Research and Innovation for
Smart Specialization Strategy
Paulo Guilherme Correa
The World Bank
Policy paper series
June 2016
Irem Güçeri
Oxford University Centre
for Business Taxation
Research and Innovation for
Smart Specialization Strategy
Concept, Implementation Challenges and Implications
by
Paulo Guilherme Correa1 and İrem Güçeri2
1
TheWorldBank.E-mail:[email protected].
2
UniversityofOxfordCentreforBusinessTaxation.E-mail:[email protected]
AuthorswouldliketothanktheEuropeanCommission,theInstituteforProspectiveStudies,andtheOECDfor
valuablediscussionsduringthepreparationstageofthisreport.Weareespeciallygratefulforthecommentsby
PascalBoijmans,DominiqueForay,CarloGianelle,GregoryT.Jedrzejczak,IngerMidtkandal,LuisaSanches,
FedericaSaliolaandtotheparticipantsoftheSmartInvestmentforSmartSpecializationWorkshopinWarsaw
whoprovidedinvaluableinputs.
ThisnotewaspreparedinMay2013toguidepolicymakinginlightoftheEuropeanCommission’sSmart
Specializationagenda.
1
Contents
Summary...................................................................................................................................................3
Introduction..............................................................................................................................................4
ResearchandInnovationforSmartSpecializationStrategies(RIS3):Definitions....................................7
Incompleteinformationandthepoliticaleconomyoflobbying.............................................................11
OperationalizationofRIS3:AccesstoInformation.................................................................................14
Implications.............................................................................................................................................18
Conclusions.............................................................................................................................................25
Annexes...................................................................................................................................................26
Bibliography..............................................................................................................................................32
2
Summary
Theprioritizationofselectedeconomicactivitiesthroughabroadconsultativeprocessisoftenconsidereda
centralaspectofresearchandinnovationpoliciesaimingatpromotingsmartspecialization (RIS3). This is
seen as an effort to reduce fragmentation and a way to increase the impact of research and innovation
investmentsonregionalandnationaldevelopment.Thisnotearguesthattargetingresearchandinnovation
policiesmaynotbealwaysrecommendableduetoinherentproblemsofincompleteinformationandthe
inevitablerepresentationbiastowardsincumbentinterests.Whentheeconomicspecializationofaregionis
not evident from observation of market dynamics, policies should aim at enabling the process of market
selection, allowing such specialization to emerge as a result of entry, exit and experimentation. Such an
approachwouldhavetwoimmediateimplicationsforthedevelopmentofRIS3:(i)replacingtheemphasison
ex ante definition of activities with a focus on a results-based approach for research and innovation
investments; and (ii) fully integrating monitoring and evaluating mechanisms in the development of the
strategy,allowingforpolicyexperimentation,structuredlearningandsystematicadjustmentofprograms
andpoliciestowardsthepre-definedobjectives.
3
Introduction
AbalancedregionaldevelopmenthasalwaysbeenapriorityforsocialcohesionintheEuropeanUnion(EU)
and resources were consistently made available by the European Commission (EC) to Member States’
catchingupwithleadingeconomies.ResourcesconcentratedmainlyintheEUStructuralFundsaresupposed
to help, in particular, the convergence of lagging regions within Member States. Structural funds are
disbursed within the framework of “Operational Programs” and have been used to finance innovation
policies,amongotherinvestments.Inthissense,innovationpolicieshavebeenanimportantcomponentof
theEUregionaldevelopmentstrategies3.
Researchandinnovationpoliciesalsobecameoneofthemaintoolsavailabletopromoteeconomicrecovery
andsustaingrowthinrecentyears.TheLisbonAgendaestablishedatargetforR&Dexpenditures(3%of
GDP)thatwassubsequentlyconfirmedbytheEurope2020Strategy.In2010,theEuropeanCommission
announced the availability of € 6.4 billion for research and innovation -- the largest budget ever made
available–whileurgingMemberStatestostrengthentheirpoliciesandprogramstoreducetheinnovation
gapbetweentheEUandotherworldleadingeconomies(TheEuropeanCommissionEurope2020Flagship
InitiativeInnovationUnion(InnovationUnion)--COM(2010)546)).
The adoption of a “smart specialization strategy” was one of the recommendations put forward by the
InnovationUnionflagshipinitiativetoincreasetheimpactofresearchandinnovationpoliciesofMember
Statesoneconomicgrowth.ItbuildsontheideasdevelopedinDavid,ForayandHall(2009),whichinclude
‘increasingthecorrelationbetweenR&Dandtrainingspecializationandthestructuresoftheeconomy’;and
facilitating economic “self-discovery” through a broad governance driven and fact based consultative
process.4 Member States have been therefore encouraged to define research and innovation policies for
smartspecializationstrategies(orsimplyRIS3).
TheCommissionproposedthatthesubmissionofaSmartSpecializationStrategybeanexanteconditionality
foraccesstoStructuralFundsinthe2014-20period.5ThePartnershipAgreementofeachMemberState
withtheCommission,tobesignedin2013,willdeterminethethematicobjectivesandthemonitoringand
resultsframeworktotracetheperformanceofeach.Inthiscontext,anumberofclientcountriesaswellas
theEuropeanCommission(EC)haveexpressedinterestincooperatingwiththeWorldBanktofosterthe
policy dialogue, practitioner exchange and peer networks to support the design, implementation and
evaluationofSmartSpecializationStrategies.
This paper focuses on a rather specific aspect of the smart specialization agenda, as it concentrates on
understanding research and innovation policies for smart specialization (RIS3) and discusses some key
challenges entailed to their design and implementation. This study does not attempt to provide a
comprehensiveanalysisoftheconcept,nordoesitlayoutafullsetofindicatorsthatcanbeusedtomeasure
smartspecialization.TheseissuesaretackledinanumberofdocumentsproducedbytheEC,IPTSandthe
OECDwhichcomplementthisreport.
3
SeeError!Referencesourcenotfound.forfurtherdetailontheEUCohesionPolicyinstruments.
KnowledgeEconomists’PolicyBriefno.9:“SmartSpecialisation–TheConcept”;David,Foray,&Hall,(2009);
MeasuringSmartSpecialisation:TheConceptandtheNeedforIndicators,David,Foray,&Hall,(2011);p.1.
4
5
SeeCommission’sCohesionPolicyproposal--COM(2011)615for2014-2020(Appendix7.1).
4
This paper does not address the overall set of policies possibly needed for the implementation of smart
specialization.ItreviewsaseriesofdocumentsissuedbytheECtoguideitsMemberStates6,andbuildsona
numberofrecenteffortstoinformthisprocess,especiallytwosimilarinitiativespreparedbytheJRC-IPTS
(Research and Innovation Strategies for Smart Specialization (RIS3) Guide7) and the OECD (TIP Project on
SmartSpecialization,DSTI/STP/TIP/AH(2012)1).
TheRIS3Guide,inparticular,presentsageneraldefinitionoftheconceptandoutlinessixstepstoestablish
anationalorregionalstrategy,startingwithananalysisoftheeconomicspecializationofthecountryorthe
region;continuingwiththeestablishmentofpriorityareasandoftheconsultativeprocessthroughwhich
theseprioritiesshouldbedetermined;concludingwiththesetofmonitoringandevaluationmechanisms
necessaryforimplementation.8ThispaperfullysupportstheemphasisoftheRIS3Guideongovernance,
transparencyandaccountabilityanditsproposalthatmonitoringandevaluationbefullyintegratedintothe
designandimplementationofthestrategy.
Thispaperexplorescircumstancesinwhichtargetedresearchandinnovationpoliciesmaybeuseful,namely,
wheninformationontheeconomicspecializationoftheregionorthecountryisfullyorpartiallyavailable,
through some sort of “market test”. In such cases, research and innovation policies may help sound
companies to keep a competitive edge in international markets or to cope with growing international
competition (shaping comparative advantages) or simply support potentially high growth companies to
realizetheirpotential(unleashinglatentcomparativeadvantages).
When the economic specialization of a region is not immediately evident from observation of market
dynamics, policies could aim to enable the process of market selection, allowing such specialization to
emerge as a result of entry, exit and experimentation. This paper therefore proposes that research and
innovationpoliciestargetingfewactivitiesmaynotbealwaysrecommendableinthefirstinstanceduetothe
natureofinformationavailabletopolicy-makers(incompleteandasymmetric).
Theinformationalasymmetrymentionedherereferstothenatureofthedecision-makingprocessthatis
inevitably biased towards “incumbents” (and thus against new entrants or potentially successful
entrepreneurs) despite the quality of the consultative process. As a result, policy-makers should be
encouragedtoenableinnovation(includingnewbusinessmodels)andefficientmarketselection–aprocess
throughwhichsuccessfulcompanieswilldefineanewspecializationpatternforthecountryortheregion.
Such policies then go beyond standard boundaries of research and innovation to include improving
regulationstofacilitatefirmentry;firminnovationandexports.Inanutshell,thesearepoliciesthatpromote
“high growth potential entrepreneurship”. This paper therefore underlines the importance of the
entrepreneurial environment, which includes (but is not limited to), entry and exit conditions, access to
6
Europe2020FlagshipInitiativeInnovationUnion;RegionalPolicyContributingtoSmartGrowthinEurope;EUBudget
Review;Europe2020FlagshipInitiativeDigitalAgenda;AnIntegratedIndustrialPolicyfortheGlobalisationEra;Proposal
forCohesionFunds2014-2020;StateoftheInnovationUnion2011.(COM(2010)546;COM(2010)553;COM(2010)700;
COM(2010)245;COM(2010)614;COM(2011)615;COM(2011)849)
7
MostrecentversionpublishedinMay2012.
