inthelate1990sthis marketboomoccurredinthelate1990s. ratiojumpedup,reflectinglargecapitalgainson Overmostofthepost-WorldWarIIperiod, corporateequities,fellin2000-2001reflecting thepersonalsavingrateaveragedabout6percent, theendofthestockmarketboom,buthasmoved withsomehigheryearsfromthemid-1970sto aboveitslong-termaveragein2003-2005. mid-1980s.ThenegativetrendintheNIPAsaving ratestartedinthemid-1990s,aboutthesametime thestockmarketboomstarted.Thusitishardto CAPITAL GAINS AND LONG-TERM dismissthehypothesisthatthedeclineinthe REAL INTEREST RATES measuredsavingrateinthelate1990sreflected theresponseofconsumptiontolargecapitalgains Theevidencepresentedabovesuggeststhat fromcorporateequity.Evidencefrompaneldata thebehaviorofaggregateconsumptioninthe ofhouseholdsalsosupportstheconclusionthat UnitedStatesrelativetohouseholdnetworth thedeclineinthepersonalsavingratesince1984 overthepasttwodecadesisconsistentwithlong- 16F.ThomasJuster,JosephP.Lupton,JamesP.Smith,andFrankStafford,“TheDeclineinHouseholdSavingandtheWealthEffect,”Boardof GovernorsoftheFederalReserveSystem,FEDSDiscussionPaper2004-32. 17AndreasLehnert,“Housing,Consumption,andCreditConstraints,”BoardofGovernorsoftheFederalReserveSystem,FEDSDiscussion Paper2004-63.AlanGreenspanandJamesKennedy,“EstimatesofHomeMortgageOriginations,Repayments,andDebtonOne-to-Four- FamilyResidences,”BoardofGovernorsoftheFederalReserveSystem,FEDSDiscussionPaper2005-41. 9 ECONOMICGROWTH establishedpatterns.Thus,mostoftheobserved expectations.SuchdataareshowninFigure5. negativetrendintheNIPAsavingrateseemsto Thefirstseriesshownthereisthe10-yearTreasury beattributabletotheomissionofcapitalgains constantmaturityyieldlessthe10-yearinflation andlossesfrommeasuredpersonaldisposable expectationreportedintheSurveyofProfessional income.Theevidencesuggeststhatitisnota Forecasters.18Thesecondseriesisthemonthly coincidencethatincreasesinhouseholdwealth averageyieldonthe10-yearU.K.indexedbond andalong-termdownwardtrendintheNIPA forthemiddlemonthofeachquarter.Thethird savingratehaveoccurredtogether.Increasesin seriesisthemonthlyaverage10-yearconstant householdwealthhavebeendrivenbyincreases maturityTreasuryyield,againforthemiddle instockpricesafter1982,albeitwithsomesig- monthofeachquarter.19Theseriesinferredfrom nificantfluctuations,andbythemorerecent theSurveyofProfessionalForecastersisquite increaseinhomevaluesafter2001. consistentwiththemarketyieldontheindexed Thisobservation,however,leavesaquestion: bondfortheperiodswherebothseriesare Whyhaveweobservedsuchhighcapitalgains available. oncorporateequitiesandhousing?Partofthe Thesedatasuggestthatasignificantdecline storyisthesubstantialdeclineininflationand inlong-termrealinterestratesstartedintheearly greaterconfidenceinsustainedlowinflation 1990s,whentheestimatedrealyieldaveraged after1982.Atthattime,also,changesintaxand around3.5percent.Morerecently,theaverage otherpoliciescontributedtohighereconomic realyieldhasbeenaround2percent.Thisnega- growthandincreasesincorporateprofits,which tivetrendisalsovisibleintheU.K.indexedyield, havebeengrowingmostyearssince1982.But whichdeclinedrapidlyinthelate1990sfrom anotherpartofthestorymaybeadownward around3.5percentto2percentorbelowinrecent trendinrealinterestratessincethelate1990s. years.Adeclineinglobalrealratesofthissize, Decliningrealinterestratesincreaseassetvalues. ifexpectedtobepermanent,shouldproducea