日本軟件行業(yè)：什么搞錯了以及我們如何吸取教訓

TheJapaneseSoftware

Industry:

WHATWENTWRONGANDWHAT

CANWELEARNFROMIT?

RobertE.Cole

YoshifumiNakata

RecentfindingsindicatethattheJapaneseITsectorincreasinglylagstheU.S.ITsectorinsoftwareinnovationandthatthisunderliesJapan’sweakeningcompetitiveperformancevis-à-visU.S.IT.Thisarticleexploresalter-nativeexplanationsforthisoutcomeandanalyzeswhatexplainstheJapanesesoftwareindustry’strajectory.Thesourcesarefoundinthelateunderstandingofthetransformationalroleofsoftwareanditsvalue-creatingpotentialaswellasintheevolutionoftheindustry’sstructure.Finally,thisarticleconsiderswhatpolicymakersinothernationsmightlearnfromtheJapaneseexperienceinbuildingamorevibrantsoftwareindustry.(Keywords:HighTechnology,SoftwareIndustry,InformationTechnology,PathDependency,StrategicManagement,Japan,Learning)

I

tishardtoevenimaginethatjustsometwenty-oddyearsago,AmericanswerebeingwarnedthatJapanwaspositionedtobecomethenextsoft-waresuperpower,followinguponitssuccessinmanufacturing.SmallfragmentedSiliconValleyfirmswereseenasnomatchforthefinancial

resourcesoflargeintegratedJapanesehigh-techmanufacturers.Moreover,thefactoryapproachtosoftwaredevelopment,favoredbylargeJapanesefirms,wasseenassuperiortothecraftapproachdominantintheU.S.1

Theauthorsgratefullyacknowledgeresearchfundingfrom:theInstituteforTechnology,EnterpriseandCompetitiveness,DoshishaUniversity,Kyoto,Japan;theMinistryofEducation,Culture,Sports,ScienceandTechnologySupportedProgramfortheStrategicResearchFoundationatPrivateUniversities,2009-2013,andJSPSKAKENHIGrantNumber23243059;theGarwoodCenterforCorporateInnovation,InstituteforBusinessInnovation,attheHaasSchoolofBusiness,UCBerkeley;andtheCaliforniaManagementReview.Weareindebtedtoouranonymousreviewersfortheirrevisionproposals.Wearealsoindebtedtothe

encouragementandadvicereceivedfromthefollowingindividuals:EvaChen,TakahiroFujimoto,MichaelGer,KenjiHiranabe,DavidHodges,ToshiroKita,HidekoKunii,KazuyukiMotohashi,ShinjiTakai,TetsuoTamai,MasatoTakeichi,HirofumiTatsumoto,HughWhittakerandHideyukiYamagishi.WeoweaspecialdebttoKatsutoshiShintani,formerChiefAdvisor,SoftwareEngineeringCenter,Information-technologyPromotionAgency(IPA).Heshowedenormouspatiencewithourendlessquestionsandpointedustomanyrelevantdatasources.

16UNIVERSITYOFCALIFORNIA,BERKELEYVOL.57,NO.1FALL2014CMR.BERKELEY.EDU

RobertE.ColeisEmeritusProfessor

ofBusinessandSociologyatUCBerkeley

andJointResearcherattheInstituteof

Technology,EnterpriseandCompetitiveness(ITEC),DoshishaUniversity,Kyoto,Japan.

<cole@>

YoshifumiNakataisProfessorofHumanResourceManagementattheGraduate

SchoolofPolicyandManagementaswellastheDoshishaBusinessSchooland

Director-GeneraloftheInstituteof

Technology,Enterpriseand

Competitiveness(ITEC),Doshisha

University,Kyoto,Japan.<ynakata@mail.doshisha.ac.jp>

TheJapaneseSoftwareIndustry

Notwithstandingthesepredictions,thelasttwodecadeshavenotbeenkindtoJapaneseITfirmsaswellasmanyhigh-techmanufacturingfirms.Inthisarticle,weexplorerecentresearchontheimpactofsoftwareinnovationonnationalcompetitivenessusingJapaneseexperiences,focus-ingprimarilyonfactorsthatunderliethesereporteddevelopments.2Inparticular,weexplorehowisitthatintheworld’sthird-largesteconomy,andinoneconsideredtobeamongthemostcompetitiveinmanufacturing,softwareintensiveproductsare,toalargeextent,increasinglynotgloballycompe-titive.Japaneseexecutives,governmentofficials,andeducationalleadersgrosslyunderestimated

theimportanceofsoftwareforcompetitivenessacrossawiderangeofindustriesandproductsinboththeconsumerandenterprisearenas.