8
TheOECDhaslaunchedadiscussionspaceforsmartspecializationwiththeEuropeanCommissionandtheIPTSalso
to elaborate on the ‘how to’ of smart specialization. This project aims to guide the design and evaluation of smart
specializationstrategiesatboththenationalandtheregionallevels,throughtheformulationofindicator-basedstrategy
profiles.
Background
documents
from
these
activities
are
mostly
available
online
https://community.oecd.org/community/smartspecialisation
5
finance (especially early stage financing), product market regulation, and access to inputs in the smart
specializationagenda.
In such cases, the role of smart specialization should be flexible and should endorse iterative learning by
emphasizingtheintegrationofmonitoringandevaluationmechanismsinthedevelopmentofthestrategy,
replacing the emphasis on targeting certain activities. This should allow for policy experimentation,
structuredlearningandsystematicadjustmentofprogramsandpoliciestowardsthepre-definedobjectives.
Thestrategymayconsistoftwosteps:first,toidentifybottlenecksandmarketfailures;then,toexperiment
withpilotinstrumentsbasedonhypotheses.Impactevaluationoftestedpoliciesshouldleadtothelearning
andadaptationprocesseswhichaddressthebottlenecksthatwereidentifiedinthefirststage.
Thispaperiscomposedoffourothersections.Thenextsectiondefinesthekeyconceptsandrelationships
drawn from the existing documents, and it identifies the relevance of targeted research and innovation
policiesandabroad-basedconsultativeprocessforRIS3.Thethirdsectiondiscussesthechallengesrelated
toincompleteinformationandconsultativeprocesses.Thefourthsectionaddressesthosechallengesatthe
operationallevel.Thefifthsectionfocusesonthepublicpolicyimplications.Thefinalsectionsummarizesthe
mainconclusions.
6
ResearchandInnovationforSmartSpecializationStrategies(RIS3):Definitions
TheRIS3Guidedefinesnationalorregionalresearchandinnovationstrategiesforsmartspecialization
(RIS3)as:
“integrated, place-based economic transformation agendas that […]: focus policy support and
investments on key national/regional priorities, challenges and needs for knowledge-based
development,includingICT-relatedmeasures;buildoneachcountry's/region’sstrengths,competitive
advantagesandpotentialforexcellence;supporttechnologicalaswellaspractice-basedinnovation
and aim to stimulate private sector investment; get stakeholders fully involved and encourage
innovation and experimentation; are evidence-based and include sound monitoring and evaluation
systems”.(RIS3Guide,p.8)
SmartSpecialization.TheideaofsmartspecializationwasputforwardasaninstrumentintheEurope2020
Agenda9.ItbuildsontheconceptsdevelopedinForayandvanArk(2007)andDavid,Foray&Hall(2009).As
a first approximation, smart specialization strategies can be understood as development strategies that,
building on existing comparative advantages, promote a larger contribution of the knowledge factor to
economicgrowth.Attheenterpriselevel,theknowledgeusedbyfirmstranslatesintobetterprocessesand
products,newbusinessmodelsorsimplyinnovations;raisingproductivity,exportsand,insomecases,more
and better jobs. In a nutshell, smart specialization strategies could be understood as knowledge-driven
growthstrategies.
Knowledgecreationanduse(ICTuse,R&Dinvestmentsandinnovation)havebeenatthecoreoftheLisbon
AgendaandtheEurope2020Strategy.In2010,theEuropeanCommissionannouncedtheavailabilityof€
6.4billionforresearchandinnovation-thelargestbudgetevermadeavailableforthispurpose-andurged
MemberStatestostrengthentheirresearchandinnovationpolicies(“InnovationUnion”10:COM(2010)546).
Inthissense,RIS3maybedefinedasanintegratedsetofmeasures(policies,programsandreforms)aiming
toincreasetheimpactofresearchandinnovationoneconomicgrowth.
Knowledgeandotherproductiveassets.Inordertoincreasetheeffectivenessofresearchandinnovation
policies, David, Foray & Hall (2011) propose the general idea that policy makers should increase ‘the
correlation between R&D and training specialization and the structures of the economy11’, where “the
structures”referstothelocaleconomicactivityindifferentsectors.Byfocusingoncertainsectors,policy
makerswouldavoidthedispersionofR&Dfundingandincreasetheprobabilityofmatchingresearchresults
with market demand. The approach is further developed in the RIS3 Guide, which proposed a specific
framework for concentrating the vast amount of resources that are allocated to R&D, innovation and
technologypoliciesinmembercountries.
Regionaldevelopmentandeconomictransformation.RIS3isexpectedtobeapplicableattheregionallevel,
asthecornerstoneofregionaldevelopmentstrategies.Theobjectiveistotriggerstructuraltransformation
in the regional economy supposedly to generate a ‘cluster of firms’ with enough spillover effects and
economiesofagglomerationtotransformtheregionfrom‘periphery’toa‘center’.AccordingtotheRIS3
Guide,R&D,innovationandtechnologypolicyshouldtargetcertainactivitieswiththepotentialtogenerate
9
COM(2010)2020.
10
EuropeanCommissionEurope2020FlagshipInitiativeInnovationUnion.
11
MeasuringSmartSpecialisation:TheConceptandtheNeedforIndicators,David,Foray,&Hall,(2011);p.1.
7
clustersoffirmsandthustheexpectedtransformationaleffectasopposedtosimplypromotingscattered
innovation.PrioritizationofsectorsthenbecomesacoreelementofRIS3.
ThefocusofresearchandinnovationpolicyonsmartspecializationisevenendorsedasacriterionforwellperforminginnovationsystemsbytheInnovationUnionFlagship(Box1).
Box1:Featuresofwellperformingnationalandregionalresearchandinnovationsystems:
“[…]Designandimplementationofresearchandinnovationpoliciesissteeredatthehighestpoliticallevelandbased
on a multi-annual strategy. Policies and instruments are targeted at exploiting current or emerging
national/regionalstrengthswithinanEUcontext(‘smartspecialisation’)
–Aneffectiveandstablecentre-of-governmentstructure,typicallysteeredbythetoppoliticallevel,definesbroad
policyorientationsonamulti-annualbasisandensuressustainedandproperlycoordinatedimplementation.This
structureisbackedupbynetworksinvolvingallrelevantstakeholders,suchasindustry,regionalandlocalauthorities,
parliamentsandcitizens,therebystimulatinganinnovationcultureandbuildingmutualtrustbetweenscienceand
society.
–Amulti-annualstrategydefinesalimitednumberofpriorities,precededbyaninternationalanalysisofstrengths
andweaknessesatnationalandregionallevelandofemergingopportunities(‘smartspecialisation’)and
marketdevelopments,andprovidesapredictablepolicyandbudgetaryframework.ThestrategydulyreflectsEU
priorities, avoiding unnecessary duplication and fragmentation of efforts, and actively seeks to exploit
opportunities for joint programming, cross-border co-operation and exploiting the leverage effects of EU
instruments.Bilateralco-operationwithnon-EUcountriesisbasedonaclearstrategyand,wherepossible,iscoordinatedwiththeotherEUMemberStates.
–Aneffectivemonitoringandreviewsystemisinplace,whichmakesfulluseofoutputindicators,international
benchmarkingandex-postevaluationtools.[…]”
Source:InnovationUnion,AnnexI:Selfassessmenttool:Featuresofwellperformingnationalandregionalresearchandinnovationsystems
Selecting(targeting)sectors.David,Foray&Hall(2009)arguethattargetactivitiesshouldbethoseresulting
from the entrepreneurial self-discovery process, as defined in Hausmann and Rodrik (2003). This is the
decision process where entrepreneurs discover the markets in which to operate from a set of “modern
sectors”.Thisconceptspecificallyrulesouttheuseoftop-downforesightexercisesandsimilarinstruments
inselectingthetarget-sectors,emphasizingtheimportanceofa“bottom-up”approach.
Definitionofa“priorityarea”.Smartspecializationaimstotargetfewpriorityareaswhichareneitherbroad
sectorsnorasinglefirm,butneworexistingactivitieswithhighmarketpotential.Theproponentsofthe
smart specialization concept support that this is one of the distinctive features of smart specialization in
comparisontotraditionalindustrialpolicy.
Theroleofbroad-basedconsultations.Thenotionofa“top-downapproach”isfurtherdevelopedinthe
RIS3Guide.Thedocumentsuggeststhatresearchandinnovationpoliciesshouldaimtoprioritizecertain
areas based on evidence, consultation and analysis. It is also highlighted that in this process, initial
stakeholderconsultationshouldinvolveallthekeyplayersininnovativesectors:publicandprivatesector
actors,academia,thecivilsociety,peernetworksandtheusersofinnovation.Researchandinnovationpolicy
shouldthustargetsectorsthatresultfromthisprocess.
OECD(2012)summarizesthedifferencesbetweentraditionalindustrialpolicyandsmartspecializationasthe
emphasisofthelatteronknowledgeandtechnology-ledgrowth,thepresenceofaconsultationprocessin
8
whichpolicymakersareopentomarketsignalsandtheuseofmarket-basedtoolssuchaspublicprivate
partnerships,surveysandSWOTanalysesbythepolicymakers12.
Thereseemstobesomeambiguityintheroleoftheself-selectionprocessandbroadbasedconsultationsin
deciding on the activities to be targeted. The process of self-discovery is sometimes presented as both a
marketandan“off-market”process,asforexampleinthefollowingstatement:“Itisofcrucialimportance
thatRIS3governancebodiesfocusonalimitednumberofinnovationandresearchprioritiesinlinewiththe
potential for smart specialization detected in the analysis phase that is anchored in entrepreneurial
discoveries”(RIS3Guide,p.22).