Recentyieldsonindexedbondsbothinthe majorupwardrevaluationofthevalueoflong- UnitedStatesandgloballyappearverylow,rela- livedassetssuchascorporateequitiesandhous- tivetotheconventionalwisdomonthehistori- ing.Hencetheexplanationoftheobservedtrend calbehaviorofrealyields.Withtheexceptionof intheconventionalsavingrateintheU.S.can theUnitedKingdom,whereindexedbondswere likelybetraced,inconsiderablepart,toglobal introducedintheearly1980s,marketsininfla- changesinrealratesofinterest. tionindexeddebtarerelativelynew.TheU.S. Wheredotheseobservationsleaveus?I’ll Treasurystartedissuingsuchsecuritiesin1997. offertwotentativeconclusions.First,household TheFrenchintroducedabondthatisdenomi- savingbehaviordoesnotseemtohavechanged natedinEurosandindexedtotheEuroareahar- inanyfundamentalway.Whathaschangedtoa monizedCPIinflationin2002.Forthepastfour degreeisthetrendinassetvalues.Households years,therealyieldonthesebondshasfluctuated haveconsumedsomeoftheincreaseinassetval- aroundtwopercentorslightlyhigher. uesinaboutthesamewaytheyalwayshave. ToinferthebehaviorofU.S.realyieldssince Mysecondtentativeconclusionisthatthe theearly1990s,wecanuseU.K.indexedbond behaviorofhouseholds,thoughperfectlysensi- dataandassumethatU.S.yieldshavemovedin bleandresponsibleforhouseholdsasawhole, similarfashionorwecanuseU.S.datathatare hasledtoasituationinwhichtheUnitedStates constructedfromsurveymeasuresofinflation asawholeissavingtoolittleofitsnationalout- 18TheSurveyofProfessionalForecastersdataarecollectedonceaquarterbytheFederalReserveBankofPhiladelphia.Thedataareavailable sinceNovember1991.Thedataplottedarethemonthlyaverage10-yearconstantmaturityTreasuryyieldforthemiddlemonthofeachquarter lessthe10-yearinflationexpectationreportedintheSPF. 19ThoughindexedTreasurieswerefirstissuedin1997,the10-yearconstantmaturityyieldisconstructedonlysince2002. 10 U.S.Saving 80.0% 60.0% 40.0% 20.0% 0.0% -20.0% -40.0% 1 946 1 949 1 952 1 955 1 958 1 961 1 964 1 967 1 970 1 973 1 976 1 979 1 982 1 985 1 988 1 991 1 994 1 997 2000 2003 11 tnecreP Figure 4 Changes in Household Net Worth Relative to Personal Disposable Income AnnualData Five-yearAverageData 6 5 4 3 2 1 0 Nov-91 Nov-92 Nov-93 Nov-94 Nov-95 Nov-96 Nov-97 Nov-98 Nov-99 Nov-00 Nov-01 Nov-02 Nov-03 Nov-04 Nov-05 Nov-06 tnecreP Figure 5 10-Year Real Interest Rate Measures US10Year-SPF10YearInflation US 10YearIndexed UK10YearIndexed ECONOMICGROWTH put.U.S.domesticinvestmenthasnotsuffered, becausecapitalhasbeenflowingintotheUnited Statesfromabroad.However,atsomepointthe U.S.netinternationalinvestmentpositionwill stopbecomingevermorenegative.U.S.saving willthenfinancealargerfractionofU.S.domes- ticinvestmentand,perhaps,repurchasesome U.S.assetsnowheldbyinternationalinvestors. Thereisnoreasonwhythisadjustmentshould bedifficultordisorderly,butitwillrequirethat U.S.consumptionoutlaysexpandmoreslowly thanU.S.GDPforatime. 12

Cite this document

APA

William Poole (2007, February 14). Speech. Speeches, Federal Reserve. https://whenthefedspeaks.com/doc/speech_20070215_poole

BibTeX

@misc{wtfs_speech_20070215_poole,
  author = {William Poole},
  title = {Speech},
  year = {2007},
  month = {Feb},
  howpublished = {Speeches, Federal Reserve},
  url = {https://whenthefedspeaks.com/doc/speech_20070215_poole},
  note = {Retrieved via When the Fed Speaks corpus}
}