Thünen and the New Economic Geography

RegionalScienceandUrbanEconomics42(2012)907–912

ContentslistsavailableatSciVerseScienceDirect

RegionalScienceandUrbanEconomics

journalhomepage:www.elsevier.com/locate/regec

ThünenandtheNewEconomicGeography☆

MasahisaFujita?

ResearchInstituteofEconomy,TradeandIndustry,1-3-1Kasumigaseki,Chiyoda-ku,Tokyo100-8901,Japan

articleinfoabstract

Inthispaper,IexplainThünen'spioneeringworkonindustrialagglomeration.Inmyopinion,Thünen'sthink-ingonindustrialagglomerationwasnotonlyamazinglyadvancedforhistime,butinmanyrespectsremainsnoveleventoday.ItisshownthatifweunifyThünen'swell-knowntheoryonagriculturallandusewiththispioneeringworkonindustrialagglomerationbyusingmoderntools,thenweessentiallycomeupwithapro-totypeofNewEconomicGeographymodel.

?2011ElsevierB.V.Allrightsreserved.

Articlehistory:

Received20July2011

Receivedinrevisedform18November2011Accepted20December2011Availableonline5January2012JELclassi?cation:R10Q10D50

Keywords:ThünenCities

Neweconomicgeography

1.Introduction

oneofthe?rstmodelsofgeneralequilibriumanddidsointermsofreal-isticeconometricparameters.”(Samuelson,1983,p.1468,emphasesbytheoriginalauthor)Morespeci?cally,Samuelsonassertsthat“Thünen'smodelhasinitelementsofallofthefollowingsystems:

1.TheRicardo–Torrenstheoryofcomparativeadvantage.2.TheMalthus–West–Ricardotheoryofrent.

3.TheHecksher–OhlinandStolper–Samuelsontheoryoffactors-and-goodspricing.

4.TheMarx–Dimitriev–Leontief–Sraffasystemofinput–output.”(p.1481)Thisispraisesograndthatnoothereconomistinhistory,exceptpos-siblyAdamSmith,couldpossiblydeserveit.Asalocationtheoristmyself,however,Iwouldliketoemphasizeinthispaperthatfromtheviewpointoflocationtheory,Thünen'sworkcontainsmorethanthat.InadditiontothefourelementscitedabovebySamuelson,Iwouldliketonotethatinhislaterwork,Thünenalsoanticipatesthefollowingtheories:5.TheMarshall–Webertheoryofindustrialagglomeration6.TheChristaller–L?schtheoryofcentralplacesystem7.Therecentdevelopmentofneweconomicgeography

Beforeelaboratingthesepointsinthesubsequentsections,letusre-callthatThünenwasnotanordinaryarmchairscholar.Asiswell-known,heneveroccupiedanacademicpositioninhislife.TheinitialideaofthesocalledThünenringshasevolvedtoagrandtheoryofthegeneralequilibriumofaspatialeconomywhilehewasworkingonhisownTellowestate,engaginginceaselessagriculturalimprovementonhisland.Thünenwassatis?edwithhisabstractmodelonlyaftertaking

“Wherevereconomictheoryisstudiedtoday,hisideas,hisworkingmethods,theproblemsheposed,haveprovedseminalrightuptothepresentday—eventhere,wherehisnameseemstohavebeenforgotten,Thünenhasworked.Hisworkshinesbrighterthanevertoday.Much,whichseemsself-evidenttous,goesbacktohim.Andstillnotallthericheswhichareburiedinhisworkarebroughttolight.To?ndthem,youmustwithoutdoubtreadTheIsolatedStatecarefullyandoften.”(ErichSchneider,1959,pp.27–8)1AccordingtoPaulSamuelsonin1983inhiscommemorativepaperatthetwo-hundredthanniversaryofThünen'sbirth,Thünen“notonlycreatedmarginalismandmanagerialeconomics,butalsoelaborated

☆ThispaperisbasedonlecturenotespresentedattheInternationalThünenConfer-ence2000,ontheoccasionofthe150thanniversaryofJohannHeinrichvonThünen'sdeath,theUniversityofRostock,September21–24,2000.IwouldliketothankJacquesThisse,IkedaMariaP.Makabenta,andtwoanonymousrefereesforhelpfulcommentsonearlierdraftsofthispaper.IamalsogratefultoGrants-in-AidforScienti?cResearchGrantA18203016(fromtheJapaneseMinistryofEducationandScience)for?nancialsupport.

?Tel.:+8133511361;fax:+81335018391.E-mailaddress:fujita-masahisa@rieti.go.jp.1ThisquotationisfromSchneider(1959)whichwaswrittentohonorThünen'snameonthe175thanniversaryofhisbirthin1958.TheEnglishtranslationquotedhereisfromtheintroductionbyPeterHalltotheEnglishtranslationofvonThünen(1826)byWartenberg(1966,p.xliv).

0166-0462/$–seefrontmatter?2011ElsevierB.V.Allrightsreserved.doi:10.1016/j.regsciurbeco.2011.12.002

908M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912

RentR1(r)R2(r)R3(r)distance fromthe TownvegetableswheatcattleFig.1.Thelandrentpro?leandvonThünenringswhenn=3.

laboriousinvestigationsofcostsandreturnsonhisTellowestateover10years,andthencon?rmingthatthecollecteddata?tteddirectlyintohismodel.Nowonder,Schumpeter(1954,p.446)calledThünen“oneofthepatronsaintsofeconometrics.”Inshort,Thünen'stimelessmodelofagriculturallanduseandrenthasbeen“cultivatedonland”lit-erallywhilehewasworkingasafarmer.Thispartofhisworkonagri-culturallanduseandrentwaspublishedasTheIsolatedStatein1826,latercalledPartIinordertoseparateitfromlatereditions.Today,al-mostanytextbookofeconomicgeographyandlocationtheoryexplainsthebasicideaofThünen'stheoryonagriculturallanduseandrent,usingadiagramofamonocentriceconomyillustratedinFig.1.

