Economic Growth - Our World in Data

PPP-adjusted GDP: spatial comparisons InthenewsExploredataforcontextofthewarinUkraineEconomicGrowthThehistoryofeconomicgrowthFrompovertytoprosperity:TheUKoverthelongrunTheeconomybeforeeconomicgrowth:TheMalthusiantrapEconomicgrowthoverthelongrunTheworldeconomyoverthelasttwomillenniaGrowthatthetechnologicalfrontierandcatch-upgrowthGDPGrowthsince1950Whichcountriesachievedeconomicgrowth?Andwhydoesitmatter?DifferentdatasetsongrowthinthelastdecadesTotaloutputofeconomiesProductivityincreasesmakeeconomicgrowthpossibleRisingproductivitybysectorUrbanizationandprosperityReligiosityandprosperityDataQuality&DefinitionsGDPpercapitaandaverageincomesHowareincomesadjustedforinflation?InternationalGDPComparisons:marketvs.PPPexchangeratesPPP-adjustedGDP:spatialcomparisonsHistoricalreconstructionsofnationalaccounts–thecaseoftheUKDataSourcesLong-rundatasetsPost-1800Post-1950OtherGDP-relateddataEndnotesLicenceCitationContentsEconomicgrowthdescribesanincreaseinthequantityandqualityoftheeconomicgoodsandservicesthatasocietyproducesandconsumes.Whilethedefinitionofeconomicgrowthisstraightforward,itisextremelydifficulttomeasureit.Growthisoftenmeasuredasanincreaseinhouseholdincomeorinflation-adjustedGDP,butitisimportanttokeepinmindthatthisisnotthedefinitionofit–justlikelifeexpectancyisameasureofpopulationhealth,butcertainlynotthedefinitionofpopulationhealth.Incomemeasuresaremerelyonewaytounderstandtheeconomicinequalitybetweencountriesandthechangingprosperityovertime.TheGrossDomesticProduct(GDP)ofaneconomyisameasureoftotalproduction.Moreprecisely,itisthemonetaryvalueofallfinalgoodsandservicesproducedwithinacountryorregioninaspecifictimeperiod.Comparisonsovertimeandacrossbordersarecomplicatedbyprice,qualityandcurrencydifferences,asexplainedbelow.Fromthelong-termperspectiveofsocialhistory,weknowthateconomicprosperityandlastingeconomicgrowthisaveryrecentachievementforhumanity.Inthisentrywewillalsolookatthismorerecenttimeandwillalsostudytheinequalitybetweendifferentregions–bothinrespecttotheunequallevelsofprosperitytodayandtheunequaleconomicstartingpointsforleavingthepovertyofthepre-growthpast.AllourchartsonEconomicGrowth AccesstoafinancialaccountorservicesAccesstoelectricityvs.GDPpercapitaAnnualgrowthofGDPpercapitaAverageGDPgrowthrate1960-2011versusGDPpercapitain1960Averageyearsofschoolingvs.GDPpercapitaAverageyearsofschoolingvs.expectedyearsofschoolingCO₂emissionspercapitavsGDPpercapitaChildmortalityvsGDPpercapitaChildrenperwomanbyGDPpercapitaCompositionofnationalgrossdomesticproductbysectorDailypercapitaincomes:Meanvs.MedianExpectedyearsofschoolingvs.GDPpercapitaGDPMaddison2020worldregionsstackedareaGDPinBritishpoundsGDPinEngland(since1270)GDPpercapitaMaddisonProjectDatabase2020(map)GDPpercapitaPennWorldTablesGDPpercapitaWorldBankGDPpercapitaMaddison(2020)-linechartGDPpercapita(inflation-andPPP-adjusted):WorldBankdatavs.PennWorldTabledataGDPpercapitafromtheWorldBankvs.GDPpercapitafromtheMaddisonProjectGDPpercapitainEnglandGDPpercapitainUS$WorldBankGDPpercapitavspopulationdensityGDPpercapitavs.DailymedianincomeGDPperpersonemployedGlobalHungerIndexvs.GDPpercapitaGovernmentspendingvsGDPpercapitaGrossDomesticProductWorldBankGrowthofGDPandtradeGrowthrateofrealGDPperemployedpersonHiddenHungerIndexvs.GDPpercapitaHistoricalIndexofHumanDevelopment(withoutGDPmetric)vs.GDPpercapitaHistoricalIndexofHumanDevelopmentvs.GDPpercapitaHistoricalIndexofHumanDevelopmentwithGDPmetricvs.withoutGDPmetricHumanDevelopmentIndexvs.GDPpercapitaLaborforceparticipationrateofmen,65yearsandolderintheUSALifeexpectancyvs.GDPpercapitaLiteracyratevs.GDPpercapitaMedicaldoctorsper1,000peoplevs.GDPpercapitaNationalGDPNationalpovertylinesvs.GDPpercapitaNominalwages,consumerprices,andrealwagesintheUKOutputofkeyindustrialsectorsinEnglandandtheUKOutputofkeyindustriesinEnglandOutputofkeyserviceandindustrialsectorsinEnglandandtheUKPricelevelrelativetotheUSvsGDPpercapitaProductivityperhourworkedRealGDPpercapitaPennWorldTablesSelf-reportedlifesatisfactionvsGDPpercapitaSettlermortalityfacedbyEuropeansettlersatthetimeofcolonizationvsGDPpercapitatodayShareofadultswhosmokevsGDPpercapitaShareofchildrenwhoarestuntedvsGDPpercapitaShareofconsumerexpenditurespentonfoodvs.GDPpercapitaSharesofGDPbyeconomicsectorTaxrevenuevsGDPpercapitaThedeclineofchildmortalitybylevelofprosperityUrbanpopulationvs.GDPpercapitaWorldBank'sIncomeGroupsWorldGDPoverthelasttwomillennia ThehistoryofeconomicgrowthInthissectionFrompovertytoprosperity:TheUKoverthelongrunTheeconomybeforeeconomicgrowth:TheMalthusiantrapEconomicgrowthoverthelongrunTheworldeconomyoverthelasttwomillenniaGrowthatthetechnologicalfrontierandcatch-upgrowthFrompovertytoprosperity:TheUKoverthelongrunTheUKisparticularlyinterestingasitwasthefirsteconomythatachievedsustainedeconomicgrowthandtherebypreviouslyunimaginableprosperityforthemajorityofthepopulation.OutputpercapitaoftheUKeconomyThechartbelowshowsthereconstructedGDPpercapitainEnglandandtheUKoverthelast7centuries.Economichistoryisaverysimplestory.Itisastorythathasonlytwoparts:Thefirstpartistheverylongtimeinwhichtheaveragepersonwasverypoorandhumansocietiesachievednoeconomicgrowthtochangethis.Incomesremainedalmostunchangedoveraperiodofseveralcenturieswhencomparedtotheincreaseinincomesoverthelast2centuries.Lifetoochangedremarkablylittle.Whatpeopleusedasshelter,food,clothing,energysupply,theirlightsourcestayedverysimilarforaverylongtime.Almostallthatordinarypeopleusedandconsumedinthe17thcenturywouldhavebeenveryfamiliartopeoplelivingathousandorevenacoupleofthousandyearsearlier.Averageincomes(asmeasuredbyGDPpercapita)inEnglandbetweentheyear1270and1650were£1,051whenmeasuredintoday’sprices.Thesecondpartismuchshorter,itencompassesonlythelastfewgenerationsandisradicallydifferentfromthefirstpart,itisatimeinwhichtheincomeoftheaveragepersongrewimmensely–fromanaverageof£1051incomesperpersonperyearincreasedtoover£30,000a29-foldincreaseinprosperity.ThismeansanaveragepersonintheUKtodayhasahigherincomeintwoweeksthananaveragepersoninthepasthadinanentireyear.Sincethetotalsumofincomesisthetotalsumofproductionthisalsomeansthattheproductionoftheaveragepersonintwoweekstodayisequivalenttotheproductionoftheaveragepersoninanentireyearinthepast.Thereisjustonetrulyimportanteventintheeconomichistoryoftheworld,theonsetofeconomicgrowth.Thisistheonetransformationthatchangedeverything.AsthischartoftotalGDPintheEnglandoversevencenturiesshows,theincreaseofthetotaloutputoftheUKeconomygrewbyevenlargerextent,becausenotonlyaverageincomesincreasedsincetheonsetoftheIndustrialRevolution,butthenumberofpeopleinthecountryincreasedaswell. Clicktoopeninteractiveversion