relationship between markets

internet marketing writing question and need the explanation and answer to help me learn.

it should include the following parts:
A summary of the paper, which discusses the main research question(s), the methodology used by the authors, and the conclusions and contributions of the paper.
A discussion of whether the question or questions asked are important (or not) and whether you find the paper convincing or not (and why?)
A discussion of the concerns you have with the paper. These concerns could be related to the data being used, the empirical methods, sample selection issues (do the data seem to be a representative sample?), possible alternative explanations for the results that the authors do not consider (or do not consider sufficiently), and any other weaknesses of the paper.
Finally, a referee report is meant to be constructive criticism, so, whenever possible, provide possible suggestions for improvement.
Requirements: markets
ContentslistsavailableatScienceDirectExplorationsinEconomicHistoryjournalhomepage:www.elsevier.com/locate/eehCleanslate:Land-usechangesinSanFranciscoafterthe1906disasterJamesSiodlaDepartmentofEconomics,ColbyCollege,5245MayflowerHill,Waterville,ME04901,UnitedStatesARTICLEINFOJELclassification:N91R14R31Keywords:DisasterLanduseNaturalexperimentPersistenceSanFranciscoABSTRACTThe1906SanFranciscofire,whichdestroyedthousandsofbuildings,providedablankcanvasuponwhichtoreshapethecity.Afterreconstruction,andatatimeofimmensegrowthinthecity,developersshiftedlandoutofresidentialusesandintononresidentialusesinburnedareasrelativetounburnedareas.Theyfacilitatedthistransitionbyrebuildingfarfewersingle-familydwellingscomparedtoothertypesofhousing,whichsuggeststhathousesinhibitedtheconversionoflandtononresidentialusesbeforethefire.Asidefromthesebroadeffects,thefirealsoreleasedneweconomicpotentialinareasthathadshownlittleindicationofshiftingintononresidentiallandusesbefore1906,therebycreatingnewclustersofbusinessactivity.Theseimpactsofthefirearestillevidenttoday—inroughlythesamemagnitudesandplaces—whichsuggeststhattheeconomicbenefitsrealizeduponreconstructioncontinuetodrivethecity’sland-usepatterns.1.IntroductionInanycity,thepastcanconstrainthepresent.Oldbuildings,forinstance,reflectdecisionsmadeinabygoneeraandactasanchorstothemovingvesselofchange.Whilethesebuildingsreflecttheeconomicconditionsoftheirdayanddevelopers’expectationsforthefuture,theynonethelesscouldmissthemark.Butacatastrophiceventcanprovideanopportunitytorevisitthesepastdecisionsandmakechangesmoreinlinewithcontemporaryneeds.OneofthelargestcatastrophesinU.S.historyoccurredinSanFranciscoin1906whenanearthquakejoltedthecityandignitedamassivefire.Thefireitselfrazedthousandsofbuildings,therebycreatinganenormousamountofvacantlandatacrucialtimeinthecity’shistory.Withthiscleanslate,developerscouldmoreeasilyadjustlanduseastheyrebuiltfromthedevastation.Theimpactofthefirethusprovidesinsightintohowlargeshockscanreshapeurbansettingsandthedegreetowhicholdbuildingsstandinthewayofchange.Citiesoftenburnedinthenineteenthandearlytwentiethcenturies.Whiledestructive,urbanfiresalsoprovidednewopportunitiesinrebuilding.InSanFrancisco,residentialdensityincreasedsignificantlyinburnedareasrelativetounburnedareasafterthefirein1906(Siodla,2015).AnecdotalevidencesuggeststhatthespatialpatternsofbusinessesandresidentschangedaftermajorfiresinChicago,Boston,andBaltimore(Rosen,1986).AfterBoston’sfirein1872,landvaluesrosesignificantly,reflectingthevalueofvacanturbanlandandthepositivespilloversgeneratedbynewerandbetterbuildings(HornbeckandKeniston,2016).InChicago,industrialandcommercialinterestsestablishedafootholdinprimeareasafterthe1871fire(FalesandMoses,1972).1Largefiresoftenresultedinsignificantlyalteredurbanenvironments.Bigshockscanalsoimpactcitysystems,ifonlytemporarily.WorldWarIIbombinginJapanesecitiesimpactedpopulationgrowthandindustrialemploymentintheshort-term(DavisandWeinstein,2002,2008),andthesameholdsforcitypopulationsinhttp://dx.doi.org/10.1016/j.eeh.2017.04.001Received26February2016;Receivedinrevisedform20January2017;Accepted3April2017E-mailaddresses:jrsiodla@colby.edu.1FalesandMoses(1972)donotconsiderpre-fireland-usepatterns,andthusthereisnorelativecomparisontoaperiodbeforethefire.Explorations in Economic History 65 (2017) 1–16Available online 12 April 20170014-4983/ © 2017 Elsevier Inc. All rights reserved.MARK
WestGermanyafterthewar(Brakmanetal.,2004).InVietnam,U.S.bombingdidnotpermanentlyimpacturbanpovertylevelsorpopulationdensities(MiguelandRoland,2011).Theseresultssuggestthaturbangrowthpatternsareoftendeterminedbylocationalfundamentals,orgeography.Butself-reinforcingagglomerationeconomies,orincreasingreturns,canalsoplayaroleindeterminingurbanoutcomes(Boskeretal.,2007,2008;Reddingetal.,2011;Ahlfeldtetal.,2015;Imaizumietal.,2016).Citiesstillexistinareaswheretheinitialdriverofsettlementnolongermatters,suchasalongoldportagesitesintheU.S.(BleakleyandLin,2012),andnearhistoricalrailroadinvestmentsinAfrica(JedwabandMoradi,2016;Jedwabetal.,2017)andSweden(BergerandEnflo,2017).Pathdependenceisalsoseenwithincities.Forexample,areasneardefunctstreetcarstopsarestillverydenselybuiltinLosAngeles(BrooksandLutz,2014),andthehousingdensitygapthatemergedinSanFranciscoafterthe1906firestilllargelyexists(Siodla,2015).Theageofthehousingstockcanalsoinfluenceneighborhooddevelopmentcyclesandthuscontributetothelong-lastinglocationpatternsoftheurbanrichandpoor(Rosenthal,2008;BruecknerandRosenthal,2009).Allthissuggeststhathistorymattersinthestudyofcities.Thispaperaddstotheliteratureonshocks,persistence,andurbansettingsbyestimatingthe1906fire’sshort-andlong-runimpactsonlanduseinSanFrancisco.Usingburnedandunburnedcityblocksattheboundaryofthefire—whichwereotherwisesimilarexceptforthefire’simpact—thestudyemploysaborder-discontinuityapproachanddifferences-in-differences(DID)methodologytoestimatechangesinlanduseovertheyears1900,1905,1914,1931,and2011.Overall,relativetounburnedblocks,residentiallandsharesonburnedblocksfellwhilenonresidentiallandsharesroseby1931.Thestudyalsoprovidesinsightintowhatheldthecitybackfrommakingthesechangesbefore1906:thepresenceofoldresidentialbuildings.Inreconstruction,developersbuiltrelativelyfewerofthesebuildings,andthemajorityofthereductioncamethroughsingle-familyhouses.Also,asidefrommerelyexpandingnonresidentialusesinmanyneighborhoods,thefirecreatedeconomicopportunitiesinnewareas,resultinginclustersofbusinessactivitythatemergedonlyinthewakeofthedisaster.Theseeffectsofthefirestillremaintoday,andthuslargeshockscanbesufficientcatalystsforpermanentlyreshapingurbansettings.2.Historicalbackground2.1.LanduseinSanFranciscobefore1906MigrantsstormedSanFranciscointhenineteenthcenturyinsearchofeconomicopportunity(Stewart,2012;Walker,2000).Withthisfoundation,thecityflourished.SanFrancisco’spopulationgrew15percentbetween1890and1900,makingittheeighth-largestcityintheU.S.asthetwentiethcenturybegan(IsselandCherny,1986,pp.23–24,Table1).Thisgrowthinresidentswasmatchedbyanexpansionoftheworkforce,whichroseby11percentinthatdecade(IsselandCherny,1986,p.54,Table3).2Manufacturing,inparticular,wasstrongbetween1899and1904,atimeinwhichthenumberoffactoriesincreasednearly30percent(Douty,1977,p.366,Table29).Asaconsequenceofallthisgrowth,assessedrealestatevalues—abarometerofacity’seconomicactivity—rosesteadilybetween1900and1906(SFMR(1904–1916)).3Leadinguptothefire,therealestatemarketwasexperiencinghistoricallyhighpricesandunprecedentedtransactionvolumes(SanFranciscoRealEstateCircular,December1906,p.4).Theriseinrealestatevalueswasgreatestforpropertiesusefulforretailandwholesalebusinessin1903(SanFranciscoRealEstateCircular,November1903,p.3).4Andby1905,whileinhighdemand,thesepropertieswerescarcelyforsaleindesirableareas(SanFranciscoRealEstateCircular,February1905,p.1).Buyersinthesemarketswereprevalentwhilesellerswerenot,thusindicatinganeedtoallocatemorelandtousesotherthanhousing.Thisneedwassogreatthatinvestorsinbusinesspropertybegangoingtopartsofthecityoutsidethecentralcoreatthistime(SanFranciscoRealEstateCircular,December1903,p.2).Asinmanyoldcities,firmsandindustriesinSanFranciscopreferredcentrallocationsforthebenefitsofagglomerationandaccesstotransportterminalstosaveonthecostofshippingandmovinggoods.Butconsumersalsodemandedlocationsnearthecenterofthecitytosaveoncommutingcosts.Thus,forfirmsandresidents,landrent—orwillingnesstopay—fallswithdistancetothecity’scenter.Sinceitisrelativelymoreexpensivetomovegoodsthanmovepeople,landrentforfirmsfallsfasterwithdistancetothecenterthanitdoesforconsumers,afeaturethathelpsexplaincentralareabusinessclusteringanddistanthousinginoldcities(MosesandWilliamson,1967).5InSanFrancisco,heavyinvestmentinelectricstreetcarlinesbeganchangingthelandscapebefore1906,connectingoutlyingresidentialneighborhoodswiththecenterofthecity(IsselandCherny,1986,p.30).Alongwitholdcablecars,streetcarsloweredthecostofcommuting,allowingresidentstolivefartherfromdowntownandthusencouragingsteadysuburbangrowthintheearly1900s.Consequently,outlyingareasexperiencedheightenedrealestateactivityjustbeforethefire(SanFranciscoRealEstateCircular,April1905,p.1).Meanwhile,landnearerthecitycenterwasstillhighlydevelopedwithbuildingsandusesthatmightotherwiseserveabetterpurpose.TalkofbetterpurposesforlandwasprevalentinSanFranciscobefore1906.AnillustrativeexampleconcernsthechangesproposedbyDanielBurnham,anarchitectandurbanplanner,inanambitiousplanpresentedtothecityjustdaysbeforethefire.2Theworkforceconsistsofthoseemployedinprofessionalservices,domesticandpersonalservices,tradeandtransportation,andmanufacturingandmechanicalindustries.3AsSiodla(2015)shows,higherlandvaluesencouragedhousingdeveloperstobeginbuildingdenserformsofhousing(i.e.,apartmentbuildings)intheseyears.Thismovetowarddensityinhousingwasanationaltrendintheearlytwentiethcentury(Barrows,1983).4Whileitisnotclearwhetherthepriceappreciationofretailandwholesalepropertieswasforplotswithbuildingsusedforthesepurposes,orforpropertieslocatedinareasadvantageousforthesetypesofbusinesses,itnonethelesssuggestsastrongoveralldemandformorenonresidentialspaceinthecity.5SeeChapter5inDiPasqualeandWheaton(1996)foradetaileddiscussionofthemodeldescribedhereinthecontextofoldcities.J.SiodlaExplorations in Economic History 65 (2017) 1–162