Later,themonocentriceconomymodelofThünenwasdevelopedinvariousformsutilizingmodernanalyticaltools.2Inparticular,Alonso(1964)reinterpretedthemonocentriceconomymodelofThü-nenbysubstitutingcommutersforfarmersandthecentralbusinessdistrict(CBD)forthetown,andgeneralizedThünen'scentralconceptofbigrentcurvestoanurbancontext.This“monocentriccitymodel”providedmodernurbaneconomicswiththetheoreticalfoundation.

Foralongtime,Ibelievedthatfromtheviewpointofgeographyandlocationtheory,thiswastheendofthestory.Ofcourse,evenifitwereso,Thünen'sstoryisalreadyoneofeternalscienti?cfascina-tion.Then,aterribleshockcametomewhenIreadrecentlySection2oftheso-calledPartIIofTheIsolatedState,editedbyHermannSchumacherandpublishedin1863,whichcontainsThü-nen'sposthumouspapersmainlydealingwiththeproblemsofspatialeconomyrelatedtotheoriginalPartI.3IfIborrowSchneider'swordsagain,myamazementwasthat“eventhere,Thünenhasworked”!“There”meansnootherthanNewEconomicGeography.

Inmyopinion,Thünenisindeed“thefoundinggod”ofmoderneconomicgeography,4whichincludesnotonlytraditionaleconomic

2ForearlycontributionstoasystematictreatmentofThünen'sideasinfullymathe-maticalform,seeLaunhardt(1885,ch.30),L?sch(1940,ch5)andDunn(1954).Itmustbenoted,however,thattheaspectofwagedeterminationoffarmershasbeencompletelyneglectedinalmostallsubsequentvariationsofThünen'smodelexceptintherecentgeneralequilibriumanalysesbySamuelson(1983)andNerloveandSadka(1991).3Actually,IreadonlyextractsofSection2becauseIamunabletoreadGerman,andonlyextractsareavailableintheEnglishtranslationbyCarlaM.Wartenberg(1966).4Samuelson(1983,p.1468)notesthat“Amonggeographersandlocationtheorists,ThünenisafoundingGod.”

geographyandlocationtheory,butalsothemodernurbaneconomicsaswellastheso-calledNewEconomicGeography.Asexplainedinthenextsection,Thünenwasconcernednotonlywiththeworkingoftheagriculturalhinterlandsurroundingasingletown,butalso,atleastinhislateryears,concernedwiththeworkingoftheentirespatialecon-omyincluding“theorderanddistributionoftownsintheIsolatedState.”52.Thünenandagglomerationeconomies

Whenwediscussthereasonsfortheconcentrationofanindustry(orindustries)inaspeci?clocation,ormoregenerally,theagglomer-ationofpeopleandindustriesinacity(orinasystemofcities),ithasbeencustomarytogobackasfarastoMarshall(1890,1920,Ch.x),andthentoWeber(1929)andHoover(1937),andtothecentralplacetheoryofChristaller(1933)andL?sch(1940).Tothebestofmyknowledge,thereisnoarticleintheeconomicsliteraturewhichassociatesthetopicofagglomeration(ordistributionofcities)withThünen.6Locationtheoristsandeconomicgeographers(both“tradi-tional”and“new”)alwaysreferredtoThünen,butneverinthecon-textofagglomerationeconomiesorcityformation.Thus,itisagreatsurprisetorealizethat(usingSchneider'swordsagain)“eventhere,wherehisnameseemstohavebeenforgotten,Thünenhasworked”!

ToseewhereThünenhasworked,itmaybesuf?cienttoglanceoverthefollowingcontentsofPartII,Section2oftheEnglishtransla-tion(Wartenberg,1966),whichcontaintheextractsofposthumouspapersonlocationtheorywrittenbyThünenbetween1826and1842.7PartTwo,SectionTwo,…Subsection5Changesinourassump-tions(manytownsofthesamesizeequi-distantfromeachother),Subsection6TheorderanddistributionoftownsintheIsolatedState,Subsection7Theroleofpopulationdensity,Subsec-tion8ThedynamicsoftheIsolatedState:arethereobstaclestoitsexpansion?

GiventhatIhaveaccessonlytotheextractsofThünen'sworkinthispart,ImustbecarefulwithwhatIsayhere.Nevertheless,letmediscussmyunderstandingofThünen'sworkwithanemphasisonSubsection6.Asweallknow,inPartI,ThünenassumedthattheIsolatedStatecontainedonlyasinglelargetown.Incontrast,inSub-section5,Thünenassumesthattheplaincontainsmanysmalltownsofthesamesizelocatedequidistantfromeachother,andinvestigateshowthesizeoftownsandthedistancebetweenthemaffectthepat-ternofagriculturalproductionandlandrent.However,inthissubsec-tion,Thünendoesnotdiscusshowthesizeofeachcityisdeterminednorhowthedistancebetweencitiesisdecided,whicharethemaintopicsinthenextsubsection.

InSubsection6,Thünenasks“Whatdeterminestherelativeposi-tionofthetownsintheIsolatedStateinrespectofsizeanddistancefromeachother?”(p.285)8Instudyingthisfundamentalproblem,Thünenfurtherdividestheproblemintoanumberofmoreconcretequestions.

5ThiscitationisfromthetitleofSection2inPart2(Extract)intheEnglishtransla-tionbyWantenberg(1966).6ThisexcludestheoverallintroductionofTheIsolatedStatebyPeterHallintheEnglishtranslationbyWartenberg(1966),inwhichHallnotesthat“justbeforehisdeath,hewasalreadyreachingoutintogenerallocationtheory,andgropingto-wardsconceptswhichwereproperlydevelopedbyothersdecadeslater.”(p.xliv)7Thetitlesofsubsections1to4areomittedheresincetheyarenotrelatedtothepresenttopicofagglomeration.8Eachquotationbelowwithouttheauthor'snameisfromtheEnglishtranslationbyWartenberg(1966).

