The timing of licensing: theory and empirics

Description

Niveau: Supérieur, Doctorat, Bac+8
The timing of licensing: theory and empirics? Marie-Laure Allain†, Emeric Henry‡and Margaret Kyle April 27, 2009 Abstract Markets for technology licenses are thought to create efficiency gains by allowing for division of labor in research and development of innovations. However, these gains depend on the timing of technology transfer: the licensee should take over development at the stage at which he has an efficiency advantage. We show that in an environment with asymmetric information about the value of the innovation and where information becomes available over time, deviations from the optimal timing of technology transfer will occur. Competition between potential licensees has an ambiguous effect on this timing. For concentrated markets, in which there are few potential licensees, an increase in the number of potential licensees may delay licensing. The opposite is true for very competitive markets. We test these predictions with data on contracts signed between biotechnology firms and large pharmaceutical firms, and find evidence consistent with our theory. PRELIMINARY AND INCOMPLETE Jel Codes: L13, L24, L65. Keywords: Innovation, Licensing, Market structure, Pharmaceuticals, Biotechnology. 1 Introduction Licensing has become an increasingly popular means of transferring innovative technologies over the last decades.1 As markets for technology grow in importance, the timing of licensing becomes an essential consideration. For instance, in an industry with two firms and two main stages of research, one firm may be more efficient in conducting early stage research and the second more efficient in the final stage.

• major pharmaceutical


• innovation


• licensing has become


• licensing


• licensing between


• over development


• costs ∆

Sujets

• École polytechnique
• Allain
• Innovation
• Licence (juridique)

Informations

• Publié par: profil-zyak-2012
• Nombre de lectures: 12
• Langue: English

Extrait

Thetimingoflicensing:theoryandempirics ∗ Marie-LaureAllain, † EmericHenry ‡ andMargaretKyle § April27,2009 Abstract Marketsfortechnologylicensesarethoughttocreateefficiencygainsbyallowingfordivision oflaborinresearchanddevelopmentofinnovations.However,thesegainsdependonthetiming oftechnologytransfer:thelicenseeshouldtakeoverdevelopmentatthestageatwhichhehas anefficiencyadvantage.Weshowthatinanenvironmentwithasymmetricinformationabout thevalueoftheinnovationandwhereinformationbecomesavailableovertime,deviationsfrom theoptimaltimingoftechnologytransferwilloccur.Competitionbetweenpotentiallicenseeshas anambiguouseffectonthistiming.Forconcentratedmarkets,inwhichtherearefewpotential licensees,anincreaseinthenumberofpotentiallicenseesmaydelaylicensing.Theoppositeistrue forverycompetitivemarkets.Wetestthesepredictionswithdataoncontractssignedbetween biotechnologyfirmsandlargepharmaceuticalfirms,andfindevidenceconsistentwithourtheory. PRELIMINARYANDINCOMPLETE JelCodes:L13,L24,L65. Keywords:Innovation,Licensing,Marketstructure,Pharmaceuticals,Biotechnology. 1Introduction Licensinghasbecomeanincreasinglypopularmeansoftransferringinnovativetechnologiesoverthe lastdecades. 1 Asmarketsfortechnologygrowinimportance,thetimingoflicensingbecomesan essentialconsideration.Forinstance,inanindustrywithtwofirmsandtwomainstagesofresearch, onefirmmaybemoreefficientinconductingearlystageresearchandthesecondmoreefficientinthe finalstage.Itissociallyoptimaltotransfertheinventionattheendofthefirststage.Anyother ∗ WethankJoshGans,DenisGrombandJ.P.Benoitforcommentsonanearlierdraft,andseminarparticipantsat Duke,EARIE2008,LondonBusinessSchool,Lucca,Wharton,theUniversityofMissouri,HarvardBusinessSchool, Emory,INSEAD,Universite´LibreDeBruxellesandHECLausanneforusefulsuggestions. † CNRS,EcolePolytechnique(91128Palaiseau,France,email:[e-mail]) ‡ LondonBusinessSchool(SussexPlace,Regent’sPark,LondonNW14SA,England,email:[e-mail])and RPEC § LondonBusinessSchool(SussexPlace,Regent’sPark,LondonNW14SA,England,email:[e-mail]), NBERandCEPR. 1 Estimatesofthesizeoftheglobalmarketfortechnologylicensingrangefrom $ 5.6billioninthe1980sand $ 36billion to $ 100billioninthelate1990s,andJapanesefirmsreportedearning ¥ 340billion( $ 3.2billion)fromlicensestoforeign firmsin2002(OrganisationforEconomicCo-operationandDevelopment,2005). 1

timingincreasesthecostofinnovatingandmightleadtotheinnovationbeingabandoned.Thus,the timingoflicensingmayimpacttheoverallinnovationrate,andmayultimatelyaffecteconomicgrowth. Specializationindifferentphasesoftheinnovationprocessiscommoninmanyindustries.For instance,smallbiotechnologycompaniesmighthaveacomparativeadvantageinachievingearlystage discoveriesincertainfieldswhilelargepharmaceuticalfirmsareconsideredmoreefficientinconducting laterstageclinicaltesting.GuedjandScharfstein(2004)showcleardifferencesinthesuccessrates ofdrugcandidatesincancerbetweenexperienced,largerfirmsandsmallbiotechfirms.Inanarticle inNatureReviews:DrugDiscovery,KalamosandPinkus(2003)claimthat“[f]orthepharmaceutical industry,innovativebiotechcompoundshaveservedtobuttresslaggingR&Dproductivity...pharma bringsclinicaldevelopment,portfoliomanagementandcommercializationskillsthatarelackingin manybiotechcompanies”.Boththeacademicliteratureandthepopularpressnotethatasignificant proportionofthedrugsmarketedbymajorpharmaceuticalcompaniesoriginatefromlicensingdeals withsmallerbiotechnologyfirms.Angell(2004)claimsthatonethirdofthedrugsmarketedbymajor pharmaceuticalcompaniesoriginatefromlicenseswithbiotechsoruniversities,andina2006surveyof innovation,TheEconomistnotesthat“BigPharma’sR&Dactivityisnowconcentratedasmuchon identifyinganddoingdealswithsmall,innovativefirmsasitisontryingtodiscoveritsownblockbuster drugs”(Economist,2006).Understandingthefactorsthatinfluencethetimingoflicensingbetween biotechandlargepharmaceuticalfirmsisessential.Inefficientdelaycouldcontributetothedecrease inproductivityinthisindustry. Thereissomeevidencethatpharmaceuticallicensingcontractshavebeensignedwithincreased delaysinrecentyears,aperiodalsocharacterizedbyalownumbersofnewdrugslaunched.This delaydoesnotmerelyreflectanincreaseinthetotaltimerequiredfordrugdevelopment;rather,the technologytransferisoccurringatlaterstagesofdevelopment,afterthecompletionofmoreadvanced clinicaltrials.ThisisillustratedinFigure1,whereweshowthefractionofdealssignedineach developmentphaseoverthelastthreedecades.Thisdelayintechnologytransferalsocoincideswitha periodofincreasedmarketconcentration,astheindustryhasknownsubstantialmergeractivity.The linkbetweenthenumberofcompetitorsandthetimingoflicensingisoneofthecentralconsiderations inthispaper.Itisanimportantissuethatcouldbecomeanessentialfactorinmergerreviews. Ifmarketsfortechnologyworkefficiently,thenthetimingoflicensingdependsonlyonthepro- ductiveefficiencyofthecontractingfirms.However,wefocusontwofactorsthataretypicalinmany innovativeindustriesandthatmaycreatedeviationsfromthesociallyoptimaltimingoftechnology transfer.First,innovativeenvironmentsareoftencharacterizedbyasymmetricinformation.Theinno- vatorisoftenbetterinformedaboutthevalueofherideathanapotentiallicensee.Second,asresearch progresses,informationisrevealedabouttheunderlyingvalueoftheinventionandtheinformation asymmetryshrinks.Forinstance,drugcandidatesundergoaseriesofclinicaltrialsrequiredforregu- latoryapproval.Onceaclinicaltrialphaseissuccessfullycompleted,outsideobserversbecomemore confidentofthedrugcandidate’svalue. Wecapturetheelementspreviouslydiscussedinatwoperiodmodelinvolvingonesoleinnovator andnproducers.Thefirstperiodischaracterizedbyasymmetricinformation:theinnovatorknows thevalueofherinventionwhereastheproducersareuninformed.Inthesecondperiod,theuncertainty isresolved.Theproductiscostlytodevelopfromthefirsttothesecondperiodfortheinnovator, butiscostlessfortheproducers.Itisthussociallyoptimaltotransfertheinnovationtooneofthe producersinthefirstperiod.However,theexistenceofasymmetricinformationcancreatedeviations fromthissociallyoptimaltiming. Innovationsaretransferredbysigninglicensingcontracts.Anexclusivelicensecanbesigned 2

betweentheinnovatorandoneoftheproducersateitherperiod.Theinnovatorbargainssequentially withthenproducers.Thismodelcapturesinanintuitivewaytheinfluenceofthenumberofpotential buyersofthelicenseontherespectivebargainingpowers.Indeedweshowthatanincreaseinn increasesthebargainingpoweroftheinnovator. Inthecontextofthismodel,weidentifyanecessaryandsufficientconditionforalicensetobe signedinthefirstperiod.Weexaminehowthenumberofcompetitorsinfluencesthisdecision.Ifthe numberofproducersnaffectstherespectivebargainingpowers,butdoesnotimpacttheprofitsonthe downstreammarket,wefindthatanincreaseinthenumberofproducersdelayslicensing.Theintuition isthefollowing.Anincreaseinthenumberofpotentialbuyersnincreasesthebargainingpowerofthe innovatorandthepricesheobtainsinthesecondperiod.Theinnovatorthuswantstodelaysignature, whileproducerswanttosignearly.Theincentivesoftheinnovatortodelayneverthelessdominateasa consequenceofherprivateinformation.Sheknowsthequalityoftheinvention,whereastheproducers considerthepossibilitythattheinventionisbad(andgeneratesnoprofits). Whenprofitsalsodependonn,anincreaseinthenumberofproducershastwocountervailingeffects onthesecondperiodpriceforthelicense.Greatercompetitionincreasesthebargainingpowerofthe innovator,butdecreasestheprofitsderivedfromtheinnovation.Theinnovatorobtainsalargerslice ofasmallerpie.Forinitiallynonconcentratedmarkets,thesecondeffectdominatesandthesecond periodpricedecreaseswiththenumberofproducers,thusleadingtoearliersigning.Theoppositeis generallytrueforinitiallyconcentratedmarkets.WethereforefindaninvertedU-shapefortheeffect ofthenumberofcompetitorsonthedelayinlicensing. Thesetheoreticalpredictionsareconfirmedbyourempiricalanalysisonlicensesinthepharma- ceuticalindustry.Wecombinedataonlicensingdealsandtheirstageofdevelopmentatsigningwith dataonthenumberoffirmsindifferenttherapeuticclasses(firmswithdrugstreatingsimilardiseases). Controllingforvariousmeasuresoffinancialconstraintsandotherfactors,weconfirmempiricallythe invertedU-shaperelationshipbetweendelayinsigningandthenumberofcompetitors. Thoughthequestionaddressedinthispaperwasmotivatedbytheparticularapplicationtothe pharmaceuticalindustry,thetheoreticalmodelisquitegeneral.Theresultsshouldberelevantin industrieswiththefollowingcharacteristics.First,thereshouldexistsomeasymmetryofinformation betweentheinnovatorandpotentialbuyers,areasonablefeatureinmostinnovativesectors.Second, informationshouldberevealedduringthedevelopmentoftheinvention.Third,thereshouldbesome specializationintheinnovationprocess. 2 Finally,exclusivelicensingshouldbeprevalent;thistends tobethecasewhendevelopmentcostsaresignificant,sothatnofirmhastheincentivetosinkthose costswithoutexclusiverights. 3 Accordingly,webelievethattheresultscouldbeappliedmoredirectly tosectorssuchaschemicalsthantothecomputerindustry. Thepaperproceedsasfollows.Insection2,wepresentthemodelanddeterminethemainthe- oreticalresultsinsection3.Insection4,weexamineanumberofrobustnesschecks.Wetestthese resultsondataonlicensingcontractsinthepharmaceuticalindustryinsection5and6.Insection7 wecompareourpapertotheexistingliteratureandconcludeinsection8.Allproofsarepresentedin theappendix. 2 SeeAroraetal.(2001)foradiscussionofspecializationandthedivisionofinnovativelabor. 3 AccordingtoAnandandKhanna(2000),exclusivecontractsaremostcommoninchemicalsandleastcommonin computers. 3