Citation
The Offshoring of Innovative Research and Development to the Developing World

Material Information

Title:
The Offshoring of Innovative Research and Development to the Developing World
Creator:
Cohle, Zachary H
Place of Publication:
[Gainesville, Fla.]
Florida
Publisher:
University of Florida
Publication Date:
Language:
english
Physical Description:
1 online resource (134 p.)

Thesis/Dissertation Information

Degree:
Doctorate ( Ph.D.)
Degree Grantor:
University of Florida
Degree Disciplines:
Economics
Committee Chair:
DINOPOULOS,ELIAS
Committee Co-Chair:
SANDOVAL GUTIERREZ,HECTOR HUGO
Committee Members:
SAPPINGTON,DAVID
LEE,GWENDOLYN K

Subjects

Subjects / Keywords:
innovation -- ipr-protection -- mobility -- multinational
Economics -- Dissertations, Academic -- UF
Genre:
bibliography ( marcgt )
theses ( marcgt )
government publication (state, provincial, terriorial, dependent) ( marcgt )
born-digital ( sobekcm )
Electronic Thesis or Dissertation
Economics thesis, Ph.D.

Notes

Abstract:
Conducting R and D in a foreign country is an established practice among multinational firms. While multinational firms from developed countries have used researchers from emerging areas to assist in the adaption of an existing product, few multinational firms have carried out innovative R and D, or R and D for the creation of a new product, in these areas. Using the threat of imitation and wage differences across regions relating to R and D, this study proposes a partial equilibrium model to explain the lack of innovative R and D in developing countries. I build a North-South model examining a single firm's choice of research locations in a number of different competitive environments. The model predicts that weak IPR-protection in developing countries does not necessarily explain the lack of Southern research. In some situations, reduced IPR-protection can even increase Southern research. Harsh competition resulting from information leaks coupled with weak IPR-protection can explain much of the lack of innovative research investment in the developing world. My model also predicts that smaller firms that require less innovative research locate their R and D in the North. Medium sized firms locate in both countries while the largest firms locate research in just the South. While many multinationals firms are reluctant to locate innovative R and D labs in developing countries, the practice has steadily grown over the past decade. I build a general equilibrium model to determine the welfare effects that R and D offshoring has on both developed and developing countries. Using two countries, a developed North and a developing South, I examine Northern firms offshoring research and development to the South in order to produce new varieties of a good. Using employee mobility as the channel of imitation, Northern firms risk their product being imitated by Southern firms who absorb Southern researchers. In high-tech industries with a high R and D intensity, firms offshore a large number of tasks to the South. For these high-tech industries, stronger IPR-protection in the South increases both industry-level and firm-level offshoring. A larger Southern market decreases offshoring to the South. The South gains from being able to place a tariff on Northern goods. I construct a dataset using U.S. patent data to examine the determinants of R and D offshoring. By using the number of inventors from a developing countries working on one patent, I capture the research intensity that firms are willing to have in developing regions. The study uses the NBER database along with WDI and NBER data to explore firm-level and industry-level offshoring. I use OLS and Tobit regression models in the study. In order to test the sensitivity of the results and address endogeneity concerns, I also employ a difference-in-difference model using years of major IPR-reform. Import tariffs do not affect firm-level offshoring, but a large and developed Southern market deters offshoring from a firm. Both high-tech firms and industries react more to changes in Southern IPR-protection than their low-tech counterparts. ( en )
General Note:
In the series University of Florida Digital Collections.
General Note:
Includes vita.
Bibliography:
Includes bibliographical references.
Source of Description:
Description based on online resource; title from PDF title page.
Source of Description:
This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Thesis:
Thesis (Ph.D.)--University of Florida, 2017.
Local:
Adviser: DINOPOULOS,ELIAS.
Local:
Co-adviser: SANDOVAL GUTIERREZ,HECTOR HUGO.
Electronic Access:
RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2017-11-30
Statement of Responsibility:
by Zachary H Cohle.

Record Information

Source Institution:
UFRGP
Rights Management:
Applicable rights reserved.
Embargo Date:
11/30/2017
Classification:
LD1780 2017 ( lcc )

Downloads

This item has the following downloads:


Full Text

PAGE 1

THEOFFSHORINGOFINNOVATIVERESEARCHANDDEVELOPMENTTOTHE DEVELOPINGWORLD By ZACHARYH.COHLE ADISSERTATIONPRESENTEDTOTHEGRADUATESCHOOL OFTHEUNIVERSITYOFFLORIDAINPARTIALFULFILLMENT OFTHEREQUIREMENTSFORTHEDEGREEOF DOCTOROFPHILOSOPHY UNIVERSITYOFFLORIDA 2017

PAGE 2

c 2017ZacharyH.Cohle

PAGE 3

Idedicatethistothetwopeoplewhosupportedandencouragedmethroughoutmyentire eduction:EdwardandDianeCohle.IloveandappreciatebothofyoumorethanIcanever possiblyexpress.

PAGE 4

ACKNOWLEDGMENTS Iwouldliketothankmychair,ProfessorEliasDinopoulos,forguidingmeandsupporting mewithbothmydissertationandmylifeingeneral.IalsothankProfessorStevenSlutsky, ProfessorDavidSappington,ProfessorGwendolynLee,andProfessorHectorHugoSandoval Gutierrezfortheirhelpfulcommentsandsupport.IamparticularlyindebtedtoProfessor StevenSlutskyforhistirelesseortstosecureresourcesforthegraduatestudents.Iwouldalso liketothankmyclassmatesforprovidingmotivation,entertainment,support,andfriendship. Finally,Iwouldliketothankmyfamily,myfriends,andmypartnerforbelievinginme whileIpursuedmypassion. 4

PAGE 5

TABLEOFCONTENTS page ACKNOWLEDGMENTS...................................4 LISTOFTABLES......................................7 LISTOFFIGURES.....................................8 ABSTRACT.........................................9 CHAPTER 1MARKETSTRUCTUREANDRESEARCHANDDEVELOPMENTOUTSOURCING11 1.1Introduction...................................11 1.2RelatedLiterature................................13 1.3TheModel....................................17 1.3.1SetUp..................................17 1.3.2Equilibrium................................22 1.4Results......................................27 1.4.1 R `sEectonRandDLocations....................27 1.4.2IPR-Protection`sEectonRandD...................28 1.4.3MarketStructureandWage`sEectonRandD............31 1.5MarketStructureEects.............................34 1.5.1BetrandwithUndierentiatedProducts.................35 1.5.2CournotwithUndierentiatedProducts.................36 1.5.3WageandMarketSizeEectsonRandD...............36 1.6Conclusion....................................38 2INNOVATIVERESEARCHANDDEVELOPMENTOFFSHORINGINNORTH-SOUTH TRADE........................................41 2.1Introduction...................................41 2.2RelatedLiterature................................43 2.3TheModel....................................45 2.4IndustryEquilibrium...............................51 2.4.1Researchlevels..............................51 2.4.2Prices...................................52 2.5GeneralEquilibrium...............................54 2.6ComparativeStatics...............................56 2.6.1IndustryTypeandFirmSize.......................57 2.6.2IPR-Protection..............................59 2.6.3SouthernLaborForce..........................62 2.6.4SouthernImportTari..........................64 2.7Conclusion....................................65 5

PAGE 6

3INNOVATIVERESEARCHANDDEVELOPMENTOFFSHORINGTODEVELOPING COUNTRIES:EVIDENCEFROMUSMULTINATIONALFIRMS..........67 3.1Introduction...................................67 3.2RelatedLiterature................................69 3.3TheModel....................................71 3.3.1AnalyticalFramework..........................71 3.3.2EstimationMethodology.........................74 3.4Data.......................................76 3.4.1DependentVariable............................76 3.4.2IndependentVariable...........................79 3.5EmpiricalResults.................................80 3.5.1Firm-LevelResults............................80 3.5.2Industry-LevelResults..........................85 3.5.3SensitivityTests.............................90 3.6DiscussionandConclusion............................94 APPENDIX AAPPENDIXFORCHAPTER1............................98 A.1EquilibriumProt................................98 A.2Including asanIPR-ProtectionParameter...................98 A.3ProofofInteriorSolutionParameterConditions.................100 A.4ProofOfResult2................................101 BAPPENDIXFORCHAPTER2............................103 B.1Consumers....................................103 B.2Producers....................................107 B.2.1FirstStage................................107 B.2.2SecondStage...............................109 B.2.3FreeEntryConditions..........................110 B.2.4PriceIndexandNumberofFirms....................113 B.2.5AlternativeEquilibria...........................116 B.3ComparativeStatics...............................117 B.3.1ProofofProposition1..........................117 B.3.2ProofofProposition2..........................120 B.3.3ProofofProposition3..........................124 B.3.4ProofofProposition4..........................125 REFERENCES........................................129 BIOGRAPHICALSKETCH.................................134 6

PAGE 7

LISTOFTABLES Table page 3-1SummaryStatisticsForFirm-LevelRegressions....................80 3-2SummaryStatisticsForIndustry-LevelRegressions..................81 3-3High-TechFirms,HostCountryPatentsLaws,andNumberofInventorsOshored, Firm-Level.......................................82 3-4LargeFirm,HostCountryPatentsLaws,andNumberofInventorsOshored,Firm-Level83 3-5High-TechFirms,HostCountryPatentsLaws,andRandDIntensityinHostCountry, Firm-Level,TobitRegression.............................85 3-6LargeFirms,HostCountryPatentsLaws,andRandDIntensityinHostCountry, Firm-Level,TobitRegression.............................86 3-7High-TechIndustry,HostCountryPatentsLaws,andRandDIntensityinHostCountry, Industry-Level.....................................88 3-8ParentRandDExpenditures,HostCountryPatentsLaws,andRandDExpenditures inHostCountry,Industry-Level............................89 3-9Di-in-Di:High-TechIndustry,PatentsReformYear,andRandDIntensityin HostCountry,Firm-Level...............................91 3-10Di-in-Di:LargeFirm,PatentsReformYear,andNumberofInventorsOshored, Firm-Level.......................................92 3-11Di-in-Di:High-TechIndustry,PatentsReformYear,andRandDExpenditures inHostCountry....................................93 3-12ParentRandDOshoringwithPreandPostReformDummies,Industry-Level...95 7

PAGE 8

LISTOFFIGURES Figure page 1-1OptimalResearchRatios................................26 1-2WeakeningofIPR-protectionyieldingadecreasein R S ................30 1-3WeakeningofIPR-protectionyieldinganincreasein R S ...............31 1-4Anincreaseinthepenaltyofimitation........................33 2-1TradingEquilibrium..................................57 2-2AnIncreasein R ....................................58 2-3AStrengtheningofIPR-ProtectionwithaLarge R ..................60 2-4AnIncreaseintheSouthernLaborForce.......................62 2-5IntroducingaSouthernImportTari.........................64 B-1NoSouthernFirms...................................115 B-2NoDierentiatedGoods................................115 8

PAGE 9

AbstractofDissertationPresentedtotheGraduateSchool oftheUniversityofFloridainPartialFulllmentofthe RequirementsfortheDegreeofDoctorofPhilosophy THEOFFSHORINGOFINNOVATIVERESEARCHANDDEVELOPMENTTOTHE DEVELOPINGWORLD By ZacharyH.Cohle May2017 Chair:EliasDinopoulos Major:Economics ConductingresearchanddevelopmentRandDinaforeigncountryisanestablished practiceamongmultinationalrms.Whilemultinationalrmsfromdevelopedcountries haveusedresearchersfromemergingareastoassistintheadaptionofanexistingproduct, fewmultinationalrmshavecarriedoutinnovativeRandD,orRandDforthecreation ofanewproduct,intheseareas.Usingthethreatofimitationandwagedierencesacross regionsrelatingtoRandD,thisstudyproposesapartialequilibriummodeltoexplainthe lackofinnovativeRandDindevelopingcountries.IbuildaNorth-Southmodelexamininga singlerm`schoiceofresearchlocationsinanumberofdierentcompetitiveenvironments. ThemodelpredictsthatweakIPR-protectionindevelopingcountriesdoesnotnecessarily explainthelackofSouthernresearch.Insomesituations,reducedIPR-protectioncaneven increaseSouthernresearch.Harshcompetitionresultingfrominformationleakscoupled withweakIPR-protectioncanexplainmuchofthelackofinnovativeresearchinvestmentin thedevelopingworld.Mymodelalsopredictsthatsmallerrmsthatrequirelessinnovative researchlocatetheirRandDintheNorth.Mediumsizedrmslocateinbothcountrieswhile thelargestrmslocateresearchinjusttheSouth. WhilemanymultinationalrmsarereluctanttolocateinnovativeRandDlabsin developingcountries,thepracticehassteadilygrownoverthepastdecade.Ibuildageneral equilibriummodeltodeterminethewelfareeectsthatRandDoshoringhasonboth developedanddevelopingcountries.Usingtwocountries,adevelopedNorthandadeveloping 9

PAGE 10

South,IexamineNorthernrmsoshoringresearchanddevelopmenttotheSouthinorderto producenewvarietiesofagood.Usingemployeemobilityasthechannelofimitation,Northern rmsrisktheirproductbeingimitatedbySouthernrmswhoabsorbSouthernresearchers.In high-techindustrieswithahighRandDintensity,rmsoshorealargenumberoftasksto theSouth.Forthesehigh-techindustries,strongerIPR-protectionintheSouthincreasesboth industry-levelandrm-leveloshoring.AlargerSouthernmarketdecreasesoshoringtothe South.TheSouthgainsfrombeingabletoplaceatarionNortherngoods. IconstructadatasetusingU.S.patentdatatoexaminethedeterminantsofRand Doshoring.Byusingthenumberofinventorsfromadevelopingcountriesworkingon onepatent,Icapturetheresearchintensitythatrmsarewillingtohaveindeveloping regions.ThestudyusestheNBERdatabasealongwithWDIandNBERdatatoexplore rm-levelandindustry-leveloshoring.IuseOLSandTobitregressionmodelsinthestudy. Inordertotestthesensitivityoftheresultsandaddressendogeneityconcerns,Ialsoemploy adierence-in-dierencemodelusingyearsofmajorIPR-reform.Importtarisdonotaect rm-leveloshoring,butalargeanddevelopedSouthernmarketdetersoshoringfromarm. Bothhigh-techrmsandindustriesreactmoretochangesinSouthernIPR-protectionthan theirlow-techcounterparts. 10

PAGE 11

CHAPTER1 MARKETSTRUCTUREANDRESEARCHANDDEVELOPMENTOUTSOURCING 1.1Introduction Innovativeresearchanddevelopment,orRandDforthecreationofanewproduct,has largelybeenkeptinthedevelopedworld.WhiletwoofthelargestattractionsforRandD oshoringinthedevelopingworld,IndiaandChina,largelyremainsitesforadaptiveresearch,a fewrmsintheforefrontoftechnologyintheirrespectiveindustrieshavestartedlocatingmore advancedtypesofresearchintheseregionsLundinandSeger,2007;Walsh2007;Atkinson 2007.Forexample,surveyevidencefromSun,Du,andHuangfoundamajorityof foreignRandDlabsinShanghaiengageinadaptiveresearchanddevelopment;however,GE isamongthefewmultinationalrmsconductinginnovativeRandDinthecity.Demirbag andGlaisterndevidencethatmoreadvancedresearchtaskstendtogetoutsourced todevelopedcountries.Regionswithhigherknowledgeprolesseemoreadvancedresearch projectsassignedtothemwhencomparedtoclustersofdevelopingcountries.Barefoot notesthatover60 % ofoutsourcedRandDtasksoutoftheUnitedStatesin2010wentto technicallyadvancedEuropeancountries.Only0.6 % ofthemoneyspentonoutsourcingRand DwenttoIndia,oneofthelargestrecipientofRandDoutsourcinginthedevelopingworld. Usinglaborcosts,intellectualpropertyrightsIPRprotection,andthethreatofimitation fromacompetingrm,IexplaintheinnovativeRandDlocationdecisionsofrmsacross dierentcountries.Mymodelexplainsthelimitedamountofmultinationalrminnovative RandDinvestmentinthedevelopingworld.UsingaNorth-Southmodel,Iexamineasingle rm,theNorthernrm,andtherm'sdecisiontolocateinnovativeRandDineitherthe NorthortheSouthinthecontextofanumberofdierentcompetitiveenvironments.The Northernrmhasaxednumberofresearchtasksthatmustbecompletedinordertodevelop anew,uniqueproduct.ResearchersintheSouthareexactlyascapableasresearchersinthe North;however,Southernresearchersearnasignicantlylowerwage.Iuseapartialequilibrium modelinwhichtheNorthernrmrstchoosesresearchlocationstocarryoutitsxedamount 11

PAGE 12

ofnecessaryresearch.Acompetingrm,theSouthernrm,hasthechancetoabsorbsome SouthernresearchersinordertoemulatetheNorthernrm'sproduct.Afterthexedcost ofentryhasbeenpaid,theNorthernrmchoosesproductiondecisionsandeitheractsasa monopolistorcompetesagainsttheSouthernrm. Inexplainingthesmallamountofoshoringofinnovativeresearchtothedeveloping world,thismodeldistinguishesitselffrompastliteratureconcerningIPR-protectionthrough itstreatmentofinnovativeRandDasaxedcostofentry.SimilartoMelitz,Imodel RandDasanecessaryxedcostofentry;however,thewayinwhichthermpaysthexed costhasrealramicationsonproductionprots.FollowingLai,Riezman,andWang, IincludethenotionofinformationleakageswhenoutsourcingRandD.Iuseaprobabilistic viewofinformationleakagerelatedtotheintensityoftheresearchdoneintheSouth.My modelusesthenotionofSouthernresearchersdefectingtocompetingrmsasthechannel ofimitation.ResearchershiredintheSouth,unliketheirNortherncounterparts,canleaveto eitherstarttheirownrmorjoinacompetingrm.Furthermore,theprobabilityofSouthern researchersleavingasateamincreasesastheNorthernrmhiresmoreSouthernlabor.The probabilityoftheSouthernrmsuccessfullycopyingtheproductandenteringthemarketis modeledasaTullockcontest.InaTullockcontest,eachplayerputsforthacertainamountof eortinordertoattempttowinthecontest.Iintroduceaproxyforplayereort:thenumber ofresearchersemployedbytheNorthernrmineachcountry.Finally,IapplysimpleCournot andBertrandmarketstructuresinordertoshowthatresearchdecisionsdependonthenature ofcompetitionandthemarketparameters. Myprimaryndingsareasfollowed.First,arm'sresearchneedsdetermineitslocation decisions.SmallrmsthatrequirelittleresearchlocateintheNorth,whilelargerrmswith largeresearchneedslocateentirelyintheSouth.Intuitively,thermswithalargeneedfor researchanddevelopmenthavealargexedcostinordertoenterthemarket.Thesermsare morewillingtorisktheirproductbeingimitatedinordertoreducethexedcostofentryby hiringlowerwageresearchers.Firm'swithintermediateresearchneedslocateinbothcountries. 12

PAGE 13

AmainresultofthemodelistheambiguouseectofIPR-protectiononRandDlocation decisions.Smallerrms,orrmsthatrequirelesstotalRandD,respondtoaweakeningof SouthernIPR-protectionbylocatinglessresearchintheSouth.Theincreaseintheprobability ofimitationcausedbytheIPR-protectionreforminducestheNorthernrmtoavoidtheriskof imitationbylocatingmoreresearchintheNorth.Ontheotherhand,largerrmsrespondtoa weakeningofIPR-protectionbyshiftingmoreresearchtotheSouth.Armwithhighresearch needs,oralargexedcostofentry,reactstothedecreaseinexpectedrevenuebyfurther decreasingexpectedrevenueinfavorofdecreasingthexedcostofentry.Thisreductionof xedcostminimizesthelossinexpectedprot.Thenalndingrelatespossibleoperating protandresearchlocations.Alargedierencebetweenmonopolyandduopolyprotinduces theNorthernrmtolocatemoreresearchtasksintheNorth.Imitationleadstoadecreasein operatingprotduetoincreasedcompetition.Whenthedecreaseislarge,theNorthernrm ndsitmoreprotabletoprotecttheirpatentthantocutcostsintheRandDphase.Holding thepenaltyofimitationconstant,alargewagegapbetweentheNorthandSouthencourages Southernresearch;however,thelargewagegapmightalsoimplyalargepenaltyofimitation. Thischapterisorganizedasfollows.Section1.2summarizesrelatedliterature.Section 1.3presentstheassumptions,functionalformsofthemodel,andthepartialequilibrium solutions.Section1.4presentstheresultsofthemodel.Section1.5introducesdierent marketstructurestothemodel.Section1.6concludesthechapter. 1.2RelatedLiterature EndogenousgrowthliteraturetypicallyassumesallinnovativeRandDiscarriedoutin theNorthdespitethegrowingtrendofresearchbeingoutsourcedtotheSouth.Ekholmand HakkalaexploredendogenousRandDlocations;however,theiranalysisexamined theknowledgeproleofacountryandnotIPR-protection. 1 Intheirmodel,thecapabilityof 1 Oshoringofinnovativeresearchinthe90'swaslargelydrivenbythedesiretoreduce costsandtheneedtoincreasethespeedtomarkets;followingthisperiod,accesstoquality 13

PAGE 14

researchersinacountrygrowsasmorermsbeginresearchinginthatcountry.Basedonthe parametervaluesofthemodeli.e.thedegreeofknowledgespillover,RandDlocationsmay beconcentratedinonecountryorspreadoutacrossbothcountries.AlacerandZhao ndthattheleadingrmsinthehigh-techindustriestendtobethelargestandhaveRand Dlabsspreadthroughoutmultiplelocationsandcountries.Zhaoalsondsevidenceof researchtasksbeingsplitbetweendevelopinganddevelopedcountries. 2 Mymodelndsthat aNorthernrmwithlargerRandD,likethoseinthehigh-techsector,splitsRandDbetween countries. ItoandWakasugindthatparentrmswithhighRandDspendinghavea tendencytodecentralizetheirresearchtasksandthusdomoreresearchoverseas.Otherstudies haveshownthatlargeresearchneedscanresultinmoretasksbeingassignedtoresearchersin developingregions.DemirbagandGlaisterndthatasrms'researchneedsincrease theystartoshoringmoretoemergingregions.Usingauniquemeasureofintra-industry oshoringconstructedfrommeasuresofimports,homeconsumption,andinputvalue,Canals andSeneralsondrmsinthelow-techindustrieshavelessoshoringwhencompared totheirhigh-techcounterparts.Thisresultisconsistentwiththepresentmodel'sprediction thatrmswithsmallresearchneedsdonotoshoretheirresearchtotheSouth.Hedgeand Hicksalsoshowhigh-techindustriesengaginginthebulkofRandDoutsourcing.My modelalsopredictsthepossibilityofmediumsizedrmsassigningsomeresearchtaskstoboth theNorthandtheSouth.Consistentwiththepredictionofmydissertation,manystudieshave researchersbecomesthedrivingfactorLewin,Massini,andPeters,2009.Theknowledge proleseemstobelessofamotivatorascapableresearcherscannowbefoundinanumber ofemergingeconomies.Forexample,China'srstyeardoctoralstudentshaveincreasedbya factorofsixduringthatsametimeperiodLewin,Massini,andPeters,2009. 2 Zhao'sanalysissuggeststhatrmscanworkaroundweakIPR-protectionthroughinternal organization.Inotherwords,thevalueofaninformationleakfromaresearchlabinahost countryislessenedbythetechnology'sstrongrelianceonothertechnologyorknowledgethata rmhasgathered. 14

PAGE 15

foundapositiverelationshipbetweenrmsizeandRandD.Whilesomestudiesndthat innovationincreasesmorethanproportionatelywithrmsize,Symeonidissurveysall literatureonthemattertoconcludethatthegeneralconsensusisthatrmspendingonR andDincreasesaboutproportionatelywithrmsizegiventhermhassurpassedacertain thresholdofsize.MydissertationusestherequiredamountofRandDnecessaryforthe inventionofanewproductasaproxyforrmsize. Moncada-Paterno-Castelloetal.,DemirbagandGlaister,andItoand WakasugipresentevidenceofstrongIPR-protectiondrivinginvestmentdecisions forinnovativeRandD.CanalsandSenerlooksatrmsinhigh-techindustriesthat requirelargeamountsofinitialresearchinordertocreatenewproducts.Consistentwithmy model'spredictionofonlysignicantlylargerrmsevenbeinginuencedbyIPR-protection changes,theauthorsndevidenceofthesermsexpandingtheiroshoringactivitiesafter periodsofIPRreformwhilelow-techsectorsseemedunaectedbythestrengtheningof IPR-protection.EconomicliteraturehaslongexaminedtheeectofIPR-protectionon innovation.Previousliteratureconcerningendogenousgrowththeoryhasfoundconicting eectsofstrongIPR-protectionindevelopingcountries.TighterIPR-protectioncouldleadto increasesininnovationGustasonandSegerstrom,2011;Branstetteretal,2007;Dinopoulos andSegerstrom,2010;GlassandWu,2007;GustasonandSegerstrom,2011ordecreasesin theinnovationrateDinopoulosandSegerstrom,2007;Helpman,1993. Mymodelcontributestoendogenousgrowthliteraturebyintroducinganewmechanism forimitation:employeemobility.ResearchersintheSouthcanuseinformationobtainedwhile workingonaninnovationandproducetheirownversionoftheproducttocompetewiththe parentrm.Gambardella,Ganco,andHonorendapositiverelationshipbetween employeeentrepreneurshipandthevalueasassessedbythermand/ortheemployeeof aninvention.MorevaluablepatentswillthenberesearchedmoresointheNorthtoavoid employeemobility.Anumberofstudieshavealsofoundevidenceofinformationleakages 15

PAGE 16

throughemployeemobility. 3 Theriskofexposingnewproductdesignstofuturecompetition canbeosetbythecostadvantageofusingSouthernlabor.Mymodeluseslowwagesto explaintheallureofRandDlocationindevelopingcountriesasconsistentwiththendings ofLewin,Massini,andPeters,Moncada-Paterno-Castelloetal.,andDemirbag andGlaister.Alloftheseauthorsnoteadierenceinwagebetweenthehomeandhost countryasastrongpredictorofresearchlocation. Whilemydissertationfocusesontheoutsourcingofresearchtasks,itrelatesto outsourcingliteratureasawhole.AntrasandHelpmanuseaNorth-Southmodel withendogenousinputlocationsandoutsourcingdecisions.Theauthorsndthatindustries thatareinput-intensivehavermsonlyengaginginoutsourcinginbothcountrieswhile industriesthataremorebasedaroundNorthernheadquarterserviceshaveinputproduction frombothoutsourcingandverticalintegrationinbothcountries.WhileAntrasandHelpman ndproductivityoftheparentrmdrivesthedecisiontooutsourceintheSouth,mymodel introducesthenotionofrmsizeasafactorinvolvedinoutsourcingchoices.Grossmanand Helpmanalsouseamodelwithendogenousorganization;however,rmsmustsearch forotherrmswhoarewillingtoproduceanintermediateinput. RandDgameswithspilloversandendogenouspayoshavebeenexploredinthepast d'AspremontandJacquemin,1988;ChowdhuryandSheremeta,2010.Forexample,Kamien, Muller,andZanguseatheoreticalmodelwitharm`sspendingonRandDspilling overintothatrm`scompetitor.So,arm`sRandDexplorationmayinfactmakeatougher competitorforitself.UsinganasymmetricTullocklottery,mymodelbuildsinthisexternality byconnectingthecreationofacompetingrmwiththedecisiontoemployresearchersinthe 3 Ganco,usingaresearcher`spriorworkatthermasaproxyfortheknowledge internalizedwithintheresearcher,ndsevidencesuggestingresearcherswithahigherdegree ofinternalizedknowledgearemorelikelytobegintheirownrm.Knowledgecomplexityalso seemstobepositivelyassociatedwithanentireteamofresearchersstartingtheirownrm ratherthanindividualentrepreneurship.YangandJiangprovidealiteraturereviewof employeemobilityinRandDoutsourcingforemergingcountries. 16