M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912909

The?rstquestionisabouttheconcentrationofpeopleinlargetownssuchastheCapital.Ifthe(urban)populationoftheeconomyweredividedamongmanysmalltowns,theneveryconsumerwouldbeeasilyaccessibletoagriculturalsupply.Thus,Thünenasks:

“WhythenisthepopulationofthelargeTownnotdividedamongmanysmallones?Thereasonsareasfollows:Inpractice,depositsofore,saltandcoalaremostunevenlydistributed....Thefocalcen-terofacountryisthenaturalresidenceoftheheadofthegovern-ment:theseatofthehighestof?cesofjusticeandadministration,ofarmyheadquarters,thehigherinstitutesoflearning,artcollec-tions,etc.Thepresenceofthecapitalofthecourt,theconcourseofscholars,menofscienceandstateof?cials,thetheaters,museums,etc.,affordmanymoresocialattractionsandamenitiesthantheprovincescouldeveroffer…Tocaterfortheneedsandpleasuresofallthecitizensassembledinthecapitalforanyofthesereasons,agreatmanypeopleoftheartisanandserviceclassarere-quired…”(p.286)

Intermsofmodernurbaneconomics,the?rstreasonrefersto?rstnature,thesecondtocentralmanagementfunctionsandpublicser-vice,thethirdtosocialandculturalamenities,andthelasttonon-tradableconsumergoodsandservices.Itisinterestingtonotethatifthenaturalamenitieswereincludedinthe?rstfactor,thenallfourfactorsabovehavebeenincreasinglyemphasizedrecentlyasim-portantdeterminantsofpopulationagglomerationinfuturecities(e.g.,Glaeser,1998).Notice,however,thatwhenThünenaskstherea-sonsforagglomerationofpeopleinalargetown,hedoesnotconsiderthosereasonsfortheagglomerationofindustriesandtheirworkers.Thisisthequestionheposesnext:

“Offargreaterimportanceanddif?cultyisthequestion:whetherindustrieswhichdrawtheirrawmaterialsfrom,andsellmostoftheirproductsto,theprovinces,arealsobetterofflocatedinthecapital?”(p.286)

Ininvestigatingthisquestion,?rstheasks“thereasonsagainstthelocationofindustriesinthecapital,”orthecentrifugalforces(usingtheterminologyoftheNewEconomicGeography)againstindustrialagglomeration.Thünen'sansweris:

“1.Rawmaterialsaremoreexpensivethaninthecountrytownsonaccountofthehighercostoftransport.2.Manufacturedarti-clesincurthecostofhaulagetotheprovincialtownswhentheyaredistributedtotheruralconsumers.3.Allnecessities,espe-cially?rewood,aremuchmoreexpensiveinthelargetown.Soisrentfor?atsandhouses,fortworeasons(1)constructioncostsarehigherbecauserawmaterialshavetobebroughtfromadistanceandareconsequentlymoreexpensive,and(2)sitesthatmaybeboughtforafewthalersinasmalltownareverydear.Sincefood,aswellasfuelandhousing,costsomuchmoreinthelargetown,thewageexpressedinmoney,mustbemuchhigherthaninthesmallone.Thisaddsappreciablytopro-ductioncosts.”(pp.286–7)

Theconsiderationofcentrifugalforcesaboveissurprisinglycom-prehensiveevenincomparisonwiththerecentliteratureofNewEco-nomicGeography.Inparticular,theeffectsofhighlandrentsandhighfoodpricesonmonetarywagesinlargetownsareexplicitlyconsidered.

Next,Thüneninvestigatesindepththecentripetalforcesofindus-trialagglomeration:

“Thefollowingfactors,ontheotherhand,favorthelocationofin-dustriesinlargetowns:1.Onlyinlarge-scaleindustrialplantsisitpro?tabletoinstalllabor-savingmachineryandequipment,whicheconomizeonmanuallaborandmakeforcheaperandmore

ef?cientproduction.2.Thescaleofanindustrialplantdependsonthedemandforitsproducts.3.Thenumberofbuyersdepends,inprovincialtowns,onthenumberofcountrymencomingintoselltheirproducts,orpassingthroughontheirwaytothecapital.Forinstance,acountrymanmayvisitthecapitaltosellhisprod-ucts,anddecidetobuysomeliquor.Itwillbecheaperforhimtobuythisinthecapital,evenifitcostshimhalfathalermorethanhewouldpayintheprovincialtowntwomilesfromhisfarm,becausehewouldhavetomakeaspecialjourneytofetchthelocalalcohol.”4.“Forallthesereasons,largescaleplantsareviableonlyinthecapitalinmanybranchesofindustry.Butthedi-visionoflabor[andAdamSmithhasshowntheimmensein?u-encethishasonthesizeofthelaborproductandoneconomiesofproduction]iscloselyconnectedwiththescaleofanindustrialplant.Thisexplainswhy,quiteregardlessofeconomiesofmachine-production,thelaborproductperheadisfarhigherinlargethaninsmallfactories.”(pp.287–88)

Turningtothenatureoflabor-and-productmarketsinlargetowns,Thünencontinues:

“5.Peopleawareofpossessinganexceptionalskillortalentwillnotwishtowastetheirtimeonotherwork,wheretheycanachievenothingoutstanding,butwillmovetothecapital,tode-votealltheirenergytotheirparticularskill;inreturntheywillreapamplereward.Thusthecapitalattractsoutstandingtalents–amongbusinessmen,artisansandlaborersaswellasamongscholarsandcivilservants–andinthiswayisabletoobtainasig-ni?cantadvantageovertheprovinces.6.Thelargetownoffersbuyersandsellersfarmoreguaranteeofbeingabletobuyandsellatcurrentprices.Thegreatmerchanthasnotthetimetocon-siderthespecialsituationofhiscustomerand?xthepriceofthearticlehewantstosellaccordingtothebuyer'sneedsorknowl-edge.Hehasanestablishedprice;whichprotectsthecustomerfromsharppractice.Besides,inthepresenceofsomanycompeti-torstheattempttocheatthecustomerwouldbescarcelyworththetrouble.”(p.288)

Finally,Thünenconsidersthelinkageor“association”amongin-dustries:

“7.…Sinceittakesmachinestoproducemachines,andthesearethemselvestheproductofmanydifferentfactoriesandwork-shops,machineryisproducedef?cientlyonlyinaplacewherefac-toriesandworkshopsarecloseenoughtogethertohelpeachotherworkinunison,i.e.inlargetowns.Economictheoryhasfailedtoadequatelyappreciatethisfactor.Yetitisthiswhichexplainswhyfactoriesaregenerallyfoundcommunally,why,evenwheninallotherrespectsconditionsappearsuitable,thosesetupbythemselves,inisolatedplaces,sooftencometogrief.Technicalin-novationsarecontinuallyincreasingthecomplexityofmachinery;andthemorecomplicatedthemachines,themorethefactorofas-sociationwillenterintooperation.”(pp.289–90)

Thünen'sanswerabovetothequestionofcentripetalforces(to-wardstheindustrialagglomerationinlargetowns)isamazingforthreereasons.First,whenThünenwrotethearticleonthistopic,Germany(inparticular,Tellowarea)hadnotyetexperiencedtheIndustrialRevolution,thusindustriesinmostGermancitieswereratherprimitive.Hence,forafarmer-scholarathisTellowestateiniso-lation,thiswritingrepresentsanamazinglyimaginativeworkbasedsolelyonhisinsights.Second,however,Thünen'sexplanationaboveissosystematicandcomprehensivethatitcouldbecomeagoodbasisforwritingonagglomerationeconomiesinamoderntext-book.Finally,fewscholarsseemtohavepaidattentiontothistrulypioneeringwork.

910M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912

Inhisin?uentialmonograph,GeographyandTrade,whichmarkedthebirthoftheNewEconomicGeography,PaulKrugmanstatesthefollowing(1991b,pp.14–15):

“ThebasicstoryofgeographicconcentrationthatIwillproposeherereliesontheinteractionofincreasingreturns,transportationcosts,anddemand.Givensuf?cientlystrongeconomiesofscale,eachmanufacturewantstoservethenationalmarketfromasin-glelocation.Tominimizetransportationcosts,shechoosesaloca-tionwithlargelocaldemand.Butlocaldemandwillbelargepreciselywherethemajorityofmanufacturerschoosetolocate.Thusthereisacircularitythattendstokeepamanufacturingbeltinexistenceonceitisestablished.”

NoticethatwhenThünen's?rstfouragglomerationfactorsarecombined,itcoincidesalmostexactlywiththeaboveKrugman's“basicstory”fortheemergenceofacore–peripherystructureonana-tionwidescale.9Furthermore,ifwecombinethe?rstfourfactorsofThünenwithhislast(7th)agglomerationfactor,whichconcernstheinter-industrylinkageor“association”,thenitnowagreeswithanotherbasicstoryofNewEconomicGeographywhichex-plainsthelocalizationofparticularindustries,bothinternationallyandintranationally,basedontheavailabilityofintermediategoodsandinput–outputlinkages.10Moreover,Thünen's?fthfactor,whichconcernstheself-selectingmigrationprocessandtheimpactofthesizeofthelabormarketonjob-matchingamongheterogeneousworkers,hasbeenmodeledjustrecentlyinNewEconomicGeographyusingagametheoreticap-proach(HelsleyandStrange1990,andHamiltonetal.,2000).ThesamenoteappliestothepartofThünen'sfourthfactorofagglomera-tionwhichreferstotheintra-industryspecializationincities(see,forexample,BeckerandHenderson,2000).Likewise,inthethirdfactorofagglomeration,Thünenmentions,ineffect,joint-tripsasanimpor-tantcauseofindustrialagglomerationinlargetowns,whichhasnotbeenexplicitlymodeledyetintheNewEconomicGeography.Last,Thünen'ssixthfactorofagglomerationreferstotwoseparateeffectsofmarketsize:(i)morecompetitors(inanindustry)inatownwillleadtolowerprices,hencebene?ttingbuyers,and(ii)thepooling-effectsofalargermarketwillleadtomorestableprices,bene?ttingbothsellersandbuyers.Althoughthe?rsteffectiswellconsideredinNewEconomicGeography,thesecondoneisnotexplicitlymod-eledyet.Therefore,wecanconcludethatThünen'sthinkingaboutin-dustrialagglomerationwasnotonlyquiteadvancedforhistime,butalsoitremainsnovelinseveralwayseventoday.

Finally,concerningthelocationofindustriesinthecontextofasystemofcities,Thünenstates:

“Whenweweightheargumentsforthelocationoffactoriesandtradesincountrytownsagainstthoseinfavoroftheirconcentra-tioninthecapital,we?ndthenatureoftheindustrydeterminesitslocation.Factoriesandworkshopsprocessingrawmaterialsoflittlevalueinrelationtotheirbulkandweight,whichneednocomplicatedmachinery,noextensivedivisionoflabour,andwhichthereforecansupplytheirproductsalmostascheaplyonasmallscaleasonalargescale,belongproperlytotheprovincialtownsoreventothecountrysideitself.These,asIshowedinPartOne,includedistillingandLinenwearing.Allotherindustries,

9Tobeprecise,thereexistsomedifferences.Inparticular,Krugmanisexplicitinem-phasizingthecircularcausalitybetweentheagglomerationofindustriesandtheag-glomerationofworkersthroughdemandexternalities.But,Thünenisnotexplicitaboutthiscircularcausality.Thismayre?ecttheprimitivestageofmanufacturinginGermanybeforetheIndustrialRevolution,wheresuchdemandexternalitiesareweakbecauseofsmallexpenditure-shareofworkers'incomeonmanufactures.10Forarecentexpositiononindustrialclustering,see,forexample,Fujitaetal.(1999,PartIII),andFujitaandThisse(2002,Ch9and10).

whereoppositeconditionsobtain,havetheirrightfulplaceinthecapital.”(pp.290–291).

ThisobservationanticipatesthetheoryofhierarchicalcentralplacesystemdevelopedlaterbyChristaller(1933)andL?sch(1940).

Beforeturningtothenextsection,itisinterestingtocompareThünenandMarshallonindustrialagglomeration.Itiswell-knownthatMarshall(1890,1920,Ch.X)introducedthetrinityoftheso-calledMarshallianexternaleconomies:(i)linkages,(ii)thickmar-kets,(iii)knowledgespilloverandotherpureexternaleconomies.AmongthethreelocalizationforcesofMarshall,Thünennevermen-tionedthethird.Wecanconjectureseveralreasonsforthis,butthereisnowaytoknowtherealone.Inanyway,asageneralequilib-riumtheorist,Thünenwasclosertotoday'sNewEconomicGeogra-phythanMarshall.