Thereisalessonhereforallnationaleconomiesabouttheimportanceofinstitutionalleadersnotonlysensingandmonitoring,butthenactingonemer-genttechnologicalshifts,lesttheyloseout.Tobesure,itisnotalwayseasytodis-tinguishbetweenwhatmightappearasatechnologicalfadorminordevelopmentandtheearlysignsofaseismicshift.Moreover,itiseasyforinstitutionalleaderstobecaughtupintheirexistingtechnologyvisions,whichcontinuetoyieldposi-tiveifdecliningrevenuestreams.Japaneseleaders,withtheirstrongsuccessinmanufacturinghardware,founditmuchmoredifficulttoenvisionsoftwareasafullpartner,muchlessalternativemodel.ItisalsoaplausiblehypothesisthattheweaknessofAmericanmanufacturingglobally,relativetoJapanesefirms,gaveAmericanfirmsstrongerincentivestosearch,sense,monitor,andrespondtothenewopportunitiescreatedbysoftware.Inourconclusion,wewillengageinasimplecounterfactualanalysisbyconsideringwhatJapanesefirmsmighthavedonedifferentlytoincreasetheprobabilityofbettercapturingthereturnsassociatedwiththegrowingimportanceofsoftware.Ouraimistoprovideguid-ancetothosepolicymakerseverywherewhoseektobuildtheireconomy’sITcapabilities.

JapaneseITfirmshavebeensteadilylosingglobalshareinmajorconsumerproducts.Notonlyhavetheyfailedtocompetegloballyinthemosttransformationalconsumerproductslikesmartphonesandtablets,buttheyhavesufferedstronginroadsintheirdomesticmarketinjusttheseproducts.Moreover,collateralcapa-bilitiesandfeaturesoftheseproducts,includingtheriseofmobilegamesandcamerasmartphones,aresavagingtheirpreviouslysuccessfulglobalfirmsinconsole-basedgamesoftwareandpoint-and-shootcamerasandmorerecentlyappeartobeerodingsalesofsophisticatedSLRcameras.

Japan’selectronicproducts’outputproducedinJapan,muchofwhichcontainsembeddedsoftware,declined50%from2000to2011.Electronicexportsaredown37%inthatsameperiod.Japaneseelectronicsfirmsareincreasinglyengagedinintermediategoodstransactionsproducingpartsandcomponentsforothermanufacturers.Consumerelectroniccomponentsanddevicesaccounted

CALIFORNIAMANAGEMENTREVIEWVOL.57,NO.1FALL2014CMR.BERKELEY.EDU17

TheJapaneseSoftwareIndustry

for45%ofproductionin2000,risingto56%by2011;inthesameperiod,industrialcomponentsrosefrom8to12%.3Theproblemisthatthebulkofprofitsforthesekindsofproductstypicallygotosellersoffinalproducts,theApple’sandSamsung’softheworld;whilepartsandcomponentssuppliers(exceptingthosewithstrongproprietarytechnology,includingsystemintegrationcapabilities,and/orstrongIPprotection)aremostlikelytobeleftwiththecrumbs.4Thecombinedmar-ketcapitalizationofJapaneseelectronicsmanufacturerswasmorethanhalvedbytheendofDecember2011fromitsyear2000level.Noristhisdeclinesimplyareflec-tionoftheoveralldeclineinmarketcapitalizationduringthisperiod;thefirstsection(largefirms)oftheTokyoStockExchangedeclinedjust16%overthesameperiod.5EvenwhereJapanesetechnologyfirmsretainstrongsalesglobally,theirprofitmargins,alreadylowcomparedtocompetitors,havefallenevenlowerinrecentyears,manyinthe1-2%rangeby2009.6Thatmeansthatlessfundsbecameavailableforfuturecapitalinvestmentrelativetomanyoftheirforeigncompetitors.Towhatextenttherecentlyweakenedyencanreversethistrendlongtermisanopenquestion.

