You are on page 1of 21

AgglomerationofinventionintheBayArea:NotjustICT*

ChrisForman,GeorgiaInstituteofTechnology
AviGoldfarb,UniversityofToronto
ShaneGreenstein,HarvardUniversity

December2015

Abstract
WedocumentthattheBayArearosefrom4%ofallsuccessfulUSpatentapplicationsin1976to16%in
2008. This is partly driven by the increase in the prevalence of information and communication
technology;however,evenforpatentsunrelatedtoinformationandcommunicationtechnology,wesee
adisproportionateincreaseintheshareofallUSpatentsfromtheBayArea.Weinterpretthistosuggest
that there has been a trend to coagglomeration in invention across technologies. We explore several
possible explanations for this trend, and conclude that the size of firm or simple measurement error
cannotexplainit.
JELCode:R12,O31

*MaxKovenandYasinOzcanprovidedexcellentresearchassistance.WethankseminarparticipantsattheUniversityofToronto

forhelpfulcomments.AviGoldfarbthanksSSHRCforresearchsupport.

1. Introduction
Doesinventionagglomerate,andifso,wheredoesitagglomerate?Inthispaperweexaminechangesin
patternsofagglomerationininventionovertime,usingdataonallUSpatentapplications.
There are plenty of reasons to expect invention to agglomerate. Carlino and Kerrs (2015) recent
handbookchaptersummarizesmanysuchresults,emphasizingtheroleofinputsharing,labormarket
matching, and knowledge spillovers, among others. Knowledge spillovers received an especially large
fraction of attention in their chapter, and in the literature overall (e.g., Audretsch and Feldman 1996;
AgrawalandHenderson2002;Moretti2004,2012;RosenthalandStrange2008;Jaffe,Trajtenberg,and
Henderson1993;ThompsonandFoxKean2005;KerrandKominers2014;Moser2011;Waldinger2012;
Azoulayetal2010).
Simpleeconomicsmightforecastthatmostinventionagglomeratesinthesameareaastheprimaryusing
industry(CarlinoandKerr2015).Forexample,patentsrelatedtoautomotivetechnologyareclusteredin
Detroit(Hannigan,CanoKollmann,andMudambi2015).Orcausalitycouldbereversed:Thelocationofa
breakthroughinventioncanleadtoindustryagglomerationandlocalizedfollowoninvention(Duranton
2007;Kerr2010).Welabelthiscolocationbetweeninventionandindustry.
However,otherforcespushawayfromcolocation.Inventionitselfisaneconomicactivityanditshares
inputs,suchasspecializedlaborinstitutions,particularintellectualpropertycontracts,andinformation
spillovers from one type of invention to another. If such forces are strong, they could lead to
agglomeration of lots of different types of invention in one place. We call this coagglomeration of
inventionFormanyindustries,thekeyinventionscouldbeinalocationdistinctfromtheplacewhere
productionforthedownstreamusingindustriesreside.
Using patent data to measure invention, there are two approaches to investigate colocation and
coagglomerationofinvention.Oneistomaptheagglomerationofdownstreamindustriesandinvention
and measure the geographic correlation. We take another approach. We look for evidence of
coagglomeration of invention namely, invention from distinct areas appearing in the same location,
irrespectiveofdownstreamusingindustry.
We find evidence consistent with the hypothesis of coagglomeration. We demonstrate a strong trend
towardtheclusteringofpatentingtotheSanFranciscoBayAreafrom4%ofUSpatentsin1976to16%of
USpatentsin2008,atimeperiodwhenthefractionoftheUSpopulationintheBayAreadidnotincrease
substantiallyrelativetotheUSpopulationasawhole.1WhilethisincreaseinBayAreapatentingispartly
drivenbytheincreasingfractionofpatentsininformationandcommunicationtechnologies(ICTs),ICTs
cannotfullyexplainthetrend.TheSanFranciscoBayAreahasseenasubstantialincreaseinitsshareof
patents, even for patents that seem quite distant from ICTs. Our broadest definition of ICT patents
includesallpatentsininformationtechnology,communicationstechnology,electronics,abroadmeasure
ofsoftwarebasedonpatentclassesandatextualsearchofpatenttitlesandabstracts,andallpatents
thatciteanyofthesepatents.TheremainingnonICTpatentswere66%ofallpatentingin1976butjust

AccordingtotheUSCensus,theBayAreagrewfrom2.54%oftheUSpopulationin1980(5.74millionresidents)to
2.65%oftheUSpopulationin2010(8.15millionresidents).