3.Uni?cationofThünen'sideasandtheNewEconomicGeographyHistoricallyspeaking,theNewEconomicGeographyrepresentsarenewedinterestinthe“generaltheoryoflocationandspace-economy,”usingtheterminologyofIsard(1956),orinshort,thegen-erallocationtheory.11AccordingtoIsard,agenerallocationtheoryissupposedtoembrace“thetotalspatialarrayofeconomicactivities”inaneconomy.Or,ifweusethewordsofKoopmans(1957),insuchatheory,thelocationofalleconomicactivitiesandhence“thedistributionitselfisavariable.”(p.154).

Historicallymovingbackward,themostrecentwaveinthegener-allocationtheoryis,ofcourse,theNewEconomicGeographyinitiatedintheearly1990s.Asiswell-known,itscentraltopichasbeenhowtoexplaintheemergenceofacore–peripherystructureonanationwidescale,oraninternationalscale.BeforetheNewEconomicGeographyofthe1990s,however,thereappearedintheearly1980sanequallysuccessful,butlesswell-known,attemptatagenerallocationtheoryinurbaneconomics.Asnotedbefore,themostimportantlimitationofmonocentricurbanmodels(adualversionofThünen'smodel)istheiraprioriassumptionoftheexistenceoftheCBDitself.Anumberofurbaneconomists,includingmyself,havetriedtoovercomethislimitation,andsuccessfullydevelopedseveralprototypesoftheso-called“NonmonocentricUrbanModels”inwhichtheformationofthetotalspatialarrayofametropolis(includingtheformationofaCBDorCBDs)hasbeenendogenouslydetermined.12Often,theNewEconomicGeography(startedinearly1990s)andNonmonocentricUrbanModelstogetherarecalled“theNewEconomicGeography”(BaumontandHuriot,2000).Fromhereon,wefollowthisbroaderde?nitionoftheNewEconomicGeography.

Goingfurtherbackhistorically,therewereseveralearlierattemptstodevelopagenerallocationtheory.AnotableoneisbyWalterIsardhimselfinhisseminalbook,LocationandSpaceEconomy(1956).Asindicatedbyitssubtitle(i.e.,AGeneralTheoryRelatingtoIndustrialLo-cation,MarketAreas,LandUse,Trade,andUrbanStructure),Isard'sbookwaswrittenwiththeaimofnothinglessthaninitiatingthede-velopmentofagenerallocationtheory.Actually,Isard'sattemptre-?ectsanearlierideaofOhlin(1933)whoproposedthedevelopmentofa“generallocalizationtheory”byintegratingtradetheoryandloca-tiontheory.13However,asexplainedlater,suchearlierattemptstode-velopagenerallocationtheorywere,unfortunately,doomedtobeincomplete.

11ForacomprehensiveexpositionoftherecentdevelopmentintheNewEconomicGeography,seeFujitaetal.(1999),FujitaandThisse(2002)andCombesetal.(2008).12EarlyexamplesofNonmonocentricUrbanModelsareOgawaandFujita(1980),FujitaandOgawa(1982),andFujita(1988).Forasystematicexpositionofsuchmodels,see,forexample,Fujita(1990),BaumontandHuriot(2000),andFujitaandThisse(2002,Chs.6and7).13Forasystematicdiscussionofearlyattemptsforagenerallocationtheory,seeIsard(1956,Ch.2).ReferalsotoFujita(1999).

M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912911

Formydiscussiontoday,themostinterestingattemptatdevelop-ingagenerallocationtheoryistheonebyThünenhimself.AlthoughThünendidnot,ofcourse,usetheterm“generallocationtheory”,Iamsurethatwhenhewrotethesubsectionon“TheOrderandDistri-butionofTownsinTheIsolatedState”(pp.285–295),heintendedtoexplain(usingIsard'swords)“thetotalspatialarrayofeconomicac-tivities”intheIsolatedState.Indeed,Thünen'sattempttodevelopagenerallocationtheoryisnotonlytheoldestone,butalsotheclosesttotheNewEconomicGeographyforseveralreasons.First,inKrugman's(1991a)seminalpaper,“IncreasingReturnsandEconomicGeography”,hiscentralquestioniswhentheeconomywouldbecomedividedbetweenamanufacturing“core”andanagricultural“periph-ery”.Asidefromthetreatmentofspace(acontinuousspacebyThü-nenandadiscreteonebyKrugman),thecore–peripherystructureofKrugmanisessentiallythesameasthemonocentricspatialstruc-tureofThünen.Second,itshouldbeclearfromwhatIexplainedintheprevioussectionthatThünenattemptstoexplaintheentirespatialstructureoftheIsolatedStateastheoutcomeofaprocessin-volvingtwoopposingforces:centripetalforceswhichfavortheag-glomerationofindustriesattheTown(or,inthecore),andcentrifugalforceswhichworkagainstindustrialagglomeration.And,asIalsoexplainedintheprevioussection,Thünen'stwosetsofforcesareessentiallythesameasthoseoftheNewEconomicGe-ography.Third,and?nally,bothThünenandtheproponentsoftheNewEconomicGeographyalwaysconsiderthedeterminationoftheeconomy'sspatialstructureintheframeworkofageneralequilibri-umtheory.Inparticular,thedeterminationoftheequilibriumwage(ateachlocation)formobileworkersisanessentialissueforbothofthem.14Wecan,therefore,concludethatifThünen'soriginaltheoryofmonocentricspatialeconomywerecombinedwithhispioneeringtheoryofindustrialagglomerationsuchthatauni?edmodeloftheIsolatedStateweredeveloped,thenitwouldbecomeatypicalgenerallocationmodelofNewEconomicGeography.