TheplanwasinspiredbytheprogressiveCityBeautifulmovementthatwassweepingthenation,whichheldthebeliefthatproperplanningandurbanaestheticscanenhanceresidents’qualityoflifeandpromoteavirtuouslivingenvironment.TheplancalledforasignificantrejiggeringofSanFrancisco’slanduse,includingwideningandcreatingnewstreets,whichwouldmeanthelossoflandformanypropertyowners(Fradkin,2005,p.227).Recognizingtheimmensecostsofsuchanendeavor,Burnhamhimselfsaid,“Itwilltakemoreyearsthanwewilllive;itwilltakemoremillionsthanwecanguess!”(SunsetMagazine,1905,p.120).Theplanwasneverimplemented,evenwiththewell-timedopportunitythefireprovided.AnotherprominentcaseconcernsBakerStreet.In1900,themayorofthecity,backedbythemajorityofvoters,triedtoissuebondstopurchaseamileofresidentialblocksalongBakerinordertocreateapleasantthoroughfarebefittingitspopularuseasagatewayintothePresidiopublicpark.However,thecourtsruledthatitwasanillegaluseofauthorityandnullifiedtheattempt(SunsetMagazine,1905).Whiletheseexamplesconcerntheuseofpublicauthority,landusewasalsoinhibitedforprivatedevelopers.Inordertoconvertaresidentiallotintononresidentialuse,adevelopermustalteranexistingstructuretofitbusinessneedsortearitdownandbuildasuitablebuilding.ManydifferentfactorscouldhaveinhibitedthischangeinSanFrancisco.Renovationandredevelopmentcostsmayhavebeenveryhigh.AnobserverwroteofMarketStreet,“Theshantiesrentwithoutdifficultyonthatstreet,andalowerrentpaysthelandlordaswellasamuchhigheroneplustheaddedcostofamodernbrickbuilding”(SanFranciscoRealEstateCircular,February1898,p.2).Propertyownersinanareacouldhaveheldoutinalandassemblycase(BrooksandLutz,2016),althoughthereisnoanecdotalevidencetosuggestthatthiswasasignificantprobleminSanFranciscoatthetime.Nonresidentialpropertydevelopersmayhavebeenreluctanttoerectbusinessbuildingsinpredominantlyresidentialneighbor-hoods.However,evenlargelyresidentialneighborhoodsinthecitywereintroducingmorenonresidentialusesbefore1906.6ButanotherfrictionsuggestedbycontemporariesinSanFranciscowasthepresenceofoldhousingthatownerssimplyhadnodesiretoimprove.In1899,onejournalistprescientlywroteaboutdowntownareas,“Tumbledownshanties,decayandrottennessmorecompletelycharacterizethebuildingstherenowthantheydidin’49and’50.Asweepingfirewouldbeofgreatbenefittothatportionofthecity[emphasisadded].Ownersseemdeterminednottoimprove”(SanFranciscoRealEstateCircular,March1899,p.2).Ownersandresidentsofthesepropertiesmayhavederivedsignificantutilityfromthestructuresthemselves,theneighborhoodsinwhichtheywerelocated,orevenfromthespatialconcentrationofpeoplesimilartothem,therebyprohibitingchangesinlandusebymaintainingthestatusquo(Rosen,1986).Repurposingoldresidentialpropertieswouldhaverequiredovercomingthesesignificantobstacles,andthuslanduseremainedstuckbefore1906.7Butthencameadisasterthatpulledupoldanchorsandprovidedopportunitiestoreshapetheurbanenvironment.2.2.The1906earthquakeandfireThedevastatingfireinSanFranciscobeganwithamassiveearthquake,whichjoltedthecityawakeonemorninginApril1906.Theshockrupturedgaslinesandtoppledlamps,causingsmallfiresthatquicklycoalescedintoalargeblaze.Ensuringitsspreadwasthetemblor’sdestructionofmuchofthecity’swaterlines,whichnotonlycutoffwatersupplyinmanyareas,butalsoreducedwaterpressureelsewhere,thusweakeningtheprimarydefenseagainstfire.Overthreedays,morethan28,000buildingswereburnedonnearlyfivesquaremilesofland(NOAA,1972).Fig.1depictsthefire’scoverage.Notonlydiditdestroycommercialandindustrialareasofthecity,butitalsoreachedintoresidentialareas:almosthalfofthecity’spopulationlivedwithinthefire’sboundaryin1905(Haasetal.,1977,pp.95–96).Thesuccessoffiresuppressionvariedacrossneighborhoods,someofwhichhadworkingwatercisternsorevenbarrelsofwinefordousingtheflames.WiththeexceptionofChinatown,whichmayhavebeenneglectedforracialreasons,firefightersandcitizensaimedtosuppressthefireonallfronts.8Althoughthemagnitudeoftheearthquakewassubstantial,fire-ravagedbuildingsborethebruntoftheshock.Thedestructionfromthefirealoneaccountedforatleast80percentofthetotaldamageinflictedbythedisaster,whichwasbetween$350and$500million,orroughly1.3to1.8percentofnominalU.S.GNP(OdellandWeidenmier,2004,p.1003).Thevastmajorityofthesebuildingswereinsuredagainstlossfromfire.9Intheprocessofreachingsettlements,insurancecompaniesoftenpaidamountsthatexceededtheirobligationsinordertocurrygoodfavorwiththepublic(Tobriner,2006).Thesesettlementsweresolarge—amountingtoalmost70percentoftotaldamagesbysomeestimates—thatitcontributedtothePanicof1907asgoldflowedintotheU.S.frominsurancecompaniesinBritain,whichhadmanyaccountsinthecity.TheBankofEnglandreactedbyraisinginterestrates,thusproddingtheU.S.economyintorecessionandbroadeningtheimpactofalocalizeddisaster(OdellandWeidenmier,2004).2.3.ReconstructionEarlyrebuildingeffortstendedtofavorbusinessinterests,whichlobbiedhardtoremovethepermittingprocessinreconstruction,andwealthycitizens,whoquicklyreceivedinsurancesettlementsandoftenrebuilttheirhousesinpre-disasterlocations(Fradkin,2005;Davies,2012).Buthousingwasnotlongunfavorablytreatedduringreconstruction:by1911,thecity’spre-6OnlythePacificHeightsneighborhoodsawadeclineinnonresidentiallandusebetween1900and1905.Theothersevenneighborhoodsinthestudy,someofwhichweremostlyresidentialbeforethefire,showedpositivegrowthinnonresidentialusesleadingupto1906.7Acorollarytohousingasafrictioninagrowingcityiswhatitinhibitsinadecliningcity(GlaeserandGyourko,2005).Thedurablenatureofhousingimpedesadjustmenttoasmallercity,leavingvacantmanyoldhomesandthuslengtheningtheprocessofurbanchange.Similarly,Field(1992)showsthatmuchofthefrenziedsubdivisionactivityofthe1920sconstrainedrecoveryfromtheGreatDepressionbycreatingfrictionsthatimpededlandreorganization.8Chinatownisoutsidethisstudy’ssamplearea.Davies(2012)givesaniceaccountofthefirefightingeffortsbyneighborhood.9Areportofinsurancesettlementsestimatedthatonlyabout5percentofburnedpropertieswereuninsured(Whitney,1972,p.11).J.SiodlaExplorations in Economic History 65 (2017) 1–163
disasterhousingcapacitywasattained(Haasetal.,1977,p.73).10Inlinewiththisrecoveryinthehousingstock,SanFrancisco’spopulationgrew21percentbetween1900and1910,evenwithsignificantout-migrationafterthedisaster,andtheworkforce,withthehelpoftradesneededinreconstruction,increasedby37percent(IsselandCherny,1986,p.24,p.54).Thecityhadquicklybouncedback—almostasifnodisasterhadoccurred—andreestablisheditselfastheeconomiccenteroftheWest.Whileshakenbutunburnedbuildingswereoftenrepaired,burnedsiteswerelargelyrebuilt.Acrossthecity,28,507buildingswereconstructedbetween1906and1914,thusreplacingthe28,188thatwerelostinthedisaster.Thenewbuildingscontainedvirtuallythesamematerialsastheoldbuildingstheyreplaced,sothatsimilarbuildingtechnologieswereusedbeforeandafterthefire(seeSiodla,2015,p.51,Table2).Somehavesuggestedthatthecitywaslargelyrebuiltby1910,withareturntonormalcyin1915whenithostedthePanama-PacificInternationalExposition(Fradkin,2005,p.196).Evenso,itwasnotuntilsometimeaftertheExpositionthatburnedareashadbeencompletelyredeveloped.11Indeed,theperiodbetween1906and1930isconsideredoneofthemostintensivebuildingspreesinSanFrancisco’shistory(Godfrey,1997).Allthisrebuildingrequiredcapital,asidefromthesubstantialinsurancepaymentsthatpropertyownershadreceived.Contemporariesfeltthatdevelopershadreadilyobtainedmoneytorebuildthecity,andeventhelostliquidityduringthePanicof1907helpedbypreventing“over-building”(SanFranciscoRealEstateCircular,September1910,p.2).12Ultimately,capitalwasnotaconstraintinfinancingnewstructures,andanystringencythatdidexistwouldhavefavoredtheconstructionofcheaperhousingovermoreexpensivenonresidentialbuildings.Withcapitalinhand,developerswantedfreedomtoconstructthetypesofbuildingstheyenvisionedforthecity.Undernormalcircumstances,buildingregulationscouldstandintheway.Butthiswasnotthecaseafterthefire.Newcodessuggestedbycityofficials,suchasheightlimitsandfire-resistantwalls,weredefeatedorignoredintherushtorebuild(Fradkin,2005,p.244).Thecityonlyslightlyenlargeditsfireproofconstructionzone,andimplementedanewfireproofroofareathatdidlittletochangeconstructionpracticesinceitstillallowedwood-frameconstruction(Tobriner,2006).13Zoningwasalsolargelyabsentatthetime.Fig.1.Burnedareafromthe1906SanFranciscofire.Note:Thedarkportionofthemaprepresentstheburnedarea.Source:SEIC(1908).10Someofthishousingwasbuiltinoutlyingareasofthecityasaworking-classsuburbanization,whichwasatlargeinmanyU.S.citiesatthetime,washastenedbythefire(Davies,2012,p.123)11Roughly30percentoflandontheaverageburnedblocknearthefire’sboundary,whichmakesupthisstudy’ssample,wasstillvacantin1914,whereaslessthan10percentwasvacantin1900.By1931,only6percentoflandonburnedblockswasvacant,whichmatchestheaverageproportionofvacantlandinunburnedareasatthetime.12Itisnotclearfromwhoseperspective,asocialplanner’sorbuildingowners,contemporariesmayhaveconsideredbuildingtobeexcessive.13Thechangestothefireproofconstructionzoneandfireproofroofareaarenotaccountedforinthemainanalysisbecausetheyareresponsestothedisasteritself.Nevertheless,controllingfortheseregulatorychangesdoesnotalterthemainresultsofthestudy.J.SiodlaExplorations in Economic History 65 (2017) 1–164