26%in2008.In2008,6.2%ofsuchpatentshaveinventorsbasedintheBayAreasecondonlytoNewYork
Citys8.1%.2
Our results are consistent with coagglomeration of invention in the Bay Area. While others have
documentedatendencytowardagglomerationofpatentingbyindustry,webelievewearethefirstto
documentageneraltendencytowardagglomerationinpatentingacrossindustriesandpatentclasses.
Further,ourstudyisuniqueinitsdocumentationofagglomerationinoneparticularregion,theBayArea.
Coagglomerationhasbeendocumentedinothersettingsandotherindustries.Forexample,Rosenberg
(1963) analyzes how sewing machines, bicycles, and automobiles located in Northern Ohio and
southeastern Michigan as they shared the same set of inventions in machine tools, and the growing
downstreamindustriesinducedadditionalimprovementsinthoseinnovationsovertime.Glaeser(2005)
discusses coagglomeration of many industries in New York City, starting in the nineteenth century.
Summarizingpriorliterature,hearguesthatNewYorksdominancestartedwithshipping.Thereisaclear
reasonwhyNewYorkcoulddominateinshipping:NewYorkhasaparticularlyappealingnaturalharbor
besideaninlandwaterway.Shippingledtorisksharingandinsurance,whichledtofinance.Shippingalso
ledtomanufacturingandearlybookpublishing.Populationgrowth,combinedwithmanufacturingand
financeledtootherservices.Anumberofrecentresearchershaveexploredthecausesandconsequences
ofsuchcoagglomeration,includingEllison,Glaeser,andKerr(2010)andHelsleyandStrange(2015).Of
particular relevance to our study, Delgado, Porter, and Stern (2014) document complementarities
betweenemploymentandpatentinginregionswithmultiindustryclusters.
At thispoint,ourresults donotprovideadefinitive conclusion onthe causeof this broadincreasein
coagglomerationininvention.Avarietyofmechanismsarepossibleincludingregulationssuchasnon
enforcement of noncompete clauses (Franco and Mitchell 2008; Marx, Singh, and Fleming 2015),
agglomerationorexpertiseinstartupfinancing(e.g.,Chen,Gompers,Kovner,andLerner2010),shared
labor markets across invention types (Almeida and Kogut 1999), and knowledge spillovers across
inventiontypes.
Oneimportantlimitationofouranalysisistheuseofpatentsasameasureofinvention.Patentsmeasure
inventionimperfectly,andtheeasewithwhichtheycanbemeasuredmeansthateconomistshavebeen
perhapsoverlyfocusedonpatentstomeasureinvention.Somepatentsaremoreimportantthanothers
andmanyinventionsareneverpatented.Still,patentsareausefulmeasurebecausetheyareobservable,
andcomparableacrosstimeandcategories.ThiswillbiasourresultsiftheinventorsintheBayAreahave
becomeincreasinglylikelytopatentwhentheyinvent.
Ofcourse,wearenotthefirsttodocumenttheagglomerationofICTintheBayArea.GarciaVicenteetal
(2014) show that such agglomeration took place primarily in the 1980s and 1990s. Our results are
consistentwiththistiming.Avarietyofauthorshaveexploredthereasonsbehindtheagglomerationof
theICTindustryintheBayAreaanditsdynamicsingeneratingnewfirmsandnewideas(Almeidaand

OneunusualaspectofpatentingintheSanFranciscoBayAreaisthatinventionisnotcenteredinthecitybutin
SiliconValley.Therefore,whilewerefertoothercitiesbythecitynames,werefertotheBayArearatherthan
SanFranciscotodescribetheSanFranciscoConsolidatedMetropolitanStatisticalArea.Weusethe2013definition,
whichincludesthefollowing12counties:Alameda(fips06001),ContraCosta(fips06013),SanFrancisco(fips06075),
San Mateo (fips 06081), Marin (fips 06041), Santa Clara (fips 06085), San Benito (fips 06069), San Joaquin (fips
06077),Sonoma(fips06097),Solano(fips05095),SantaCruz(fips06087),andNapa(fips06055).

Kogut1999,KerrandKominers2015,Saxenian1994;FrancoandMitchell2008;Marx,SinghandFleming
2015, etc.). Our contribution relates to the finding of the increasing role of the Bay Area in patenting
overall.
2. Dataandempiricalstrategy
WeusepatentsgrantedbytheUSPatentandTrademarkOffice(USPTO)asourmeasureofinvention.
Because of the delay between patent application and grant date, we date patents using the year of
application.Wehavedataonpatentsgrantedbetween1976and2012,andouranalysisdatasetincludes
patentswithapplicationdatesbetween1976and2008.Wecutoffthelastfouryearsofthedatabecause
oflagsbetweenyeargrantedandyearfiled.Generally,westarttoseeadeclineinpatentingin2008,
suggestingrighttruncation may beanissueforthelastfewyearsofourdata.Thetrendsweidentify
appearlongbefore2008.
Patentshavebeenshowntoprovideausefulmeasureofafirmsintangiblestockofknowledge(Hallet
al.2005).Theirlimitationsarewellknown.NotallpatentsmeettheUSPTOcriteriaforpatentability(Jaffe
andTrajtenberg2002).Notallinventorsseektopatent,andmanyusealternativemeanstoappropriate
valuefromtheirinventions.Furtherthepropensitytopatenthaschangedovertimeduringoursample
(e.g.,HallandZiedonis2001),thiswasparticularlythecaseforpatentsrelatedtosoftwarewhichgrew
rapidlytowardtheendofoursampleperiodduetolegalchangeswhichstrengthenedthelegalrightsof
patentsinthisarea(e.g.,GrahamandMowery2003,HallandMacGarvie2010).Ouruseofpatentcitations
as a measure of knowledge flows between successive generations of inventions can also create
measurement error (Roach and Cohen 2013). We are comfortable with using patents in this context
because our primary focus is on changes in the geographic distribution of patenting within broad
technologyareasovertime.Whilethepropensitytopatenthaschangedacrosspatentclassesovertime,
wedonotbelieveithaschangedsignificantlyacrossgeographiclocationspatentingwithinapatentclass.
We map inventors to counties and MSAs using the zip code of the location of the inventor. We used
consolidatedMSAs(CMSAs)wherethosewerepresent.Thiswillbeparticularlyimportantforouranalysis
oftheBayArea,whichincludesseveralcomponentPMSAssuchasOakland,SanFrancisco,SanJose,Santa
CruzWatsonville,SantaRosa,StocktonLodi,VallejoFairfield,andNapa.
Formostoftheanalysisthatfollows,wedonotweightbycitations.Formultiauthorpatents,wedivide
bythenumberofauthors.Forexample,ifapatenthas1authorintheBayAreaand2authorsinBoston,
itwouldcountas1/3ofapatentintheBayAreaand2/3ofapatentinBoston.Ourresultsaregenerally
robust,andoftenstronger,usingthreeyearandfiveyearcitationweightedmeasures.Forexample,using
eitherthreeorfiveyearcitationweights,theBayAreasurpassesNewYorkCityasthelocationwiththe
mostpatentsthreeyearsearlierthanwiththeunweightedmeasure.
Someofourresultsrequireustorecognizeaconsistentidentifierforassignees(especiallyfirms)withina
particularapplicationyear.Forthispaper,wedonotseektoidentifychangesinpatentingactivitywithin
firmsovertime.Becauseassigneenamesarenotcodedconsistentlywithinthepatentdata,thechallenges
of mapping patents to assignees is well known. No prior data set provides a complete set of cleaned
assigneenamesduringoursampleperiod.UsingthedatafileofstandardizednamesintheNBERdatabase