Reconciling sectoral growth policy with regional development and research and innovation policies.
Designingandimplementingsmartspecializationattheregionallevelisachallengingtaskduetothestrong
forces of agglomeration that arise from fundamental differences between the core and the periphery
(Krugman,1991).Regionalboundariesaremuchmoreopentoaccommodatingtheseforcesthannational
boundaries, and this constitutes a challenge for the implementation of RIS3 (McCann & Ortega-Argiles,
2011).
Relocationofentrepreneurs.Inthepresenceofbothadiversifiedcoreandaspecializedperiphery,new
firmsmayfinditmoreprofitabletolocateinthecitiesduringthelearningandsearchprocessfortheirideal
domain(Duranton&Puga,2004).Althoughcostly,thismayallowfirmstoeasilyemploythedifferentkinds
ofinputsneededwhenanexperimenteddomainfailsandanewoneneedstobetried.Suchrelocation
maycurbtheemergenceofnewactivitiesintheperiphery.
Relocationofthelaborforcewhichisskilledindifferentdomainsthantheoneprioritizedintheregion.
Suchlaborwillbemoregeographicallymobile,andislikelytorelocatetothediversifiedcitiesortoother
regionswithdemandfortheirexistingskills.Generationoflocalhumancapitalfitforthelocal
specializationwillbecostlyandmayrequiretimetobuildup.Localtrainingprogramswhichreinforcethe
formationofgeneralaswellasspecializedskillsmayfurtherinducetheemigrationoflocally-produced
humancapital(McCann&Ortega-Argilés,2011).Thequestionhereiswhethercountriesarereadyto
accommodatetheseeffects.
EmphasisonICT.AnothercomponentoftheRIS3agendaistoimplementpolicieswhichaimtoincreasethe
ICTuse.Sincemid-1990s,theEUhasremainedinacatching-uppositionlaggingbehindtheUSandJapan,
especiallyinthefieldofinvestmentsininnovationandtechnologywhileChinaandtherestofthedeveloping
world have been progressing fast. One of the eleven thematic objectives of the 2014-2020 programming
periodisthereforeidentifiedbytheCommissionas“enhancingaccesstoanduseandqualityofinformation
andcommunicationtechnologies”13.
SmartSpecializationStrategyasan“exanteconditionality”.TheCommissionproposedtheacceptanceof
aSmartSpecializationStrategyasanexanteconditionalityforaccesstoStructuralFundsinthe2014-2020
12
“DraftSynthesisReportonInnovationDrivenGrowthinRegions”(2012)
13
SeeAnnexII.
9
period.14ThePartnershipAgreementofeachMemberStatewiththeCommission,tobesignedin2013,will
determinethethematicobjectivesandthemonitoringandresultsframeworktotracetheperformanceof
each.ANNEXIIpresentstheexanteconditionalitiesfortheuseofStructuralFundsin2014-2020includinga
roadmapforincreasingtheICTuseinMemberStates.
Theproposedapproachtoregionaldevelopmentisjustifiable.Inter-regionalincomedisparitiesacrossEU
regionshavenotsubstantiallynarroweddespitelargeamountsallocatedthroughpreviousregionalpolicies
(Puga,2002).15ThemaincauseforregionaldisparitiesbetweenthecoreandtheperipheryintheEUrelate
to agglomeration effects, strengthened by rapid innovation and technological advancement in the core
(Farole, Rodriguez-Pose, & Storper, 2009). The transformational effect aims to address precisely this
structuralproblemand,ideally,targetedresearchandinnovationpolicieswouldpromoteit.Yetthisnote
willarguethatitmaybenotalwaysfeasible.
14
SeeCommission’sCohesionPolicyproposal--COM(2011)615for2014-2020(Appendix7.1).
15
Bysomemeasures,distributionofpercapitaincomeacrossregionsintheEUismoreunequalthanacrosstheUS
states,contrarytowhatonewouldhaveexpectedgiventhehigherspatialconcentrationofproductionintheUS(Puga,
2002).
10
Incompleteinformationandthepoliticaleconomyoflobbying
Inthissectionwearguethat,duetoincompleteandasymmetricinformation,apublicsectorgovernance
structurethatdiscouragesefficientriskmanagementandthecollectivedecisionmakingprocessesthatare
inevitably biased towards incumbents’ interests may hinder the government’s capacity to properly select
sectorsorproductsthatmayinduceaneconomictransformation.
Competitive rent-seeking, a concept whose roots date back to the 1970s when targeted policies were
widespread (Krueger, 1974; 1997), explains the incumbent bias arising from the asymmetric information
problem.Despitetherigorintheconsultationprocessbetweenthegovernmentandtheincumbents,the
information resources available to the policy maker are limited. This is partly due to a demand side
coordinationproblem:inatypicalsetting,thereexistaverylargenumberofconsumers,whoconsumea
relativelysmallamounteach,andafewproducerswhoservealltheseconsumers.Thereisastrongincentive
for the producers to lobby, disclose only limited information and compete for the scarce government
resourcestoreducetheircostsandstayinthemarket.Ontheotherhand,itismuchmoredifficultforthe
largenumberofconsumerscoordinateandlobbytoenableademand-drivenspecializationstrategy.
The capacity to obtain and process information related to supply and demand conditions is a classical
problemofeconomicpolicymaking.Information-relatedproblemsarelargerwhentheproductsorservices
willonlyexistinthefuture–asinthecaseoftargetedinnovationpolicy.Thetaskbecomesoverlydifficult
whenthoseproductsneedtogeneratea“transformational”effect–asexpectedinthepresentcontext.The
trickier part here is identifying the sufficient conditions – for example, relevant economies of scale, and
spillovereffectsbasedonexistingcomparativeadvantages.Inthe1990s,thedifficultytoidentifysectors
withtheattributesthatwouldrendertargetedsupportjustifiablewasoneofthereasonswhythe“strategic
tradepolicies”didnotdevelopmuchasapracticaloption.
Theproblemofcollectingandprocessinginformationaboutsupplyanddemand,evenwhenthemarket
conditionsareknown,isaclassicalchallengeforeconomicplanning.Thebestknowninstrumenttocollect
andprocessinformationaboutcostandpreferences,supplyanddemand,isthemarket-despitethewellestablished case of market failures. In the case of innovations or entrepreneurial activities that could
potentiallyleadtoatransformationintheeconomybygeneratingaclusterofspecialization,theproblemis
evenmoresevere.Thereasonisbecausetheseactivities,bydefinition,sufferfromtheinexistenceofaprice
and,moregenerally,frommissingmarkets.Sincetheproductisexanteunknown,themarketcannotbe
createdeither.
Collectivedecisionmaking,evenwhenbasedonabroadconsultativeprocess,doesnothelpmuchinthis
context for two reasons. The first is asymmetric information. Self-motivated sector or project advocates
naturallypossessmoreinformationabouttheirownsectororprojectthanpolicy-makers.Thequalityand
relevanceofinformationprovidedorgeneratedinthiscontexthasinevitablylimitedusefordecisionmaking.
Thesecondreasonisthatrepresentationinthecollectivedecision-makingprocesswillbebiasedtowards
incumbentpoliticalandeconomicinterests.Incumbentshavemoreincentivetovoiceorlobby(moretowin,
clear goals, and defined coalition) as opposed to new entrants, particularly future entrepreneurs, simply
becausetheseentrepreneursdonotexistatthetimeoftargetedpolicymakingandhencehavenolobbying
power.
The critical factors are the inexistence of market generated information and any alternative source of
reliableinformationfordecisionmaking.ThisispreciselythereasonwhyHausmannandRodrik(2003)argue
in favor of subsidizing entry into new markets: the information generated by the first-comer is a non11
proprietarygoodandcanbequicklyutilizedbyfollowers(imitators).Thisinturnreducessignificantlythe
returns to investments in the “discovery” process, which makes private investment levels inferior to the
sociallyoptimal(asinthecaseofR&Dinvestmentsforexample).16
When reliable market information is not available, should governments bet public money in targeting
specificsectors?Oneargumentoftenheardisthatprivatesectorinvestors,includingventurecapitalistsand
serial entrepreneurs also have incomplete information, face high uncertainty and still profit from those
investments.Couldthepublicsectornotachievethesameresults?Thisisveryunlikelygiventhedifferent
incentiveregimesunderwhichthepublicandtheprivatesectorsoperate.
Privateinvestorsinvesttheirownresources(ortheyaretheagentsofindirectinvestors),profitfromsuccess
andbearthefinanciallossesassociatedtofailure,facinganincentivestructurethatismoreconduciveto
better risk management and thus entrepreneurial success. For instance, private investors are quick in
selectingaprojectandbringingthefinancialandnon-financialresources(suchasmentoringandnetworking)
requiredforitsdevelopment(toexitandthuscashintheexpectedreturns).Yet,theyarealsofastinassessing
theoddsofprojectsuccessandareeagertoexitassoonasfailureisperceivedasthelikelyoutcomesothat
investors’lossesarecontained.
Publicservantsinvesttaxpayers’money,theydonotprofitfromitssuccess,anddonotbearthecostsof
baddecisions.Underthesecircumstances,favoritismbecomesmorefrequentthanusual.Insteadofaquick
exit,supporttounlikelysuccessfuldealswillbeunnecessarilyextended.Indeed,asarguedmorebroadlyby
Rodrik(2004),identifyingcredibleexitstrategiesisoneofthecriticalchallengesfortheindustrialpolicyof
the21stCentury.