Theeconomybeforeeconomicgrowth:TheMalthusiantrapThepre-growtheconomywasazero-sum-game:LivingstandardsweredeterminedbythesizeofthepopulationInthepreviouschartwesawthatitwasonlyafter1650thatlivingstandardsintheUKdidstarttoincreaseforasustainedperiod.Beforethemoderneraofeconomicgrowththeeconomyworkedverydifferently.Nottechnologicalprogress,butthesizeofthepopulationdeterminedthestandardsofliving.IfyougobacktothechartofGDPpercapitaintheEnglandyouseethatearlyinthe14thcenturytherewasasubstantialspikeinthelevelofincomes.Incomesincreasedbyaroundathirdinaperiodofjustafewyears.Thisistheeffectthattheplague–theBlackDeath–hadontheincomesoftheEnglish.Theplaguekilledalmosthalf(!)oftheEnglishpopulation.Thepopulationdeclinedfrom8millionto4.3millioninthethreeyearsafter1348.Weevenseeitinthechartfortheworldpopulation.Butthosethatsurvivedtheepidemicweremateriallymuchbetteroffafterwards.Theeconomywasabrutalzero-sumgameandthedeathofyourneighbourwastothebenefitforthosethatdidsurvive.Thishappenedprimarilybecausefarmersnowachievedanhigheroutput.Whilefarmersbeforetheplaguehadtouseagriculturallandthatwaslesssuitedforfarming,afterthepopulationdeclinetheycouldfarmonthemostproductiveareasoftheisland.IntheverylongtimeinwhichhumanitywastrappedintheMalthusianeconomyitwasbirthsanddeathsthatdeterminedincomes.Morebirths,lowerincomes.Moredeaths,higherincomes.Weseethiscouplingofincomeandpopulationinthechartbelowthatplotsthesizeofthepopulation(onthex-axis)againstthetotaloutputoftheEnglisheconomy(toppanel)andagainsttheincomeperperson(bottompanel).LookingatthebottompanelweseethespikeofincomesthatwasassociatedwiththekillingofhalfofthepopulationintheBlackDeath.Afterthisthepopulationandtheincomeperpersonstagnateuntilaround1500.Inthefollowingperiodweseetheeconomygrowing–totalGDPincreasesbymorethan280%from1500to1650–butthisincreaseinoutputisnotassociatedwithanincreaseinincomeperperson,butonlyanincreaseofthetotalpopulationoftheUK.Itisonlyafter1650thattheEnglisheconomybreaksoutoftheMalthusianTrapandthatincomesarenotdeterminedbythesizeofthepopulationanymore.Fortheperiodafter1650weseethatboththepopulationandtheincomeperpersonaregrowing.Theeconomyisnotazero-sumgameanymore;economicgrowthmadeitapositive-sumgame.WhenMalthusraisedtheconcernsaboutpopulationgrowthin17981hewaswrongabouthistimeandthefuture,buthewasindeedrightinhisdiagnosisofthedynamicsofhispast.TheworldbeforeMalthuswasMalthusianandpopulationincreaseswereassociatedwithdecliningnutrition,declininghealth,anddecliningincomes.TheworldafterMalthusbecameincreasinglylessMalthusian.WhatMalthusdidnotforeseewasthattheincreasingoutputoftheeconomywilldecouplefromthechangeofthepopulationsothattheoutputavailableforallwillincreaseoveralongperiod.Thisdecouplingofincomeandpopulationisshowninthechart.Technologicalchangeinthepre-growtheconomyTechnologicalinnovationthatincreasesproductivityisthekeytoincreasedprosperity.Butthereweretechnologicalbreakthroughsbeforethe17thcentury.Windmills,irrigationtechnology,andalsonon-technicalnoveltiesespeciallythenewcropsfromtheNewWorld.Whydidthesenotleadtosustainedeconomicgrowth?Whathappenedasaconsequenceoftheseinnovationswereindeedincreasesinproductivity,andtheoutputincreasesledtoincreasedprosperity.Butonlyforashorttime.ImprovementsintechnologyhadadifferenteffectintheMalthusianpre-growtheconomy.Theyraisedlivingstandardsonlytemporarilyandinsteadraisedthesizeofthepopulationpermanently.TheeconomichistorianGregoryClarksumsitupcrisply:“Inthepreindustrialworld,sporadictechnologicaladvanceproducedpeople,notwealth.”2Technologicalimprovementsleadtolarger,butnotricherpopulations.Ifthisanalysisofthepre-growtheconomyistruethanwewouldexpecttoseeapositivecorrelationbetweenproductivityandthedensityofthepopulation.AshrafandGalor(2011)3studiedtheMalthusianeconomytheoreticallyandempiricallyinapaperpublishedintheAmericanEconomicReview.Thechartbelowistakenfromtheirpublicationandconfirmsthetheoreticalpredictionforthepre-growtheconomiesintheyear1500.Allthedataisreportedinthecurrentbordersoftheworld.Onthex-axisofbothchartsyoufindthesamemetric:Theproductivityoftheagriculturallandasmeasuredbythequalityofthesoilandtheclimate.Onthechartontheleftweseethatthoseworldregionswithalowproductivityofagriculturehadverylowpopulationdensity.Theregionswiththehighestpopulationdensityontheotherhandareallregionswithveryproductiveland.Andonthechartontheleftweseethatthehigherproductivityofthelanddidnotmatterforpeople’slivingstandards.TheagriculturalsectorinSpain,India,orMoroccowasmuchmoreproductivethaninFinland,Egypt,andNorway,butthepeoplewerenotbetteroff.Themoreproductiveregionswerethemorepopulousregionsandthepeopletherehadtosharewithsomanysothateveryoneremainedatdismallevelsofprosperity.Inthelonghistorybeforemoderneconomicgrowth,higherproductivityleadtolarger,butnotricherpopulations.IntheMalthusianEconomyproductivityproducespeoplenotprosperity4Incomeswerenotflat–HistorysawseveralepisodesofgrowthwhichwerenotsustainedThroughouthistorytherewereseveralepisodesinwhichcertaineconomiesachievedeconomicgrowth,butincontrasttothesustainedgrowthsincetheIndustrialRevolutiontheseepisodeswereallshort-lived.Whatisnewaboutmoderntimesisthatthegrowthofincomeslastedforaverylongtime–untiltoday–andthatthisgrowthdidnotonlyincreasetheincomesinoneeconomy,butinsteadspreadtoothereconomiesaswell.TheoriginofthistransformationisNorth-WesternEurope.ItwasinEngland(andHolland)intheearly17thcenturywhereitbecamefirstpossibletogrowincomesoverasustainedperiodoftime.Thechartshowsthis.Inthelongtimebeforesustainedeconomicgrowthincomesneverexceeded$3.50perday[3.50*365=1277.5]inpricesof1990.5FortheUKthischangesinthe17thcentury,thefluctuationofincomesthatweseeinthefourprecedingcenturiesgivewaytoasteadyincreaseofaverageincomes.GDPpercapitainEuropeaneconomies6EconomicgrowthoverthelongrunDataoneconomicgrowthisnowroutinelypublishedbystatisticaloffices,butresearchershavehadtoreconstructaccountsoftheeconomicproductivityforthepast.ThereareseveralreconstructionsofGDPpercapitaoverthelastcenturies;mostwidelyusedwereforalongtimethereconstructionsbytheBritisheconomistAngusMaddison.MaddisonwasworkinginGroningenintheNetherlandsandafterhisdeathin2010theGroningenGrowthandDevelopmentCentreiscontinuingthiswork.Maddisonattemptedtoreconstructeconomicgrowthinallregionsoftheworldandsomeoftheestimates,especiallyinearlypublications,weremorecrudethanothers.Inrecentyearsseveralresearchteamshaveproducedseveralmuchmoredetailedreconstructionsofeconomicgrowthoverthelongrun.Theseresearchersputextensiveworkinthesereconstructionsandtypicallyfocusedononeparticularregionorcountryonly.Anexamplearethelong-runreconstructionsoftheeconomyofGreatBritainatthebeginningofthisentry.ThesuccessorsofMaddisoninGroningenhavethenextendedhisoriginalworkbycombiningthemwiththemanynewreconstructionsthatwerepublishedinrecentyears.ThisprojectisreferredtoastheMaddisonProjectDatabaseandthisisthemainsourceoflong-runreconstructionsofeconomicgrowthusedtoday.GDPpercapitaoverthelongrunThetwochartsbelowpresenttheestimatesofeconomicprosperityoverthelongrun,astheywerepublishedbytheMaddisonProjectDatabaseintheir2020update.7Thefirstchartshowsthetheeconomicgrowthofdifferentworldregionssince1820.ThedatasetalsoincludesestimatesofGDPpercapitaforindividualcountries–someofwhichextendevenfurtherbackinhistory,asshowninthesecondchart.Youcanaddfurthercountriesbyselecting‘ Addcountry ’.Whatwelearnfromthesechartsisthatonaveragethepeopleofthepastweremanytimespoorerthanwearetoday.In1820theglobalGDPpercapitaisestimatedtohavebeenaround1,102international-$peryearandthisisalreadyaftersomeworldregionshadachievedsomeeconomicgrowth.Forallthehundreds,andreallythousands,ofyearsbefore1820,theaverageGDPpercapitawasevenlower.Prosperityisaveryrecentachievementthatdistinguishesthelast10or20generationsfromalloftheirancestors.In2018,theaverageGDPpercapitawas15,212international-$–almost15-timestheaverageofthepast.Ifwecomparetheeconomicprosperityofeveryregiontodaywithanyearliertimeweseethateverysingleregionisricherthaneverbeforeinitshistory.Economicgrowth,however,hasnothappenedequallyasfastinallregions.Thishascreatedsubstantialinequalitiesglobally,whichwestudyinmoredetailinourentryonglobalincomeinequality.Economicgrowthisaveryrecentphenomenon–wealreadysawthisinthedatathatwediscussedearlierinthisentry.Itistruethatinthepre-growtherasomepeoplewereverywelloff–butthiswasthetinyeliteofthetriballeaders,pharaohs,kingsandreligiousleaders.Whilstglobalinequalitieswerelowerinaworldwheresustainedeconomicgrowthhadyettooccuranywhere,economicinequalitywithinpre-modernsocietieswasextremelyhighandtheaveragepersonwaslivinginconditionsthatwewouldcallextremepovertytoday.Thedestitutionofthecommonmanonlychangedwiththeonsetofeconomicgrowth.Thetimewhenthischangehappenedinvariouscountriescanbeseeninthesetwocharts.Economicprosperitywasonlyachievedoverthelastcoupleofhundredyears.Infact,itwasmostlyachievedoverthesecondhalfofthelasthundredyears.Theriseofglobalaverageincomes–globalGDPpercapita–showsthattheworldeconomyhasmovedfromazero-sumgametoapositive-sumgame.Thismadeitpossiblethatwhenpeopleinoneplacebecamericher,otherpeopleinotherplacescouldbecomericheratthesametime. Clicktoopeninteractiveversion