andthenamesinICTindustriescompiledbyOzcanandGreenstein(2013)asstartingpoints,wecleaned
theassigneenamestocreateourownassigneeidentifier.
OuranalysisrequiresustoidentifypatentsthatrepresentinventionsrelatedtoICT,orinventionsthat
drawuponthestockofknowledgerelatedtoICT.Asiswellknown,identifyingsuchinventionsthrough
thepatentdataisnotoriouslydifficult(see,e.g.,GrahamandMowery2003,BessenandHunt2007;Hall
andMacGarvie2010).Asaresult,weuseseveraldifferentdefinitionsbasedontheprimaryclassofthe
patentandexploretherobustnessofourresultstofouralternatives.Wediscusstheconstructionofthese
alternativesintheonlineappendix.
Ourdatacontainatotalof2,213,271patents.In1976,therewere41,100newpatentsissuedfromthe
PTO.Atthepeakofourdatain2007,therewere100,832patents.
We present our results at the year level, as aggregated means over the 33 years from 1976 to 2008
inclusive.Inparticular,ourresultsarepresentedasgraphsoftimetrendsofthefractionofpatentseach
yearthatmeetsomecriteriasuchasbeingbasedintheBayArea.Thisisthereforeadescriptiveexercise
thattestswhethertheresultsareconsistentwithincreasingcoagglomerationintheSanFranciscoBay
Areaovertime.Wehavenotdeterminedtheprimarycause(s)oftheobservedpatterns.

3. Results
a) Patentingacrosslocations
Giventheoverallriseinthepropensitytopatent,allmajorcitieshadanincreaseinthenumberofpatents.
WeexplorethefractionofallUSpatentsbycity,therebycontrollingfortheoveralltrend.
Figure1showstheincreasingimportanceoftheBayAreaasafractionofUSpatenting.Figure1acompares
thetop10citiesintheUnitedStates,definedbythetotalnumberofpatentsbetween1976and2008.In
1976,NewYorkCitywasthedominantcenterforpatenting,withjustunder15%ofallpatents.LosAngeles
wassecondandChicagowasthird.Generally,patentingwashighlycorrelatedwithpopulation.TheBay
Arearosesteadilyasafractionofpatentinginthe1970sand1980s,andthenthetrendincreasedinthe
1990sbeforesettlingdownattheearlierrateofincreaseinthe2000s.In1995,theBayAreasurpassed
NewYorkCityastheUSlocationwiththelargestnumberofpatents.Figure1bcontraststhe11ththrough
20thcitiesinpatentingwiththeBayAreainordertoshowthatnoothercityhasarisesimilarinscale.
Figure1ccombineslocationsintofourgroups:theBayArea,NewYorkCity,the18othercitiesinthetop
20,andallotherlocations.GenerallywhileNewYorkandlocationsoutsidethetop20arefallingasa
proportionofpatenting,theBayArea isrisingquickly,and the other18 citiesinthetop20arerising
slightly(42.6%in1976to46.1%atthepeaklevelin2004).

b) Patentingacrosstypesofpatents
TheBayAreahashadaclusterofICTfirmsformanyyears.Therefore,onereasontheBayAreaisbecoming
an increasing large fraction of patenting is that overall increase in ICT patents. Figure 2 displays this