Inotherwords,inretrospect,whattheNewEconomicGeographyhasachievedistheuni?cationofthetwopioneeringideasofThünenforthe?rsttime.Indeed,whatKrugmandidinhispaperin1991isjustsuchauni?cationofThünen'stwoideasusingatwo-regionecon-omy(butwithoutknowingThünen'sideaonagglomerationecono-mies).Moresurprisingly,whatFujitaandKrugman(1995)didintheirpaper,“Whenistheeconomymonocentric?:vonThünenandChamberlinuni?ed,”in1995isexactlytheuni?cationofThünen'stwoideasintheoriginalframeworkoftheIsolatedStateincontinu-ousspace(again,withoutpreviousknowledgeofThünen'sworkonagglomerationeconomies).

Youmay,then,naturallyaskwhyThünendidnotdevelopsuchauni?edmodelhimself?IfwereadthefollowingsentencesbyThünen(p.295)abouttheimpactofthedevelopmentoftransportdevelop-mentonthedistributionoftownsintheIsolatedState,thenwecanseethatThünenhadactuallyagooduni?ed-modelinmind.

“Itisworthnotingthatrailwayconstructionwillrobofalltheirforcetheargumentsagainstthedevelopmentofthecapital,and

14ItisalsointerestingtonotethatbothThünenandKrugmandevelopedtheirspatialmodelbysimilarlydeviatingfromthethenmainstreamtradetheory.Inexplaininghisworkoneconomicgeography,Krugman(1991b,p.X)statesasfollows:“AsIworkedonthesubject,however,IfoundthatmyanalysiswasdriftingfurtherandfurtherawayfrominternationaleconomicsIknewit.…WhatIfoundmyselfgravitatingtowardwasastyleofmodelinwhichfactorsofproductionwereperfectlymobilebutinwhichtherewerecoststotransportinggoods.Inotherwords,Ifoundmyselfdoingsomethingclosertoclassicallocationtheorythantointernationaltradetheory.”While,incharac-terizingThünen'sspatialtheory,Samuelson(1983,p.1482)statesasfollows:“Ricar-diantradetheorytraditionallyassumeszerofactormobilitybetweencountriesorregionsand100percentcommoditymobilitybetweencountriesorregions.Thünen'smodelworksouttheoppositecase.Withinaregion,labormovesfreely(onimmobileland);goodsmoveonlyatacost.Wherelaborwilllocatewasnotaquestionthattradetheoryconsidered,butThünendid.”

willstrengthenthoseinfavorofsuchgrowth.Thuswemaysaywithcertaintythatrailwayswillmakeanimportantcontributiontothedevelopmentofthelargetowns,andthat,butforthefactthatrailwayswillpromotealsotheprosperityoftheruraldistrictssurroundingtheprovincialtowns,thelatterwoulddecayinconsequence.”

Infact,Thünen'sstatementabovecoincideswithoneofthemostimportanttheoretical?ndingsoftheNewEconomicGeography:thatis,(contrarytotheintuitivebeliefofmostpeople)thedevelop-mentoftransporttechnology(atleastinitsinitialphase)willstrengthentheagglomerationofeconomicactivities(operatingunderincreasingreturns)inthecoreregionorinlargecities.

Then,again,whydidn'tThünenwritedownhisuni?edmodel?Itismyconjecturethatbecauseoftheoreticallimitationsineconomicsathistime,Thünenwasunabletodoso.Toelaboratethepoint,letusrecallthatThünen'smodelofamonocentriceconomyisbasedonthethreeessentialassumptions:A1—exogenouslygivencenter,A2—con-stantreturns,andA3—perfectcompetition.And,asIexplainedbefore,inordertodevelopagenerallocationmodeloftheIsolatedStateinwhichtheemergenceofthetown(ortowns)itselfisendogenouslydetermined,thenwemustdropAssumption1(ofexogenouslygivencenter).But,ifwedropAssumption1,thenwecannotkeepAs-sumption2(ofconstantreturns)anymore.Thisisbecauseifmanufacturingindustrieswereunderconstantreturns,thentherewouldbenoreasonforthemtoconcentrateinasinglelocation;onthecontrary,theywoulduniformlydistributethemselvesinordertoavoidunnecessarytransportcosts,thustheeconomywoulddegen-eratetotheso-called“backyardcapitalism”.

Therefore,asThünenhimselfclearlyunderstood,inordertode-velopaninterestinggenerallocationmodeloftheIsolatedState,theexistenceofincreasingreturnsinmanufacturingisessential.But,asiswell-knowninmoderneconomictheory,increasingreturns(atin-dividual?rm-level)areinconsistentwithThünen'sthirdassumptionofperfectcompetition.Hence,inordertodevelopauni?edgenerallocationmodeloftheIsolatedState,Thünenneededanon-competitivegeneralequilibriummodeloftheIsolatedState.IfThünenhadinventedsuchanon-competitivegeneralequilibriummodelbyhim-self,ofcourse,hecouldhavedevelopedauni?edmodeloftheIsolat-edState.But,thisisaskingtoomuchofThünen.Infact,the?rstoperationalnon-competitivegeneralequilibriummodel,calledthemonopolisticcompetitionmodel,wasinventedonlyin1977byDixitandStiglitz(1977).Thismonopolisticcompetitionmodelprovidedthegenerallocationtheorywithanappropriatetheoreticaltool,andtheNewEconomicGeographyrepresentsthe?rstprototypewhichactuallysucceededinunifyingThünen'stwopioneeringworks.

4.Conclusion

Wehaveseenthatinthe?eldoflocationtheoryandeconomicge-ography,Thünenachieved(atleast)twodistinctpioneeringworks:Oneisthewell-knowntheoryoflanduseandlandrentintheagricul-turalhinterlandsurroundingtheTown,andtheotheristhelesswell-knownworkonthemechanicsofindustrialagglomerationandcityformation.And,inretrospect,whattheNewEconomicGeographyhasachievedisjusttheuni?cationofthetwopioneeringideasofThünenbyutilizingmoderntoolsofeconomicmodeling.

Still,however,amysteryremains.GiventhatThünen'sstudyonthemechanicsofindustrialagglomerationandcityformationhasbeensocomprehensiveandadvanced,whydidalmostnobodypayseriousattentiontoitoverthelastoneandahalfcenturies?Inthisre-spect,ImaycitefamouswordsbyJohannWolfgangvonGoethe.“Whycanweseethesunlight?”askedGoethe.And,hisanswerwas,“wecanseethesunlightonlybecausewehaveeyesthatresemblethesun.”