Clicktoopeninteractiveversion TheworldeconomyoverthelasttwomillenniaThetotaloutputoftheworldeconomyoverthelasttwothousandyearsAndfinallyIhaveincludedonereconstructionofglobaltotalGDPoverthelasttwomillennia.ForthisIhaveusedrecentdatafromtheWorldBankfortheperiodfrom1990onwardandforthehistoricalestimatesbefore1990IrelyonthehistoricalestimatesconstructedbytheeconomichistorianAngusMaddison.8 Clicktoopeninteractiveversion Growthatthetechnologicalfrontierandcatch-upgrowthGrowthatthetechnologicalfrontierThefollowingchartshowseconomicgrowthintheUSAadjustedforinflation.GDPpercapitaintheUSAattheeveofindependencewasstillbelow$2,500–adjustedforinflationandmeasuredinpricesof2011itisestimatedto$2,419.In2018–roughly240yearsafterindependence–GDPpercapitahasincreasedbymorethan20timesto$55,335.Thismeansthattheoutputperpersoninoneyearinthepastwaslessthantheoutputoftheaveragepersoninthreeweekstoday.Itisremarkablehowsteadyeconomicgrowthwasoverthisverylongperiod.From1870to2018GDPperpersonintheU.S.economyhasgrownonaverageat1.67percentperyearwithonlyveryshortdeviationsfromthisverysteadytrend. Clicktoopeninteractiveversion CatchupgrowthcanbefastThechartbelowcomparestheeconomicgrowthatthetechnologicalfrontierwiththegrowthofcountriesthatarefurtherawayfromthetechnologicalfrontier.InthischartthesteepnessofthegrowthpathcorrespondstothegrowthrateasGDPpercapitaisplottedonalogarithmicaxis.Economiesthatarefarawayfromthetechnologicalfrontiercangrowveryrapidly.Catchupgrowthcanbemuchfasterthangrowthatthetechnologicalfrontier. Clicktoopeninteractiveversion GDPGrowthsince1950InthissectionWhichcountriesachievedeconomicgrowth?Andwhydoesitmatter?DifferentdatasetsongrowthinthelastdecadesTotaloutputofeconomiesWhichcountriesachievedeconomicgrowth?Andwhydoesitmatter?Theaveragepersonintheworldis4.4-timesricherthanin1950.Butbeyondthisglobalaverage,howdidincomeschangeincountriesaroundtheworld?Whogainedthemost,whogainedtheleast?Andwhyshouldwecareaboutthegrowthofincomes?ThesearethequestionsIanswerinthispost.ThechartshowsthelevelofGDPpercapitaforcountriesaroundtheworldbetween1950and2016.Ontheverticalaxisyouseethelevelofprosperityin1950andonthehorizontalaxisyouseeitfor2016.GDP–GrossDomesticProduct–measuresthetotalproductionofaneconomyasthemonetaryvalueofallgoodsandservicesproducedduringaspecificperiod,mostlyoneyear.DividingGDPbythesizeofthepopulationgivesusGDPpercapitatomeasuretheprosperityoftheaveragepersoninacountry.Becauseallexpendituresinaneconomyaresomeoneelse’sincomewecanthinkofGDPpercapitaastheaverageincomeofpeopleinthateconomy.HereatCore-Econyoufindamoredetaileddefinition.Lookattheworldaverageinthemiddleofthechart.Theincomeoftheaveragepersonintheworldhasincreasedfromjust$3,300in1950to$14,574in2016.Theaveragepersonintheworldis4.4-timesricherthanin1950.Thistakesintoaccountthatthepricesofgoodsandserviceshaveincreasedovertime(itisadjustedforinflation;otherwisethesecomparisonswouldbemeaningless).Similarly,toallowustocompareprosperitybetweencountries,allincomesareadjustedfordifferencesinthecostofgoodsbetweendifferentcountries(usingpurchasingpowerparityconversionfactors).Asaconsequenceofthesetwoadjustmentsincomesareexpressedininternational-dollarsin2011prices,whichmeansthattheincomesarecomparabletowhatyouwouldhavebeenabletobuywithUS-dollarsintheUSAin2011.EconomicgrowthmeansthatpopulationgrowthandrisingprosperitycangotogetherWhiletheglobalaverageincomegrew4.4-fold,theworldpopulationincreased3-fold,fromaround2.5billiontoalmost7.5billiontoday.9It’seasytomisswhatthismeans:Hadtheworldeconomynotgrown,a3-foldincreaseoftheworldpopulationwouldhavemeantthatonaverageeveryoneintheworldwouldnowbe3-timespoorerthanin1950.Theaverageincomeintheworldwouldhavefallento$1,100.Beforeeconomicgrowththeworldwasexactlythis:azero-sumgameinwhichmorepeoplemeantlessforeveryoneelse,andifonepersonisbetteroffinastagnatingeconomythenthatmeansthatsomeoneelseneedstobeworseoff(Iwroteaboutithere).Whateconomicgrowthmakespossibleisthateveryonecanbecomebetteroff,evenwhenthenumberofpeoplethatneedtobeservedbytheeconomyincreases.10Analmost3-foldincreaseofthepopulationmultipliedbya4.4-foldincreaseinaverageprosperitymeansthattheglobaleconomyhasgrown13-foldsince1950.11Whichcountriesachievedgrowth?TheworstofftodayarethosethatlaggedbehindandremainedinpovertyIncomesdidnotgroweverywhereintheworld.Ifyoulookatthepoorestcountriesintheworldtodayyou’llseethatthesecountriesdidnotstandoutbackin1950;theirincomeswereaslowastheincomesofmanyothercountriesintheworld.Buttodaytheydo–whilemanyeconomiesachievedstronggrowthsomestagnatedaroundtheirlevelfrom1950.Thesearethecountriesthatremainedatthebottomofthechart.Thedifferencebetweenstagnationorevendeclineinsomeplacesandrapidgrowthinotherplacesleadtoadramaticincreaseininequalityintheworld.Norwegiansarenowonaveragemorethan100-foldricherthanpeopleinLiberia,Burundi,andtheCentralAfricanRepublic.Thisfailuretogrowtheeconomyandtoprovidethegoodsandservicesthattheyneedisoneofthelargestfailuresinrecentdecades.Itmeansthatpopulationsintheseplacesarenowmuchworseoffthanpeopleintherestoftheworld–theyarelesshealthyanddiesooner,educationispoorer,andmanysufferfrommalnutrition.12InmostplacespeopletodayaremuchricherthanourancestorsEconomicgrowthhasallowedustobreakoutoftheconditionsofthepastwheneveryonewasstuckinpoorhealth,hardandmonotonouswork,andmalnutrition.Thechartshowsalleconomiesthathaveachievedgrowthsince1950abovethediagonal45°-line.Taiwanisoneofthemostimpressiveexamples.TheTaiwanesehadanincomeof$1,400in1950.AllcountriesdirectlybelowTaiwan–Malta,Bolivia,SierraLeone,andtheDemocraticRepublicofCongoforexample–weresimilarlypoorin1950.By2016GDPpercapitainTaiwanhadincreasedto$42,300.TheTaiwanesearenowamongtherichestpeopleintheworld,30-timesricherthantheywerein1950.Itishardtoimaginewhatthismeantforlivingconditionsinthecountry.Totakejustoneexample.EverysixthchildborninTaiwanin1950diedbeforeitwasfiveyearsold(13%).Todaythechildmortalityratehasdeclinedtohalfapercent(1-in-200children).Anotherwaytolookatitistostartwiththerichestpeopleinthepast–shownfurthesttotherightinthechart.In1950thecountrywiththehighestaverageincomewastheUSAwithaGDPpercapitaof$15,241(andtheyhadjustbecameprosperousinthefewdecadesbefore;beforesomeeconomiesachievedsustainedeconomicgrowth,incomedifferencesbetweendifferentregionswereverysmallandthevastmajorityofpeoplewereextremelypoor).Ifyoulookatincomestodaythenyoufindthattheincomeintherichestcountryin1950isveryclosetotheaverageincomeoftheaveragepersonintheworldtoday($14,570).Todaytheaveragepersonontheplanetisasrichasthetheaveragepersonintherichestcountryin1950.AndallthosecountriesthathaveanincomehigherthantheglobalaveragetodayaremoreprosperousthantheUSin1950:Iran,Mexico,Bulgaria,…Havealookatthatlist.Thesameistrueforhealthglobally.Theaveragelifeexpectancyintheworldtodayis71years,just1yearlessthanthelifeexpectancyintheverybestoffplacesin1950.Iwroteaboutithere.Whobenefitsfromeconomicgrowth?InthispostIlookedatthepopulation-wideaverageincome.But,thequestionofhowprosperityissharedamongthepopulationisanimportantoneandithasbeencentraltomyresearchoverthelastyears:Ifyouareinterestedinthisquestion,havealookattheshortarticleonVox.comthatIhavewrittenwithmycolleagueStefanThewissenhereorotherrecentresearchofmine.13BrianNolanhaseditedtwobooksonthequestionwhichwerepublishedjustrecently.Whatthisresearchshowsisthatitverymuchdiffersbetweencountriesandovertimewhoisbenefitingfromeconomicgrowth.WhileintheUS,forexample,mostoftheincomegainswenttotherichestmembersofsocietythisisnottrueofothercountrieswhereeconomicgrowthwaswidelysharedamongall.Whyshouldwecare?Thedatavisualizedinthesetwochartsshowsthattheworldisnotthezero-sumeconomythatitwasinourlongpast.Itisnotthecaseanymorethatoneperson’soronecountry’sgainisautomaticallyanotherone’sloss.Economicgrowthtransformedtheworldintoapositivesumeconomywheremorepeoplecanhaveaccesstomoregoodsandservicesatthesametime.Itwouldbewrongtofocusoneconomicgrowthonly.ThatisthereasonwhyOurWorldinDatadoesnotonlylookatthismetric,butathundredsofaspects–includinghealth,education,humanity’simpactontheenvironment,andhumanandpoliticalrights.AndtherearealternativestoGDPpercapitaasakeymetricandwe’vewrittenaboutsomeofthembefore(hereandhere).Economicgrowthhastobeachievedatatimewhenweurgentlyhavetoreduceourimpactontheenvironment.Thismeansthatitisnotonlytherateofgrowththatmatters.AsMariana