increase using the Hall, Jaffe, and Trajtenberg (HJT) definitions of patent classes. Computers and
Communication(Class2)wentfromunder10%ofpatentstoover30%ofpatentsbetween1980and2005.
SomeofthisgrowthmayreflectchangesinthepropensitytopatentsoftwareandotherICTinventions
(e.g., Graham and Mowery 2003, Hall and Ziedonis 2001) that have been encouraged by sympathetic
treatmentinthecourtsandthePTO.DrugsandMedical(Class3)trackedtheincreaseinComputersand
Communicationuntilthemid1990sbutthensettledbacktoaround13%ofpatents.
WeofferthefirstevidenceofthecoagglomerationhypothesiswithFigure3,whichshowsthefractionof
patentsthatareintheBayAreabybroadclass.TheincreaseissharpestinComputersandCommunication
and in Electrical and Electronic (Class 4). It is also visibly noticeable in Chemicals (Class 1), Drugs and
Medical,andMechanical(Class5).InOther(Class6)theincreaseissmaller,risingfrom3.8%in1976to
peakof6.3%in2004beforefallingbackto4.4%in2008.Thus,forfiveofsixbroadpatentclasses,wesee
anoticeableriseintheproportionofpatentscomingfromtheBayArea.
OnepossibilityisthatmanyofthepatentsinChemicals,DrugsandMedical,andMechanicalclassesare
ICTbased.Softwarehasincreasinglybeenusedasaninputintoawiderarrayofinventionsinotherpatent
categories(Arora,Branstetter,andDrev2013;Branstetter,Drev,andKwon2015),takingadvantageof
increasinglyinexpensiveandmorecapableelectronics,especiallyprocessors.Figure4providesalternative
measuresofICTandnonICTpatentstoaccountforthispossibility,andexaminesthetrendovertime.
PanelAofFigure4providesthenarrowestdefinition.ItdefinesICTpatentsaspatentsinHJTClass2:
ComputersandCommunication.Thesolidlineatthetopofthegraphshowstheincreasingproportionof
Class2patentsthatareintheBayArea,replicatingthesolidlineatthetopofFigure3.PanelBprovides
awiderdefinition,includingsoftwarepatentsasdefinedbyGrahamandMowery(2003).PanelCprovides
astillwiderdefinition,addingtothedefinitioninpanelBallElectricalandElectronicpatents(HJTClass
4).PanelDprovidesourwidestdefinition,whichusesthedefinitioninPanelCasastartingpointandthen
widensittoincludesoftwarepatentsidentifiedthroughakeywordsearchasinBessenandHunt(2007)
andsoftwarepatentsidentifiedin GrahamandVishnubhakat (2013). Thisbroadestdefinitionincludes
74%ofallpatentsin2008andislikelytoincludemanyfalsepositives.Usingallfour,increasinglybroad,
definitionsofICTpatents,thesolidlinesshowthattheproportionofallICTpatentsintheBayAreahas
risensharplysincethe1970s.
ThedottedlineidentifiesallpatentsthatciteICTpatents.Ineachpanel,theICTpatentsaredefinedasin
thepreviousparagraph.ThesepatentsarenotexplicitlycategorizedasICTusingthedefinitionabove,but
theyareconnectedthroughcitationandthereforebuildonICTinvention.Thereisacleartrendtoward
anincreasingproportionofthesepatentsintheBayArea,providinganotherexplanationfortheriseof
BayAreapatents.
Together,theabovesuggestthefollowing:ICTisanincreasinglylargefractionofpatents;theBayAreais
anincreasinglylarge(andevendominant)fractionofICTpatents;andtheBayAreaisanincreasinglylarge
fractionofpatentsthatciteICTpatents.GivenpriorresultsonagglomerationoftheICTindustryinthe
BayArea,perhapsnoneoftheseresultsaresurprising,thoughwebelievethattheresultsongeography
of patents that cite ICT are not previously documented. These all could result from agglomeration of
softwareinventionnearthelocationofthefirmsproducingelectronics,computing,andcommunications.

TheevidenceforcoagglomerationofinventionappearsinthedashedlineinFigure4:TheBayAreaisan
increasingfractionofnonICTUSpatents,evenforthebroadestdefinitionsofsoftware.PanelDshows
thatthefractionofnonICTpatentsintheBayArearisesfrom3.9%to6.2%from1976to2008,a59%
increase.UnderthenarrowersoftwaredefinitioninPanelC,thevaluerisesfrom3.9%to6.9%.Dropping
electronics,asinPanelB,theproportionofnonICTpatentsintheBayArearisesfrom4.3%to9.5%.
WhilethesefiguresaremoremodestthantheincreaseinICTpatents,theystillsuggestanincreasingly
importantrolefortheBayArea,relativetoallotherareas,inUSnonICTpatenting.Figure5comparesthe
Bay Area to the four other top patenting cities in the United States. Panel A uses the narrower ICT
definitionthatincludesSoftwareandComputersandCommunicationspatentsinFigure4panelB.Under
thisdefinition,theBayAreaovertookNewYorkasthetoplocationfornonICTpatentingin2000.PanelB
includesElectricalandElectronicpatentsasinFigure4panelC.Underthisbroaddefinition,theBayArea
wassecondbehindNewYorkformostoftheperiodfrom1997to2008.Usingthebroadestdefinition(as
inFigure4panelD)yieldsasimilarpattern(though,asnotedabove,thatdefinitionwillincludemanyfalse
positivesonsoftwarepatents).
Overall,weinterprettheseresultstosuggestthatwecannotrejectcoagglomerationofinvention.The
increaseinpatentingintheBayAreaisnotentirelyattributabletotheincreasingfractionofICTpatents
inoverallpatenting.
c) Patentingacrosstypesofpatentees
Does the evidence for coagglomeration reflect some other factors, such as firm size? We look at the
fractionofpatentsintheBayAreabysizeofpatentee.Wefindthatsizedoesnotexplaintheresults.
We split patentees into four categories: Independent inventors, firms in the 99th percentile of firm
patentees, firms between the 50th and 99th percentile of firm patentees, and firms below the 50th
percentileofpatentees.Figure6PanelAshowsthatthefractionofpatentsintheBayAreaisrisingsharply
overtimeforallfourgroups.The99thpercentilegrouphasamorediscontinuousincrease,largelydriven
by a sharp increase in the proportion of top patenting firms that are ICT firms in the 1990s. The only
exceptiontothegeneraltrendisthat,since2000,theproportionofallpatentsfromsmallerpatenting
firms(belowthe50thpercentile)intheBayAreahasdeclinedtothelevelsoftheearly1990s.PanelB
looksatICTpatentsonly(definedtoincludesoftwareandelectronicsasinFigure4panelC)andshowsa
steadyriseinthe proportionofsuchpatentsintheBayAreabetween1976 and2006,againwith the
exceptionofsmallerpatentingfirmssince2000.
PanelClooksatnonICTpatents,againdefinedasinFigure4panelC.ThefractionintheBayAreahas
been rising steadily for firms below the 99th percentile and for independent inventors. In the 99th
percentile, results are noisier as the entry or exit from the 99th percentile of one firm can make a
meaningfuldifference.Forexample,thevalueisgenerallyunder3%withtheexceptionoftheperiod1997
to 2002. During this period, at least one of Genentech, Advanced Cardiovascular Systems, and Incyte
Pharmaceuticalswasinthe99thpercentileofnonICTpatenteeseachyear.
4. Conclusions