912M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912

Well,onlynowweareabletofullyappreciateThünen'sworkonthemechanicsofindustrialagglomerationandcityformationbecauseourscienceofeconomicgeographyhasadvancedjustthatmuch.Thismeans,however,usingSchneider'swordsagain,thatquitecertainly“stillnotallthericheswhichareburiedinhisworkarebroughttolight.To?ndthem,youmustwithoutdoubtreadTheIsolatedStatecarefullyandoften.”References

Alonso,W.,1964.LocationandLandUse.HarvardUniversityPress,Cambridge(Mass.).Baumont,C.,Huriot,J.-M.,2000.Urbaneconomicsinretrospect:continuityorchange?

In:Huriot,J.-M.,Thisse,J.-F.(Eds.),EconomicsofCities.CambridgeUniversityPress,Cambridge.

Becker,R.,Henderson,V.,2000.Intra-industryspecializationandurbandevelopment.

In:Huriot,J.-M.,Thisse,J.-F.(Eds.),EconomicsofCities.CambridgeUniversityPress,Cambridge.

Christaller,W.,1933,DieZentralenOrteinSüddeutschland,Jena,GustavFischerVerlag.

Englishtranslation:TheCentralPlacesofSouthernGermany,EnglewoodCliffs(N.J.),Prentice-Hall(1966).

Combes,P.-P.,Mayor,T.,Thisse,J.-F.,2008.EconomicGeography:TheIntegrationofRe-gionsandNations.PrincetonUniversityPress.

Dixit,A.K.,Stiglitz,J.E.,1977.Monopolisticcompetitionandoptimumproductdiversity.

AmericanEconomicReview67,297–308.

Dunn,E.S.,1954.Theequilibriumofland-usepatterninagriculture.SouthernEconom-icJournal21,173–187.

Fujita,M.,1988.Amonopolisticcompetitionmodelofspatialagglomeration:adiffer-entiatedproductapproach.RegionalScienceandUrbanEconomics18,87–124.Fujita,M.,1990.Spatialinteractionsandagglomerationinurbaneconomics.In:Chatterji,

M.,Kunne,R.E.(Eds.),NewFrontiersinRegionalScience.Macmillan,London.

Fujita,M.,1999.Locationandspaceeconomyathalfacentury:revisitingprofessor

Isard'sdreamonthegeneraltheory.TheAnnalsofRegionalScience33,371–381.Fujita,M.,Krugman,P.,1995.Whenistheeconomymonocentric?vonThünenand

Chamberlinuni?ed.RegionalScienceandUrbanEconomics25,505–528.

Fujita,M.,Ogawa,H.,1982.Multipleequilibriumandstructuraltransitionofnon-monocentricurbancon?gurations.RegionalScienceandUrbanEconomics12,161–196.

Fujita,M.andJ.-F.Thisse,2002,EconomicsofAgglomeration,CambridgeUniversity

Press.

Fujita,M.,Krugman,P.,Venables,A.J.,1999.TheSpatialEconomy:Cities,Regionsand

InternationalTrade.MITPress,Cambridge(Mass).

Glaeser,E.L.,1998.Arecitiesdying?JournalofEconomicPerspectives12,139–160.Hamilton,J.,Thisse,J.-F.,Zenou,Y.,2000.Wagecompetitionwithheterogenous

workersand?rms.JournalofLaborEconomics18,453–472.

Helsley,R.W.,Strange,C.,1990.Matchingandagglomerationeconomiesinasystemof

cities.RegionalScienceandUrbanEconomics20,189–222.

Hoover,E.M.,1937.Spatialpricediscrimination.ReviewofEconomicStudies4,

182–191.

Isard,W.,1956.LocationandSpace-Economy.MITPress,Cambridge(Mass).

Koopmans,T.C.,1957.ThreeEssaysontheStateofEconomicScience.McGraw-Hill,

NewYork.

Krugman,P.R.,1991a.Increasingreturnsandeconomicgeography.JournalofPolitical

Economy99,483–499.

Krugman,P.R.,1991b.GeographyandTrade.MITPress,Cambridge(Mass).

Launhardt,W.,1885,MathematischeBegründungderVolkwirtshafslehre,Leipzig,B.G.

Teubner.Englishtranslation:MathematicalPrinciplesofEconomics,Aldershot,EdwardElgar(1993).

L?sch,A.,1940,DieR?umlicheOrdnungderWirtschaft.Jena:GustavFisher.English

translation:TheEconomicsofLocation.NewHaven,CN:YaleUniversityPress,1954.

Marshall,A.,1890.PrinciplesofEconomics.Macmillan,London.8theditionpublished

in1920.

Nerlove,M.L.,Sadka,E.,1991.VonThünen'smodelofthedualeconomy.JournalofEco-nomics54,97–123.

Ogawa,H.,Fujita,M.,1980.Equilibriumlandusepatternsinanon-monocentriccity.

JournalofRegionalScience20,455–475.

Ohlin,B.,1933.InterregionalandInternationalTrade.HarvardUniversityPress,Cam-bridge(Mass).revisedversionpublishedin1968.

Samuelson,P.A.,1983.Thünenattwohundred.JournalofEconomicLiterature21,

1468–1488.

Schneider,E.,1959.JohannHeinrichvonThünenunddiewirtschaftstheoriedergegen-wart,SchriftendesVereinsfürSozialpolitik,N.F.14,Berlin.Schumpeter,J.A.,1954.HistoryofEconomicAnalysis.

vonThünen,J.H.,1826,DerIsolierteStaatinBeziehungaufLandwirtschaftundNatio-nal?konomie,Hamburg,Perthes.EnglishtranslationbyC.M.Wartenberg:vonThü-nen'sIsolatedState,Oxford,PergammonPress(1966).

Wartenberg,C.M.,1966.vonThünen'sIsolatedState.PergammonPress,Oxford.

Weber,A.,1909,UeberdenStandortderIndustrien,TübingenJ.C.B.Mohr.English

translation:TheTheoryoftheLocationofIndustries,Chicago,ChicagoUniversityPress,1929.