Morerecentlytoescapethecarnageinconsumerelectronics,Hitachi,Toshiba,andMitsubishiElectrichaveshiftedtheirprimarybusinessfocustolargesocialinfrastructureprojects(energy,transportation,andutilities).Fortheseprojects,thesefirmsseesoftwaremoreasaglueholdingdifferentprojectcomponentstogetherratherthanasprimaryvaluecreators.Yet,itisonlyamatteroftimebeforesomecompetitorswithstrongsoftwarecapabilities,likeSiemens,buildcompetitiveadvantageintheseinfrastructuremarketsbyalsousingsoftwaretodrivefront-endfunctionalitiesandservices.

TheITsectorismorethanelectronics.ITproductsareincludedintheInfor-mationandCommunicationTechnology(ICT)sector.AccordingtoWorldBankdata,servicesaccountedfor5%ofglobalICTvaluein1996butroseto22%by2009.MostofthevalueofICTservicesiscontributedbysoftware.Yet,adisturbingportentforJapan’sfuturecompetitiveperformanceisthatservicesaccountedforalittlelessthan2%ofJapan’sICTexportsin2009.7Clearly,Japanislagginginthehardwaretoservicestransitionandalargepartoftheweaknesscanbeattributedtoaweaknessinsoftware.AnadditionalfactorintheirweakexportofICTservicesisthatproductcustomizationandsystemintegrationareoftendoneinthelocallanguagebasedonknowledgeoflocalbusinesspractices.ItisplausiblethatJapanesefirmshavebeenatadisadvantageintheirserviceexporteffortsbecauseofthetypicallyweakEnglishlanguagecapabilitiesofJapanesemanagement.

ConsistentwiththesejustmentionedobservationsarelargeJapanesesoft-wareandinformationservicesimports,andapaucityofpath-breakingsoftwarestartups,indeedanabsenceofalargeandvibrantindependentsoftwaresector.Theminisculesizeofsoftwareexportsandrelianceonhighlevelsofsoftwareimportsisassociatedwiththeweaknessoftheirdomesticpackagedsoftwaremarket,withthegreatmajorityofapplicationsbeingone-of-a-kindcustomizedsolutions(seeFigure1).8Eighty-sixpercentoftotalJapanesesoftwaresalesintheinformationservicesmarketin2012wereaccountedforbycustomsoftwarewhereasonly14%camefromsoftwareproducts.Notably,gamesoftwareaccountedfor37%oftotalsoftwareproductsalesor5%oftotalsoftwaresales.

18UNIVERSITYOFCALIFORNIA,BERKELEYVOL.57,NO.1FALL2014CMR.BERKELEY.EDU

TheJapaneseSoftwareIndustry

FIGURE1.InformationServicesMarketinJapan(2012)*

Totalinformationservicessales=$124.94billion**(100%)

Ofwhichtotalsoftwaresales=$89.66billion(71.7%)

Ofwhichcustomsoftwaredevelopment=$77.07billion(86%)

Ofwhichsystemintegrationdevelopment=$49.02billion(64%)

Ofwhichtotalsoftwareproducts=$12.59billion(14%)

OfwhichGameSoftware=$4.66billion(37%)

InformationProcessing$8.51billion(06.8%)

SystemManagersonCommission$16.66billion(13.3%)

DatabaseServices$1.72billion(01.4%)

VariousSurveys$1.72billion(01.4%)

Others$6.67billion(05.3%)

*Thecomputerprogrammingandothersoftwareservicescoveredinthissurveyarefrombusinessestablishments,whichareinvolvedinservicessuchasproduction,surveys,analysis,andadviceconcerningcomputerprograms[made-to-ordersoftware,softwareproducts(e.g.,packagedsoftwareforbusinessuseorsoftwarepackagedinabox),embeddedsoftware,andsoftwareforgames.]