WehavedocumentedanincreaseofthefractionofUSpatentingofallkindsthatoccurstheBayAreathat
isdisproportionatetopopulationgrowthandoccurswithinavarietyofpatentclasses.Thispartlyresults
from the agglomeration of invention near the production of firms who use the invention, and who
themselvesagglomerateinonearea.Wealsothinkitoffersevidenceofcoagglomeration,theclustering
ofinventionfrommanydistincttypesofinventionintoonegeographicarea.
WhilewedonotknowthecauseoftheriseincoagglomerationofmanypatenttypesintheBayArea,our
resultssuggestthatanypossibleexplanationmustbebroadbased.Inparticular,anyexplanationmust
accountforgrowthinthefractionofICTandnonICTpatentsintheBayAreaandfortheincreasetobe
trueoflargeinventingfirms,smallinventingfirms,andindependentinventors.

Figure1:Fractionpatentsbylocation
PanelA:Top10cities

PanelB:Next10citiesplusBayArea

PanelC:BayArea,NewYorkCity,Othertop20cities,andelsewhere

Figure2:Informationandcommunicationtechnologyisbecomingincreasinglyimportant

Figure3:FractionofinnovationintheBayAreabyHJTClass

Figure4:AlternativedefinitionsofICT
PanelA:Narrowestdefinition

PanelC:Includesoftware&electronics

PanelB:Includesoftware

PanelD:Includebroadsoftware&electronics

Figure5:AlternativedefinitionsofnoICT,comparisonacrosstop5cities
PanelA:NoICTwhereICTdefinedasICTormainsoftwaredefinition,orcitesthose

PanelB.NOICTwhereICTdefinedasICTormainsoftwaredefinitionorelectronicsorcitesthose

Figure6:BayAreaFractionpatentingbytypeofpatentee
PanelA:Allpatents

Fraction of patents in the Bay Area


.05
.1
.15
.2
.25

PanelB:ICTpatents,definedbybestsoftwaredefinitionandelectronic

1970

1980

1990
Year

2000

99th percentile firms

Firms between 50th and 99th percentile

Firms below the 50th percentile

Independent inventor

2010

Fraction of patents in the Bay Area


.05
.1
.15

.2

PanelC:NONICTpatents,definedbybestsoftwaredefinitionandelectronic

1970

1980

1990
Year

2000

99th percentile firms

Firms between 50th and 99th percentile

Firms below the 50th percentile

Independent inventor

2010

ExtendedReferenceList

Aarland,Kristin,JamesC.Davis,J.VernonHenderson,andYukakoOno.2007.Spatialorganizationoffirms:
thedecisiontosplitproductionandadministration.RANDJournalofEconomics38(2):480494.

Almeida,PaulandBruceKogut.1999.LocalizationofKnowledgeandtheMobilityofEngineersinRegional
Networks.ManagementScience45(7):905917.

Arora,Ashish,LeeG.Branstetter,andMatejDrev.2013.GoingSoft:HowTheRiseOfSoftwareBased
InnovationLedToTheDeclineOfJapansItIndustryAndTheResurgenceOfSiliconValley.Review
ofEconomicsandStatistics,95(3):757775.

Arora, Ashish and Alfonso Gambardella. 2005. From Underdogs to Tigers: The Rise and Growth of the
SoftwareIndustryinBrazil,China,Ireland,andIsrael.Oxford,UK:OxfordUniversityPress.

Audretsch, D., Feldman, M., 1996. R&D spillovers and the geography of innovation and production.
AmericanEconomicReview.86,630640.

Azoulay,P.,JGraffZivin,andJ.Wang.2010.Superstarextinction.QuarterlyJournalofEconomics.125(2),
549589.

Barro, Robert J. and Xavier SalaiMartin. 1991. Convergence across States and Regions. Brookings
PapersonEconomicActivity,(1991)1:107158.

Bessen,James,andRobertHunt.2007.AnEmpiricalLookatSoftwarePatents.JournalofEconomicsand
ManagementStrategy.16(1),157189.