Mazzucatosays“economicgrowthhasnotonlyaratebutalsoadirection”.Andmanypathsforgrowthpointinadirectionthatdoesnotincreaseourenvironmentaldamageandinsteadcanoftenreducetheimpact(bettercareforthesickandelderly,bettereducationalinstitutions,alternativestomeat, careformentalhealth,improvedsolartechnology;alltheseimprovementswouldmeanmoregrowth).Economicgrowthisnottheonlythingthatmatters,butitdoesmatter.IncontrasttomanyoftheothermetricsonOurWorldinData,economicgrowthdoesnotmatterforitsownsake,butbecauserisingprosperityisameansformanyends.Itisbecauseapersonhasmorechoicesastheirprosperitygrowsthateconomistscaresomuchaboutgrowth.Risingprosperitygivespeopleaccesstoawiderangeofthingstheyvalue:food,healthcare,accesstoeducation,entertainment,holidays,freetime,andmore.TheconcernwithGDPpercapitaisbasedontheideathatrisingprosperitymakesforaricherlife.Ifindthemetricimportantbecauseitisameasureofmeansonlyandtherebyrespectsthefreedomofeveryonetochooseforthemselves.Itisbecauseofthisthatissoimportanttotrackhowincomeshavechangedaroundtheworld.DifferentdatasetsongrowthinthelastdecadesForrecentdecadesseveralinternationaldatasetsonGDPareavailable.Thetwomostwidelyusedarethe‘PennWorldTable’andtheestimatespublishedbytheWorldBank.PennWorldTableThePennWorldTable(PWT)isadatabaseonthelevelofincomes,output,input,andproductivityovertime.Itisnowcoveringmorethan180countriesanddataisavailablefrom1950onwards.Thedatabaseisproducedbyagroupofresearchersfromaroundtheworldandpublishedbythe‘GroningenGrowthandDevelopmentCentre’attheUniversityofGroningen–thewebsiteishere.ThemapbelowshowsthePWTestimatesofinflation-adjustedGDPpercapita.WorldBankdataTheWorldBankalsopublishesestimatesofGDPpercapitaforcountriesaroundtheworld.TheWorldBankdatainconstant2011international-dollarsisavailablefrom1990onwards.ThisscatterplotshowsthecomparisonbetweentheWorldBankandPWTdata. Clicktoopeninteractiveversion GDPpercapitainUS-DollarAdjustingforthedifferentpricelevelsindifferentcountriesisnecessaryifonewantstocomparelivingstandardsofpeople.Theinternational-dollar–usedinthemapsabove–makesthesecomparisonspossible.Butifoneisinterestedincomparingthepercapitaoutputofdifferenteconomiesitisusefultoconsidertheoutputbysimplyusingtheexchangeratesofthedifferentcurrencies.Thisisdoneforthedatathatisshowninthemapbelow,theGDPpercapitaisconvertedtoonesinglecurrency:theUS-$. Clicktoopeninteractiveversion TotaloutputofeconomiesThediscussionabovefocussedmostlyonoutputpercapita,themapbelowshowsthetotaloutputbycountry. Clicktoopeninteractiveversion ProductivityincreasesmakeeconomicgrowthpossibleInthissectionRisingproductivitybysectorUrbanizationandprosperityReligiosityandprosperityTherearetwowaystoincreaseoutputovertime:Increaseinputsortoincreaseproductivity,theratioofoutputtoinput.Howitispossibletoraiseproductivitycanbemostclearlyseenwhenoneconsidersasingleindustryonly.Thinkabouttheproductionofbooks:Beforetheprintingpresswasinventedtheonlywaytocopyabookwasforascribetocopyit.GregoryClark14estimatesthatthescribeswhoweredoingthisworkbackthenwereabletocopy3,000wordsofplaintextperday.ThisimpliesthattheproductionofonecopyoftheBiblemeant136days(4.5months)ofwork.ThischangedfundamentallywhenJohannesGutenbergadoptedthetechnologyofthescrew-typewinepressesoftheRhineValleywherehewasfromtodevelopaprintingpress.Thehoursofworkaprinterhadtoputinwasnowmeasuredhoursratherthanmonths. Estimatesarethataworkerwasabletoproducearound2.5booksinaday.OvertimeprintingpresseswereimprovedandduringtheIndustrialRevolutiontheyweremechanizedandproductivityofworkersincreasedfurther.TheInternetstandsinthislongtraditionandastextscannowbeseenbymillionsinaninstanttheproductivityinthebusinessofmakingtextsavailableisoffthecharts.Productivity–OutputovertimeThisvisualizationshowslaborproductivityperhour.Switchingtothemapviewshowsthelargedifferencesbetweencountries. Clicktoopeninteractiveversion RisingproductivitybysectorRisingoutputbyindustryThevisualizationbelowshowstherisingoutputoftheeconomybyindustry.Eachtime-seriesisindexedtotheyear1700sothatthefocushereisonthechangeovertimeasallchangesarerelativetothatyear.Therisingoutputofkeyindustrialandservicesectorsisshownhere. Clicktoopeninteractiveversion UrbanizationandprosperityUrbanizationandeconomicprosperityarestronglycorrelatedasthefollowingvisualizationshows.InmostcountrieswithaGDPpercapitahigherthan30,000international-$morethan60%ofthepopulationlivesincities. Clicktoopeninteractiveversion ReligiosityandprosperityAsurveyaskedthequestion“Howimportantisreligioninyourlife?”andthepossibleanswerswere“veryimportant”,“somewhatimportant”,“nottooimportant”and“notatallimportant”.Thefollowingchartplotsthesharethatanswered“veryimportant”againsttheaverageprosperityofthepopulationforeachcountryinthesurvey.Thereisaclearcorrelationbetweenpovertyandreligiosity.Inpoorcountriesthehugemajoritysaythatreligionisveryimportantintheirlife:incountrieslikeUganda,Pakistan,andIndonesiaitistheanswerofmorethan90%.InEthiopiaitistheanswerof98%ofthepopulation.Inrichercountriestheshareofthepopulationforwhomreligionisveryimportantismuchlower.IntheUK,SouthKorea,Germany,orJapanitislessthan1in5forwhomreligionisveryimportant.ThebigoutlierinthiscorrelationistheUSA,averyrichcountryinwhichmorethan50%answerthatreligionisveryimportantintheirlife.RetirementbecamepossiblewhenpeoplegotricherThevisualizationshowstheverysubstantialdeclineinthelaborforceparticipationofmenof65yearsandolderintheUSAsincetheendofthe19thcentury. Clicktoopeninteractiveversion AccesstofinancialservicesToallowsavingandfacilitatetransactionsaccesstofinancialservicesisimportant.Weknowthatinpoorercountriesthisaccessisoftenverylimited.Wedon’tknowmuchabouthowthisaccesshaschangedovertimeandtounderstandthischangebetterwehaveattemptedtocombinedifferentsources–theresultofwhichisshownintheworldmapbelow.TheWorldBankGlobalFinancialInclusiondataisavailablefor2011and2014.Thedataisveryscarceonthispre-2011,butWorldBankestimatesprovideanadditionalsinglepointforcountries.Thishasbeenrepresentedasapointfor2005,howeverit’simportanttonotethatthisvariesbetween2000-2005acrosscountries.Thechallengeisthatitisnotexactlythesamemeasureasthe2011and2014data,butinsteadacompositemeasureofaccesstoabankaccountandfinancialservices.TheWorldBankdefinethiscompositeindicatorasmeasuring“thepercentageoftheadultpopulationwithaccesstoanaccountwithafinancialintermediary.Theindicatorisconstructedasfollows:foranycountrywithdataonaccessfromahouseholdsurvey,thesurveyedpercentageisgiven.Forothercountries,thepercentageisconstructedasafunctionoftheestimatednumberandaveragesizeofbankaccountsasdiscussedinHonohan(2007).Thesenumbersaresubjecttoestimationerror.”Theuseofcompositemeasuresis,ofcourse,notideal.However,wethinkitshouldstillgiveafairlyreasonablebasisoftheearly2000stouseasanearlierestimateandthedirectionofprogresstrends.Wehavecombinedthiscompositemeasurewith2011and2014datainthefollowingchart. Clicktoopeninteractiveversion