Thecitypasseditsfirstzoningordinancein1921,butnoofficialregulatorybodyoversawituntil1928.ZoningthusplayedarelativelyminorroleuntilaftertheGreatDepression,whenbuildingconstructionslowlypickedupwiththeeconomy(Weiss,1988).Withtheabsenceofstrictzoningandbuildingregulations,SanFrancisco’sredevelopmentoccurredwithinafreemarketsettinginwhichprivateinterestsguidedthecity’srisefromtheashes.Anotherimportantconsiderationinpost-disasterreconstructionarechangesinthepublictransportationsystem.Someneighborhoodsexperiencedalossincablecaraccess,includingMissionDistrictandpartsofWesternAddition.14Buttheseareasnonethelesscontinuedtoflourishafterthefire.Encouragingthisgrowthwastheelectricstreetcar,whichgrewinprominenceduringtherebuildingprocess.Justonemonthafterthefire,inahastydesiretorebuild,electricoverheadlinesweredeemedthequickestroutetorecoveryandmanysuchinvestmentsweremadeonoldcablecarroutes(Fradkin,2005,p.327).15Municipaltransportationservicesalsobeganin1909,competingwithprivatecompaniesthatwereoperatingmanyofthecity’slinesuptothatpoint.Intheend,noareathatwaspreviouslyservedbysomeformofpublictransportationwaswithoutaccesstoserviceafter1906.2.4.Landuseandthefire:ahypothesisThedisaster’stimingprovidesuniqueinsightintothenaturaleconomicforcesatworkincitiesattheturnofthetwentiethcentury.SanFrancisco’sconstructiontechnology,buildingandland-useregulations,andtransportationnetworkremainedlargelyconstantintheimmediateyearsbeforeandafter1906.Theprimaryagentofchangeatthistimewasthusthefireitself,whichthroughthedestructionofthousandsofbuildings,clearedawayfrictionstochanginglanduseintheburnedarea.Manyobserverscertainlyviewedthefireasablessing.AccordingtotheSanFranciscoRealEstateCircular,“ThefiregaveSanFranciscoagreatcityhousecleaning.Alltheold,undesirablebuildingsweresweptaway;fourandahalfsquaremilesof‘improvements,’consistingofoldwoodenandbricktwo-storyshacks…werepurifiedbyfire”(September1910,p.2).AdvocatesoftheCityBeautifulmovementfeltthatthedestructionofSouthofMarket,havingundesirablefeaturesassociatedwiththeunmarriedworking-classcitizenswhosettledthere,clearedthewayforpositivechangesintheneighborhood.EventhePortlandOregonianreportedthatthelossofoldshacksintheareawasnolosstothecity(Davies,2012,p.114).Withdecrepithousesandotherbuildingsreducedtoashes,landownerswereatanimpasseandforcedtomakeadecisionaboutthefutureoftheirproperties.Thepotentialforchangewasimmense.HistoricalanecdotessuggestthatthedisasterhadthecapacitytosignificantlyreshapeSanFrancisco.Ifso,whatwouldbetheexpectedresult?Intheheartofthecity,wherethefireburned,SanFrancisco’snonresidentialpropertymarketwasstrongerthanitshousingmarketleadingupto1906.Thisstrengthsuggeststhatthecityneededmoreoverallspaceforcommercial,industrial,wholesale,andretailuses.Also,amongotherthings,stand-alonehouses(i.e.,shantiesandshacks)wereseenasanuisancethatinhibiteddesirablechangesinlanduse.Giventheseconditions,thefire’sdestructionshouldhaveresultedintwooutcomesuponredevelopmentofburnedareas:ashiftinlandusetowardsnonresidentialusesandawayfromhousing,andalargereductioninthenumberofsingle-familyhouses.16Furthermore,iftheland-usechangesgeneratedstrongpositivespilloversacrossnewbuildings,whichtendtoremainfordecades,thenthefire’simpactislikelytopersistthroughtime.3.DataThisstudyreliesmostlyondatacollectedfromhistoricalmapsforcityblocksalongthefire’sboundary.17Theshapesoftheseblocksremainedvirtuallyunchangedafterthedisaster,makingthemreliableunitsofcomparisonacrosstime.Thefocusofthestudyisontheshareoftotalblockacreagedevotedtoresidentialandnonresidentialuses.18Residentiallandshareistheproportionoflandusedexclusivelyforsingle-familyhouses,flatbuildings,multi-familydwellings,andapartmentbuildings.Nonresidentiallandshareistheproportionoflandusedforcommercial,office,retail,industrial,institutional,educational,medical,andsmallopen-spaceuses(e.g.,parks).19TheselandsharesweredeterminedusingfireinsurancemapsproducedbytheSanborn-PerrisMapCompanyfortheyears1899/1900,1905,1913/1914,and1928/1931(Sanborn,1899–1914,1905,1928–1931),thusyieldingtwoperiodsthatestablishpre-disasterdevelopmenttrendsandtwopost-disasterperiodsthathighlighttheshort-runeffectsofthefire.Forsimplicity,thepaperwillrefertotheyears1900,1905,1914,and1931.Thefocusattheendofthestudyisonthelong-runimpactofthefire,whichisestimatedin2011usingmodernland-usedataobtainedfromthecity’splanningdepartment(SFLUD,2011).Fig.2showsfourcityblocks—eightpolygons—fromthe1905Sanbornsurvey.Byprovidinginformationaboutbuildinguse,the14Fig.3depictstheseneighborhoods.Mapsdepictingthepost-disasterchangesareavailableattheSanFranciscoCableCarMuseum’swebsite,http://www.cablecarmuseum.org/archive/8c/8c.htm.15Effortstoupgradethecity’stransportationsystembyconvertingcablecarlinestostreetcarlinesbeganbeingdiscussedin1904(Fradkin,2005,p.327).AreportissuedbyaSanFranciscoengineerin1905recommendedthatthecityinvestinelectrictrolleylinesandabolishhorse-drawncarsandcablecars(SanFranciscoCall,6December1905,p.1).16Mostresidentialbuildingsatthetimeweresingle-familyhouses.Attheboundaryofthefire,roughly58percentofresidentialbuildingswerehousesin1900.InAugust1921,theSanFranciscoRealEstateCircularpublishedtheresultsofacensusofhousingtakenbytheCityPlanningCommissionthatshowedthat,city-wide,65percentofresidentialbuildingsweresingle-familyhouses(p.2).17Blockdefinitionsarebasedonpre-fireassessorrecords,whichassignedanumbertoeachblock.18Anotherpossibleoutcomeislandacreage.However,landshareswereusedbecausetheyeasilyaccountforthesizedifferencesacrossblocks.Calculatingsharesalsorequiredfewerstepsandhencefewerpossibilitiesforerror.Ultimately,thegeneralconclusionsofthestudydonotdependonwhichoutcome—landacreageorlandshares—isconsidered.19Severalblocksweredevotedcompletelytoparkuseinallperiodsunderstudy.Theseblockswereexcludedfromtheanalysis.J.SiodlaExplorations in Economic History 65 (2017) 1–165
mapsshowtheamountofacreagedevotedtoresidentialandnonresidentiallanduseforeachcityblock.20Theremainderofablock’sland,ifany,isdevotedtomixeduses,semi-residentialuses(e.g.,tenements,hotels,boardinghouses,andhousekeepingrooms),andvacantland.21Theseothercategoriesmakeupaverysmallproportionoflanduseinanygivenyearandarenotconsideredintheanalysis.The1905maps,whichwerehousedinabuildingdowntownatthetimeofthefire,werefuelforthefireitself.Smallportionsofthesemapsweresinged,sothat59blocks(35burnedand24unburned)werepartiallyburned,affectingroughlyhalfanacreoneach.Thestudyassumesthattheseunreadablemapportionsremainedthesamebetween1900and1905,andthusreferencesthe1900mapsforland-useinformationinthesecases.Fortheremainingreadableportionsoftheseblocks,the1905mapswereused.ThemapshowninFig.1wascreatedin1908bytheStateEarthquakeInvestigationCommissionandwasusedtodistinguishbetweenburnedandunburnedblocks.Predominantlyburnedblockscomprisethetreatmentgroup,whilepredominantlyunburnedblocksmakeupthecontrolgroup.22Thesamplecontainsblockslocatedclosetothefire’sboundary,allofwhichexistedbeforeandafterthefire—i.e.,noneareonlandbuiltupafter1906.23Includingmoreblocksfarawayfromtheboundaryincreasesstatisticalpower,butpotentiallyintroducesdissimilarblocksintotheanalysis.Consideringthesetrade-offs,dataweregatheredfor421blocksclosetothefire’sboundary,185ofwhichwereburnedand236wereunburned.Fig.3showstheseblocks.24Blocksmaybelargelyvacantinanygivenyear.Sincethefocusofthisstudyisonwell-developedsites,thesampleisrestrictedtoblocksforwhichatleasthalfofthelandisdeveloped.Developmentcyclescausesomeblockstoenterorexitthesampleovertime,sothatthedatasetisanunbalancedpanel.254.Estimationmethodsandresults4.1.Pre-disasterlandsharesIdentificationofthefire’simpactrequiresthatotherwisesimilarblocksweretreateddifferentlybythefire.IflandshareswereFig.2.Mapsheetsfromthe1905SanbornMap,Volume5.Note:Thefigureshowsatotaloffourcityblocks(eightpolygons)intheMissionDistrictneighborhood.Source:Sanborn(1905).20Seethedataappendixfordetailsonthedatacollectionandconstructionprocess.21Streetsandalleysarenotpartofablock,andsowereexcludedinconstructingthedata.22Inotherwords,ifthemajorityofablock’slandwasburned,itwasassignedtothetreatmentgroup.Ifthemajorityofablock’slandwasunburned,itwasassignedtothecontrolgroup.Relativelyfewblockswerepartiallyburnedasstreetsthatseparatedblocksoftenactedashelpfulfirebreaks.23Fig.1showsseveralblockswithintheburnedareathatweresparedfromtheflames.Theseblockscontainedfederalbuildings(suchasthePostOfficeandtheU.S.Mint)andwerethussparedfromtheflamesthroughconcentratedfirefightingefforts.Theyarenotincludedinthesample.Additionally,someblocksintheNorthBeachareawerepartofthebayatthetime,andsothesamplestopsshortofthewaterfront(seeFig.3).24SomeblocksarecomprisedofmultiplepolygonsbutconsideredasingleunitifsoshownintheSanbornmaps.25Thedatasetisonlyslightlyunbalanced:theaveragenumberofobservations(i.e.,timeperiods)persampleblockis3.7outofamaximumof4.J.SiodlaExplorations in Economic History 65 (2017) 1–166