Blanchard,OlivierJeanandLawrenceF.Katz.1992.RegionalEvolutions.BrookingsPapersonEconomic
Activity,1992(1):161.

Bloom,Nicholas,RaffaellaSadun,andJohnVanReenen.2012.AmericansDoITBetter:USMultinationals
andtheProductivityMiracle.AmericanEconomicReview102(1):167201.

Bloom, Nicholas, Luis Garicano, Raffaella Sadun, and John Van Reenen. 2009. The Distinct Effects of
InformationTechnologyandCommunicationTechnologyonFirmOrganization.NBERWorking
Paper#14975.

Bresnahan, Timothy F., Erik Brynjolfsson, and Lorin M. Hitt. 2002. Information Technology, Workplace
Organization, and the Demand for Skilled Labor: FirmLevel Evidence. Quarterly Journal of
Economics117(1):339376.

Bresnahan, Timothy and Shane Greenstein. 1996. Technical progress in computing and in the uses of
computersBrookingsPapersonEconomicActivity:Microeconomics(1996):178.

Branstetter,Lee,MatejDrev,andManhoKwon.2015.GetwiththeProgram:SoftwareDrivenInnovation
inTraditionalManufacturing.Workingpaper.

Brynjolfsson, Erik and Lorin Hitt. 2003. Computing Productivity: FirmLevel Evidence. Review of
EconomicsandStatistics,85(4):793808.

Bresnahan, Timothy, and Franco Malerba. 1999. Industrial Dynamics and the Evolution of Firms and
NationsCompetitiveCapabilitiesintheWorldComputerIndustry.InDavidMoweryandRichard
Nelsoneds,SourcesofIndustrialLeadership:StudiesofSevenIndustries.CambridgeUniversity
Press.Chapter3.

Cairncross,F.1997.TheDeathofDistance.Cambridge,MA:HarvardUniversityPress.

Carlino, Gerald, and William Kerr. 2015. Agglomeration and Innovation. In Gilles Duranton, Vernon
Henderson,andWilliamStrangeeds.HandbookofRegionalandUrbanEconomicsVolume5.p.
349404.

Charlot,SylvieandGillesDuranton.2006.Citiesandworkplacecommunication:SomequantitativeFrench
Evidence.UrbanStudies43(8):13651394.

Chen,Henry,PaulGompers,AnnaKovner,andJoshLerner.2010.Buylocal?Thegeographyofventure
capital.JournalofUrbanEconomics67(1):90102.

Chinitz,B.,1961.Contrastsinagglomeration:NewYorkandPittsburgh.AmericanEconomicReview.51
(2),279289.

Delgado,M.,Porter,M.,Stern,S.,2010.Clustersandentrepreneurship.JournalofEconomicGeography.
10(4),495518

Delgado, Mercedes, Michael Porter, and Scott Stern, 2014. Clusters, Convergence, and Economic
Performance.ResearchPolicy43(10),17851799.

Duranton,G.,2007.Urbanevolutions:thefast,theslow,andthestill.AmericanEconomicReview.97,
197221

Duranton,GillesandDiegoPuga.2005.Fromsectoraltofunctionalurbanspecialization.JournalofUrban
Economics57:343370.

Ellison, Glenn, Edward L. Glaeser, and William R. Kerr. 2010. "What Causes Industry Agglomeration?
EvidencefromCoagglomerationPatterns."AmericanEconomicReview,100(3):11951213.

Feldman,M.,1994.TheGeographyofInnovation.KluwerAcademic,Boston,MA.

Forman,Chris,AviGoldfarb,andShaneGreenstein.2005.HowDidLocationAffecttheAdoptionofthe
CommercialInternet?GlobalVillagevs.UrbanDensity.JournalofUrbanEconomics58(3):389
420.

Forman,Chris,AviGoldfarb,andShaneGreenstein.2008.UnderstandingtheInputsintoInnovation:Do
CitiesSubstituteforInternalFirmResources?JournalofEconomicsandManagementStrategy
17(2):295316.

Forman, Chris, Avi Goldfarb, and Shane Greenstein. 2012. The Internet and Local Wages: A Puzzle.
AmericanEconomicReview102(1):556575.

Forman,ChrisandNicolasvanZeebroeck.2012.Fromwirestopartners:HowtheInternethasfostered
R&Dcollaborationsamongfirms.ManagementScience58(8):15491568.

Franco, April, and Matthew Mitchell. 2008. Covenants not to Compete: Labor Mobility and Industry
Dynamics.JournalofEconomicsandManagementStrategy17(3),581606.

GarciaVicente, Florencia, Daniel GarciaSwartz, and Martin CampbellKelly. 2014. The Hydra of Lerna:
EconomicGeography Perspectives on Americas Early Computer Clusters. Working paper,
WarwickUniversity.

Gaspar, J. and E. L. Glaeser. 1998. Information Technology and the Future of Cities. Journal of Urban
Economics43:136156.

Glaeser,Edward.2005.UrbanColossus:WhyisNewYorkAmericasLargestCity?FRBNYEconomicPolicy
Review.December.p.724.

Glaeser, Edward L., Hedi D. Kallal, Jos A. Scheinkman, and Andrei Shleifer. 1992. Growth in Cities.
JournalofPoliticalEconomy,100(6):112652.

Glaeser,E.,Kerr,W.,Ponzetto,G.,2010.Clustersofentrepreneurship.JournalofUrbanEconomics.67(1),
150168

Glaeser,E.,Kerr,S.,Kerr,W.,2015.Entrepreneurshipandurbangrowth:anempiricalassessmentwith
historicalmines.Rev.Econ.Stat.,forthcoming.

Glaeser,E.L.andGiacomoA.M.Ponzetto.2007.DidtheDeathofDistanceHurtDetroitandHelpNew
York?NBERWorkingPaper13710.

Graham,Stuart,andDavidMowery.2003.IntellectualPropertyProtectionintheU.S.SoftwareIndustry.
InWesleyCohenandStephenMerrilleds,PatentsintheKnowledgeEconomy,NationalResearch
Council,WashingtonDC.p.219258

Graham,Stuart,andSaurabhVishnubhakat.2013.OfSmartPhoneWarsandSoftwarePatents.Journalof
EconomicPerspectives27(1),6786.

Greenstein,Shane.2015.HowtheInternetBecameCommercial:Innovation,Privatization,andtheBirth
ofaNewNetwork.PrincetonUniversityPress;Princeton,NJ.

Hall,Bronwyn,andMeganMacGarvie.2010.Theprivatevalueofsoftwarepatents.ResearchPolicy39(7),
9941009.

Hall, B. H., R. Ziedonis 2001. The Patent Paradox Revisited: An Empirical Study of Patenting in the US
SemiconductorIndustry,19791995.RANDJournalofEconomics32:101128.

Hannigan, Thomas J., Marcelo CanoKollmann, and Ram Mudambi. 2015. Thriving innovation amidst
manufacturingdecline:theDetroitautoclusterandtheresilienceoflocalknowledgeproduction.
IndustrialandCorporateChange24(6):122.

Henderson,Vernon.2003.MarshallsScaleEconomies.JournalofUrbanEconomics53(1):128.

Hensley, Robert W. and William C. Strange. 2014. Coagglomeration, Clusters, and the Scale and
CompositionofCities.JournalofPoliticalEconomy.122(5),10641093.

Higgins,MatthewJ.,DanielLevy,andAndrewT.Young.2006.GrowthandConvergenceacrosstheUnited
States:EvidencefromCountyLevelData.ReviewofEconomicsandStatistics,88(4):67181.

Hsieh,ChangTai,andEnricoMoretti.2015.WhyDoCitiesMatter?LocalGrowthandAggregateGrowth.
NBERWorkingPaper21154.

Isaacson,Walter.2011.SteveJobs.Simon&Schuster,NewYork.

Jaffe, Adam and Manuel Trajtenberg. 2002. Patents, Citations, and Innovations: A Window on the
KnowledgeEconomy.MITPress,Cambridge,MA.

Jaffe, A., M. Trajtenberg and R. Henderson 1993. Geographic Localization of Knowledge Spillovers as
EvidencedbyPatentCitations.QuarterlyJournalofEconomics108(3):577598.

Kerr,W.,2010.Breakthroughinventionsandmigratingclustersofinnovation.JournalofUrbanEconomics.
67(1),4660.

Kerr,William,andScottKominers.,2015.Agglomerativeforcesandclustershapes.ReviewofEconomics
andStatistics,forthcoming

Levy,FrankandRichardJ.Murnane.2005.TheNewDivisionofLabor:HowComputersareCreatingthe
NextJobMarket.Princeton:PrincetonUniversityPress.

Macher,J.andD.Mowery.2008.InnovationinGlobalIndustries:U.S.FirmsCompetinginaNewWorld.
Washington,DC:NationalAcademiesPress.

Magrini, Stefano. 2004. Regional (Di)Convergence. In Handbook of Regional and Urban Economics,
Volume4,ed.J.VernonHendersonandJacquesFrancoisThisse,224392.Amsterdam:Elsevier
NorthHolland.

Marshall,Alfred.1920.PrinciplesofEconomics.8thed.NewYork:PorcupinePress.

Marx,M.J.Singh,andL.Fleming,RegionalDisadvantage?EmployeeNoncompeteAgreementsandBrain
Drain.ResearchPolicy44(2):394404(2015).

Moretti,E.,2004.Workerseducation,spilloversandproductivity:evidencefromplantlevelproduction
functions.AmericanEconomicReview.94,656690.

Moretti,E.,2012.TheNewGeographyofJobs.HoughtonMifflinHarcourt,NewYork.

Moretti,E.,2013.Realwageinequality.AmericanEconomicJournal:AppliedEconomics.5(1),65103.

Moser,P.,2011.Didpatentsweakenthelocalizationofinnovations?EvidencefromWorldsFairs.Journal
ofEconomicHistory.71(2),363381.

Ozcan,YasinandShaneGreenstein.2013.The(de)ConcentrationofSourcesofInventiveIdeas:Evidence
fromICTEquipment.WorkingPaper.

OSullivanandStrange.2015.TheemergenceofCoagglomeration.Workingpaper,UniversityofToronto.

Perlman,ElizabethRuth.2015.DenseEnoughtobeBrilliant.Workingpaper.BostonUniversity.

Powell,WalterW.,andEricGiannella.2009.CollectiveInventionandInventorNetworks.Handbookof
EconomicsofInnovation.BronwynHallandNathanRosenbergEds.