TheEUhastriedtoaddressthechallengesabovebefore,asforexample,inthecaseofstateaidregulations
forrestructuringsectors.Inordertoreceivesupporttorestructuringactivities17,firmsorcountriesareasked
toprovideadetailedplanwithmeasurableandmonitorablemilestonesdemonstratinglongtermfinancial
sustainabilityoncethesupporthasended.Anamountofsubsidyisdefinedexante,andrespondstothe
criteria of minimum necessary. Based on these, a clear time frame is set with a defined date for the
suspensionofthesubsidy,afterwhichthefirmisnotallowedtousesimilarsupportforalongperiod.TheUS
approachtothefirmsintheautomotivesectorisanotherexampleofdirectsupport(withaverydifferent
governancestructure).Inmostcases,however,suchapproachesaretheexceptionandnottheruleofpolicymaking.
Therearecircumstanceswhenthelackofinformationposesnoseriousrisktopolicymakers.Weargue
that,dependingontheamountofmarketgeneratedinformation,itmaynotbedesirabletotargetR&Dand
innovation policies to specific sectors. The argument does not exclude the possibility of targeting sectors
whenmarketdriveninformationissufficientlyavailable.Inthesecases,focusingR&Dpoliciesinsectorsin
whichtheeconomyisspecializedandgloballycompetitiveispossibleandmayhelptoincreasetheimpactof
R&Dandinnovationinvestments.AschematicviewispresentedinFigure1.Inthenextsectionwedevelop
atentativetypologybasedontheexistenceofevidenceabouttheeconomicspecializationofaregion.
16
HausmannandRodrik(2003)justifypolicyinterventioninthecasewhereentrepreneursaresearchingfortheirideal
domainofoperationunderuncertaintyaboutcosts,whichdependonsomeunobservedproductivityparameterfor
eachgoodinthe‘modern’sector(p.608).Theysolveforthegeneralequilibriuminvestmentinthemodernsector,and
showthatthedecentralizedequilibriumresultsinunderinvestment,coupledwithtoohighproductdiversification.
17
Anexampleistheshipbuildingsectorinthe1980s.
12
Risk level for policy makers
Figure1.Accesstoinformation,risklevelandpolicymaking
HIGH
Unleash latent
comparative
advantage
MED
Modernization
LOW
NO INFO
Discovery
through
experimentation
Enabling
specialization
PARTIAL INFO
FULL INFO
13
OperationalizationofRIS3:AccesstoInformation
IndevelopingRIS3,policy-makerswillhavetoaddressthetypeofeconomicspecializationoftheregion
and its determinants or binding constraints. Based on the nature of the economic specialization and its
underlyingcausesor“bindingconstraints”,policymakerscandecidewhetherresearchandinnovationpolicy
isarelevantinstrumenttofosterdevelopmentandifsowhethertargetedinstrumentsarerecommendable
ornot.Thereislittlethatresearchandinnovationpoliciescandotohelpregionsovercomeproblemsrelated
to“distance”orlackofeconomicresourcesorpoorinstitutionalframeworkbutthereisprobablyalarge
contributioninhelpingcompaniesmovetheirproductsuponthequalityladder.Inthissectionweassume
researchandinnovationarepotentiallyrelevanttofocusonissuesofsectortargeting.
In what follows, we present a simplified typology relating availability of information on the economic
specializationofaregiontothedegreeofconvenienceofadoptingtargetedpolicies.Wearguedinthe
previoussectionthattheabsenceofreliableinformationaboutsupplyanddemandisoneofthereasons
whythegovernmentoftenavoidstargetingsectorsandproducts.Thelevelofinformationavailableonthe
specializationofthecountryortheregionisacrucialfactorforthefuturesuccessoftargetedpolicies.We
proposethreecasesinwhichdifferentdegreesofinformationabouteconomicspecializationimplydifferent
chances of success with sector targeting and provide some practical examples: regions with apparent
comparativeadvantages;regionswithlatentcomparativeadvantageorregionswithunclearspecialization.
TheargumentispicturedinFigure1.
Regions with apparent comparative advantage. These are the regions where a few globally competitive
industriesarealreadyinstalled.Forsuchregions,thekeyindicatorsandconsultativeprocessesshouldlead
tosimilarconclusionswithregardtotheregion’scomparativeadvantages.Yetregionsofthistypemaybe
experiencing growth or decline and, consequently, the pattern of specialization may or may not be
sustainable. Targeted R&D and innovation policies will “complement existing productive assets”, helping
firmstomaintainacompetitiveedgeinthesectorbyinvestinginR&Dorregaincompetitiveadvantagelost
tonewplayersintheglobalmarket.TheRIS3guidepresentsthecaseoftheFinnishpulpandpaperindustry,
whichisincreasingitsR&Deffortsinnanotechnologytoimprovetheefficiencyandqualityofproduction.
SpanishCITES,publicprivatelaboratoriesprovidinganumberoftechnologyrelatedservicestotheprivate
sector,wereinstrumentalinimprovingproductdesignandqualityinthelocalshoeindustrytherebyhelping
thesectortorefocusexportstohighendmarkets(particularlyJapan).
Box2:HowpublicprivatepartnershipscanhelptoavoidinadequateinvestmentsinR&Dfacilities:Thecaseof
CITES
HowcanpublicprivatepartnershipshelptoavoidinadequateinvestmentsinR&Dfacilities?Onewayofassessing
theprivatesector’sdemandforknowledgeservicesinaspecificdomainisthrough“public-privatepartnerships
(PPP)”,whichallowthepolicymakertoco-investinhigh-returnprojects.AnexampleofthiskindofsetupisSpain’s
programforTechnologyCenters(CITEs).Thesecentersareco-financedbytheprivatesectorandthegovernment,
inanarrangementwherethegovernmentfinancingisenvisagedtophaseoutgradually.Theyhaveavery
specializedsectororevenproductfocus.CITEscarryoutthefollowingfunctions:
‘(i)facilitatethetransferofknowledgeandexistingtechnologies(off-the-shelf)toenterprises;(ii)addressmissing
linksinsectoralvaluechainsandqualityissues;(iii)identifybottlenecksandopportunitiesforfurtherinnovationof
14
productsandprocessesatthesectorallevel;(iv)facilitatethecommercializationofnewproducts;(v)providevalue
addedservicesnotreasonablyavailable;and(vii)providespecializedtraining.’
Thesefunctionsrespondtothetechnologicalneeds,aswellasskillformationoradaptation,forthespecificareaof
specializationintheregion.
Morebroadly,inthistypeofregions,thekeyistofosterbackwardandforwardlinkagesandproductdifferentiation
tomoveupinthevaluechainbyexploitingexistingscientificpotential,fosteringacademicentrepreneurship,
researchcommercializationandcollaborationbetweenpublicandprivatesectors.
Regions with latent comparative advantage. In this second category of regions, there is no substantial
economic activity in the potential area of specialization, but the required knowledge ‘partially’ exists
typically, but not only, due to (i) availability of a non-tradable, location-specific input, such as a natural
resource,oranimmovableasset(landandclimateforexample);or(ii)localcommonknowledgeaboutthe
economic activity, a tradition prevalent in the region that indicate potential for specialization. Again,
assumingthatthebottleneckisrelatedtolackofknowledgewithlocalcontent,R&Dandinnovationpolicies
(andprobablyinvestmentsinskills-formationandotherbusinessdevelopmentservices)maybeusefulto
‘unleash’existingcomparativeadvantages.ThecasesofthesoybeansindustryinBrazilandofwineinChile
aregoodillustrationsofhowtechnologyorientedpoliciesmayenablethedevelopmentofentiresectors,
triggeringstructuralchange.
Box3:ThecaseofsoybeanexportsinBrazil
Tradeorientation,entrepreneurshipandtechnologypolicywerethekeyfactorsinBrazil’stransformationfroma
netimporterintotheworld’ssecond-largestexporterofsoybeansinthespanofthreedecades.
Intheearly1970s,theBraziliangovernmentencouragedthecultivationofsoybeansintheSouthernregionofthe
country,whereatemperateclimateandfertilesoilseemedtoprovideadequateconditions.Bythemid-1980s,
however,fertilelandinthatareahadbecomescarceandrentalpriceshadrisen.Agriculturalentrepreneursbegan
lookingforalternativesinthethen-cheapandvirtuallyunexplored“Cerrado”–theBrazilianprairies–anarea
aboutthesizeofFrancebutwithcompletelydifferentsoilandclimaticconditions.Atthattime,Embrapa,a
governmentagriculturalresearchinstitute,wasinstrumentalinrebalancingsoilacidityandcultivatingcrops
suitedtothecountry’stropicalclimate,therebyexpandingtheareaeffectivelyavailableforcultivationof
soybeans.Embrapanowgrowsmorethan200varietiesofsoybeanstosuitthecountry’sdiversesoilandclimatic
conditions.
Highagriculturalproductivityratesandtheeffectiveutilizationoflandenabledentrepreneurstoexplorenew
exportopportunities.Accesstobetterandcheaperagriculturalinputs,stemmingfrombroadtradeliberalization,
werealsopivotalinraisingagriculturalproductivity,andsoybeanexportsinparticular,afterthe1990s.
Regions with unclear specializations. When there is no obvious local asset that induces an economic
specializationinaregion,informationfromdirectobservationofmarketdynamicsisnotavailableandsector
targeting becomes less recommendable for the reasons discussed in the previous section. Under these
circumstances,werecommendthatpolicy-makersfocusoncreatinganenablingenvironmentforefficient
15
marketselection(fromwhichaneconomicspecializationshouldemergeovertheyears,asaresult).This
meanscombiningmeasuresthatpromotefirmentryandstartups–possiblyhighgrowthpotentialfirms–
and allowing firm exit. Research and innovation policies play a central role in promoting entry but other
policiessuchasfacilitatingaccesstocredit,skillsandinformation;andimprovingthebusinessenvironment
(such as adopting pro-competition regulation in the service sectors) will be equally important. When
economic specialization is not evident, promoting entrepreneurship across the board seems to be the
dominantstrategyforpolicymakers.
Thechallenge,therefore,istodefinetherightpolicymixandtheprogramsthatbetterfittheeconomic
and institutional circumstances of each region or country within a given pre-defined objective. Yet
predefinedentrepreneurshippoliciesdonotexistassuch;rather,existingeconomicpoliciesconstitute(or
not)aneco-systemthatismoreorlessfavorabletoentrepreneurship(Acs,2012).Overall,aneco-system
thatpromotesentrepreneurshipshouldallowfortheidentificationofbusinessopportunities,providethe
right incentives and access to inputs required for their development by entrepreneurs. For instance, the
inadequateregulationofthetelecomorlogisticssectorsmayprecludethedevelopmentofe-commerce;in
othercases,thelackofearlystagefinancingmayhinderthedevelopmentofscience-basedstartups.More
broadly,countriesandregionsvaryaccordingtotheirentrepreneurialstrengthsandweaknesses.
Figure2.Understandingeconomicspecialization
Inordertoidentifythetypeofaregion,theavailabilityofinformationcanbeassessedduringthedesignof
itssmartspecializationstrategy.Analyticalexercisesthataimtouncoverlocalcomparativeadvantagesmay
provideconsistentandreliablesignalstothepolicymaker,revealingtothepolicymakertherisksoftargeting
16
specific activities, and if such risks are low, the areas of specialization which may be considered18. The
emphasishereisontheoppositescenario,wherethepolicymakerdoesnotobtainreliableinformationfrom
the analytical exercise and should draw the strategy on uncovering the latent comparative advantages
throughenablingpolicies,whicharebrieflydiscussedinthenextsection.
18
Diagnosticactivitiesmayinclude,butarenotlimitedto,analysesoftradecompetitiveness,transport
infrastructureandconnectivitywithmarkets,economicgeographyaspects,productivityandresearchand
technologicaldevelopmentinfrastructure.SomegeneraltoolsforthispurposeareoutlinedinAnnexI.
17
Implications
Exantetargetingismorelikelytoachieveitsdesiredgoalsiftheregion’scomparativeadvantageisknown
to the policy maker. This is the case for regions with apparent or latent comparative advantage, where
informationisavailableaboutthespecializationsoftheregionandtheriskofmoralhazardandopportunistic
lobbyingislow.Intheseregions,specializationcanbewelfare-improvingasitfacilitatestheefficientuseof
theavailableresources.
On the other hand, peripheral regions with less clear emerging or existing trends need their future
specializations to be revealed through a flexible strategy which encompasses enabling policies for
entrepreneurship and market selection, rather than ex ante targeting. While forces of agglomeration
stimulatethecreationofadiversifiedcoreandaratherspecializedperiphery,thechallengefortheperiphery
istocreateasuitableenvironmentforentrepreneurship.Thisisdifficultbecauseentrepreneurshipismore
likely to flourish in regions where: (i) a diverse set of sectors exists, (ii) many multinational companies
operate,(iii)marketsarecompetitive,(iv)populationisdense,and(v)marketpotentialishigh(McCann&
Ortega-Argiles,2011).
Inthesetypicallylessdevelopedregions,theprimegoalsofthepolicymakermayincludefacilitatingfirms’
accesstoinformation,improvingmarketentryandexitconditions,buildingtheinfrastructureforinnovation
financing and helping the buildup of knowledge assets (such as a large research university with
commercializationpotential).Researchandinnovationpolicymayfocusoncommercializationofresearch
throughuniversity-industrycollaboration,researchstartupsandspinoffs,improvementoftheintellectual
propertyregime,developmentoftheearlystagefinancinginfrastructure,strengtheningknowledgenetworks
andfacilitatingspillovers.
Achieving structural transformation of the region is a challenge that extends beyond research and
innovation policies. Trade policy, technology absorption and adoption, business environment and the
regulations that govern the human capital input are directly related to the entrepreneurial environment.
Policysettingintheseareasnaturallyinvolvesamorecomplexinstitutionalframeworkthantheboundaries
ofresearchandinnovationpolicy.Tobemorespecific,theentrepreneurialenvironmentisshapedbyawhole
host of policies including labor market regulation, education policy (strong primary, secondary, tertiary,
vocational and higher education); health policy (which helps provide security for the entrepreneur); the
regulatoryenvironment;businesstaxationandthejudicialsystem(Acs,2012).Rigidbusinessregulationmay
deteriorateentryconditions,hinderingtheentrepreneurialcapacityoftheregion.
“Self discovery” is heavily dependent on entry conditions, with fewer discoveries occurring in systems
burdened by barriers to entry. Klinger and Lederman (2011) document the strong correlation between
governmentregulationandselfdiscovery.Thehighertheregulatoryburden,thefewerexportdiscoveries
aremade,despitethethreatofallowingimitatorsinthenewlydiscoveredsectors.Infact,imitationtendsto
increasewelfareinmanynewsectorsasitincreasesthescopeforsocialreturnsfromtheseactivities.The
government, which has an incentive to support such discoveries should then consider addressing entry
barriers.
Is knowledge the binding constraint against the structural transformation of the region? This is a key
questionthatneedstobeansweredbytheregionwhiledesigningitssmartspecializationroadmap.Ifthe
bindingconstraintrelatestoamorefundamentalstructuralbottleneck,suchastheregulatoryenvironment
or physical infrastructure, then the region should prioritize addressing such constraint, while taking into
considerationthenextstepsinresearchandinnovationinitsmediumorlongtermagenda.Forinstanceif
18
thebusinessregulatoryframeworkforestablishingacompanyisheavilycomplicatedbyredtape,thentaking
measurestogenerateuniversityspinoffsmayprovetobeinsufficientinstimulatingthedesiredstructural
transformationofthelocaleconomy.
Identificationofregionalbottlenecksagainststructuraltransformation.Havingestablishedwhetherornot
knowledge is a binding constraint, the strategy should then sketch more specifically what the binding
constraintsare,andwhichonestoaddressintheshort,mediumandlongterms.Knowledgebottleneck(s)
mayrelatetoanumberofinstitutionalfactors,mainlyarisinginoneormoreofthefollowingaspects:(i)
constraints to increasing private R&D and innovation investment (access to finance, intellectual property
protection,incentiveschemes),(ii)constraintstoimprovingtechnologyadoptionbySMEs(skillmismatches,
easeofaccesstotechnology,awareness),(iii)constraintstoresearchexcellence(skilledlabor,partnerships).
Therightpolicymixisnotnecessarilyrelatedtoresearchandinnovationpolicy,butitisonethatenables
the“selfdiscovery”process.Thenatureofentrepreneurialactivityistoexperimentwithnewproductniches,
discoveringthecostoftheseactivitiesandabandoningthosethatcannotbeundertakenatasufficientlylow
costtoproveprofitable.AsRodrik(2004)elaborates,thediscoveryofsuchcostdoesnotnecessarilyarise
frominventingaproductorprocessthatisnewtotheworld,althoughitmaywellbe.Suchdiscoverymay
evenarisefromadoptingatechnologytoproduceatraditionalproductorundertakeawell-knownprocess
atalowercost.Anotherrouteistoadapttheproductorprocesstolocalconditionsinamoreefficientway
thanthecompetition.
Thesmartspecializationstrategycanbedesignedasalongtermplantoinstallandcontinuallyreviewthe
policyframeworkthroughintegratedimpactevaluationanditerativelearning.Thislongtermstrategyis
envisagedto:(i)determinetheoverallobjectiveandrelevantmeasurablegoals,(ii)identifybottlenecksand
marketfailures,and(iii)experiment,learnfromtheseexperiments,andthenadaptthepoliciesaccordingly.
The“identificationstage”consistsofthetasks(i)and(ii),afterwhichthe“experimentationandadaptation
stage”follows.Eachstagecanbeguidedbyachecklistofquestionsaslistedlater.
Importantly,thestrategyidentifieschannelsthroughwhichmeasurabledevelopmentobjectivescanbe
achievedviaflexiblepolicyinterventions.Themonitoringandevaluationframeworkshouldthereforebe
embeddedinallstagesofthesmartspecializationstrategy.Progresstowardsthesemeasurabledevelopment
objectives can be monitored through intermediate outcomes and influenced by means of flexible
interventions.Inotherwords,thepolicymakerworksbackwardsfromthedesiredfinaloutcomes,identifying
themainchannelsthroughwhichtheseoutcomescouldbeachievedviaproposedinterventions.
Understandingtheincentivesforlobbyingforallkeystakeholders,bothpublicandprivate,mayhelpto
mitigate(butnotcompletelyeliminate)someissuesconcerningthepoliticaleconomyoflobbying.Several
privateandpublicactors,ordifferentlevelsofgovernmentmaycompeteorcollaboratewhiledefendingthe
interestsofalargesetofrelatedparties.Inregionswherelimitedornoinformationisavailabletothepolicy
maker,theproblemisparticularlysevereandmoredifficulttoalleviatebysuchexamination.Itwouldbe
expectedintheseregionsthatthebenefitstorelatedpartiesarelessobvious,andhencemoredifficultto
analyze.
19
Thefollowingchecklistprovidesasetofquestionsthatmayguidethecharacterizationofaregionpriorto
designingasmartspecializationstrategy.
Fully integrating monitoring and evaluation to policies and programs is essential to allow systematic
learningandimprovement.Ratherthanexantetargetingcertainareas,researchandinnovationpolicies
couldfocusonidentifyingtheintermediategoals,thechannelsthroughwhichthesegoalsaremeanttobe
achieved,andthebottlenecks–marketandinstitutionalfailures–thatreducethechancesofsuccess.Public
actions would then tentatively aim at correcting such failures; be monitored and evaluated, as well as
adjustedbasedonlessonsacquiredthroughsystematiclearning.Recognizingthattherearenopre-defined
policy recipes when economic specialization is not immediately evident implies raising the relevance of
results-orientation as well as the integration of proper monitoring and evaluation mechanisms into
policy/programimplementation.
In order to raise the impact of research and innovation on regional and national development,
governments could focus their interventions around a small number of measurable intermediate
objectiveswithacleardirectlinkwiththebroaderdevelopmentgoalofhighergrowthandjobcreation,as
forinstance:
Acceleratingthecommercializationofpublicresearchandacademicentrepreneurshipthroughlicensingand
spinoffcompanies;
Increasing the collaboration between public research organizations with the local private sector through
organicresearchcollaboration;
PromotingbusinessexpenditureinR&Dandinnovation,includingnewbusinessmodels,andthenumberof
firmsengagedininnovationactivities;
Reformingthesystemofmanagingpublicresearchorganizationstofavortalentedresearchandperformance
(excellence and productivity), as well as skills formation, all underlying factors for successful smart
specialization.
Enabling faster adoption of updated technologies and organization processes (especially general purpose
technologiesbysmallandmediumsizedcompanies,ICTbytheservicesector).
20
Table1.Identificationstage,examples
Goal
Bottleneck
Instruments
Measureofsuccess
IncentivizingR&D
expenditure
Financing
Alternativecreditlines
Incentives
Matchinggrants
#firmsengagedin
innovation;proportionof
spendingonR&D
AccelerateR&D
commercialization
Businessskills(academia)
Entrepreneurialtraining
EarlyStageFinancing
Angelinvestors
Improvetechnology
adoptionbySMEs
Awareness
Informationcampaigns
Skills
TrainingPrograms
Retentionofskilledlabor
force
Non-financialincentive
schemes
Coordination/
collaboration
Workshops/international
conferences
ResearchExcellence
Licensingofspinoff
companies;patents
Takeupratesoftargeted
technologies
Academicqualifications;
Qualityofpublication
output;Publication
collaborations
Selectedresearchandinnovationpolicies,programsandinvestmentswouldthereforedirectlycontribute
tooneofthepreviouslyselectedintermediategoalsandbeevaluatedaccordingly.Theywouldaddress
identifiedmarketandinstitutionalfailures,andbedesignedandimplementedinawayconsistentwithrobust
evaluation,thusinformingsubsequentadjustmenttowardsahigherimpact.Suchapproachwouldhelp,in
particular,torationalizetherelevanceofinvestmentsinresearchinfrastructurethattendtoencompassa
largeshareofresearchandinnovationinvestmentsfinancedbythestructuralfunds.
Inthecaseofcommercializationofpublicresearch,forinstance,regionscouldexperimentwithprograms
addressing the supply of ideas originating from public research organizations -- from supporting better
intellectualpropertymanagementpracticesbytechnologytransferofficestotheprovisionofsmallgrants
forthedevelopmentofproofofconcepts;and/orwithprogramspromotingthedemandforsuchideas,such
as supporting angel investments (for example, accelerator programs), according to their specific
circumstances.
InthecaseofadoptionofICT,thelowerusagelevelsofEuropeanfirmsascomparedtotheirUSpeersderives
from relatively lower returns from ICT investments in Europe due to the adoption of less efficient
management (organizational) practices and higher product market regulation – which prevent local firms
fromreapingthesamelevelofbenefitsastheirUScounterparts.19Hence,thereisnoobviousindicationof
whetherandwhichtypeofpoliciescouldhelpincreasetheproductivityofICTandincreaseitsuseinEurope.
Experimentation may also be related to the design of support mechanism. For instance, in promoting
collaborationbetweenpublicresearchorganizationsandtheprivatesector,oneoftheimportantvariables
maybethetargetedpopulation.Undercertainconditions,whicharedifficulttoidentifyupfront,eliminating
19
(Bloom,Sadun,&vanReenen,2012).TheEurope2020FlagshipInitiativeInnovationUnion(COM(2010)546)argues
thatanotherreasonforthegapisthelowlevelofinvestmentinICTequipmentwhichremainbelowthethresholdICT
thatcouldpotentiallyallowtheprivatesectortoreapthebenefitsofICTinvestments;andmodestprogressininnovative
activitybyEuropeanbusinessesandinadequateresearchinfrastructurewhichunleashtheproductivitygainsfromICT
investment.IftheunderlyingreasonisthelevelofICTinvestmentthatremainsbelowapotentialefficiencythreshold,
thenthereinforcementsuggestedbytheEUcouldhelp
21
thepossibilityoftheprivatesectorsponsoringprojectapplicationsmayresultinalargeshareofthefunds
beingunused(especiallywhenotherstandardsourcesofpublicfinancingforpublicresearcharegenerously
available).
Table2.ExperimentationandAdaptationStage
Question/Goal
Target group
•Selection/picking
•R&D promotion
•Commercialization
•…
Identify market failure
•Entrepreneurs
•Researchers /
academics
•Public/private
Institutions
•Venture capitalists
•Angel investors
•Banks
•…
Market failure critical
assessment
•Lack of credit
•Poor property rights
•Coordination failure
•Labor market
mismatch/skills
shortage
•Lack of insurance
market
•…
Intervention design
•Eg. lack of credit:
•Why?
•Intangible collateral
•High (uninsured) risk
•Asymmetric
information
•…
•Matching grant
•Tax break
•Scholarships
•Investment climate
reforms
•Networking events
•Academic incentives
(financial and nonfinancial)
•Guaranteed loans
Experimentation and
iteration
•Within and across
interventions (eg.
combinations vs.
variations)
•Eg. Variations in costsharing agreements;
criteria for credit,
alternative insurance
schemes, various
information
campaigns…
The assessment of innovation policies is still in its initial stages, especially in new member countries and
despitethesignificantemphasisplacedbytheEC.Yet,motivatedbytheLisbonAgendaandmorerecently
theEurope2020Strategy,newmembercountriesareencouragedtocommitmoreresourcestoresearchand
innovation (for example, 1.5 and 2 percent of GDP in Bulgaria and Romania respectively). The proposed
results-basedframeworkcouldhelpgovernmentsbetterdeployadditionalresourcestopromoteresearch,
developmentandinnovation.Itmayalsohelpgovernmenttoestablishaclearstrategytoachieverelated
goalsestablishedundertheirinnovationstrategies(suchas1percentofGDPinvestedbytheprivatesector,
asinthecaseofRomania).Finally,itwillhelpgovernmentsofNewMemberStatestoidentifyremaining
areasofcriticalinstitutionalreformstocompletethetransitionoftheirinnovationsystems.
22
SmartSpecializationProposedChecklist
-IdentificationStage-
Isthereanyobviouscomparativeadvantageoftheregionbasedonexistingeconomicactivity?
Istherealatentcomparativeadvantagebasedonanimmovableassetthatcouldbeasourceofwealth
creation?
Isknowledgeabindingconstrainthinderingthestructuraltransformationoftheregion?
Ifknowledgeisabindingconstraint:
Willresearchandinnovationpolicyhelpinaddressingthesebottlenecks?
Doesthebottleneckrelatetoresearchanddevelopmentinfrastructure?
Doesthebottleneckrelatetotechnologyadoptionandabsorption?
Doesthebottleneckrelatetotheentrepreneurshipenvironment?
Whatarethealternativesfortheappropriatepolicymixtoaddresstheidentifiedmarketfailure?
Ifthereareotherfundamentalbindingconstraintsthatrelatetothesuccessofacluster20inachieving
structuraltransformation,aretheserelatedto:
Inadequateinfrastructureservicedelivery?
Financialmarketimperfectionsthatlimitfinancingtofirms?
Weakstatecapacity(adverseregulatoryframework/poorimplementationofregulations)?
Weakprivateinvestorcapacity(SMEs:inadequatetechnicalormarketknowledge)?
Statecoordinationfailure(differentministries,municipalities,etc)?
Privatecoordinationfailure(firstmovercosts)?
Availabilityofland,landtenuresystemsandmechanismstoaccessland,pricingmechanisms,etc.
(inallagribusinessgrowthpolestheissueoflandandthedeterminantsofaccessingitshouldbe
givenacentralrole)?
Lackofcriticalmass?
Inadequatelearningmechanisms?
Which are the areas of high risk? Identify specific, potential sources of risks based on existing data.
Externalrisks(suchasmonetaryshocks,economicrecession,pricefall),countryrisks(suchaspolitical
instability,landtenure/propertyrights),andregionalrisks(suchasdrought,naturaldisasters)directly
affect the likelihood of success. Some of these could be addressed through mitigation actions and
integratedintotheprojectdesign.
Howisthecausalchainbetweenactions,intermediateoutcomesandfinaloutcomescharacterized?
Whatisthedevelopmentobjectiveofthenationalsmartspecializationstrategy?Inwhatways
doesthatconcerntheregion?
Whatisthedevelopmentobjectiveoftheregionalsmartspecializationstrategy?
Howcanthedevelopmentobjectivebebrokendownintooneorseveralintermediateoutcomes?
Whattypesofactionsshouldbetakentoaddresstheidentifiedbottlenecks?
Whatisthesequenceofactions,andthroughwhichmechanismsaretheseactionsplannedto
solvetheidentifiedbottlenecks?
Whatarethekeymonitoringindicatorsidentified?
Whoisaccountableformonitoring?Whomonitorstherelevantvariablesandinwhatperiodicity?
20
Forthecasewithotherfactorsthanknowledgeasthebindingconstraintandtheimpactevaluationframework,
componentsareadaptedfromtheWorldBankAfricaFinancialandPrivateSectorDevelopmentFrameworkfor
theDesignandImplementationofGrowthPolesProjects
23
Aretheindicatorsderivedfromthedevelopmentobjectives,intermediateoutcomesoractions?Arethey
verifiableandobjective?
How is the institutional setting that may affect the political economy of various components of the
strategyandindividualprojectsputinplaceaspartofthestrategy?
Whoarethekeystakeholders?
Whatistheirinterestintheproject?Arethereimportantofficialsorentitieswithlimitedorno
interestinthesuccessofthestrategy,orinparticularprojectcomponents?Forexample,local
officialsmaybemotivatedbytheincreaseinlocalrevenuesgeneratedbysomeeffectsofthe
policy,withlimitedornointerestin,addressingthebottlenecks.
Arethereimportantofficialsorentitiesthatmightwanttoblockorslowdowntheproject,orany
particularprojectcomponent?Aretherefirmsthatarelikelytobenefitfrommonopoly/oligopoly
rents?
Arethereimportantstakeholdersthatmightwanttocapturetheproject,oranyparticularproject
component?
-ExperimentationandAdaptationStage-
Whatarethefeedbackmechanismsavailableinthegivenpolicymix?
Isthepolicymixtargetingsectorsorinstruments?Orisacombinationofinterventionsputinplace?
Has the pilot policy mix been successful in achieving the intermediate goals determined in the
identificationstage?
Havetheintermediategoalschangedsincethepilotpolicylaunch?(Ifso,revisittheidentificationstage)
Have the monitoring indicators changed since the pilot policy launch? (If so, revisit the identification
stage)
Whataretheexternalfactorsaffectingtheresults?
Aretherisksidentifiedintheidentificationstagelikelytoberealized?Didadditionalrisksemerge?
Arethereanyotherconcernsthatmayinfluencetheimpactsuchascoordinationproblemsandtimingof
interventions,ordataavailability?
Wereimpactevaluationstudiesundertakenfortheexperimentedpolicies?
24
Conclusions
Theprioritizationofselectedeconomicsectorsthroughabroadconsultativeprocessisoftenconsidereda
centralaspectofresearchandinnovationpoliciesaimingatpromotingsmartspecialization(RIS3).Targeted
researchandinnovationpoliciesareseenasanefforttoreducefragmentationandawaytoincreasethe
impactofresearchandinnovationinvestmentsonregionalandnationaldevelopment.Theyalsorelateto
theunderstandingthatthestrategiesofthecurrentProgrammingPerioddidnotachievethedesiredgoals.
Targetingresearchandinnovationpolicies,however,maynotbealwaysrecommendableduetoinherent
problems of incomplete information and the inevitable representation bias towards incumbent interests.
When the economic specialization of a region is not immediately evident from observation of market
dynamics, policy-makers should be encouraged to adopt measures that enable the process of market
selection,allowingsuchspecializationtoemergeasaresultofentry,exitandexperimentation.
Researchandinnovationpoliciescanplayamajorroleinfosteringentranceofknowledge–basedstartups,
innovationandtechnologyadoption–i.e.contributingtoaneco-systemthatnurturesthecreationofhigh
growthpotentialfirms.Otherpoliciesthatfacilitateentryandencouragetechnologicalmodernization–
e.g.promotingcompetition,providingaccesstocredit,skillsandinformation,andfacilitatingmarketexit
–areequallyrelevant.Theiradoptioncanleverage,insomecasesbecomeapre-conditionfor,theimpact
ofresearchandinnovationmeasuresoneconomicdevelopment.Becausesuchpoliciesandprogramsare
hardtodefineupfront,someexperimentation(andfailure)willbeinevitable.
Thepaperalsoexploredcircumstancesinwhichtargetedresearchandinnovationpoliciesmaybeuseful,
namely,wheninformationontheeconomicspecializationoftheregionorthecountryisfullyorpartially
available,throughsomesortof‘markettest’.Insuchcases,researchandinnovationpoliciesmayhelpsound
companies to keep a competitive edge in international markets or to cope with growing international
competition (shaping comparative advantages) or simply support potentially high growth companies to
realizetheirpotential(unleashinglatentcomparativeadvantages).Fordifferentreasons,thesesituations
arelikelytobetheexceptionratherthantherule.
Inourview,theaboveapproachwouldhaveatleasttwoimmediateimplicationsforRIS3worthtakinginto
account:
Replacingtheemphasisonex-antedefinitionofsectorsandfullcommitmentofresourcesup-front
witharesults-basedapproachinwhichsomeflexibilityforpolicy/programexperimentationandex-post
allocationofresourcesbasedonresultsisgranted;
Fully integrating monitoring and evaluating mechanisms in the design and implementation of the
research and innovation strategy, allowing for structured learning and systematic adjustment of
programsandpoliciestowardsthepre-definedobjectives.
These recommendations seem not too far from the views conveyed by the Commission on different
occasions. It is also consistent with an emphasis placed by the Commission, the OECD and other
organizations on the importance of adequate monitoring, evaluation and learning for the adaptation of
researchandinnovationpolicyfordiverseandchangingcircumstances.Asmartspecializationstrategywould
thusbea“living”document,asrecommendedbytheGuidelinesandamanagementtoolforbetteruseof
publicexpendituresinresearchandinnovation,asadvocatedbytheInnovationUnionReport.
25
Annexes
ANNEXI
Dataanalysis:informationonthespecializationoftheeconomy
IdentificationoflatentcomparativeadvantageintheR&D,innovationandtechnologycontexthastwo
components.Thefirstcomponentisanassessmentoftheexistingstrengthsoftheeconomy,whichisan
analysisoftheinstitutionalsetupandtheobservedcomparativeadvantage.Thesecondcomponentinvolves
anattempttoinvestigatethelatentcomparativeadvantage.
Thefirstcomponentimpliesastocktakingexerciseoftheregion’sobservedstrengthsandweaknesses.
Thisinvolvesthereviewofexistingindustrytrendsandofeconomicspecialization21inthecountryorregion
withacomplementaryanalysisofthepotentialdistortionsgeneratedbycurrenteconomicpolicies.Examples
ofdistortionarypoliciesinclude:stateprotectionintheindustry;tariffs,subsidies,technicalorregulatory
barrierstoentry;thedegreeofdominanceofstate-ownedenterprisesintheeconomy;andothermonetary
ornon-monetarysubsidiesdirectlyavailabletostimulatethegrowthofthesector.Benchmarkingtheregion’s
intra-industryinnovativeactivityagainstinternationalcomparatorsisacrucialcomponentofthisstocktaking
exercise.
Thesecondcomponentinvolvesananalysisofthepotentialtounlocklatentcomparativeadvantages.A
surveyofemergingglobaltrendsintechnologyandinnovation,complementedwithanassessmentofthe
region’s potential in advancing in some areas may shed light on the potential avenues for development.
David,ForayandHall(2009)conjecturethatthelocalpotentialwillberelatedtothealreadyexistingsectors
intheregionorcountry.Asanexampletoafollowerstrategy,theysuggestanexaminationofthepatenting
activity by firms in the same sector located in other regions of the world to measure the prospects for
catchingup.
JointworkbytheCommissionandtheOECD22isunderwaytocompileageneralsetofindicatorstoassess
thedegreeofspecializationineachregion.Theprojectconsistsoffourstages:(i)identificationofindicatorbased specialization profiles23, (ii) preparation of templates for strategic governance profiles of countries
involvedintheproject,(iii)constructionofstrategicgovernanceprofiles,and(iv)preparationofoneortwo
casestudiespercountry.
Indicatorsusedtoassesstradecompetitivenessprovidevaluableguidance,asthecapacityofaregionin
terms of innovation and technology is part of the ‘trade promotion infrastructure’. A Trade
Competitiveness Diagnostic (TCD) (Farole & Reis, 2012) comprises three modules: (1) Trade outcomes
21
Economicspecializationcanbemeasuredbytherevealedcomparativeadvantage(RCA).TheRCAforproductk
incountryiistheexportofproductkrelativetothecountry’stotalexportsdividedbytheworldexportof
productkrelativetototalworldexports(FaroleandReis,2012).
22
TIPProjectonSmartSpecialization,DSTI/STP/TIP/AH(2012)1,introducedearlier.
23
Indicatorsidentifiedinstage(i)aredividedintocategories,accordingtothedataused:(a)dataonscientific
research:an‘activityindex’tracingdowntheimportanceofacountry’spublicationsrelativetoitssizeandfield,
andspecializationbasedonaclassificationofscientificdisciplines;(b)dataontechnology,namely,patent-based
measures;(c)economicspecializationindicators.
26
analysis; (2) Competitiveness diagnostics; and (3) Policy options for competitiveness and case studies. To
attainacomprehensiveviewoftheregion’sstrengthsandweaknesses,afullTCDanalysiswouldbehighly
recommended.ATCDmayhelpidentifywhetherthebindingconstraintintheregionistechnology-related
ornot,therebyincreasingtheprospectsforestimatingthepotentialforsmartspecializationtotriggergrowth
intheregion.
Finally, field work involving interviews, surveys, and tools such as value chain and/or product space
analysescanbeundertakentocomplementthedeskworkoutlinedabove.Thedrawbackofinterviewsis
thateachstakeholdermightpresenttheirsubjectiveviewandmayhavepersonalorinstitutionalbenefits
fromsupportingacertainstandpoint.FaroleandReis(2012)presentasetofinterviewtargetsandissuesfor
discussion as part of the qualitative analysis of the ‘trade promotion infrastructure’. By the same token,
surveys need to be carefully designed to lead to an objective assessment. Value chain analyses provide
product-level information with a detailed description of the range of processes from the source of the
producttoitsend-user.
Box4:Valuechainanalysis
Source:FaroleandReis(2012),fromWorldBank(2009),ClustersforCompetitiveness:APracticalGuide&PolicyImplicationsfor
DevelopingCountries,Washington,DC:WorldBank;Gereffi,G.andFrederick,S.(2010),GlobalApparelValueChina,Trade,andtheCrisis:
ChallengesandOpportunitiesforDevelopingCountries,PolicyResearchWorkingPaper5281,WorldBank
27
Box5.Productspaceanalysis
[…]Thehypothesisisthatcountriesthatbuildupcompetenceinproducingacertaingood
canredeploytheirhuman,physicalandinstitutionalcapitalmoreeasilyiftheyseekto
producegoodsthatare“nearby”thosethattheyareproducingalready.Proximitybetween
productsonaproductspaceiscomputedfromthepair-wiselikelihoodthatacountry
exportsaproductgiventhatitalsoexportsanotherproduct.
Proximitybetweenanytwogoods(mandn)istheminimumofthepair-wiseconditional
probabilitiesofhavingcomparativeadvantage.
ϕ m,n = min⎧⎨P⎛⎜ RCAm RCA ⎞⎟, P⎛⎜ RCAn RCA ⎞⎟⎫⎬
⎩ ⎝
n
⎠
⎝
m
⎠⎭ […]Whetheracountry’sexportsinwhichithascomparativeadvantagearelocatedinthe
denserpartoftheproductspaceorintheperipherycanpredicttheeasewithwhichthat
countrytransformsitselfeconomically.Structuraltransformationsarenotsmooth
movementsalongacontinuumbutamessyprocessbesetbymarketfailures.When
suchmarketfailuresarebinding,itisharderforfirmstohoplongerdistanceswithout
governmentcoordinationandsupport.Becauseproductsdonotevolveinsequence,
i.e.,havingironoredepositsdoesnotnecessarilymakeacountryanefficientsteel
producer,lateral-linkagesareasormoreimportantthanforward-linkageswith
downstreamindustries.
Source:FaroleandReis(2012),p.71
28
ANNEXII
ThematicexanteconditionalityfortheuseofStructuralFundprograms
COM(2011)615: Proposal for a REGULATION OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL
laying down common provisions on the European Regional Development Fund, the European Social Fund, the
Cohesion Fund, the European Agricultural Fund for Rural Development and the European Maritime and Fisheries
Fund covered by the Common Strategic Framework and laying down general provisions on the European Regional
Development Fund, the European Social Fund and the Cohesion Fund and repealing Regulation (EC) No 1083/2006,
Annex IV:
First two of eleven ex ante conditionalities:
Thematic objectives
Ex ante conditionality
Criteria for fulfilment
1. Strengthening
research, technological
development and
innovation (R&D
target)
1.1. Research and innovation: The
existence of a national or regional
research and innovation strategy for
smart specialisation in line with the
National Reform Program, to
leverage private research and
innovation expenditure, which
complies with the features of wellperforming national or regional
research and innovation systems.
–A national or regional research and innovation strategy for
smart specialisation is in place that: is based on a SWOT
analysis to concentrate resources on a limited set of research
and innovation priorities; outlines measures to stimulate private
RTD investment; contains a monitoring and review system.
2.1. Digital growth: The existence
within the national or regional
innovation strategy for smart
specialisation of an explicit chapter
for digital growth to stimulate
demand for affordable, good quality
and interoperable ICT-enabled
private and public services and
increase uptake by citizens,
including vulnerable groups,
businesses and public administrations
including cross border initiatives.
A chapter for digital growth within the national or
regional innovation strategy for smart specialisation is in
place that contains:
(referred to in Art.9(1))
2. Enhancing access
to and use and
quality of information
and communication
technologies
(Broadband target)
(referred to in Art.9(2))
–A Member State has adopted a framework outlining available
budgetary resources for research and innovation;
–A Member State has adopted a multi-annual plan for
budgeting and prioritization of investments linked to EU
priorities (European Strategy Forum on Research
Infrastructures -ESFRI).
– budgeting and prioritisation of actions through a SWOT
analysis carried out in alignment with the Scoreboard of the
Digital Agenda for Europe;
– analysis of balancing support for demand and supply of ICT
should have been conducted;
– measurable targets for outcomes of interventions in the field
of digital literacy, skills, e-inclusion, e-accessibility, and ehealth which are aligned with existing relevant sectoral
national or regional strategies.
– assessment of needs to reinforce ICT capacity-building.
2.2. Next Generation Access (NGA)
Infrastructure: The existence of
national NGA Plans which take
account of regional actions in order
to reach the EU high-speed Internet
access targets, focusing on areas
where the market fails to provide an
open infrastructure at an affordable
cost and to an adequate quality in
line with the EU competition and
state aid rules, and provide accessible
services to vulnerable groups.
A national NGA Plan is in place that contains:
– a plan of infrastructure investments through demand
aggregation and a mapping of infrastructure and services,
regularly updated;
– sustainable investment models that enhance competition
and provide access to open, affordable, quality and future
proof infrastructure and services;
– measures to stimulate private investment.
29
30
ANNEXIII
FullDecisionTree:Atypologyforsmartspecializationandpolicyimplications
STEP I: Specialization
STEP II: Bottlenecks
31
Bibliography
Acs,Z.(2012).PublicPoliciesinanEntrepreneurialSociety.In:Dietrich,M.andKrafft,J.HandbookontheEconomics
andTheoryoftheFirm.Chapter36p.515-28.EdwardElgar.
Duranton,G.,&Puga,D.(2004).Micro-foundationsofurbanagglomerationeconomies.InJ.Henderson,&J.Thisse,
HanbookofRegionalandUrbanEconomics(pp.Volume4,Chapter48).Elsevier,B.V.
Bloom,N.,Sadun,R.,&vanReenen,J.(2012).AmericansDoITBetter:USMultinationalsandtheProductivity
Miracle.AmericanEconomicReview,167-201.
David,P.,Foray,D.,&Hall,B.(2009).KnowledgeEconomists'PolicyBriefno.9:"SmartSpecialisation-TheConcept".
David,P.,Foray,D.,&Hall,B.(2011).MeasuringSmartSpecialisation:TheConceptandtheNeedforIndicators.
Farole,T.,&Reis,J.G.(2012).TradeCompetitivenessDiagnosticToolkit.WashingtonDC:TheWorldBank.
Farole,T.,Rodriguez-Pose,A.,&Storper,M.(2009).CohesionPolicyintheEuropeanUnion:Growth,Geography,
Institutions.
Foray,D.,&vanArk,B.(2007).KnowledgeEconomists'PolicyBriefno.1:"Smartspecialisationinatrulyintegrated
researchareaisthekeytoattractingmoreR&DtoEurope".
Guasch,L.(2011).PresentationtitledTechnologyCenters(CITEs)–APublic-PrivatePartnershipforTechnologyand
Innovation:LessonsforRussiainWorldBankSouth-SouthKnowledgeExchangeworkshopRussiaSeriesSession3:
Innovation,CommercializationofResearchandSpin-offSMEs.
Hausmann,R.,&Rodrik,D.(2003).Economicdevelopmentasself-discovery.JournalofDevelopmentEconomics(72),
603-633.
Klinger,B.,&Lederman,D.(2011,March).Exportdiscoveries,diversificationandbarrierstoentry.Economic
Systems,35(1),64-83.
Krueger,A.O.(1974).Thepoliticaleconomyoftherent-seekingsociety.TheAmericanEconomicReview,291-303.
Krueger,A.O.(1997).Tradepolicyandeconomicdevelopment:howwelearn.TheAmericanEconomicReview,1-22.
Krugman,P.R.(1991).Increasingreturnsandeconomicgeography.JournalofPoliticalEconomy,483-99.
McCann,P.,&Ortega-Argiles,R.(2011).SmartSpecialisation,RegionalGrowthandApplicationstoEUCohesion
Policy.Groningen:EconomicGeographyWorkingPaper,FacultyofSpatialSciences,UniversityofGroningen.
OECD(2012).DraftSynthesisReportonInnovation-drivenGrowthinRegions:TheRoleofSmartSpecialisation,
versionDecember2012,Paris.
Puga,D.(2002).Europeanregionalpoliciesinlightofrecentlocationtheories.JournalofEconomicGeography,2,
373-406.
Rodrik,D.(2004).IndustrialPolicyfortheTwenty-FirstCentury.CentreforEconomicPolicyResearchDiscussion
PaperSeriesno.4767.
SmartSpecialisationPlatform,(2012).RIS3Guide,versionMay2012.
32