similarinburnedandunburnedareasbeforethefire,andfollowedsimilartrendsupto1906,thentheseareaswerelikelysimilarinother(unobservable)ways.Undertheassumptionthatthedevelopmenttrendswouldhavecontinuedintheabsenceofthefire,itseffectcanbeidentified.Ontheotherhand,dissimilarsharesandtrendsbeforethefirewouldsuggestthatblocksdirectlyacrossfromoneanotherweresomehowdifferent,potentiallyleadingtobiasedresults.Toestablishsimilarity,Table1comparesmeanresidentialandnonresidentiallandsharesacrossburnedandunburnedareasintheyearsbeforethefire.Column1showstheaveragesharesfortheentiresample,whilecolumns2and3comparesharesacrossburnedandunburnedareas.Column4showsthetestforpre-disasterdifferencesinlandsharesinboth1900and1905,whilecolumn5showsthetestfordifferencesintrendsacrossburnedandunburnedareasbetween1900and1905.Standarderrorsaregiveninparentheses.Overall,blockswerehighlyresidentialbeforethedisaster.Inboth1900and1905,theburnedareashowshigherresidentiallandsharesandlowernonresidentialsharesthantheunburnedareaonaverage.However,thesedifferencesarenotstatisticallysignificant,asshownincolumn4,andsoburnedandunburnedareasnearthefire’sboundarysharedsimilardevelopmentoutcomesbeforethedisaster.Itisalsoevidentthatbetween1900and1905,asthecitywasgrowing,nonresidentialuseswereontherisewhilehousingdevelopmentremainedrelativelystagnant.Althoughsubtle,thischangesuggestsagrowingdemandfornonresidentiallandrelativetoresidentiallandintheheartofthecityleadingupto1906.Asidefrombeingsimilarlydevelopedin1900and1905,burnedandunburnedareasalsoexperiencedsimilarchangesinlandsharesbeforethe1906disruption.Asshownincolumn5,thereisnosignificantdifferenceintrendsacrossburnedandunburnedareasforeitherresidentialornonresidentialsharesbetween1900and1905.Thisresult,combinedwiththeoutcomethatburnedandunburnedareasalsosharedsimilarlandsharesbeforethedisaster,suggeststhatanysignificantdeviationinland-useoutcomesonburnedblocksafter1906isduetothefireitself.Fig.3.Cityblocksnearthefire’sboundary.Note:Polygonsoftenrepresentindividualcityblocks,althoughsomeblockscontainmultiplepolygons.Source:Seetext.J.SiodlaExplorations in Economic History 65 (2017) 1–167
4.2.Thefire’simpactonlanduseFig.4displaysovertimetheaverageresidential(4a)andnonresidential(4b)landsharesforblocksatthefire’sboundary.Thereisarelativeshiftawayfromresidentiallandandintononresidentiallandinburnedareasafter1906,andthisshiftisstillevidentin2011.Whiletheserawdatasuggestalargeeffectofthefire,thereareotherpotentialsourcesofvariationtoconsider.Inthefollowingmodeloflanduse,letYitdenotetheland-useoutcomeonblockiattimet:YαγdBurneddδdβθε=++(×)′+(X×)′++,ittititiit(1)wheredtaretimedummies,Burnediisanindicatorforblocksburnedinthefire,Xiaretime-invariantcontrols,θiareblockfixedeffects,andεitisanerrorterm.Sinceθiabsorbsalltime-invarianteffects,BurnediandthecontrolsinXiareinteractedwithdttoestimateyear-specificeffectsonlanduse.Eventhough1905isthepre-disasteryearclosestto1906inthedata,coefficientsarereportedrelativeto1900(i.e.,1900istheomittedyear)sincethedatafrom1905maybesubjecttoasmallamountofmeasurementerror.26Also,sinceblockoutcomesarelikelytobespatiallycorrelated,standarderrorsarereportedthataccountforbothheteroskedasticityandcontemporaneousspatialcorrelation(Conley,1999).27ThemaincontrolsinXiincludedistancetodowntown(inquadraticform),neighborhooddummies,geologicalandtopographicalcharacteristics,andavariablerepresentinginitialconditions.Distancetodowntown(i.e.,thecitycenter),animportantpredictoroflandvaluationacrossresidentialandnonresidentialuses,ismeasuredinastraightlinetothemostprominentdowntownproperty,thePhelanBuilding.28Landvalues,andhencelanduses,changeovertimeasthecityevolvesandsoyear-specificdistanceeffectsareimportantcontrols.29Landusecanalsovarybyneighborhood,andsoincludingneighborhood-yearinteractionssweepsoutbetween-neighborhoodvariationovertime.Neighborhoodboundariesarebasedonhistoricaldefinitions—whicharestilllargelythesametoday—providedinIsselandCherny(1986)andDavies(2012)andaredepictedinFig.3.Theseneighborhoodsalsoactasaproxyforsocioeconomiccharacteristics,sincetheywere(andare)oftendefinedbycertainworkingclassesandethnicgroups.Geologicalandtopographicalcharacteristics,suchaselevationrangeandlandquality,alsolikelyhaveyear-specificeffectsonlanduseasbuildingtechnologiesimproveovertime.Ablock’selevationrangeismeasuredinmeters,sothat,allelseequal,largernumbersrepresentsteeperterrainandhencemoredifficultaccessandbuildingconditions.30LandqualityisproxiedbyliquefactionTable1Pre-disasterlanduse.DifferenceDifference:inchanges:AllBurnedUnburned(2)–(3)1900to1905(1)(2)(3)(4)(5)PanelA:Residentialshare19000.5100.5500.4730.078(0.016)(0.019)(0.025)(0.049)19050.5040.5350.4770.058−0.017(0.016)(0.019)(0.025)(0.054)(0.026)PanelB:Nonresidentialshare19000.3470.3160.376−0.060(0.017)(0.020)(0.027)(0.053)19050.3690.3390.396−0.057−0.006(0.017)(0.021)(0.027)(0.058)(0.023)Notes:Residentialshareistheaverageproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtoresidentialuse.Nonresidentialshareistheaverageproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtononresidentialuse.Column1reportssummarystatisticsforallblocksinthesample,whilecolumns2and3reportaveragelandsharesacrossburnedandunburnedareas.Column4reportstheestimateddifferenceinlandsharesacrossburnedandunburnedareasin1900and1905.Column5reportstheestimateddifferenceinchangesinlandsharesbetween1900and1905acrossburnedandunburnedareas.Includedintheanalysisareonlythoseblocksforwhichthemajorityofacreageisdeveloped.Thereare378oftheseblocksin1900,ofwhich183areburnedand195areunburned,andthereare392blocksin1905,ofwhich185areburnedand207areunburned.Standarderrorsarereportedinparentheses,andarecorrectedforspatialcorrelationaccordingtoConley(1999)intheregressionresultsshownincolumns4and5.Theregressionsincolumn4includeaconstantandanindicatorforburned.Theregressionresultsincolumn5arereportedfromcolumn1inTable3.26ThisissueisdescribedinSection3.27Inthismethod,spatialweightsareassignedthatdecaylinearlywithinaquarter-mileradiusaroundeachblock.Thisdistancecut-offcapturesimmediatelyneighboringblocksinneighborhoodswithlargeblockareas(e.g.,SouthofMarket)andagreaternumberofsurroundingblocksinneighborhoodswithsmallerblockareas(e.g.,WesternAddition).StandarderrorswerecorrectedusingStatacodedevelopedinHsiang(2010).28Althoughtraveldistanceisabettermeasure,usingstraight-linedistanceshouldproducesimilarresults.Themainobjectiveistoaccountforthefactthatunburnedareasarefartherfromdowntown,whichstraight-linedistanceaccomplishes.29AtackandMargo(1998)estimatechangesinlandvaluesacrossdecadesinNewYorkCitywhilecontrollingfordistancetothecity’scenter.30SanFranciscoisknownforitshills.Aconcernthenistheextenttowhichcontouredorterracedstreetsimpactlanduseonblocksthatfrontsuchstreets.However,streetswerenotterracedinSanFranciscountilwellafterreconstructionfromthefire.ThequintessentialexampleisLombardStreet,whichwasnotbuiltwithitshairpinturnsuntil1922.AnotherexampleisTelegraphHill,whereCoitTowernowsits.StreetsnearthislandmarkwerenotredesigneduntilsometimeaftertheGreatDepression.Untilthen,streetsonhillsweresubjecttosteepangles,andwereconsideredimpassibleforhorseteams,astheSanbornmapsclearlyportray.Asitcapturestheextentofablock’shilliness,elevationrangeisagoodproxyfortheneedforterracedstreets,ortheimpassibilityofhorseteams,whichcouldhaveinfluencedlandusedecisionsforcertainblocks.J.SiodlaExplorations in Economic History 65 (2017) 1–168