Powell, Walter W., Kelley Packalen, and Kjensten Whittington. 2011. Organization and Institutional
Genesis:TheEmergenceofHighTechClustersintheLifeSciences.

Roach,MichaelandWesleyM.Cohen.2013.LensorPrism?PatentCitationsasaMeasureofKnowledge
FlowsfromPublicResearch.ManagementScience59(2):504525.

Rosenberg,Nathan,1963,TechnologyChangeintheMachineToolIndustry,18401910.TheJournalof
EconomicHistory,23(4),pp.414443.

Rosenthal,S.,Strange,W.,2004.Evidenceonthenatureandsourcesofagglomerationeconomies.In:
Henderson, J.V., Thisse, J.F. (Eds.), Handbook of Regional and Urban Economics, vol. 4.
NorthHolland,Amsterdam,pp.21192171.

Rosenthal,S.,Strange,W.,2008.Theattenuationofhumancapitalspillovers.JournalofUrbanEconomics.
64(2),373389.

Saxenian,A.,1994.RegionalAdvantage:CultureandCompetitioninSiliconValleyandRoute128.Harvard
UniversityPress,Cambridge,MA.

Storper, M. and Venables, A. J. 2004. Buzz: facetoface contact and the urban economy. Journal of
EconomicGeography.4(4):351370.

Thompson, P., FoxKean, M., 2005. Patent citations and the geography of knowledge spillovers: a
reassessment.AmericanEconomicReview95(1),450460

von Hippel, Eric. 1998. Economics of Product Development by Users: The Impact of Sticky Local
Information.ManagementScience44(5):629644.

Waldinger,F.,2012.Peereffectsinscience:evidencefromthedismissalofscientistsinNaziGermany.
Rev.Econ.Stud.79,838861.

Watzinger,Martin.2015.Antitrust,Patents,andCumulativeInnovation:EvidencefromBellLabs.Working
paper.


Zook,MatthewA.2005.TheGeographyoftheInternetIndustry:VentureCapital,Dotcoms,andLocal
Knowledge.BlackwellPublishers.

OnlineAppendix:Identificationofpatentsrelatedtoinformationandcommunicationtechnologies(ICT)

Asnotedinthepaper,wetakefourapproachestoidentifyinventionsthatarerelatedtoICT,orthatdraw
uponthestockofknowledgerelatedtoICT.Allfourapproachesarebasedontheprimaryclassofthe
patent.
OurfirstapproachusesthecategoriesfromtheNBERPatentDataprojects2006updateofthecategories
developedbyHall,Jaffe,andTrajtenberg(HJT)(2001).WeidentifyICTpatentsasHJTcategory2,which
includessubcategoriessuchascomputerhardware&software,communications,computerperipherals,
informationstorage,andelectronicbusinessmethodsandsoftware,amongothers.
Oursecondapproachaddspatentsspecificallyidentifiedassoftwarepatentstothesetdefinedinourfirst
approach. It relies upon identifying softwarerelated patent classes in the International Patent
Classification (IPC) system. Specifically, we use the approach defined in Graham and Mowery (2003),
which identified software patents as being in the class/subclasses Electric Digital Data Processing
(G06F),RecognitionofData;PresentationofData;RecordCarriers;HandlingRecordCarriers(G06K),
and Electric Communication Technique (H04L). Compared to USPTO classes, software is more
concentrated in the IPC system. The GrahamMowery approach has been used as inputs into the
definitionsusedinavarietyofpapersthathaverequiredidentificationofsoftwarepatents,includingHall
andMacGarvie(2010)andArora,Branstetter,andDrev(2013),amongothers.
OurthirddefinitionaddspatentsinHJTcategory4(electronics)tothesetofICTpatentsdefinedinour
seconddefinition.ThisdefinitionwillcaptureelectronicsinventionsthatarecloselyrelatedtoICT,such
asinventionsrelatedtosemiconductordevices.
OurlastapproachwidensthesetofICTpatentsdefinedinapproach3intwoways.First,itaddspatents
identified as software through a keyword search. Specifically, we identify software patents using the
keywordsfromBessenandHunt(2007)tosearchforspecificsoftwarerelatedstringsinthetitlesand
abstractsofpatentsinoursample.ThisapproachalsoincludespatentsfromtheUSPTOclassesidentified
byGrahamandVishnubhakat(2013)ascontaininggeneralpurposesoftwareorsoftwarethatisspecific
tosometypeofhardware.

References

Arora,Ashish,LeeG.Branstetter,andMatejDrev.2013.GoingSoft:HowTheRiseOfSoftwareBased
Innovation Led To The Decline Of Japans IT Industry And The Resurgence Of Silicon Valley.
ReviewofEconomicsandStatistics,95(3):757775.

Bessen,James,andRobertHunt.2007.AnEmpiricalLookatSoftwarePatents.JournalofEconomicsand
ManagementStrategy.16(1),157189.

Graham,Stuart,andSaurabhVishnubhakat.2013.OfSmartPhoneWarsandSoftwarePatents.Journalof
EconomicPerspectives27(1),6786.

Hall,BronwynH.,AdamJaffe,andManuelTrajtenberg.2001.TheNBERpatentcitationsdatafile:Lessons,
insights,andmethodologicaltools.NBERWorkingPaper8498.

Hall,Bronwyn,andMeganMacGarvie.2010.Theprivatevalueofsoftwarepatents.ResearchPolicy39(7),
9941009.

You might also like