DataQuality&DefinitionsInthissectionGDPpercapitaandaverageincomesHowareincomesadjustedforinflation?InternationalGDPComparisons:marketvs.PPPexchangeratesPPP-adjustedGDP:spatialcomparisonsHistoricalreconstructionsofnationalaccounts–thecaseoftheUKThemostcommonmeasureforeconomicprosperityistheGrossDomesticProductorGDPforshort.Itmeasuresthemonetaryvalue–theprice–ofallgoodsandservicesproducedinacountry.Toallowforcomparisonsbetweencountriesandovertime,thetotaleconomicoutputofacountryisputinrelationtothenumberofcitizensinthatcountry.ThisisGDPpercapita.Thechangefromoneyeartothenextisreferredtoaseconomicgrowth.Aseveryonewhohadabeerorahaircuttenyearsagowillrememberthepriceofgoodsandservicesusuallyincreasesovertime,thisiscalledinflationandismostcommonlymeasuredwiththeconsumerpriceindex(CPI).Comparisonsofprosperityovertimearethereforeonlymeaningfulwhenthesepricechangesaretakenintoaccountsothatthegrowthratedoesnotcapturemerechangesinprices.Measuresofincomesareonlymeaningfulwhentheyareputinrelationtomeasuresofpricesthattheseincomereceiversface.Whenincomesareadjustedforpriceseconomistsspeakoftherealvalueofagoodorservice.(Butsincecomparisonsonlymakesensewhenoneadjustsforpricechanges,itisusuallythecasethatadjustmentsforinflationhavebeenmadeevenwhenitisnotexplicitlysaid.)TheproductionboundaryGDPispublishedinacountry’sNationalAccounts.Thesestatisticscomplytoprotocolslaiddowninthe1993versionoftheSystemsofNationalAccounts,SNA93.TheSNA93establishedthe‘productionboundary’,acrucialdefinitionofwhatisandisnotincludedinGDPfigures:includedareallgoodsandservicesthatareexchanged,includedareallgoodsthatarenotexchanged(forexamplefoodproducedforhomeconsumption)butexcludedareservicesthatarenotexchanged.Amongtheseexcludedservicesarefoodpreparation,educationofchildrenathome,andminorhomerepairs. Animportantexceptiontotheservicesthatareincludedisthehousingservicesconsumedbyowner-occupiers.Thisserviceisimputed(imputedrent)andincludedintheGDPfigure. GDPpercapitaandaverageincomesInprincipletherearethreeequivalentwaystocalculateGDP:Production:thevalueoffinaloutputsproducedinaneconomylessthevalueofalltheinputsusedintheirproductionExpenditure:theexpenditureonfinalgoodsandservicesproducedinaneconomybyhouseholds,corporationsandgovernmentsIncome:theincomeearnedbyindividualsandbusinessesfromtheproductionofgoodsandservicesintheeconomyIttheorythesethreemeasuresshouldbeequal;theyconstituteanaccountingidentity.Acompany’srevenueistheincomeitgeneratesfromsellingthegoodsandservicesitproducestoconsumers;yetthatsamerevenueisalsotheexpenditureofconsumersonthosegoodsandservices.PaulKrugmanmakesthepointthat“ourincomemostlycomesfromsellingthingstoeachother.Yourspendingismyincome,andmyspendingisyourincome.”15ThissymmetryallowsustouseGDPtoapproximateaverageincomesbydividingGDPbythetotalpopulation.Inreality,averageincomesandGDPpercapitawillnotbeequal.16DifferentmeasuresofaverageincomeintheUS1914-2004–VisualizingEconomics17Howareincomesadjustedforinflation?Tomakemeaningfulcomparisonsofprosperityovertimeitisnecessarytoadjustforinflation.Buthowisthisactuallydone?Asimplepossibilitywouldbetotakeaproductthatiscommonlypurchased–let’ssayaloafofbread–andtonotedownthepricechangesofthisproductovertime.Ifyouthenfindthatthepriceofbreaddoubledoveraperiod,butyouremployerstillpaysyouthesameincome,thenyoucanonlybuyhalfasmanybreadsfromyourincomeandyourincomeintermsofbreadhashalved.Ahalvingofyourincomeintermsofbreadisyourincomeadjustedfortheinflationofbreadprices.Buttomeasurepricesbyrelyingononeproductonlyhastheobviousproblemthatyoucouldenduppickingaproductthatwasnotrepresentativeofthepricechangesofalltheotherproductsandservicesthatconsumerswanttobuy.Whilethepriceofbreadmayhaveincreased,thepricesofthemajorityofothergoodscouldhavedecreased.Theideaforinflationadjustmentforincomesisthereforetoinsteadrelyonacommoditybundleofgoodsandservicesthatarerepresentativeoftheconsumptionoftheaveragehousehold.Byrelyingonarepresentativecommoditybundleinsteadofbreadaloneallowsyoutoadjustincomesnotonlyforbread,butforthecost-of-livingmorebroadly.Thebasketusedischosentoreflecttheexpenditureofthetypicalhousehold,sothatchangesofthisbundlemeasurethechangestopricesthetypicalconsumerfaces.WhatweareinterestedinisthepricechangeofthisbundleofgoodovertimeandthehistoryofpricesisthenexpressedinanindexcalledtheConsumerPriceIndex,whichisindexedto100forachosenbaseyear.Asconsumptiondiffersindifferentcountries,thesehouseholdconsumptionbasketsvaryfromcountry-to-countryandovertimeasnewtechnologiesemergewhichmakenewgoodsandservicesavailableandbecauseconsumptionpreferenceschange.18IncomesrelatedtoapriceindexgiveusrealincomesOnlyincomesinrelationtopricesgivesusanideaabouthowtheprosperityofapopulationchanges.Incomesontheirownorpricesontheirowncannotgiveusanideaaboutchangingprosperity.Thepricesthatweseeonthepricetagsintheshoparethenominalpricesandsincewealmostalwayshavesomeinflationthesepricestendtogoup.Nominalpricesandincomesareexpressedintermsofmoney,inthiscaseBritishPound,andinincomestatisticsnominalincomesarereportedasincomesin‘currentprices’.Theinflationadjustmentofincomeisdonebyexpressingincomerelativetothepriceofacommoditybundlesuchastheonedescribedbefore.ThenominalincomerelativetothenominalpricelevelasmeasuredbytheCPIistherealincome.Realincomesmeasurethechangetoincome,adjustedforthefactthatnominalpriceshaveincreased(ordecreased)andinincomestatisticsrealincomesarethereforereportedasincomesin‘constantprices’.Anexample:AdjustingwagesforinflationintheUKThechartbelowshowstwoseriesincurrentprices,nominalweeklywagesandnominalpricesintheUK,andathirdseries,whichisconstructedfromtheinformationinthenominalseries:‘inflationadjustedwages’or‘realwages’.Let’slookatnominalpricesfirst.TheindexthatmeasuresthetypicalconsumptionbundleofgoodsandservicesintheUKhasavalueof100in2015andavalueof0.66in1750.Theratiobetweenthesetwoobservationsis100/0.66=152,whichtellsusthatthenominalpricesthatconsumersfacehaveincrease152-foldinthese265years.ButoverthesameperiodthenominalweeklywagespaidintheUKeconomyincreasedaswell.From£0.29to£492perweek.Thisisa1695-foldincrease.Tocalculatetherealincreaseinwagesweneedtolookatthenominalwageincreaseinrelationtothenominalpriceincrease.191695dividedby152is11.2.Thistellsusthataveragerealwagesare11.2-timeshighertodaythanfortheaveragewageearnerbackin1750.Iftheirgreat-great-grandfathersin1750hadtoworkforayeartobuyarepresentativeconsumptionbundle,Britstodayhavetoworkforonlyabitmorethanamonthtobuythecomparablebundleofgoodsandservices. Clicktoopeninteractiveversion NewproductsandservicesbecomingavailableItisnotjustpricesthatchangeovertime.Theveryproductsandservicesthatweproduceandbuychange.Technologicalprogresshasmeantthatthegoodsandservicesavailabletodayareinvariablysuperiortothoseavailableseveralhundredyearsago,withalmostnoexampletothecontrary.Theintroductionofnewgoodsandservicescreatesseriousproblemsforintertemporalcomparisonsofwealththataremostrelevanttoday;itislessofaproblemforthelongpre-modernworldwhenalmostalleconomicproductionconsistedof