riskandlandsliderisk,whicharemeasuredastheproportionofablock’sacreagethatissubjecttoeachrisk.Althoughrepresentingthelevelofrisktoday,thesevariablesarebasedonhistoricalliquefactionandlandslideevents.Theyarehighlycorrelatedwiththeunderlyingsoil,whichisstableovertime,andrepresenttheabilitytoallowforcertainlandusesandbuildingdensities.EachoftheseserieswasobtainedfromSanFrancisco’sdatawebsite.31Itisalsoimportanttoaccountforinitialconditionswhichcouldsetblocksoffondifferentdevelopmentpathsthroughtime.Forinstance,blockswithmorevacantlandin1900beganwithagreatercapacityfordevelopmentsincemakingchangesdidnotnecessarilymeantearingdownbuildings.Thisstudyaccountsfortheseinitialvacantlandconditionsintwoways.First,sincethefocusisonwell-developedland(asdescribedinSection3),blocksthatwerepredominantlyvacantinanygivenyearareexcludedfromtheanalysis.Second,initialvacantlandshare(in1900)isincludedinXi,thussweepingoutanyyear-specificvariationinresidentialandnonresidentiallandsharesthatisduetodifferentstartingpointsintheamountofvacantland.SummarydataforthecontrolsareshowninTable2,withstandarddeviationsgiveninparentheses.Column1reportssummarystatisticsforallblocksinthesample,whilecolumns2and3separatethedataacrossburnedandunburnedareas.Onaverage,burnedandunburnedareassharesimilarelevationranges,liquefactionrisks,andlandsliderisks,aswellassimilarrepresentationintheSouthofMarketneighborhood.However,themeansofothervariablesdiffermoresubstantiallyacrossgroups,thushighlightingtheirusefulnessascontrols.4.2.1.Landuseandthefire,1900–1931Table3presentstheresultsofestimatingEq.(1)forlandsharesinresidential(PanelA)andnonresidential(PanelB)uses.1906 Fire0.2.4.6.81Residential land share190019201940196019802000YearUnburnedBurned(a)Residential1906 Fire0.2.4.6.81Nonresidential land share190019201940196019802000YearUnburnedBurned(b)NonresidentialFig.4.Averagelandshares,1900–2011.Notes:Residentiallandshareistheaverageproportionoftotalblockacreage(netofstreetsandalleys)thatisresidential.Nonresidentiallandshareistheaverageproportionoftotalblockacreage(netofstreetsandalleys)thatisnonresidential.Source:Seetext.31Thewebsiteishttps://data.sfgov.org.J.SiodlaExplorations in Economic History 65 (2017) 1–169
Averagetreatmenteffectsareestimatedin1914and1931,with1905actingasaplaceboyearinwhichthereshouldbenoeffectofthefire.Thebestyeartofocusonis1931,bywhichtimetheburnedareahadbeencompletelyredeveloped.32Firstconsiderresidentialland.Forallspecifications,burnedandunburnedareasexperiencedsimilartrendsindevelopmentintheyearsleadinguptothefire,asshownbythenulleffectfor1905.Thebasicspecificationincolumn1ofPanelAshowsalargerelativedeclineinresidentiallandsharein1914.By1931,blocksinburnedareassawarelativedeclineof16percentagepointsinresidentiallandsharecomparedtounburnedareas.Theseeffectsshrink,butremainstatisticallysignificantatthe1percentlevel,whenaddingmorecontrols.Intherichestspecificationincolumn5,therelativedeclineinresidentiallandsharewasnearly9percentagepointsby1931.Thiswasasubstantialreductioninacreage,astheaverageburnedblockinthesamplewasnearlyfouracresinsize.Theoppositeeffectoccursfornonresidentialland.Thebasicspecificationincolumn1ofPanelBshowsasignificantrelativeincreaseinnonresidentiallandshareof6percentagepointsby1914and12percentagepointsby1931.Inotherspecifications,theresultsfor1914weakenorwashoutandthe1931effectdiminishes,butremainsstatisticallysignificant.Therichestspecificationincolumn5showsthat,by1931,nonresidentiallandshareroseby8percentagepointsinburnedareasrelativetounburnedareas.Thus,onaverage,nonresidentialacreageonburnedblocksrosebyroughlythesizeoftherelativelossinresidentialacreage.Afterthefire,developersinburnedareasshiftedlandoutofhousingandintononresidentialuses.33Table2Summarystatisticsforcontrols.AllBurnedUnburned(1)(2)(3)Distancetodowntown4.9524.5305.348(1.467)(1.355)(1.462)Elevationrange26.50325.57427.375(33.359)(32.128)(34.533)Landsliderisk0.0080.0050.010(0.058)(0.035)(0.074)Liquefactionrisk0.3640.3700.358(0.445)(0.444)(0.447)Vacantsharein19000.0930.0710.114(0.123)(0.097)(0.140)Neighborhoods:Downtown0.0610.1260(0.239)(0.332)(0)MissionDistrict0.1980.1690.226(0.399)(0.376)(0.419)NobHill0.0320.0660(0.176)(0.248)(0)NorthBeach0.0930.1370.051(0.290)(0.344)(0.221)PacificHeights0.0630.0110.113(0.244)(0.104)(0.317)RussianHill0.0820.1200.046(0.275)(0.326)(0.210)SouthofMarket0.2490.2510.246(0.433)(0.435)(0.432)WesternAddition0.2220.1200.318(0.416)(0.326)(0.467)Notes:Column1reportssummarystatisticsforallblocksinthesample,whilecolumns2and3reportsummarystatisticsacrossburnedandunburnedareas.ThelocationofdowntownisconsideredthePhelanBuildingonMarketStreet,themostprominentcentralpropertyatthetime.Distancetodowntownisgiveninquarter-mileunits.Elevationrangeisgiveninmeters.Landslideandliquefactionriskrepresenttheproportionofblockacreagethatissubjecttoeachtypeofrisk.Vacantsharein1900istheproportionoftotalblockacreage(netofstreetsandalleys)thatisvacantin1900.Standarddeviationsarereportedinparentheses.32SeeSection2.3forthediscussiononredevelopment.33AsimilarshiftfromresidentialtocommercialoccupantsoccurredinBoston’sbuildingsafterthe1872Fire(HornbeckandKeniston,2016).J.SiodlaExplorations in Economic History 65 (2017) 1–1610
ThereversalinlanduseisconsistentwiththeriseofrealestatevaluesandtheneedformorebusinessspaceinSanFranciscoattheturnofthetwentiethcentury.Thefiresweptawayoldbuildingsandmadewayfortheexpansionofbusinessactivities.Buthowexactlydidresidentiallandusechange?Diddevelopersmakeroomforotherusesbyalteringthecompositionofhousingstructures?Theanswerprovidesinsightintotheobstaclestochangebefore1906.Ofallhousingtypes—whichincludedsingle-familyhouses,flatbuildings,multi-familydwellings,andapartmentbuildings—housesweretheleasteconomicalstructurestobuildonvaluableland.34Andsincemanywereshantiesandshacksthatownerswerenotinterestedinredevelopingorrenovating,theywerebigimpedimentstoconvertinglandtootheruses.Thus,housingcompositionshouldhavechangedprimarilythroughalargerelativereductioninsingle-familyhouses.Table4showstheresultsoftestingthishypothesis.Firstconsidertheimpactontotalhousingunitsfortheaveragecityblock.Ascolumn1shows,by1931,thenumberofresidentialunitsinburnedareashadrecoveredrelativetounburnedareas.35Thus,thechangeinlandusewasnotduetoaccommodatingfewerresidentsorfamiliesinburnedareasrelativetounburnedareas.Thechangeinlandusewasinsteaddrivenbychangesinhousingcomposition.Column2showsthatrelativelyfewersingle-familydwellingswerebuiltontheaverageburnedblockuponreconstruction.Asdepictedincolumns3and4,flatbuildingsandmulti-familydwellings,whichweregenerallysimilarinsizetohouses,alsofellinburnedareasrelativetounburnedareas,butonlyslightly.Thenumberofapartmentbuildings,however,rose(shownincolumn5).Addingupcoefficientsincolumns2through5showsanetrelativereductionof12.2residentialbuildingsfortheaverageburnedblockin1931,andabout70percentofthisreductioncamethroughthedeclineinsingle-familyhouses.ThisresultsuggeststhathousesimpededtheprocessofrepurposinglandinSanFranciscobeforethefire.Nonresidentiallanduseswereunabletoexpanduntiltheshantiesandshacksweregone.4.2.2.TheshiftinlanduseacrossneighborhoodsAlthoughhailedasanequalizerofsorts,thedisasterleftintactbroadpatternsofresidence,work,andethnicity(IsselandCherny,1986,p.78).Nonetheless,thefireexpandedbusinessspaceandcreatedopportunitiesinnewareas.Toillustrate,Figs.5aand5bshowchangesinnonresidentiallandsharesthatoccurredonblocksattheboundaryofthefirebeforeandafter1906.Themapsshowblockswithpositivechanges,wherebiggercirclesrepresentlargerchanges.Asthefiguresdepict,theWesternAdditionneighborhoodexperiencedsomeofthelargestchangesfollowingthefire,eventhoughtherewasnogeneralindicationbefore1906thatthiswouldhappen(i.e.,landwasnotquicklybecomingmorenonresidentialleadinguptothefire).MostofthesechangesoccurredalongVanNessAvenue,wheremanyretailbusinessesrelocatedinwhatmanythoughtwouldbeatemporarymove(SanTable3Landuseandthefire.(1)(2)(3)(4)(5)PanelA:Residentialshare1905×burned−0.017−0.009−0.002−0.0000.004(0.026)(0.025)(0.021)(0.020)(0.019)1914×burned−0.225***−0.193***−0.202***−0.199***−0.187***(0.026)(0.024)(0.021)(0.020)(0.019)1931×burned−0.163***−0.108***−0.112***−0.105***−0.087***(0.033)(0.031)(0.025)(0.024)(0.024)R20.3790.4260.5650.5740.590PanelB:Nonresidentialshare1905×burned−0.006−0.011−0.019−0.020−0.017(0.023)(0.021)(0.022)(0.022)(0.022)1914×burned0.055**0.046**0.0300.0230.040*(0.023)(0.021)(0.024)(0.023)(0.024)1931×burned0.118***0.091***0.070**0.062**0.076***(0.029)(0.027)(0.029)(0.028)(0.029)R20.3260.3360.4160.4260.443Controls:BlockfixedeffectsXXXXXYearindicatorsXXXXXYear×distancetodowntownXXXXYear×neighborhoodXXXYear×geology/topographyXXYear×vacantsharein1900XNotes:Residentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtoresidentialuse.Nonresidentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtononresidentialuse.Allregressionsincludeaconstant.Thecoefficientsareinterpretedrelativeto1900,theomittedyear.Downtownistheomittedneighborhood.Thereare1,545block-yearobservations.StandarderrorsarereportedinparenthesesandarecorrectedforspatialcorrelationaccordingtoConley(1999).*,**,and***indicatestatisticalsignificanceatthe10%,5%,and1%levels.34Single-familyhousesarestand-aloneunits,flatbuildingsconsistofflatsthatsharefloorsorceilings,multi-familydwellingshaveunitsthatshareacommonwall,andapartmentbuildingsconsistonlyofapartments.35Siodla(2015)showsthatnetresidentialdensity,orresidentialunitsperresidentialacre,roseinburnedareasafterthefire.Thisoutcomeisconsistentwiththeresultthatthenumberofhousingunitsremainedconstantwhileresidentialacreagefell.J.SiodlaExplorations in Economic History 65 (2017) 1–1611