**U.S.$1=￥81(averagerateofexchangein2012).

Source:MinistryofEconomy,TradeandIndustry,“TokuteiSabisuSangyouDoutaiTokeiGeppou[MonthlyReportontheCurrentSurveyofSelectedServiceIndustries,InformationServiceIndustry],”Tokyo,2013.

Notwithstanding,Japan,theworld’sthird-largesteconomy,issecondonlytotheU.S.insoftwaresales.Itsembeddedsoftwarecapabilitiesareimpressive.Machinetools,robotics,andautosarejustthreeofthemajoroutletsfortheirhigh-qualitygloballyrecognizedembeddedsoftwareproducts.IndependentresearchranksJapanesesoftwarequality(asmeasuredindefectsoverawiderangeofdifferentkindsofsoftwareprojects)andtheproductivityofthesoftwaredevelopmentprocess(codeproductivity)atlevelsfarabovethesampledAmericanprojects.9Theseareimpressiveaccomplishments.

Itisthegapbetweentheirstrongprocesscapabilitiesinsoftwaredevelop-mentandtheirweaknessinproductinnovationasreflectedinthelackofaglobalpresenceinsoftwareproductsandservices,whichliesattheheartofthisanalysis.MichaelCusumanotalksaboutthisjuxtapositionintermsof“thepuzzleofJapanesesoftware.”10

RecentPatentStudyFindings

Inarecentarticle,Arora,Branstetter,andDrevreportedfindingsfromtheirstudyofU.SandJapaneseITpatentsgrantedbytheU.S.PatentOffice

CALIFORNIAMANAGEMENTREVIEWVOL.57,NO.1FALL2014CMR.BERKELEY.EDU19

TheJapaneseSoftwareIndustry

(USPTO)from1983-2004.11TheauthorsexaminedU.S.andJapanesenon-hardwarepatents(ineffect,embedded-application,system,andenterprise-applicationsoftwarepatents)inthehigh-techsector.TheyfoundthatdespitethegrowthincitationofsoftwarepatentsinbothJapanandtheU.S.,JapanesefirmsfilefewersoftwarepatentsthantheirU.S.counterpartsandthatthedifferencehasgrownsteadilysincethelate-1980sandespeciallyafterthemid-1990s.TheirfindingsindicatefirstthatITinventionshavebecomemuchmoresoftwareintensiveovertime.Second,U.S.firmshavemoreactivelyincorporatedsoftwareinventionsintotheirproductsandservicesthanhavetheirJapanesecounterparts.Third,theyfindthatU.S.firmshaveimprovedtheirinnovationperformancevis-à-visJapanesefirmsinjustthoseITsegmentsthataremostsoftwareintensive(computersandperipherals).Finally,theresearchersfoundthatwithinIT,Japanesefirmsarenowdisproportionatelylocatedinless-software-intensivesectors.

Theyconcludethattheirfindingsprovideanewexplanationforthe“precipitousglobaldecline”ofJapan’sITindustryinthe1990s.TheauthorsdiscusstwopossiblesourcesforJapan’sshortfallinsoftwareinnovation.First,theypointtoaconstrainedsupplyofsoftwareknowledgeandskills(humanresourceconstraints)andsecondtotheslownessofJapaneseITfirmstorecognizethetransformationalnatureofsoftware.TheyconcludethattheformerratherthanthelatterisprimarilyresponsibleforJapan’sgrowingshortfallinsoftwareinnovationrelativetotheU.S.

UniversityComputerScienceEducation

Webuildontheiranalysisandaddapiecetothepuzzlesolutionfirstbysur-veyingovertimeU.S.andJapaneseuniversityeffortstoprovideanewgenerationofstudentswiththeknowledge,analyticskills,andtoolstoaddresstherapidlyevolvingfieldofcomputerscience(CS).WhilewefocusonCS,ouranalysisalsocoversthebroadfieldofinformationtechnology.

Giventhecriticalityofskilledhumancapitaltosoftwaredevelopment,anexaminationofITsoftwareeducationallowsustopartiallyevaluateIThumancap-italintherespectiveeconomiesthroughthetimingandqualityoftheirsoftwareinvestments.12SteinmuellerarguesthattheU.S.softwareindustry’sadvantageoverJapanlayinitsfirst-moveradvantage,onethatwasenabledbygovernmentR&Dpolicy,andtheearlydevelopmentofCSeducationattheuniversitylevel.13ThelatteradvantageoverJapan,andindeedoverallotheradvancedindustrialnations,persiststothisday.

Some20%ofsoftwaredevelopersintheU.S.havesomeamountofgraduateschooleducationcomparedto10%inJapan.14ThegapinPhDsisevenlargerthansuggestedbythis10%difference.Untilrecently,themajorityofJapanesePhDsinengineeringwereearnedbyindustrialresearchersinJapanesecompanies.Thesedegrees(ronbunhakase)areawardedbytheemployees’formeruniversity,typicallyaftermanyyearsofR&Dresearch.Nocourseworkisnecessary,onlysubmissionofadissertationwithsomearticlespublishedinwell-regardedjournals.ThesePhDdegreeholdersareveryunlikelytobeontheforefrontofsoftwareinnovation.Puttingasidetheronbunhakasedegrees,mostCSPhDs(andmostJapaneseengineeringPhDsforthatmatter)areearnedbythoseaimingforan

20UNIVERSITYOFCALIFORNIA,BERKELEYVOL.57,NO.1FALL2014CMR.BERKELEY.EDU

TheJapaneseSoftwareIndustry

academiccareer.BehindthelimitednumbersofJapanesePhDsinsoftwarerelatedprivatesectorjobsisthelong-termlackofdemandbyJapaneseindustryforengi-neeringPhDsoverallandthepaucityofsoftwarestartups.

Bycontrast,mostCSPhDsfromU.S.universitiesarehiredbyindustryorgovernment.15AgreatdealofnewvalueintheprivatesectorhasbeenandisbeingcreatedintheU.S.byPhDandMastersofScienceengineers.Some22,000PhDshavebeengrantedinCSandITintheUnitedStatesbetween1978and2008.16PhDstudentsatthebestU.S.universitiesexperiencerigorouscourseworkandpracticumsandnotafewgoontoenvisionandthencreateandsustainawholenewworld,e.g.,UNIX,relationaldatabases.StartupsfoundedbythosewithPhDsincludeAdobe,Qualcomm,Google,Sun,Cadence,Synopsys,VMware,andSymantec.StillotherPhDslikeEdgarCodd,withaPhDinCS,workedatIBMbutinspiredstartupfirmssuchasOracletodevelopcommercialproductsbuildingonhispathbreakingresearchonrelationaldatabases.PhDcontributionsarewheretheAmericanshaveexcelled.OneishardputtocompileacomparablelistofpathbreakingequivalentsinJapan.

Notablealsohasbeenthelong,slowprocessofincorporatingstateoftheartsoftwareknowledgeintothecurriculumofJapaneseFacultiesofScienceandEngineering.ItwaswellrecognizedbyCSprofessorsinJapaninthe1990sthatCSfacultiesinJapaneseuniversitieswereweak.17TheyreliedheavilyonU.S.createdstandardsofIEEE-CS(InstituteofElectricalandElectronicsEngineers-ComputerScience)andACM(AssociationofComputingMachinery)forsuccessivecurriculumupgrades,butwithlongdelayseachtimerightupto2007.Sixyearswasthetypicalgap.Moreover,thestandardswereoftenoutdatedevenwhentheywereadoptedintheU.S.18ThislagsuggestsalackofappreciationbytheJapaneseeducationalestab-lishmentofthespeedatwhichITtechnologywaschangingandtheimportanceofquicklyaccommodatingtothesechanges.

Criticaltotheincorporationandcommunicationtostudentsofup-to-datecurriculumisfacultywithup-to-dateknowledge.Here,Japaneseuniversitieswereatadisadvantage.MuchnewITknowledgewasbeingdevelopedintheU.S,espe-ciallywiththeadventofthePCeraintheearly1980s.Atthistime,JapaneseITfirms,whichhadachievedsignificantcompetitivesuccessthroughmodelingIBM’smainframearchitecture,remainedcommittedtobuildingnewgenerationsofstilllargermainframecomputers.19ThismisplacedfocusputthemstillfurtherbehindtheU.S.indeveloping,integrating,anddeployingnewsoftwareknowledge.ThemodestnumberofJapanesestartups,moreopentonewknowledge,ledtolostopportunitiestoclosethegap.

TheexpansionofITdepartmentsinJapaneseuniversitiesinthe1980swasassociatedwithashortageofwell-trainedCSandinformationengineeringfaculty.Tofillpositions,manyuniversitieshired“retired”ITexecutivesfrommajorcompanieslikeNEC,Fujitsu,Hitachi,andToshiba.Thesenewacademicshelpedfilloutthefac-ultyinITdepartmentsinthe1980sand1990s.WhiletheyhadpracticalexperienceinITbusinessissuesandsometechnicalmatters,mostoftheircorporatetrainingandexperiencewasinmainframehardwareandsoftware.IntheU.S.,mainframeswerequicklybeingdisplacedbyPCs,workstations,newerprogramminglanguages,networkedsystems,theinternet,andthelike.

CALIFORNIAMANAGEMENTREVIEWVOL.57,NO.1FALL2014CMR.BERKELEY.EDU21

TheJapaneseSoftwareIndustry

MostoftheJapanesefacultywiththesecorporatecareers,werenotcompe-tenttoteachuniversitystudentsstate-of-the-artsoftware.Norweremostofthosewithcorporatebackgroundscapableofmakingcontributionsontheresearchfrontatatimewhenresearchscientists,mostlyfromtheU.S.,weretransformingthefieldofCS.20AllthiscontributedfurthertotheJapanese,andothernations,fallingfurtherbehindtheU.S.insoftwareinnovationduringtheseformativeyears.

ProfessorNobumasaTakahashi,akeyplayerintheInformationProcessingSocietyofJapan,analyzedtheNationalUniversities’activities,astheytypicallysetthedirectiononnewtechnicalcurriculumforallotheruniversities.HeobservedthattheJapanesepostwaruniversitydepartmentalstructureswereshapedbycontinualexpansioninthefieldsofcivilengineering,machinery,electricalengi-neering(EE),andchemistry.Inmanycases,newdepartmentswereformedfromtheMinistryofEducation’sbudgetaryappropriationspushedbytheseexistingdepartments.“Asaresult,thenewdepartmentswerecreatedwiththestrongcol-orationofcoloniesoftheolddepartments.”21

Thenewdepartmentsofinformatics(jouhougaku),roughlyequivalenttowhatAmericanscallinformationscienceandincludingboththecoreofCSaswellascontentfromrelatedscientificdisciplines,weresubjecttostillotherconstraints.Thesearosefromanunanticipatedinteractionoftwodevelopments.In1991,adefactoMinistryofEducationderegulationofuniversitiestookplacethatnolongerrequiredscienceandengineeringschoolstoofferliberalartsforthefirsttwoyearsofundergraduateeducation.Thus,manyofthescienceandengineeringschoolssubstitutedengineeringandsciencecoursesintheirstead.Asaconsequence,alargenumberofredundantfacultywerecreated.Sincetheycouldnotbefired,theyneededtoberelocatedfromtheiroriginaldepartments,adilemmathatfounda“fortuitous”solution.

Atroughly,thesametime,theMinistry,concernedaboutdecliningnationalbirthrates,imposedlimitsontheexpansionofuniversitydepartmentsandfaculties.Theymadeafewexceptions,however,forgrowingfieldsofnationalimportance,oneofwhich,totheircredit,wasforinformationtechnology.Theuniversitiestookadvantageofthisexception,however,to“offload”manyoftheaforementionedredundantfacultytothesegrowingfields.ManyofthetransferredfacultycontinuedtoteachtheirexistingspecialtiesdespitenowbeinginCSandrelateddepartments.

Prof.MasatoTakeichifoundthatthetotalfaculty(2,615)oftheJapaneseinformationsciencedepartmentsinSchoolsofScienceandEngineeringin1998wereconstitutedroughlyasfollow:only30%werecoreCSfacultywhoseacademicspecialtieswereinCSandmathematicalinformatics.Viewinginformationsciencemorebroadlytoincludesysteminformatics,electricalinformatics,andintelligentinformaticsaddedanother45%.Aremarkable25%,however,hadeducationalspecializationswithlittleornothingtodowithCS.22SoalthoughtheMinistryrecognizedtheimportanceofthenewdisciplinesforJapan’sfuturebydesignat-ingthemagrowtharea,theythenignoredtheblatantsubversionofthenewfieldinthefillingofthenewfacultyslots.Asaresult,thenewdepartmentswereweigheddownbynon-contributingCSfaculty.Inturn,thismeantthatstudentshadfarlessaccesstoup-to-dateknowledgeinCS.Moresubtly,itmadeitmoredifficultforCStoberecognizedasadistinctivescientificdiscipline.23

22UNIVERSITYOFCALIFORNIA,BERKELEYVOL.57,NO.1FALL2014CMR.BERKELEY.EDU

TheJapaneseSoftwareIndustry

Writinginthemid-1990swithprimaryfocusonmainframesoftware,Baba,Takai,andMizutaobservethatcomputerscienceeducationisnotwellacceptedinJapanwithmostsoftwarepersonnelacquiringtheirprofessionalskillsthroughon-the-jobtraining.Theyfurthernotethat,consistentwithheavycustomization,knowledgeintheJapanesesoftwareindustry,unlikeintheU.S.,tendstobelocallyaccumulatedandexploitedprimarilywithintheboundariesofafirm.24AsonereadstherecentadmonitionsoftheInformationTechnologyPromotionAgency(IPA),asubsidiaryoftheMinistryofEconomyTradeandIndustry(METI),itisnothardtoseethatthesecharacteristicspersisttothisdayasmajorbarrierstosoftwareinnovation.Prof.FumihikoKimura,amemberoftheScienceCouncilofJapan,atapublicforumheldattheUniversityofTokyoinMarch2013,statedthatCS(calledjouhoukagakuinEngineeringSchools)isstillnotrecognizedbymostinterestedpartiesinJapanasaclearlydefineddiscipline.25

Onemeasureofanewfield’scoalescenceintoarecognizeddisciplineisthewillingnessofthoseinestablisheddisciplinestocontributetotheemergingdiscipline.Suchcontributionsbothreflectpositiveassessmentsoftheemergingdisciplineandbestowfurtherlegitimacyonthenewfield.InthecaseoftheU.S.,boththeoreticalandappliedmathematiciansplayedakeyroleinbuildingthenewCSdiscipline.Mathematicallogic,thetheoremsofTuringandGodel,Booleanalgebraforcircuitdesign,andalgorithmsforsolvingequationsandotherclassesofproblemsinmathe-maticsplayedstrongrolesintheearlydevelopmentoftheCSasadiscipline.26Bycontrast,Prof.MasatoTakeichi,DeanoftheGraduateSchoolofInformationScienceandTechnologyattheUniversityofTokyofrom2004-2007,reportsthatmathematicianshavenotbeenasdrawntothenewITfieldsinJapanaswasthecaseintheU.S.HeattributesthistoCSnotbeingseenasadistinctdisciplineinJapan.27Similarly,informationengineeringintheFacultyofEngineeringtypicallyhasrelativelylowstatusinthehierarchyofengineeringrelatedfieldsinJapan.

Overall,some63,300U.S.CSandmathematicsmajorsreceivedbachelordegreesin2009comparedtosome16,300inJapan.28ThisisamuchlargergapthanwouldbeexpectedgiventhatthatJapan’stotallaborforceissome43%thatoftheU.S.anditsnominalGDPisroughly37.5%thatoftheU.S.

AtleadingU.S.engineeringschools,thosedepartmentslikeCS,whichexpe-riencesharpgrowth,drawingonahighlyqualifiedapplicantpool,havethediscre-tionandstrongincentivestogrowtheirdepartment.Facultycanbemotivatedto“puttheirbrandon”anewgenerationofsmartyoungpeopleandtheyalsousegrowthtoclaimgreaterdepartmentalresources.ThehistoricallydecentralizededucationsystemintheUnitedStatesmeansthattheDepartmentofEducation,unlikeitscounterpar