food,shelterandclothing.Toemphasizethispointconsiderthefollowingexample:In1836therichestmanintheworldwasprobablyNathanRothschild.Rothschildcouldaffordwhateverhewanted–everygoodandserviceavailableintheworldof1836.Yetinthatsameyear,the56yearolddiedofaninfectionthatiscurabletodaybyantibiotics,whichareavailablearoundtheworldforeventhepoorestpeopletoday.20QualitychangesTheStiglitz,Sen,Fitoussireportexplains“Qualitychangecanbeveryrapidinareaslikeinformationandcommunicationtechnologies.Therearealsoproductswhosequalityiscomplex,multi-dimensionalandhardtomeasure,suchasmedicalservices,educationalservices,researchactivitiesandfinancialservices.Difficultiesalsoarisein1.EvidenceandreferencesinsupportoftheclaimspresentedinthisSummaryarepresentedinacompaniontechnicalreport.22SHORTNARRATIVEONTHECONTENTOFTHEREPORTcollectingdatainanerawhenanincreasingfractionofsalestakeplaceovertheinternetandatsalesaswellasdiscountstores.Asaconsequence,capturingqualitychangecorrectlyisatremendouschallengeforstatisticians,yetthisisvitaltomeasuringrealincomeandrealconsumption,someofthekeydeterminantsofpeople’swell-being.Under-estimatingqualityimprovementsisequivalenttoover-estimatingtherateofinflation,andthereforetounderestimatingrealincome.Forinstance,inthemid-1990s,areportreviewingthemeasurementofinflationintheUnitedStates(BoskinCommissionReport)estimatedthatinsufficientaccountingforqualityimprovementsingoodsandserviceshadledtoanannualoverestimationofinflationby0.6%.ThisledtoaseriesofchangestotheUSconsumerpriceindex.”21Priceindices:TheConsumerPriceIndex(CPI)andtheGDPdeflatorAdjustingforincomesistoisolatevolumechangesfrompriceeffectsInsteadoflookingatpriceadjustmentforincomes,wecanalsolookatpriceadjustmentsforoutput.NominalGDPisameasureofthevalueofoutputproducedinacountryorregionoveraspecifiedperiod(usuallyoneyear).Thevalueofoutputiscomposedoftwofactors:thevolumeproducedandtheprice,,whererepresentsthepriceofoutputandrepresentsthevolumeofoutput.ThereforeincreasestoGDPareeithertheresultofmoreoutput,higherpricesoracombinationofthetwo.Whenlookingattheperformanceofasingleeconomyovertimeitisimportantthatwecontrolforpriceeffectssincetheycanmaskchangestothevalueofoutput.InthesecasesweadjustforthepricechangesandlookatrealGDP.TherealGDPisconstructedby‘deflating’thenominalGDPbyapriceindexthattrackschangestopricesintheeconomyrelativetoachosenbaseyear.Thistransformationattemptstoisolatevolumechangesbyeliminatingpriceeffects.ThetwomostcommonpriceindicesusedtodeflateincomesandnominalGDPare:ConsumerPriceIndex(CPI):Thebasketisusedmeasurethepricechangesofthegoodsandservicesconsumedbythetypicalhousehold.Thetypicalhouseholdconsumptionbasketwillvaryfromcountry-to-countryduetodifferentpreferencesandincomelevelsaswellasovertimeasnewtechnologiesemergeandpreferenceschange.CPIindicesareproducedbynationalstatisticalagencies,withonemajorexceptionbeingtheEuropeanUnion’sHarmonisedCPI(HCPI).TheHCPIusesconsistentdefinitionsforallEurozonememberstoaidcomparability.TheCPIusedtoadjusthouseholdincomestrackschangestothepriceofabasketofgoodsthatthetypicalhouseholdpurchases.Itismeasuringthepricechangesofconsumption.GDPdeflator:TheGDPdeflatoralsocalledNationalAccountsdeflatormeasuresthepurchasingpowerrelativetothepricesofalldomesticallyproducedfinalgoodsandservicesinaneconomy.Itismeasuringthepricechangesofdomesticproduction.DifferencesbetweenCPIandGDPdeflatorTheCPIindexmeasurespricechangesofconsumptionwhereastheGDPdeflatormeasurespricechangesofdomesticproduction.Consequentlythereareseveralimportantdifferencesofthetwopriceindices:22UnliketheCPI,theGDPdeflatordoesnotadjustforchangesofgoodsthatareimportedfromothercountriesandthepricechangesofimportpricesarenotdirectlytakenintoaccount.Secondly,theGDPdeflatorcoverscapitalgoods,goodsthatarenotboughtbyconsumers.TheCPIisalsoavailablemonthlyformostcountries,whiletheGDPdeflatorismostlyavailableonlyquarterly.InternationalGDPComparisons:marketvs.PPPexchangeratesInordertohighlightthedifficultlyofmakinginternationalcomparisonsbetweencountries,considertheaverageincomeofsomebodylivinginIndiacomparedwiththeUS.In2011theaverageIndianearned72,000rupees,whiletheaverageAmericanearned$50,000.Sincetheaverageincomesarestatedindifferentcurrencies,comparingthenumericalvalueismeaninglessanddoesnothelpusdeterminewhoisricherandbyhowmuch.ConvertingtherupeeamountintoUSdollarsusingmarketexchangeratesgivesusanaverageincomeof$1,500inIndia.Thisnumberisover33timessmallerthantheincomeoftheaverageAmerican!However,itisobviousthatthecostoflivingintheUSismuchhigherthaninIndia,whichimpliesthatthecomparisonofincomesmadeatmarketexchangeratesisalsonotafaircomparisonofhowrichorpoorthepeoplereallyareincomparison.23AsolutionistoconverttheamountsusingthePurchasingPowerParity(PPP)exchangerate.Thisconversiontakesintoaccountdifferencesinthepricelevelsofbothcountries.MakingthePPP-adjustmentrevealsthattheaverageincomeofsomeonelivinginIndiais$4,800(internationaldollars),only10timessmallerthanintheUS.TheprecisenatureofPPPadjustmentsisexplainedinthesectionbelow.ThisdoesnotmeanthatcomparisonsofGDPevaluatedatmarketexchangeratesareuninformative.Infact,whencomparingfinancialflows,PPP-adjustmentsaremeaninglessandGDPevaluatedatthemarketexchangerateisthemostappropriatemeasure.Whencomparingdevelopmentandlivingstandards,theconverseistruesinceweneedtoeliminatepriceeffects.PPP-adjustedGDP:spatialcomparisonsGDPcomparisonsmadeusingmarketexchangeratesfailtoreflectdifferencesinthepurchasingpowerofdifferentcurrencies.Ingeneral,pricesarehigherindevelopedeconomies,24andsoexchangerateadjustedGDPmeasureswillunderestimatethesizeoflowincomeeconomies.Inasimplifiedworld,thepricesoftradedgoodsaredeterminedbyglobaldemandandsupplyforces,whilethepricesofnon-tradedgoodsaredeterminedbylocaldemandandsupplyforces.Sincewagesandsalariesarelowerindevelopingcountries,thepricesofnon-tradedgoodsalsotendtobelower.ThisfeatureismissedbyexchangerateadjustedGDPcalculationsasnodistinctionismadebetweentradedandnon-tradedgoods.Inaworldwhereallgoodsaretraded,exchangerateadjustedGDPwouldbeamoreinformativewaytomakeinternationalcomparisons.Nevertheless,anothercomplicatingfactoristhatexchangeratesarehighlyvolatileanddeterminedbycurrencyspeculation,interestratesandinternationalcapitalflows.PurchasingPowerParity(PPP)adjustmentstoGDPareanattempttoisolatedifferencesinthevolumeofoutputoftwoeconomies.Thatis,theyeliminatedisparitiesinthepricelevelsofdifferenteconomies.APPPexchangeratecanbethoughtofasthecostratioofacomparable(butnotidentical)basketofgoodsintwocountries.Hence,themethodologyisanalogoustothatusedincomputingCPIinflationtocalculaterealGDP,exceptthatherethecomparisonsaremadebetweencountriesratherthanovertime.Moreprecisely,PPP-adjustedGDPisaspatialmeasureandnotatemporalonelikerealGDP:abasecountryisusedasopposedtoabaseyear.TheUSdollaristhemostcommonunitofcurrencyusedtomakeinternationalcomparisonsand,forclarity,PPP-adjustedquantitiesarequotedininternationalorGeary-Khamisdollars.25WecandecomposetheGDPratiooftwoeconomiesintowhereisthepricelevelincountry,isthecurrencylevelincountry,andisthevolumeofoutput(realoutput)incountry.Whenmakinginternationalcomparisonsweareinterestedintheratioofoutputvolume.ItispossibletoremovethecurrencydifferencesbyusingtheexchangeratetoconverttheGDPintoacommoncurrency,butthiswouldleavepriceleveldifferences.ThePPPexchangerateadjustsforboththecurrencyandthepricelevelratio.ThecreationofPPPsbytheInternationalComparisonsProgramme(ICP)ThePPPsthatareusedininternationalcomparisonstodayarecreatedbytheInternationalComparisonsProgramme(ICP)conductedbytheWorldBank.ThelatestroundoftheICPwascompletedin2014andhasestimatedPPPsfor2011.Thestudycovers199countriesandisthemostextensivestudyofPPPseverconducted.Sincetheresearchishighlyintensiveandrequirescooperationwithmanydifferentcountriesandstatisticalagencies,theexistingPPPdataissparse(8benchmarkyearssincethefirstICPstudyin1970).Whatismore,themethodologyusedandthegroupofparticipatingcountrieshasdifferedbetweeneachroundoftheICP.TherearesixyearsseparatingthemostrecentICPestimates(2011and2005).Herisadetailedexplanationofthemethodologyandfindingsofthe2011ICP.ComputingPPP-adjustedGDPforyearsforwhichnoICPdataisavailableComputingPPP-adjustedGDPforyearswhereICPdataisavailableisstraightforward,however,inyearswherethereisnodata,thereisnoconsensusregardingthebestwaytoproduceestimates.SincePPPisaspatialmeasure,eachICPestimateisindexedtoabenchmarkyear.ProducingPPP-adjustedGDPestimatesfornon-benchmarkyearsrequireseitherextrapolationofPPP-estimatesfromasingleroundofICPdata,interpolationbetweendifferentroundsoracombinationofthetwo.ExtrapolationtakesthePPP-adjustedGDPinasingleyearandassumesthatitevolvesaccordingtorealGDPgrowthratesortheinflationratioofthecountryofinterestwiththeUS.Interpolationattemptsto“fill-inthegaps”usingobservedICProundsinconjunctionwithinflationorgrowthdataaccordingtosomestatisticalmodel.Thetwofigurespresentedbelowaredesignedtoaidunderstandingofthesetwomethods.Understandingthemethodologyandpurposeofeachdatasetisimportantwhenconductingdataanalysis.ThefourmostimportantdatasetsforPPP-adjustedGDPdatahavethefollowingmethodology:WorldBankdata:ExtrapolatesfromthemostrecentICProundusingnationalaccountsdeflators.Thedataandabriefdescriptionofthemethodologycanbefoundathttp://data.worldbank.org/indicator/PA.NUS.PRVT.PP. Sincethedataisextrapolatedfromthe2011ICP,dataisonlypresentedfortheyears1990-2014asanyfurtherextrapolationislikelytogiveunreliableestimates.Thisdatasetismostusefulwhenconsideringtheveryrecentpastasthe2011ICProundusesthemostsophisticatedmethodologytodate.Italsohastheadvantageofbeingextrapolatedforwardintimeto2014. Notesfromthedocumentation:“FormosteconomiesPPPfiguresareextrapolatedfromthe2011InternationalComparisonProgram(ICP)benchmarkestimatesorimputedusingastatisticalmodelbasedonthe2011ICP.” PennWorldTables:Interpolatesandextrapolatesusingnationalaccountsdeflators.Themethodologyusedcanbefoundinthedocumentationavailableathttp://www.rug.nl/research/ggdc/data/pwt/pwt-8.1. Thisdatasetismostcommonlyusedforstatisticalanalysisbyeconomistsandcoverstheyears1950-2011.Itisarguablythemostreliable,long-rundataavailableonPPP-adjustedGDP. Gapminder:Interpolatesandextrapolatesusingrealgrowthrates.Themethodologyusedcanbefoundinthedocumentation. Gapminderaimstogivethewidestcoveragepossibleattheexpenseofrobustestimates.MuchofthehistoricdatahasbeenestimatedfromtrendsandotherForthisreason,wepresentthedatahereforthepurposesofgraphicalpresentationandnotasfact. Notesfromthedocumentation:“Themainpurposeofthedataistoproducegraphicalpresentationsthatdisplaythemagnitudeofincomedisparitiesintheworldovertime…Hencewediscouragetheuseofthisdatasetforstatisticalanalysis…Theobservationsfortheperiodbefore1950are,inthemajorityofcases,basedonroughestimateswithinarangeoflikelyvalues.Inmanycaseswehavenoinformation,what-so-ever,ontherelativerankingofcountries.” MaddisonProject:Dataisdrawnfromvarioussources.Acompletelistofsourcescanbefoundinthedocumentationavailableathttp://www.ggdc.net/maddison/maddison-project/home.htm. Thisdatasethasmanydifferentsourcesandiscuratedbytheprojectteam.Althoughthelong-rundataismuchlessreliablethanmorerecentestimates,themorerigorousnatureoftheestimatesallowsforbetterspatialcomparisons. Notesfromthedocumentation:“TheMaddisonProjectdatabaseplacestheemphasismoreontheinternationalcomparabilityoftheestimatesthanontheirconsistencyovertime.” DiscrepancybetweenincomesreportedinhouseholdsurveysandGDPper capitaThisdiscrepancyreflectsseveralfactors:–GDPincludesitemssuchasdepreciation,retainedearningsofcorporations,andgovernmentrevenuesthatarenotdistributedbackbythegovernmentorcorporationstohouseholdsascashtransfers.–ParticularlyincomesattheverytopoftheincomedistributionarenotfullyaccountedandthiscontributestothegapbetweentheNationalAccountfigure (GDP)andhouseholdsurveyfigures.–Particularlyindevelopingcountries–wherethisdiscrepancytendstobelarger– alargeshareofgovernmentrevenuemayfirstlyendupinsovereignwealthfundsor secondly representsprofitsofforeignmultinationalsthatarerepatriatedbuttakenintoaccountintheGDPcalculation.26Historicalreconstructionsofnationalaccounts–thecaseoftheUKHowdoeconomichistoriansgoaboutestimatingincomesinthedistantpast?Inbroadterms,thestrategyistoextendbacktoearlierperiodsthesystemofnationalincomeaccountingthatcountriesusetodaytoestimatethetotaloutputoftheeconomy.Themainobjectiveistoapplyamethodologythatreconstructsthismetricconsistentlyovertimeandacrosscountries.Intheabsenceofdatacollectedatthetime,researchershavetobringtogetherwhatevidencetheycanfromhistoricalsources,butthebasicprinciplesarethesame.BecauseGreatBritain’seconomywasthefirsttoachievepersistenteconomicgrowthitistheeconomythathistorianshavestudiedinmostdepth.ThereconstructionofBritain’seconomichistoryThevisualizationshowstheoutputoftheEnglisheconomyperpersonsincetheMiddleAges.Asexplainedbelow,itisnotonlycapturingtheproductionofworkerspaidinthelabormarket,butalsotheproductionofsubsistencefarmersandotherproducerswhichwerenotpaidamonetarysalary.AssuchitgivesusaperspectiveonthehistoryofmateriallivingconditionsoftheEnglishpopulationoverthelast746years.ThedatainthechartistakenfromtheseminalbookonthehistoryofmateriallivingconditionsinBritain–BritishEconomicGrowth1270-1870byBroadberry,Campbell,Klein,Overton,andvanLeeuwen.Itpresentsafantasticoverviewofthisworkandisverymuchrecommendedforanyonewhowantstostudytheoriginsofeconomicgrowthindetail.27Wheninterpretingthesereconstructionsitisimportanttobearinmindthefundamentalidentityinthishistoricalaccounting:“Withinthemethodologicalframeworkprovidedbynationalincomeaccounting,theestimationofGDPcanbeapproachedinthreedifferentways,viaincome,expenditureandoutput,allofwhichoughttoyieldbroadlysimilarresults.”28Fortheimportantcaseofthesubsistencefarmerforexample,thevalueofthefoodtheyproducerepresentsboththeeconomicoutputoftheactivityandtheincomereceivedbythefarmer.Consumptionofthatproducethenrepresentsaformofexpenditure,asitisusinguppartofthefarmer’sincome.BecauseofthisidentitythemeasurementofGDPcanbeapproachedfromanyoneofthesethreeangles:output,income,orexpenditure.Forhistoricalestimates,theoutputapproachisoftenconsideredthemorereliableinpracticegiventheavailableevidence,thoughinformationonincomesandexpenditurestillprovidebenchmarkstocross-checktheplausibilityofestimates. Clicktoopeninteractiveversion Itwouldbewrongtobelievethathistoriansdonottakenon-monetaryincomesintoaccountFirstly,it’simportanttogetclearfromtheoutsetthathistoricalreconstructionsofpovertyandprosperitydonotjustconcerntheamountofmoneypeoplehadinthepast.Thisisacommonmisunderstandingthatisoftenattheheartofmisinformedcritiquesofhistoricalresearch.Forinstance,inadiscussionofourglobalextremepovertychartonreddit,oneusersuggestedthatitwas“indicativeofthefactthatquitealotoftheworld[…]didnotusefiatcurrency.”Thisinterpretationisincorrect.Yes,overthelasttwohundredyears,therehasbeenamajorshiftfrompeoplefarmingfortheirownconsumptiontowardspeopleworkingforawageandpurchasinggoodsinthemarket.Buthistoriansknowabouthistoryandwherenon-marketsourcesofincomemakeupasubstantialpartoftotalincome,itisveryobviousthatmoneywouldrepresentarathersillyindicatorofwelfare.Justasweneedtoadjustforpriceinflation,accountingfornon-marketsourcesofincomeisanessentialpartofmakingmeaningfulwelfarecomparisonsovertime.Estimatesofpovertyandprosperityaccountforbothmarketandnon-marketsourcesofincome,includingthevalueoffoodgrownforownconsumptionorothergoodsandservicesthatenrichedthelivesofhouseholdswithoutbeingsoldinamarket.Thisissueisnotjustofimportanceforhistoricalestimates,butitisalsoofcentralrelevanceforpovertymeasurementtoday,giventheimportancethatfoodproducedathome,orotherwisereceivedin-kind,continuestoplayintheincomesoftheruralpoor,especiallyinlow-incomecountries.Accordingly,theseflowsareaccountedforinhouseholdsurveysofbothconsumptionandincome,andinthehistoricalestimates.AdjustingforpricechangesandnewproductsItisstraightforwardtocomparematerialprosperityovertimerelativetoallthosegoodswhichremainedrelativelyunchangedoverthecourseofhistory–economichistorianscantracktheaffordabilityofproductslikebread,shirt,beer,nails,meat,booksorcandlesovertime.Thishoweverisnoteasilypossiblewhenentirelynewproductswereintroducedorwhenthequalityofproductsandserviceschangedverymuch.Thefactthatsomeofthemostimportantgoodsandservicesverymuchchangedinqualityordidnotexistatallinthepastrepresentsthebiggestproblemofanylong-termcomparisonsofpovertyandprosperitybecauseitmakespriceadjustmentsdifficult.Manyofthemostvaluablegoodstodaywerenotavailableatall:nokingorqueenhadaccesstoantibiotics,theyhadnovaccines,nocomfortabletransportintrainsorplanes,andnoelectronicdevices–nocomputerandnolightatnight.Historiansofcourseattempttotakethisintoaccountasmuchaspossible29butthiscaveatshouldbekeptinmind:nomatterhowhighsomeone’sincomemighthavebeen,someofthegoodsyoumightvaluethemost–orwouldvalue,whenyougetsick–werenotavailableatall.30ThedetailofthehistoricalreconstructionsThesimplestructureofhowthiseconomichistoryispresented–asasinglelinethatisflatformostofthetimeandverysteepfortherecentperiod–shouldnotfoolusintobelievingthatitisonlyalooseandapproximatehistoricalanalysis.Theamountofworkthatwentintothereconstructionofthishistoryisextraordinary:itistheculminationofdecadesoftediousandextremelycarefulacademicworkbylargeteamsofdedicatedresearchers.Thelevelofdetailthatgoesintosuchestimatesalthoughthisishardtodoinashortoverviewlikethisone.Whilst,thereareindeedmanysourcesofuncertaintyinsuchaprocess,itwouldbeverywrongtothinkthathistoricalGDPestimatesarebasedonflimsyevidence.Itisrecommendedtoreadthisworkinfulllength,butapassageonagriculturaloutputgivessomeinsightalready:“[Theoutputmethod]hasentailed,first,estimatingtheamountsoflandunderdifferentagriculturallanduses…and,then,derivingvalidnationaltrendsfromspatiallyweightedfarm-specificoutputinformationoncroppedareasandcropyieldsandlivestocknumbersandlivestockyields…Thelattertaskisfurthercomplicatedbytheneedtocorrectfordatabiasestowardsparticularregions,periodsandclassesofproducer.”Andbelowisoneofthemanytablesfromthisbook,showingtheauthors’estimatesofoutputofjustonepartoftheagriculturalsectorofEngland.Thisisoneofhundredsofdatasetsthatarerequiredtoconstructthetimeseriesinthechartabove.Andthistable–andallothers–inturnbuilduponasubstantialbodyofhistoricalresearch,asissuggestedbythelistofsourcesitcites.Therearetwokeytakeaways:First,thathistoricalreconstructionsofGDParetheoutcomeofveryseriousacademicwork.Andsecond,theserepresentestimatesoftotaloutput,notjustthatpartofproductionsoldonmarkets.ItwouldbewrongtobelievethattheseGDPseriesdonotaccountforthevalueofnon-marketproduction,includingdomesticproductionforhouseholds’ownuse.AtablefromBroadberryetal.(2015)showingestimatesofhistoricalarableoutputinEngland DataSourcesInthissectionLong-rundatasetsPost-1800Post-1950OtherGDP-relateddataLong-rundatasetsBigdatasetsAngusMaddisonHistoricalStatisticsData:GDPpercapitaandtotalGDP(+populationdata)Geographicalcoverage: Global–bycountries&worldregionsTimespan: 1-2008CE(5observationsbefore1820andthenannualifpossible)Availableat:OnlineattheGroningenGrowth&DevelopmentCentrehere FoundedbytheeconomichistorianAngusMaddisonandlaterupdated.Anupdatedversion–availablethroughtheCLIOInfraproject–isdescribedbelow.AnothercorrectionoftheMaddisondataisthe‘Barro-UrsuaMacroeconomicData’,whichisavailableatRobertBarro’swebsitehere.MaddisonProject–availablethroughtheCLIOInfraProjectData:GDPpercapitaGeographicalcoverage:Globalcoverage–bycountry Timespan:Between1500and2010Availableat:Onlinehere ThisdatabasebuildsontheworkofAngusMaddison.Thedatabase andtheaccompanyingpaper31areaproductofTheMaddisonProjectandarepartlyarevisionoftheearlierwork.TheauthorsareJuttaBoltandJanLuitenvanZanden.ThemanysourcesareverywelldocumentedatthewebsiteoftheCLIOproject. TheGlobalPriceandIncomeHistoryGroupData: NominalGDPGeographicalcoverage:Manyworldregions(includingAfrica,AsiaandOceania)Timespan:Variesbyworldregion–longforsome(e.g.Italysince1310)butmostlysincethe19thcenturyAvailableat:OnlinehereWelldocumented,readilyavailableinExcelspreadsheets.IncludesmanyotherdatasetsSmallDatasetsWikipedia’s‘listofregionsbypastGDP(PPP)percapita’includesMaddison’sestimatesforcountriesandworldregionsbetween1CEand2003CE.ItalsoincludesBairoch’sestimatesforEuropebetween1830–1938.AndsomeestimatesfortheprosperityoftheRomanandByzantineempires.StephenBroadberry’sDataData:GDPpercapitaGeographicalcoverage:Britain,Italy,SpainandHolland1270-1870andEuropefor1870-2000Timespan:1270-1870and1870-2000Availableat:TheearlydataispublishedinresearchpapersbyBroadberryandothers.AnoverviewisgiveninBroadberry(2013)32ThedataforseveralEuropeancountriesforthetimeperiod1870–2000isavailableonlineatBroadberry’swebsitehere. Thequalityofthisdataishigh.Thedatawasrecentlypublishedindetailedresearchpapersbyvariousauthors.Post-1800NationalAccountsdatainthe‘InternationalHistoricalStatistics’Data:GDPpercapitaandsomeotherNationalAccountsdataGeographicalcoverage: Global–bycountryTimespan:Variesbycountry.Oftensincetheearly19thcentury(USsince1789)Availableat:Thestatisticsarepublishedinthreevolumescoveringmorethan5000pages.33AtsomeuniversitiesyoucanaccesstheonlineversionofthebookswheredatatablescanbedownloadedasePDFsandExcelfiles.Theonlineaccessishere.Thesestatistics–originallypublishedundertheeditorialleadershipofBrianMitchell(since1983)–areacollectionofdatasetstakenfrommanyprimarysources,includingbothofficialnationalandinternationalabstractsdatingbackto1750.JerryDwyer’sdatasetData: Nationalaccountdata&otherrelevantdata(e.g.averageageandexperienceoftheworkforce)Geographicalcoverage: GlobalbycountryTimespan: Mostlysince1880–observationsateveryevendecadeAvailableat: Onlinehere.Downloadableasanexcel-fileTheHistoricalNationalAccountsdatabaseoftheGroningenGrowthandDevelopmentCentreData:NationalAccountdatawithmanyinterestingmeasuresGeographicalcoverage: (Mostlybutnotonly)EarlyindustrializedcountriesTimespan: Sincetheearly19thcenturyAvailableat: Itisonlinehere.TheHistoryDatabaseoftheGlobalEnvironment(HYDE)Data:GDP(percapita)andprivateconsumptionasshareofGDPGeographicalcoverage:Timespan:Since1890Availableat:GDPdataisonlinehere.Privateconsumptionishere.EarlyestimatesarebasedonMaddison’sdata.Post-1950PennWorldTableData:GDPandmanymeasuresrelatedtorelativelevelsofincome,output,inputsandproductivity.Geographicalcoverage:Global–bycountryTimespan:Since1950Availableat: Itisonlinehere.Thedocumentationcanbefoundin Feenstra,InklaarandTimmer(2015),34inwhichthevalidityofthesemeasuresisalsodiscussed.35WorldDevelopmentIndicators(WDI)publishedbytheWorldBankData:GDPandGNI(grossnationalincome)constructedandcorrectedforpricedifferences(acrosstimeandcountries)indifferentwaysGeographicalcoverage:Global–bycountryandworldregionTimespan:Theearliestdataisavailablefor1960.Availableat:ThedifferentmeasuresarepublishedaspartoftheWorldDevelopmentIndicators.ThedataoncurrentGDPinUS-$ishere.TotalEconomyDatabase(TED)Data: GDP,population,employment,hours,laborquality,capitalservices,laborproductivity,andtotalfactorproductivityGeographicalcoverage:Global–bycountryTimespan:Since1950Availableat: Itisonlinehere.OtherGDP-relateddataSubnationaldataNicolaGennaioli,RafaelLaPorta,FlorencioLopez-de-Silanes,andAndreiShleiferpresentsubnationaldataofincomeintheirpublication36Thedatacovers 1,569subnationalregionsfrom110countriescovering74%oftheworld’ssurfaceand97%ofitsGDP.SpatialDataGeographicallybasedEconomicdata(G-Econ)byWilliamNordhausand XiChen.Thedatasetcovers“grosscellproduct”forallregionsfor1990,1995,2000,and2005andincludes27,500terrestrialobservations.Thebasicmetricistheregionalequivalentofgrossdomesticproduct.Grosscellproduct(GCP)ismeasuredata1-degreelongitudeby1-degreelatituderesolutionataglobalscale. WordpressEditPageOurWorldinDataisfreeandaccessibleforeveryone.Helpusdothisworkbymakingadonation.Donatenow