FranciscoCall,7May1906,p.1;SanFranciscoRealEstateCircular,December1906,p.4).Butovertimethestreetbecamearetaildistrictuntoitself,havingbeenradicallychangedbythefire(Grant,1909,p.365).ButdidVanNessandotherareassimplystealeconomicactivityfromotherstreetsandneighborhoods?Theevidencesuggestsnot.Whilecertainareasattractedbusinessesfromelsewhereinthewakeofthefire,mostbusinesseseventuallymovedbacktotheirpreviouslocations.Almosttwoyearsafterthefire,manynewbuildingsinretailportionsofDowntownandSouthofMarketwerenearlycompleted(SanFranciscoRealEstateCircular,December1907,p.2).Bytheendof1908,realestateactivityinWesternAdditionhadslowedduetothereturnofretailbusinessfromVanNesstoDowntown(SanFranciscoRealEstateCircular,December1908,p.2).In1909,itwassuggestedthat“Theyearjustclosedhaswitnessedthereturntothedown-townretaildistrictofallofourretailmerchants”(SanFranciscoRealEstateCircular,December1909,p.2).Wholesalemerchantsalsolargelyreturnedtotheiroldlocationswithinafewshortyearsofthefire(Grant,1909,p.365).Ifeconomicactivitywasnotstolenfromotherareas,thentheexpansionofnonresidentiallandusemusthavebeenduetothecreationofnewbusinessopportunities.Theconstructionofnewbuildingslikelygeneratedpositivespillovers,whichwouldhavedrivenoutinferiorlandusesinareasdesirableforcommercialandindustrialuses.Inastudyofpost-disasterbusinesslocations,oneTable4Housingcompositionandthefire.Dependentvariable:HousingSingle-familyFlatMulti-familyApartmentunitshousesbuildingsdwellingsbuildings(1)(2)(3)(4)(5)1905×burned5.087−1.227*1.949*−0.060−0.003(4.554)(0.734)(1.020)(0.287)(0.191)1914×burned−18.402***−8.704***−4.462***−2.453***0.448***(4.115)(0.741)(0.956)(0.272)(0.162)1931×burned−4.591−8.552***−2.933***−2.172***1.470***(6.031)(0.901)(1.119)(0.265)(0.348)R20.4150.5300.3950.4550.501Controls:BlockfixedeffectsXXXXXYearindicatorsXXXXXYear×distancetodowntownXXXXXYear×neighborhoodXXXXXYear×geology/topographyXXXXXYear×vacantsharein1900XXXXXNotes:Housingunitsarethetotalnumberofresidentialunitsinthefourprimarybuildingtypeslistedincolumns2through5.Single-familyhousesarestand-aloneunits,flatbuildingsconsistofflatsthatsharefloorsorceilings,multi-familydwellingshaveunitsthatshareacommonwall,andapartmentbuildingsconsistonlyofapartments.Allregressionsincludeaconstant.Thecoefficientsareinterpretedrelativeto1900,theomittedyear.Downtownistheomittedneighborhood.Thereare1,545block-yearobservations.StandarderrorsarereportedinparenthesesandarecorrectedforspatialcorrelationaccordingtoConley(1999).*,**,and***indicatestatisticalsignificanceatthe10%,5%,and1%levels.Fig.5.Nonresidentialland-sharechanges.Notes:Nonresidentiallandshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisnonresidential.Thefigureshowsblocksthatexperiencedpositivechangesinnonresidentiallandsharesbetween1900and1905,1900and1931,and1900and2011.Largercirclesrepresentlargerchangesinthesedirections.Blockswitheithernegativechangesornochangesarenotshown.Source:Seetext.J.SiodlaExplorations in Economic History 65 (2017) 1–1612
observernotedthatbuildingspacehadexpandedasstructuresbecamemuchtaller(Grant,1909,p.366).Inaddition,firmsinSanFranciscoincreasedinbothsizeandnumber,andthreedistinctretaildistrictshademergedwhereasbefore1906therewasone(Grant,1909,p.366).Thus,ifthefirecreatednewbusinessopportunities,thenitwouldhavehadalargerimpactinunexpectedareas.Inotherwords,burnedblocksinneighborhoodsthatwerenotquicklymovingtowardnonresidentiallandusebeforethefirewouldhaveexperiencedlargerland-usechangesafterward.Considerthefollowingvariationofthepreviousmodeloflanduse.Again,letYitdenotetheland-useoutcomeonblockiattimet:YαγdBurneddδTrenddρBurnedTrenddϕdβθε=++(×)′+(×)′+(××)′+(X×)′++,ittitntintitiit(2)whereallvariablesareasdefinedinEq.(1)exceptforTrendn,whichrepresentsneighborhoodn’spre-firegrowthrateinnonresidentiallandshare(inpercentageterms)between1900and1905.36Also,XidoesnotincludeneighborhooddummiessincetheyarecollinearwithTrendn.Inthismodel,thefire’seffectisnowdependentonthepre-fireneighborhoodgrowthtrendinnonresidentiallanduse.Asbefore,forallregressions,1900istheomittedyearandstandarderrorsarecorrectedforspatialcorrelation(Conley,1999).Table5presentstheresultsoftestingwhetherneighborhoodpre-firetrendstowardnonresidentiallandhelpexplainpost-firechangesinlanduse.Thespecificationsshownincolumns1and2showtheresultsforresidentiallandshareandcolumns3and4showresultsfornonresidentiallandshare.Thefirsttwocolumnsrevealthatburnedblocksinareastrendingtowardnonresidentiallandusebeforethefiresawsmallerchanges(i.e.,lessnegativechanges)inresidentiallanduseafterward.Thissmallerchangealsoholdsfornonresidentiallanduse.Thus,burnedblocksinareaswithlittlegrowthinnonresidentiallandusebeforethedisastersawlargershiftsafter1906.Theseresultssuggestthatthefirehelpedcreatebusinessopportunitiesinnewareas,evenwhileexpandingnonresidentiallanduseoverall.Wasthisoutcomeasurprise?Notnecessarily.Withpropertydowntowninhighdemandandlowsupplybefore1906,investorsinbusinesspropertieslookedtootherpartsofthecitysuchasVanNess(SanFranciscoRealEstateCircular,December1903,p.2).InDecember1904,“Itwasonlyaquestionoftimewhenbusinessmustarrivethere”(SanFranciscoRealEstateCircular,p.2).Thattimecamein1906.Afterthefire,economicactivitydidnotre-sortwithoutaguidingforce,butrathermovedtolocationsthatwerehighlydesirable.4.2.3.Thelong-runimpactofthefireBythetimeSanFranciscowasrebuilt,thedisasterhadleftalargeimprintonthecity.Isitstillevidenttoday?WhiletherawdatashowninFigs.4and5csuggestso,theeconometricspecificationsinTable6testwhetherthefirehashadapersistenteffectonlandusewhilecontrollingforothersourcesofvariation.Thefocusisontheimpactofthefireto2011,andforsimplicity,observationsfromtheyears1905,1914,and1931areexcludedintheregressions.Theresultsforresidentialandnonresidentiallandsharesareshownincolumns1and2.Thefirehashadalastingimpactbypermanentlyshiftinglanduseinburnedareasawayfromresidentialuseandtowardnonresidentialuse.Additionally,ascolumn3shows,burnedareasstillfeaturerelativelyfewersingle-familyhouses.Theseeffectsareallsimilarinmagnitudetothoseseenin1931.Additionally,theneighborhoodsthatweremostimpactedthenremainsotoday.Columns4and5showthatblocksinneighborhoodswithlowgrowthinnonresidentiallandusebeforethefirestilldepictbiggershiftsinlandusetoday.Thus,thechangesmadeinthewakeofthefirehavelastedoveracentury,andhavepersistedinthesameplaces.37Neweconomicopportunitiescamefromthecleanslatecreatedbythefire,whichhelpedformbusinessclustersthathavegeneratedpositivespilloversfordecades.Thefire’spersistenteffectonlandusesuggestsanumberofthings.First,therewereimmensebenefitsthatcamefromprovidingeconomicopportunitiesinnewareas.Ifthiswerenotso,100yearswouldsurelyhavebeensufficienttimetorearrangeeconomicactivityandrepurposelandforotheruses.Second,thefrictionstochangeinacityaresubstantial:developersinburnedareaspermanentlyshiftedoutofbuildingsingle-familyhousesafter1906.SanFranciscowouldhavedevelopedmuchdifferentlyhaditnotbeenforthecleansingeffectofthefire.Finally,largeshockshavethecapacitytodrasticallyreshapeurbansettings.InthecaseofSanFrancisco,economicactivitywasdisruptedbyfire,butwasultimatelyrearrangedtoformnewbusinessclustersthathaveshapedthecitysincebeingformedshortlyafter1906.WhilethefirehascastalongshadowonSanFrancisco’sland-usepatterns,itisambitioustoputalltheweightoflong-runcausalityonthisfatefulevent.Regulationssuchaszoning,whichonlybecameasignificantconstraintaftertheGreatDepression,andthemanyspecialusedistrictsandhistoricpreservationzonesestablishedinrecentyears,havecertainlyplayedaroleindetermininglanduseinthecityoverthepastcentury.38Nevertheless,disastersthatgreatlydisruptacity’sdevelopmentpatternsareenormouslyinfluentialharbingersofchange.5.ConclusionSanFranciscowasthrivingwhen,in1906,amassiveearthquakesparkedafirethatprovidedacleanslateinmuchofthecity.Landusewasalteredasbuildingsrosefromtheashes.Developersrebuiltburnedareasusinglesslandforhousingandmorelandfor36Theresultsdonotchangeifthetrendvariableisinsteadcalculatedasachangeinlevelsornormalizedbyresidentialornonresidentiallandsharein1905.37Theseresultsholdwhenusingtoday’sneighborhoodboundariesgiveninthemodernlandusedatabase(SFLUD,2011).38InChicago,the1923zoningordinancehashadpersistenteffectsonthecity’slanduse(Shertzeretal.,2016).J.SiodlaExplorations in Economic History 65 (2017) 1–1613