912M.Fujita/RegionalScienceandUrbanEconomics42(2012)907–912

Well,onlynowweareabletofullyappreciateThünen'sworkonthemechanicsofindustrialagglomerationandcityformationbecauseourscienceofeconomicgeographyhasadvancedjustthatmuch.Thismeans,however,usingSchneider'swordsagain,thatquitecertainly“stillnotallthericheswhichareburiedinhisworkarebroughttolight.To?ndthem,youmustwithoutdoubtreadTheIsolatedStatecarefullyandoften.”References

Alonso,W.,1964.LocationandLandUse.HarvardUniversityPress,Cambridge(Mass.).Baumont,C.,Huriot,J.-M.,2000.Urbaneconomicsinretrospect:continuityorchange?

In:Huriot,J.-M.,Thisse,J.-F.(Eds.),EconomicsofCities.CambridgeUniversityPress,Cambridge.

Becker,R.,Henderson,V.,2000.Intra-industryspecializationandurbandevelopment.

In:Huriot,J.-M.,Thisse,J.-F.(Eds.),EconomicsofCities.CambridgeUniversityPress,Cambridge.

Christaller,W.,1933,DieZentralenOrteinSüddeutschland,Jena,GustavFischerVerlag.

Englishtranslation:TheCentralPlacesofSouthernGermany,EnglewoodCliffs(N.J.),Prentice-Hall(1966).

Combes,P.-P.,Mayor,T.,Thisse,J.-F.,2008.EconomicGeography:TheIntegrationofRe-gionsandNations.PrincetonUniversityPress.

Dixit,A.K.,Stiglitz,J.E.,1977.Monopolisticcompetitionandoptimumproductdiversity.

AmericanEconomicReview67,297–308.

Dunn,E.S.,1954.Theequilibriumofland-usepatterninagriculture.SouthernEconom-icJournal21,173–187.

Fujita,M.,1988.Amonopolisticcompetitionmodelofspatialagglomeration:adiffer-entiatedproductapproach.RegionalScienceandUrbanEconomics18,87–124.Fujita,M.,1990.Spatialinteractionsandagglomerationinurbaneconomics.In:Chatterji,

M.,Kunne,R.E.(Eds.),NewFrontiersinRegionalScience.Macmillan,London.

Fujita,M.,1999.Locationandspaceeconomyathalfacentury:revisitingprofessor

Isard'sdreamonthegeneraltheory.TheAnnalsofRegionalScience33,371–381.Fujita,M.,Krugman,P.,1995.Whenistheeconomymonocentric?vonThünenand

Chamberlinuni?ed.RegionalScienceandUrbanEconomics25,505–528.

Fujita,M.,Ogawa,H.,1982.Multipleequilibriumandstructuraltransitionofnon-monocentricurbancon?gurations.RegionalScienceandUrbanEconomics12,161–196.

Fujita,M.andJ.-F.Thisse,2002,EconomicsofAgglomeration,CambridgeUniversity

Press.

Fujita,M.,Krugman,P.,Venables,A.J.,1999.TheSpatialEconomy:Cities,Regionsand

InternationalTrade.MITPress,Cambridge(Mass).

Glaeser,E.L.,1998.Arecitiesdying?JournalofEconomicPerspectives12,139–160.Hamilton,J.,Thisse,J.-F.,Zenou,Y.,2000.Wagecompetitionwithheterogenous

workersand?rms.JournalofLaborEconomics18,453–472.

Helsley,R.W.,Strange,C.,1990.Matchingandagglomerationeconomiesinasystemof

cities.RegionalScienceandUrbanEconomics20,189–222.

Hoover,E.M.,1937.Spatialpricediscrimination.ReviewofEconomicStudies4,

182–191.

Isard,W.,1956.LocationandSpace-Economy.MITPress,Cambridge(Mass).

Koopmans,T.C.,1957.ThreeEssaysontheStateofEconomicScience.McGraw-Hill,

NewYork.

Krugman,P.R.,1991a.Increasingreturnsandeconomicgeography.JournalofPolitical

Economy99,483–499.

Krugman,P.R.,1991b.GeographyandTrade.MITPress,Cambridge(Mass).

Launhardt,W.,1885,MathematischeBegründungderVolkwirtshafslehre,Leipzig,B.G.

Teubner.Englishtranslation:MathematicalPrinciplesofEconomics,Aldershot,EdwardElgar(1993).

L?sch,A.,1940,DieR?umlicheOrdnungderWirtschaft.Jena:GustavFisher.English

translation:TheEconomicsofLocation.NewHaven,CN:YaleUniversityPress,1954.

Marshall,A.,1890.PrinciplesofEconomics.Macmillan,London.8theditionpublished

in1920.

Nerlove,M.L.,Sadka,E.,1991.VonThünen'smodelofthedualeconomy.JournalofEco-nomics54,97–123.

Ogawa,H.,Fujita,M.,1980.Equilibriumlandusepatternsinanon-monocentriccity.

JournalofRegionalScience20,455–475.

Ohlin,B.,1933.InterregionalandInternationalTrade.HarvardUniversityPress,Cam-bridge(Mass).revisedversionpublishedin1968.

Samuelson,P.A.,1983.Thünenattwohundred.JournalofEconomicLiterature21,

1468–1488.

Schneider,E.,1959.JohannHeinrichvonThünenunddiewirtschaftstheoriedergegen-wart,SchriftendesVereinsfürSozialpolitik,N.F.14,Berlin.Schumpeter,J.A.,1954.HistoryofEconomicAnalysis.

vonThünen,J.H.,1826,DerIsolierteStaatinBeziehungaufLandwirtschaftundNatio-nal?konomie,Hamburg,Perthes.EnglishtranslationbyC.M.Wartenberg:vonThü-nen'sIsolatedState,Oxford,PergammonPress(1966).

Wartenberg,C.M.,1966.vonThünen'sIsolatedState.PergammonPress,Oxford.

Weber,A.,1909,UeberdenStandortderIndustrien,TübingenJ.C.B.Mohr.English

translation:TheTheoryoftheLocationofIndustries,Chicago,ChicagoUniversityPress,1929.

本文来源：https://www.bwwdw.com/article/eaja.html