PAGE 17

South;however,extraeortputforthbytheNorthernrminthecontesttopreventimitation willdirectlydecreasetheeortofthecompetingrminattemptingtocopytheproduct.This aspectofthemodeldiersfromthestandarduseofTullockcontestsinRandDgames. 4 1.3TheModel 1.3.1SetUp Consideraworldwithtworegions,NorthandSouth.AssumethatNorthernworkersenjoy ahigherwagethanSouthernworkers.Ifocusontheresearchlocationdecisionofasingle multinationalrmherebyreferredtoastheNorthernrm.Thismodelisnotconcerned withthiseconomyatlarge,onlyaselectrmanditscompetitor.Thetimingofthismodel isasfollows.First,theNorthernrmdiscoversthetotalamountofresearchneededtomake anewproduct.Then,thermchoosesthenumberofNorthernresearchers, R N ,andpays thexedcostsofentry.Finally,acompetingrmreferredtoastheSouthernrmenters themarketwithprobability ,andnallytheNorthernrmproducesandsellsitsproduct. ThistimingimpliesthattheNorthernrmhastwostagestomakedecisions.First,therm choosesresearchallocationsacrosstwocountries.Then,theNorthernrmfacesproduction relateddecisionsinordertomaximizeoperatingprots.SincetheNorthernrmobservesthe entryofitscompetitor,ithasnoneedtomaximizeexpectedoperatingprots.Instead,the Northernrmeitheractsasamonopolyifnoimitationoccursorengagesincompetitive 4 SkaperdasusesanaxiomaticapproachtoprovidesupportforusingtheCSF proposedbyTullockinanumberofrentseekingcontests.Tulluckcontestshavebeenapplied toR & D racesinthepastBayeandHoppe,2003.Leiningerexaminestheasymmetric Tullocklotteryfunctionsimilartotheoneusedinthisdissertation.Leiningerincludesascaling parameterontheeortofoneplayerinordertoreectdierentialadvantages.Baik andFonsecaexplorethepropertiesofasymmetricTullocklotteries;however,few authorshavelookedatTullocklotterieswhereasingleplayerchoosesbothplayers'eort levels.Jia,Skaperdas,andViadyaoutlinemanydicultiesinempiricallyexamininga contest.Specically,eorttendstobeunobservable.Seeingastheactualresearchersworking ontheprojectlargelydrivesuccessfulinnovation,Imodeleortasthenumberofresearchersin eachregion. 17

PAGE 18

behaviorifimitationoccursandthusentryoftheSouthernrmensues.TheNorthernrm maximizesexpectedprotbychoosingRandDlevelsforeachregion.Therearenoxedcosts ofproduction,onlyaxedcostofentry.Ifimitationoccurs,theNorthernrmcompeteswith theSouthernrm,yieldinganoperatingprotof .Likewise,noSouthernentryallowsthe Northernrmtoactasaglobalmonopolyandgain fromsellingitsproduct.Therefore, expectedprotscanbeexpressedas: E = )]TJ/F24 11.9552 Tf 11.955 0 Td [( + )]TJ/F38 11.9552 Tf 11.955 0 Td [(f e {1 Inordertobeabletoproducetheproduct,thermmustrstinvestininnovative researchanddevelopment.Researchiscomposedofavarietyoftasks,eachrequiringone unitoflaborfromeithertheNorthortheSouth.OneunitofNorthernlabor,oroneNorthern researcher,costs w N ,andoneunitofSouthernlaborcosts w S .So,beforethermisableto enterthenalgoodsmarket,itmustrstinvestinRandDandcombinethediscoveriesto createanewproducttosell.Therefore,theentirecostofRandDandthecreationofthenew productcanbeseenasaxedcostofentryfortherm. 5 f e = w N R N + w S R S 0 < b < 1 WhiletheNorthernrmcancutcostsbyusingSouthernlaborforresearch,aSouthern competitorimitatestheproductwithaprobability .Theprobabilityofimitationfrom aninformationleakispositivelyrelatedtotheamountofRandDdoneintheSouth. 6 5 LinkingRandDcoststothexedcostofentryhasbeenlongestablishedbyprevious internationaltradeliterature.Forexample,Melitz3usesamodelwithaxedcostarm mustpayinordertoexport.Thisxedcostofexportingisexplainedasthecostofadaptive researchinordertopreparetheproducttobesoldinnewmarkets. 6 Theprobabilityofimitationisbasedaroundthelargeamountofxedcoststhatthe SouthernrmmustpayifitisnotabletocopytheproductthroughmobileSouthern researchers.IassumethattheSouthernrmwillfacesucientlylargecoststofurtherresearch theproductifthermlosestheimitationlottery.Thisadditionalxedcostofentryishigh 18

PAGE 19

Employeemobilityactsasthechannelofimitation.ASouthernworkercanleavetowork attheSouthernrmafterassimilatingknowledgeusedinthecreationofthenewproduct; however,institutionalbarriers,thelackofotherRandDknowledge,andlabormovement barrierspreventthisworkerfromfullythreateningtheNorthernrm'sclaimonthepatent.I denetheprobabilitiesas: = R S R N + R S 0 < < 1 {2 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( = R N + R S )]TJ/F24 11.9552 Tf 11.955 0 Td [( R N + R S {3 Theexogenousparameters and measurebarrierstoimitationintheSouth.These barrierstoimitationcanbecertainIPR-protectionbarrierssuchasinstitutionalenforcement mechanisms,durationofpatentprotection,involvementininternationalIPR-agreements, andtheabsenceofpatentrestrictionssuchasworkingrequirements.Thisinterpretationof thebarrierparametersisespeciallyrelevantforinnovationsthatleadtopatentsinsteadof innovationsthatareprotectedthroughothermeans.Thebarrierparametersalsorepresent thelackofnecessaryinformationforfullyimitatingtheproduct.Southernresearchersmay nothaveaccesstofundamentalcomponentsoftheinventionwhileresearchingit.That is,NorthernresearcherscanresearchimportantaspectswhiletheNorthernrmallocates non-integraltaskstotheSouth.Withemployeemobilityasthechannelofimitation,these barrierparameterscanalsobeviewedasconstraintsonlabormovementsuchaslabor contractsandnoncompeteclauses.Therefore,havingstronglegalinstitutionsthatcan enforcepatentprotectionandlaborcontractsdetersimitation.Furthermore,theavailabilityof lawyersinthecountrytopursueinfringementcasesdetersimitation.Theburdenoflitigation detersworkersfromattemptingtobreakformallaborcontracts. enoughtokeepthermoutofthemarket.Also,Iassumethatsuccessfulimitationrequires verylittleextracostinordertoenterthemarket.Thus,successfulimitationleadstotheentry oftheSouthernrm. 19

PAGE 20

Theparameter scalestheimitationprobabilitytoreectthedicultiesinimitating aproductdespitehavingaccesstotheresearchgoneintomakingtheproduct.Thus, representstheabilitylevelofSouthernresearchersasindependentresearcherstorecreatethe innovation.So,ifallresearchisdonebySouthernworkers,andthustheSouthernresearchers andSouthernrmhaveknowledgeonallnewcomponentsoftheinnovation,theprobabilityof imitationwillbe .At =0 ,copyingtheproductisimpossibleandtheSouthernrmcannot enterthemarket.Forexample,theinternalknowledgethattheNorthernrmkeepsintheir headquartersmaybefundamentaltothecreationoftheproduct.DespitetheSouthernrm havingaccesstoalltheknowledgecreatedintheSouth,itwillnotbeabletocombinethe knowledgeintoanactualproduct. When =1 ,aTullocklotterybetweentheNorthernandSouthernrmssolelydetermines whetherimitationoccurs.ThenumberofNorthernresearchers R N canbeseenasthe eortextendedbytheNorthernrmtoprotectitspatent.Likewise,thenumberofSouthern researchers R S istheeortextendedbytheSouthernrmtosuccessfullycopythenew product. 7 AlthoughtheprobabilityfunctioninthismodeldrawsfromTullock'sproposed successfunction,theNorthernrmsetsbotheortlevels. 8 Tosimplifythemodel,Iset =1 whichsetstheresearchabilityoftheSouthernresearchersatthehighestvalue.When < 1 model'sresultsremainunchanged. 9 7 Althoughtherehasbeensomeevidencetosuggestthatemployeemobilityand entrepreneurshipistiedtoindividualperformanceCarnahan,Agarwal & Campbell,2012, ItreateachSouthernresearcherashomogenousintheirthreatofleavingtobringtheir knowledgetoacompetingrm. 8 Becauseonlyoneplayersetsbotheortlevels,theprobabilityfunctionusedhereis incompatiblewithotherformsoftheTullockcontestprobabilityfunctions.Forexample, secondorderconditionsdonotholdwhenapowerTullockCSFisused. 9 Ingeneral, addsacertainamountofrealismtothemodelinprovidinganadditional barriertoimitationwhenallresearchiscarriedoutintheSouth.AllowingSouthernresearchers tobelesseectivedoesnotchangethecomparativestaticresultsofthismodel.Iexplorethe casewhere 0 << 1 inAppendixA.2. 20

PAGE 21

Theparameter becomesthesolemeasureofbarrierstoimitationintheSouth.I interpret asthebarrierstoimitationrelatedtoIPR-protectionthatareexogenoustothe Northernrm.Specically, representstheinstitutionalbarriersassociatedwithprotectinga patent.So,theSouthernrm'seort, R S ,tocopytheproductisnotweightedasmuchas theNorthernrm'seorts, R N ,toprotectitsproductfromimitation.At =1 ,theeorts ofthetwormsareweightedequally.A nearzeroindicatesstrongSouthernIPR-protection, whilea near1indicatesweakIPR-protection.IfallresearchisdoneintheSouth,imitation followedbycompetitionoccurswithabsolutecertaintyregardlessofthestrengthoftheSouth's IPR-protectionlaws.As increases,theprobabilityofimitationincreases.Thetotalamount ofresearchcarriedoutisxedbysomeproductionfunctionforthecreationofnewproducts. ConsiderthefollowingCobb-Douglasknowledgeproductionfunction: Y = b 1 R 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( 1 {4 whereYisthenumberofproductsdiscovered.Toproduceonenewproduct,or Y =1 ,the Northernrmneedstoemploysomenumberoftotalresearcherssuchthattheproduction functioncanbesatised.Isolveforthetotalresearchlevelthatwillallowforthecreationofa newproduct: R = 1 b 1 1 1 )]TJ/F26 5.9776 Tf 5.756 0 Td [( 1 {5 Thefollowinganalysisuses R asaxedparameter.Giventhisxedtotalresearchlevel, theNorthernrmchoosesthenumberofresearchersinbothregionsbymerelysettingthe valueinoneregion.Thatis, R S canbeexpressedintermsofthexedtotalresearchvalueand Northernresearch: R S = R )]TJ/F38 11.9552 Tf 11.955 0 Td [(R N {6 21

PAGE 22

Giventheseprobabilitiesandtotalresearchlevels,expectedprotscanbedened: E = R N R N + R S )]TJ/F24 11.9552 Tf 28.494 8.088 Td [( R S R N + R S )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N R N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R S = R N R N )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ]+ )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N R N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R )]TJ/F38 11.9552 Tf 11.955 0 Td [(R N {7 Theactualvaluesof and dependonthemarketstructureandthenatureofcompetition betweentheNorthernrmandtheSouthernrm.SincetheNorthernrmobservesthe Southernrm'sentrancebeforemakingproductiondecisions,theproductiondecisionsdonot dependontheRandDchoices.First,Isolvefor R N intermsofgeneric and values. 1.3.2Equilibrium TheNorthernrmsolvesthefollowingmaximizationproblem: max R N E Thermdoesnothiremoreresearchersthanisnecessarytocompletetheinnovation. Likewise,thermcannothireanegativeamountofresearchers.So, R N isboundedby0and R .Inordertoguaranteeconcavityoftheexpectedprotfunction,severalassumptionsare needed.First,monopolyprot, ,isgreaterthanduopolyprot, ,asisstandardineconomic literature.Second, w N > w S .ThewagepaidtoNorthernworkersisgreaterthanthewage paidtosimilarworkersintheSouth.Finally,Iassume 0 << 1 .Duetothemanychallenges ofimitation,theeortbySouthernresearcherstoemulatetheinnovationisinherentlyless eectivethantheNorthernrm`seorttokeeptheinnovation.Giventheseassumptionshold, arstorderconditionyieldsamaximumvalueofNorthernresearch.Fortheremainderofthe discussion,Iassumethattheseassumptionsdoinfacthold. 10 10 Thesecondderivativeofprotfunction-7isequalto @ 2 @ R N 2 = )]TJ/F22 7.9701 Tf 6.586 0 Td [( )]TJ/F25 7.9701 Tf 6.586 0 Td [( R [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] R N )]TJ/F25 7.9701 Tf 6.587 0 Td [( + R 3 .Itistrivial toshowthattheprotfunctionisconcavegiventheassumptionsontheparametervaluesand theboundson R N 22

PAGE 23

Proposition1. Theprotfunction-7isconcavefor R N 2f 0, R g Proposition1impliesthatrstorderconditionsyieldsavalueof R N thatmaximizesprot function-7ifthisvalueexistsbetween0and R .IfinfacttheFOCdoesnotyieldavalue withinthesebounds,thenoptimalNorthernresearchamountislocatedatoneofthesebounds dependingonwhethertheprotfunctionstrictlyincreasesordecreaseswith R N betweenthe rangeofinterest,0and R .Thetypeofsolution,interiororcorner,canbedeterminedby lookingattherelationshipbetweentotalresearch,thepenaltyofimitation,IPR-protection,and thewagegap. 11 Inordertobetterdiscussthecomparativestaticsofthismodel,Ifocuson threeassumptionsonparametervalues: 12 Case1: TheNorthernrmlocatesallresearchintheNorth Case2: TheNorthernrmlocatesresearchinboththeNorthandtheSouth Case3: TheNorthernrmlocatesallresearchintheSouth TakingtheFOCandrearrangingyieldstheoptimalNorthernresearchlevel: 13 R N = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R 11 TheseparameterrelationshipsareexploredinAppendixA.3. 12 Someinnovationsarenotabletobesplitacrossmultiplecountries.Acomplexinnovation mayrequirethatallresearchtasksarepursuedtogether.Firmsthenlocateallinnovative activitieswithinonelocation.However,complexinnovationscanhavemodularsoftasks thatcanbepursuedindividuallyandthenreassembledtocreatethelargerinnovationSimon, 1962. 13 FOCsactuallyyieldtwosolutionsfor R N ;however,theunreportedvalueofNorthern researchersisanegativevalueandexistsontheportionoftheprotfunctionthatisconvex. Thisunreportedsolutionisthusaminimum. 23

PAGE 24

Takingthevaluefor R N alongwithequation-6andtheboundsontheNorthern researchlevelsyieldsauniquesolutionfortheoptimalNorthernandSouthernresearchlevels: R N = 8 > > > > > > < > > > > > > : R if R < R N 1 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] R w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( R if 0 < R N < R 0 if R N < 0 {8 R S = 8 > > > > > > < > > > > > > : 0 if R < R N 1 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( R )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] R w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S 1 2 if 0 < R N < R R if R N < 0 {9 Equations-8and-9showtheoptimalNorthernandSouthernresearchlevels forCase1,Case2,andCase3,respectively.Thewagegap,thepenaltyofimitation, IPR-protection,andthetotalamountofresearchneededtoinnovateaecttheoptimal levelofNorthernresearchersemployed.InitialinspectionrevealsthattheamountofNorthern researchersincreaseswiththepenaltyofimitation.Thewagegap,ontheotherhand,has anegativerelationshipwithNorthernlabor.TheIPR-protectionparameterandthetotal researchvariablehasanambiguouseects.NotethatinCase2,theNorthernresearchlevel increasesonlyforverysmallvaluesof R 14 Northernresearchsteadilydecreaseswith R until theNorthernrmmovesentirelytotheSouth.Whensettingtheoptimalresearchlevels,the 14 Specicallyfor R < 1 4 [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S 24

PAGE 25

Northernrmtherebysetstheprobabilityofimitationinequilibrium: = 8 > > > > > > < > > > > > > : 0 if R < R N 1 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( R w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( if 0 < R N < R 1 if R N < 0 {10 Asexpected,theprobabilityofimitationweaklydecreasesasthepenaltyofimitation increases.TheNorthernrmattemptstopreventemployeesfromleavingthermand imitatingvaluableinnovationsbylocatingmoreresearchintheNorthandthusloweringthe probabilityofimitation.Again,theIPRparameterhasanambiguouseect.Theprobability alsoweaklyincreasesastotalresearchincreases.Thisresultseemstoconictwiththe relationshipbetween R N and R .Forlowvaluesof R andaweakIPRenvironment,Northern researchincreasesastotalresearchincreaseswhiletheprobabilityofimitationalsoincreases. Inthissituation,thermlocatesonlyaportionofthenewrequiredresearchintheNorth. TherestoftheneededresearchiscontractedtotheSouth,whichcontinuestoincreasethe probabilityofimitation.Infact,theratioofNorthernlabortototallaborweaklydecreases withtotalrequiredlabor.Figure1-1showstheratioofNorthernresearchtototalresearch andtheratioofSouthernresearchtototalresearchasfunctionsof R .TheshareofNorthern researchexceedstheshareofSouthernresearchforsmallvaluesof R WhentheNorthernrmhasasmalllevelofneededresearch,thermlocatesentirely intheNorth.As R increases,theNorthernrmshiftsmoreresourcestotheSouth.After acertainpoint,theNorthernrmonlyusesSouthernresearch.Alargeamountofresearch requiredtocreateanewproductmeansalargexedcostofentry.Theonlywayinwhich entrywouldbefeasibleisiftheNorthernrmconductedtheresearchintheSouth.Therm attemptstoloweritsxedcostbylocatingmoreresearchintheSouth,wherelaborischeap. Figure1-1showstheintervaloftotalresearchinwhichthermwilllocateinbothcountries. 25

PAGE 26

Figure1-1.OptimalResearchRatios Formally,thatintervalis: [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < R < [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S {11 When R fallsintheintervalexpressedincondition-11,theNorthernrmlocates researchinboththeNorthandtheSouth.Noticethatfor 0 << 1 ,thelefthandside ofcondition-11islessthantherighthandside.Whetherthisconditionholdsdepends ontherelationshipbetweenthedemandspecicparameters,productiondecisions,degree ofcompetition,wagedierential,SouthernIPR-protection,andtheleveloftotalresearch. StrengtheningSouthernIPR-protectionlower leadstoadecreaseinthelefthandside andanincreaseintherighthandside.So,astrongIPR-protectionenvironmentintheSouth 26

PAGE 27

facilitatestheexistenceofSouthernresearch.Incontrast,thepenaltyofimitationinteracts withtheleftandrighthandsideinthesameway.Anincreaseinthepenaltyofimitationshifts theintervalexpressedincondition-11totheright;therefore,theNorthernrmmaybe inducedawayfromSouthernresearchasthepenaltyofimitationincreasesdramatically. 15 The dierenceinwagesintheNorthandtheSouthhasacomplicatedeect.Aswiththepenalty ofimitation,thewagedierentialactsdirectlyontheleftandrighthandsideinthesame direction;however,wagesalsohaveanindirecteect.Theprotcollectedintheproduction phasedependsonthewagesaswell. 1.4Results 1.4.1 R `sEectonRandDLocations Researchlocationdecisionsareafactorof R ,thetotalamountofresearchnecessaryto makethenewinnovation.AsseeninFigure1-1,rmswiththesmallestresearchneedslocate entirelyintheSouthwhilermswiththelargestresearchneedslocateentirelyintheNorth. Firmswithmediumresearchneedslocateinbothcountries: Result1. Largevaluesoftotalresearch,specically R [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S ,inducetheNorthernrm tolocateresearchonlyintheSouth.Lowvaluesoftotalresearch, R [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S ,inducethe NorthernrmtolocateallresearchintheNorth.TheNorthernrmlocatesinbothcountries formediumvaluesofresearch, [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S < R < [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S Result1relatestheextracostofresearchinlocatingintheNorthovertheSouthto thepenaltyofimitation.TheIPR-protectionparameter, ,weighsthepenaltyofimitation. Thatis,thecostofresearchinginonlytheNorthdoesnotneedtoexceedthefullpenalty 15 Anincreaseinthepenaltyofimitationinthismodelcanbeseenbycomparingtwo identicalrmsthatexistindierentcompetitivemarkets.Forarmthatcompetesina Cournotmarket,thepenaltyofimitationissmall.ForarmthatcompetesinaBertrand market,thepenaltyofimitationiscomparativelylarger.Monopolyprotisthesameregardless ofmarketstructure.DuopolyprotfortheNorthernrmiszeroinaBertrandmarketwhile duopolyprotispositiveandlessthanmonopolyprotintheCournotmarket.Thiscanbe seenintheanalysisinsection1.5. 27

PAGE 28

ofimitationfortheNorthernrmtostartsplittingresearchbetweennations.Likewise,the costpenaltyofresearchinginonlytheNorthmustbesignicantlygreaterthanthepenalty ofimitationinorderfortheNorthernrmtondlocatinginonlytheSouthprotable. ThestrongertheIPR-protectionenvironmentintheSouth,thegreaterthecostpenaltyof researchinginonlytheNorthmustbefortheNorthernrmtobewillingtolocateallresearch intheSouth.Again,thetotalresearchneededtocreateanewproductcanbeseenasaproxy forrmsize.LargerrmshavetheresourcestosecureinitialfundingforRandDforadvanced projectsthatrequirelargeamountsofresearch.Whiletherelationshipdoesnotalwayshold, thecorrelationbetweenthetwofactorsgenerallyimpliesthatlargerrmsinvestintheresearch projectsthatrequirealargeamountofresearchtaskstobecompleted.Theselargerms arealsomorelikelytoalreadyhavetheinfrastructuretobothfacilitatethecoordinationof researchacrossmultiplelocationsandproperlymarketandsellthenewproduct.Whenlooking acrossdierentindustries,rmsinhigh-techindustrieswillhavelargevaluesof R whilethe oppositeistrueforrmsinlow-techindustries. 1.4.2IPR-Protection`sEectonRandD Therearetwomaindrivers,aftertotalneededresearch,oftheNorthernrm'sactions foragivenmarketstructure:wageratesandIPR-protection.Theuseoflaborinboththe productionandresearchstagescomplicatetheinterpretationofwageeectsonthesystem. AlthoughthewageineachcountryfactorsintoboththeproductiondecisionsandtheRand Ddecisions,IPR-protectiononlyaectsresearchchoices.IPR-protectionhasadualeecton theoptimalresearchlevels.Fromequation-8,astrengtheningofSouthernIPR-protection adecreasein canaect R N bothnegativelyandpositively.Thatis,astrengtheningof IPR-protectionincreasestheprobabilityofobtainingmonopolyprot.Theincreaseinexpected protcausedbyachangingIPRregimepromptstheNorthernrmtoreallocatesomeresources intothelessexpensiveSouth;however,ahighmonopolyprotmayincentivesthermto increasetheprobabilityofgainingthatprotbyshiftingevenmoreresourcesintotheNorth. Thechangein R N dependsonwhethertheexpectedprotisincreasedmorebyloweringthe 28

PAGE 29

xedcostsorbyincreasingthechancetogainthemonopolyprot.Again,theamountoftotal researchneededplaysaroleindetermininghowtheNorthernrmreactstoIPRchanges: Result2. WeakeningSouthernIPR-protectionhasanambiguouseecton R N and R S Specically,aweakeningofSouthernIPR-protection,orincreasing ,decreases R N i )]TJ/F22 7.9701 Tf 6.675 -4.428 Td [(1+ 2 2 [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S < R < [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S .Otherwise, @ R N @ 0 and @ R S @ 0 WhentheNorthernrmlocatesentirelyintheNorth,IPR-protectionhasnoinuenceon theNorthernrm'sresearchlocationdecision.Theprobabilityofimitationinthesesituations doesnotrelatetoIPR-protection.IftheNorthernrmlocatesresearchandproductioninonly theNorth,theSouthernrmneverhasachancetocopyinformationandproduceitsown productinthismodel.So,thelevelofIPR-protectionintheSouthshouldhavenoeecton theNorthernrm'sdecisions.Likewise,locatingallresearchintheSouthiscertaintoresultin imitation. InCase2,ontheotherhand,aweakeningofIPR-protectionintheSouthcouldcause theNorthernrmtoeitherincreaseNorthernresearch,asonewouldexpect,ordecrease Northernresearchinordertoincreaseexpectedoperatingprotsevenmoreso.Inline withpreviousliterature,themodeldoesnotpredictadenitiveeectofIPR-protection oninnovation.Whenarelativelysmallamountofresearchisneeded,theNorthernrm reactstoanincreasein byincreasingNorthernresearchanddecreasingSouthernresearch. Thelowlevelsofrequiredresearchindicatealowxedcostofentry.Thedecreaseinthe expectedoperatingprotcausedbytheworseningoftheIPRenvironmentwouldthencause theNorthernrmtoshiftmoreresourcesintotheNorthasthemarginalchangeinthe expectedoperatingprotoutweighsthemarginalchangeinxedcosts.Figure1-2showsthe resultofaweakeningofIPR-protectionresultinginadecreaseinSouthernlabor.Afterthe weakeningofIPR-protection,thermmovesfrompointAtopointB.Forhighervaluesof totalresearch,thexedcostofentrywillbehigh.AweakeningofIPR-protectionyieldsa decreaseinexpectedrevenue,whichtheNorthernrmthenrecoupsbydecreasingthealready highxedcostsofentry.So,thermthenrespondstotheweakeningofIPR-protectionby, 29

PAGE 30

Figure1-2.WeakeningofIPR-protectionyieldingadecreasein R S counterintuitively,increasingSouthernresearchers.Figure1-3showsthiscase.Theintervalin whichthermexistsinCase2shrinksasIPR-protectionweakens,meaningthatsmallrms whohadresearchintheSouthmightbemovedentirelyintotheNorthbyaweakeningof IPR-protection. Theprobabilityofimitationwillincreasewhen increasesi 2 1+ 2 [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S < R .This minimumvalueisalsothepointatwhichtheSouthernresearchsharestartstoovertakethe NorthernresearchshareasseeninFigure1-1.Sincethispointislessthan )]TJ/F22 7.9701 Tf 6.675 -4.428 Td [(1+ 2 2 [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S therewillbearegioninwhichtheNorthernrmincreasesNorthernresearchinthefaceofa weakeningofIPR-protection,buttheprobabilityofimitationincreasesstillduetothechange in 30

PAGE 31

Figure1-3.WeakeningofIPR-protectionyieldinganincreasein R S 1.4.3MarketStructureandWage`sEectonRandD ThestructureofthecompetitivemarketdictatesanumberofdecisionstheNorthernrm faces.Themodelisbuiltaroundthedierencebetweenmonopolyprotsandduopolyprots, orthepenaltyofimitation.Thevalueofthisdierencedependsonthewagedierential, marketparameters,andtypeofcompetition.Thissubsectionexaminesthecasewhenthe penaltyofimitationincreasesholdingallelseequal.Result3canbeseenasacomparison betweentwodierentmarketstructures,suchasaBertrandmarketstructureandaCournot marketstructure.WhileBertrandmarketstructureswithundierentiatedgoodsyieldaduopoly protofzerofortheNorthernrm,Cournotcompetitionwithundierentiatedgoodsyields 31

PAGE 32

apositiveduopolyprotfortheNorthernrm. 16 Therefore,aNorthernrmhasahigher penaltyofimitationintheBertrandmarketthanintheCournotmarket.Result3alsooutlines thecaseofanincreaseinmarketsize. 17 Result3. Anincreaseinthedierencebetweenmonopolyprotandduopolyprotweakly increases R N andweaklydecreases R S .Ingeneral,anincreaseinthepenaltyofimitation detersinvestmentinSouthernresearch. Result3impliesthatcompetitivestructuresthathavehigherduopolyprotscausethe NorthernrmtolocatemoreresearchintheSouth.Giventhatthepenaltyofimitationis small,theNorthernrmtakesadvantageofthecostdierencebetweenSouthernresearch andNorthernresearch.Ontheotherhand,alargepenaltyofimitationcausesthermto locateintheNorthtobeabletoobtainamonopolyprotandcapturetheentiremarket. Likewise,whentheNorthernrmpreferstolocateinbothcountries,anincreaseinthepenalty ofimitationinducestheNorthernrmtolocatemoreresearchintheNorth.Thischangethen decreasestheprobabilityofimitationandincreasestheprobabilityoftheNorthernrmbeinga monopolyintheproductionstage. Figure1-4showstheeectofanincreaseinthepenaltyofimitationontheoptimal R N and R S .NoticethatasubstantialincreaseinthepenaltyofimitationresultsintheNorthern rmnotlocatingintheSouthatall.Theintervalthatrepresentscondition11shiftsrightwith anincreaseinthepenaltyofimitationandshiftsleftwithadecrease.So,harshcompetition discouragesinvestmentinSouthernresearch.Thedierencebetweenthemonopolyand duopolyprotlevelsdependsonthenatureofcompetitionintheduopolyphase.Monopoly protsremainsconstantacrossallduopolystructures.Aduopolyprotofzeroresultsinthe NorthernrmlocatingallresearchintheNorth.LocatingintheSouthinthissituationresults innooperatingprotwhatsoevertohelppaythexedcostofentry. 16 Thisisshowninsection1.5. 17 Thisassertioncanbeseenfromtheanalysisinsubsection1.5.3. 32

PAGE 33

Figure1-4.Anincreaseinthepenaltyofimitation Holdingthemonopolyprotandduopolyprotequal,alargewagegapinducesthe NorthernrmtoinvestheavilyinSouthernresearch,whichisexpected.Relativelycheaplabor intheSouthdrivesSouthernresearchinvestment.Anincreaseinthewagegapcausesboth thelefthandsideandrighthandsideofcondition-11todecrease.Anincreaseinthewage gapthenmaypusharmfromlocatingentirelyintheNorthtolocatinginbothcountries. RelativelycheaplaborintheSouthhelpsdriveSouthernresearchinvestment. Result4. Holdingallelseequal,alargewagegapwillweaklyincrease R N Thewageratemayinfactberelatedtothewageoftheworkersusedtoproducethe good.So,anexogenousriseinthewagesofresearchersmayalsobereectedinthewageof laborers.AlargewagegapisnotsucientforSouthernresearchtotakeplace.Themonopoly 33

PAGE 34

andduopolyprotspossiblychangewiththewagerates.So,theeectofawagechangeis unclearwithoutdeningamarketstructurerst.Overall,theeectofawagechangeson researchallocationisambiguousinthismodel. 1.5MarketStructureEects Thepenaltyofimitationdependsheavilyonthemarketstructureofthenalgoods market.Whenthereisnoinformationleak,theNorthernrmistheonlyrmonthemarket andactsasamonopoly.WhenanoutsideSouthernrmisabletoimitate,theNorthernrm mustcompetewiththisrm. 18 TheoperatingprotfortheNorthernrmwiththepresence oftheSouthernrmwilldependonhowthetwormscompeteforcustomersintheirindustry. ConsiderthepossibilityofonlytheNorthernrmandtheSouthernrmproducing q 1 and q 2 ,respectivelysellingtheirgoodtoacontinuumofhomogeneousconsumers.Consumers purchasegoodsfromthismonopolisticsectorandanoutsidegood, q 0 ,fromacompetitive sector.Consumerschoose q 1 and q 2 tomaximizetheutilityfunction: U q 0 q 1 q 2 = q 0 + A q 1 + q 2 )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 1 q 2 1 +2 q 1 q 2 + 2 q 2 2 {12 Let A and i bepositivefori=1,2.Also,assume 1 2 > 2 and i > for 8 i A i is amarketsizeparameter.Theparameter measuresthedegreeofsubstitutabilitybetween thetwoproducts.Iassumethatthetwoproductsareperfectsubstitutes. 19 Assumethatthe SouthernrmexclusivelyusesSouthernlaborinproductionwhiletheNorthernrmusesonly 18 TheSouthernrm,liketheNorthernrm,hasnoxedcostofproduction.Furthermore, thexedcostofentrytotheSouthernrmwouldberelatedtothecostofstealinginformation fromtheNorthernrm.TheotheraspectoftheSouthernrm'sxedcostofentryisthe additionalresearchneededtocreatetheproduct.Itispossiblethatevenafteraninformation leak,theadditionalxedcostofentryistoohighfortheSouthernrmtoenterthemarket. 19 Ialsoexploredmarketswithslightlydierentiatedproducts.Thedegreeofsubstitutability changesthecomparativestaticsofanymarketparameter.Forexample,theinclusionof dierentiatedproductsmakesduopolyprotmoresensitivetowageandtotalresearchchanges thanmonopolyprotgivensmallwagedierential. 34

PAGE 35

Northernlabor.Also,ittakesonworkerstoproduceoneunitofthenalgood.Assumethat A > w N .Inotherwords,themarketmustbesucientlylargeinorderforthermtobe abletonditprotabletoproduce.WhentheSouthernrmdoesnotenterthemarket,the Northernrmgathersamonopolyprotof: = A )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N 2 4 1 {13 AsseeninEquation-13,themonopolyprotincreaseswithmarketsize,oranincrease inA.Themoreconsumersbuyingtheproduct,thehigherprotstheNorthernrmwill make. 1 relatesthepriceandquantityofagoodforconsumers.Anincreasein 1 decreases theconsumersquantitydemandedforagivenprice.Thepricesetbythemonopolistis unaectedby 1 ;however,thequantitysoldisinverselyrelatedtotheparameter.Notethat themonopolyprotwillnotchangeregardlessofthemarketstructure,withtheexceptionof 1 whichwillchangewhentheproductsproducedbybothrmsareundierentiated.Sections 1.5.1and1.5.2explorethiscase. 1.5.1BetrandwithUndierentiatedProducts InBertrandcompetition,rmscompeteinpricespace.Inordertohaveundierentiated products,let = 1 = 2 = .Withundierentiatedproducts,thermthatsetsthelowest pricegainstheentiremarket.So,demandfortheNorthernrmcanbere-speciedas: q 1 p 1 p 2 = 8 > > > > > > < > > > > > > : Q p 1 2 if p 1 = p 2 Q p 1 if p 1 < p 2 0 if p 1 > p 2 Given w N > w S ,theSouthernrmwillbeabletosetapriceat w N )]TJ/F24 11.9552 Tf 12.272 0 Td [( ,where > 0 isasmallnumber,andtherebypricetheNorthernrmoutofthemarket.TheNorthernrm 35

PAGE 36

wouldearnnegativeaprotifitproducedatthispriceandzeroprotifitdoesnotproduce. Therefore, =0 20 1.5.2CournotwithUndierentiatedProducts Again,let = 1 = 2 = .Unlikethemonopolyprot, willdependontheSouthern rm'smarginalcost,or w S .Usingthisinversedemand,theNorthernrm'sprotunderthis competitiveframeworkare: = A )]TJ/F21 11.9552 Tf 11.956 0 Td [(2 w N + w S 2 9 {14 Aswiththemonopolyprots,theprotfortheNorthernrmundercompetitionis positivelyrelatedtomarketsize.Northernduopolyprotsarealsorelatedtopositivelyrelated to w S .Furthermore, decreasesas increasesinasimilarmannerasmonopolyprots. AnincreaseintheNorthernrm'smarginalcost, w N ,expectedlydecreases ;however,the decreaseinmonopolyprotsforthesameincreasein w N willbelarger.Also,notethatthe monopolyprotsarehigherthantheprotsundercompetition,whichisexpected. 1.5.3WageandMarketSizeEectsonRandD MarketsizeonlyfactorsintotheNorthernrm'sdecisionthroughmonopolyandduopoly prots.So,thechangeinthedierenceofthesetwoprotlevelsduetoachangeinmarket sizedictatesthechangeintheNorthernrm'sresearchdecision. Proposition2. Anincreaseinthemarketsizeincreasesthedierencebetweenthemonopoly andduopolyprots. UsingProposition2,anincreaseinthemarketsizefacilitatestheNorthernrmlocating more,ifnotall,researchintheNorth.Alargermarketmeansmorecustomers.Themonopoly 20 IftheNorthernrmwasabletoproduceintheSouth,bothrmswouldhavethesame marginalcostandwouldthusbeunabletodriveeitheroutofthemarket.TheSouthernrm setpriceequaltomarginalcost, w S ,topreventtheotherrmfromundercuttingtheirprice andgainingtheentiremarket.So,thermssplitthemarketbutmakenoprotoneveryunit sold,and =0 36

PAGE 37

cangainallofthesecustomerswhiletheduopolymustsplitthenewconsumerswiththe Southernrm.So,bothprotlevelsincrease;however,themonopolyprotincreasesmore thantheduopolyprot.Theeectofanincreaseinmarketsizeonresearchlevelsisthenthe sameResult3. Proposition3. ForrmsinCase2,ariseintheNorthernwageratedecreases R N and increases R S .AriseintheSouthernwageratehasanambiguouseectintheCournotcase. IntheBetrandcase,anincreasein w S increasesNorthernresearchanddecreasesSouthern research. ForbothBetrandandCournotenvironmentswithundierentiatedproducts,anincreasein theNorthernwageratereducesthepenaltyofimitation.ThehighercostofNorthernresearch alsomakesSouthernresearchersmoreattractivetotheNorthernrm.Sincethepenalty ofimitationislowered,theNorthernrmislessconcernedwithdecreasingtheprobability ofimitation.Therefore,theNorthernrmcontractsmoreresearchintheSouthinorderto takeadvantageoftherelativedecreaseinthepriceofSouthernlabor.When w S increases intheBertrandcase,duopolyprotfortheNorthernrmremainsatzero.IntheCournot case,thedierencebetweenprotsshrinksastheSouthernwageincreases.TheSouthern rm'smarginalcostrises,makingthemalessercecompetitortotheNorthernrm.So,the NorthernrmmayincreaseNorthernresearchinthefaceofthischangeinordertoincrease theprobabilityofgainingthemonopolyprot.Ifthedierencebetweenpossibleprotsis smallenoughasaresultofthechangein w S ,theNorthernrmwillnditscurrentamount ofNorthernresearcherstobeexcessiveandshiftresourcestotheSouth.Sincethepenaltyof imitationdecreases,theappealofNorthernresearchersdecreasesgreatly.Finally,whenthe NorthernrmlocatesallresearchintheNorth,theSouthernwagewillnotaectexpected prot.TheNorthernrmemploysnoSouthernlaborandimitationwillnotoccur;therefore, theSouthernwageratewillnotaectanyoftherm'sdecisions. Proposition4. ForBetrandcompetitionwithundierentiatedproducts,theNorthernrm neverlocatesresearchonlyintheSouth. 37

PAGE 38

TheBetrandmarketstructurewithundierentiatedproductscasegivesthemaximum dierencebetweenmonopolyandduopolyprots.WhentheNorthernrmlocatessolelyinthe South,theyearnnooperatingprot.Therefore,thetotalprotfunctiononlyconsistsofthe xedcostofentry.TheNorthernrmearnsanegativeprotandthuswouldbeworseothan stayingoutofthemarketcompletely.IftheprobabilityofimitationislowenoughtheNorthern rmcouldstilllocatesomeresearchintheSouth.Notethattheothermarketstructuressee theNorthernrmearningapositiveduopolyprot,meaninglocatingentirelyintheSouthis stillanoption. 1.6Conclusion Thismodel'smainpurposeistoexplainthelackofSoutherninnovativeresearchinthe faceofalowSouthernwageandcapableresearchers.Innovativeresearchinthedeveloping worldisdiscouragedbytheharshpenaltiesofproductimitation.Ifimitationleadstothe Northernrmbeingdrivenoutofthemarketcompletelybyrmswhodidnothavetoinvest incostlyresearchtoinventanewproduct,thentheNorthernrmonlylocatesresearchinthe North.Mymodelpredictsthatamediumsizedrmcanbenetfromsplittingresearchtasks, whenpossible,betweenbothnations.Thispredictionmatchesthegrowingtrendinthereal worldofconductingresearchinbothlowincomeandhigh-incomenations.Themodelalso predictsthatsmallerrmsarelesslikelytooutsourceinnovationwhilelargerrmsoutsourceall innovationtothedevelopingworld. InexplainingthelackofinnovativeRandDdoneintheSouth,IPR-protectionplaysa distinctbutsomewhatambiguousrole.Again,IPR-protectiondoesnotfactorintotheNorthern rm'schoiceswhenitlocatesallresearchwithinonecountry.InthecasethattheNorthern rmlocatesinbothcountries,IPR-protectionhasanambiguousrole.Whilestrengthening IPR-protectionintheSouthcouldresultintheNorthernrmmovingresearchtotheSouth, theincreasedprobabilityofgainingamonopolyprotcouldcausetheNorthernrmtolocate evenmoreresearchintheNorthtofurtherincreasethisprobability.Inthiscase,theNorthern rmwillinglypaysahighercostofentryinordertoincreaseexpectedoperatingprot.My 38

PAGE 39

modelalsoexplainsthelackofSouthernresearchasaresultofwagedierences.Inthemodel, wagesfactorintotheproductionprocessandthepre-productionprocessortheresearch phase.AhighenoughNorthernwagerelativetotheSouthernwagecausestheNorthernrm tolocateamajority,ifnotall,researchintheSouth;however,harshenoughcompetitioninthe SouthpreventstheNorthernrmfromlocatingintheSouth,regardlessofthesizeofthewage gap. Thegrowingvalueofhavingresearchtasksinbothcountriesshouldbekeptinmindwhen policymakersbeginthinkingaboutintellectualpropertyrightsreform.Thischapterprovides theframeworkforanumberoffutureresearchprojectsaddressingavarietyofissues.While thismodeldealswithasingleNorthernrmanditsSoutherncompetitor,theresultscanbe appliedtomultiplermsofdierentvarieties.Furthermore,theinterplaybetweenresearch labsandproductionsitesincontexttoinformationleakagehasnotbeenexploredwiththis modelandwouldbeapossibleextensionofthischapter.Thatis,giventhecolocationofa productionsiteandaresearchsite,theprobabilityofaninformationleakwouldbehigherthan ifthetwositeswerelocatedindierentregionswithdierentIPRlaws.Whilethismodelonly presentstheNorthernrm'sresearchchoices,includingalookattheSouthernrm'sxed costofentrycouldbeapossibleadditionaswell.Thatis,byperformingonlyafewRandD tasksintheSouth,theNorthernrmincreasesthexedcostofentryoftheSouthernrm. AfterabsorbinginformationfromtheSouthernresearchsites,theSouthernrmwouldstill needtodoadditionalresearchinordertocreatetheproduct.Byhavingintegralresearchdone instrongIPRcountries,theNorthernrmcanseverelylimitthedamageofaninformation leak.Myanalysistreatsallresearchtasksthesame;however,sometasksaremorekeytothe functionalityoftheoverallproductthanothers.TheNorthernrmhastheabilitytolessen thevalueofsomeresearchtasksbyhavingitrelyheavilyonothertasksnotknowntothe Southernresearchers. Whilethemodelpresentedaboveshowstheeectsofdierentmarketstructureson researchdecisions,eachmarketstructurecontainsonlytwormsatmostinthemarketwith 39

PAGE 40

undierentiatedproducts.Aninformationleakthatleadstoperfectcompetitionwouldyield similarresultstoBertrandcompetitionwithundierentiatedproducts.Iftheinformation leakleadstoonlyafewrmsenteringthemarketwithdierentiatedproducts,thedegreeof substitutionbetweentheproductsproducedbythermdictatestheNorthernrm`sresponse tocompetition.Finally,thismodelcouldbeempiricallytested.DatameasuringIPR-protection strengthexistsalongwithsomedataonresearchprojectspendingalthoughthisdatamay bediculttondformultinationalrms.SimilartoBranstetteretal,momentsof majorIPR-reformwithinacountrycanbeusedempiricallyinordertoseetheresponseto rms.Aspredictedinthismodel,thelargerrmsshouldactuallybedrawnawayfromlocating innovationwithinacountrywhenIPR-protectionisstrengthened.Mymodelalsosuggests anempiricallyobservableanaloguetoeort:researchers.Inconclusion,mymodelprovidesa springboardforanumberofotherinquiresintoRandDacrossnations. 40

PAGE 41

CHAPTER2 INNOVATIVERESEARCHANDDEVELOPMENTOFFSHORINGINNORTH-SOUTH TRADE 2.1Introduction DespitethelowlevelsofresearchanddevelopmentRandDtasksoshoredtothe developingworld,thepracticeisontherise. 1 Usually,thisresearchisadaptiveandforthe purposeofmodifyinganalreadyestablishedproductforsaleinadierentmarket;however, newproductdesign,orinnovativeRandD,forglobaldistributionisthefastestgrowing oshoringsegmentinIndiaLewin,Massini,andPeeters,2009.Intheirindependentsurvey, Garcia-CanalandMartnez-Noyand15.7 % ofoutsourcingrmsoshoredRand Dtodevelopingcountries.MultinationalrmshavebeenexpandinginnovativeRandD operationsintoemergingeconomiesbysplittingRandDprojectsintoseparatetasksthatcan beallocatedtodierentresearchsitesZhao,2006.Withthegrowingtalentofresearchers inthedevelopingworldcoupledwiththelowcostoflabor,multinationalrmsarestartingto locatesomeresearchtasksincountrieslikeChinaandIndia. 2 OshoringRandDtaskstothe developingworldprovidesemploymentopportunitiesinemergingeconomieswhilemakingrms indevelopedcountriesmorecompetitive;however,employeesindevelopingregionsmaytake informationgatheredwhileresearchingthenewproductandstarttheirowncompetingrm. ThisdissertationseekstoexaminetheroleinnovativeRandDoshoringhasonthewelfare ofbothdevelopedcountriesanddevelopingcountriesastheseemergingeconomiesbecome 1 Forexample,between1998-2004,theshareofRandDpersonnelinFDIrmsinChina increasedfrom7percentto20percentLundinandSerger,2008. 2 Moncada-Paterno-Castello,Vivarelli,andVoigtnoteastarkincreaseinRandD outsourcingoverseasandtasksplittinginRandD.Theauthorsalsondthatanincreasing supplyofqualiedscientistsandengineersinthedevelopingworlddrivestheincreaseinR andDoshoringintheseregions.DemirbagandGlaisteruseofdatasetofRandD projectsbyMNEsinbothdevelopedanddevelopingcountries.Lowwagesandthequality ofresearchers,inscienceandengineering,facilitatethelocationofresearchtasksinthe developingworld. 41

PAGE 42

moreattractivelocationsforRandDlabs.Furthermore,thisdissertationpredictstheeects ofglobalizationoninnovativeRandDoshoringtoemergingeconomies. ThemodelpresentedinthisdissertationusesaNorth-Southgeneralequilibriummodel whereNorthernrmsinnovateandSouthernrmsimitate.Northernrmsinventanew productvarietybyconductingRandDtasks.ANorthernrmcanbothconductresearch init'sowncountryandoshoresometaskstothedevelopingSouth.Southernresearchers demandamuchlowerwagethanNorthernresearchers;however,theseSouthernworkersmay alsostealthedesignsforthenewproduct.ANorthernrmfacesaprobabilityofimitation relatedtoitsRandDintensityinboththeNorthandtheSouth.Successfulimitationofa productshiftstheproductionofthatproducttotheSouthernrm.Aftertheproducthas beencreated,rmsselltheirspecicvarietiesinboththeirdomesticandtheforeignmarket. ThenumberofNorthernandSouthernrms,andthereforethenumberofproductvarieties availabletoconsumers,isendogenous.Thedissertationexaminesbothrm-leveloshoring andtheindustry-leveloshoringinequilibrium.Finally,theroleofindustrycharacteristics, SouthernIPR-protection,Southernmarketsize,andtarisonRandDoshoringareexplored. Themodeldeliversseveralndings.FirstProposition1,industrieswithhighRandD intensitiesengageinmoreRandDoshoringperrm;however,thesehigh-techindustriessee largeimitationratesandSoutherncompetition.Withintheseindustries,therearefewNorthern rmsengagedinthishighintensityoshoring.SecondProposition2,forhigh-techindustries, astrengtheningofIPR-protectionlawsintheSouthcauseseachrmtooshoremoreresearch taskstotheSouth.TheSouthishurtfromthestrengtheningofpatentprotectionasless SouthernrmssuccessfullystealNortherninnovations. NextProposition3,alargerSouthernlaborforcedecreasestheamountofoshoring totheSouth.TheincreaseinthelaborforceincreasesthenumberofSouthernrmsin themarketwhichforcesNorthernrmsoutofthemarket.Thus,oshoringdecreases.The growinglaborforceintheSouththenbenetstheSouthwithouthurtingtheNorth.Finally Proposition4,theSouthbenetsfromplacingatarionNortherndierentiatedgoods.As 42

PAGE 43

aresult,theNorthmustchargeahigherpriceintheSouthwhichallowsmoreSouthernrms tocompeteinthemarket.SouthernconsumersbenetatthecostofNorthernconsumers. Furthermore,oshoringtotheSouthdecreasesonanindustrylevelassomeNorthernrms exitduetotheincreasedcompetition. Severalempiricalstudiesareconsistentwiththemainresultsofthisdissertation. DemirbagandGlaistershowrmswhoconductlargeamountsofresearchtendto locatesomeresearchtasksinemergingregions.Whileestimatingalloshoringinanindustry, CanalsandSenershowhigh-techindustriesincreaseoshoringinresponsetoincreased IPR-protectioninthehostcountry.UsingdataonJapanesealiates,ItoandWakasugi ndthathostcountryIPR-protectionhasastrongpositiveeectontheprobabilityoflocating aninnovativeRandDresearchlabinthathostcountry.Finally,CanalsandSenershow anopennesstotradeinahostcountryincreasesthehomecountries`oshoringinthathost country.TheirndingsupportsProposition4inthischapter,whichstatesthattradebarriers decreaseoshoring. Thechapterisorganizedasfollows.Section2.2summarizesrelatedliterature.Section 2.3presentstheassumptionsandfunctionalformsofthemodel.Section2.4establishesthe industryequilibrium.Section2.5showsthegeneralequilibrium.Section2.6presentsthe comparativestatics.Section2.7concludesthechapter. 2.2RelatedLiterature Thisdissertationcontributestoanumberofdierentstrandsofliterature.AsRandD oshoringisatypeofoshoring,thisdissertationcomplementstheglobaloutsourcingand oshoringliteratureasawhole.Acemoglu,Gancia,andZilibottidevelopamodel withoshoringofintermediateinputsinaglobaleconomy.AntrasandHelpman alsopresentaNorth-Southmodelwithoutsourcinganddierentiatedproducts.Likethe modelpresentedinthisdissertation,onlytheNorthernrmscaninnovate;however,rms canoutsourceineithercountryorverticallyintegrateinputproduction.FeenstraandTaylor showtherelationshipbetweenthepriceofoutsourcedinputsandRandDaectthe 43

PAGE 44

amountofoutsourcinginequilibrium.DinopoulosandTsouluhasmodeltheoshoring ofheterogenoustaskstotheSouthwhileassumingSouthernworkersarelessskillfulthantheir Northerncounterpartsandproduceinputsatagreaterqualityvariance.Theauthors'analysis doesnotmaketheformaldistinctionbetweenRandDtasksandcomponentproductiontasks. Themodelpresentedinthisdissertationassumesanothernegativeconsequenceofcontracting Southernresearchers:productimitation.Also,thismodelintroducesanewmechanismfor whichimitationmayoccur:employeemobility.Employeemobilitythenalsoactsasthechannel oftechnologytransfer. 3 Whiledealingwithadierenttypeofoshoring,innovativeRandD, themodelinthisdissertationdealswithboththeoshoringcostandIPR-protection. IndealingwithinnovativeRandD,thisdissertationaddstothegrowingdebateon SouthernIPR-protection.ProductionoshoringandIPR-protectionhavebeenexploredin thepast.Forexample,HelpmanshowsthataSoutherncountrywithNorthernFDI investmentishurtbytighterintellectualpropertyrights.GustafssonandSegerstrom useaNorth-SouthmodelwithincreasingproductvarietytoshowstrongerIPR-protection increasesadaptiveRandDbyforeignaliatesandstimulatesinnovativeRandDfor multinationalrms.ThemodelinthisdissertationshowsthatstrongerIPR-protectionfor high-techindustriesincreasinginnovativeRandDoshoring.Theresultingincreasein oshoringallowsformoreNorthernrmstoenterthemarketwhichbenetsNorthernwelfare atthecostofSouthernwelfare. Thismodelcontributestotheliteratureconcerningtarisandmonopolisticcompetition. Whileexaminingalargecountrywithamonopolisticallycompetitivesector,Felbermayr,Jung, andLarchshowimporttarisincreasetheamountofhomevarietiesanddecreasethe amountofimportedvarieties.DemidovaandRodrguez-Clareshowthatanimporttari 3 Thisdissertationfocusesonintra-countrylabormovementasachanneloftechnology transfer;however,anumberofotherchannelshavebeenexploredinthepast.Alternative channelsoftechnologytransferincludetechnologylicensing,foreigndirectinvestmentFDI, andthetradeofproductsHoekman,MaskusandSaggi,2005. 44

PAGE 45

inasmallcountryonamonopolisticallycompetitiveindustrycanraisethenumberofactive rms.BytransformingthesymmetrictwocountrymodeldevelopedbyVenablesand allocatingtarirevenuetoconsumers,themodelpresentedinthisdissertationexpandson thisliteraturebyintroducingatractablemodeltoexamineoptimaltarisintheNorth-South frameworkwithoshoring. Finally,thepresentmodeladdstotheliteratureonRandDoutsourcingandoshoring dAspremontandJacquemin,1988;ChowdhuryandSheremeta,2010;Kamien,Muller,and Zang,1992.Lai,Riezman,andWangintroducethenotionofinformationleakage whenoutsourcingcost-reducingprocessRandD.EkholmandHakkalahaveexplored endogenousRandDlocationsusingtheknowledgeproleofacountry.Themodelinthis dissertationcontributestotheRandDliteraturebyprovidingalinkbetweenRandD oshoring,welfare,andIPR-protection. 2.3TheModel ThismodelconvertsthesymmetrictwocountrymodeldevelopedbyVenables intoaNorth-Southtrademodel.Twocountriespopulatetheglobaleconomy:Northnand Souths.TheNorthhasrmscapableofcreatingnewvarietiesofadierentiatedproduct tosellinamonopolisticallycompetitivemarket.FirmsintheSouthcannotinnovateandare forcedtoimitate.ImitationofasinglevarietystartscompetitionbetweentheNorthernrm thatinnovatedthevarietyandtheSouthernrmthatimitatedthatvariety.Whenimitationof avarietyoccurs,theSouthernrmpricestheNorthernrmoutofbothmarkets. Laboristheonlyfactorofproduction.TheSouthisendowedwith L s workerswhilethe Northhas L n workers.Bothcountriesalsohaveoutside-goodsectorsproducingproductsunder perfectcompetitionwithconstantreturnstoscale.ThetechnologicallyadvancedNorthern rmsproduceeachunitoftheoutside-goodusingoneunitoflabor.Southernrmsmust use 1 c unitsoflabortoproducetheoutside-good,where c < 1 .Workersaremobilebetween bothsectorsbutnotcountries.Settingthepriceoftheoutside-goodequalto1allowsthe wagerateintheNorthtobexedatunityandthewagerateintheSouthtobexedat c 45

PAGE 46

ConsumerpreferencesineachcountrytaketheformofaCobb-Douglasutilityfunctionforthe representativeconsumer: U i = z 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( i X i i =[ n s ] {1 where z i istheconsumptionoftheoutside-goodincountryiand X i isthesub-utility functionfordierentiatedgoods.Letthissub-utilityfunctionbedenedas: X n = h M n a nn x nn )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 + M s a sn x sn )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 i )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 {2 X s = h M n a ns x ns )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 + M s a ss x ss )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 i )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 {3 where M n istheexpectednumberofrmsoperatingintheNorthand M s istheexpected numberofSouthernrms.Let a ji describeconsumerpreferencesfromcountryiforgoodssold byrmsfromcountryj.Likewise, a ii describesconsumerpreferencesfromcountryiforgoods soldbyrmsfromcountryi.TheseparametersallowthemarketsharesforNorthernrms todierfromthemarketsharesofSouthernrmsthroughafactorotherthanprice. 4 Thus, NorthernrmscanstillcaptureamajorityoftheNorthernmarketdespiteoeringproducts atahigherpricethantheSouthernrms.Also,lettheelasticityofsubstitutionbetween products, ,begreaterthan1.Correspondingpriceindexestoequations-2and-3can nowbedenedas: P n = M n p nn a nn 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( + M s p sn a sn 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( 1 1 )]TJ/F26 5.9776 Tf 5.756 0 Td [( {4 P s = M n p ns a ns 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( + M s p ss a ss 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( 1 1 )]TJ/F26 5.9776 Tf 5.756 0 Td [( {5 4 TheseparametersmayalsobeproductqualityparametersFeenstraandRomalis,2014. Giventhatthermsmustadapttheirproductstoeachspecicmarket,qualityofaproduct mayvaryfrommarkettomarket.TheseparametersleadtoHomeconsumerspurchasingmore ofHomegoodsthanforeigngoods. 46

PAGE 47

where p ji and p ii arethepricesforgoodsincountryiproducedbyrmsfromcountryj andrmsfromcountryi,respectfully.Thepriceindexes-4and-5representthecostof livingintheNorthandSouth,respectively.So,welfareineachcountryisinverselyrelatedto thepriceindexinthatcountry.Eachrmproducesadierentvarietyofthegood.Thus, M n representsthenumberofvarietiessoldbyNorthernrms,and M s representsthenumberof varietiessoldbySouthernrms.Allrmsexport,soconsumersineachregionhavethesame varietyofgoodsavailable;however,thepriceofthegoodsvaryacrosscountries. EveryNorthernrmsetsprice p nn intheNorthwhileeachSouthernrmsetsalowerprice p sn intheNorth.Thesepricesaresetproportionallytomarginalcost.Likewise,intheSouth, consumerscanpurchaseNorthernproductsfor p ns andSoutherngoodsfor p ss .Consumersin theNorththenfaceadierentcostoflivingthanconsumersintheSouthbasedontheprice dierencebetweenthetwotypesofgoods.Finally,thepriceofeachproductisweighedby theconsumerpreferencesforgoodsfromeachcountry.Consumersdonotmindpayinghigher pricesforproductswithaperceivedhigherquality.TheNorthmayhaveamuchlowerprice indexduetothegreaterperceivedqualityoftheproducts,moreNorthernvarietiesavailablein equilibrium,andlowtransportcostsforSouthernrms. Usingatwo-stagemaximizationmethod,consumersallocateexpendituresbothwithin sectorsandacrosssectors. 5 Assumingfree-entryineachsector,theincomeintheNorth measuredintermsoftheoutside-goodis L n .Likewise,theSouth'sincomeis cL s .Usingthe indirectutilityfunctionsforeachcountry,welfareperconsumercanbedened: W n = B P n {6 W s = Bc P s {7 5 SeeAppendixB.1fordetailsonconsumermaximization. 47

PAGE 48

where B = )]TJ/F24 11.9552 Tf 12.249 0 Td [( 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( .Asseenfromequation-6,eachNorthernworker'swelfare, W n ,increasesastheNorthernpriceindexdecreases.Likewise,welfareperworkerintheSouth, W s ,increasesastheSouthernpriceindexdecreases.Consumerdemandsforthedierentiated sectorare: P n X n = L n {8 P s X s = cL s + TR {9 where TR isSouthernimporttarirevenuedispersedtoeachconsumer.Untilsubsection 2.6.4,Iassumethereisnoimporttari.Inotherwords, TR =0 .Withinsectorbudgeting resultsinconsumersallocatingtotalspendingondierentiatedgoodsbetweentheindividual productstakingintoaccounteachproduct'sprice.Thisbudgetingyieldsthedemands: x nn = p )]TJ/F25 7.9701 Tf 6.587 0 Td [( nn a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 nn P n X n {10 x sn = p )]TJ/F25 7.9701 Tf 6.587 0 Td [( sn a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 sn P n X n {11 x ns = p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ns a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ns P s X s {12 x ss = p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ss a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ss P s X s {13 Northernrmsmustpayaxedcostofenteringthemarket.Thisxedcosttakesthe formofresearchanddevelopmentRandD. 6 Firmsmustpayworkersinordertocomplete researchtasks.Allrmsmustcompleteacertainnumberoftasks, R ,inordertocreateanew product.Eachresearchercompletesoneresearchtask.Finally,rmscanoshoresomeresearch taskstotheSouth.Northernxedcostofentrycanthenbewrittenas: f e = w n R n + w s R s = R n + c R s 6 InternationaltradeliteraturehaspreviouslyinterpretedthexedcostofentrywithRand DMelitz,2003. 48

PAGE 49

where R n denotesthenumberofNorthernresearchers,and R s denotesthenumberof Southernresearchersemployedbytherm.WhiletheNorthernrmcancutcostsbyusing Southernlaborforresearch,weakIPR-protectionintheSouthcanresultinimitationfroma Southerncompetitorwithaprobability .Theprobabilityofimitationfromaninformation leakispositivelyrelatedtotheamountofRandDdoneintheSouth.Morespecically,I denetheprobabilitiesas: = R s R n + R s 0 << 1 {14 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( = R n R n + R s {15 UsingthefunctionalformofaTullocklottery,theamountofNorthernresearchers canbeseenastheNorthernrm'seorttoretainthepatent. 7 Likewise,thenumberof SouthernresearchersworkingfortheNorthernrmistheeortextendedbytheSouthern rmtosuccessfullycopythenewproduct.Southernresearcherscanleavetoworkatthe Southernrmafterinternalizingknowledgeusedinthecreationofthenewproduct;however, governmentIPR-protectionbarriersandotherbarrierstoimitationpreventthisworkerfrom fullythreateningtheNorthernrm'sclaimonthepatent.Barrierstoimitationcanberelated tothecomplexityofproject.Increasingthecomplexityofaproductcancomplicatethereverse engineeringofthatproductbycompetitors.Likewise,barrierstoimitationcanbeformallabor agreements,suchasnon-competeclausesincontracts,thatpreventcurrentemployeesfrom contributingtocompetitors. Theparameter capturesthebarriersassociatedwithimitation.Ispecicallyfocuson theinstitutionalbarriersthatpreventformalemployeesfromcreatinganimitation.So, 7 TulluckcontestshaveoftenbeenappliedtoRandDracesinpreviousstudiesBayeand Hoppe,2003.Leiningerincludesascalingparameterontheeortofoneplayerin ordertoreectdierentialadvantages.BaikandFonsecaexploretheproperties ofasymmetricTullocklotteries.Thismodelusesasingleplayerchoosingbotheortlevels,or thenumberofresearchersineachregion. 49

PAGE 50

capturestheexogenousIPR-protectionenvironmentintheSouth.A nearzeroindicates strongSouthernIPR-protection,whilea near1indicatesweakIPR-protection.Oncethexed costofentryhasbeenpaid,rmsmustproduceandsellproducts.NorthernandSouthern rmshavethesametechnologyofproduction.Producingaunitrequires laborers.Therefore, themarginalcostofproductionforNorthernrmsistheNorthernwagemultipliedbythe numberofworkersneededtomakeoneproduct: .Themarginalcostofproductionforthe Southernrmisthenc .Northernexportingrmsfaceicebergtradecosts n whenexporting thedierentiatedgoodtotheSouth.TheNorthernrmalsofacesthepossibilityoflosingthe SouthernmarkettoaSouthernimitator.TheNorthernrmfacesaprobabilityofimitationof .Northernexpectedprotcanthenbedenedas: E n = )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ x nn p nn )]TJ/F24 11.9552 Tf 11.955 0 Td [( + x ns p ns )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.956 0 Td [( n )]TJ/F38 11.9552 Tf 11.956 0 Td [(t ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R s {16 TheNorthernrmmustpayatariofamountttotheSouthwhenexportingitsproduct. Untilsubsection2.6.4,thetariissettozero.Southernrmsareonlyabletoenterthemarket byattemptingtocopyavariety.ExportingSouthernrmsmustpayatariof s 8 Finally, Southernrmshireanumberofadditionalworkers, f s ,toattempttoimitatetheproduct successfully: E s = [ x sn p sn )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F24 11.9552 Tf 11.955 0 Td [( s + x ss p ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(c ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s {17 8 Tradecostscanvaryvastlybetweencountries.DierenttradecostsfacedbytheNorth andtheSouthhelpfacilitateanequilibriumwherebothcountriestrade.Thisdierencein tradecostsmaybeduetoadierenceininfrastructureBeharandVenables,2011. 50

PAGE 51

Fillinginequations-14and-15andreplacing R s with R )]TJ/F38 11.9552 Tf 12.208 0 Td [(R n ,theexpectedprot functionscanberedened: E n = R n R n )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R [ o n ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R )]TJ/F38 11.9552 Tf 11.956 0 Td [(R n {18 E s = R )]TJ/F38 11.9552 Tf 11.956 0 Td [(R n R N )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R [ o s ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s {19 where o n and o s denoteoperatingprotforNorthernandSouthernrms,respectively. Formally,theseequationsare: o n = x nn p nn )]TJ/F24 11.9552 Tf 11.955 0 Td [( + x ns p ns )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( n {20 o s = x sn p sn )]TJ/F38 11.9552 Tf 11.956 0 Td [(c )]TJ/F24 11.9552 Tf 11.955 0 Td [( s + x ss p ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(c {21 2.4IndustryEquilibrium Inthissection,Iexploretheindustryequilibrium.EachNorthernrmmustdecideon howmuchresearchtooshoretotheSouth.Southernrmsdecideonentryintothemarket. Anyrminthemarketmustalsosetprices.Subsection2.4.1addresseachindividualNorthern rm'sresearchchoicewhilesubsection2.4.2examineshoweachrmsetspricerelativeto marginalcost.IonlyexaminethecasewhereeachNorthernrmsplitsresearchtasksbetween theNorthandtheSouth.Insection2.2.5,thenumberofrmsthatcanexistineachcountry inequilibriumisoutlined. 2.4.1Researchlevels TheNorthernrmsolvesthefollowingmaximizationproblemtakingoperatingprotas given: max R n E n Southernrmsonlydecidewhethertoenterthemarketortonotenterinthisstage.The Northernrmdoesnothiremoreresearchersthanisnecessarytocompletetheinnovation. Likewise,thermcannothireanegativeamountofresearchers.So, R n isboundedby0and 51

PAGE 52

R .With 0 << 1 ,arst-orderconditionyieldsamaximumvalueofNorthernresearch.That is,theexpectedprotfunctionisconcavefor R n 2f 0, R g 9 Thetypeofsolution,interior orcorner,canbedeterminedbylookingattherelationshipbetweentotalresearch,operating prot,IPR-protection,andthewagegap.Aninteriorsolutionexistsi: [ n ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(c < R < [ n ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(c {22 ThevalueofequilibriumNorthernoperatingprot n isdeterminedinsection2.5.I assumethatcondition-22holdsinequilibrium;therefore,everyNorthernrmlocates researchtasksinboththeNorthandtheSouth. 10 TakingtheFOCandrearrangingyieldsthe optimalNorthernresearchlevel: R n = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F21 11.9552 Tf 11.956 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R {23 Theprobabilityofimitationisthen: = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( {24 2.4.2Prices Afterrmspayxedcosts,theymustchoosethepriceoftheproducttomaximize operatingprot.Maximizationofoperatingprotequations-20and-21yieldthe followingpricesinequilibrium: p nn = )]TJ/F21 11.9552 Tf 11.956 0 Td [(1 p ns = + n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 {25 p ss = c )]TJ/F21 11.9552 Tf 11.956 0 Td [(1 p sn = c + s )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 {26 9 ThesecondorderconditionsareestablishedintheAppendixB.2. 10 AppendixB.2.3denesthisconditioninjusttheparametersofthemodel. 52

PAGE 53

TheSouthernrmmustbeabletosetalowenoughpricethattheNorthernrmcannot competeineithermarket.IassumethehighestpricechargedbytheSouthernrmislower thantheNorthernrm'smarginalcost.So, p sn < ensuresthattheNorthernrmmust dropoutofthemarketafterimitationoccurs.Thisassumptionholdsunderthefollowing condition: 11 c + s < )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 {27 JustasinVenables987,Idenetheconvenientparameters, s nn s sn s ns ,and s ss ,to simplifythediscussionoftheequilibrium: s nn = p nn = a nn 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s ns = p ns = a ns 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( {28 s ss = p ss = a ss 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s sn = p sn = a sn 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( {29 Thesevaluescanbeusedtoshowthemarketsharesthatagivenrmhasinagiven marketrelativetoarminthesamemarketfromtheothercountry. 12 Imaketwoassumptions thatimplyarelationshipbetweenthemarketshares.First,Iassumethatconsumersinthe Northslightlypreferproductsmadebyrmsfromtheirowncountry.Second,consumersin theSouthpreferproductsfromSouthernrms.Thatis, a sn < a nn and a ns < a ss .Itproves convenienttodenethisimplicationintheformofasingleassumption: Assumption1. s ns s sn < s nn s ss 11 Condition-27isdeterminedusingequation-26and p sn < 12 Thesevaluesdonotimmediateresemblemarketshares.Usingequations-10,-11, -12,and-13,thesevaluescanberewrittenas: s ij = p ij x ij P j X j .So, s ij s jj = p ij x ij p jj x jj .Thus,thisratio expressesthemarketsharearmfromcountryihasinmarketjrelativetothemarketshareof armfromcountryjinthesamemarket. 53

PAGE 54

Assumption1impliesthateachrm'sdomesticmarketshareislargerthanitsforeign marketshareduetoconsumerpreferencefordomesticrms. 13 2.5GeneralEquilibrium Theequilibriumofthiseconomycannowbedenedusingfreeentryconditions.With theoptimalNorthernresearchlevelandtheexpectedprotfunctions-18and-19,two free-entryconditionscanbedenedas: E n = n )]TJ/F21 11.9552 Tf 11.955 0 Td [(2 )]TJ/F24 11.9552 Tf 5.48 -9.684 Td [( R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c n 1 2 )]TJETq1 0 0 1 327.445 562.087 cm[]0 d 0 J 0.478 w 0 0 m 8.995 0 l SQBT/F38 11.9552 Tf 327.445 552.111 Td [(R c )]TJ/F24 11.9552 Tf 11.955 0 Td [( =0 {30 E s = s )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( cf s n 1 2 h R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( n 1 2 i )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 =0 {31 Thesetwoequationsdeterminetheequilibriumlevelof n and s .ThelevelofNorthern operatingprotintermsofthemodel'sparametersis: n = R h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 {32 Usingequations-8and-9,theindividualdemandsfordierentiatedgoods -10--13,andequations-28and-29,equations-20and-21canbeused todenetheoperatingprotofagivenrminequilibrium: n = 1 s nn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n L n + s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s cL s {33 s = 1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n L n + s ss P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s cL s {34 13 Intradeliterature,thisimplicationisreferredtoasa"homebias"inconsumption.Many authorshavedocumentedindividualcasesofhomebiasBaltaandDelgado,2009;Obstfeld andRogo,2000.Generally,thehomebiasinconsumptionisexplainedbytradecostsand apreferencefordomesticproducts.Friberg,Paterson,andRichardsonshowstrong preferenceforU.S.wineintheU.S.marketdrivesthehomebias.Inthemodelpresentedin thisdissertation,strongpreferencesfordomesticgoodswouldbedescribedas a ii > a ji .Evans ndsevidencethattradetarisexplainalargeamountofthehomebias.Inthiscase, a ii and a ji wouldberelativelysimilar. 54

PAGE 55

Notethatequilibriumprotlevelsmustbepositive.Ifoneofthevalueswerenegative, thenoneorbothofthepriceindexeswouldhavetobenegative,implyinganegativenumber ofrmsinatleastoneregion.Iusetwosucientconditionstoguaranteeauniqueinterior solution: 14 Assumption2. < c Assumption3. c + < 1 Assumption2statesthatthetheIPR-environmentmustbesucientlystronginorderto attractanyinvestmentintotheSouth.Assumption3indicatesthatSouthernresearchrequires bothasmallenoughwageandasucientlystrongIPR-environmenttobeprotable. 15 Giventheseassumptions,rmsarewillinglytoriskimitationbyusingSouthernresearchers. Inotherwords,Assumptions2and3guaranteethatcondition-16holdsandthatthe equilibriumoperatingprotsforNorthernandSouthernrmsarepositive;however,this doesnotguaranteepositivepriceindexes.Equations-23,-24,andtheknownvaluesof operatingprotinequilibriumdeterminethepriceindexes.Bothindexesmustbepositive.This canbeensuredifthefollowingconditionholds: s sn n s nn < s < s ss n s ns {35 Condition-35guaranteesthatbothpriceindexesarepositive;however,thecondition doesnotimplythattheequilibriumvaluesof M n and M s arepositive.Anecessaryand 14 IderivetheseconditionsandtheexplicitformforNorthernoperatingprot-32in AppendixB.2.3. 15 Assumptions2and3havesomeempiricalsupport.Otherauthorshaveshownthatthese conditionsfacilitateRandDoshoring.Forexample,DemirbagandGlaistershow thatlowwageattractsmultinationalRandDinvestmentinthedevelopingworld.Itoand WakasugindthatIPR-protectionhasastrongpositiveeectonthelocationofRand Dresearchlabs. 55

PAGE 56

sucientconditionforbothregionstohaveactivermsinthedierentiatedgoodssector: s sn s ss < P s P n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 < s nn s ns {36 Usingequations-4and-5withtheoptimalpriceindexes,theequilibriumnumberof activermsineachcountrycanbefound: M n = s ss P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn {37 M s = s nn P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn {38 Inspectionofcondition-36showsanumberoffactorsinuencetheexistenceof bothcountrieshavingactivermsinthedierentiatedgoodssector.First,ahighSouthern consumerpreferenceforSoutherngoodsfacilitatesSouthernentry.Second,alowSouthern productioncostrelativetotheNorthernproductioncostandtransportcostmakeSouthern entrymorelikely.Finally,alargepopulationintheSouthrelativetotheNorthfacilitatesthe existenceofSouthernrms.Iexaminethecasewhereconditions-35and-36hold. Figure2-1graphicallyshowsthiscase.Theintersectionofthetwoequilibriumoperating protlinesdeterminetheequilibriumpriceindexesandexpectednumberofrms.The M n =0 and M s =0 linesshowthecombinationofNorthernandSouthernpriceindexesthatyieldzero rmsintheNorthandSouth,respectively.Anycombinationofpriceindexesthatfallbelow the M n =0 lineyieldsapositivenumberofNorthernrms.Likewise,anycombinationofprice indexesthatlieabovethe M s =0 lineyieldsapositivenumberofSouthernrms. 16 2.6ComparativeStatics Assuminganequilibriumwherebothcountrieshavermsengagedinproducingandselling dierentiatedgoods,thetwofreeentryconditionscanbeusedtoanalyzethecomparative staticsfromanyexogenouschange.Thissectionexaminestheeectofanumberofexogenous 16 AppendixB.2.5showsthecaseswhereonlyonecountryhasapositivenumberofrmsin thedierentiatedgoodssector. 56

PAGE 57

Figure2-1.TradingEquilibrium changesonwelfare,oshoring,andwageinequality.LetthetotalnumberofSouthern researchersemployedbe O s .ThisvariablecapturestheamountofoshoringintheSouth. Idene O s asthenumberofNorthernrmsmultipliedbyeachrm'sSouthernRandD intensity,or O s = M n R s .Finally,lettherealwagegapbetweenNorthandSouthbedened as = 1 = P n c = P s = P s cP n .Subsection2.6.1examinestheroleofNortherntechnologyandrmsize. Subsection2.6.2examinestheeectofstrengtheningSouthernIPR-protectionwhilesubsection 2.6.3increasestheSouthernmarketsize.Finally,subsection2.6.4introducesaSoutherntari. 2.6.1IndustryTypeandFirmSize Thissubsectiondealswithindustrycharacteristicsandtheireectonoshoring.Firms withhigherRandDspendingtendtospreadtheirresearchacrossmultiplecountriesIto andWakasugi,2007.Inordertoexaminetheroleofindustrytype,Iexaminethecaseof anincreaseinthenecessaryamountofresearchneededtoproduceanewproduct.Inother words,theproductrequiresrmstoincorporateanotherelementintothenalgoodinorder tobesuccessfulinthemarketplace.IndustriesfacinghighRandDrequirementscanbe seenashigh-techindustries.While R hasbeeninterpretedasaNorthernrm'stotalRand 57

PAGE 58

Figure2-2.AnIncreasein R Dintensity,itcanalsobeseenasaproxyforrmsize. 17 Proposition1canalsobeusedto examinetheoshoringdierencebetweenindustriespopulatedbylargermsandindustries withmanysmallrms. Proposition1. Anincreaseintotalresearch R ,oranincreaseinNorthernrmsize, decreasesNorthernwelfarepercapita W n # ,increasesSouthernwelfarepercapita W s andmayincreaseordecreaseoshoringtotheSouth O s "# .Realwageinequalitybetween theNorthandtheSouthdecreases # Proof. SeeAppendixB.3.1. Theresultsofanincreasein R areoutlinedinFigure2-2.TheNorthernequilibriumprot lineshiftsright.They-interceptshiftsupas n increases.Likewise,thex-interceptshiftsright. TheSouthernequilibriumprotlineshiftstotheleft.Thenewequilibriumpoint E 0 therefore mustbeaboveandtotheleftoftheoldequilibriumpoint E 17 FirmsizetendstobepositivelyrelatedtoRandDspending.SeeSymeonidis. 58

PAGE 59

Intuitively,asrmsrequiremoretotalresearchtasks,theyassignmoretaskstothe relativelycheapSouthernresearchers.Thisincreasestheprobabilityofimitation;however,the probabilityofimitationdecreasesasNorthernrmsemploymoreNorthernresearchersaswell. Thesetwoeectsoseteachother,andtheprobabilityofimitationdoesnotchange.Some Northernrms,facinghigherxedcosts,exitthemarketwhichallowsadditionalSouthern rms.LessdomesticvarietiesintheNorthdecreasesNorthernwelfarewhileSouthernwelfare increasesasSouthernconsumersgainSouthernvarieties.ThenumberofNorthernrms decreases;however,eachrmemploysmoreSouthernresearchers.TheRandDintensity withineachcountryincreasesforeveryNorthernrm.Oshoringthenincreasesithe decreaseinNorthernrmsoutweighstheincreaseineachrm'soshoring.Employmentin theproductionofdierentiatedgoodsisdirectlyrelatedtothenumberofrmsoperating ineachregion.So,thedecreaseinthenumberofNorthernrmsimpliesthatemployment inproductionintheNorthdecreases.Likewise,employmentintheproductionofSouthern dierentiatedgoodsincreases. Proposition1impliesthatrmsinhigh-techindustriesoshoretotheSouthmoresothan rmsinlow-techindustries.Likewise,largermsengageinmoreoshoringthansmallrms. Proposition1isconsistentwithanumberofempiricalstudies.DemirbagandGlaister ndthatthelikelihoodoflocatingresearchtasksinemergingregionsincreasesasthenumber ofRandDprojectsincreasesforamultinationalrm.HedgeandHicks007noteintheir datasetthathigh-techU.S.industries,likecomputersandelectronics,engageinsignicantly moreoverseasRandDthanlowtechindustries,likeagriculture. 2.6.2IPR-Protection Thenumberofpatentshasbeenontheriseintherecentyears.Forexample,theUS patentocegranted326,033patentsin2014andonly181,299in2004U.S.PatentStatistics 59

PAGE 60

Figure2-3.AStrengtheningofIPR-ProtectionwithaLarge R Report,2015. 18 Asknowledgeintensivegoodshavebecomemorecommonlytraded,IPRlaws indevelopingcountrieshavebeenthesubjectofdebatebyworldleaders.StrongerIPRlawsin theSouthcanencourageinvestmentandinnovationintheSouth;however,theselawscould hinderthegrowthofindustriesintheSouthbypreventingrmsfromenteringthemarket. SimilartoHelpman'smodelpresentingstrongerIPR-protectionasadecreaseinthe rateofimitation,SouthernIPR-protectionaectstheprobabilityofimitationinthemodel presentedinthisdissertation.AstrengtheningoftheSouthernIPRenvironment,oradecrease intheSouthernrm'seectivenessatimitatingadecreasein ,placesmorelegalbarriers preventingemployeemobilityandproductimitation.Forhigh-techindustries,thetighteningof IPR-protectionbenetsNortherninnovatorsatthecostofSouthernimitators. Proposition2. Foranindustrywithlargeresearchneedslarge R ,astrengtheningofIPRprotection # increasesNorthernwelfarepercapita W n ,decreasesSouthernwelfareper 18 Likewise,patentrightsworldwidehaveexpandedinthepastcentury.Intheirdatasetof 110countries,GinarteandParkshowtheaveragevalueoftheirpatentrightsindex increasesby15.5 % from1960to1990. 60

PAGE 61

capita W s # ,andincreasesoshoringtotheSouth O s .Realwageinequalitybetweenthe NorthandtheSouthworsens # Proof. SeeAppendixB.3.2. Figure2-3showstheresultofadecreasein .Theequilibriumpointshiftsdowncausing adecreaseintheNorthernpriceindexwhiletheSouthernpriceindexincreases.MoreNorthern rmsenterthemarket,andSouthernrmsareforcedtoexit.Theintuitionbehindthe downwardshiftsinbothequilibriumprotlinesliesintheprobabilityofimitation. NorthernrmsrespondtothisincreaseinIPR-protection,andtheexogenousdecrease intheprobabilityofimitation,byshiftingresearchtaskstotheSouth.Thisshiftofresearch tasksdecreasesthexedcostofentryforNorthernrmsbutincreasestheprobabilityof imitation.Forrmsfacingalargevalueof R ,theexogenouschangeintheprobabilityof imitationoutweighstheendogenouschange.NorthernrmsareheavilyinvestedintheSouth tosaveonxedcosts.Thus,theNorthernrmsarelessreactivetochangesintheprobability ofimitation.Atleastonecountry'swelfareincreasesasIPR-protectionstrengthens.Both countriescanindeedbenetfromthereductioninIPR-protection.When R islargeenough, toomanySouthernrmsexit,andSouthernwelfaredecreases.EachNorthernrmincreases theamountofresearchthatitoshorestotheSouth.Oshoringthenmustincreaseonan industrylevel.Southernemploymentinproductionofthedierentiatedgooddecreasesthough asthestrongIPR-environmentforcesSouthernrmstoexit.So,thedierentiatedgoods sectorshrinksintheSouthwhileitexpandsintheNorth. OtherauthorshavefoundthattighterIPRlawscanharmtheSouthundercertain conditionsHelpman,1993;DinopoulosandSegerstrom,2007.StrongerIPR-protection isoftenassociatedwithmoreRandDinvestmentbytheNorthandadecreasesinthe relativewageofNorthernworkersGustafssonandSegerstrom,2011.Additionally,a numberofempiricalstudiesndthatstrongIPR-protectioninahostcountryincreases innovationoshoringtothatcountryDemirbagandGlaister,2010;ItoandWakasugi,2007; Moncada-Paterno-Castelloetal,2011.Thisdissertationpositsthatrmsinhigh-tech 61

PAGE 62

Figure2-4.AnIncreaseintheSouthernLaborForce industries,orthosewithalargexedcostofentry,increasesoshoringmoreduetoa strengtheningintheIPR-environmentthanrmsinlow-techindustries.WhileCanalsand Senershowhigh-techindustriesincreaseoshoringduetothestrengtheningof IPR-protection,noempiricalstudieshaveconrmedthatinnovativeRandDoshoringis moreresponsivetotheIPR-environmentthaninnovativeRandDoshoringfromlow-tech industries.Furthermore,noempiricalstudyhaslookedathowrmsizeaectsarm's oshoringdecisionwhenfacedwithIPR-protectionchanges. 2.6.3SouthernLaborForce DevelopingnationssuchasBrazil,Russia,India,China,andSouthAfricaortheBRICs haverapidlygrownasglobalpowers.Developedcountriesexpandedtheirinteractionswith thedevelopedworldthroughtheadditionofnewcustomerswithdisposableincomeandnew oshoringopportunities.ThetypicalfearassociatedwiththeemergenceoftheseBRICsas possibleoshoringsitesisthatNorthernworkerswillloseemployment.Inthismodel,the emergenceofthesecountriesontheglobalscalecanbeseenasanincreaseintheSouthern laborforce L s .Theincreasein L s eliminatesbothNorthernRandDjobsandNorthern productionjobs. 62

PAGE 63

Proposition3. AnincreaseintheSouthernlaborforce L s doesnotchangeNorthern welfarepercapita W n $ ,increasesSouthernwelfarepercapita W s ,anddecreases oshoring O s # .RealwageinequalitybetweentheNorthandtheSouthdecreases # Proof. SeeAppendixB.3.3. AnincreaseinthesizeoftheSouthincreasestheamountofincomeintheSouth.Figure 2-4illustratesthiscase.AnincreaseinSouthernincomedoesnotaecttheresearchintensities ofeachNorthernrm;however,theincreaseintheSouthernmarketaectsthenumberof Northernrmsinequilibrium.Thex-interceptofeachequilibriumoperatingprotlineshifts down.TheSouthernpriceindexmustdecreaseasadditionalSouthernrmsenterthemarket. Proposition3reectsthefollowingconsiderations.SincetheSouthhasbecomea largermarket,moreSouthernrmsenter.Thegrowthinpopulationincreasesdemandfor dierentiatedgoodsintheSouthbutleavesthedemandintheNorthunchanged.The additionalSouthernrmsforceoutsomeNorthernrms.WelfareintheNorthdoesnotchange whiletheSouthernwelfareincreases.Sincetheprobabilityofimitationisunaected,each rm'soshoringdecisionisnotchangedbythesuddenincreaseinthesizeoftheSouth. Therefore,theamountofoshoringdecreasesasthenumberofNorthernrmsshrinks.Inthe shortrun,manyNorthernworkersaredisplacedbytheincreaseinSouthernsizewhileonly Southernresearchersaredisplaced. Proposition3qualiesanumberofpreviousndings.Forexample,Dinopoulosand Tsoulouhas015showthatincreasesinthesizeoftheSouthinaperfectlycompetitive economywithoshoringofheterogenoustasksresultsinanincreaseintherangeoftasks oshoredtotheSouthwhileworseningthewageinequalitybetweenthetwonations.With monopolisticallycompetitivemarketsandoshoringsearchcosts,GrossmanandHelpman ndtheamountofinputoutsourcing,notjustRandDoutsourcing,totheSouth increasesasthesizeoftheSouthincreases.Dierentiatingitselffromthesepreviousmodels, themodelinthisdissertationusestheadditionalSouthernrmsthatresultfromalargerSouth asanexplanationforindustryleveldecreasesinoshoring. 63

PAGE 64

Figure2-5.IntroducingaSouthernImportTari 2.6.4SouthernImportTari Withmoredevelopingcountriesenteringfree-tradeagreementswiththedevelopedworld, Southerncountriesarebecomingrestrictedintheirabilitytofreelyplaceanimporttari.In amonopolisticallycompetitivemarket,atarifortheSouthcanbeanimportanttoolfor generatingtaxrevenueandprotectingdomesticrms.Thefollowingsubsectionshowsthatthe Southbenetsfromsettinganimporttari. Proposition4. TheSouthintroducinganimporttarionNortherndierentiatedgoods t > 0 decreasesNorthernwelfarepercapita W n # ,increasesSouthernwelfarepercapita W s ,anddecreasesoshoringtotheSouth O s # .RealwageinequalitybetweentheNorth andtheSouthdecreases # Proof. SeeAppendixB.3.4. Figure2-5showsProposition4graphically.WhentheSouthplacesatarionNorthern imports,theNorthernequilibriumprotlineshiftsoutduetotheincreasethepricesetby NorthernrmsintheSouth.TheNorthernequilibriumprotlineshiftsdownaswellduetothe increaseinSouthernincome;however,theshiftinincomealsoshiftstheSouthernequilibrium 64

PAGE 65

protlinedown.MoreSouthernrmsenterthemarketasaresultwhileNorthernrmsexit. TheincreaseintheSouthernmarketplaceisnotenoughtoosettheforcedincreaseinprice thatNorthernrmsmustendurefromthetari. ThetariontheNortherngoodintheSouthincreasesthepricechargedbyNorthern exporters,makingthemlesscompetitiveintheSouth.SomeNorthernrmsarenotableto stayinthemarket.TheNorthernconsumerslosewelfarebythelackofhomevarietieswhile theSouthernconsumersgainwelfarefromanincreaseinSouthernvarieties.TheSouthalso gainstarirevenuebyintroducingatariwhichisthentransferredtoconsumers.Thetari thenincreasestheincomeofeachconsumerwhileincreasingthenumberofSouthernrms. AsNorthernrmsexitthemarket,researchersineachcountrylosetheirjobs.Theoshoring decisionofeachrmleftinthemarketisunaectedbythechangeintari.Acontracting dierentiatedgoodssectorintheNorthmeansNorthernproductionandresearchworkersare displacedasrmsclose.Thus,theSouth'starionNortherngoodsbenetsSouthernworkers andharmsNorthernworkers. SimilartoothermodelsoftarisinmonopolisticallycompetitivemarketsDemidova andRodrguez-Clare,2009;Felbermayr,Jung,andLarch,2013,theimporttariinduces moreSouthernrmstoenterthemarket.Thereisalsoanunambiguouspositiveeectonthe South'swelfarewhentheSouthsetsanimporttari.Increasingthetarimaybegindecreasing tarirevenueandSouthernwelfareperworkerasthelackofNorthernrmsdecreasesfurther. Whilethemodelinthisdissertationdoesnotlookatthecaseofincreasingthetari,the frameworkforexamininganoptimaltariisinplaceforfutureresearch.Anempiricalstudy ontarisandRandDoshoringisstillneeded;however,CanalsandSenerndan increaseintariscandecreaseindustryoshoringofalltasks. 2.7Conclusion IexamineaglobaleconomywhereNorthernrmsinnovateusingNorthernandSouthern researcherswhileSouthernrmsimitate.TomatchthegrowingtrendinRandDoshoring, thisdissertationassumesthatrmsareabletosplitresearchtasksbetweenmultipleresearch 65

PAGE 66

sites.Northernrmsizehasanambiguouseectonindustryleveloshoring.Firms'R andDintensityintheSouthincreaseswithrmsize;however,fewerNorthernrmscan existinthemarket.ANorthwithahigh-techdierentiatedindustrybenetsfromstrong IPR-protectionlawsintheSouthwhileSouthernworkersareharmeddespitetheincreasing oshoringemployment.AlargerSouthdecreasestheoverallamountofoshoringwhile increasingSouthernwelfarewithoutaectingNorthernwelfare.AnimporttariintheSouth alsodecreasesoshoring. Asglobaltradeincreases,afewoutcomesareevident.Free-tradeagreementsstrengthen SouthernIPR-protectionandlimitstheSouth'sabilitytosetimporttaris.ThestrongIPR lawsandreductioninNortherntradecostsraiseNorthernwelfare,lowerSouthernwelfare,and increasetheamountofoshoring.Globalizationmaythenleadtomoreoshoringandgreater NorthernwelfareatthecostofSouthernwelfare.Therefore,developingcountrieshavean interestinkeepingcontrolovertarisonNortherngoodsandIPR-protection. Severalextensionsofthisanalysiswarrantfurtherresearch.Forexample,themodelcan beusedtoexamineheterogenousrmsthatdierinsize.Themodelmayalsobeadapted toexamineotherRandDtypessuchasqualityincreasingRandD.Themodelcanalsobe expandedtoincludemorethantwocountries.Multinationalrmsoftennditadvantageous tosplitresearchacrossmanydevelopedanddevelopingcountriesinordertopreventimitation furtherZhao2006.Finally,thisdissertationsuggestsempiricalpossibilitiesthathavenotyet beenexplored.Thatis,fewstudieshaveexaminedtheroleoftarisandRandDoshoringin North-Southtrade.FirmlevelRandDintensityshouldbeunchangedasaresultofchanges intaris.ThesizeoftheSouthisalsoassociatedwithindustryleveloshoringandnotrm intensity.Also,thedissertationsupposesthattheNorthcountry`sRandDemployment changesasfactorsintheSouthchange.Forexample,anincreaseinthesizeoftheSouth shouldyieldlessNorthernRandDemployment.Theserelationshipshaveyettobeveriedin tradeliterature. 66

PAGE 67

CHAPTER3 INNOVATIVERESEARCHANDDEVELOPMENTOFFSHORINGTODEVELOPING COUNTRIES:EVIDENCEFROMUSMULTINATIONALFIRMS 3.1Introduction Untilrecently,amajorityofresearchanddevelopmentRandDtasksoshoredtothe developingworldwasfortheadaptationofproductsforsaleinlocalmarkets.Forexample, Sun,Du,andHuangshowamajorityoftheforeignRandDin2002inChinawas adaptiveRandD.WeakIPR-protectionhasdeterredmultinationalrmsfromlocating higherlevelsofRandDtasks,suchasinnovativeRandDforthecreationofanewproduct; however,lowwagesandtheavailabilityofscientistsandengineersintheseregionshas driventherelocationofresearchtaskstodevelopingcountriesDemirbagandGlaister,2010; Moncada-Paterno-Castello,Vivarelli,andVoigt,2011.Multinationalrmshavebegunlocating theirrepresentativeRandDtasksindevelopingcountriesGarcia-CanalandMartnez-Noya, 2014.Withanincreaseinglobalization,IPR-protectioncanbeexpectedtostrengthenin thedevelopingworld,andimporttariscanbeexpectedtofall.Thisdissertationempirically examinestheeectsofrmsize,technologytype,IPR-protection,importtaris,andmarket sizeonmultinationalinnovativeRandDoshoringtodevelopingcountries.Furthermore,this studyshowsthathigh-techrms,orrmsfromindustriesthatdolargeamountsofRandD, increaseIPR-environmentbasedontheirsizeandresearchneeds. Thisstudycollectstwosetsofdatatotestanumberofhypotheses.UsingtheUSPTO andNBERdatabases,therstdatasetcollectsrmleveldataonpatentsledforUS corporationsfrom1970to2006.Aninventorwithacurrentaddressinadevelopingcountry thatisrecordedonthepatentsignalsthatarmhassoughtresearchersinadeveloping country.UsinganOLSmodelandTobitmodelwiththenumberofforeigninventorsandthe RandDintensityasthedependentvariables,Iexplorehowmultinationalrmschoosethe numberofresearchtaskstolocateinthedevelopingworld.Independentvariablesofinterest includeIPR-protection,high-techandlow-techindustries,importtaris,andhostcountrysize. UsingBEAdata,IcollectaseconddatasettoexamineindustrylevelUSmultinationalRand 67

PAGE 68

Dactivitiesin35developingcountriesfrom1999to2013.Ialsouseadierence-in-dierence modeltoexaminethesensitivityoftheresultsofthisdissertationtodierentIPR-protection measurementmethods. Theempiricalanalysisdeliversseveralndings.First,high-techrmsaremorereactiveto IPR-protectioninhostcountriesthanlow-techrms.High-techrmsaremorelikelytohave largeresearchneedsandthusaremorewillingtolocateresearchtasksindevelopingcountries asIPR-protectionstrengthens.TheIPR-reformsallowthesehigh-techrmstogreatlydecrease theirresearchcostsbyshiftingresourcesintothedevelopingworldwithlessriskofproduct imitation.High-techindustriesalsoincreasetheiroshoringactivitiesmorethanlow-tech industrieswhenIPR-protectionstrengthens. Second,largehostcountries,intermsofGDPpercapitaandGDP,determultinational rmsfrominnovativeRandDoshoring.Thelargemarketsizeislledwithcompetingrms thatcanattractresearcherswhopreviouslywerecontractedforworkonamultinationalrm's patent.Thus,themarketsizeaectstheprobabilityofimitationfacedbymultinationalrms locatinginnovativeRandDtasksindevelopingcountries.Finally,importtarisindeveloping countrieshaveaverysmallbutpositiveeectonattractingmultinationalrms'innovativeR andDactivities.ImporttarisallowformoreSouthernrmstoenterthemarketwhileforcing outsomeNorthernrms.Anincreaseinimporttarisdoesnotthenchangetheexpected protforthesaleofthenalgood;therefore,thermdoesnotchangeresearchlevelsas importtarisrise. FewpreviousstudieshaveexaminedthesedeterminantsofinnovativeRandDoshoring tothedevelopedworld.HedgeandHicksexamineinnovativeRandDchoicesasthey relatetoforeignsubsidiarycharacteristics.Berryestablishesalinkbetweentheuseof foreigninventorsandparentrmcharacteristics,suchasRandDintensity.Demirbagand GlaisterndtheestablishmentofRandDinforeigncountriesdependsgreatlyonthe availabilityofqualiedresearchers.Thisstudyshowsthathostcountrysizealsodetermines innovativeRandDoshoringwhileimporttarishaveasmalleect.Themainresultsof 68

PAGE 69

thisdissertationalsocontributetoanumberofdierentpreviousndingsonIPR-protection's roleinmultinationalactivities.Severalauthorshavenotedthathigh-techmultinationalrms responddierentlytoIPR-protectionchangesinahostcountrythanlow-techmultinational rmsCanalsandSener,2014;Branstetteretal,2006;Branstetteretal,2011;Bilir,2014. Thisstudyexpandsonthatliteraturebynarrowingthescopeofmultinationalactivitytoonly innovativeRandDoshoring. Thechapterisorganizedasfollows.Section3.2summarizesrelatedliterature.Section 3.3presentstheanalyticalframeworkandestimationstrategyofthemodel.Section3.4 summarizesthedataused.Section3.5presentstheresults.Section3.6concludesthechapter. 3.2RelatedLiterature Thisstudycontributestoliteratureexploringtheconnectionbetweenmultinational activities,high-techindustries,andIPR-protection.JavorcikndsweakIPR-protection detersFDIinhigh-techindustries.BilirusestheNBERUSpatentcitationdatale toderiveasector-levelindexofproductlife-cyclelengths.Theauthorusesmultinational activity,suchasforeignaliatessales,asadependentvariableandtheinteractionofthe productlifecyclewiththeIPRvariableastheindependentvariableofinterest.Bilirshows thatcountrieswithstrongIPR-protectionseehighRandDintensityindustriesincreasing activitiesinhostcountriesmorethantheirlowRandDcounterparts.Bilirdoesnotexamine themultinationalactivityofRandDoshoring.CanalsandSeneridentify16countries whichunderwentanIPR-reformandexaminetheeectofthereformonoshoringtothose countries.Theauthorsndthathigh-techindustriesintheUSincreasetheirintra-industryand broadoshoringfollowingIPR-reform. Branstetteretal.alsousethetimeofamajorIPR-reforminacountrytoshow thatparentrmswithagreaterthanaveragenumberofpatentapplicationsresponddierently toIPR-protectionchanges.Theauthorsndthatthesehighpatentrmsincreaseintrarm royaltypaymentsfortechnologytransferredtoforeignaliatesmorethanlowpatentrms whenfacedwithastrengtheningofIPR-protection.TheauthorsalsofoundaliateRand 69

PAGE 70

DspendingincreaseswithIPR-protectionathigherratesforhighpatentuseparentrms. Branstetteretal.expandsonthismodelbyalsoexaminingtheeectIPRreformhas onaliateemployeecompensation,aliateassets,andaliateRandDspendingforparent rmswithhightechnologytransferrates.WhiletheauthorsexplorealiateRandD,Ifocus oninnovativeRandDtasks.SinceinnovativeRandDistheearlieststepofthecreationand saleofanewproduct,thereisahighpenaltyofimitation.Acloserexaminationofinnovative RandDtasksisneededinthisstrandofliterature. Thendingsofthisdissertationcontributetotheliteratureconcerningmarketsizeand multinationalactivity.TheoreticalstudieshaveshownthesizeoftheSouthispositivelyrelated totherangeoftasksoshoredtotheSouthDinopoulosandTsoulouhas,2015andthe amountofinputsoutsourcedGrossmanandHelpman,2005.Empiricalstudieshavefound apositiverelationshipbetweenthesizeofthehostmarketandRandDoshoring. 1 Todo andShimizutaniusesurveydataonJapanesermstoshowthesizeofahostcountry's marketpositivelyaectsbothadaptiveandinnovativeRandDoshoring.HedgeandHicks ndthatalargehostcountrysizepredictsUSmultinationalrmengagementofR andDinthehostcountry.Berryndsthathostmarketsizeincreasesthelikelihoodof usingaforeigninventorforapatent.Ifocusonlyondevelopingcountriesandthenumberof inventorsusedinthecreationofapatent.Previousstudiesonoshoringcannotpredicthow Southernresearcheremploymentrespondstomarketsize. Finally,themodelpresentedinthisdissertationconnectsinnovativeRandDoshoring andSouthernimporttaris.Whilenoempiricalstudieshavepreviouslyconnectedindustry specicimporttarisandinnovativeRandDoshoringtotheSouth,anumberofauthors 1 Otherauthorshaveshownthatthesizeofahostcountrypositivelyinuencesothertypes ofmultinationalactivity.Forexample,Ivus,Park,andSaggishowthathostcountry sizeispositivelyrelatedtotechnologylicensingtodevelopingcountries.Bilirndsthat aliatesales,assets,andemploymentincreasewithhostcountrysize.Javorciknds thatalargemarketsizeincreasesthelikelihoodofforeigndirectinvestment.Branstetteretal showthatmarketsizeispositivelyassociatedwithmultinationalactivity. 70

PAGE 71

haveconnectedahostcountry'sopennesstotradewithincreasesinmultinationalactivityinto ahostcountryCanalsandSener,2014;Branstetteretal,2011. 3.3TheModel 3.3.1AnalyticalFramework Thismodelexploresanumberofhypothesesdevelopedin"InnovativeResearchand DevelopmentOshoringNorth-SouthTrade,"whichdevelopsageneralequilibriumexplaining researchintensitydecisionsbymultinationalrms.InnovativeRandDistheprocessof conductingresearchtasksinordertocreateanewproductforsaleoneitheraglobalorlocal market.MultinationalparentcompaniesinadevelopedcountrytheNorthoshoresome innovativeRandDtaskstodevelopingcountriestheSouth. 2 Oncethemultinational rmchoosestolocatesomeresearchintheSouth,themultinationalrmmustdetermine howmanyresearcherstohireintheSouth.Theoshoringintensitydecisionisaectedbya numberoffactorssuchasintellectualpropertyrightsprotection. First,thisstudyexaminesthelinkbetweenarm'sRandDspendingonanewpatent andtheirreactiontotheIPR-protectionchangesinthehostcountry.Thereisevidence tosuggestthatrmsarelesswillingtoexposevaluableproductstomarketswithweak IPR-protection.Ivusndstheincreaseinpatentrightsindevelopingcountriesincreases thevalueofpatentsensitiveexportstothosecountries.However,rmswhohavelargeR andDneedstendtospreadouttheirresearchbetweenmultiplenationsItoandWakasugi, 2007;DemirbagandGlaister,2010.Likewise,high-techindustries,likecomputersand electronics,engageinmoreRandDoshoringthanlow-techindustriesHedgeandHicks, 2007.D'Agostino,Laursen,andSantangeloshowrmsoshoremedium/low technologyRandDtaskstoemergingeconomieswhichcomplimentshomehigh-techR 2 Evenhighlycomplexinnovationscanbedecomposedintomodularsinordertosplitthe RandDamongdierentsitesSimon,1962.Theresultofthisdecompositionalsoprevents individualresearchteamsfromfullyunderstandingtheproductandthuslowerstheprobability ofimitationZhao,2006. 71

PAGE 72

andDtasks.Northernrmswithhighresearchcostscanoftentakeadvantageofthediering costofresearchersintheSouthtolowerthecostofentry.Northernparentrmsrespondtoa strengtheningofIPR-protectioninahostcountrybyshiftingmoreresearchtaskstothenow lessriskySoutherncountry.Forrmswithlargerresearchneeds,theexogenousdecreasein theprobabilityofimitationfromastrengtheningofIPR-protectionallowsfortheopportunity tosaveonresearchcosts.Thesermshavelargeresearchcostsbasedonthenatureoftheir product. High-techproductsaremorecomplexthanlow-techproductsandthusrequiremore initialresearchtaskstoinventtheproduct.Firmsinhigh-techindustriesshouldhavelarger researchcoststhanrmsinlow-techindustries.Decreasingresearchcostsismoreimportant forthesermsthanrmswithlowresearchcosts.Therefore,rmswithlargeresearchneeds areexpectedtoincreaseRandDoshoringmoresothanrmswithsmallerresearchneeds. Sinceeachparentrminhigh-techindustrieslocatesmoreRandDtasksinthehost countryasaresultofthestrongerIPR-protectionlaws,theindustrylevelofRandDoshoring tothehostcountryincreasesmoresothanlow-techindustries.Also,theamountofRand Dconductedbyarmrelatespositivelytothatrm'ssizeSymeonidis,1996.Smallerrms cannotaordtohavetheirproductimitatedanddonotspendmuchonresearch.Therefore, thesesmallrmsbenetmorefromthedecreaseintheprobabilityofimitationthanthe possibleopportunitytodecreaseresearchcosts.Theselargermsalsoaremoreabletohandle thexedcostsoftransferringmoreresearchtaskstotheSouth.Thatis,largermsaremore likelytoalreadyhavechannelsofcommunicationopenwithSouthernresearchersfromprevious researchprojects. Hypothesis1 :SouthernIPR-protectionpositivelyincreaseshigh-techorlargerms innovativeRandDoshoringatahigherratethanlow-techorsmallrms. Hypothesis2 :SouthernIPR-protectionpositivelyaectsinnovativeRandDoshoring ofhigh-techindustriesatahigherratethanlow-techindustries. 72

PAGE 73

Next,Hypothesis3and4relateinnovativeRandDoshoringandhostcountrysize. LargeSouthernhostcountriesmayalreadyhavealargeamountofSouthernrms.Theserms representriskfortheNorthernparentrmswholocateresearchtasksinthehostcountry. TheSouthernrmsinthehostcountryhaveanincentivetoimitatethemultinationalrm's product.Imitationallowsarmtosellaproductwithoutpayingthehighxedcostsofentry. Locatingresearchtasksinthehostcountrythenstrengthensaparentrm'scompetition. Withlargerhostcountries,thereismorepotentialcompetition.Successfulimitationmayforce someNortherncompaniesoutofthemarketentirely.Thus,onewouldexpectindustry-level innovativeRandDoshoringtodecreaseashostcountrymarketsizeexpands.Thereare lessNorthernrmsinthemarkettooshoreactivitiestotheSouth.Individualparentrm's oshoringdecisionsarenotaectedbythehostcountrysize.ForaNorthernparentrm,the threatofimitationcomesfromindividualemployeeswhomayleakinformationormovetoa competingrm.Firm-leveloshoringwouldbeaectediftheemployeesaremorelikelyto leavethermgivenalargernumberofoptions;however,thislinkhasyettobeestablished intheliterature.Therefore,Iexpectedrm-leveloshoringtobeunaectedbythesizeand characteristicsoftheSouthernmarket. Hypothesis3 :Southernmarketsizedoesnotaectrm-levelinnovativeRandD oshoring. Hypothesis4 :Southernmarketsizereducesindustry-levelinnovativeRandDoshoring. Finally,highimporttarisinhostcountrieshavebeenconnectedtothefacilitationof rmentryinmonopolisticallycompetitivemarketsDemidovaandRodrguez-Clare,2009; Felbermayr,Jung,andLarch,2013.Theextratradecostsincreasethemarginalcostofthe Northernhomerms,whoarethecompetitionoftheSouthernhostcountry'srms.With highermarginalcosts,fewerrmsfromtheNorthernhomecountrycansurviveintheSouthern hostmarket.Therefore,someoftheNorthernparentrmsgooutofbusinessandarenot abletooshoreanyRandDtotheSouth.Industry-levelinnovativeRandDoshoringthen decreasesasaresultoflessrmsengaginginoshoring.Fortheparentrmsthatstayin 73

PAGE 74

business,theirRandDdecisionsshouldnotbeaectedbyhightarisinthehostcountry. FirmsonlychangetheirRandDintensityiftheybelievethatthetariwillchangethevalue oftheproductwhenitbecomesavailableonthemarket.Theadditionalmarginalcostinthe hostmarketdecreasesthereturnoneachunitsold;however,rmssellmoreunitsintheir domesticmarket.FewerNorthernparentrmsexistallowingtheremainingNorthernrmsto expandsalesintheNorth.Theseeectsmoveagainsteachotherandresultinthevalueof theproductremainingrelativelyunchangeddespitethelevelofthetariplacedbythehost country.Withoutachangeintheexpectedreturnoftheproduct,tarisdonotaectthe researchlevelssetbyeachrm. Hypothesis5 :Firm-levelinnovativeRandDoshoringisunaectedbytaris. 3.3.2EstimationMethodology Inordertotestthehypothesesputforthinthisdissertation,Iuseastatisticalmodel fortherm-levelnumberofresearchersintheSouthemployedbytheNorthernparentrm speciedasfollows: Y i jkt = 0 + 1 Tari jkt + 2 HostSize jt + 3 IPR jkt + 4 High-Tech k + 5 High-Tech k IPR jkt + 6 X jt + 7 X kt + i jkt {1 whereY i jkt isthenumberofSoutherninventorsworkingonrmi'spatentinhostcountry jinindustrykattimet.Taris jkt istheaverageadvaloremtariinindustrykforimportsinto countryjattimet.Likewise,IPR jkt isthelevelofIPR-protectionincountryjforindustryk attimet.HostSize jt isthesizeofthemarketincountryjattimet.Finally,High-Tech k isa dummyvariableindicatingwhetherornotindustrykishigh-tech. i jkt istheerrortermforeach individualproject.X jt isavectorofcountryspeciccontrolswhileX kt isavectorofindustry speciccontrols.Bothsetsofcontrolsvarywithtime.Eachobservationmeasuresaspecic patent.Thus,therecanbemultipleobservationsforarminasingleyearandcountry.The estimationmethodforthenumberofforeigninventorsisOLS. 74

PAGE 75

IalsoestimateEquation-1whereY i jkt isRandDintensity,ortheratioofSouthern researcherstototalresearchersworkingonrmi'spatent.IuseaTobitmodelwhenRand Dintensityisthedependentvariable.GiventhattheRandDintensityisbynaturebetween 0and1,aTobitmodelisappropriate.TheTobitmodelcanbeusefulasmanyrmsmaybe willingtolocatemorethan100 % oftheirresearchtasksintheirhostcountry.Thus,thedata iscensoredat1. 3 Also,IexaminetheeectthatrmsizehasonRandDoshoring: Y i jkt = 0 + 1 Taris jkt + 2 HostSize jt + 3 IPR jkt + 4 LargeFirm ikt + 5 LargeFirm ikt IPR jkt + 6 X jt + 7 X kt + i jkt {2 whereLargeFirm jkt isthesizeofrmiinindustrykattimet.LargeFirmisadummy variableindicatingwhetherornotarmhasemployedamorethanaverageamountof inventorsbeforethetimeofthepatent.Equations3-1and-2areusedtotestHypothesis 1,Hypothesis3,andHypothesis5. Finally,thisstudypresentsanOLSmodelwiththeindustrylevelofRandDspendingin ahostcountryfromUSmultinationalsasthedependentvariabletoexamineHypothesis2and Hypothesis4: Y jkt = 0 + 1 Tari jkt + 2 HostSize jt + 3 IPR jkt + 4 High-Tech kt + 5 High-Tech kt IPR jkt + 6 X jt + 7 X kt + jkt {3 whereY jkt isinnovativeRandDtasksforhostcountryjinindustrykattimet. jkt isthe errortermforeachindustrykoperatingincountryjattimet. 3 Thedataalsodoesnotshowanyheteroskedasticity;therefore,theTobitmodelwillnot presentanybiasedcoecients. 75

PAGE 76

3.4Data Threemaindatabasesareusedtoprovideobservationsforthisstudy.First,boththe USPTOandtheNBERdatabasesprovidermleveldataonpatentsledandinventorsused tocreatethosepatents.TheWDItablesprovidecountryspecicinformation.Finally,the BEA'sdatabaseonUSmultinationalenterprisesisusedforbothindustryspeciccontrolsand theindustryRandDspendinginaparticularhostcountryperyear.Industriesaredened usingtwodigitNAICScodes.Tenindustriesareusedwithindividualpatentsmatchedtotheir correspondingindustry. 3.4.1DependentVariable Thedependentvariablesareindustry-levelandrm-levelinnovativeRandDoshoring todevelopingcountries.Forrm-levelstatistics,thisstudyusesaproxythatiscommonly usedinpreviousliterature:patents. 4 Patentsindicatethatresearchershavecreatedanew andvaluableproductorcomponent.AsarguedbyHedgeandHicks,inventionsfor thepurposeofadaptingproductsforforeignmarketsareunlikelytowarranttheexpenseof protectionintheUSmarket.Thus,patentsrepresentthecompletionofinnovativeresearch andnotadaptiveresearch. Thisapproachhasthreemajorproblems.First,thepatentmayonlybeapartofa largerproduct.Therefore,rmsmightneedtopatentanumberofinnovationsjustfor oneproduct.Thispossibilityshouldnotchangermbehaviorgreatly.Thisstudysupposes thatrmsriskinformationleakswhenconductingresearchabroad.Informationleaksfor individualcomponentsofaproductstilldamagetherm.Second,patentsmaybeacquiredby 4 SomeauthorsthatusepatentsasaproxyforinnovativeRandDincludeHedgeandHicks ,Zhao,Berry,D 0 Agostino,Laursen,andSantangelo,andQian .ItoandWakasugiexamineinnovationbylookingattheoddsofmultinationals choosing3typesofforeignaliates:noRandD,RandDwithnolab,RandDwithalab. DemirbagandGlaisteruseRandDlablocationannouncementsinsteadofpatentdata. 76

PAGE 77

multinationalsrmsasawayofsuppressinganinnovationtoprotectapreviousinvention. 5 Thus,thepatentmaynotbeintendedforthesaleofanewproduct;however,rmsshould stillfearimitationofthispatentasathreattoanexistingproduct.Third,somepatentsare assignedforprocessinnovationandnotproductinnovations. Keepingtheseconcernsunderconsideration,theuseofinventorsforindividualpatents stillsucientlycaptureinnovativeRandD.ThismodelsupposesthatinnovativeRand Doshoringattherm-levelisinuencedbytheprobabilityofimitation.Withanyof theconcernsraisedabove,theprobabilityofimitationstillmotivatesRandDlocation decisionstoanextent.Thus,Ihavenoreasontobelievethatrmscollectingpatentsusing foreignresearchersforpurposesotherthaninnovativeRandDreactdierentlytochangesin IPR-protection,marketsize,andimporttarithanrmsconductinginnovativeRandD. Inordertocreatethedataset,Itakeinventorrecordsforeachpatentapplicationfrom theUSPTOdatabaseandeliminateanyapplicationthatdoesnothaveatleastoneforeign inventorfromadevelopingcountry.Theremainingpatentsarethenmatchedwiththepatent dataintheNBERdatabase.Thenumberofforeigninventorsinahostcountryisusedasthe dependentvariable.IalsousetheRandDintensityasadependentvariable.Toconstructthis variable,Idividethenumberofinventorsinahostcountrybythetotalamountofinventors listedonthepatent.Forexample,apatentassignedtotheUSbasedcompanyNalcoChemical Companyin1988hastwoinventorslistedonthepatentcoverpage.Oneinventorisbasedin theUnitedStates.TheotherinventorisbasedoutofBrazil.Inmydataset,thisobservation readsthatthereisoneSoutherninventor.Thedependentvariablefortheregressionanalysis isthenequalto1.FortheRandDintensityvariable,thevalueis1/2.Forthisobservation, countryspecicdataisforBrazil,andindustryspecicdataisforthechemicalindustryinthe US. 5 Saundersdiscussesthenonusepatents.Also,theauthorprovidesadiscussionon patentacquisitionforthepurposeoflicensing. 77

PAGE 78

EachforeigninventorlistedonaUSmultinational'spatentapplicationisevidencethatR andDhasbeenoshoredtoaforeigncountry.Therefore,alargeamountofoshoringtakes placeonaninventionifmanyforeigninventorswerelistedonthepatentapplication.Patent applicationdataisfoundintheUSPTOdatabase.First,theinventorsforeachinventionare categorizedandmatchedwiththeeventualpatentgrantnumber.Applicationsthatdonot yieldasuccessfulpatentarethrownoutofthesample.Patentsareassignedforinventionsthat haveacertainlevelofnovelty.Therefore,theinventionrepresentsaprotabletechnological advance.Patentshelpseparatetheadaptiveresearchthatmultinationalsoftenassigntotheir foreignsubsidiariesfromtheinnovativeresearch.TheNBERDatabasetakesinformationfrom theUSPTOdatabaseandmatchespatentswithboththeindustryandthetypeofassignee: i.e.UScorporation,foreigncorporation,USindividual,etc.Limitingthesampletoonly patentsbeingassignedtomultinationalUScorporations,Ilinkthetwodatabasesviapatent number.Oncelinked,thesamplecontainsdatafrom1970-2006with6,501uniquepatents thatuseatleastoneinventorfromthedevelopingworld.Ofthoseobservations,6,416patents onlyusedinventorsfromasingledevelopingcountry.87ofthesepatentshadinventorsfrom morethanonedevelopingcountry.Thesepatentsweresplitintomultipleobservationsto createuniquecountry,rm,industryobservations.Thenalsamplecontains6,590unique observations. Industry-leveloshoringiscapturedbyaBEAstatisticondollarsspentbymultinationals onRandDindevelopingcountries.Forexample,in2007,aliatesofUSmultinational chemicalrmsinBrazilspent147milliondollarsinRandDexpenditures.Thus,the dependentvariableforthisobservationis147.RandDexpenditureshavebeenusedin thepastasaproxyvariableforinnovationQian,2007;HedgeandHicks,2008.Oneaw withusingRandDexpendituresbymultinationalrmsinahostcountryincludestheinability tofullydistinguishbetweenadaptiveresearchandinnovativeresearch.Using35dierent developingcountriesaspossiblehostsforRandD,thereare4,308observationsspanning between1999and2013. 78

PAGE 79

3.4.2IndependentVariable ThisstudyexaminestheeectIPR-protectionhasonhigh-techorlargerms'innovative RandDoshoringdecisions.Additionally,thisstudyexaminestheeectofhostcountrysize andtarilevelsoninnovativeRandDoshoring. IPR-protectiondatafromeachcountryduringspecictimesisfoundusingthePark index,whichisanupdatedversionoftheGinarteandParkindex. 6 Dierentiating frompastliterature,Iproposeanewproxyforrmsize:thetotalnumberofresearchers previouslyemployedbyeachrm. 7 Thatis,largermsconductmoreresearchthansmaller rms.UsingtheUSPTOdatabase,Ideterminetheamountofinventorsemployedbytherm beforethecurrentpatent.Firmswithagreaterthanaverageamountofresearchersemployed aredesignatedaslarge.RandDexpendituresforUSparentcompaniesapproximatesindustry researchneedsintheindustryregressions. HostcountrysizeismeasuredbyGDPpercapitacollectedfortheWDI.Thelevelof importtarisfacedbyUSmultinationalrmssellingaproductintoacountrycanbeobserved. FollowingKellerandYeaple13,industryandcountryspecictarisarefoundfromthe UNCTAD'sTradeAnalysisandInformationSystemTRAINSdataset.Tarilevelsarenot givenbyindustry.Instead,thetaridatausesbroadproductcategoriestodescribetarilevels. Importtarilevelsinacountryforeachofthe10industriesaredeterminedusinganaverage oftheadvaloremtarisplacedonnalproductswithinanindustry.Table3-1presentsthe summarystatisticsfortherm-leveldataset.Table3-2presentsthesummarystatisticsforthe industry-leveldataset. 6 ThisupdatedindexhasbeenusedinanumberofpreviousstudiessuchasKellerand Yeaple,Berry,Naghavi,Spies,andToubal,andBilir. 7 Inpreviousstudies,thesalesofthermhavebeenusedasawayofestimatingrm size.ThenumberofrmemployeeshasalsobeenusedtomeasurermsizeBussand Peukert,2015.Sincethisdatawasnotavailable,Iproposeanewproxyforrmsizebased ontheestablishedconnectionbetweenrmsizeandresearchanddevelopmentspending.See Symeonidis. 79

PAGE 80

Table3-1.SummaryStatisticsForFirm-LevelRegressions VariableMeanStd.Dev.Min.Max. NumberofForeignInventors1.81.4123 USFirmRandDIntensityinHostCountry0.60.30.021 High-techIndustry0.50.4901 LargeFirm0.50.501 IPRIndex2.90.804.88 Tari13.210.50100 PreviousExperience0.50.501 GDPpercapita2681.93810.615358884 IndustryEmployment1415.92064.808808.1 IndustryAssets410729.5264711.401741805 ProductLife9.30.578.410.4 HostTrade58.144.111.5220.4 HostSciencePublications17305.722456.40181690 HostCountryPopulation4.59e+085.11e+08559301.31e+09 HostGDPgrowth4.54.8-22.988.9 Note:HostcountryspecicdataisfoundusingWDItables.Industryspecicdataisfound usingBEAdataandonlyconcernstheindustryintheU.S.Thenumberofforeigninventors, rmsize,andrmRandDintensityinahostcountryaredeterminedusingtheUSPTO database.TaridataisfoundusingtheTRAINSdatabase.TheIPRindexisfromPark008. ProductlifedataisfromBilir. 3.5EmpiricalResults 3.5.1Firm-LevelResults Tables3-3and3-4presenttheresultsoftheregressionanalysiswiththenumberof inventorsasthedependentvariable.Table3-3reportsthesetofregressionsthatusethe high-techindicator.Thereislittleevidencethatrmsinhigh-techindustriesoshoremore thanrmsinlow-techindustries.Thecoecientonthehigh-techandIPRinteractionvariable isalsonotstatisticallysignicant.Thereisthennoevidencetosupportthatrmsinhigh-tech industriesresponddierentlytoIPR-protectioninhostcountries.ThelackofevidenceinTable 80

PAGE 81

Table3-2.SummaryStatisticsForIndustry-LevelRegressions VariableMeanStd.Dev.Min.Max. RandDExpendituresinHostCountry8.746.40822 IPR3.40.64163471.724.68 GDPpercapita9535.314954.9299.493605.75 High-techIndustry0.310.4701 Tari8.78098.8 IndustryEmployment1287.92046.608808.1 IndustryAssets371051.2415370.901741805 IndustrySales404120.9333240.501493181 IndustryRandDexpenditures13415.517264.72058859 HostSciencePublications13556.2418445.5401434.5 HostGDPpercapitaGrowth2.94.1-15.130.3 HostTrade7842.821430.4 Note:HostcountryspecicdataisfoundusingWDItables.Industryspecicdataisfound usingBEAdataandonlyconcernstheindustryintheU.S.Taridataisfoundusingthe TRAINSdatabase.TheIPRindexisfromPark08. 3-3isexplainedbythevarianceinprojectsizeforrmsinhigh-techindustries.Thisstudyonly measuresproductcomplexityforanindustryandnotthespecicprojectcomplexity. 8 Table3-4includesthelargermdummyvariable;however,thecoecientonthedummy variableisonlystatisticallysignicantincolumnandcolumn.Likewise,thecoecient forthelargermandIPRinteractionvariableisonlystatisticallysignicantinthesecolumns aswell.Thecoecientforthelargermdummyvariableisnegativewhilethecoecienton theIPRandlargerminteractionvariableispositive. 9 Thesignofthecoecientonthe 8 TheregressionswithRandDintensitycontrolforthisinabilitytomeasuretheproject complexitybydividingthenumberofSouthernresearchersbythetotalnumberofresearchers workingonthepatent. 9 WhentheinteractiontermistakenintoaccountwiththemeanIPRvalue,largermstend tooshoremorethansmallrmsaslongastheIPR-environmentisstrongerthanaveragefor developingcountries. 81

PAGE 82

Table3-3.High-TechFirms,HostCountryPatentsLaws,andNumberofInventorsOshored, Firm-Level Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.High-techisadummyvariableindicatingifarmisinahigh-techindustryChemistry, Computers,Transport,andWholesale.Thenumberofforeigninventorsoneachpatentisdeterminedusing theUSPTOdatabase.HostcountryspecicdataisfoundusingWDItables.Hostcountrycontrolvariables includeGDPgrowth,numberofsciencepublications,andtheamountoftradeasdenedbyexportsand importsdividedbyGDP.Industryspecicdata,fromtheBEAdatabase,includesU.S.industry-levelparents sales,numberofemployees,assets,andRandDexpendituresforUSparentcompanies.Taridataisfound usingtheTRAINSdatabase.TheIPRindexisfromPark.Productlife,fromBilir,controlsfor productcomplexityineachindustry. interactionvariableimpliesthatlargermsaremorereactivetochangesinIPR-protectionthan smallrms.ThisprovidessupportforHypothesis1. ThroughoutallsetsofregressionsinTables3-3and3-4,thecoecientonGDPper capitaisnegativeandoftenstatisticallysignicant.Asinglestandarddeviationincreasein thehostcountrysizeyieldsadecreaseofupto0.39researchers.Thisprovidesevidence 82

PAGE 83

Table3-4.LargeFirm,HostCountryPatentsLaws,andNumberofInventorsOshored, Firm-Level Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.Largermisadummyvariableindicatingifrmshavemorethanaverageinventors employedpreviouslyasseenfromtheUSPTOdata.Thenumberofforeigninventorsoneachpatentis determinedusingtheUSPTOdatabase.HostcountryspecicdataisfoundusingWDItables.Hostcountry controlvariablesincludeGDPgrowth,numberofsciencepublications,andtheamountoftradeasdenedby exportsandimportsdividedbyGDP.Industryspecicdata,fromtheBEAdatabase,includesU.S. industry-levelparentssales,numberofemployees,assets,andRandDexpendituresforUSparentcompanies. TaridataisfoundusingtheTRAINSdatabase.TheIPRindexisfromPark.Productlife,fromBilir ,controlsforproductcomplexityineachindustry. 83

PAGE 84

againstHypothesis3. 10 Thisndingsuggeststhatthereisaconnectionbetweenemployee mobilityandthemarketsize.Alargemarketsizemeanstherearemoreopportunitiesfor employeestomigrate.Thus,theprobabilityofimitationincreaseswithmarketsize.Thereis alsosomeevidencetosuggestthatIPR-protectionpositivelyaectsrm-levelinnovativeRand Ddecisions. Thecoecientonahostcountry'starilevelispositiveandoccasionallystatistically signicant.Theeectoftarisisverysmall.Anincreaseofonestandarddeviationofthe tarilevelincreasestheamountofpeopleoshoredbyonly0.26inventorsatthemost.A possibleexplanationforthisistherelationshipbetweenimporttarisandthepreferences ofconsumersinthehostmarket.Ahighimporttarimaysignalthatacertainindustryis thrivinginthehostcountry.Therefore,theproductisvaluedinthemarketplace.Furthermore, qualiedresearchersinaspecicindustrywouldthenalreadyexistinthehostcountry. Giventhesmallmagnitudeofthecoecient,thereissomesupportforHypothesis5.Also, supportingthendingsofDemirbagandGlaister,thepreviousexperiencevariableisa positivepredictorofRandDoshoring. Tables3-5and3-6presenttheresultsofregressionswithRandDintensityasthe dependentvariable.Table3-5providesadditionalsupportforHypothesis1.Thecoecientfor thehigh-techandIPRinteractionvariableispositiveandstatisticallysignicant.High-tech rmsaremorelikelytoincreaseoshoringactivityduringastrengtheningofIPR-protection thanlow-techrms.Table3-6doesnotyieldanystatisticallysignicantresultsregarding rmsize.ThecoecientsonthetariandGDPpercapitavariablesareneverstatistically 10 Thisresultcontradictsmanypreviousndingsonhostcountrysizeandmultinational activities.Forexample,Ivus,Park,andSaggishowthathostcountrysizeispositively relatedtotechnologylicensingtodevelopingcountries.Bilirndsthataliatesales, assets,andemploymentincreasewithhostcountrysize.HedgeandHicksndthat hostcountrysizepositivelyinuencesUSmultinationalrmengagementofRandDinthe hostcountry. 84

PAGE 85

Table3-5.High-TechFirms,HostCountryPatentsLaws,andRandDIntensityinHost Country,Firm-Level,TobitRegression Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.High-techisadummyvariableindicatingifarmisinahigh-techindustryChemistry, Computers,Transport,andWholesale.RandDintensityisdenedasthenumberofforeigninventorsdivided bythenumberoftotalinventorsonapatentandisconstructedwiththeUSPTOdatabase.Hostcountry specicdataisfoundusingWDItables.HostcountrycontrolvariablesincludeGDPgrowth,numberofscience publications,andtheamountoftradeasdenedbyexportsandimportsdividedbyGDP.Industryspecic data,fromtheBEAdatabase,includesU.S.industry-levelparentssales,numberofemployees,assets,andR andDexpendituresforUSparentcompanies.TaridataisfoundusingtheTRAINSdatabase.TheIPRindex isfromPark.Productlife,fromBilir,controlsforproductcomplexityineachindustry. signicantineitherTable3-5orTable3-6.Thus,thissetofregressionsdoesnotyieldsupport fororagainstHypotheses3and5. 3.5.2Industry-LevelResults Tables3-7and3-8usetheindustry-levelofmultinationalRandDexpendituresinahost countryasthedependentvariable.Table3-7againusesabinaryvariabletoindicatewhether 85

PAGE 86

Table3-6.LargeFirms,HostCountryPatentsLaws,andRandDIntensityinHostCountry, Firm-Level,TobitRegression Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.Largermisadummyvariableindicatingifrmshavemorethanaverageinventors employedpreviouslyasseenfromtheUSPTOdata.RandDintensityisdenedasthenumberofforeign inventorsdividedbythenumberoftotalinventorsonapatentandisconstructedwiththeUSPTOdatabase. HostcountryspecicdataisfoundusingWDItables.HostcountrycontrolvariablesincludeGDPgrowth, numberofsciencepublications,andtheamountoftradeasdenedbyexportsandimportsdividedbyGDP. Industryspecicdata,fromtheBEAdatabase,includesU.S.industry-levelparentssales,numberofemployees, assets,andRandDexpendituresforUSparentcompanies.TaridataisfoundusingtheTRAINSdatabase. TheIPRindexisfromPark.Productlife,fromBilir,controlsforproductcomplexityineach industry. 86

PAGE 87

anindustryishigh-tech.High-techindustriesaredeterminedbyhavingahigherthanaverage parentRandDexpenditure.Thecoecientontheinteractiontermbetweenthehigh-tech dummyandtheIPR-protectionvariableispositive,statisticallysignicant,andlarge.These high-techindustriesrespondtoastrengtheningofIPR-protectionwithmoreinvestmentthan low-techindustries.Table3-8testsHypothesis2usinganexaminationoftheinteraction variableconstructedfromtheIPRvariableandtheRandDexpendituresofUSparentrms. ThecoecientforparentRandDexpendituresisnegativeandstatisticallysignicant.The coecientontheinteractiontermisstatisticallysignicantandpositiveinallregressions. Industrieswithlargeamountsofresearcharereluctanttolocateinahostcountrywith extremelyweakIPR-protection;however,theseindustriesbeginmovingresearchtotheSouth asIPR-protectionstrengthensatagreaterratethanindustrieswithsmallresearchneeds. Alongwiththeresultsoftheprevioussubsection,thesetwotablesprovidestrongevidencefor Hypothesis2. Tables3-7and3-8alsotestHypothesis4.GDPpercapita,whichcapturesthehost countrysize,isnotstatisticallysignicant.Thereisnoevidencetoeitherconrmorrefute Hypothesis4.Also,thecoecientonthetarilevelisoftennotstatisticallysignicant.When thecoecientisstatisticallysignicant,thevalueispositiveandeconomicallysignicant. Importtarismayindicateamarketthathighlyvaluesthatindustry.Thus,parentrms expandintothemarketwithadaptiveresearchasawayofbothenteringthemarketand adaptinganexistingproducttottheuniquepreferencesandtastesofconsumersinthat market.ByusingRandDexpendituresasadependentvariable,adaptiveandinnovative researcharetiedtogether.Highimporttarismayalsoindicatethatqualiedresearchers alreadyexistinthehostcountryforthatindustry.Also,highimporttarismaydecrease SouthernconsumerincomeandunintentionallyforceSouthernrmstoclose.MoreNorthern rmscouldthenenterthemarket. 87

PAGE 88

Table3-7.High-TechIndustry,HostCountryPatentsLaws,andRandDIntensityinHost Country,Industry-Level Note:Standarderrorsinparentheses.Starsindicatesignicancewith***p < 0.01,**p < 0.05,*p < 0.1.Rand DexpendituresfromU.S.multinationalsinahostcountrydataisfromtheBEA.High-techisadummy variableindicatingifarmisinahigh-techindustryChemistry,Computers,Transport,andWholesale.Host countryspecicdataisfoundusingWDItables.HostcountrycontrolvariablesincludeGDPgrowth,numberof sciencepublications,andtheamountoftradeasdenedbyexportsandimportsdividedbyGDP.Industry specicdata,fromtheBEAdatabase,includesU.S.industry-levelparentssales,numberofemployees,assets, andRandDexpendituresforUSparentcompanies.TaridataisfoundusingtheTRAINSdatabase.TheIPR indexisfromPark.Productlife,fromBilir,controlsforproductcomplexityineachindustry. 88

PAGE 89

Table3-8.ParentRandDExpenditures,HostCountryPatentsLaws,andRandD ExpendituresinHostCountry,Industry-Level Note:Standarderrorsinparentheses.Starsindicatesignicancewith***p < 0.01,**p < 0.05,*p < 0.1.Rand DexpendituresfromU.S.multinationalsinahostcountrydataisfromtheBEA.U.S.RandDexpendituresin theU.S.areusedasaproxyfortheresearchneedsoftheindustry.ThisdataalsocomesfromtheBEA.Host countryspecicdataisfoundusingWDItables.HostcountrycontrolvariablesincludeGDPgrowth,numberof sciencepublications,andtheamountoftradeasdenedbyexportsandimportsdividedbyGDP.Industry specicdata,fromtheBEAdatabase,includesU.S.industry-levelparentssales,numberofemployees,and assetsforUSparentcompanies.TaridataisfoundusingtheTRAINSdatabase.TheIPRindexisfromPark .Productlife,fromBilir,controlsforproductcomplexityineachindustry. 89

PAGE 90

3.5.3SensitivityTests Inordertodeterminetherobustnessoftheresultspresentedabove,anumberof sensitivitytestsareperformed.First,twoindustries,ChemicalsandWholesale,areremoved fromthesample.Thisremovaloftheseindustriesensuresthattheresultsarenotdrivenby specicindustriesinwhichpatentsarefoundtobeespeciallyeective.Inanadditionaltest,I removethe5countriesthatarethelargestrecipientsofRandDtasksfromUSmultinational rms.Again,thistestistodeterminewhethertheresultscanbeappliedingeneralorare drivenbytheseunobservedcharacteristicsofthesevecountries.Inbothofthesetests,the resultsofthisdissertationremainlargelythesame.Thisstudyusesindustryxedeects andcountryxedeects.Controllingforthepossibilityofsignicantchangesinthesizeor technologyofindustriesthroughouttheyear,Iaddsector-yearxedeects.Country-yearxed eectsareaddedtotheregressionstocontrolforsimilarreasons.Theinclusionofthesexed eectsdoesnotsignicantlychangeanyoftheresults. TestingthestabilityoftheresultstotheIPRindex,theIPRvariableisremovedinfavor ofdierence-in-dierencemethodusedbyBranstetteretal.,2011,CanalsandSener ,andPark,Ivus,andSaggi.UsingtheyearofmajorIPR-reformasthetimeof treatmentandrmsfromhigh-techindustriesasthetreatmentgroup,Iestimatetheeect IPR-reformhasoninnovativeRandDoshoring.Thespecictimeofreformfromeach countrycomesfromthepreviousliteraturementionedaboveinsection3.2.Table3-9uses thisdi-in-dimethodwiththetreatmentgroupbeinghigh-techrmsandthedependent variablebeingthenumberofforeigninventors.Theresultsaresimilartotheonespresentedin Table3-3.Table3-10presentsresultssimilartoTable3-4.Usinglargermsasthetreatment group,thecoecientfortheinteractiontermbetweenthetreatmentgroupandtheIPR-reform dummyvariableisstatisticallysignicantandpositive.Largermstendtooshorelessthan smallrmsbeforetheIPR-reform;however,largermsoshoremoreafterthereform.Table 3-11displaysresultsofregressionsusingtheIPRreformdummyvariable,high-techdummy variable,andindustry-leveldata.Theinteractiontermbetweenthetwodummyvariablesis 90

PAGE 91

Table3-9.Di-in-Di:High-TechIndustry,PatentsReformYear,andRandDIntensityin HostCountry,Firm-Level Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.High-techisadummyvariableindicatingifarmisinahigh-techindustryChemistry, Computers,Transport,andWholesale.Thenumberofforeigninventorsoneachpatentisdeterminedusing theUSPTOdatabase.HostcountryspecicdataisfoundusingWDItables.Hostcountrycontrolvariables includeGDPgrowth,numberofsciencepublications,andtheamountoftradeasdenedbyexportsand importsdividedbyGDP.Industryspecicdata,fromtheBEAdatabase,includesU.S.industry-levelparents sales,numberofemployees,assets,andRandDexpendituresforUSparentcompanies.Taridataisfound usingtheTRAINSdatabase.TheIPRreformdummyindicatesthetimeofamajorstrengtheningof IPR-protectioninahostcountry.Productlife,fromBilir,controlsforproductcomplexityineach industry. statisticallysignicantandpositive.Thehigh-techdummyvariableisnolongerstatistically signicant.Theresultsofthedierence-in-dierenceregressionsisnotdrasticallydierent fromtheregressionsusingtheIPRindex;therefore,theresultsofthisdissertationarenot sensitivetotheIPRindexusedinthisstudy. 91

PAGE 92

Table3-10.Di-in-Di:LargeFirm,PatentsReformYear,andNumberofInventorsOshored, Firm-Level Note:Standarderrorsinparenthesesandareclusteredbyrm.Starsindicatesignicancewith***p < 0.01,** p < 0.05,*p < 0.1.Largermisadummyvariableindicatingifrmshavemorethanaverageinventors employedpreviouslyasseenfromtheUSPTOdata.Thenumberofforeigninventorsoneachpatentis determinedusingtheUSPTOdatabase.HostcountryspecicdataisfoundusingWDItables.Hostcountry controlvariablesincludeGDPgrowth,numberofsciencepublications,andtheamountoftradeasdenedby exportsandimportsdividedbyGDP.Industryspecicdata,fromtheBEAdatabase,includesU.S. industry-levelparentssales,numberofemployees,assets,andRandDexpendituresforUSparentcompanies. TaridataisfoundusingtheTRAINSdatabase.TheIPRreformdummyindicatesthetimeofamajor strengtheningofIPR-protectioninahostcountry.Productlife,fromBilir,controlsforproduct complexityineachindustry. 92

PAGE 93

Table3-11.Di-in-Di:High-TechIndustry,PatentsReformYear,andRandDExpenditures inHostCountry Note:Standarderrorsinparentheses.Starsindicatesignicancewith***p < 0.01,**p < 0.05,*p < 0.1.Rand DexpendituresfromU.S.multinationalsinahostcountrydataisfromtheBEA.Thenumberofforeign inventorsoneachpatentisdeterminedusingtheUSPTOdatabase.Hostcountryspecicdataisfoundusing WDItables.HostcountrycontrolvariablesincludeGDPgrowth,numberofsciencepublications,andthe amountoftradeasdenedbyexportsandimportsdividedbyGDP.Industryspecicdata,fromtheBEA database,includesU.S.industry-levelparentssales,numberofemployees,assets,andRandDexpendituresfor USparentcompanies.TaridataisfoundusingtheTRAINSdatabase.TheIPRreformdummyindicatesthe timeofamajorstrengtheningofIPR-protectioninahostcountry.Productlife,fromBilir,controlsfor productcomplexityineachindustry. 93

PAGE 94

ThereisacommonconcernamongpreviousliteraturethattheIPRvariableisendogenous Branstetteretal.,2006;CanalsandSener,2014;Bilir,2014;Park,Ivus,andSaggi, 2015.Sofarinthisstudy,IPR-protectionistreatedasexogenous;however,theremaybe omittedvariablescorrelatedwiththeIPRenvironmentandtheinowofRandDoshoring. Furthermore,countriesmayfeelobligatedtostrengthenIPR-protectionifalargenumberof USmultinationalrmsalreadylocateRandDinthecountry.Inanattempttoruleoutthese concern,thisdissertationtestswhethertherehasbeenaclearupwardtrendintheoshoring ofinnovativeRandD.Inordertotestwhethertherewasanupwardtrend,Icreateaseriesof dummyvariablesforthepre-andpost-reformyears.Pre5equals1for5yearsormorebefore thereformtakesplace.Post5equals1for5yearsormoreafterthereform.Pre4equals1for 4yearsbeforethereformwhilePost4equals1for4yearsafterthereform.Pre3,Pre2,Post3, Post2,andPost1aredenedinsimilarways.Pre1isleftoutasabaseline.ROisadummy variableindicatingtheyearofreform.Table3-11showstheresultsoftheseregressions. Columnsandusetherm-leveldependentvariableofthenumberofforeigninventors. Columnsandusetheindustry-leveldependentvariableofRandDexpendituresin ahostcountry.ThereisnostatisticallysignicanttrendinRandDoshoringbeforeeach reform;therefore,IconcludethattheIPRreformisnotendogenous. 3.6DiscussionandConclusion Theregressionanalysisyieldsanimportantresultabouttheconnectionofimporttaris andRandDtasks.ThepreviousliteratureconcerningRandDoshoringdoesnotinclude theeectofindustryspecicimporttarisoninnovativeRandDintensitywithinaSouthern hostcountry.Anumberofauthorshaveconnectedahostcountry'sopennesstotradewith increasesinmultinationalactivityintoahostcountryCanalsandSener,2014;Branstetter etal,2011.AmainresultofthestudypresentedhereisthatlargeSouthernimporttaris predicttheuseofinnovativeRandDtasksinthathostcountrybymultinationalrms.A possibleexplanationforthisperceivedrelationshipistheassociationbetweenimporttaris andthepreferencesofconsumersinthehostmarket.Ahighimporttariindicatesthata 94

PAGE 95

Table3-12.ParentRandDOshoringwithPreandPostReformDummies,Industry-Level 95

PAGE 96

certainindustryisthrivinginthehostcountry.Therefore,theproductisvaluedinthemarket place.Furthermore,qualiedresearchersinaspecicindustrywouldthenalreadyexistinthe hostcountry.So,acountrymaybestiinginnovationwithinitsbordersbysettinglowimport taris.Freetradeagreements,liketheTPP,mayhaveunintendedconsequencesonRandD employment. Theresultsofthisdissertationconrmandaddtopreviousndingsintheempirical literature.Thisdissertationndsthathigh-techrmsandindustriesresponddierentlyto IPR-protectionchangesthanlow-techrmsandindustries.Specically,high-techrmsincrease RandDintensitygreatly,relativetotheirlow-techcounterparts,asSouthernIPR-protection strengthens.Likewise,high-techindustriesasawholeincreaseRandDoshoringmore thanlow-techindustriesasIPR-protectionstrengthensintheSouth.Thisresultrelatesto thendingsofanumberofstudiesthatconnectmultinationalactivity,high-techrms,and IPR-protectionCanalsandSener,2014;Branstetteretal,2006;Branstetteretal,2011; Bilir,2014.ThestudyinthisdissertationfurtherstheevidencethatUShigh-techindustries increasetheiroperationsabroadasSoutherncountriesstrengthentheirIPR-protectionlaws. Dierentiatingitselffrompreviousstudies,thisdissertationonlyexaminesaspecictypeof oshoring:innovativeRandD.Informationleaksthatresultfromimitationaremoredamaging attheveryearlystagesofproductdevelopment.Amainresultofthisdissertationisthat high-techrmsarewillingtolocatemoreRandDtasksintheSouthasIPR-protection strengthensdespitethehighpenaltyofimitation. Finally,theresultsofthisdissertationindicatethatrmsarenegativelyinuencedby hostcountrymarketsizewhenmakinginnovativeRandDoshoringdecisions;however,the eectissmall.Employeemobilityisadrivingforcebehindimitation.Alargermarketinthe hostcountryindicatesmoreSouthernrmsforemployeestoseekemployment.Thisresult qualiesmanypreviousndingsonhostcountrysizeandmultinationalactivities.Forexample, Ivus,Park,andSaggishowthathostcountrysizeispositivelyrelatedtotechnology licensingtodevelopingcountries.Bilirndsthataliatesales,assets,andemployment 96

PAGE 97

increasewithhostcountrysize.HedgeandHicksndthatalargehostcountrysize predictsUSmultinationalrmengagementofRandDinthehostcountry.Thecontradicting resultssuggeststhatinnovativeRandDintensitydecisionsaremoresensitivetothethreatof imitationthanotherRandDactivities.OtherRandDactivities,suchasadaptiveRandD oshoring,arefocusedonestablishingrelationshipsinuntappedmarkets. ThisstudyraisesfurtherquestionsonthesubjectofinnovativeRandDoshoring.While thisdissertationexaminestheresponseofhigh-techrmsandlargermstoIPR-protection changes,itdoesnotseparateintoindustryspecicIPR-protectionchanges.Onepossible avenueforfurtherresearchistonarrowthescopetothepharmaceuticalindustryafterthe TRIPSagreement.TRIPSpromotedIPRreforminanumberofdevelopingnationsinthe pharmaceuticalindustry.Firmleveldatacanalsobeaddedtotheanalysisinthisdissertation toprovidearicherpictureofmultinationalinnovativeRandDlocationdecisions.TheBEA keepsdetailedrmleveldatathancanbeconnectedtotheNBERpatentdatabase. 97

PAGE 98

APPENDIXA APPENDIXFORCHAPTER1 A.1EquilibriumProt Withthecalculated R N R S ,andprobabilityofimitation,thethreepossibleexpected protlevelsare: E N = )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N R A{1 E = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] )]TJ/F21 11.9552 Tf 11.955 0 Td [( 2 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] 1 2 + 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S + )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R A{2 E S = )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R A{3 EquationA-1representstheprotlevelofaCase1rm.EquationA-2istheprot ofarminCase2,andarminCase3haveaprotasshowninequationA-3.While stayingoutofthemarketcompletelyisaviableoption,itremainsrelativelyuninteresting. Thenorthernrmwillonlychoosetonotenterthemarketwhentheexpectedprotisalways negative.Thispossibilitydependsontheparametersofthemodelandmarketstructure.I assumeawaythiscaseandonlyfocusonthetwocornercasesandtheoneinteriorcase. A.2Including asanIPR-ProtectionParameter Let 0 << 1 .Thisyieldsanewexpectedprotfunction: E = R N + R S )]TJ/F24 11.9552 Tf 11.955 0 Td [( R N + R S )]TJ/F24 11.9552 Tf 25.604 8.088 Td [( R S R N + R S )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N R N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R S 98

PAGE 99

= R N + R S )]TJ/F24 11.9552 Tf 11.956 0 Td [( R N + R S [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ]+ )]TJ/F38 11.9552 Tf 11.955 0 Td [(w N R N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R )]TJ/F38 11.9552 Tf 11.955 0 Td [(R N A{4 MaximizingEquationbychoosing R N yields: R N = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.956 0 Td [(w S 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R A{5 While doesdecreasethevalue R N atanygiven R ,theinclusionoftheIPR-protection parameterdoesnotaectthedirectionalresponseof R N toanyothergivenparameter; however,with 0 << 1 ,themagnitudeofanycomparativestaticswillchange.Since enters EquationA-5inonlyoneplace,theparametercanbeabsorbedintothepenaltyofimitation variable.So,theinclusionofthevariablewillbeakintoloweringthepenaltyofimitation. Thus,theintervalofpossibleinteriorsolutionsshifts: [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < R < [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S A{6 FromConditionA-6,includingtheextraIPR-parameter, ,shiftstheintervaltothe left.SmallerrmsthatlocatedresearchonlyintheNorthwillthereforebeabletolocatesome researchintheSouthiftheIPR-protectionstrengthens.Also,notethatas increases, R N decreasesinEquationA-5.While doesnotgreatlyimpactthecomparativestaticsofthis model,theprotfunctionwhenthenorthernrmlocatesresearchsolelyintheSouth: E S = )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ]+ )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S R A{7 ComparedwithEquationA-3,theprotfunctioninEquationA-7hasanextraterm reectingthenewpossibilityofimitationnotoccurring.Theprotfunctionfortheopposite case,whenthenorthernrmlocatesintheNorth,remainsunchanged.So, buildsinthe realismthatimitationisnotguaranteed,themodeldoeslittletochangethemainresultsof thispaper.ThisextraIPRparameterwillchangesomeoftheresultswhendierentmarket structuresareappliedtothemodel.Forexample,eventheharshestcompetitiveenvironments couldnowseeanorthernrmwithahighresearchneedlocatingresearchfullyintheSouth. 99

PAGE 100

A.3ProofofInteriorSolutionParameterConditions Takingthederivativeofprotfunction-7withrespectto R N yields: @ @ R N = R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] R N )]TJ/F24 11.9552 Tf 11.956 0 Td [( + R 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S A{8 SettingA-8lessthanzero: @ @ R N < 0 R [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < R N )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R 2 R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.956 0 Td [(w S 1 2 < R N )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 )]TJ/F24 11.9552 Tf 23.595 8.088 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R < R N R N < R N R N representsthequantityofnorthernresearchersfoundfromsettingequationA-8 equaltozero.Again, R N 2f 0, R g .Ifthe R N valuethatsetsEquation-8equaltozerois itselflessthanzero,thentheslopeoftheprotfunctionovertheintervalofinterestwillbe strictlynegative.Ifthisisthecase,then 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 )]TJ/F24 11.9552 Tf 23.595 8.087 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R < 0 R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 < R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 < R 1 2 [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < R Therefore,whentotalresearchisgreaterthan [ )]TJ/F25 7.9701 Tf 6.586 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.587 0 Td [(w S ,thermwillndthehighest protbyemployingzeronorthernresearchersandlocatingtheresearchtasksentirelyinthe 100

PAGE 101

South.Likewise,if R N > R ,thentheprotfunctionwillbestrictlyincreasingintheintervalof interest.Inthiscase,thermwilllocateentirelyinthenorthandset R N = R : 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 )]TJ/F24 11.9552 Tf 23.595 8.087 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R > R R [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 > R [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 > R 1 2 [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S > R So,forrmswithsmallresearchneeds,oshoringtotheSouthwillnotbenecessary. Therefore,theconditionneededforaninteriorsolutionis: [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < R < [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S A{9 A.4ProofOfResult2 Again,thenorthernrminCase2istheonlytypeofrmtobeaectedbyIPRchanges. So,thederivativeof R N wrt inCase2is: @ R N @ = [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.956 0 Td [(w S 1 2 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( + 1 2 1 2 )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1+ 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( @ R N @ = [ )]TJ/F24 11.9552 Tf 11.956 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.956 0 Td [(w S 1 2 1+ )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( A{10 101

PAGE 102

ForthenorthernrmtoincreasenorthernresearchinthefaceofweakerIPR-protection, or @ R N @ > 0 ,thefollowingmustbetrue: @ R N @ > 0 [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 1+ )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 > 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] R w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1 2 > 2 1+ [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1+ 2 2 > R So,forarmwithavalueoftotalresearchthatissmallerthan [ )]TJ/F25 7.9701 Tf 6.587 0 Td [( ] w N )]TJ/F39 7.9701 Tf 6.586 0 Td [(w S )]TJ/F22 7.9701 Tf 6.675 -4.428 Td [(1+ 2 2 weakeningofIPR-protectionresultsinmorenorthernresearch.Since islessthanone,itis notdiculttoseethat: [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S < [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S 1+ 2 2 < [ )]TJ/F24 11.9552 Tf 11.955 0 Td [( ] w N )]TJ/F38 11.9552 Tf 11.955 0 Td [(w S Therefore,whenthenorthernrmlocatesinbothcountries,itsreactiontoIPRreform willbeambiguous.SmallerrmsinCase2willincreasetheamountofnorthernresearchwhen IPR-protectionweakens,whilethelargerrmswillreactbyloweringtheamountofnorthern researchers. 102

PAGE 103

APPENDIXB APPENDIXFORCHAPTER2 B.1Consumers Consumersmaximizeutilityfunction-1intwostages.First,Isolvethewithinsector consumeroptimization.Thatis,consumersdecidewhatquantitiesof x ii and x ji toconsumein ordertomaximizetheCESsub-utilityfunction, X i .Theconsumermustdeterminehowmuch ofeachvarietytoconsumer: max x ii x ji X i = h M i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 + M j a ji x ji )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 i )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 s t M i p ii x ii + M j p ji x ji = E X i B{1 where i 6 = j 2 n s and E X i istheexpenditurespentondierentiatedgoods.The Lagrangiancanbedenedas: L = h M i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 + M j a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 i )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ M i p ii x ii + M j p ji x ji )]TJ/F38 11.9552 Tf 11.956 0 Td [(E X i ] B{2 Takingrstorderconditions: @ L @ x ii = X 1 i a ii M i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( p ii M i =0 B{3 @ L @ x ji = X 1 i a ji M j a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( p ji M j =0 B{4 @ L @ = )]TJ/F21 11.9552 Tf 9.299 0 Td [([ M i p ii x ii + M j p ji x ji )]TJ/F38 11.9552 Tf 11.955 0 Td [(E X i ]=0 B{5 Notethat: h M i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 + M j a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 i 1 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 = h X )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 i i 1 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 = X 1 i CombiningB-3andB-4yields: 103

PAGE 104

a ji X 1 i a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 p ji = a ii X 1 i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 p ii B{6 a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 = p ji p ii a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ji a ii B{7 p ji x ji = p ii x ii a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ji a 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( ii p 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( ji B{8 Likewise, p ii x ii = p ji x ji a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ii a 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( ji p 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( ii B{9 UsingequationB-5withB-8,B-9andthedenitionofthepriceindexfrom equations-4and-5,theindividualdemandsfor x ii and x ji canbefound: x ji = a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ji p )]TJ/F25 7.9701 Tf 6.587 0 Td [( ji P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 i E X i B{10 x ii = a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ii p )]TJ/F25 7.9701 Tf 6.587 0 Td [( ii P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 i E X i B{11 PluggingequationsB-10andB-11backintothethesub-utilityfunctionyields: X i = h M i a ii x ii )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 + M j a ji x ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 i )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 = M i a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 ii a )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 ii p )]TJ/F25 7.9701 Tf 6.587 0 Td [( +1 ii P )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 i E )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 X i + M j a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 ji a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 ji p )]TJ/F25 7.9701 Tf 6.586 0 Td [( +1 ji P )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 i E )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 X i )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 = E X i P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 i M i a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 ii a )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 ii p )]TJ/F25 7.9701 Tf 6.587 0 Td [( +1 ii + M j a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 ji a )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 ji p )]TJ/F25 7.9701 Tf 6.586 0 Td [( +1 ji )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 Usingthedenitionofthepriceindex,thesub-utilityfunctionsimpliesto: 104

PAGE 105

X i = E X i P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 i P )]TJ/F25 7.9701 Tf 6.586 0 Td [( i = E X i P i B{12 So,theexpendituresondierentiatedgoodsincountryiequalsthepriceindexmultiplied bythequalityindex: X i P i = E X i .Thedemandsforeachvarietycanbefurthersimplied: x ji = a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ji p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ji P i X i B{13 x ii = a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ii p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ii P i X i B{14 Thiscompletesthewithinsectorconsumeroptimization.Notethattheexpenditureonthe outside-goodgoodisjust z i .Thepriceoftheoutside-goodgoodissetequalto1.Consumer thenfacetheacrosssectormaximizationproblem: max z i X i z 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( i X i s t z i + P i X i = w i L i B{15 where w i isthewageincountryi.Assumingfreeentryinthedierentiatedgoodsector, consumerincomeisequaltothewagemultipliedbythenumberofworkers.TheLagrangian associatedwiththisproblemcanthenbedened: L = z 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( i X i )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ z i + P i X i )]TJ/F38 11.9552 Tf 11.955 0 Td [(w i L i ] B{16 Takingrstorderconditions: @ L @ z i = )]TJ/F24 11.9552 Tf 11.955 0 Td [( z )]TJ/F25 7.9701 Tf 6.586 0 Td [( i X i )]TJ/F24 11.9552 Tf 11.955 0 Td [( =0 B{17 @ L @ X i = z 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( i X )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 i )]TJ/F24 11.9552 Tf 11.955 0 Td [( P i =0 B{18 @ L @ = )]TJ/F21 11.9552 Tf 9.299 0 Td [([ z i + P i X i )]TJ/F38 11.9552 Tf 11.955 0 Td [(w i L i ]=0 B{19 105

PAGE 106

CombiningequationsB-17andB-18: X i = z i P i )]TJ/F24 11.9552 Tf 11.955 0 Td [( B{20 PluggingequationB-20intoB-19yields: z i + z i 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( = w i L i z i 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( = w i L i z i = )]TJ/F24 11.9552 Tf 11.955 0 Td [( w i L i B{21 EquationB-21representstheshareofincomethattherepresentativeconsumerspends ontheoutside-goodgood.Likewise,combiningB-21andB-20showstheshareofincome devotedtodierentiatedgoods: X i P i = w i L i B{22 UsingB-15,B-21,andB-22,theindirectutilityfunctioncaneasilybefound: V P i =[ )]TJ/F24 11.9552 Tf 11.955 0 Td [( I i ] 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( [ w i L i P i ] B{23 = )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( w i L i P i = Bw i L i P i B{24 DividingB-24bythetotalpopulationincountryiyieldsthewelfareperworkerin countryi: 106

PAGE 107

W i = Bw i P i B{25 B.2Producers B.2.1FirstStage Intherststage,Northernrmsmustdecideonthelevelofresearchtasktoassignto NorthernandSouthernresearchers.ANorthernrmfacestheexpectedprotfunction: E n = R n R n )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R [ o n ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R )]TJ/F38 11.9552 Tf 11.956 0 Td [(R n B{26 Takingtherstderivative: @ E n @ R n = R R n )]TJ/F24 11.9552 Tf 11.955 0 Td [( + R 2 [ o n ] )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c B{27 SettingB-27equaltozeroyieldstheoptimalresearchleveli @ 2 E n R n 2 < 0 .Note: @ 2 E n R n 2 = )]TJ/F21 11.9552 Tf 9.298 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( R R n )]TJ/F24 11.9552 Tf 11.956 0 Td [( + R 3 [ o n ] B{28 GiventherestrictionontheIPRparameter, ,thesecondorderconditionhold.Therefore, theFOCyields: R n = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R B{29 ThisvalueofNorthernresearchthenimpliesthatthevalueofSouthernresearchisequal to R )]TJ/F38 11.9552 Tf 11.956 0 Td [(R n : 107

PAGE 108

R s = 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 B{30 Notethatforbothlevelstobepositive, R n cannotbegreaterthanthetotalresearch level, R or: R > R n R > 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R > 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 R > [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 = R > [ o n ] 1 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c B{31 ConditionB-31alsoguaranteesthat R s ispositive.Finally,thelevelofNorthern researchmustalsobepositive: 0 < R n 0 < 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R < 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 R < [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 = R < [ o n ] )]TJ/F38 11.9552 Tf 11.955 0 Td [(c B{32 108

PAGE 109

Thesetwoconditionstogetherimplythenecessaryandsucientcondition-22foran interiorsolution. B.2.2SecondStage Inthesecondstage,rmschoosestheoptimalpricelevelstomaximizeoperatingprot.A rmfromtheNorthmaximizesthefollowingoperatingprotfunction: n =[ x nn p nn )]TJ/F24 11.9552 Tf 11.956 0 Td [( + x ns p ns )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( n ] B{33 Firmsmustset p nn and p ns tomaximizeprot.Usingthedemands-10and-12,the protfunctioncanberewritten: n = [ p )]TJ/F25 7.9701 Tf 6.586 0 Td [( nn a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 nn P n X n p nn )]TJ/F24 11.9552 Tf 11.956 0 Td [( + p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ns a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ns P s X s p ns )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( n ] B{34 Takingthederivativewrt p nn and p ns yieldstworstorderconditions: @ n @ p nn = )]TJ/F24 11.9552 Tf 9.298 0 Td [( [ p )]TJ/F25 7.9701 Tf 6.586 0 Td [( )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 nn a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 nn P n X n p nn )]TJ/F24 11.9552 Tf 11.956 0 Td [( )]TJ/F38 11.9552 Tf 11.956 0 Td [(p )]TJ/F25 7.9701 Tf 6.587 0 Td [( nn a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 nn P n X n ]=0 B{35 @ n @ p ns = )]TJ/F24 11.9552 Tf 9.298 0 Td [( [ p )]TJ/F25 7.9701 Tf 6.587 0 Td [( )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ns a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ns P s X s p ns )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( n )]TJ/F38 11.9552 Tf 11.955 0 Td [(p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ns a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ns P s X s ]=0 B{36 RearrangingtheFOC'syields: @ n @ p nn = )]TJ/F24 11.9552 Tf 9.298 0 Td [( [ p )]TJ/F25 7.9701 Tf 6.586 0 Td [( )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 nn a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 nn P n X n p nn )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(p )]TJ/F25 7.9701 Tf 6.587 0 Td [( nn a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 nn P n X n ]=0 B{37 @ n @ p ns = )]TJ/F24 11.9552 Tf 9.298 0 Td [( [ p )]TJ/F25 7.9701 Tf 6.586 0 Td [( )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ns a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ns P s X s p ns )]TJ/F24 11.9552 Tf 11.956 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( n )]TJ/F38 11.9552 Tf 11.956 0 Td [(p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ns a )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 ns P s X s ]=0 B{38 TheFOC'syield: 109

PAGE 110

p nn = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 B{39 p ns = + n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 B{40 AsimilarprocesscanbeusedtoshowtheoptimalpricelevelssetbyaSouthernrm. B.2.3FreeEntryConditions Usingequation-18,IndtheequilibriumNorthernexpectedprot: E n = o n R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( [ o n ] R 1 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( o n )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 5.48 -9.684 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 + )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( o n )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 5.48 -9.684 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 + )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R TakingthisequationandsettingitequaltozeroyieldstheNorthernfree-entrycondition: E n = n )]TJ/F29 11.9552 Tf 11.955 9.684 Td [()]TJ/F21 11.9552 Tf 5.479 -9.684 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 )]TJ/F29 11.9552 Tf 11.955 9.684 Td [()]TJ/F21 11.9552 Tf 5.48 -9.684 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 + )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R =0 B{41 Usingequation-19,theSouthernfreeentryconditioncanbefoundinasimilar manner: 110

PAGE 111

E s = o s R )]TJ/F38 11.9552 Tf 11.955 0 Td [(R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s = o s )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 R )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( o n R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s = o s )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 R 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( o n R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s = o s 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s = o s )]TJ/F38 11.9552 Tf 11.955 0 Td [(c o n R 1 2 R 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( o n R 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s = o s 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( o n 1 2 o n 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s TakingthisequationandsettingitequaltozeroyieldstheNorthernfree-entrycondition: E s = s 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( n 1 2 n 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(cf s =0 B{42 Next,IndtheoptimalNorthernoperatingprotlevelintermsofthemodel'sparameters inordertofullysolvefortheinteriorcondition-22.Letting x = n 1 2 andrearranging B-42: x 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [(2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 x )]TJ/F21 11.9552 Tf 11.955 0 Td [( c )]TJ/F24 11.9552 Tf 11.955 0 Td [( R =0 Usingthequadraticformula: x = 1 2 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 q 4 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R +4 c )]TJ/F24 11.9552 Tf 11.955 0 Td [( R = )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 R 1 2 p )]TJ/F38 11.9552 Tf 11.955 0 Td [(c + c )]TJ/F24 11.9552 Tf 11.955 0 Td [( 111

PAGE 112

Therearetwopossiblesolutionsforx.Forxtobebothpositiveandunique,Iassume c > .Thatsis,thefollowingequationwillyieldapossiblesolution: x = )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 )]TJETq1 0 0 1 297.305 634.342 cm[]0 d 0 J 0.478 w 0 0 m 8.995 0 l SQBT/F38 11.9552 Tf 297.305 624.366 Td [(R 1 2 p )]TJ/F38 11.9552 Tf 11.956 0 Td [(c + c )]TJ/F24 11.9552 Tf 11.956 0 Td [( B{43 NotethatB-43yieldsanegativex: x < 0 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 < R 1 2 p )]TJ/F38 11.9552 Tf 11.956 0 Td [(c + c )]TJ/F24 11.9552 Tf 11.956 0 Td [( < c Thus,theonlysolutionforxis: x = )]TJ/F38 11.9552 Tf 11.956 0 Td [(c R 1 2 + R 1 2 p )]TJ/F38 11.9552 Tf 11.956 0 Td [(c + c )]TJ/F24 11.9552 Tf 11.956 0 Td [( B{44 Therefore, n = R h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 B{45 Condition-22canthenberewrittenas: < )]TJ/F38 11.9552 Tf 11.956 0 Td [(c h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 < 1 B{46 TheupperboundforB-46holds. 112

PAGE 113

)]TJ/F38 11.9552 Tf 11.955 0 Td [(c < h 1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 1 2 < 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 0 < c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 Thelowerboundholdsonlygivencurrentrestrictions: h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 1 2 i 2 < 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 1 2 < 1 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 1 2 c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 < )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 1 2 c 1 2 < )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 c < )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c = c + < 1 B{47 B.2.4PriceIndexandNumberofFirms Solvingfortheequilibriumpriceindexesusestheoperatingprotequations2-33and -34.Rearrangingequation-33yields: P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n L n = n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s cL s s nn B{48 FillingB-48intotheSouthernoperatingprotinequilibrium: cL s P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s = s ss )]TJ/F38 11.9552 Tf 13.15 8.088 Td [(s sn s ns s nn )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s )]TJ/F38 11.9552 Tf 13.15 8.088 Td [(s sn n s nn B{49 113

PAGE 114

Sincethe s ss )]TJ/F39 7.9701 Tf 13.151 4.708 Td [(s sn s ns s nn > 0 byassumption, P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s ispositivei: s )]TJ/F38 11.9552 Tf 13.151 8.087 Td [(s sn n s nn > 0 B{50 Likewise, P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n ispositivei: cL s P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s < n s ns B{51 UsingequationB-49,thisconditioncanberewrittenas: s < n s ss s ns B{52 Thiscondition,takenwitharearrangedconditionB-50yieldsasucientandnecessary conditionfor P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s > 0 .Theequilibriumnumberofrmswillbepositivefromboth nationsgivenequations-37and-38arepositive.Since s nn s ss )]TJ/F38 11.9552 Tf 10.989 0 Td [(s ns s sn > 0 byassumption, M n > 0 i: s ss s sn > P n P s )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 B{53 Likewise, M s > 0 i: s nn s ns > P s P n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 B{54 Takingthesetwoconditionstogetheryields: s sn s ss < P s P n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 < s nn s ns B{55 114

PAGE 115

FigureB-1.NoSouthernFirms FigureB-2.NoDierentiatedGoods Finally,notethat P s P n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 = L n s nn cL s s s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn n n s ss s nn )]TJ/F24 11.9552 Tf 11.955 0 Td [( s s ns s nn B{56 So,theplausibilityofanequilibriumwherermsexistinbothcountriesthendependson therelativeincomesofbothregions. 115

PAGE 116

B.2.5AlternativeEquilibria FigureB-1showsthecasewherethefreeentryconditionsintersectabovethe M s =0 line.Theequilibriumpointinthiscaseisthepointofintersectionofthe M s =0 lineandthe Northernfreeentryline.Inthiscase,SouthernrmsneverenterthemarketasNorthernrms areabletocaptureahighenoughshareofbothmarketstomakeanyentrybySouthernrms yieldoperatingprotthatdoesnotcoverthexedcostofentry.Asmallnumberofworkersin theSouthandalargeworkforceintheNorthmakesthiscasemorelikelytooccur.Northern rmscanthenearnahighoperatingprotfromtheNorth,wherethermsareabletocapture moreofmarketplacebasedonpreferenceforNortherngoods.Ahighxedcostofentryinthe Southcoupledwithalow R intheNorthalsofacilitateanequilibriumwithoutSouthernrms. FigureB-2showsthecaseinwhichthisintersectionisabovethe M n =0 line.Inthis case,noNorthernrmswillexistinthemarketinequilibrium.Northernrmsdonotenterthe marketduetotheexpectationofanegativeprot.Sincethesermsdonotinnovateanynew products,theSouthernrmsdonothaveachangetoimitateanyoftheproducts.Therefore, thedierentiatedgoodssectorwillnotexist.Lowvaluesof f s and L n andhighvaluesof R and L s increasethelikelihoodofthiscase.Northernrmsthathaveahighxedcostofentryare notabletocoverthexedcostofentryfromthesmallamountofNorthernconsumersandthe largeamountofSouthernconsumerswhoprefertheSouthernproductmore.Furthermore,low xedcostsintheSouthmeanmanySouthernrmsareabletoenterandthuscapturealarge amountoftheSouthermarket.Thus,Northernrmsarenotabletocoverthexedcostsand refusetoenterthemarket.Inthefollowingsection,Iassumebothcountrieshaveactiverms inequilibrium. 116

PAGE 117

B.3ComparativeStatics B.3.1ProofofProposition1 Dierentiatingthefreeentrycondition-30wrt R yieldsthechangeinNorthern operatingprotastotalnecessaryresearchchanges: @ n @ R )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 )]TJ/F24 11.9552 Tf 13.15 8.088 Td [(@ n @ R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n 1 2 = c )]TJ/F24 11.9552 Tf 11.955 0 Td [( B{57 @ n @ R 1 )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n 1 2 # = c )]TJ/F24 11.9552 Tf 11.955 0 Td [( + )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 B{58 Notethat 1 )]TJ/F29 11.9552 Tf 11.955 13.27 Td [( )]TJ/F39 7.9701 Tf 6.587 0 Td [(c R n 1 2 > 0 : 1 )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( )]TJ/F38 11.9552 Tf 11.956 0 Td [(c R n 1 2 # > 0 1 > )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c > R Thisconditionholdsbyinequality-22.Finally,notethat c )]TJ/F24 11.9552 Tf 12.264 0 Td [( + )]TJ/F39 7.9701 Tf 6.586 0 Td [(c n R 1 2 > 0 Thisclearlyholdsif c )]TJ/F24 11.9552 Tf 9.956 0 Td [( > 0 ,whichisassumedwithassumption-36.Therefore, @ n @ R > 0 Now,partiallydierentiatingfree-entrycondition-34wrt R : @ s @ R = 1 2 s n @ n @ R )]TJ/F24 11.9552 Tf 11.955 0 Td [( s [ R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( n 1 2 ] )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 [ 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 n )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 @ n @ R ] B{59 Notethat @ s @ R =0 : @ s @ R =0 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 = 1 2 s n @ n @ R )]TJ/F24 11.9552 Tf 11.956 0 Td [( s [ R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( n 1 2 ] )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 [ 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 n )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 @ n @ R ] 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R 1 2 = @ n @ R 1 n [ R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( n 1 2 + 2 n 1 2 # @ n @ R = n R 117

PAGE 118

FillingintheNorthernfree-entrycondition: @ n @ R = c )]TJ/F24 11.9552 Tf 11.956 0 Td [( +2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 FromB-58: c )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c R n 1 2 = )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n R 1 2 1 )]TJ/F21 11.9552 Tf 11.955 0 Td [(2 R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c n 1 2 c )]TJ/F38 11.9552 Tf 11.955 0 Td [(z R = n )]TJ/F21 11.9552 Tf 11.955 0 Td [(2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c R 1 2 0=0 So, @ s @ R =0 .TakenalongwithB-58,thisimpliesthattheNorthernpriceindexincreases whiletheSouthernpriceindexdecreases.Totallydierentiatingequations-33and-34to examinethechangeof n : 2 6 4 )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s 3 7 5 2 6 4 @ P n @ P s 3 7 5 = 2 6 4 0 3 7 5 @ n Letthedeterminantofthelefthandsidebe j D j .Notethat j D j > 0 : )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s )]TJ/F21 11.9552 Tf 11.956 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n > 0 s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s > s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n s nn s ss > s sn s ns B{60 InequalityB-60holdsbypreviousassumption.UsingCramer'srule,thechangein P n and P s duetoincreasesin n canbefound: @ P n @ n = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s j D j > 0 B{61 @ P s @ n = )]TJ/F24 11.9552 Tf 9.299 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n j D j < 0 B{62 118

PAGE 119

Therefore,theNorthernpriceindexincreaseswhiletheSouthernpriceindexdecreases fromanincreaseinNorthernoperatingprot.SinceSouthernoperatingprotdoesnot change,thereisnoneedtocalculateitseectonthepriceindexesatthistime.Inorderto examinetheeectonthenumberofrms,Ipartiallydierentiateequations-37and-38 wrt P n and P s @ M n @ P n = )]TJ/F21 11.9552 Tf 9.299 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F25 7.9701 Tf 6.587 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn s ns < 0 B{63 @ M n @ P s = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F25 7.9701 Tf 6.586 0 Td [( s s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn s ns > 0 B{64 @ M s @ P n = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ns P )]TJ/F25 7.9701 Tf 6.586 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn s ns > 0 B{65 @ M s @ P s = )]TJ/F21 11.9552 Tf 9.299 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F25 7.9701 Tf 6.587 0 Td [( s s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn s ns < 0 B{66 Usingtheseequations,clearly @ M n @ R < 0 and @ M s @ R > 0 .Theresultscansummarized: @ n @ R > 0 @ s @ R =0 @ P n @ R > 0 @ P s @ R < 0 @ M n @ R < 0 @ M s @ R > 0 Now,dierentiatingequation-23wrt R : @ R n @ R = 1 2 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( n )]TJ/F38 11.9552 Tf 11.956 0 Td [(c R 1 2 )]TJ/F24 11.9552 Tf 23.595 8.088 Td [( 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( B{67 EquationB-67ispositivei: 1 4 n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 > R B{68 119

PAGE 120

Thisinequalityholdsbythepreviousassumptionthat c > .UsingB-67torewrite inequalityB-68: 0 B @ h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 C A > 4 h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 > 4 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 > 2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 > 1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 c )]TJ/F38 11.9552 Tf 11.955 0 Td [(c > )]TJ/F24 11.9552 Tf 11.955 0 Td [( c c > Therefore, @ R n @ R > 0 .Now,Ishowthat @ R s @ R > 0 aswell.FromB-61,dierentiating R s wrt R : @ R s @ R = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 1 )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 n R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 13.15 8.088 Td [(@ n @ R 1 2 R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 # B{69 So, @ R s @ R > 0 i: R > 1 4 n 1 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c B{70 Thisclearlyholdsbylookingatcondition-22.Therefore, R s increasesas R increases. B.3.2ProofofProposition2 Partiallydierentiating-30wrt yieldsthechangeinNorthernoperatingprotas IPR-protectionlawschange: @ n @ =2 R h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 1 2 i 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 c 1 2 B{71 120

PAGE 121

Notethat @ n @ > 0 : @ n @ > 0 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 > 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 c 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 > c 1 2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F24 11.9552 Tf 11.955 0 Td [( > c 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c + c > c 1 > + c Byassumption3, @ n @ > 0 .NowtodeterminethechangeinSouthernoperatingprot,I dierentiatetheprobabilityofimitation-24wrt : @ @ = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( )]TJ/F21 11.9552 Tf 23.465 8.088 Td [(1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( + 1 2 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 )]TJ/F21 11.9552 Tf 11.955 0 Td [( n )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 @ n @ B{72 Notethat @ @ > 0 i: )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 > )]TJ/F21 11.9552 Tf 10.494 8.088 Td [(1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 )]TJ/F21 11.9552 Tf 11.955 0 Td [( n )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 @ n @ R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [(> )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 13.151 8.088 Td [(1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 )]TJ/F21 11.9552 Tf 11.955 0 Td [( n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 @ n @ )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 )]TJ/F21 11.9552 Tf 11.955 0 Td [( n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 @ n @ > 2 R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c n )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [(2 B{73 Notethat n )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 @ n @ = h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 1 2 i )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 h )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F21 11.9552 Tf 11.956 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 c 1 2 i Also, R )]TJ/F38 11.9552 Tf 11.956 0 Td [(c n = )]TJ/F38 11.9552 Tf 11.955 0 Td [(c h 1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 2 121

PAGE 122

Therefore, and @ @ areindependentof R .Now,inequalityB-73canberewritten: )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 )]TJ/F29 11.9552 Tf 11.955 13.27 Td [(h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 h )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.956 0 Td [( )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 c 1 2 i h )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 i > 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 [ 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 ] )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 )]TJ/F21 11.9552 Tf 9.298 0 Td [(2 2+ 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( > h 1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 h )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 c 1 2 +2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 i h 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 i 1+ > )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 c 1 2 +2 )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F38 11.9552 Tf 11.956 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 + 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 ++ c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 > )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 c 1 2 +2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 + )]TJ/F21 11.9552 Tf 11.955 0 Td [(1+ 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 [1+2 ] > 2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 [1+2 ] > 2 1 2 )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F22 5.9776 Tf 5.757 0 Td [(1 2 2 [ )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.956 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 ]+ c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 +2 > 0 = c 1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 +2 > 2 1 2 [ )]TJ/F38 11.9552 Tf 11.955 0 Td [(c )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 ] B{74 ConditionB-74isthenecessaryandsucientconditionfor @ s @ < 0 .Thechangein SouthernoperatingprotisthenambiguousanddependsonB-74.Thechangeinprice indexesisambiguousaswell.Totallydierentiatingequations-33and-34toexamine thechangeof : 2 6 4 )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.956 0 Td [(1 s ns P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s 3 7 5 2 6 4 @ P n @ P s 3 7 5 = 2 6 4 @ n @ @ s @ 3 7 5 @ UsingCramer'srule,thechangein P n and P s duetoincreasesin canbefound: @ P n @ = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 @ n @ s ss P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s )]TJ/F24 11.9552 Tf 11.955 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 @ s @ s ns P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s j D j # B{75 @ P s @ = )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 @ s @ s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n )]TJ/F25 7.9701 Tf 13.151 5.699 Td [(@ n @ )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n j D j # B{76 122

PAGE 123

Thedirectionalchangeofthepriceindexesthendependsontherelativemarketshares andthedegreeofchangeintheoperatingprot. @ P n @ > 0 i: @ n @ s ss s ns > @ s @ B{77 Likewise, @ P s @ < 0 i: @ n @ > @ s @ s nn s sn B{78 Since R onlyaects @ n @ andnot @ s @ ,alarge R increasethelikelihoodofB-77and B-78holding.Likewise,alarge s ss s ns andsmall s nn s sn increasethelikelihoodoftheabove conditionsholding.Note,thatbypreviousassumption, @ P n @ < 0 and @ P s @ < 0 isimpossible: @ s @ s ns s ss > @ n @ > @ s @ s nn s sn Since s ns s ss > s nn s sn doesnothold,thiscasecannothappen.Itispossiblefor @ P n @ > 0 and @ P s @ > 0 .Also, @ P n @ < 0 and @ P s @ > 0 ispossible.Finally, @ P n @ > 0 and @ P s @ < 0 ispossible.So, atleastonepriceindexmustgoupasIPR-protectionbecomesweaker.Whenthepriceindexes moveinoppositedirection,thechangeinthenumberofrmsisunambiguous.AsSouthern operatingprotincreases,thechangeinSouthernprotasIPRchangesbecomesgreater.As seeninSectionC.4,Southernoperatingprotincreasesas f s increases.So,atechnologically advancedSouthncreasesthechancesof @ P n @ > 0 and @ P s @ < 0 .FromequationsB-63,B-64, B-65,andB-66,anincreasingNorthernpriceindexcoupledwithadecreasingSouthern priceindeximplies @ M n @ < 0 and @ M s @ > 0 .Thechangeinthenumberofrmscanbewritten as: @ M n @ = @ M n @ P n @ P n @ + @ M n @ P s @ P s @ B{79 @ M s @ = @ M s @ P n @ P n @ + @ M n @ P s @ P s @ B{80 Todeterminethedirectionofoutsourcing,Ilookathoweachrmrespondstochangesin IPR-protection.AsIPR-protectionchanges,theresearchallocationschangeaswell.Unlikethe 123

PAGE 124

caseofaNortherntechnologychange,theresearchtaskchangesinoneregionmustequalthe oppositeofthechangeintheotherregion.Thatis, @ R s @ = )]TJ/F25 7.9701 Tf 10.494 4.707 Td [(@ R n @ .Ionlyexaminethechangein Southernresearchasaresult.FromB-30,dierentiating R s wrt : @ R s @ = 1 1 )]TJ/F24 11.9552 Tf 11.956 0 Td [( 2 R )]TJ/F29 11.9552 Tf 11.955 16.857 Td [( R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 # )]TJ/F21 11.9552 Tf 11.955 0 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 @ n @ R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 # B{81 So, @ R s @ > 0 i: 2 R s > @ n @ R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 + R n )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 2 R > R n )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 h 1 2 + )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 i + @ n @ R n )]TJ/F38 11.9552 Tf 11.955 0 Td [(c 1 2 2 R n )]TJ/F38 11.9552 Tf 11.956 0 Td [(c 1 2 > n h 1 2 + )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 2 i + @ n @ 1 2 n )]TJETq1 0 0 1 174.63 425.411 cm[]0 d 0 J 0.478 w 0 0 m 8.995 0 l SQBT/F38 11.9552 Tf 174.63 415.435 Td [(R c )]TJ/F24 11.9552 Tf 11.955 0 Td [( > n + + @ n @ )]TJ/F29 11.9552 Tf 11.291 16.857 Td [( c )]TJ/F24 11.9552 Tf 11.956 0 Td [( R > n + @ n @ B{82 Frompreviousassumptions,theLHSisnegativewhiletheRHSispositive.So,B-82will neverhold.Therefore, @ R s @ < 0 and @ R n @ > 0 B.3.3ProofofProposition3 Fromequations-30and-31,achangein L s hasnoeecton n and s .Totally dierentiatingequations-33and-34toexamineanexogenouschangeinSouthernsize: 2 6 4 )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s cL s )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n L n )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s cL s 3 7 5 2 6 4 @ P n @ P s 3 7 5 = 2 6 4 )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s c )]TJ/F38 11.9552 Tf 9.298 0 Td [(s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s c 3 7 5 @ L n 124

PAGE 125

UsingCramer'srule,thechangein P n and P s duetoincreasesin L s canbefound: @ P n @ L s = )]TJ/F38 11.9552 Tf 9.299 0 Td [(s ns P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 s c 2 L s + s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s 2 c 2 )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 s L s j D j =0 B{83 @ P s @ L s = )]TJ/F21 11.9552 Tf 9.299 0 Td [( )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s nn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(2 n 2 L n cs ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s + s ns P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 s c )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 s sn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(2 n L n j D j < 0 B{84 FromequationB-83,thechangeinsizeintheSouthhasnoeectonthepriceindexin theNorth.ThepriceindexintheSouthdecreases.UsingequationsB-63,B-64,B-65, andB-66,theeectsonthenumberofrmscanbeseen. M n decreaseswhile M s increases whentheSouthernmarketgrowsinsize.Theresultscanbesummarizedas: @ n @ L s =0 @ s @ L s =0 @ P n @ L s =0 @ P s @ L s < 0 @ M n @ L s < 0 @ M s @ L s > 0 Inowshowthatresearchlocationdecisionsdonotdependonmarketsize.FromB-30, dierentiating R s wrt L s : @ R s @ L n = )]TJ/F29 11.9552 Tf 11.291 16.857 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 @ n @ L s =0 B{85 Since @ n @ L n =0 @ R s @ L n equalszeroaswell.Therefore, @ R n @ L n =0 aswell. B.3.4ProofofProposition4 Forthisproof,Icomparethethelevelsof P n P s M n ,and M s when t =0 andwhen t > 0 .First,theinclusionofatarichangesconsumerdemandsintheSouthasthetari addstoSouthernconsumerincome.Thus,Southernconsumerincomebecomes cL s + M n x ns t Consumerdemands-12and-13canberewritten: x ns = p )]TJ/F25 7.9701 Tf 6.586 0 Td [( ns a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ns cL s + M n x ns t B{86 x ss = p )]TJ/F25 7.9701 Tf 6.587 0 Td [( ss a )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 ss cL s + M n x ns t B{87 125

PAGE 126

ThepricesetbytheNorthernrmintheSouthernmarketincreasestoincludethecostof thetari. p ns = + n + t )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 B{88 Inspectionofthetwofree-entryconditions-30and-31,showthattheinclusionofa taridoesnotchangethe n and s .Noticethatapositiveimporttaridoesnotaecteach rm'sresearchintensity.FromB-5,dierentiating R s wrt t : @ R s @ t = )]TJ/F29 11.9552 Tf 11.291 16.857 Td [( 1 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( 1 2 R )]TJ/F38 11.9552 Tf 11.955 0 Td [(c n 1 2 @ n @ t =0 B{89 So,theresearchallocationdecisionisnotaectbythetariamount.Theequilibrium protequations-33and-34become: n = 1 [ s nn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n L n + s ns P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s cL s + M n x ns t ] B{90 s = 1 [ s sn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n L n + s ss P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s cL s + M n x ns t ] B{91 CombiningB-90andB-91tond P n rearrangingtermsyields: s ss n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ss s nn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n L n = s ns s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n L n B{92 s ss n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s = P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn B{93 = P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 n = L n )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s ss n )]TJ/F38 11.9552 Tf 11.956 0 Td [(s ns s s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 B{94 Theintroductionoftonlyaects s ns inequationB-94.Introducingataridecreases s ns : @ s ns @ t = 1 )]TJ/F24 11.9552 Tf 11.955 0 Td [( + n + t + n + t )]TJ/F21 11.9552 Tf 11.955 0 Td [(1 a ns 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( B{95 126

PAGE 127

Thus,theaectthetarihasontheNorthernpriceindexcanbefoundbydierentiating B-94wrtto s ns : @ P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n @ s ns = s sn s ss n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 + )]TJ/F24 11.9552 Tf 9.298 0 Td [( s L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn B{96 = s sn s ss n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn s ns s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ss s nn s + s ns s sn s L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 B{97 = s sn s ss n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ss s nn s L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 B{98 = s sn n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s nn s s ss L n s ss s nn )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 B{99 Bycondition-35, @ P )]TJ/F22 5.9776 Tf 5.756 0 Td [(1 n @ s ns < 0 .So,introducingthetariincreasestheNorthernprice indexanddecreasesNorthernwelfare.Asthetariincreases,theNorthernpriceindexwill unambiguouslyincreaseaswell.RearrangingequationB-91yields: P )]TJ/F22 7.9701 Tf 6.586 0 Td [(1 s = s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn P )]TJ/F22 7.9701 Tf 6.587 0 Td [(1 n L n s ss cL s + M n x ns t B{100 TheinclusionofatariincreasesSouthernincomeandNorthernpriceindex.Therefore, theSouthernpriceindexmustdecrease.Asthethetariincreases,thechangeintheSouthern priceindexisambiguous.Increasesinthetaricanconceivablyreducetarirevenueas Northernrmsexitthemarket.Dierentiatingequations-37and-38wrt s ns : @ M n @ s ns = s sn s ss P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( n )]TJ/F38 11.9552 Tf 11.955 0 Td [(s sn P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( s s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 B{101 @ M s @ s ns = s sn s nn P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( s )]TJ/F38 11.9552 Tf 11.956 0 Td [(s ns P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.956 0 Td [(s ns s sn 2 )]TJ/F38 11.9552 Tf 35.572 8.087 Td [(P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn B{102 InspectingofB-101showsthatthenumberofNorthernrmsdecreaseas s ns decreases. Since M n decreasesfromboththepriceindexchangesandthemarketsharechanges,the increasein t unambiguouslydecreasesthenumberofNorthernrms.EquationB-102canbe 127

PAGE 128

simplied: @ M s @ s ns = s sn s nn P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ss s nn P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 B{103 = s nn s sn P 1 )]TJ/F25 7.9701 Tf 6.587 0 Td [( s )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ss P 1 )]TJ/F25 7.9701 Tf 6.586 0 Td [( n s nn s ss )]TJ/F38 11.9552 Tf 11.955 0 Td [(s ns s sn 2 < 0 B{104 NotethatequationB-104isnegativefromtheassumptionthat M n > 0 .Using equationsB-63-B-66,thenumberofrmsintheNorthdecreaseswhilethenumberof Southernrmsincreasesasthetariisintroduced.So,theSouthbenetsfromatarion NorthernimportsatthecostoftheNorth.Raisingthetariafteritsintroductiondecreases NorthernwelfareandhasanambiguouseectonSouthernwelfare. 128

PAGE 129

REFERENCES Acemoglu,D.,Gancia,G., & Zilibotti,F..OshoringandDirected TechnicalChange. AmericanEconomicJournal:Macroeconomics 7,84-122. http://dx.doi.org/10.1257/mac.20130302 Antras,P., & Helpman,E..Globalsourcing. JournalofPoliticalEconomy 112, 552-580. Baik,K.H..Eortlevelsincontestswithtwoasymmetricplayers. SouthernEconomic Journal 61,367-378.http://dx.doi.org/10.2307/1059984 Balta,N., & Delgado,J..HomebiasandmarketintegrationintheEU. CESifo EconomicStudies ,55,110-144.http://dx.doi.org/10.1093/cesifo/ifn037 Barefoot,K.B..USMultinationalCompanies:OperationsofUSParentsandTheir ForeignAliatesin2010. SurveyofCurrentBusiness ,91,51-74. Baye,M.R., & HeidrunC.H..Thestrategicequivalenceofrent-seeking, innovation,andpatent-racegames. GamesandEconomicBehavior 44,217-226. http://dx.doi.org/10.1016/s0899-8256-7 Baylis,K.,Garduo-Rivera,R., & Piras,G..ThedistributionaleectsofNAFTAin Mexico:Evidencefromapanelofmunicipalities. RegionalScienceandUrbanEconomics 42,286-302.http://dx.doi.org/10.1016/j.regsciurbeco.2011.09.006 Behar,A., & Venables,A.J..5Transportcostsandinternationaltrade. AHandbookof TransportEconomics ,97. Berry,H..Globalintegrationandinnovation:Multicountryknowledgegenerationwithin MNCs. StrategicManagementJournal ,35,869-890. Bilir,L.K..Patentlaws,productlife-cyclelengths,andmultinationalactivity. The AmericanEconomicReview ,104,1979-2013. Buss,P., & Peukert,C..R & Doutsourcingandintellectualpropertyinfringement. ResearchPolicy ,44,977-989. Branstetter,L.,Fisman,R., & Foley,C.F..Dostrongerintellectualpropertyrights increaseinternationaltechnologytransfer?EmpiricalevidencefromUSrm-leveldata. The QuarterlyJournalofEconomics ,21,321-349. Branstetter,L.,Fisman,R.,Foley,C.F., & Saggi,K..Doesintellectualpropertyrights reformspurindustrialdevelopment? JournalofInternationalEconomics ,83,27-36. Branstetter,L.,Fisman,R.,Foley,C.F., & Saggi,K..Intellectualpropertyrights, imitation,andforeigndirectinvestment:TheoryandevidenceNo.w13033. National BureauofEconomicResearch 129

PAGE 130

Canals,C., & Sener,F..Oshoringandintellectualpropertyrightsreform. Journalof DevelopmentEconomics 108,17-31.http://dx.doi.org/10.1016/j.jdeveco.2014.01.001 Carnahan,S.,Agarwal,R., & Campbell,B.A..Heterogeneityinturnover:Theeect ofrelativecompensationdispersionofrmsonthemobilityandentrepreneurshipofextreme performers. StrategicManagementJournal ,33,1411-1430. Chowdhury,S.M., & Sheremeta,R.M..AGeneralizedTullockcontest. PublicChoice 147,413-420.http://dx.doi.org/10.2139/ssrn.1755604 d'Aspremont,C. & Jacquemin,A..CooperativeandnoncooperativeR & Dinduopoly withspillovers." TheAmericanEconomicReview 78,1133-1137. DAgostino,L.M.,Laursen,K., & Santangelo,G.D..TheimpactofR & Doshoringon thehomeknowledgeproductionofOECDinvestingregions. JournalofEconomicGeography 13,145-175. Demidova,S., & Rodriguez-Clare,A.009.Tradepolicyunderrm-levelheterogeneityina smalleconomy. JournalofInternationalEconomics ,78,100-112. Demirbag,M., & Glaister,K.W..FactorsdeterminingoshorelocationchoiceforR & D projects:AComparativestudyofdevelopedandemergingregions. JournalofManagement Studies 47,1534-1560.http://dx.doi.org/10.1111/j.1467-6486.2010.00948.x Dietl,H.M.,Franck,E.P.,Grossmann,M., & Lang,M..Contesttheoryandits applicationsinsports. Dinopoulos,E., & Tsoulouhas,T..PerformancePayandOshoring. Journalof EconomicsandManagementStrategy ,Forthcoming. Dinopoulos,E., & Segerstrom,P..Intellectualpropertyrights,multinationalrmsand economicgrowth. JournalofDevelopmentEconomics ,92,13-27. Dinopoulos,E., & Segerstrom,P..North-southtradeandeconomicgrowth. Stockholm SchoolofEconomics .http://dx.doi.org/10.11130/jei.2002.17.3.554 Dinopoulos,E., & Syropoulos,C..Internationaldiusionandappropriabilityof technologicalexpertise. AdvancesinAppliedMicroeconomics ,7,115-137. Ekholm,K., & Hakkala,K.LocationofR & Dandhigh-techproduction byverticallyintegratedmultinationals. TheEconomicJournal 117,512-543. http://dx.doi.org/10.1111/j.1468-0297.2007.02028.x Evans,C.L..HomeBiasinTrade:LocationorForeign-ness?. FRBofNewYorkSta Report ,. Ewerhart,C..MixedequilibriainTullockcontests. EconomicTheory ,60,59-71. 130

PAGE 131

Feenstra,R., & Kee,H.L..Exportvarietyandcountryproductivity:Estimatingthe monopolisticcompetitionmodelwithendogenousproductivity. Journalofinternational Economics ,74,500-518.http://dx.doi.org/10.1016/j.jinteco.2006.11.006 Feenstra,R. & Taylor,A.. InternationalTrade .NewYork,NY:Worth. Feenstra,R.C., & Romalis,J..InternationalPricesandEndogenousQuality. Quarterly JournalofEconomics ,129. Felbermayr,G.,Jung,B., & Larch,M.3.Optimaltaris,retaliation,andthewelfare lossfromtariwarsintheMelitzmodel. JournalofInternationalEconomics ,89,13-25. http://dx.doi.org/10.1016/j.jinteco.2012.06.001 Friberg,R.,Paterson,R.W., & Richardson,A.D..Whyistherea homebias?AcasestudyofWine.JournalofWineEconomics,6,37-66. http://dx.doi.org/10.1017/S193143610000105X Fonseca,M.A..AnExperimentalinvestigationofasymmetric contests. InternationalJournalofIndustrialOrganization 27,582-591. http://dx.doi.org/10.1016/j.ijindorg.2009.01.004 Gambardella,A.,Ganco,M., & Honor,F.014.Usingwhatyouknow:Patentedknowledge inincumbentrmsandemployeeentrepreneurship. OrganizationScience ,26,456-474. Ganco,M..CuttingtheGordianknot:Theeectofknowledgecomplexityonemployee mobilityandentrepreneurship. StrategicManagementJournal ,34,666-686. Garcia-Canal,E., & Martinez-Noya,A..InternationalevidenceonR & Dservices outsourcingpracticesbytechnologicalrms. MultinationalBusinessReview 22,372-393. http://dx.doi.org/10.1108/MBR-08-2014-0042 Glass,A.J., & Wu,X.07.Intellectualpropertyrightsandqualityimprovement. Journalof DevelopmentEconomics ,82,393-415. Ginarte,J.C., & Park,W.G..Determinantsofpatentrights:Across-nationalstudy. Researchpolicy,26,283-301. Grossman,G.M., & Helpman,E..Integrationvs.Outsourcingin IndustryEquilibrium. TheQuarterlyJournalofEconomics ,117:85-120. https://doi.org/10.1162/003355302753399454 Grossman,G.M., & Helpman,E..OutsourcinginaGlobalEconomy. Reviewof EconomicStudies ,135-159. Gustafsson,P., & Segerstrom,P.S..North-southtradewithmultinationalrms andincreasingproductvariety. InternationalEconomicReview ,52,1123-1155. http://dx.doi.org/10.1111/j.1468-2354.2011.00662.x 131

PAGE 132

Helpman,E..Innovation,imitation,andintellectualpropertyrights. Econometrica 61.http://dx.doi.org/10.2307/2951642 Hedge,D., & Hicks,D..TheMaturationofglobalcorporateR & D:Evidence fromtheactivityofUSforeignsubsidiaries. ResearchPolicy 37,390-406. http://dx.doi.org/10.1016/j.respol.2007.12.004 Hoekman,B.M.,Maskus,K.E., & Saggi,K..Transferoftechnologytodeveloping countries:unilateralandmultilateralpolicyoptions. WorldDevelopment ,33,1587-1602. Hummels,D., & Klenow,P.J..Thevarietyandqualityofanation'sexports. American EconomicReview ,95,704-723. Ito,B. & Wakasugi,R..Whatfactorsdeterminethemodeofoverseas R & Dbymultinationals?Empiricalevidence. ResearchPolicy 36,127512-87. http://dx.doi.org/10.1016/j.respol.2007.04.011 Ivus,O.,Saggi,K., & Park,W..PatentProtectionandtheIndustrialCompositionof MultinationalActivity:EvidencefromUSMultinationalFirms. Javorcik,B.S..Thecompositionofforeigndirectinvestmentandprotectionof intellectualpropertyrights:Evidencefromtransitioneconomies. EuropeanEconomicReview, 48,39-62. Kamien,M.I.,Muller,E., & Zang,I..ResearchjointventuresandR & Dcartels. The AmericanEconomicReview 82,1293-1306. Keller,W., & Yeaple,S.R..Thegravityofknowledge. TheAmericanEconomic Review ,103,1414-1444. Lai,E.L.-C.,Riezman,R., & Wang,P..Outsourcingofinnovation. EconomicTheory 38,485-515.http://dx.doi.org/10.2139/ssrn.935003 Leininger,W..Moreecientrent-seeking:AMunchhausensolution. PublicChoice 75,43-62.http://dx.doi.org/10.1007/978-1-4757-5055-3 Lewin,A.Y.,Massini,S., & Peeters,C..Whyarecompaniesoshoringinnovation? TheEmergingglobalracefortalent. JournalofInternationalBusinessStudies 40, 901-925.http://dx.doi.org/10.1057/jibs.2008.92 Lundin,N., & Serger,S.S..GlobalizationofR & DandChina. IFNWorkingPaper 710, ResearchInstituteofIndustrialEconomica,Sweden. Melitz,M.J..TheImpactoftradeonintra-industryreallocationsandaggregate industryproductivity. Econometrica ,71,1695-1725.http://dx.doi.org/10.3386/w8881 Moncada-Paterno-Castello,P.,Vivarelli,M., & Voigt,P..Driversandimpactsin theglobalizationofcorporateR & D:AnIntroductionbasedontheEuropeanexperience. IndustrialandCorporateChange 20,583-603.http://dx.doi.org/10.1093/icc/dtr005 132

PAGE 133

Naghavi,A.,Spies,J., & Toubal,F..Intellectualpropertyrights,productcomplexity andtheorganizationofmultinationalrms. CanadianJournalofEconomics ,48,881-902. Nitzan,S..Modellingrent-seekingcontests. EuropeanJournalofPoliticalEconomy 10,41-60. Obstfeld,M., & Rogo,K..Thesixmajorpuzzlesininternationalmacroeconomics:is thereacommoncause?. InNBERMacroeconomicsAnnual2000 ,15,339-412 Qian,Y..Donationalpatentlawsstimulatedomesticinnovationinaglobalpatenting environment?Across-countryanalysisofpharmaceuticalpatentprotection,1978-2002. The ReviewofEconomicsandStatistics ,89,436-453. Rosen,S..PrizesandIncentivesinEliminationTournaments. TheAmericanEconomic Review ,76,701-715.? Saunders,K.M..Patentnonuseandtheroleofpublicinterestasadeterrentto technologysuppression. Harv.JLTech ,15,389. Simon,H.A..Thearchitectureofcomplexity. ProceedingsoftheAmericanphilosophicalsociety ,106,467-482. Skaperdas,S..Contestsuccessfunctions. Economictheory ,7,283-290. Snyder,J.M..Electiongoalsandtheallocationofcampaignresources. Econometrica: JournaloftheEconometricSociety ,637-660. Symeonidis,G.,Innovation,FirmSizeandMarketStructure:Schumpeterian HypothesesandSomeNewThemes. OECDEconomicsDepartmentWorkingPapers No.161,OECDPublishing,Paris.http://dx.doi.org/10.1787/603802238336 Todo,Y., & Shimizutani,S..R & Dintensityforinnovativeandadaptivepurposes inoverseassubsidiaries:EvidencefromJapanesemultinationalenterprises. Researchin InternationalBusinessandFinance ,23,31-45. Torstensson,J..Countrysizeandcomparativeadvantage:Anempiricalstudy. WeltwirtschaftlichesArchiv ,134,590-611. U.S.PatentStatisticsReport.,June.Retrievedfrom http://www.uspto.gov/web/oces/ac/ido/oeip/taf/us stat.htm UnescoScienceReport2010 .. Venables,A.J..Tradeandtradepolicywithdierentiatedproducts:A Chamberlinian-Ricardianmodel. TheEconomicJournal ,97,700-717. http://dx.doi.org/10.2307/2232931 Zhao,M..ConductingR & Dincountrieswithweakintellectual propertyrightsprotection. ManagementScience 52,1185-1199. http://dx.doi.org/10.1287/mnsc.1060.0516 133

PAGE 134

BIOGRAPHICALSKETCH ZacharyCohlewasborninasuburbofPhiladelphia,Pennsylvania.Heobtainedhis BachelorofScienceineconomics,mathematics,andphilosophyfromtheUniversityof Pittsburghin2011.AftergraduatingfromtheUniversityofPittsburgh,heimmediatelybegan studyingattheUniversityofFloridaandobtainedaMasterofArtsinEconomicsin2014.He receivedhisPh.D.fromtheUniversityofFloridainthespringof2017. ZacharyCohle'sresearchmainlyfallsintheInternationalTradeeld.Hisdissertation primarilydiscussesandexaminestherelativelynewpracticeofoshoringinnovativeR & D todevelopingcountries.HisworkcombinesaspectsofInternationalTrade,Development Economics,andIndustrialOrganization.Hehaspresentedpapersatthe94th,95th,and96th SouthernEconomicsAssociation'sAnnualConference.HehastaughtbothInternationalTrade andGameTheoryandApplicationundergraduateclassesattheUniversityofFlorida. 134