nonresidentialactivities,primarilybybuildingrelativelyfewersingle-familyhouses,thussuggestingthathousesimpededchangesinlandusebefore1906.Comparingthesechangeswiththedevelopmenttrendsinunburnedareassuggeststhattheland-useshiftwouldnothaveoccurredintheabsenceofthefire;theblazewasindeedacatalystforchangeinSanFrancisco,freeingthecityfromthefettersofoldbuildingsandthefrictionstheyrepresent.Thesignificantchangesrealizeduponreconstructionthusproveimportanttheinfluenceofsuchfrictionsonurbansettings.Table5Theland-useshiftacrossneighborhoods.Dependentvariable:ResidentialshareNonresidentialshare(1)(2)(3)(4)1905×burned−0.042−0.024−0.004−0.009(0.036)(0.028)(0.032)(0.027)1914×burned−0.352***−0.278***0.123***0.115***(0.035)(0.027)(0.032)(0.028)1931×burned−0.267***−0.138***0.204***0.158***(0.052)(0.037)(0.041)(0.036)1905×burned×pre-firetrend0.0040.003−0.001−0.001(0.003)(0.002)(0.003)(0.002)1914×burned×pre-firetrend0.016***0.012***−0.011***−0.009***(0.003)(0.002)(0.003)(0.002)1931×burned×pre-firetrend0.015***0.008***−0.013***−0.009***(0.004)(0.003)(0.004)(0.003)R20.4130.5380.3530.420Controls:BlockfixedeffectsXXXXYearindicatorsXXXXYear×distancetodowntownXXYear×geology/topographyXXYear×vacantsharein1900XXYear×pre-firetrendXXXXNotes:Residentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtoresidentialuse.Nonresidentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtononresidentialuse.Thepre-firetrendvariablereferstothegrowthrate(inpercentageterms)innonresidentiallandsharesforeachofthestudy’seightneighborhoodsbetween1900and1905.Allregressionsincludeaconstant.Thecoefficientsareinterpretedrelativeto1900,theomittedyear.Thereare1,545block-yearobservations.StandarderrorsarereportedinparenthesesandarecorrectedforspatialcorrelationaccordingtoConley(1999).*,**,and***indicatestatisticalsignificanceatthe10%,5%,and1%levels.Table6Thefire’slong-runimpact.Dependentvariable:ResidentialNonresidentialSingle-familyResidentialNonresidentialsharesharehousesshareshare(1)(2)(3)(4)(5)2011×burned−0.116***0.099**−8.996***−0.206***0.205***(0.044)(0.049)(1.031)(0.049)(0.052)2011×burned×pre-firetrend0.013***−0.012***(0.003)(0.003)R20.3740.2460.6430.3250.197Controls:BlockfixedeffectsXXXXX2011indicatorXXXXX2011×distancetodowntownXXXXX2011×neighborhoodXXX2011×geology/topographyXXXXX2011×vacantsharein1900XXXXX2011×pre-firetrendXXNotes:Residentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtoresidentialuse.Nonresidentialshareistheproportionoftotalblockacreage(netofstreetsandalleys)thatisdevotedtononresidentialuse.Single-familyhousesarestand-aloneunits.Thepre-firetrendvariablereferstothegrowthrate(inpercentageterms)innonresidentiallandsharesforeachofthestudy’seightneighborhoodsbetween1900and1905.Allregressionsincludeaconstant.Thecoefficientsareinterpretedrelativeto1900,theomittedyear.Downtownistheomittedneighborhoodincolumns1through3.Thereare779block-yearobservations.StandarderrorsarereportedinparenthesesandarecorrectedforspatialcorrelationaccordingtoConley(1999).*,**,and***indicatestatisticalsignificanceatthe10%,5%,and1%levels.J.SiodlaExplorations in Economic History 65 (2017) 1–1614
Whilethefirebroadlyshiftedlanduseinthecity,italsocreatedeconomicopportunitiesinnewareas:itsimpactwasgreatestinneighborhoodsthatexperiencedlessgrowthinnonresidentiallandusebefore1906.Furthermore,thefire’sinitialimpactonlandusestillremainstoday.Theeconomicbenefitsthatformeduponreconstructionhavesurvivedoveracentury.Largeshocksthatdestroyphysicalcapitalcandramaticallyreshapecities,andonceareaschange,theydonoteasilygoback.Thepastthusholdssignificantswayincities,suggestingacentralroleforhistoryinexplainingtheirdevelopmentovertime.AcknowledgementsIamgratefultoDanBogartandJanBruecknerfortheirguidance.Forvaluablecomments,Ithanktwoanonymousreferees.IthankKrisTacsikandStephenKallemeynfromtheGeographyMapLibraryatCaliforniaStateUniversity,NorthridgeforaccommodatingmeasIphotographedthephysicalmapdocumentsusedinthestudy.IalsothankTonySoelleratUCIrvineandMannyGimondatColbyCollegefortheirvaluableGISassistance,aswellasAkselOlsenfromtheSanFranciscoPlanningDepartmentforprovidingmodernland-usedata.MostofthedatainthestudywerecollectedwiththebenefitofresearchsupportfromtheAll-UCGroupinEconomicHistory,theDepartmentofEconomicsatUCIrvine,andtheInstituteforHumaneStudies.DataAppendixAccesstodigitalSanbornmaps(1899/1900and1913/1914)wasobtainedthroughtheLosAngelesPublicLibrary,whichsubscribestoDigitalSanbornMaps,1867–1970,theonlinedigitaldatabasecreatedbyProQuest,LLC(Sanborn(1899-1914)).Physicalmapsfor1928/1931weredigitallyphotographedintheGeographyMapLibraryatCaliforniaStateUniversity,Northridge(Sanborn(1899-1914)).The1905mapsweredownloadedfromDavidRumsey’swebsite(Sanborn,1905).Themapsforaparticularsurveyyearhaveseverallargevolumes,eachofwhichcontainsmanysheetsthatcoverseveralcityblockseach.Therearesixvolumesinthe1899/1900editionandtenvolumesinthe1913/1914edition.Themapsfrom1905arecorrectedversionsofthe1899/1900volumesandthe1928/1931mapsarecorrectedversionsofthe1913/1914volumes.Thesheetsprovideadetailedaccountofthebuildingsthatexistedoneachcityblockatthetimeofthesurvey.Cityblocksandbuildingfootprintsaredrawnatascaleoffiftyfeettooneinch.Asidefromfootprints(whichmustbephysicallymeasuredonthemapstoobtaintheirperimeter),detailsatthebuildinglevelincludeconstructioninformationandtypeofuse.ThehistoricallandsharemeasuresusedinthestudywerecalculatedfromtheinformationgivenintheSanbornmaps.Residentiallandiscomposedoflotsoccupiedbyresidentialbuildings—single-familyhouses,flatbuildings,multi-dwellingdwellings,andapartmentbuildings—aswellastheopenspaceandoutbuildingsthattogethercomprisetheacreagedevotedtoresidentialuse.39Nonresidentiallandincludesonlylotsoccupiedbynonresidentialbuildings,suchascommercial,industrial,andofficebuildings,andanyopenspace“attached”tothesebuildings.Ifaparticularlothadaresidentialbuildingandaseparatenonresidentialbuildingdevotedtocommercialorindustrialuse,thelandacreagewassplitaccordingtoeachbuilding’sproportionalcoverageonthelot.Theproportionofablock’slanddevotedtoresidentialandnonresidentiallanduse(andvacantland)wascalculatedusingacomputerprogram—calledScreenTracingPaper—thatmeasuresshapesonacomputerscreen.Engineersanddesignerswhoneedtocalculatemeasurementsfromdigitalblueprintsoftenusesuchprograms.Datafor2011wereobtainedfromtheSanFranciscoPlanningDepartment(SFLUD,2011).Thedatacontaininformationonthetypeoflanduseforeachparcelinthecity.Landuseisgiveninseveralcategories,includingresidentialandnonresidential.Thedefinitionsusedbytheplanningdepartmentareverysimilartothoseusedtocalculatethehistoricaldata,andthusthereismuchsymmetryacrossdatasets.Parcelsontherelativelyfewblocksthatchangedsizeandshapeoverthelastcenturywereassignedtothehistoricalblockonwhichtheyarepredominantlylocated.GeographicInformationSystems(GIS)facilitatedthecalculationofstraight-linedistancestovariouspointsofinterest,thedeterminationofgeologicalandtopographicalvalues,andtheassignmentofvariousspatialreferences(i.e.,neighborhoods,burnedareas,etc.).Thelayoutofthecity’sblocks,aswellasthegeologicalandtopographicaldata(elevationrange,liquefactionrisk,andlandsliderisk),wereobtainedasshapefilesfromthecity’sdatawebsite(https://data.sfgov.org/).AppendixA.SupplementarydataSupplementarydataassociatedwiththisarticlecanbefoundintheonlineversionathttp://dx.doi.org/10.1016/j.eeh.2017.04.001.ReferencesAhlfeldt,G.M.,Redding,S.J.,Sturm,D.M.,Wolf,N.,2015.Theeconomicsofdensity:evidencefromtheBerlinWall.Econometrica83(6),2127–2189.Atack,J.,Margo,R.A.,1998.Location,location,location!thepricegradientforvacanturbanland:NewYork,1835to1900.J.RealEstateFinanceEcon.16,151–172.Barrows,R.G.,1983.Beyondthetenement:patternsofAmericanurbanhousing,1870–1930.J.UrbanHist.9(4),395–420.Berger,T.,Enflo,K.,2017.Locomotivesoflocalgrowth:theshort-andlong-termimpactofrailroadsinSweden,J.UrbanEcon.98,124–138.39Inlateryears,residentialbuildingsoftenhadautogaragesonfirstfloorsorbasements.However,ifanautogaragewaslocatedonalotbyitselfwithitsownuniqueaddress,itwasconsideredanonresidentialbuildingasitlikelywasaplaceofbusiness.J.SiodlaExplorations in Economic History 65 (2017) 1–1615
Bleakley,H.,Lin,J.,2012.Portageandpathdependence.Q.J.Econ.127(2),587–644.Bosker,M.,Brakman,S.,Garretsen,H.,Schramm,M.,2007.Lookingformultipleequilibriawhengeographymatters:GermancitygrowthandtheWWIIshock.J.UrbanEcon.61(1),152–169.Bosker,M.,Brakman,S.,Garretsen,H.,Schramm,M.,2008.Acenturyofshocks:theevolutionoftheGermancitysizedistribution1925–1999.Reg.Sci.UrbanEcon.38(4),330–347.Brakman,S.,Garretsen,H.,Schramm,M.,2004.ThestrategicbombingofGermancitiesduringWorldWarIIanditsimpactoncitygrowth.J.Econ.Geogr.4(2),201–218.Brooks,L.,Lutz,B.,2014.Vestigesoftransit:urbanpersistenceatamicroscale.Workingpaper.Brooks,L.,Lutz,B.,2016.Fromtoday’scitytotomorrow’scity:anempiricalinvestigationofurbanlandassembly.AmericanEconomic.J.:Econ.Policy8(3),69–105.Brueckner,J.K.,Rosenthal,S.S.,2009.Gentrificationandneighborhoodhousingcycles:willAmerica’sfuturedowntownsberich?Rev.Econ.Stat.91(4),725–743.Conley,T.G.,1999.Gmmestimationwithcrosssectionaldependence.J.Econom.92(1),1–45.Davies,A.R.,2012.SavingSanFrancisco:ReliefandRecoveryafterthe1906Disaster.TempleUniversityPress,Philadelphia.Davis,D.R.,Weinstein,D.E.,2002.Bones,bombs,andbreakpoints:thegeographyofeconomicactivity.Am.Econ.Rev.92(5),1269–1289.Davis,D.R.,Weinstein,D.E.,2008.Asearchformultipleequilibriainurbanindustrialstructure.J.Reg.Sci.48(1),29–65.DiPasquale,D.,Wheaton,W.C.,1996.UrbanEconomicsandRealEstateMarkets.Prentice-Hall,Inc.Douty,C.M.,1977.TheEconomicsofLocalizedDisasters:The1906SanFranciscoCatastrophe.ArnoPress.Fales,R.L.,Moses,L.N.,1972.Land-usetheoryandthespatialstructureofthenineteenth-centurycity.Pap.Reg.Sci.28(1),49–82.Field,A.J.,1992.UncontrolledlanddevelopmentandthedurationofthedepressionintheUnitedStates.J.Econ.Hist.52,785–805.Fradkin,P.L.,2005.TheGreatEarthquakeandFirestormsof1906.UniversityofCaliforniaPress.Glaeser,E.L.,Gyourko,J.,2005.Urbandeclineanddurablehousing.J.Polit.Econ.113(2),345–375.Godfrey,B.J.,1997.UrbandevelopmentandredevelopmentinSanFrancisco.Geogr.Rev.87(3),309–333.Grant,C.,1909.SanFranciscomovesdowntown.OverlandMonthly53(5),365–377.Haas,J.E.,Kates,R.W.,Bowden,M.J.,eds.,1977.ReconstructionFollowingDisaster.Cambridge,MA:MITPress.Hornbeck,R.,Keniston,D.,2016.Creativedestruction:barrierstourbangrowthandthegreatBostonfireof1872.AmericanEconomicReview(forthcoming).Hsiang,S.M.,2010.TemperaturesandcyclonesstronglyassociatedwitheconomicproductionintheCaribbeanandCentralAmerica.Proc.Natl.Acad.Sci.USA107(35),15367–15372.Imaizumi,A.,Ito,K.,Okazaki,T.,2016.Impactofnaturaldisastersonindustrialagglomeration:thecaseofthegreatKantōearthquakein1923.Explor.Econ.Hist.60,52–68.Issel,W.,Cherny,R.W.,1986.Politics,Power,andUrbanDevelopment,1865–1932.UniversityofCaliforniaPress,SanFrancisco.Jedwab,R.,Kerby,E.,Moradi,A.,2017.History,pathdependenceanddevelopment:evidencefromcolonialrailroads,settlersandcitiesinKenya.Econ.J.,forthcoming.Jedwab,R.,Moradi,A.,2016.Thepermanenteffectsoftransportationrevolutionsinpoorcountries:evidencefromAfrica.Rev.Econ.Stat.98(2),268–284.Miguel,E.,Roland,G.,2011.Thelong-runimpactofbombingVietnam.J.Develop.Econ.96(1),1–15.Moses,L.,Williamson,H.F.,1967.Thelocationofeconomicactivityincities.TheAmericanEconomicReview:PapersandSeventy.In:ProceedingsoftheSeventy-NinthAnnualMeetingoftheAmericanEconomicAssociation57(2),pp.211–222.NOAA,1972.Nationaloceanicandatmosphericadministration.AstudyofearthquakelossesintheSanFranciscobayarea-dataandanalysis,areportpreparedfortheofficeofemergencypreparedness.Tech.rep.,U.S.DepartmentofCommerce.Odell,K.A.,Weidenmier,M.D.,2004.Realshock,monetaryaftershock:the1906SanFranciscoearthquakeandthepanicof1907.J.Econ.Hist.64(4),1002–1027.Redding,S.J.,Sturm,D.M.,Wolf,N.,2011.Historyandindustrylocation:evidencefromGermanairports.Rev.Econ.Stat.93(3),814–831.Rosen,C.,1986.TheLimitsofPower:GreatFiresandtheProcessofCityGrowthinAmerica.CambridgeUniversityPress.Rosenthal,S.S.,2008.Oldhomes,externalities,andpoorneighborhoods:amodelofurbandeclineandrenewal.J.UrbanEcon.63(3),816–840.Sanborn,1899–1914.InsuranceMapsofSanFrancisco,California.Sanborn-PerrisMapCompany,〈http://sanborn.umi.com〉.Sanborn,1905.InsuranceMapsofSanFrancisco,California.Sanborn-PerrisMapCompany,〈http://www.davidrumsey.com〉.Sanborn,1928–1931.InsuranceMapsofSanFrancisco,California.Sanborn-PerrisMapCompany,GeographyMapLibrary,CaliforniaStateUniversity,Northridge.SEIC,1908.Stateearthquakeinvestigationcommission.MapofthecityofSanFranciscoshowingthestreetsandtheburntarea,1906.Tech.rep.,CarnegieInstitutionofWashington.SFLUD,2011.SanFranciscoLandUseDatabase.SanFranciscoPlanningDepartment,CityandCountyofSanFrancisco.SFMR,1904–1916.SanFranciscoMunicipalReports.CityofSanFrancisco,https://archive.org.Shertzer,A.,Twinam,T.,Walsh,R.P.,2016.Zoningandtheeconomicgeographyofcities.Workingpaper.Siodla,J.,2015.RazingSanFrancisco:the1906disasterasanaturalexperimentinurbanredevelopment.J.UrbanEcon.89,48–61.Stewart,J.I.,2012.MigrationtoU.S.frontiercitiesandjobopportunity,1860–1880.Explor.Econ.Hist.49,528–542.SanFranciscoCall,December61905.Newstreettransportationideasaresuggestedtocity’smerchants.SanFranciscoCall,May71906.Toshopontheavenue:ManyretailstoresarebeingrelocatedonVanNess.SanFranciscoRealEstateCircular,variousdates.SunsetMagazine,June1905.Tobriner,S.,2006.BracingforDisaster:Earthquake-ResistantArchitectureandEngineeringinSanFrancisco,1838–1933.HeydayBooks,Berkeley.Walker,T.R.,2000.Economicopportunityontheurbanfrontier:wealthandnativityinearlySanFrancisco.Explor.Econ.Hist.37,258–277.Weiss,M.A.,1988.TherealestateindustryandthepoliticsofzoninginSanFrancisco,1914–1928.Plan.Perspect.3,311–324.Whitney,A.J.,1972.OnInsuranceSettlementsIncidenttothe1906SanFranciscoFire.CaliforniaInstituteofTechnology.J.SiodlaExplorations in Economic History 65 (2017) 1–1616
1  Economics 111B Referee Report Assignment (Due Tuesday, June 4 at the beginning of class)   Department of Economics            Professor Siegler UC Davis              Spring 2019  The purpose of this assignment is to introduce you to published research in economic history.  While a textbook summarizes the main results of thousands of published articles and books, it is important to see how economic history research is actually done.  The referee report assignment is due in paper form at the beginning of class on Tuesday, June 4.  It is worth 75 points (15 percent of your course grade).  If you miss the deadline at the beginning of class, but provide a hard copy of the report to me later in the period or via e‐mail as a PDF file (e‐mailed to msiegler@csus.edu) at any time on June 4, then there will be an automatic 15‐point deduction (60 possible points).  The deduction will increase to 30 points on June 5, 45 points on June 6, 60 points on June 7, and reports will not be accepted after June 7 (the day of the final exam).  Details on the composition and length of the report are described below.   What is a Referee Report?  A referee report is a critical part of the peer‐review process in academic research.  When an article is submitted to an academic journal for possible publication, the editor of the journal sends out the paper to several experts who are familiar with the topic and who can evaluate the contributions and limitations of the submitted article.  Each referee writes a report, which is the basis for a recommendation to the editor on whether to accept the paper for publication, to suggest that the authors revise and later resubmit the paper for publication, or to reject the paper for publication.  Since each paper below has already been published, you do not need to provide this summary recommendation, but your report should consist of all of other aspects of a referee report.  Your report is intended for the authors of the paper.  It should be approximately three double‐spaced pages in length, and it should include the following parts:  
2   A summary of the paper, which discusses the main research question(s), the methodology used by the authors, and the conclusions and contributions of the paper.  A discussion of whether the question or questions asked are important (or not) and whether you find the paper convincing or not (and why?)  A discussion of the concerns you have with the paper.  These concerns could be related to the data being used, the empirical methods, sample selection issues (do the data seem to be a representative sample?), possible alternative explanations for the results that the authors do not consider (or do not consider sufficiently), and any other weaknesses of the paper.  Finally, a referee report is meant to be constructive criticism, so, whenever possible, provide possible suggestions for improvement.  For more detailed guidelines on writing a referee report, see:   Berk, Jonathan B., Campbell R. Harvey, and David Hirshleifer, 2015, “Preparing a Referee Report: Guidelines and Perspectives,” available at SSRN: http://ssrn.com/abstract=2547191 or http://dx.doi.org/10.2139/ssrn.2547191    Possible Articles to Review  As is discussed in the Preface of the textbook, the course is organized around four main themes:  economic growth, distributional issues, economic fluctuations, and the relationship between markets and government.  The articles listed below have been recently published in either Explorations in Economic History or The Journal of Economic History, which are the two leading journals in the discipline, and each article (or articles) fits into one of these main themes.  All of the articles below can be found in Canvas by the author’s or authors’ last name(s).  Please select one of the following four themes for your referee report:  Theme 1:  Economic Growth  There has been substantial debate and concern about the future of economic growth.  The two papers below are written by two experts:  Robert J. Gordon and Joel Mokyr:   Gordon, Robert J., 2018, “Declining American Growth despite Ongoing Innovation, Explorations in Economic History 69, 1‐12.  Mokyr, Joel, 2018, “The Past and the Future of Innovation:  Some Lessons from Economic History,” Explorations in Economic History 69, 13‐26. 
3   In this case, your report on this this theme will cover both papers in a “compare‐and‐contrast” manner, but it must still include your reasoned judgements described in the bullet points on the top of page 2.   Theme 2:  Distributional Issues  The paper below is international and comparative in scope, and it examines the causes of long‐run economic inequality in the United States and other economies:   Bengtsson, Erik, and Daniel Waldenstrom, 2018, “Capital Shares and Income Inequality:  Evidence from the Long Run,” Journal of Economic History 78(3), 712‐743.   Theme 3:  Economic Fluctuations  While much has been written about the “Quantitative Easing” conducted by the Federal Reserve after the Great Recession, which began in December 2007, there was another period of Quantitative Easing during the Great Depression of the 1930s as described in the following paper:   Jaremski, Matthew, and Gabriel Mathy, 2018, “How was the Quantitative Easing Program of the 1930s Unwound?,” Explorations in Economic History 69, 27‐49.   Theme 4:  The Relationship between Markets and Governments  After the 1906 earthquake devastated San Francisco, city planners were largely able to “start from scratch” in terms of city planning.  This paper provided an example of a “natural experiment”:   Siodla, James, 2017, “Clean slate:  Land‐use Changes in San Francisco after the 1906 Disaster,” Explorations in Economic History 68, 1‐16. 

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *